Business Innovation Policies ||

www.oecd.org/publishing

Please cite this publication as:

OECD (2011), Business Innovation Policies: Selected Country Comparisons, OECD Publishing. http://dx.doi.org/10.1787/9789264115668-en

This work is published on the OECD iLibrary, which gathers all OECD books, periodicals and statistical databases. Visit www.oecd-ilibrary.org, and do not hesitate to contact us for more information.

-:HSTCQE=VVZ[ZV:ISBN 978-92-64-11565-192 2011 08 1 P

Business Innovation PoliciesSELECTED COUNTRY COMPARISONS

This study is concerned with trends in and key features of policies and programmes used by governments to support innovation in the business sector. In addition to identifying good practices across a range of programme types, it compares business innovation policies across several countries, with a particular focus on Canada.

Contents

Chapter 1. Business innovation policies: Findings and policy conclusions

Chapter 2. Public support for business R&D

Chapter 3. Non R&D-based public support for business innovation

Chapter 4. Demand-side policies to support innovation: Trends and challenges

Chapter 5. Evaluating public support for innovation in business: Methodologies and metrics

Business Inno

vation P

olicies S

ELE

CT

ED

CO

UN

TR

Y C

OM

PAR

ISO

NS


Business Innovation Policies

SELECTED COUNTRY COMPARISONS

This work is published on the responsibility of the Secretary-General of the OECD.The opinions expressed and arguments employed herein do not necessarily reflectthe official views of the Organisation or of the governments of its member countries.

This document and any map included herein are without prejudice to the status ofor sovereignty over any territory, to the delimitation of international frontiers andboundaries and to the name of any territory, city or area.

ISBN 978-92-64-11565-1 (print)ISBN 978-92-64-11566-8 (PDF)

Photo credits: Cover © Veer/Fancy Photography.

Corrigenda to OECD publications may be found on line at: www.oecd.org/publishing/corrigenda.

© OECD 2011

You can copy, download or print OECD content for your own use, and you can include excerpts from OECDpublications, databases and multimedia products in your own documents, presentations, blogs, websites andteaching materials, provided that suitable acknowledgment of OECD as source and copyright owner is given.All requests for public or commercial use and translation rights should be submitted to [email protected] for permission to photocopy portions of this material for public or commercial use shall be addresseddirectly to the Copyright Clearance Center (CCC) at [email protected] or the Centre français d’exploitation dudroit de copie (CFC) at [email protected].

Please cite this publication as:OECD (2011), Business Innovation Policies: Selected Country Comparisons, OECD Publishing.http://dx.doi.org/10.1787/9789264115668-en

FOREWORD – 3

BUSINESS INNOVATION POLICIES: SELECTED COUNTRY COMPARISONS – © OECD 2011

Foreword

This study is concerned with trends in and key features of policies and programmes used by governments to support innovation in the business sector. The study has been prepared at the request of, and with financial support from, Industry Canada.

In addition to identifying good practices across a range of programme types, the study compares business innovation policies across Australia, Canada, Denmark, Finland, Netherlands, Norway and the United Kingdom. The comparisons draw on publicly available material and on submissions of information received from countries to September 2010. Changes in policies or programmes occurring after that date have not been reflected in the text.

The themes examined in the study include: Key trends in business innovation policies. Policies to encourage business spending on research and develop-ment, and the respective merits and demerits of direct and indirect support measures, such as grants and tax credits. The rationales for – and good design features of – a range of additional policies to encourage and facilitate innovation in business. These policies include: financial support for innovative start-ups; innovation hubs and centres of excellence; technology-related networks; and technology incubators. A review and assessment of policies to stimulate demand for innovation and facilitate commercialisation of innovative ideas, whether through innovation-oriented regulations, standards or public procurement. Methodological approaches and metrics to evaluate policies to foster innovation in business.

The study also makes observations on issues of policy mix. In this, and in a number of other sections, there is a particular focus on conditions in Canada.

This publication was written by Alistair Nolan. Essential inputs were provided by Fabienne Cerri, particularly as regards the chapter on demand-side policies, and Dirk Pilat. Research support was had throughout from Linda Haie-Fayle. All of the above-mentioned staff belong to the Structural Policy Division of the OECD Directorate for Science, Technology and Industry.

TABLE OF CONTENTS – 5


Table of contents

Executive summary ................................................................................................... 9

Chapter 1. Business innovation policies: Findings and policy conclusions ........ 13

Key trends in business innovation policies ............................................................ 14The changing rationale for policy .......................................................................... 16Key policy instruments supporting business innovation ....................................... 17Considerations on the policy mix .......................................................................... 21Strengthening evaluation ....................................................................................... 23Canada’s policy mix for innovation ....................................................................... 25

Notes .......................................................................................................................... 26

Chapter 2. Public support for business R&D ....................................................... 29Direct support for R&D ......................................................................................... 32R&D tax credits ..................................................................................................... 37

Notes .......................................................................................................................... 46References ................................................................................................................. 47

Chapter 3. Non R&D-based public support for business innovation ................. 49Policy rationales ..................................................................................................... 50Allocating direct support ....................................................................................... 53Examples of financial support for start-ups, technology-oriented and fast-growth businesses ........................................................................................... 54Examples of support for researchers and innovative entrepreneurs to start their own business .................................................................................................. 61Support to knowledge transfer through technology-related collaboration and networks ................................................................................................................. 64Support for the creation of innovation hubs and centres of excellence ................. 75Examples of technology incubators ....................................................................... 78

Notes .......................................................................................................................... 80References ................................................................................................................. 82

6 – TABLE OF CONTENTS


Chapter 4. Demand-side policies to support innovation: Trends and challenges ................................................................................................................. 83

Evidence and policy trends .................................................................................... 84Public procurement ................................................................................................ 87Innovation-oriented regulations and standards ...................................................... 98Lead markets ........................................................................................................ 106Conclusions .......................................................................................................... 106

Notes ........................................................................................................................ 109References ............................................................................................................... 111

Chapter 5. Evaluating public support for innovation in business: Methodologies and metrics ................................................................................... 115

Introduction .......................................................................................................... 116Standard methodological challenges to policy evaluation ................................... 118Steps in undertaking an evaluation ...................................................................... 123Examples of evaluations of different forms of innovation support ..................... 129Evaluating demand-side programmes and policies .............................................. 140Evaluation of policy systems ............................................................................... 145Selected policy lessons ........................................................................................ 146

Notes ........................................................................................................................ 150References ............................................................................................................... 152

Boxes

Box 2.1. Design considerations: Fiscal support for business R&D ..................... 40Box 3.1. Programmes fostering partnerships for innovation under the

Canadian National Sciences and Engineering Research Council ......... 74Box 4.1. General action plans for demand-driven innovation: The cases

of Finland and the United Kingdom ..................................................... 84Box 4.2. The case of biometrics standardisation in the United Kingdom ......... 104Box 4.3. The European Union Lead Market Initiative ...................................... 105Box 5.1. Policy evaluation and monitoring: Examples of recent

initiatives ............................................................................................. 117

TABLE OF CONTENTS – 7


Figures

Figure 2.1 Direct government funding and tax incentives for business R&D, 2007 ............................................................................................ 30

Figure 2.2. Firms receiving public support for innovation by size, 2004-06 .......... 31Figure 2.3. Direct government funding of business R&D, as % of total

BERD (2000 and 2008) ........................................................................ 33Figure 2.4. Tax subsidy rate in 2008 for USD 1 of R&D (large firms and

SMEs) ................................................................................................... 41Figure 2.5. Change in the tax subsidy rate for USD 1 of R&D (large firms

and SMEs, between 1999 and 2008)..................................................... 41Figure 5.1. Logic model for an R&D tax credit policy ......................................... 123

Tables

Table 2.1. Overview of R&D tax incentives (2009 and 2010) .............................. 39Table 3.1. Overview of innovation voucher schemes in Europe ........................... 67Table 4.1. Level of priority accorded to demand-side policies .............................. 86Table 4.2. Key features of demand-side policy instruments .................................. 86Table 4.3. Procurement types and possible effects of public sector

interventions on innovation .................................................................. 88

EXECUTIVE SUMMARY – 9


Executive summary

This study addresses trends in and key features of public policies and programmes to support innovation in business. It was prepared at the request of, and with financial support from, Industry Canada.

Chapter 1 describes the context for the study, and reflects on some of the main findings as they relate to policy challenges faced by Canada and other countries. Key trends in business innovation policies and their rationales are summarised. Observations are also made on the merits and demerits of direct and indirect support measures, such as grants and R&D tax credits respectively. Consideration is also given to the overall policy mix – the combination of policy instruments deployed to achieve policy goals. Among other points, reference is made here to the importance of non-scientific knowledge in the creation of economic value from invention, and the contention that in some countries the importance of actors that do not produce patents or publications might not be given appropriate recognition. A high-quality system of policy evaluation should ultimately inform the choice of policy mix. Chapter 1 also notes that Canada’s current policy mix heavily emphasises certain instruments, notably indirect support for R&D, through the tax credit system, as well as direct support for some aspects of innovation, including venture capital investment and commercialisation. Canada currently places only a limited emphasis on demand-side measures. Some possible limitations to this mix are briefly discussed.

Chapter 2 addresses public support for business R&D. OECD countries provide a range of measures to support private investment in R&D. Such support aims to address the inability of firms to fully capture the returns to their investment in R&D, which may lead to too little private investment in R&D. This chapter describes both direct support instruments (e.g. direct procurement, grants and subsidies for R&D) and indirect instruments (e.g. R&D tax allowances and credits) used by the seven countries examined. Observations are made on the relative merits and demerits of direct and indirect support, as well as key features of programme design. Programme examples are cited from the countries studied. Countries’ overall mix of policies, notably the balance between direct and indirect support, is also discussed. It is seen that the mix of policies used to support private investment in R&D and innovation differs widely across countries. For instance, some

10 – EXECUTIVE SUMMARY


countries such as Sweden, Finland and Germany provide no support to private R&D through the tax system, and only provide direct support, in the form of grants or other mechanisms. Other OECD countries, such as Canada, Japan and the Netherlands provide relatively limited direct support for private R&D, but have tax credits for R&D that account for the bulk of support. Yet other countries, such as Austria, Belgium and Korea, allocate significant resources for both direct and indirect support of private R&D. The optimal balance of direct and indirect R&D support will vary from country to country, as each tool addresses different market failures and stimulates different types of R&D.

Chapter 3 reviews forms of public support for business innovation that do not target R&D. Non R&D-based innovation support is of growing importance. Such support includes: measures to facilitate access to early-stage finance, particularly for equity; initiatives to assist the commercialisation of innovation; support for the development of networks of different sorts; programmes to promote regional innovation hubs; and a range of initiatives – such as voucher programmes – to ease access to information, expertise and advice. Schemes of this sort are obviously more targeted than R&D tax credits. Careful consideration of programme logic is thus required in their creation. Among other things, policy makers must ensure that such programmes do not override or distort market mechanisms. The text considers the economic rationale for this class of policies and notes that governments sometimes make only passing or incomplete reference to the economic rationale for particular initiatives. Furthermore, some of the theoretical justifications for policy lack robust empirical undepinnings. The text likewise points to ways in which governments can harness market mechanisms in the design of direct support, such that the information contained in markets is brought to bear, while costs to the public sector are contained. This chapter also describes a selection of such programmes in the countries under review.

Chapter 4 reviews and assesses policies to stimulate demand for innovation. It is seen that, from the United Kingdom to Finland, OECD countries and the European Commission have made explicit statements on the need to give greater emphasis to demand-side innovation policies (although, in practice, the role of demand-side initiatives in the full portfolio of government policies on innovation remains relatively marginal). The text describes a number of reasons for the growing interest in demand-side policies.

The policy instruments reviewed in Chapter 4 include innovation-oriented public procurement, technology-oriented regulations, and product standards. These instruments all entail different rationales. For instance, support for the development of standards – which reflects public-good characteristics possessed by standards – has a different economic justification to innovation-oriented public procurement which, among other justifications, can help accelerate the emergence of technologies for which there is an urgent time-

EXECUTIVE SUMMARY – 11


bound societal need, as well as helping to close gaps in the supply of risk finance for small early-stage ventures. To be efficient, demand-side policies should clearly target specific market or system failures.

The text reviews the (limited) evaluative evidence on good practice in the design of demand-side approaches and the advantages and disadvantages of their use. Each of the key demand-side measures is seen to pose specific implementation challenges. Errors in the setting of key regulations, for instance, could have far-reaching economic consequences. Demand-side policies may also require the development of new capabilities and capacities in the public sector. To give just two examples: industry-specific expertise will be needed when setting innovation-oriented regulations. And, as regards public procurement, challenges can arise in terms of governance, co-ordination and strategic planning, because in many countries numerous sub-national units of government play important roles in the public procurement market. Chapter 4 describes various examples of different countries’ demand-oriented strategies and approaches.

Chapter 5 considers conceptual, practical and institutional issues in the evaluation of direct and indirect support for private sector research and innovation. Effective evaluation of policies and programmes to stimulate research, development and innovation has become increasingly important for policy makers. This increased emphasis on evaluation is driven by constraints on discretionary public spending, a greater focus on accountability and transparency in policy, and the desire to minimise distortions arising from government actions while maximising their impact. This chapter focuses on the assessment of policy and programme impacts. It also draws a clear distinction between evaluations of impact and audits and other forms of process control. The emphasis is placed on methodological and practical challenges and opportunities facing evaluators, rather than the content of findings from evaluation studies. Challenges and possible approaches to the evaluation of demand-side innovation policy are also examined. The limita-tions of ex ante evaluation are briefly reviewed, as are the shortcomings of using performance standards as measures of impact. Examples of evaluation studies are described – along with a consideration of their strengths and weaknesses – for the following programme types: research in universities and public research organisations; collaborative R&D; R&D tax credits; access to finance; technical extension programmes; access to research-related informa-tion or expertise; technology incubation; technology-oriented regulations; technical standards; technology-oriented public procurement; and SBIR-type programmes. The chapter concludes with a summary of generic lessons learned on the practice of policy and programme evaluation.

1. BUSINESS INNOVATION POLICIES: FINDINGS AND POLICY CONCLUSIONS – 13


Chapter 1

Business innovation policies: Findings and policy conclusions

This chapter describes the context for the study, and reflects on some of the main findings as they relate to policy challenges faced by Canada and other countries. Key trends in business innovation policies and their rationales are summarised. Observations are also made on the merits and demerits of direct and indirect support measures, such as grants and R&D tax credits respec-tively. Consideration is also given to the overall policy mix – the combination of policy instruments deployed to achieve policy goals.

14 – 1. BUSINESS INNOVATION POLICIES: FINDINGS AND POLICY CONCLUSIONS


Key trends in business innovation policies

All OECD countries have policies in place to strengthen their innovation performance. These policies are wide-ranging and include framework policies, specific policies to foster private investment in innovation, and policies that aim to shape demand for innovative goods and services. Typically, the mix of policies differs considerably between countries, due to structural and institutional differences, among other reasons. Among OECD countries as a whole, as well as the seven countries examined here, recent years have seen some shifts in the mix of policies used to support business innovation. These shifts include:

A growing attention to demand-side policies. Growing attention is being paid to a range of policies aimed at strengthening demand for innovation. This includes policies such as innovation-oriented public procurement, standards and regulation. The growing interest in demand-side policies has emerged in part because of greater aware-ness of the importance of feed-back linkages in the innovation process between supply and demand. Demand-oriented innovation policies are thus part of an evolution from a linear model of policy, usually focused on R&D, to a more broad-based approach that considers the full innovation cycle. Indeed, many countries have noted that a main challenge for innovation is often not the lack of knowledge or technology, but linking these innovations to a market. This is particularly the case in markets with major public goods characteristics, e.g. markets for environmental goods and services, certain health services, and other public and semi-public services. Recognition of the essential interaction between demand and supply conditions is also reflected in the broader academic literature.1 The focus on demand-side policies also reflects a frequent perception that traditional supply-side policies – despite many refinements in their design over past decades – have not been sufficient to bring about the desired levels of improvement in innovation performance and productivity. Furthermore, current pressures on governments’ discretionary spending create incentives to explore how innovation might be fostered without new programme spending. Several of the instruments used are relatively new and little is known thus far about their cost-effectiveness and impact. Moreover, the scale of these demand-side policies is typically still quite small and not considered a high priority in most countries.



A growing use of indirect support for innovation, notably a growing use of R&D tax credits and an increase in their generosity. At the time of writing, 22 OECD member countries provided some form of R&D tax credit to foster private investment in innovation, up from 12 countries in 1995 and 18 in 2004. Moreover – as shown in Chapter 2 – several countries have enhanced the generosity of their system in recent years, as measured by changes in the so-called B-index. Many countries have also made changes to their R&D tax credit schemes with the goal of expanding the number of beneficiaries. Some have enlarged the criteria of eligibility, or the degree of coverage for firms eligible for tax relief. The growing importance of R&D tax credits partly reflects that this instrument is less likely to distort private decisions and market behaviour than more targeted, direct support. At the same time, there is some concern that the growing generosity of these measures may lead to tax competition among countries.

Shifting emphases in direct support for innovation. As indirect support has grown, so direct support for R&D has declined, at least that part of direct support linked to direct procurement by govern-ments, e.g. for defence. At the same time, as described in Chapter 3, direct support is increasingly provided for different purposes, e.g. to encourage collaboration and knowledge transfer between firms or between firms and scientific institutions, to foster the growth of high-technology start-up firms, to encourage the development of venture capital activity, and to support innovation relevant to climate change and the environment. Direct support therefore remains an important policy instrument for OECD countries, as it allows governments to target specific barriers that affect innovation performance such as lack of different forms of co-operation, or specific areas considered to have high social returns.

Growing attention given to policy evaluation. The evaluation of policies is clearly essential to enhance their impact and efficiency. In recent years, and particularly following the financial crisis and the squeeze on discretionary public spending, emphasis has grown on evaluating policies to measure their impact and efficiency. However, the evaluation record is still patchy. While considerable effort has been invested in the evaluation of some forms of direct and indirect support for R&D, particularly as regards the additionality entailed in any changes in private R&D spending, more remains to be done. There has also been relative underinvestment in the evaluation of other programmes used to increase innovation. In all cases, as discussed in Chapter 5, the methodological challenges of reliably attributing



impact to government programmes are significant, and ongoing efforts are needed to meet the evaluation challenge.

The changing rationale for policy2

In considering what policies to undertake, governments need to consider in some detail the rationale for their policy interventions. The various instru-ments to foster business innovation do not all respond to the same policy rationale.3 Direct and indirect government support for R&D aims to address a range of market and system failures. These include well-documented external effects related to the inability of firms to fully capture the returns to their investment in R&D, which can lead to a level of private sector invest-ment that is socially suboptimal. It also includes system failures, such as: lack of networks in which to innovate; incomplete markets, particularly as regards the provision of small volumes of risk capital for high-tech ventures; as well as market power and scale effects, which may make it difficult for small or new firms to enter markets.

Demand-side policies aim to address other perceived market failures (Edler, 2007), namely problems related to market introduction and the diffusion of innovations, e.g. information asymmetries (producers do not know the preferences; users do not know the innovations); lack of interaction between producers and users; high switching cost to new technologies, high entry costs (especially in areas with important network effects); and techno-logical path dependencies.

Even before the recent financial crisis, efficiency considerations were becoming increasingly important for governments in evaluating their policy mix to support business innovation. This is potentially another reason for the growing interest in demand-side policies. Several of these policies, e.g. standards and regulations, and to a lesser extent innovation-oriented public procurement, involve only a limited budgetary allocation and may be effective in fostering innovation if they succeed in strengthening market demand for innovative goods and services. Thus far, however, little is known about their impact or effectiveness.

Not all potential market or system failures make government intervention required or desirable. There is no guarantee that government policy will be able to address each market or systemic failure in a way that effectively improves outcomes. Even in cases where government could potentially improve welfare, it does not always have the means to achieve such a result in practice. Governments’ sphere of action may be limited, and informational constraints can limit its capabilities to intervene effectively. Indeed, “policy failures” are often due to the fact that the government is subject to similar and sometimes even more stringent informational constraints than private actors.4



There is also a risk that government interventions can be counterproductive, if not carefully designed. For these reasons, the soundness of the foundations and the achievements of government intervention need to be scrutinised ex ante and ex post. Transparency, built-in feedback (e.g. through competitive mechanisms or, in their absence, evaluation) and associated learning processes may help keep policy on track and avoid wasteful activities.

Key policy instruments supporting business innovation

The case for direct support Direct support for R&D and innovation has the advantage that it can be

focused on activities and actors of greatest interest in meeting public policy goals and that may yield the highest social returns. A case in point is environmental amelioration and pollution abatement. Many governments have made time-bound commitments on reducing carbon emissions. The search to develop, diffuse and apply low-emission technologies is taking place within defined time-horizons. Direct support can help focus innovation efforts to meet such timeframes, in ways that generic instruments might not.

Direct support can also be focused on specific barriers to innovation. These barriers are varied, and can include: the possible undersupply of private investment in R&D; the failures of market actors to supply public goods (such as open, credible international technical standards); gaps in the market for early-stage equity finance; possible under-provision of business development services, especially for small firms and for businesses located in remote or lagging regions; co-ordination and information problems that hinder networking or other types of collaborative activity; and difficulties faced by some SMEs in accessing public procurement markets. Several governments also target resources to attracting multinational firms that tend to be more internationally mobile.

Focusing support on activities that either have high positive externalities or are prone to market or systemic failures can increase the impact of public support. However, because direct support can be focused, the problematic prospect arises of governments “picking winners”. The design of schemes to provide direct support for R&D and innovation is therefore of great importance. Good policy designs need to ensure competitive selection processes, select projects that best serve public policy objectives, avoid favouring incumbents or providing opportunities for lobbying, ensure a rigorous evaluation of policy impact, and contain costs. As focused direct support puts greater demands on government capabilities to shape programmes, direct support tends to be more expensive than indirect support in terms of the costs of executing the programme and administering the selection process.



Governments can take various steps to harness market mechanisms in the design of direct support, such that the information contained in markets is brought to bear while costs are contained. Chapter 3 for instance describes work by Murray (1999) that assesses the best designs for public support of early-stage venture capital funds (concluding that the best option is to provide public co-investment with private partners, which multiplies the financial benefits of success to the disproportionate advantage of venture capital funds and their investors while maintaining incentives for fund managers to make good investment decisions). This example illustrates how well-designed public support can operate to amplify and target market dynamics.

And governments can also introduce design features in R&D support procedures that could increase the efficiency of allocations. Giebe et al.(2005) for instance describe the resource savings and efficiency benefits that could follow the introduction of competition among applicants for R&D grants – through the use of various auction mechanisms – greater information can be extracted on the proposals and some degree of unnecessary funding can be avoided. The increasingly common practice of giving direct support to pre-competitive ventures, and to partnerships, is one means of reducing problems associated with picking winners at the level of individual firms.

Several OECD countries also have generic policies that provide direct support. For example, the innovation vouchers used in the Netherlands are a generic tool to provide direct support. Such programmes tend to have relatively low information and administrative costs, and also face a smaller risk of government failure. On the other hand, there is a greater risk of large deadweight loss, with support possibly going to firms that would have undertaken innovation efforts even without public support.

The case for R&D tax credits Traditionally, the argument for tax incentives lies in their non-

discriminatory nature and ease of use. The choice of R&D tax incentives will depend on country-level variables such as overall innovation performance, market failures in R&D, industrial structure, size of firms and the nature of corporate tax systems. R&D tax credits are neutral with respect to the type of R&D being conducted by a firm, and therefore operate more in accordance with market rationale than direct support. At the same time, as R&D tax credits tend to support any formal R&D, they are less easy to steer and may therefore be less effective in achieving public policy goals or in achieving high social returns.



R&D tax credits have become one of the most widely used instruments of innovation policy. It has been argued that R&D tax credits present many advantages since they may be relevant for all industrial sectors or research and technological fields. Tax concessions interfere little in firms’ R&D strategy and let market mechanisms determine R&D priorities. These instruments allow governments both to stimulate investments in R&D and to retain or attract foreign R&D investments. Spillovers may go well beyond the R&D performer, the firm or the industry. In addition, in crisis times tax concessions for R&D could serve as an element of economic stimulus when the relief provided helps companies improve their cash flow (OECD, 2009). Recent trends indicate that countries tend to use both indirect and direct financial support, reflecting their complementarity.

The role of demand-side policies Although demand-side policies have received increased attention in

recent years, their role in the full portfolio of government policies on innovation remains relatively marginal. The rationale for these policies is linked in general terms to the need to stimulate innovation in areas where societal needs are pressing (e.g. health, environment) and where government action can complement market mechanisms with, ideally, minimal financial outlays. But specific rationales pertain to individual demand-side instru-ments. For example, innovation-oriented public procurement can be designed to help counter gaps in the supply of risk finance for small early-stage ventures. Such procurement can also be structured to help offset biases against SMEs in public procurement. And procurement processes can help accelerate the emergence of technologies for which there is an urgent time-bound societal need. To be efficient, demand-side policies should clearly target specific market or systemic failures and their objectives, and impacts should be carefully evaluated.

Implementation of each of the key demand-side policies poses challenges. Errors in the setting of key regulations for instance could have far-reaching economic consequences. This is complicated by the fact that the effects of economic regulation on innovation – and the timing of these effects – can be complex and ambiguous a priori. Moreover, the effects of regulation on innovation are likely to be highly technology and industry specific, which implies a need for significant industry-specific intelligence among policy makers. Such intelligence also relates to the need to assess whether the market would introduce the right level technology in the absence of the regulation (the case of automotive engine fuel efficiency is described later in the text, which illustrates the complexity of determining the degree of efficiency in that specific market). Indeed, all demand-side policies should



consider whether the action undertaken is efficient from a market per-spective and whether it improves social welfare.

The text likewise highlights that the precise form that regulation takes will also shape its impact on innovation. The example is given of environ-mental regulations, showing that consideration should be given to such policy design features such as stringency, predictability, flexibility, incidence and depth. A further critical consideration is that even in cases where regulation spurs innovation, regulation-based policy might be cost-ineffective overall. Other less costly means might be employed to achieve the same goal, and conflicting regulatory goals might also exist. This underscores the importance of performing cost benefit analysis on key regulatory decisions. Indeed, innovation is usually not among the key objectives of regulators. The policy focus has typically been on avoiding regulatory burdens, rather than on the targeted use of regulation to allow the emergence of new technologies.

Demand-side policies also imply a lead role for the public sector, where new capacities may need to be developed. For instance, as regards public procurement, despite considerable policy interest, the traditional focus on value-for-money as well as the problem of fragmentation of public demand (often between different levels of government) can limit potential scale effects for innovation-oriented procurement. In many countries numerous sub-national units of government play important roles in the public procurement market. This in turn creates challenges in terms of governance, co-ordination and strategic planning. Furthermore, many agencies with responsibilities for public procurement operate separately from line ministries or government agencies with a remit to foster innovation. Technical expertise in the respective fields of innovation may be lacking in the procurement bodies. As many firms do not see public procurement as a relevant source of business, the scope of policy can also be limited. Public sector capabilities may also need to be enhanced in order to mitigate risks associated with procurement of innovation, such as technological risks, organisational and societal risks, and specific market risks. Furthermore, innovation goals must be balanced against the need for competition, transparency and accounta-bility in public procurement. OECD countries need to adhere to national competition and public procurement rules as well as related international standards and obligations (e.g. the WTO Government Procurement Agree-ment). Large player dominance is also possible and in particular occurs in areas where the potential for learning by doing is high. Governments should take measures to ensure this does not occur, sourcing competitively from different firms to prevent dominance.



The public sector’s role with respect to standards largely involves measures to include under-represented groups in the process of developing standards, and subsidisation of teams drafting international standards. Unlike regulation, the setting of standards is mainly the responsibility of industry bodies – with government acting as facilitator or co-ordinator. However, procedures in standards bodies can be slow and bureaucratic and can be held up by large players, which raises the important issue of timing. If standardisation is brought into effect too early it could shut out better technologies. But if standardisation occurs too late then the costs of transition to the new standard could be high enough to slow or prevent diffusion. If product life-cycles are shortening, the issue of timing is likely to increase in importance. Policy makers might also consider monitoring the number and age distribution of the national stock of standards, with a focus on standards developed by national standards bodies (which are open and carry additional credibility). Policy action could be called for if the number, rate of creation and age distribution of the stock of public standards were seen to diverge greatly from norms in other advanced economies. Government also plays a role in shaping the behaviour of consumers. And consumer policy regimes and consumer education can play a role in providing an arena for innovation in meeting new market demands.

Considerations on the policy mix

The term policy mix refers to the combination of policy instruments deployed to achieve policy goals. Whether by design or default, all countries have a policy mix. The seven countries examined in this report, and OECD countries more broadly, make different choices in their mix of policy instruments to support innovation, although a trend can be observed towards greater use of R&D tax credits and a growing generosity in these credits. In most OECD countries, direct and indirect support measures are used, as countries provide generic incentives to strengthen private investment in R&D, and more targeted measures to support specific goals or steer innovation towards given sectors, technologies or groups of firms.

The combination of demand-side and supply-side policies is an important consideration for the policy mix. Neither supply-side nor demand-side policies are likely to be effective in isolation. Fostering innovation requires addressing the entire innovation chain. As noted above and in Chapter 4, one reason for the growing interest in demand-side policies is their potentially low cost and their possible role in helping select “winners” on the basis of better functioning markets.



Even when countries have similar policy goals, the respective instrument mixes can be expected to differ as these mixes need to be adapted to the specific environments in which they are intended to work. These environ-ments vary in terms of the structure of the productive base, institutions and preferences. For instance, a strong preference for a simple, transparent tax system may rule out tax incentives for R&D. Different countries exhibit different degrees of acceptance of regulation. And, as Chapter 4 shows, the efficacy of various demand-side instruments can be highly sensitive to industry-specific characteristics. The key challenge for policy makers is therefore to strike a balance between the various instruments. For example, direct and indirect support for R&D may be applied as complementary tools, making best use of their respective advantages and recognising their interdependencies.

Another balance concerns the number of policy instruments deployed. The trade-off involved here is on the one hand to have a set of instruments that is sufficiently differentiated to meet the needs of complex innovations systems. On the other hand, the policy mix needs to avoid inefficiencies arising from operating too many schemes at too small a scale. This is a real concern, since instruments can develop constituencies of support and a degree of autonomy, making them less amenable to change or cancellation, even where this would be sensible. In some cases, there may be ways to streamline the range of instruments and programmes, reduce complexity, enhance transparency and lower administrative costs.

Another challenge concerns the effectiveness and efficiency of support. To the extent that evidence on the additionality of specific instruments is known, governments can of course decide to focus on instruments that have the highest cost-benefit ratios. Following the financial crisis and in a context of scarce public resources, governments are increasingly trying to focus their efforts to areas where there is good evidence that public money has high returns.

Consideration of strategic direction raises the complex but critical issue of the emphasis of support. Should government prioritise support for investment in cutting-edge R&D, or should priority be given to other dimensions of the innovation system that relate to application and diffusion? A number of points might be made in this connection.

First, using reasoning that also applies to the countries covered in this study, Bhide (2008) argues against the prioritisation of policy support for R&D in the United States. He notes that R&D support tends to be captured by industry. However, services account for the major share of GDP in most OECD economies. At the same time, much innovation in services takes non-technical incremental forms that are less reliant on R&D. And because many



services are not traded, the incorporation of innovations benefits domestic employees and consumers (the advent of electronic health records for instance improves the productivity of health workers and the care given to patients).

In addition, Bhide – and other authors – argues persuasively for the importance of non-scientific knowledge in the creation of economic value from invention. For instance, much of the economic value of inventions comes through their incorporation into existing products (such as transistors into radios). This requires the solution of engineering problems, the creation of new designs, and systems for deciding how to price, market and distribute the new product, among other inputs. A stellar illustration of this is the example of the iPod, the success of which lies largely in design rather than new technology. Indeed, as a company, Apple’s success in large part reflects its integration skills, bringing together knowledge products from firms in many countries (such as, with the iPod, hard-drives from Toshiba; soundcompression technology from Germany; the CPU from England; audio codecs from Scotland; flashdrives in the iPod nano from Japan and Korea, etc.). Overall, Bhide holds, in the translation between invention and living standards, the contribution of actors that do not produce patents or publica-tions is not given appropriate recognition.5

Bearing in mind the above arguments, it is nevertheless difficult to conceive a metric that might be used to establish a socially optimal, or even preferred, allocation of public resources to support innovation (except perhaps in cases of extreme under- or overspending on different forms of public support). But a high-quality system of policy evaluation could ultimately inform the choice of policy mix.6

Strengthening evaluation

Effective evaluation of policies and programmes to stimulate R&D and innovation has become increasingly important for policy makers. This increased emphasis on evaluation is driven by constraints on discretionary public spending, a greater focus on accountability and transparency in policy, and the desire to minimise distortions arising from government actions while maximising their impact.

Support for R&D and innovation takes many forms. Consequently, diverse methodological challenges exist for evaluators. For instance, a characteristic of many schemes to support collaborative R&D is that they involve multiple objectives which can be wide-ranging and difficult to measure (such as the disposition to collaborate). The possibility of evaluating R&D tax credits by constructing a control group using randomisation is nullified by the fact that firms cannot legally be excluded from a tax



incentive to which they are entitled. And the fact that regulation could affect all firms in a regulated sector raises issues of how to compare the effects of the regulation against some alternative benchmark. Some of the programmes considered in this report have a rich evaluation record, whereas others, such as some of the demand-side policies, are under-evaluated.

Evaluations of pro-innovation programmes – from R&D subsidies and tax credits to manufacturing extension and advisory services – ultimately aim to attribute changes in a target group (usually firms, but also institutions and sometimes individuals) to the effect of a given programme. But simply comparing the situation of target groups before and after a programme is insufficient, because the programme may be only one among a number of causes of any observed change of state.

Evaluations essentially take three generic forms: experiments; quasi-experiments; and participant opinion/expert review (structural econometric modelling has also been used to evaluate R&D tax credits). Each method has strengths and weaknesses in being able to identify unrelated effects, minimise selection bias, and accurately identify programme impact. The appropriateness of each method should be fully appreciated before pro-ceeding with an evaluation, and in some cases will require in-depth understanding of programme data and econometric method. Multi-method approaches – combining case studies and the various forms of econometric research – are often required.

Deciding when to perform an evaluation is also important. Different programmes have different gestation periods. Some initiatives offer the possibility of almost immediate benefits, while others require months or even years before change is evident. The impact of some programmes might be significant initially, but decline over time.

A strategic approach to evaluation could be beneficial. A small number of long-term well-designed evaluations and sectoral studies using controlled samples of beneficiaries and non-beneficiaries is likely to yield more generalisable policy-relevant results (with lower response burdens on bene-ficiaries) than multiple evaluations of varied quality employing disparate methods. All significant programmes should be required to develop a formal evaluation strategy. Preparing for and undertaking evaluation is itself a pedagogical exercise that has the potential to improve thinking about programme logic. And planning evaluation before beginning a programme will allow data strategies to be put in place. Placing evaluation data in the public realm (as is typical in Norway for instance), possibly in anonymised form, could also facilitate (cost-free) academic analysis and serve as a form of evaluation quality control (even if academic work need not automatically be of the highest quality). Furthermore, in the pluralistic policy systems that



characterise OECD countries, evaluation should be expected to serve the knowledge needs of a wider set of actors beyond programme managers and public policy officials.

Canada’s policy mix for innovation

Canada’s current policy mix heavily emphasises certain instruments, notably indirect support for R&D, through the tax credit system, as well as direct support for some aspects of innovation, including venture capital investment and commercialisation. Canada currently places only a very limited emphasis on demand-side measures in its overall policy mix.

These policy settings lead to a number of considerations. First, by relying primarily on the R&D tax credit to foster business investment in innovation, all actors and activities engaged in R&D (and meeting the relevant criteria) can benefit from government support. That is, this key instrument does not single out specific areas considered to be particularly promising for the future, or to have high potential social returns. However, Canada does operate a number of supply-side programmes that are sector specific. Two examples are the Strategic Aerospace Defence Initiative and the Automotive Innovation Fund. Explicit priorities are also contained in the Federal Science and Technology Strategy. A question to consider is whether the right balance has been struck between the R&D tax credit and more focused efforts to support priority areas considered likely to have high social returns.

Secondly, by emphasising the R&D tax credit, government policy also focuses primarily on R&D as the driver of innovation performance. While R&D is certainly critical, it is not the only investment that firms make in innovation. In Canada, as in many other OECD countries, a key barrier to stronger innovation performance does not so much appear to be a lack of knowledge, but a lack of take-up in the market and commercialisation. Moreover, a focus on R&D typically implies a choice in favour of sectors and firms that are heavily engaged in R&D, such as high-technology manufacturing and certain service-sector activities. Balancing the support for R&D with support for other important features of innovation (e.g. through support for venture capital, business angels or high-growth firms) as in fact Canada does to some extent; may enhance some of the other dimensions of innovation.

Third, as in other OECD countries, policy makers may wish to consider complimenting the existing supply-side policies to foster innovation with well-designed demand-side policies.



Notes

1. For example, Mowery and Rosenberg (1979) conclude that neither supply nor demand factors are necessary and sufficient for innovation. Both must exist simultaneously. Freeman (1974) surveyed a set of 40 innovations, showing that successful cases were able to link technical with market opportunities.

2. A more elaborate discussion of the rationale for government intervention can be found in Chapter 4 of the OECD Science, Technology and Industry Outlook 2010(OECD, 2010).

3. Indeed, the nature of the alleged market failure will in some cases directly shape the optimal policy response. For instance, a number of the countries considered in this report operate variants of voucher programmes, or other schemes, that seek to facilitate firms’ access to different types of technical or research-related expertise. If the assumption is that firms are unaware of the possible benefits from using external advice then the appropriate policy response should be an initiative that involves a one-time demonstration of these benefits. If the assumption (or evidence) is that markets systematically fail to supply the necessary services in the right amounts then policy could entail continuous measures.

4. A prime instance of this is the long record of failure when governments have themselves sought to act as a supplier of equity finance to new and small firms.

5. Bhide notes that a motivating factor for what he considers a pro-R&D bias in public support is the view that advanced economies depend for their continued affluence on an absolute advantage in scientific knowledge, relative to emerging economies. In policy circles there is a fear that, following large-scale off-shoring of manufacturing, and growing offshoring in services (including R&D), scientific knowledge is the remaining pillar supporting prosperity. But Bhide seeks to demonstrate the fallacy of zero-sum thinking around competition and the development of scientific knowledge. He argues convincingly that an increase in the share of leading-edge research performed in emerging markets will not reduce prosperity at home (his examples refer to the United States but are relevant to OECD economies more generally). On the contrary, they could raise living standards, by benefitting domestic production and consumption. For example, if a technology is developed abroad that helps retailers reduce inventories, then the



capacities of companies like Walmart to apply high technology will translate into greater productivity in the United States. Indeed, technologies of foreign origin – such as the World Wide Web and Skype – are beneficial to the United States’ economy. Stress is placed by Bhide on the importance of numerous intermediate producers in effectively using new know-how and products, and on consumers being ready to purchase innovative goods and services.

6. In a textbook world policy makers would allocate public resources to different programmes based on knowledge of the marginal cost of achieving given common objectives for different programme types. For instance, if increasing private investment in R&D were the principal goal of policy, an economically efficient resource allocation could be achieved if policy makers had information on the marginal costs of increasing private R&D outlays through the different available programmes. An efficient allocation would exist when the marginal costs of a unit increase in R&D spending was the same across programmes. In practice, however, most evaluations provide information on the average cost of achieving some outcome. In addition, the marginal cost of private R&D expansion – or other innovation-related objectives – will vary over time depending, for instance, on the scale and duration of the programmes concerned and the character of the enterprise population. The textbook ideal, then, would necessitate a constant cycle of complex evaluation across many programme types. For practical, methodological and budgetary reasons this goal is effectively unattainable. Nevertheless, carefully targeted evaluations of major programmes could inform the strategic goal of determining a preferred policy mix.

28 – 1. BUSINESS INNOVATION POLICIES – FINDINGS AND POLICY CONCLUSIONS


References

Bhide, A. (2008), The Venturesome Economy: How Innovation Sustains Prosperity in a More Connected World, Princeton University Press.

Edler, J., L. Georghiou, E. Uyarra and K. Blind (2009), Monitoring and Evaluation Methodology for the EU Lead Market Initiative: A Concept Development, University of Manchester/Fraunhofer ISI, Technical University, Berlin.

Freeman, C. (1974), The Economics of Industrial Innovation, MIT Press, Cambridge, MA.

Giebe, T., T. Grebe and E. Wolfstetter (2005), “How to Allocate R&D (and Other) Subsidies: An Experimentally Tested Policy Recommendation”, Discussion Paper No.108, www.sfbtr15.de/uploads/media/108.pdf.

Mowery, D. and N. Rosenberg (1979), “The Influence of Market Demand upon Innovation: A Critical Review of Some Recent Empirical Studies”, Research Policy, Vol. 8, No. 2., pp. 102-153.

Murray, G. (1999), “Early Stage Venture Capital Funds, Scale Economies and Public Support”, paper presented at the conference “Funding Gap Controversies”, Warwick University, 12-13 April.

OECD (2009), OECD Science, Technology and Industry Scoreboard 2009,OECD Publishing, Paris, www.oecd.org/sti/scoreboard.

OECD (2010), OECD Science, Technology and Industry Outlook 2010,OECD Publishing, Paris, www.oecd.org/sti/outlook.

2. PUBLIC SUPPORT FOR BUSINESS R&D – 29


Chapter 2

Public support for business R&D

This chapter assesses countries’ support to private investment in R&D. It provides an overview of current policies, describing both direct support instruments (e.g. direct procurement, grants and subsidies for R&D) and indirect instruments (e.g. R&D tax credits) used by the seven countries examined. General observations are drawn on the relative merits and demerits of direct and indirect support. Countries’ overall mix of policies, notably the balance between direct and indirect support, is also discussed.

30 – 2. PUBLIC SUPPORT FOR BUSINESS R&D


OECD countries provide a range of direct support measures for private investment in R&D. Such support aims in particular to address the inability of firms to fully capture the returns to their investment in R&D, which can lead to a level of private investment below a social optimum. Government action is also premised on system failures, such as a lack of networks in which to innovate; incomplete markets, which may pose particular problems for entrepreneurs seeking equity finance for innovative projects; as well as market power and scale effects, which may make it difficult for new firms to enter markets.

Public support to address possible private under-spending on R&D has taken a variety of forms. These include direct support for R&D – for instance through grants or direct procurement - and indirect support in the form of various sorts of R&D-related tax allowance and credit. The mix of policies used to support private investment in R&D and innovation differs widely across countries. Some comparable data have recently become available on the balance between direct support for R&D and indirect support (through the tax system). These show that some countries such as Sweden, Finland and Germany provide no support to private R&D through the tax system, and only provide direct support, in the form of grants or other mechanisms (Figure 2.1).

Figure 2.1. Direct government funding and tax incentives for business R&D, 2007

As a percentage of GDP

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.40

Unite

d St

ates

(200

8)

Fran

ce (2

008)

Kore

a (20

08)

Czec

h Re

publi

c

Spain

Swed

en

Austr

ia

Finlan

d

Norw

ay (2

008)

Germ

any

Unite

d Kin

gdom

(200

8)

Icelan

d (2

008)

Belgi

um

Luxe

mbo

urg

Denm

ark (

2008

)

Hung

ary

Austr

alia

(200

6)

Irelan

d

New

Zeala

nd Italy

Switz

erlan

d (2

008)

Japa

n

Turk

ey

Slov

ak R

epub

lic (2

008)

Neth

erlan

ds

Cana

da (2

008)

Portu

gal

Polan

d

Mexic

o

Gree

ce (2

005)

% Direct government funding of BERD Indirect government support through R&D tax incentives

Source: OECD (2010), Measuring Innovation: A New Perspective, OECD Publishing, Paris.



Other OECD countries, such as Canada, Japan and the Netherlands provide relatively limited direct support for private R&D, but have tax credits for R&D that account for the bulk of support. Yet other countries, such as Korea, Belgium and Austria, allocate significant resources for both direct and indirect support of private R&D. The optimal balance of direct and indirect R&D support will vary from country to country, as each tool addresses different market failures and stimulates different types of R&D.

The comparison above only covers direct and indirect support to business R&D. No comparable information is currently available on the other forms of financial support that governments provide for innovation in the private sector. However, innovation surveys provide some evidence on the share of innovative firms that receive public support (Figure 2.2). The evidence for Canada only covers the manufacturing sector, which in most countries typically receives more support than the services sector. Between one-tenth and one-third of innovating firms receive public support, with large firms receiving public support more frequently than SMEs.1

Figure 2.2. Firms receiving public support for innovation by size, 2004-06

As a percentage of innovating firms

0

20

40

60

80

Cana

da (2

002-

04,

man

ufactu

ring) Ita

ly

Kore

a (20

05-0

7, ma

nufa

cturin

g)

Nethe

rland

s

Austr

ia

Spain

Belgi

um

China

Japa

n (19

99-2

001)

Germ

any

Luxe

mbou

rg

Czec

h Rep

ublic

Chile

Portu

gal

Icelan

d (2

002-

04)

Esto

nia

Austr

alia

(200

6-07

)

Sout

h Afri

ca (2

002-

04)

Large firms SMEs%

Source: OECD (2010), Measuring Innovation: A New Perspective, OECD Publishing, Paris, based on innovation microdata project based on CIS-2006, June 2009 and national data sources.



Direct support for R&D

Direct support for R&D has the advantage that it can be focused on activities and actors that are of greatest interest in meeting public policy goals and that may have the highest social returns. A case in point is environmental amelioration and pollution abatement. For instance, many governments have made international and time-bound commitments on reducing carbon emis-sions. The search to develop, diffuse and apply low-emission technologies is taking place within defined time-horizons. Direct support can help focus innovation efforts to meet such timeframes, in ways that generic instruments might not.

Focusing support on activities that either have high positive externalities or are prone to market failures can increase the additionality of public support (i.e. producing desired outcomes that would not have arisen in the absence of the support). However, precisely because direct support can be more focused, it also raises the problematic prospect of governments “picking winners”. The design and implementation of schemes to provide direct support for R&D is therefore of great importance. The increasingly common practice of giving direct support to pre-competitive ventures, and to research partnerships, is one means of reducing problems associated with picking winners at the level of individual firms. Good policy designs need to ensure competitive selection processes, to select projects that best serve public policy objectives, avoid favouring incumbents or providing opportunities for lobbying, contain costs, and ensure a rigorous evaluation of policy impact.

Governments can introduce design features in R&D support that assist efficient allocations. Giebe et al. (2005) assess how R&D grants for firms are awarded. This is often on the basis of a ranking of individual projects, rather than an assessment of the returns to a set of applications. The authors describe a situation in which candidate proposals are graded – say along a spectrum A to C. Proposals above some predetermined threshold, for instance a B, are given priority, and are supported until funds are exhausted. The authors note however that such a process can be inefficient, because an allocation of fixed resources to a higher number of lower grade projects could give greater returns than allocating those same fixed resources to a smaller number of higher grade projects. Indeed, under the decision rule “only fund grade A proposals”, any number of lower grade but more numerous proposals could be rejected. A second inefficiency in such systems stems from the limited information used in selection and from the practice of financing proposals on the basis of a predetermined percentage of the project cost. Funding on a fixed percentage basis can lead to over funding of projects (which effectively excludes other projects). By inducing competition among applicants – through the use of various auction mechanisms – greater information can be extracted



on the proposals and some degree of unnecessary funding can be avoided. Controlled laboratory experiments suggest that adopting these proposals could considerably improve the efficiency of allocation.

Several OECD countries also have more generic policies that provide direct support. For example, the innovation vouchers used in the Netherlands (see Chapter 3) are a generic tool to provide direct support. Such programmes tend to have relatively low information and administrative costs, and also face a smaller risk of government failure. On the other hand, there is a greater risk of large deadweight loss, with support possibly going to firms that would have undertaken innovation efforts even without public support.

Figure 2.3 shows OECD data on direct support for R&D as a percentage of total business expenditure on R&D. It shows that France, the United States and Norway had the highest ratio of private R&D funded by governments, with Japan, Switzerland and Canada funding lower levels of private R&D directly.

Figure 2.3. Direct government funding of business R&D, as % of total BERD, 2000 and 2008

0.0

2.0

4.0

6.0

8.0

10.0

12.0

14.0

16.0

18.0

2008

2000

Note: Data are from 1999 (instead of 2000) for Denmark. Data are from 2007 (instead of 2008) for Iceland, Norway, New Zealand and Sweden.

Source: OECD, Main Science and Technology Indicators database, 2010.



Typically, some direct support is linked to public procurement. For example, in countries such as Spain, France, the United Kingdom and the United States, governments support R&D by firms in the defence industry, aimed at the development of new military equipment. This component is not easily separated from other forms of direct support. However, some OECD data indicate that this component can be significant. For example, more than 50% of the US government budget appropriations for R&D are allocated to defence and more than 30% of the US government’s direct support for business R&D went to the aircraft industry in 2000 (the latest year for which detailed data are available).

Other direct support for private R&D mainly aims at stimulating private investment in R&D. This typically takes the form of grants or subsidies and can be generic, supporting innovation in every firm that qualifies for the programme, or oriented towards specific sectors or goals.

Programme examples The text that follows describes examples of direct support programmes

in five of the countries examined. The range of support measures comprises grant schemes (Australia and Denmark), loan facilities (the Netherlands), and mixed grant/loan arrangements (Finland). Some programmes are oriented to a specific sector (Australia), while others are generic, applying in principle to any sector. The schemes are also managed by agencies with somewhat varying remits, operating at both central and sub national levels of government. Schemes also vary as concerns the balance between the sponsorship of research and the focus on commercialisation.

Australia – Green Car Innovation Fund Commencing in July 2009, Australia’s Green Car Innovation Fund

provides assistance in the form of grants to Australian companies for projects that enhance research and development and commercialisation of Australian technologies. The programmeis directed towards projects that can significantly reduce greenhouse gas emissions or lower fuel consumption in passenger motor vehicles.

The grants start at AUD 100 000 (USD 77 000) and cover up to 25% of the costs of the project. Grants are provided at a ratio of AUD 1 of government funding for every AUD 3 of eligible expenditure by the grantee. The Fund has a total value of AUD 1.1 billion (USD 857 million) and will operate over ten years.



Denmark – Danish National Advanced Technology Foundation The Danish National Advanced Technology Foundation offers grants for

co-funding of high-technology research and innovation projects. Each project must possess obvious commercial potential, opportunities for technology transfer and opportunities for collaboration between public-sector research institutions and private-sector companies. Higher education centres and/or public-sector companies can also participate. The goal is for the Foundation to have a capital base of DKK 16 billion (USD 3 billion) by 2012.

Netherlands – Innovation Credit Scheme The Innovation Credit Scheme was launched to address the problem that

SMEs find it difficult to finance their innovations. An indication of the relative importance of this scheme in the Dutch policy mix is that the budget has increased from EUR 21 million (USD 29 million) in 2008 to EUR 50 million (USD 69 million) in 2010. An SME in the Netherlands may be eligible for a loan to finance a technical development project. The project must result in the development of technologically innovative products, processes or services with good commercial prospects. The loan will account for no more than 35% of the project costs, up to a maximum of EUR 5 million (USD 7 million). The loan is made available in phases. Both start-ups and existing companies are eligible.

Finland – Foundation for Finnish InventionsThe Foundation for Finnish Inventions supports private individuals and

entrepreneurs to develop and exploit invention proposals through the provision of risk financing. The financing can take the form of grants or loans, which generally vary between amounts of EUR 2 000 (USD 2 800) and EUR 200 000 (USD 280 000).

Grants are aimed at covering the early costs of developing an invention. Funding decisions are made on the basis of the innovativeness, technical functionality and economic evaluation of the invention proposal. The inventor has to supply the Foundation with a written report on the use of the funding.

Funding is generally used to pay the costs of patenting, product develop-ment and commercialisation. It incorporates a refund to the Foundation, contingent on the success of the project and on the revenue received by the inventor.



If the inventor starts a business to exploit an invention commercially the Foundation can lend working capital to the inventor-entrepreneur. The loan is usually only granted if adequate funding is unobtainable from other sources.

Canada – National Research Council of Canada Industrial Research Assistance Program (NRC-IRAP)

NRC-IRAP supports SMEs to develop, grow or adapt technologies to improve the lives of Canadians. The IRAP team provides advice on science and technology, identifying partnerships for firms and facilitating networks.

Canada – The Strategic Aerospace and Defence Initiative (SADI) The SADI encourages strategic research and development (R&D) and

aims to enhance the competitiveness of Canadian aerospace and defence companies, notably by fostering collaboration with research institutes, universities and colleges.

United Kingdom – Matched Grant funds Grants and support schemes are made available for specific activities

and take the form of a subsidised loan or an outright cash grant. Most government grants require the business to match the funds being awarded. The amount of matching funds asked for varies across scheme (for instance a research grant may require a business to find 40% of the total cost, with 60% provided by the grant). Matching funds can come from the owners of the business, retained profits, a loan, or from a new investor.

United Kingdom – Funding by the Technology Strategy Board The Technology Strategy Board2 (TSB) matches its investments with

those made by United Kingdom businesses. By 2011 it is expected to have generated a GBP 2 billion (USD 3.1 billion) programme of investment designed to support business and drive technology-enabled innovation. The TSB is jointly supported and funded by BIS and other government depart-ments, the devolved administrations, regional development agencies and research councils. Among other activities, the TSB promotes supports and invests in technology research, development and commercialisation.

United Kingdom – Innovation, Research and Development Grants The Grant for Research and Development provides grants to help

individuals and SMEs based in England and involved in researching and developing technologically innovative products and processes. The Grant



for Research and Development is administered by the nine English Regional Development Agencies (RDAs) (Scotland, Wales and Northern Ireland run equivalent schemes).

There are five different types of project that a grant can be awarded for and each requires the applicant to contribute to project costs:

Proof of market grants to test the commercial potential of an innovative idea for a new technology. A grant of up to GBP 20 000 (USD 31 000) is available to businesses with fewer than 250 employees.

Micro projects are simple, low-cost development projects lasting no longer than 12 months. A grant of up to GBP 20 000 (USD 31 000) is available to businesses with fewer than ten employees.

Research projects aim to investigate the technical and commercial feasibility of innovative technology and last from 6 to 18 months. A grant of up to GBP 100 000 (USD 155 000) is available to businesses with fewer than 50 employees.

Development projects aim to develop a pre-production prototype of a new product or process that involves a significant technological advance. Projects last from six to 36 months. A grant of up to GBP 250 000 (USD 389 000) is available to businesses with fewer than 250 employees.

Exceptional development projects involve a significant technological advance and must be considered strategically important for a particular technology or industry. Projects can last from 6 to 36 months. A negotiable grant of up to GBP 500 000 (USD 779 000) is available to businesses with a qualifying project.

R&D tax credits3

A growing number of OECD governments are offering fiscal incentives to business in order to increase private spending on R&D. Many are also redesigning their R&D tax incentives to make them more effective. Traditionally, the argument for tax incentives lies in their non-discrimina-tory nature and relative ease of implementation. Fiscal incentives may also be more accessible than some forms of direct support, being equally available for instance to small and medium-sized as well as large firms. R&D tax credits are neutral with respect to the type of R&D conducted, and therefore conform to market rationale more closely than some types of direct support. At the same time, for the very reason that R&D tax credits can support any formal R&D, they may be less effective in targeting public resources to forms of R&D that could have the highest social returns.



Table 2.1 provides an overview of R&D tax credits offered by Australia4,Canada, Denmark, the Netherlands, Norway and the United Kingdom. The most common type of scheme used by countries is a volume-based tax incentive in which eligible expenditures are based on current R&D (e.g.United Kingdom, Czech Republic, Norway, Denmark) or the current machinery and equipment component of R&D (e.g. Canada, Australia). Countries generally provide more generous support to SMEs. Some OECD countries (e.g. Sweden, Finland) neither subsidise nor extend preferential tax treatment to business R&D, even though these countries record high levels of private R&D expenditure. Other countries (e.g. Germany) prefer R&D subsidies over taxes so as to steer research to particular goals.

Van Pottelsberghe et al. (2003) provide an overview of issues to be considered when designing fiscal support for business R&D (Box 2.1). The general fiscal environment will play a role in deciding whether to use fiscal measures at all, as fiscal incentives are less effective in a country with a low corporate income tax. In addition, the target group needs to be selected, eligible expenses must be defined and a choice made between a tax credit that applies on all R&D outlays (volume) and a credit based on additional spending on R&D (the incremental approach). A volume-based tax credit has a number of advantages: it is simple to administer and relatively predictable. But it provides less of an incentive to conduct additional R&D and is more costly for government. By contrast, incremental schemes are more complex for businesses. This has a negative effect on smaller businesses and risks diluting the intended policy effect. All the countries reviewed in this study have chosen a volume-based approach (Australia has proposed a move to an R&D tax credit to replace the current R&D tax concession. At the time of publication, this is being debated in the Australian Parliament).

Most countries provide fiscal incentives through tax credits or allowances and capital expensing. In some countries, however, including the Netherlands, fiscal incentives are used to reduce tax on the wages of workers engaged in R&D, as well as to lower national insurance contributions.5 Belgium’s tax incentives cover R&D expenditures but also include a deduction for patent income. In 2008, 21 OECD countries operated R&D tax credits, up from 18 in 2004. France and Spain provide the highest subsidy rates and make no distinction between large and small firms (Figure 2.4). Canada and the Netherlands are significantly more generous to small firms than to large ones. Tax subsidy rates for R&D by large firms increased significantly between 1999 and 2008 in France and Norway (where schemes were introduced in 2002 for small firms and 2003 for all firms), and to a lesser extent in Italy, Portugal, the United Kingdom, Belgium and Japan (Figure 2.5). Elsewhere, the tax subsidy rate remained stable (except in Mexico and Denmark, where it decreased – with Mexico repealing its tax credit in 2008).



Table 2.1. Overview of R&D tax incentives (2009 and 2010)

Rates Expense base Deducted from

AUSTRALIA 2010 R&D Tax Credit 45% on volume

refundable for small firms (aggregate turnover of less than AUD 20m) and 40% of volume non-refundable for other firms

Current expenditure and machinery and equipment (depreciation)

Tax Payable Small firms (aggregate turnover of less than AUD 20m) in tax loss position can claim an R&D tax refund equal to 45% of R&D Expenditure. No cap on R&D expenditure amount

Indefinite No ceiling but a floor of AUD 20,000 (USD 15,600)

2009 R&D tax concession (allowance)

125% on volume plus 175% on increment over past 3 years average

Current expenditure and machinery and equipment (depreciation)

Taxable income (taxable)

Small firms (aggregate turnover less than AUD 5m) with less than AUD 2m of R&D expenditure that are in a tax loss position can claim an R&D tax refund equal to tax savings from Tax Concession

Indefinite No ceiling but a floor of AUD 20,000 (USD 15,600)

2008: AUD 820M (0.07% of GDP)

CANADASR&ED Tax Credit 35% for small firms

(CCPC) and 20% on any excess amount; 20% for large firms

Current and machinery and equipment

Tax payable (benefit is taxable)

Cash refund for small firms (CCPC)

20 years Generally no ceiling but small firms get 35% credit on first CAD 3M (USD 2.80 M) R&D only

2008: CAD 3.2B (0.22% GDP)

DENMARKR&D allowance 200% on volume Current costs Taxable income Not available Indefinite 2008: DKK 1.15B

(0.06% GDP)

Foreign researchers and key staff

Taxed at lower rate (at 25 for 36 months or 33% for 60 month)

There are a number of limitations and conditions for the reduced tax rate to be allowed

NETHERLANDSR&D wage tax credit 2009: 50% reduction

on the first EUR 150,000 of R&D wage bill; 18% afterward; 2010: 50% reduction on the first EUR 220,000 of R&D wage bill; 18% afterward

Research wages Reduction of withholding tax on wages

Works like refund Not applicable

Ceiling of EUR 14M of R&D wage bill (increased from EUR 8M in 2008)

2008: EUR 445M (0.07% GDP)

NORWAY

R&D tax credit 18% on volume; 20% for SMEs

Current costs Tax payable Refund of tax credits exceeding assessed taxes (all firms)

Not relevant

Up to NOK 5.5 million from 2009 (earlier NOK 4 million); if joint project with approved R&D institution – up to NOK 11 M (earlier 8 M)

2008: NOK 1B (0.07% GDP)

UNITED KINGDOMR&D allowance Incentive R&D depreciation

175% for small firms; 130% for large firms; volume based

Current cost Taxable income Refund for small firms - 24% of the cash cost of the qualifying R&D

Indefinite No ceiling but a floor of GBR 10,000

2008: GBR 820M (0.06% GDP)

100% R&D plant & equipment

Taxable income

LimitationsCarryoverRefund Country/ Tax Incentive Description of Tax Incentive Forgone tax

revenue

Note: for the enhanced allowance, the corporate income tax (CIT) rate must be taken into account to estimate the tax liability reduction. For example, for a small UK firm with a 21% corporate CIT, the tax liability reduction of one unit of R&D would be 0.16 (1x (175%-100%) x21%). Estimation of the tax liability reduction is more direct with the tax credit. For example, for a small Canadian firm, a tax liability reduction of one unit of R&D would be 0.35 (1x35%) for the first CAD 3 million, and 0.20 (1x20%) for eligible R&D above CAD 3 million. For Dutch firms, tax liability reduction (from payroll withholding tax) of one unit of R&D wage is 0.42 for the first EUR 0.22 million, and 0.14 after that (up to EUR 14 million of eligible R&D).

Source: OECD (forthcoming), “R&D Tax Incentives and Government Forgone Tax Revenue: A Cross-Country Comparison” and responses to the OECD Science, Technology and Industry Outlook questionnaire.



Box 2.1. Design considerations: Fiscal support for business R&D The target group

Governments can make fiscal support accessible to all companies, or make support more generous for target groups of firms (e.g. SMEs). This can be done by:

Placing upper limits on the amount of tax credit that can be claimed (upper limits are more likely to be attained by larger companies than by SMEs).

Giving higher tax credit rates for SMEs, and/or greater flexibility, such as cash refunds or unused credits.

Minimum thresholds can increase the efficiency of policy as administrative costs can also be high for small applications. Labeling of activities and claiming the tax credit

While the definition of R&D is typically based on the 1993 Frascati Manual, most countries have produced lists of types of R&D that qualify. To claim the tax credit, companies can be obliged to submit their R&D projects for prior approval. Alternatively they might claim the credit at the end of the year. While mandatory upfront applications eliminate uncertainty regarding eligibility, they lack flexibility in the case of change (Norway is a case in which upfront applications address issues of research content, more than issues of cost, and where it is recognised that projects evolve as knowledge accumulates).Eligibility of the R&D expenditure

Three types of expenditure can qualify for the fiscal incentive: Expenditure on wages related to R&D (for example, this is the case in the Netherlands). This reduces social security and wage taxes and gives an incentive for investment in human capital.

Current R&D expenditure, which includes wages and all consumables used in the R&D process.

Current and capital R&D expenditure. This enlarges the incentive for companies, but increases the public cost of the policy.

Carry over provisions and cash refunds These provisions allow unused portions of the credit to be carried forward or backward

to previous fiscal years. Carry forward provisions are particularly important for SMEs, as these tend to have limited current corporate income against which the credit can be applied, while many younger firms are carrying losses from previous periods. Cash refunds can also replace carry forward provisions. The time value of funds should be taken into account when calculating refunds. Delays in effecting cash refunds need to be avoided in order to making this tool efficient. Source : Van Pottelsberghe et al. (2003).



Figure 2.4. Tax subsidy rate in 2008 for USD 1 of R&D (large firms and SMEs)1

-0.1

0.0

0.1

0.2

0.3

0.4

0.5

Large firms SMEs

Note: Data for Mexico are for 2009.

1. The tax subsidy for R&D is calculated as 1 minus the B index. The B index is defined as the present value of before-tax income necessary to cover the initial cost of R&D and to pay corporate income tax, so that it is profitable to perform research activities.

Source: OECD (2009), OECD Science, Technology and Industry Scoreboard 2009, OECD Publishing, Paris.

Figure 2.5. Change in the tax subsidy rate for USD 1 of R&D (large firms and SMEs, between 1999 and 2008)1

-0.40

-0.30

-0.20

-0.10

0.00

0.10

0.20

0.30

0.40SMEs Large firms

Note: Data for Mexico are for 2009.

Source: OECD (2009), OECD Science, Technology and Industry Scoreboard 2009, OECD Publishing, Paris.



Programme examples

Australia In 2009 the Australian government proposed to replace the R&D Tax

Concession (an allowance) with an R&D Tax Credit. This new support system would redirect assistance to those activities most likely to generate spillovers. It aims to favour smaller innovative firms more likely to respond to fiscal incentives, but also encourages investment by the growing number of multinational enterprises in Australia. Firms undertaking R&D in Australia, where the intellectual property is owned overseas, would be eligible for the tax credit.

A 45% refundable tax credit is proposed for small firms with an aggregate turnover of less than AUD 20 million (USD 16 million) per annum. A 40% tax credit is proposed for all other firms with an aggregate turnover above AUD 20 million (USD 16 million).

CanadaThe Scientific Research and Experimental Development (SR&ED)

programme is a federal tax incentive administered by the Canada Revenue Agency (CRA). It gives claimants support in the form of cash refunds and/or tax credits and is the largest single source of federal government support for industrial R&D (it provided over CAD 3 billion in tax assistance to innovative Canadian businesses in 2006). The R&D tax incentives are supplemented by similar measures in most provinces.

Claimants can apply for R&D investment tax credits for expenditures such as wages, materials, machinery, equipment and some overhead. To qualify for the R&D programme, work must advance the understanding of scientific relations or technologies, address scientific or technological uncertainty, and incorporate a systematic investigation by qualified personnel.

A Canadian-controlled private corporation (CCPC) can earn an investment tax credit of 35% up to the first CAD 3 million of expenditures for R&D (carried out in Canada), and 20% on any excess amount6. Other Canadian corporations, proprietorships, partnerships, and trusts can earn an investment tax credit of 20% of qualified R&D expenditures carried out in Canada. The investment tax credit earned by a Canadian corporation that is not a CCPC is non-refundable, but may be used to reduce any taxes payable. The ITC earned by a proprietorship or certain trusts may be partially refunded after applying these tax credits against taxes payable.



The upper limit of the taxable income phase-out range was increased several times (by Budget 2006, Budget 2008 and Budget 2009). Revisions to the tax legislation over the last five years have allowed certain R&D activities to be undertaken outside of Canada. The geographical area of the investment tax credit was also expanded to include Canada’s Exclusive Economic Zone.

In addition to the Federal R&D tax incentive scheme, there are also some sub-national R&D tax incentives to encourage R&D, notably in Québec and Ontario.

For companies performing R&D in-house Québec has introduced: a fully refundable basic tax credit corresponding to 17.5% of R&D salaries paid in Québec; a fully refundable tax credit of 37.5% on the first USD 3 million of R&D-related salaries per year for a Canadian-controlled SME. As regards R&D entrusted to a certified research centre, as well as private partnership research, and dues and fees paid to an R&D consortium, a fully refundable tax credit is available corresponding to 35% of eligible expenditures. In addition, foreign researchers employed by a company in Canada that does R&D in Québec benefit from a provincial tax holiday on their salaries for five consecutive years.

The Ontario Research and Development Tax Credit (ORDTC) is available to corporations that carry out scientific research and experimental development work in Ontario. It is non-refundable and provides a 4.5% tax credit based on eligible R&D expenses. The credit may be used to reduce payable corporate income tax. Any unused credit may be carried back three years (to tax years ending after 31 December 2008) or carried forward 20 years.

Denmark7

Denmark gives tax incentives (or allowances) for basic research con-ducted by the private sector. It allows companies to deduct expenditures for experimental research activities in either the year in which they occur or over a four-year period. In specific circumstances, expenditures that occurred prior to the official start of a business can also be deducted. Enhanced deduction (100% over the normal deduction rate of 100%) is permitted for certain R&D expenses incurred. In addition, foreign researchers and key staff have a reduced income tax rate. They can choose between a 25% tax rate over a period of 36 months or a 33% tax rate over 60 months (when the labour market contribution are included, the tax rates reach 31% and 38.4%). The system was introduced in 1991. It should also be noted that a law introduced in 2010 allows individuals to obtain a deduction for donations to charities or other organisations providing resources for research of benefit to the public.



Netherlands The Research and Development Promotion Act – WBSO – is a fiscal

scheme aiming to stimulate business R&D by lowering R&D-related labour costs. The scheme targets innovative SMEs in both industry and services.

WBSO’s budget in 2009 was EUR 550 million (USD 764 million) and was recently increased in 2010 to EUR 700 million (USD 972 million) in order to help counter the economic downturn. For the period 2011-14 WBSO’s yearly budget was set at EUR 547 million (USD 759 million).

An evaluation of WBSO in 2007 concluded that the scheme’s uptake is widespread (de Jong and Verhoeven, 2007). The impact on R&D expenditure was found to depend on company size, the effect being larger in small companies. SMEs constituted over 95% of the users.

NorwayThe R&D tax credit scheme-Skattefunn-was passed by Parliament for

the tax year 2002 for small firms, and was extended to all firms from 2003. To be eligible, businesses must be subject to taxation in Norway. Under the scheme, the R&D content of projects must be approved by the Research Council of Norway ex ante, while final costs are approved by the tax authorities ex post. The scheme offers rebates of 20% of qualifying expenses for SMEs and 18% for large enterprises. In both cases, there is a cap on expenses per enterprise of NOK 5.5 million (USD 870 000) for intramural R&D and NOK 11 million (USD 1.75 million) for projects conducted at an approved R&D institute. One should note that the calculation of cost is affected by an exogenously established figure for hourly wages (approxi-mately EUR 62) and annual hours (at 1 850). If the calculated tax credit exceeds the assessed taxes of the firm, the difference is refunded rather than carried forward. About three-quarters of total tax expenditure under the Skattefunn scheme has taken the form of such cash refunds. The total R&D tax concession for 2008 is approximately NOK 1.0 billion (USD 160 million).

In a recent evaluation, carried out by Statistics Norway, it was found that firms that received support through Skattefunn had stronger growth in their R&D investments than other firms (Cappelen et al., 2008). It was also found that firms that previously invested less than the cap (NOK 4 million; USD 640 000) increased their R&D investments more than those previously above the ceiling, and that firms that previously did not invest in R&D were more likely to start doing so since the introduction of Skattefunn. Estimates of how much additional R&D Skattefunn triggers per NOK in foregone tax revenue vary between 1.3 and 2.9, which compares well with results for other countries.



United Kingdom The Government introduced R&D tax credits for SMEs in 2000 and for

large companies in 2002. The SME scheme was extended to mid-size companies in 2008. The tax relief on allowable R&D costs is 175% for SMEs and 130% for large firms (or respectively 75% and 30% over the normal deduction rate of 100% for SMEs and large firms). SMEs not earning profits have the choice of either carrying forward their 175% tax allowance to future years when they will be profitable, or receiving a rapid refund of 24% of eligible R&D expenditures. Under the large-company scheme, tax relief is only available if the company spends at least GBP 10 000 a year (USD 16 000) on qualifying R&D (no upper limit). During 2009 over GBP 3 billion (USD 4.76 billion) of relief was claimed by 36 000 companies. The condition that any intellectual property deriving from the R&D must be owned by the company making the claim is to be dropped.

At the time of writing, consideration was being given as to whether a further fiscal support mechanism will be introduced in 2013, namely the so-called patent box. The scheme would mirror similar measures introduced in the Netherlands and Belgium, and would introduce a 10% rate of corpora-tion tax on income derived from patents held in the United Kingdom. This compares with a main rate of corporation tax of 28%. Some parts of the business community less reliant on patents, such as engineering firms, are reported to prefer reform of the existing R&D tax credit (The Times,20 August 2010, page 33).

In November 2006, the government launched a nationwide network of specialist R&D units within HM Revenue & Customs (HMRC) to improve the consistency of the claims process. The units are organised on a geographic basis, dealing with claims from companies and organisations whose main R&D base is within their postcode allocation.



Notes

1. A recent OECD econometric study using innovation surveys from 21 countries estimated that firms receiving public support for innovation invest 40 to 70% more than those that do not (OECD, 2010b).

2. The Technology Strategy Board is an executive non-departmental public body, established by the government in 2007 and sponsored by the Department for Business, Innovation and Skills (BIS).

3. For more information on R&D tax incentives in OECD countries, see OECD (forthcoming), R&D tax incentives and government forgone tax revenue: A cross-country comparison.

4. In Australia the R&D tax credit is currently under debate and has not been finalised. Therefore, the information in Table 1 refers to the proposed rates for the new R&D tax incentive and may be subject to change.

5. The OECD is working to compare countries’ R&D schemes and methodologies and to assess factors that affect the overall cost (inclusion of sub-national R&D tax credits, differences in firm eligibility, etc.).

6. Firms eligible for the 35% tax credit must have prior-year taxable income of CAD 500 000 or less and prior-year taxable capital employed in Canada of CAD 10 million or less. The CAD 3 million expenditure limit is phased out if prior-year taxable income is between CAD 500 000 and CAD 800 000 or if prior-year taxable capital is between CAD 15 million and CAD 50 million. In this paper, to ensure proper international comparison, small firms are defined as Canadian-controlled private corporations (CCPCs) with prior-year taxable income of CAD 500 000 or less and prior-year taxable capital employed in Canada of CAD 10 million or less. As stated by Finance Canada (2009), “defining small firms in terms of eligibility for the 35% federal credit rather than in terms of employment has only a minor impact on the share of small firms in total R&D spending” (Source: Finance Canada (2009), Tax Expenditures and Evaluations 2009, Ottawa, Canada).

7. Information regarding Denmark R&D tax incentives is preliminary.



References

Cappelen, A., E. Fjærli, F. Foyn, T. Hægeland, J. Møen, A. Raknerud and M. Rybalka (2008), “Evaluering av SkatteFUNN”, Rapporter 2008/2, Statistics Norway, Oslo.

de Jong, J. and W. Verhoeven (2007), WBSO evaluation 2001-2005: Impacts, target group reach and implementation, Research Series, Ministry of Economic Affairs, The Hague. Netherlands.

Finance Canada (2009), Tax Expenditure and Evaluations, Ottowa, Canada.

Giebe, T., T. Grebe and E. Wolfstetter (2005), How to Allocate R&D (and Other) Subsidies: An Experimentally Tested Policy Recommendation, Discussion Paper No.108, www.sfbtr15.de/uploads/media/108.pdf.

OECD (2009), OECD Science, Technology and Industry Scoreboard 2009,OECD Publishing, Paris, www.oecd.org/sti/scoreboard.

OECD (2010), Measuring Innovation: A New Perspective, OECD Publishing, Paris.

OECD (forthcoming), R&D tax incentives and government forgone tax revenue: A cross-country comparison, OECD, Paris.

Van Pottelsberghe, B., E. Megally and S. Nysten (2003), “Evaluation of Current Fiscal Incentives for Business R&D in Belgium”, Université Libre de Bruxelles, Solvay Business School, Centre Emile Bernheim, Working Paper WP-CEB 03/011.

3. NON R&D-BASED PUBLIC SUPPORT FOR BUSINESS INNOVATION – 49


Chapter 3

Non R&D-based public support for business innovation

This chapter reviews other forms of government support for business innovation. Non R&D-based innovation support is of growing importance. In recent years, government support has shifted towards measures to ease the creation and development of early ventures through seed investment, venture capital and commercialisation programmes. Non R&D-based support for innovation has also sought to facilitate the creation and diffusion of knowledge to firms through initiatives to promote networks of different sorts, including regional innovation hubs (centres of excellences) and a range of programmes aimed at assisting firms’ access to information and expertise.

50 – 3. NON R&D-BASED PUBLIC SUPPORT FOR BUSINESS INNOVATION


Policy rationales

As noted in the caption, support for innovation other than R&D-related schemes includes measures to facilitate access to early-stage finance, particularly for equity, initiatives to assist the commercialisation of innova-tion, support for the development of networks of different sorts, programmes to promote regional innovation hubs, and a range of initiatives – such as voucher programmes – aiming to ease access to information, expertise and advice. These schemes are obviously more targeted than R&D tax credits and careful consideration of programme logic is required when establishing schemes, such as these, that directly supplement market activities. Policy makers must ensure – among other things – that those programmes do not override or distort important market mechanisms. In this connection, as background to a description of some of the programmes in the countries under review, the following paragraphs briefly consider the economic rationale for this class of policies.

Venture capital One important area of government support described below concerns

risk capital. Many OECD countries have schemes and funds to support the venture capital industry, with some governments playing an active role in the provision of venture capital. There is much empirical evidence to support the proposition that the supply of small volumes of equity can be particularly problematic for start-ups and new-technology-based firms. This is largely because venture capitalists face costs in assessing, monitoring and managing investments that vary little with the size of the investments they make. Whether the investment is worth EUR 1 million, or EUR 50 million, the basic costs entailed in the transaction are similar. Accordingly, smaller investments become relatively unattractive for the venture capitalist.

In addition, mobilising investment resources is a time-consuming and therefore costly process. This fact, combined with regulations that in some countries restrict the share of a venture fund that can be held by any individual partner, creates incentives to raise large-volume funds. Conse-quently, selecting large investment projects makes sense for fund managers who need to invest large-volume resources.

Analysts have drawn a distinction between so-called “primary” and “secondary” equity gaps. The primary gap represents the smallest size of investment required by most formal venture capital firms. The secondary gap is the minimum equity investment given consideration by informal private equity investors (so-called “business angels”).1



Another aspect of formal equity investment that can be a concern for policy makers relates to its geographic distribution. Most formal venture capital activity is highly concentrated in locations that possess significant existing financial services activity and/or a sizeable volume of new firm births (i.e. a sizeable potential deal flow). Commercially viable new firms using innovative technology, but located outside of such places, often find it difficult, if not impossible, to access formal equity finance.

Information and advice Some of the public programmes described below seek to encourage

and/or facilitate access to research or technology-related information and advice. On the demand side, small firms in particular can incur high costs in information search and screening processes (at least relative to their turnover). A further demand-side argument is that small and micro-enterprises might not know their real assistance needs, especially in a context of rapidly changing technologies (this argument however is open to criticism, as many would view it as incumbent on entrepreneurs to seek relevant information). On the supply side, the elevated costs of marketing services to large numbers of small enterprises – relative to expected revenues – is held to be a barrier to institutional (and particularly private sector) outreach.2 Whether the market works well in providing advisory and information services to new and small firms is a contested subject, but market failure may be less frequent than is claimed. The Internet is augmenting information supply to small firms, and many providers of technical information have a strong interest in demonstrating products, equipment and services to potential small-firm clients.

The nature of the alleged market failure is directly relevant to the form that public intervention should take. If the assumption is that firms are unaware of the possible benefits from using external advice then the appropriate policy response should involve a one-time demonstration of these benefits. If the assumption (or evidence) is that markets systematically fail to supply the necessary services in the right amounts then a policy to rectify this situation could entail continuous measures.

Collaborative R&D and business networks Some of the public initiatives described here entail financial incentives

for joint or collaborative projects, including collaborative R&D. From a public policy standpoint, once a rationale for supporting private-sector R&D has been accepted, advantages arise from having that R&D occur in the form of collaborative undertakings. For instance, economies of scale and scope can be realised, while informational spillovers can be internalised among the collaborating firms. With this in mind, many countries have



decided to exempt R&D partnerships from anti-trust legislation. Support for collaborative research also reflects the realisation that many research projects can be too multifaceted and complex for any individual firm or institution.

The generic term “network” covers multiple forms of informal and formal collaboration. Of relatively recent policy attention are formal mechanisms of co-operation among groups of firms, and sometimes among firms and research institutions or centres of technical excellence.

Some network programmes aim at general sharing of information, while others tackle more specific goals. Networks can allow rapid learning and can also facilitate the reconfiguration of relationships with suppliers. In fact, networks have spurred co-operation on issues as diverse as training, techno-logical development, product design, marketing, exporting and distribution.

Governments might justify a facilitatory role in network development owing to the fact that in some places and industries there may have been no, or limited, prior familiarity with the opportunities that networks afford. In some industries the network principle may be hard to establish, with fears existing of possible unfair appropriation of the benefits from any collabora-tive undertaking. Co-ordination problems among geographically dispersed firms may also create inertia in the establishment of business networks.

Support for technology-related entrepreneurship per seSome of the programmes described below also seek to provide

generalised facilitation of entrepreneurial activity in technology-intensive sectors. One argument that could afford some theoretical justification for such schemes relies on the presupposition that the creation of new firms gives rise to positive externalities. Attention has been drawn to the fact that when an entrepreneur creates a business, valuable information is provided to other actual and potential entrepreneurs. That is, the establishment of a firm generates a positive externality. Information is supplied, for instance, on what products sell and what business strategies work. Even business failure sends useful signals to others. However, entrepreneurs are not rewarded for producing this information. Therefore, it is claimed, the amount of entre-preneurial activity is below a level that economists would describe as socially optimal. If imitation plays a significant role in diffusing entre-preneurship the plausibility of such an account is strengthened.3 However, this policy rationale is theoretical only, with as yet little empirical basis.



The rationales for policy outlines above are highly stylised. Govern-ments themselves sometimes make only passing or incomplete reference to the economic rationale for particular initiatives. Furthermore, some of the theoretical justifications for policy lack a robust empirical complement: an argument for a given market failure can be made on paper, but whether the purported failure is empirically significant is sometimes underexplored. At the same time, some policy initiatives to support innovation have multiple objectives, tackling more than one market constraint (technology incubation schemes for example provide both advice and information to entrepreneurs while addressing gaps in the supply of certain types of industrial real estate required by new/small firms). Where this is the case, careful examination of all the relevant market conditions is called for (many technology incubators have been created, for instance, in places where the market for technical works well, but the market for suitable industrial real estate does not, implying that a different form of intervention might be more cost effective).

Allocating direct support

A merit of direct support is that it can be focused on specific barriers to innovation. These barriers are varied, and can include: the possible under-supply of private investment in R&D; failure among market actors to supply public goods (such as open, credible international technical standards); gaps in the market for early-stage equity finance; possible under-provision of business development services, especially for small firms and for businesses located in remote or lagging regions; co-ordination and information problems that hinder networking or other types of collaborative activity; and difficulties faced by some SMEs in accessing public procurement markets. Several governments also target resources to attracting multinational firms that tend to be more internationally mobile.

Focusing support on activities that either have high positive externalities or are prone to market failures can also increase the additionality of public support (i.e. producing desired outcomes that would not have arisen in the absence of the support). However, precisely because direct support can be more focused, it also raises the problematic prospect of governments picking winners. The design of schemes to provide direct support for innovation is therefore of great importance. As noted earlier in this report, good policy designs need to ensure competitive selection processes, promote projects that best serve public policy objectives, avoid favouring incumbents or providing opportunities for lobbying, contain costs, and ensure a rigorous evaluation of policy impact. As focused direct support puts greater demands on government capabilities to shape programmes, direct support tends to be more expensive than indirect support in terms of the costs of executing the programme and administering the selection process.



Governments can take various steps to harness market mechanisms in the design of direct support, such that the information contained in markets is brought to bear, while costs to be borne by the public sector are contained. In the environmental sphere for example, the empirical evidence suggests that market-based instruments such as tradable emissions permits are more likely to stimulate innovation than direct regulations such as technology-based standards (OECD, 2010). Moving beyond R&D, and considering support for innovation more broadly, market mechanisms can be brought to bear in a number of ways to facilitate the financing of innovation. For instance, most formal venture capital investment in start-ups is undertaken by small-volume funds that face high unit costs and high levels of risk. Various governments have established a range of equity guarantee/enhance-ment schemes focusing on early stage investments. Murray (1999) sets out the key issues in the design of equity guarantee/enhancement schemes. The three generic options available are: providing insurance against the venture fund’s losses; providing upside leverage; and subsidising part of a fund’s operating costs. Providing a financial safety net for a small fund can be important, but may weaken the fund managers’ incentives to ensure good investment decisions. Instruments to provide upside leverage multiply the financial benefits of success to the disproportionate advantage of venture capital funds and their investors. Murray’s simulations show that a leverage programme has a superior impact on the returns of the partners in a venture capital fund when compared with a guarantee scheme. Moreover, the leverage exclusively rewards successful investment activity and does not insulate management from the consequences of unsound decision making. This example illustrates how well designed public support can operate to amplify and target market dynamics.

Examples of financial support for start-ups, technology-oriented and fast-growth businesses

Australia – The Innovation Investment Fund The Innovation Investment Fund (IIF) programme was created in 1997

to support fund managers with expertise in early-stage venture capital investing to establish new innovation funds. Private sector investors co-invest with the government in these funds to assist early stage companies to commercialise the outcomes of Australia’s strong research capability. The Round 3 objectives of the programme are to:



a) Develop fund managers with experience in the early-stage venture capital industry.

b) Address capital and management constraints to encourage the development of new companies that commercialise research and development.

c) Establish in the medium-term a “revolving” or self-funding programme.

d) Develop a self-sustaining Australian early-stage venture capital industry.

The Government licenses managers to run the IIF funds with the managers being responsible for all investment decisions. Investment decisions are made on a commercial basis and with regard to the programme guidelines.

The programme is currently in its third round of investment – five fund managers were licensed under Round 1 (1998) and four under Round 2 (2000) – with AUD 200 million (USD 156 million) to be matched on at least a one-to-one basis with private sector funding over five years. Up to AUD 20 million (USD 16 million) of government funding is provided to each venture capital fund. Previously the government invested AUD 221 million (USD 172 million) in the two first rounds, matched by AUD 133 million from the private sector.

Since 2007 four new fund managers have been licensed under two tranches of Round 3 of the IIF program. A third tranche of Round 3 will be completed later in 2010 when new IIF fund managers are expected to be licensed.

A review of venture capital conducted in 2005 found that the IIF pro-gramme was effective. In August 2008 the Australian government reviewed the National Innovation System and in its publication Venturous Australia4

recommended that the IIF programme be maintained, and that a fourth round be implemented after 2010. As Round 1 IIF fund managers have now completed their contract, a full evaluation of the IIF programme is currently being undertaken. The evaluation report is expected in late 2010.

Australia – The Victorian State Government Smart SMEs Innovation Commercialisation Program (ICP)

The Smart SMEs Innovation Commercialisation Program (AUD 12 million; USD 9 million) is part of the Victorian Government’s Boosting Highly Innovative SMEs (BHIS) program (see Chapter 5 for details). The ICP aims to: i) facilitate the development of new, export-oriented technology busi-nesses, products and services; ii) encourage the commercial take-up of



advances in science and technology that will boost the international competi-tiveness of Victorian businesses; iii) facilitate access to appropriate finance for technology businesses; and iv) promote regional development by helping to build sustainable technology commercialisation capabilities throughout the State. The programme is rolled out by contracted partner organisations specialising in technology commercialisation. The TCP informs businesses of innovation options and opportunities and seeks to enhance skills and develop knowledge through internships, mentoring and coaching. The scheme also connects Australian businesses to each other, to R&D organisations and to global businesses.

Australia – Venture Capital Limited Partnerships (VCLP) program The aim of the Venture Capital Limited Partnership (VCLP) program is

to increase foreign investment in the Australian venture capital sector. Fund managers may be eligible for VCLP registration if seeking to raise a new venture capital fund of at least AUD 10 million (USD 7.8 million) for investment in Australian businesses with assets of up to AUD 250 million (USD 195 million). A fund registered as a VCLP is governed by the Venture Capital Act 2002 and the Income Tax Assessment Act 1997, and is required to self-assess its compliance. Registration entitles a fund to flow-through taxation treatment. Further, eligible foreign investors in the fund are exempt from capital gains tax on their share of any profits made by the fund.

Canada – The Business Development Bank of Canada Venture Capital The Business Development Bank of Canada (BDC) is Canada’s small

business bank and a financial institution wholly-owned by the Government of Canada, delivering financial and consulting services to Canadian small business, with a particular focus on technology and exporting. BDC provides Canadian businesses with flexible financing, venture capital and consulting services. It works with entrepreneurs in all industries, through all economic cycles and helps businesses in their development projects, both local and global. The Co-Vision Start-up Financing programme provides customised term financing for new businesses demonstrating long-term viability. For newly created firms the programme was created to help cover the costs of the purchase of fixed assets, working capital, marketing expenses and the purchase of a franchise. Funding is up to CAD 100 000.Among the eligibility criteria are that the businesses need to be in the start-up or early growth phase (first 12 months of sales), demonstrate realistic market and sales potential, with the entrepreneurs possessing experience in the chosen field.



Canada – Alberta Innovates venture capital fund Alberta Enterprise Corporation is equipped with CAD 100 million from

the Government of Alberta to co-invest in venture capital firms with demon-strated expertise in environmental, nano, life science or communications technologies. The corporation’s focus is on venture capital fund managers with a presence in Alberta (making investments in Alberta and abroad). Alberta Enterprise also supports firms with targeted initiatives to drive a high-quality deal flow as well as with initiatives connecting entrepreneurs and start-ups to investors.

Denmark – VækstfondenVækstfonden is a publicly supported investment fund that contributes to

the promotion of Danish business and trade. The fund’s mission is to strengthen development and renewal in the Danish economy by providing financing for promising projects in small and medium-size businesses. It co-invests with private partners.

Since 1992 it has funded over 3 500 Danish companies, committing more than DKK 65 billion (USD 12 billion). It is one of the largest investors in the Danish venture market with total commitments of DKK 3.7 billion (USD 690 million) in 20 venture funds.

Finland – Vigo Accelerator Programme In 2008 Tekes launched a funding scheme for young innovative start-

ups. In Spring 2009 the Ministry of Employment and the Economy, in co-operation with Tekes and Avera (a subsidiary of Finnvera Ltd.), established the Vigo Accelerator Programme. The Programme is co-ordinated by Tekes. Finnvera – a specialised publicly owned financing company – participates in the programme via its subsidiary, the venture capital company Seed Fund Vera.

The overall aim is to boost the Finnish venture capital market, facili-tating private risk capital investments in participating companies and attracting international venture capitalists to Finland. Public support for the Vigo programme for the first three years is estimated to be approximately EUR 45 million (USD 63 million) (including Tekes grants/loans and investments by Avera).



Finland – Finnish Industry Investment Finnish Industry Investment Ltd is a government-owned investment

company whose mission is to promote Finnish business, employment and economic growth through capital investment. Finnish Industry Investment invests in funds and directly in growth companies. Capital investments are needed for financing growth, and for internationalisation, spin-offs, major industrial investments, sectoral and corporate restructurings. Target firms operate in all sectors. The company co-invests with private investors, limiting its investment to at most half the capital invested.

The government-owned investment company Finnish Industry Invest-ment Ltd (FII) and Finnish pension companies have established a common fund, FoF Growth. This is a fund of funds, placing money in funds that in turn invest in growth companies. FoF Growth has EUR 135 million (USD 187 million) in capital, of which pension companies have committed 60% and FII 40%. FoF began operations in December 2008, making investments in venture capital funds operating in Finland, with investments possible in all sectors except real estate. The aim is to selectively invest sufficient capital in the most promising ventures. FoF aims to make annual investments of EUR 30-50 million (USD 42-69 million) in 2-4 funds.

The investments of Finnish Industry Investment amounted to EUR 640 mil-lion (USD 889 million) at the end of 2009, and directly affect some 439 com-panies.

Norway – Nationwide Seed Capital Scheme Through the Ministry of Trade and Industry, Innovation Norway operates

the Nationwide Seed Capital Scheme. This can lend up to NOK 667 million (USD 106 million) to investment funds located in four university cities. The Ministry also runs the Regional Seed Capital Scheme, which can lend up to NOK 700 million (USD 111 million) to investment funds oriented to start-ups in assisted areas. In addition, the government owns a fund-of-funds invest-ment company Argentum, established in 2001 with a NOK 6.5 billion (USD 1.1 billion) currently under management. Argentum aims to facilitate access to foreign venture capital and to encourage the development of the Norwegian equity market. All of its investments require majority private ownership (OECD, 2008). Furthermore, the government owns a venture investment firm, Investinor, with a capital base of NOK 2.2 billion. In addition to supplying risk capital, Investinor takes an active ownership role in portfolio companies, mainly in start-up companies.



Netherlands – The TechnoPartner SEED-facility and the TechnoPartner Label

TechnoPartner is a joint initiative of the Ministry of Economic Affairs and the Ministry of Education, Culture and Science. This facility aims to make it more attractive for venture capitalists to invest in technology-based start-ups. Technology-oriented start-ups able to obtain finance from venture capitalists are eligible for additional funding, on a loan basis, from the Seed facility. In practice, this means that these start-ups have to obtain just 50% of their required risk capital from a private funding source. Venture capitalists are reported to perceive a decreased risk of investing in techno-starters.

United Kingdom – Innovation Investment Fund In June 2009 the creation was announced of the United Kingdom

Innovation Investment Fund (UKIIF), aimed at investing in technology-based businesses with high growth potential. The new fund will focus on investing in growing small businesses, start ups and spin outs, in digital and life sciences, clean technology and advanced manufacturing. It will operate on a Fund of Funds structure (i.e. it will not invest directly in companies, but rather in a small number of specialised private sector technology funds that themselves have the expertise and track record to invest directly in tech-nology-based businesses).

The European Investment Fund (EIF) and Hermes Private Equity were each confirmed as managers for two separate funds of funds. Backed by a capital injection of GBP 150 million (USD 234 million) from the Depart-ment for Business, Innovation and Skills, the Department of Energy and Climate Change and the Department of Health, UKIIF has already attracted private investment to more than match this amount, with the total first closing of GBP 325million (USD 506 million).

Further private investment will be secured before the UKIIF Fund closes for investors in 2011. The ambition remains to create a billion-pound 15-year fund. The Hermes Environmental Innovation Fund is up and running with GBP 125 million (USD 195 million) to invest at first closing. The fund is aimed at increasing the efficient use of resources and is a major boost in the transition to a low-carbon economy. Hermes is investing GBP 75 million (USD 117 million) alongside GBP 50 million (USD 78 million) from UKIIF.

EIF will manage the second Fund-of-Funds, which will have resource of up to GBP 200 million (USD 311 million), to invest in fields such as life sciences, digital technology and advanced manufacturing. EIF is investing up to GBP 100 million (USD 155 million) in the fund alongside an investment of



GBP 100 million (USD 155 million) from the Government. These comple-mentary mandates address investment across all technology sectors and all stages of enterprise development.

United Kingdom – National Endowment for Science, Technology and the Arts

The National Endowment for Science, Technology and the Arts (NESTA) is an independent body with the broad mission of making the United Kingdom more innovative. Among other activities, NESTA combines capital investment with non-financial support to help the United Kingdom’s innovative early-stage companies turn their ideas into commercial success. It has strict investment criteria, and only works with companies considered to have high potential for growth and that are at a seed or start-up stage, and have the potential to attract syndicated support. In 2008-09 NESTA made five new direct investments totalling GBP 1.2 million (USD 1.9 million) and 14 follow-on investments totalling GBP 1.7 million (USD 2.6 million) in early stage companies, and now has a portfolio of 50 investee companies.

United Kingdom – Venture capital schemes Enterprise Capital Funds (ECFs) are run by a relatively new body, Capital

for Enterprise Limited (CfEL). The ECFs have been created to address a gap in the supply of small volumes of equity finance to SMEs. Government funding is employed alongside private sector investment to establish funds that target the equity gap. 10 such funds have been launched since 2006. On 1 April 2008, responsibility for the management of ECFs, along with BIS’s other equity funds and the Small Firms Loan Guarantee (SFLG), was transferred to (CfEL)]. This transfer was aimed at improving the selection and management of the funds, but did not change the nature of the funds or their policy objectives. Enterprise Capital funds replace previous programmes such as the Regional Venture Capital Funds.

A University Enterprise Capital Fund will be established to provide direct support for university innovation and spin-out companies.5 The fund will be worth as much as GBP 37.5 million (USD 58 million). The Govern-ment has pledged up to GBP 25million (USD 39 million). However, under Government rules for Enterprise Capital Funds the remaining third must come from the private sector. The fund will support universities seeking to commercialise their Intellectual Property, particularly patented inventions (www.bis.gov.uk).



Examples of support for researchers and innovative entrepreneurs to start their own business

Australia – Pre-Seed Fund The competitive Pre-Seed Fund (PSF) for universities and public sector

research agencies was created in 2002. It encourages the private sector to take a more active role in funding and managing the commercialisation of research from universities and Australian Government research agencies.

Four fund managers, Allen & Buckeridge, GBS Venture Partners, SciVentures and Starfish Ventures, were licensed in 2002. The maximum investment in any project or company is AUD 1 million (USD 780 000) (although there is a facility to allow this to be exceeded in particular circumstances). Fund managers acquire an equity interest in each project or company and provide management and technical advice.

The Pre-Seed Fund has AUD 104.1 million (USD 81.6 million) in capital, of which the Australian Government is providing AUD 72.7 million (USD 57 million). The rest comes from private sector investors, universities and public sector research agencies. Selected companies and projects must be established in Australia but not yet generating sales revenue. They must be engaged in the commercialisation of research and either be controlled by a university, a public sector research agency or a qualifying researcher or use intellectual property that is at least 50% owned by a university, a public sector research agency or a qualifying researcher.

A 2008 review for the Australian government, Venturous Australia,recommended that the government immediately establish a second group of Pre-Seed Funds. One major problem with the scheme had been the AUD 1 mil-lion (USD 780 000) cap imposed on investments by the Pre-Seed Funds. The policy intent of the cap was to ensure that investments remained at the very early stage. However, the cap had had the unintended consequence of stranding investee companies when follow-on funding had been difficult to find in a timely manner. The review therefore recommended that the current AUD 1 million (USD 780 000) cap per investee firm should be changed to a maximum AUD 1 million cap (USD 780 000) on the first tranche of invest-ment, recognising the high risk nature of the early stage of investment. The review also recommended that four new funds be established at a cost of AUD 100 million (USD 78 million) over 15 years.



Australia – Commercialisation Australia Commercialisation Australia is a Government initiative that started in

early 2010. It assists researchers, entrepreneurs and innovative firms turn their intellectual property into successful commercial ventures. It was designed to ensure assistance is adjusted to meet each applicant’s needs rather than trying to make their application fit the program. The support provided by the scheme is designed to help successful applicants through the commercialisation process. Assistance is tailored to the needs of each successful applicant and is structured around the key stages in the commercialisation pathway. It is a merit based, competitive assistance program that offers:

Skills and knowledge grants of up to AUD 50 000 (USD 39 000) to pay for specialist advice and services.

Experience executive grants of up to AUD 200 000 (USD 156 000) over two years to assist with the recruitment of experienced executives so as to help build the skills, knowledge and connections required to commercialise new ideas.

Proof of concept grants of AUD 50 000 (USD 39 000) to AUD 250 000 (USD 195 000) to test the commercial viability of a new product, process or service.

Early stage commercialisation repayable grants of AUD 250 000 (USD 195 000) to AUD 2 million (USD 1.6 million) to develop a new product, process or service to the stage where it can be taken to market.

Each successful applicant is assigned a case manager to guide them through the commercialisation process and facilitate access to experienced volunteer business mentors. Applications were being accepted at the time of writing by Commercialisation Australia and are to be assessed on a continu-ous basis.

Norway – The FORNY Programme The FORNY programme has provided funding for the development of

business ideas based on R&D results from universities and university colleges (FORNY has no reached the end of its programme life). FORNY aimed to start new companies and develop new technology that could be used by existing industry. The FORNY programme awarded funding at an early stage in the commercialisation process, long before seed funds or venture capital companies might typically engage. FORNY was administered jointly by the Research Council and Innovation Norway. It was designed to commercialise



research-based business ideas with market potential. To achieve this, the programme:

Helped research institutions establish professional systems and organisations for commercialisation of R&D results.

Made competent support available for researchers with research-based business ideas.

Encouraged and contributed to increased co-operation between research communities, entrepreneurs, investors, industry and public authorities.

The programme offered four kinds of funding:

Funding activities to make researchers and research institutions aware of the commercial potential of research results.

Funding of commercialisation projects, i.e. the process of estab-lishing a new company or a licence agreement.

Funding of proof of concept.

Scholarships for researchers which will enable them to focus on the commercialisation project instead of their regular work.

In addition, bonuses were offered for successful completion of commerci-alisation projects. Funding can cover a maximum of 50% of the costs related to the various activities. Following evaluation of FORNY, a new successor programme is envisaged (details of which were not available to be included in this report).

Denmark – Proof of Concept Scheme The programme started in 2007 and aims to strengthen the further

development and documentation of inventions made by researchers in public research institutions. It is not sector specific. The funding is intended to allow researchers to concentrate on the further development of their inven-tions and to reduce their education and research-related burden. Priority is given to strengthening commercialisation of promising research results from public research institutions.

Since the programme was initiated two proof-of-concept consortia have been established: one in East Denmark that includes the University of Copenhagen, the capital region and the Statens Serium Institute, and one in the West of Denmark that includes Aarhus University, Aalborg University and the University of Southern Denmark. Every proof of concept consor-tium manages a portfolio of projects.



Funding takes the form of grants and co-financing. The overall budget is EUR 8 million (USD 11 million). From 2011 it is expected that the instru-ment will be part of a larger policy instrument.

Netherlands – The TechnoPartner Knowledge Exploitation funding programme (SKE)

This programme addresses spin-offs and new independent start-ups. A pre-seed facility provides the opportunity to dedicate more time and effort on the pre-start-up phase. A patent facility makes it possible for the know-ledge institution to professionalise the internal patent policy. Large companies and knowledge institutes can, as a consortium, obtain 50% funding for initiatives that create technological start-ups based on research programmes.

Support to knowledge transfer through technology-related collaboration and networks

Australia – Enterprise Connect Introduced in 2008, the Australian Government has committed

AUD 251 million (USD 196 million) over five years for the Enterprise Connect initiative. The initiative will provide small and medium-sized enterprises (SMEs) with better access to new ideas, knowledge and tech-nologies, to enable businesses to become more innovative, efficient and competitive and to lift productivity across Australian industry.

The Enterprise Connect network comprises a number of Manufacturing and Innovation Centres located around Australia. Firms can receive a free Business Review with skilled and experienced Business Advisers and then obtain matching funding – up to AUD 20 000 (USD 16 000) – through the Tailored Advisory Service to implement changes identified by the Business Review. Enterprise Connect provides a range of other services that transfer business and technology related knowledge to SMEs.

The Manufacturing Centres are located in Sydney, Melbourne, Brisbane, Perth, Adelaide and Burnie. There are six Innovation Centres located across the country: the Remote Enterprise Centre, Innovative Regions Centre, Mining Technology Innovation Centre, Creative Industries Innovation Centre, Clean Energy Innovation Centre, Defence Industry Innovation Centre. The initiative employs around 100 Business Advisers nationally.

To be elegible for support firms must operate in one of the following sectors or regions: manufacturing or manufacturing related services, mining technology, defence, clean energy, or creative industries. Firms also have certain turnover requirements. For example, in the manufacturing, manufac-



turing-related services, defence and mining technology sectors firms must have turned over more than AUD 2 million (USD 1.56 million) and under AUD 100 million (USD 78 million) in the current financial year or one of the two preceding financial years (if located in Melbourne, Sydney, Adelaide, Perth or Brisbane).

Denmark – Innovation Consortiums The aim of Innovation Consortiums is to strengthen co-operation

between companies, public research institutions and technological service centres in order to develop new generic technology platforms for product and service development. The programme seeks to increase the incidence of innovation with a potential benefit for enterprises and Danish society as a whole. The programme seeks to achieve its goals by enhancing R&D resources in industry, facilitating the commercialisation of research results and assisting the transfer of knowledge to Danish companies, especially SMEs. The programme started in 2002 and covers all sectors. It is managed by the Danish Council for Technology and Innovation.

Enterprises must contribute 50% of the funding. Typically a consortium has a total budget of EUR 2.5 million (USD 3.9 million) EUR 5.5 million (USD 8.6 million) and lasts 3-4 years. It has to include at least two companies, one public research institution and one counselling and knowledge transfer organisation.

Netherlands – Innovation Vouchers Scheme Introduced in 2006 The Innovation Vouchers Scheme is an initiative of

the Dutch Ministry of Economic Affairs. The scheme arose in response to the widely held view that knowledge sharing between SMEs and research institutions in the Netherlands was inadequate. Accordingly, the scheme aims to stimulate interaction between the knowledge suppliers and SMEs. Although the scheme has a relatively small budget (around EUR 26 million (USD 36 million) in 2009), the leverage effects appear to be significant.

There are two types of innovation voucher available, a small voucher and a large voucher.

The small voucher is worth EUR 2 500 (USD 3 500). Each SME may get a “small” voucher only once. The objective is to encourage the enterprise to approach a knowledge institution (the voucher being redeemed against services of equivalent value from the research body). A total of 3 000 small vouchers are available.



The large voucher is worth EUR 7 500 (USD 10 400). To qualify for these vouchers an enterprise must contribute one-third of the total project costs. The government will provide an amount not exceeding EUR 5000 (USD 6 900). A large voucher allows the enterprise to submit a more extensive request to a knowledge institution. Enter-prises may obtain one large voucher each year. A total of 3 000 large vouchers are available.

Eligibility criteria for the Innovation Voucher programme are broad, and the scheme’s administrative costs are modest. An innovation voucher can be used for a project that transfers knowledge that is new to the enterprise. The enterprise may use the new knowledge to innovate a product, production process or service. A requirement is that the results must largely benefit the Dutch economy. A knowledge question cannot be answered simply by submitting goods – like software – or giving training courses without the knowledge institution performing any further work. Similarly, the scheme excludes activities for the purpose of promoting and selling products or processes, such as the design and production of advertising material and other sales activities. The costs of internships for students of knowledge institutions are also excluded.

SME can apply at SenterNovem for a voucher. After receiving the voucher the SME can submit a question to a knowledge institution, on presentation of the voucher. After answering the question the knowledge institution in turn surrenders the voucher to SenterNovem to claim a subsidy for the costs it incurred in answering the question. A voucher is valid for twelve months. Therefore, the knowledge institution must present the voucher to SenterNovem for payment within 12 months of the date it was issued to the enterprise.

In the case of a large voucher a subsidy is also obtainable for a project for which enterprises have bundled several innovation vouchers. The maximum number of enterprises that may submit a question jointly is ten, and they are required collectively to surrender all the received vouchers to a knowledge institution. Vouchers are not tradable because they are issued in the enter-prise’s name.

SMEs can also apply for a voucher in order to refund the costs for a patent. Evaluating the scheme Cornet et al. (2007) find evidence for positive effects

on production processes, but no evidence for effects on product or process innovation. More SMEs commission a study to a public research organisation, but this is a once-only outcome. SMEs do not co-operate more closely with public research bodies in the post-programme period. The authors however caution that half of the SMEs with a voucher for 2004 did not respond to their survey, for which reason conclusions need to be drawn with care.

2. N

ON

R&

D-B

ASE

D P

UB

LIC

SU

PPO

RT

FOR

BU

SIN

ESS

INN

OV

ATI

ON

– 6

7

BU

SIN

ESS

INN

OV

ATI

ON

PO

LIC

IES:

SEL

ECTE

D C

OU

NTR

Y C

OM

PAR

ISO

NS

– ©

OEC

D 2

011

Tabl

e 3.

1. O

verv

iew

of i

nnov

atio

n vo

uche

r sch

emes

in E

urop

e

Stat

ed ra

tiona

les/

goal

s Ta

rget

s of

pol

icy

actio

n

Country/ region

Stimulate/raise level of demand for R&D in firms

Support R&D performing institutions

Promote collaboration

Make public R&D more responsive to demand signals

Match supply of and demand for knowledge in the same region

Eligibility

Eligible R&D/knowledge partners

Face value of voucher

Allocation and other conditions

Co-funding by company

Aus

tria

In

nova

tion

Vou

cher

A

ustri

a

Aus

trian

SM

Es

not i

n an

y co

ntra

ct w

ith th

e se

lect

ed p

artn

er

durin

g th

e pa

st fi

ve

year

s an

d w

hich

ha

ve re

ceiv

ed le

ss

than

a c

erta

in

amou

nt o

f pub

lic

fund

ing

Uni

vers

ities

or

non-

univ

ersi

ty

publ

ic re

sear

ch

inst

itute

s fro

m

Aus

tria,

any

EU

m

embe

r sta

te o

r fro

m a

ny

exte

rnal

cou

ntry

< E

UR

5

000

Max

. one

vo

uche

r per

ye

ar p

er

com

pany

. A

lloca

tion

on a

fir

st c

ome,

firs

t se

rved

(FC

FS)

basi

s

No

info

Bel

gium

W

allo

nia

Tech

nolo

gy

Vou

cher

s

Wal

loni

an S

ME

s Th

e su

ppor

ted

serv

ices

can

not

alre

ady

be s

ubje

ct to

pu

blic

fund

ing

One

of t

he 2

2 ac

cred

ited

Wal

loni

an

rese

arch

cen

tres

or o

ne o

f the

13

rese

arch

cen

tres

asso

ciat

ed to

the

Fren

ch-s

peak

ing

Hau

tes-

écol

es o

f B

elgi

um

EU

R50

0 M

axim

um

40 te

chno

logy

vo

uche

rs p

er

com

pany

per

ye

ar

SM

E m

ust c

o-fu

nd 2

5% o

f the

va

lue

of th

e vo

uche

r

68 –

2. N

ON

R&

D-B

ASE

D P

UB

LIC

SU

PPO

RT

FOR

BU

SIN

ESS

INN

OV

ATI

ON

BU

SIN

ESS

INN

OV

ATI

ON

PO

LIC

IES:

SEL

ECTE

D C

OU

NTR

Y C

OM

PAR

ISO

NS

– ©

OEC

D 2

011

Stat

ed ra

tiona

les/

goal

s Ta

rget

s of

pol

icy

actio

n Country/ region






Eligibility





Cyp

rus*

In

nova

tion

vouc

hers

Cyp

riot S

ME

s A

ll pu

blic

or

priv

ate

orga

nisa

tions

in

Cyp

rus

doin

g re

sear

ch a

nd/o

r te

chno

logy

tra

nsfe

r act

ivite

s

EU

R5

000

One

vou

cher

pe

r com

pany

. A

pplic

atio

ns o

n FC

FS b

asis

No

info

Den

mar

k K

now

ledg

e vo

uche

r (sm

all

inno

vatio

n pr

ojec

ts)

Dan

ish

SM

Es

that

ha

ve n

ever

co

llabo

rate

d w

ith th

e se

lect

ed p

artn

er

A p

ublic

re

sear

ch

orga

nisa

tion

or a

m

embe

r of t

he

Adv

ance

d Te

chno

logy

G

roup

(GTS

In

stitu

tion)

EU

R6

670-

13

330

No

info

S

ME

mus

t m

eet a

t lea

st

50%

of t

he c

ost

of th

e pr

ojec

t

*Foo

tnot

e by

Tur

key:

The

info

rmat

ion

in th

is d

ocum

ent w

ith re

fere

nce

to “C

ypru

s” re

late

s to

the

sout

hern

par

t of t

he Is

land

. The

re is

no

sing

le a

utho

rity

repr

esen

ting

both

Tu

rkis

h an

d G

reek

Cyp

riot p

eopl

e on

the

Isla

nd. T

urke

y re

cogn

izes

the

Turk

ish

Rep

ublic

of N

orth

ern

Cyp

rus

(TR

NC

). U

ntil

a la

stin

g an

d eq

uita

ble

solu

tion

is fo

und

with

in

the

cont

ext o

f Uni

ted

Nat

ions

, Tur

key

shal

l pre

serv

e its

pos

ition

con

cern

ing

the

“Cyp

rus

issu

e”.

*Foo

tnot

e by

all

the

Eur

opea

n U

nion

Mem

ber S

tate

s of

the

OE

CD

and

the

Eur

opea

n C

omm

issi

on:

The

Rep

ublic

of C

ypru

s is

rec

ogni

zed

by a

ll m

embe

rs o

f the

Uni

ted

Nat

ions

with

the

exce

ptio

n of

Tur

key.

The

info

rmat

ion

in th

is d

ocum

ent r

elat

es to

the

area

und

er th

e ef

fect

ive

cont

rol o

f the

Gov

ernm

ent o

f the

Rep

ublic

of C

ypru

s.”

2. N

ON

R&

D-B

ASE

D P

UB

LIC

SU

PPO

RT

FOR

BU

SIN

ESS

INN

OV

ATI

ON

– 6

9

BU

SIN

ESS

INN

OV

ATI

ON

PO

LIC

IES:

SEL

ECTE

D C

OU

NTR

Y C

OM

PAR

ISO

NS

– ©

OEC

D 2

011

Stat

ed ra

tiona

les/

goal

s Ta

rget

s of

pol

icy

actio

n Country/ region






Eligibility





Den

mar

k R

esea

rch

vouc

her f

or

SM

Es

Dan

ish

SM

Es

not i

n re

ceip

t of o

ther

pub

lic

fund

ing.

P

roje

ct m

ust m

eet

Fras

cati

mau

nal

defin

ition

of R

&D

Dan

ish

univ

ersi

ties,

R

&D

inte

nsiv

e D

anis

h ho

spita

ls, G

TS-

Inst

itute

s or

ot

her r

esea

rch

inst

itutio

n,

incl

udin

g eq

uiva

lent

ov

erse

as

rese

arch

in

stitu

tions

< E

UR

0.

2m

Par

tner

ship

of

at le

ast o

ne

SM

E a

nd a

t le

ast o

ne

rese

arch

or

gani

satio

n

SM

E m

ust

mee

t 50%

of

the

tota

l co-

fund

ing

and

the

rese

arch

in

stitu

tion

at

leas

t 25%

Gre

ece

Inno

vatio

n vo

uche

rs fo

r S

ME

s

One

or m

ore

Gre

ek

SM

Es

of th

e m

anuf

actu

ring

sect

or, s

oftw

are

indu

stry

and

re

sear

ch a

nd

deve

lopm

ent f

irms

Gre

ek u

nive

rsi-

ties,

tech

no-

logi

cal c

olle

ges,

re

sear

ch c

entre

s an

d in

stitu

tes

and

“sec

tora

l” su

pplie

rs o

f kn

owle

dge-

inte

nsiv

e se

rvic

es “o

f hig

h ad

ded

valu

e”

EU

R7

000

but t

wo

SM

Es

may

co

mbi

ne

thei

r vo

uche

rs

One

vou

cher

pe

r com

pany

al

loca

ted

on a

FC

FS b

asis

w

hile

fund

s re

mai

n av

aila

ble

No

info

70 –

2. N

ON

R&

D-B

ASE

D P

UB

LIC

SU

PPO

RT

FOR

BU

SIN

ESS

INN

OV

ATI

ON

BU

SIN

ESS

INN

OV

ATI

ON

PO

LIC

IES:

SEL

ECTE

D C

OU

NTR

Y C

OM

PAR

ISO

NS

– ©

OEC

D 2

011

Stat

ed ra

tiona

les/

goal

s Ta

rget

s of

pol

icy

actio

n Country/ region






Eligibility





Hun

gary

IN

NO

CS

EK

K

(inno

vatio

n vo

uche

r)

Hung

arian

SM

Es

regis

tere

d or

hav

ing a

br

anch

offic

e in

the

resp

ectiv

e re

gion.

El

igible

acti

vities

: pro

ject

conc

ept a

ppra

isal;

prod

uct d

evelo

pmen

t; pr

oces

s inn

ovat

ion; a

nd

othe

r inno

vatio

n de

velop

-m

ent s

ervic

es

No in

form

ation

A

rang

e fro

m E

UR

12 0

00

(pro

ject

appr

aisal)

to

EUR

12

0 00

0 (p

rodu

ct de

velop

men

t)

Appli

catio

ns m

ust

be su

mbit

ted

in on

e re

gion

only

for

one

of th

e lis

ted

activ

ities.

Not r

equir

ed –

vo

uche

r can

mee

t up

to 1

00%

of

proje

ct co

sts

Net

herla

nds

Inno

vatio

n vo

uche

rs

For s

mall

vouc

hers

, Du

tch S

MEs

not

pr

eviou

sly in

rece

ipt o

f a

small

vouc

her.

For la

rge

vouc

hers

, Du

tch S

MEs

(Sem

i-)pu

blic

know

ledge

ins

titute

s; lar

ge

com

panie

s with

R&

D ex

pend

iture

s th

at e

xcee

d €

60

millio

n p.

a.; o

ther

EU

pub

lic

know

ledge

ins

titute

s

EUR

2 50

0 (s

mall

) or

EUR

7 50

0 (la

rge)

3 50

0 of

eac

h typ

e pe

r yea

r. O

ne

small

vouc

her p

er

SME

per y

ear.

Alloc

aion

on a

FC

FS b

asis.

SM

Es ca

n co

mbin

e lar

ge

vouc

hers

for

colle

ctive

pro

jects.

For la

rge

vouc

hers

, SM

E m

ust m

eet a

t leas

t on

e-th

ird o

f the

face

value

.

Port

ugal

S

ME

ski

lls

supp

ort

syst

em –

in

nova

tion

vouc

her

Portu

gese

SM

Es. T

hree

-ye

ar lim

it of E

UR 2

00

000

per c

ompa

ny

No in

form

ation

<

EUR

25 0

00

Prior

ity g

iven

to

small

est fi

rms.

Afte

r size

, all

ocat

ion o

f FCF

S ba

sis.

No in

form

ation

Sou

rce:

Fla

naga

n et

al.

(201

0).



In part as a reaction to this scheme, there has been an increasing interest in voucher schemes across Europe. National schemes are in place in France, Poland, Greece, Slovenia, Ireland, Denmark, Austria, Netherlands, Portugal and Switzerland. Regional voucher schemes exist in Belgium, Germany and the United Kingdom, managed by regional authorities or universities (for instance, Aston University issues the vouchers in the West Midlands, in the United Kingdom). Table 3.1 presents innovation voucher schemes in a selection of European countries, showing their stated rationales, eligibility criteria and implementation modalities.

United Kingdom – Knowledge Transfer Networks A Knowledge Transfer Network (KTN) is an over-arching national

network in a specific field of technology or business which brings together experts from business, universities, finance and technology organisations. KTNs are funded by government, industry and academia. They bring together diverse organisations and provide activities and initiatives that promote the exchange of knowledge and the stimulation of innovation in these com-munities.

There are currently 24 KTNs in operation. During 2008 a review of the KTNs was carried out to assess their current effectiveness and scope.6 The comprehensive review, which obtained views from 2100 KTN users and R&D-intensive businesses, strongly confirmed the value of the networks. 75% of business respondents rated KTN services as effective or highly effective. Over 50% have developed, or are developing, new R&D or commercial relationships with people met through a KTN. 25% have made changes to their innovation activities as a result of their engagement. The most highly rated functions of KTNs, according to the survey, are monitoring and reporting on technologies, technology applications and markets, networking opportunities, and identifying and prioritising key innovation-related issues and challenges.

United Kingdom – Knowledge Transfer Partnerships Knowledge Transfer Partnerships (KTPs) seeks to encourage colla-

boration between businesses and universities, enabling companies to obtain knowledge, technology or skills which they consider to be of strategic importance. These services are obtained from either the further/higher education sector or from a research and technology organisation. The scheme is one of the United Kingdom’s most important technology transfer mechanisms, with typically over 1000 projects underway at any given time. The knowledge sought is embedded in the company through a project or projects undertaken by an individual recruited for the purpose (in other words, each KTP has three parties, a company – which can include public



bodies or voluntary agencies, a university, higher education institute or research organisation, and an associate – typically a recently qualified graduate). KTPs enable companies to identify the most appropriate source for the knowledge or capability they are seeking from within the United Kingdom’s knowledge base (universities, colleges or research organisations).

A company entering into a KTP contributes between 40 and 67% of the project cost, with the government contributing the remainder. Average annual project costs are approximately GBP 60 000 (USD 93 000). The KTP Annual Report for 2006/07 indicates that on average companies reported an annual increase in profit before tax, after project completion, of GBP 229 000 (USD 357 000). Not only does the schemes seek to help transfer knowledge, technology, technical and business skills to the companies involved, it also aims to stimulate business-relevant research and training in the relevant academic bodies, and enhance the business and technical skills of a recently qualified graduate (indeed, a large share of participating graduates find employment in their partner companies).

The KTP scheme is supported by regional advisors working across the United Kingdom. At participating universities, designated staff liaises with partner companies. A web portal, www.ktponline.org, provides information on best practice case studies and other relevant information.

United Kingdom – Collaborative Research and Development Collaborative Research and Development is designed to assist the

industrial and research communities to work together on R&D projects in strategically important areas of science, engineering and technology. The Technology Strategy Board organises and sponsors competitions for funding of collaborative R&D projects. These have been held since 2004, and by June 2007 a portfolio of over 600 projects was being supported with a combined business and government investment in excess of GBP 1 billion (USD 1.5 billion). In the past year the scope of the collaborative R&D competitions has been expanded to support large projects as well as smaller projects approved within shorter timescales. Information on the Technology Strategy Board’s funding of competitions is available on the competitions page of the Technology Strategy Board website. Applicants apply on line. (www.innovateuk.org/competitions.ashx).

Canada – Networks of Centres of Excellence Programs The Networks of Centres of Excellence (NCE) have been running for

more than 20 years. They provide opportunities for Canadian researchers and students to work with sector partners and accelerate the exchange of knowledge and transfer of technological innovations. The NCE Secretariat



fosters multi-disciplinary, multi-sectoral partnerships between universities, industry, government and non-governmental organisations. It supports academic research and the commercialisation of products and ideas. The three NCE programmes are: Networks of Centres of Excellence (NCE); Centres of Excellence for Commercialisation and Research (CECR); and, Business-Led Networks of Centres of Excellence (BL-NCE).

In 2007–08, the Government of Canada expanded on the success of the Networks of Centres of Excellence by investing CAD 46 million over four years in a new NCE program initiative called the Business-Led Networks of Centres of Excellence (BL-NCEs). The BL-NCEs are headed by industrial consortia. They are designed to support increased private sector investment in Canadian research, support the training of skilled researchers and accelerate the timeline for the transfer of ideas from the laboratory to commercialised products and services.

Selected through a rigorous competitive process, the BL-NCEs focus on developing innovative tools for drug discovery, nanotechnology-enhanced forestry products, next-generation aviation technologies and sustainability challenges relating to hydrocarbons.

The Business-Led NCE program focuses on increasing private sector investment in Canadian research and optimising the timeline between research and commercialisation stages. The BL-NCEs differ from other NCEs in a number of ways. BL-NCEs must carry out their proposed activities in a shorter timeframe (four-years) and are not eligible for a second term of funding. Their activities are not fully supported by a BL-NCE grant which means that BL-NCEs must seek additional funding from their private sector stakeholders. One of the most significant differences between the two initiatives is that BL-NCEs are led by private sector participants whereas the NCEs are led by academics.

Despite their differences, the two initiatives share common goals, including: the focus on S&T priority areas, training and development of high quality personnel, the mobilisation of Canadian research talent in academic, public and private sectors and the dissemination of benefits to Canadian society more broadly.

Canada – programmes for partnerships and innovation As part of its strategy for partnerships and innovation, the National

Sciences and Engineering Research Council of Canada (NSERC) has a whole range of programmes dedicated to collaborative research (i.e. funding academic research working with industry partners) to accelerate and transfer R&D in priority areas for industry and stimulate the innovation process (Box 3.1).



Box 3.1. Programmes fostering partnerships for innovation under the Canadian National Sciences and Engineering Research Council

Engage Grants (EG) programme

The EG Program is intended to give companies that operate from a Canadian base access to knowledge and expertise available at Canadian universities. It aims to foster the development of new research partnerships between academic researchers and companies that have never collaborated before, through short-term research and development projects that address a company-specific issue. The EG Program supports well-defined projects undertaken by eligible university researchers and their private-sector partners. A maximum grant of CAD 25 000 over a period not exceeding six months, will be awarded to the academic researcher to cover the direct project costs associated with the research activities needed to address the identified problem. Projects can be at any point in the R&D spectrum consistent with the university’s research, training and technology transfer mandate. Eligible collaborations include focused projects with specific short-term objectives. The project must be scientifically sound and technically feasible. It must be aimed at solving a company-specific problem through the generation of new knowledge, or the application of existing knowledge in an innovative manner.

Strategic Project Grants (SPG) programme

The SPG has been introduced to increase participation of Canadian companies and/or government-based organisations that are suitably positioned to strengthen Canada’s economy, industrial base, create employment and/or positively impact Canadian public policy within the next ten years. SPG funds early-stage project research in targeted areas for a period of one to three years. Partnerships in foci outside of natural sciences and engineering and international collaborations are also encouraged. For the approval of funding, NSERC requires that the project support the goals of the SPG program and belong to at least one of the seven target areas, with clearly defined objectives and scope and involve collaboration with at least one academic researcher and one supporting organisation.

Idea to Innovation (I2I) programme

I2I’s primary objective is to expedite the pre-competitive development of new technology and encourage its transfer from the research and development (R&D) stage to Canadian companies. I2I provides funds to post-secondary faculty members as support for programs and projects in their initial stages of technology validation and market connection. The R&D activities that are eligible for this grant include, refining and implementing designs, verifying application, conducting field studies, preparing demonstrations, building engineering prototypes and performing beta trials. Eligibility criteria necessitate that all projects and proposals must specify a technology transfer plan. Four different funding options, with corresponding time limitations, are available to applicants.

…/…



Box 3.1. Programmes fostering partnerships for innovation under the Canadian National Sciences and Engineering Research Council (continued)

These stages and their time limitations are as follows:

Market assessment: NSERC shares costs of an independent and professional market study with the institutions for a period of up to twelve months (maximum CAD 15 000)

Phase I (reduction-to-practice stage): NSERC provides the direct costs of research for a period of up to twelve months (maximum CAD 125 000)

Phase II (technology enhancement) falls into two categories. Phase IIA (early-stage investment partner): NSERC provides up to two-thirds of the costs of the project from six to eighteen months (funding requested cannot exceed CAD 125 000 per year). Phase IIB (Partnership with a Canadian company): NSERC equally shares the funding cost with the company (funding requested cannot exceed CAD 350 000).

Support for the creation of innovation hubs and centres of excellence

Canada – Centres of Excellence for Commercialisation and Research (CECR) program

In 2007, the Government of Canada invested approximately CAD 285 mil-lion to create the CECR program. To date, 17 CECR initiatives have been launched with funding for five years. A third competition was launched in 2010-11 that is expected to support four additional centres. Funded applica-tions to this program are chosen by international peer review and with the advice of the private sector.

The goal of the CECR programme is to create internationally recognised centres of commercialisation and research expertise in four priority areas in order to deliver economic, social and environmental benefits to Canadians. These priority areas include: environmental science and technologies; natural resources and energy; health and related life sciences and techno-logies; and information and communications technologies. In the context of the programme, commercialisation is defined as everything a firm does that transforms knowledge and technology into new goods, processes or services to satisfy market demands. Each Centre brings together people, services and infrastructure to maximise the benefits of the government’s investment in skills and research and to encourage private sector investment.



Expected commercialisation-related benefits include: i) create, grow and retain companies in Canada able to capture new markets with breakthrough innovations; ii) accelerate the commercialisation of leading edge technolo-gies, goods and services in priority areas where Canada can significantly advance its competitive advantage; iii) draw on existing commercialisation strength, infrastructure, networks and funding sources to enhance capacity; and iv) attract investment (including foreign direct investment and venture capital).

Norway – Centres for Research-based Innovation The Centres for Research-based Innovation (CRI) scheme targets the

research-intensive segment of the Norwegian business sector. The main objective is to enhance innovative capability by forging alliances between research-intensive enterprises and prominent research groups for long-term research projects. Participation by foreign companies or research institutions is welcome. Fourteen centres were initiated in 2006. A centre is financed for five years, with the possibility of a three year extension. Co-financing by government authorities and private businesses is a prerequisite. Several businesses and research institutions can be accommodated in a centre. The total budget allocation from the Research Council for the 14 existing centres under the CRI scheme amounts to NOK 1.12 billion (USD 178 million) over an eight year period. Each centre receives an allocation from the Research Council of roughly NOK 10 million a year. The host institution and partners must contribute at least as much as the Research Council. The Research Council’s share is finance by yields on the Fund for Research and Innovation (OECD, 2008).7 A recent call for 5-6 new centres had a deadline of September 2010, and the creation of these new centres is expected to be confirmed in December 2010.

Denmark – The Danish National Research Foundation Since 1991 the Danish National Research Foundation has set up and

funded a number of research centres. The strategy of the Foundation is to invest in long-term research endeavours, mainly Centres of Excellence that run for 5-10 year periods. The Foundation continuously supports 50 Centres of Excellence with an average annual budget up to DKK 400 million (USD 74 million). The centres cover a wide array of scientific and academic themes. Independent groups of scientists form these centres, which are headed by a centre leader. A board consisting of a chairman and eight members governs the National Research Foundation. The Minister of Science, Technology and Innovation appoints the board according to recommendations from competent bodies.



Denmark – Competence and Innovation Network A competence and innovation network is supported by the Danish

Council for Technology and Innovation (DCTI) under the Ministry for Science, Technology and Innovation. A network is here conceived as a flexible framework for collaboration between companies, research institutions and non-profit advisory/knowledge dissemination parties. The competence and innovation networks carry out a wide range of activities that aim to share, develop and disseminate knowledge in a way that creates value for businesses. There are nine core network services:

Bridge-building activities and meeting places

Themed networks

Matchmaking

Idea generation

Conferences, seminars, etc.

Partnership projects

Pre-projects

R&D and innovation projects

Business- to-business partnerships

Knowledge information and communication

Consultation

Skills development

The annual budget of the total network programme is approximately EUR 10 million. The annual budget of an average network is approximately EUR 0.9 million, of which the 40% is financed by the network programme of the DCTI, at least 40% is financed by companies and the rest by regional sources, universities, technological and research institutes and the European Union. To 2010 there were 23 networks under the programme.

Finland – Strategic Centres for Science, Technology and Innovation The Strategic Centres for Science, Technology and Innovation (SHOK)

have been set up recently. In these Centres, companies and research units work in close co-operation, carrying out jointly defined research. The research aims to meet the needs of Finnish industry and society. Between 2008 and 2009, six centres had begun operations. Each centre is co-ordinated by a not-for-profit company, jointly owned by the shareholders



(including companies, research organisations, funding agencies and different interest groups). In addition to centres’ shareholders, public funding organisations have made a commitment to providing funding for the centres over the long term. In each Strategic Centre, some EUR 40-60 million are invested in research annually. Tekes is committed to the operations of the Strategic Centres and their development and will finance the centres’ research programmes and projects. It is not however a shareholder of the centres or a contracting party.

The Strategic Centres represent a market and user-driven innovation policy approach, allowing industry and other interest groups to participate in decision-making related to future R&D agendas. European Commission (2009).

Examples of technology incubators

Policy makers have turned to business incubation to meet a wide range of policy goals – from raising enterprise birth rates, to commercialising university research, to expanding the supply of infrastructure. Technology incubators typically possess features that differ somewhat from mixed-use incubators. They are frequently affiliated with a University and often have entry criteria focusing on businesses with high growth potential. They can also be more expensive to establish and run on account of specialised facili-ties and higher staffing costs. Their service offering may also include a greater emphasis on services related to intellectual property. Some maintain close links to networks of informal equity investors, so-called business angels.

What potentially makes incubation a cost effective policy tool is that information sharing and synergies can be realised among the firms that use the incubator – the tenants. Such information sharing is not expected to include proprietary knowledge, of course, but rather concerns day-to-day problems that typically affect small-fast-growth firms, such as the challenge of managing cash-flow. Incubation can also lower the unit cost of delivering services to co-located firms, as against firms that are geographically dispersed. Chapter 5 considers some of the metrics used and challenges involved in evaluating incubation programmes.

United Kingdom – Oxford Innovation Ltd8

Oxford Innovation (OI) supports innovative start-up and growing companies, and also provides services to central and regional government organisations that wish to encourage innovative enterprises. OI began as a commercial subsidiary of The Oxford Trust, a private charitable foundation with a mission to encourage the study and application of science and



technology. In 2001, OI secured external investment and has been operating since then as an independent company. OI is now the leading operator of Innovation Centres in the United Kingdom, managing 12 centres used by over 300 technology-based and creative companies.

The Innovation Centres provide office and laboratory premises on flexible terms in multiple locations. OI also offers a virtual office service, Oxiflex, for companies not yet ready to move into premises. Some of the Innovation Centres attract businesses from diverse sectors, while others focus on specialist fields such as advanced engineering and materials, motor sport, ICT and healthcare. The business support services offered range from on-site business advice and help with fundraising and public relations to networking events and membership of an online community of entre-preneurs. OI also operates four business angel networks, catering to entre-preneurs from across the United Kingdom. The most well-established of these is the Oxfordshire Investment Opportunity Network.

OI also manages delivery of the Innovation Advisory Service (IAS), an initiative to help companies in the south east of England to enhance their innovation and R&D capacities. A key objective of the IAS is to link companies into a regional knowledge base of universities and research institutes.

Finland – Jyvaskyla Science Park Incubation facilities were first established at the Jyvaskyla Science Park

in 1992. The incubation services offered include pre-incubator and post incubator phases. During the pre-incubator phase the entrepreneur prepares a business plan for the company. The entrepreneur is assisted with access to tested budget and production planning tools used by the incubator. Companies that are successful in the pre-incubation phase can stay in the incubator for up to two years. Individual counselling is offered for each company during the incubation period. Mentoring services are provided in the post-incubation phase. In addition to physical incubator premises, the Jyvaskyla Regional Development Company has developed a company “clinic” tailored to the needs of firms working in knowledge-intensive business services. In part these services aim at helping firms to better plan and manage their growth (OECD, 2006).



Notes

1. While the subject of business angel investment is not addressed in the programme descriptions in this chapter, there are a number of features of angel investment that merit mention. Business angels generally invest in the early stages of enterprise start-up, often in firms that are not yet ready for formal venture capital. Informal equity investors are often successful entrepreneurs themselves. For example, in the United Kingdom, previous business ventures represent the main source of wealth for around two-thirds of business angels. Such investors therefore bring significant knowledge and experience to the investee firm. And the time horizons over which informal investments are allowed to mature are often greater than in the formal industry. Business angels are also not averse to investing in technology-based firms (Mason and Harrison, 1997). Furthermore, business angels are geographically dispersed, not being limited to those locations where formal venture capital firms tend to cluster. Informal equity investors can also play a role in raising the quality of investment proposals put to early-stage venture capital funds. This is because business angels screen projects and, through intensive interaction with entre-preneurs, help to resolve design and presentational shortcomings that might deter formal investors. In this way business angels help to address important demand-side constraints in the formal venture capital market (Mason and Harrison, 2002). Public policy towards business angels is justified, in principle, by obstacles to the efficient functioning of the informal equity market. An information barrier may exist in this market if business angels are reluctant to publicise their willingness to invest and entrepreneurs are disinclined to reveal innovative ideas. Furthermore, the time required to search for and appraise potential investments is considerable, while many business angels are – as noted – active entrepreneurs investing on a part-time basis (Mason and Harrison, 1997). Information and search-cost barriers on both the supply and demand sides of this market can be lowered through policy support for business angel networks (other policies to augment informal equity investment are also available, including various forms of tax relief and measures to facilitate exit from investments).

2. For instance, it has been seen that some management consultancy companies have first designed and marketed major service packages – such as in technology management – for and to the large-firm market. Similar products tailored to the needs of smaller companies have been developed and marketed only later.



3. Various strands of evidence suggest a role for imitation. For instance, Audretsch et al. (2002) examined the impact of the Small Business Innovation Research programme in the United States. The commercialisation of research brought about by this awards programme was found to have had an important secondary effect. That is, following examples of success among their peers, other scientists had been induced to attempt entrepreneurship. Audretsch and Stephan (1996) also showed that demonstrations of entrepreneurial success were important in creating clusters of scientists in biotechnology firms.

4. www.innovation.gov.au/innovationreview/Documents/NIS_review_ Web3.pdf

5. Businesslink (2010), Starting up, www.businesslink.gov.uk/bdotg/action/home

6. Technology Strategy Board, press release 3 December 2008. www.technitex.org/files/08125923KTN%20REVIEW%203.12.08.pdf.

7. Norway’s Centres for Environment-Friendly Energy Research operate in an identical manner. See: www.forskningsradet.no/servlet/Satellite?c=Page&cid=1222932140849&p=1222932140849&pagename=energisenter%2FHovedsidemal.

8. See www.sbs.ox.ac.uk/centres/entrepreneurship/Documents/David_Kingham.pdf.

82 – 3. NON-R&D-BASED PUBLIC SUPPORT FOR BUSINESS INNOVATION


References

Audretsch, D.B. and P.E. Stephan (1996), Company-scientist locational links: The case of biotechnology, American Economic Review, 86 (3), pp. 641- 652.

Audretsch, D.B., J. Weigand and C. Weigand (2002), “The Impact of the SBIR on Creating Entrepreneurial Behavior”, Economic Development Quarterly,Vol. 16, No.1, February.

Cornet, M., M. van der Steeg and B. Vroomen (2007), De effectiviteit van de innovatievoucher 2004 en 2005 Effect op innovatieve input en innovatieve output van bedrijven, edited by C. D. P. N. 140.

European Commission (2009), “INNO-Policy Trend Chart – Innovation Policy Progress Report: Norway”, Enterprise DG, Brussels.

Flanagan, K., E. Uyarra and M. Laranja (2010), “The ‘Policy Mix’ for Innovation: Rethinking Innovation Policy in a Multi-level, Multi-actor Context”, MPRA Paper N° 23567.

Mason, C.M. and R.T. Harrison (1997), “Business Angel Networks and the Development of the Informal Venture Capital Market in the U.K.: Is There Still a Role for the Public Sector?”, Small Business Economics, Vol. 9, No. 2, April, pp. 111-123.

Mason, C.M. and R.T. Harrison (2002), “Closing the Regional Equity Gap? A Critique of the Department of Trade and Industry’s Regional Venture Capital Funds Initiative”, Hunter Centre for Entrepreneurship, University of Strathclyde, Glasgow.

Murray, G. (1999), “Early Stage Venture Capital Funds, Scale Economies and Public Support”, paper presented at the Conference “Funding Gap Controversies”, Warwick University, 12-13 April.

OECD (2006), Entrepreneurship in the Districts Mittweida and Altenburger Land, OECD LEED Local Entrepreneurship Series, October 2006, OECD, Paris.

OECD (2008), OECD Reviews of Innovation Policy: Norway, OECD Publishing, Paris.

OECD (2010), “Environmental Policy Design Characteristics and Technological Innovation: Evidence from Patent Data”, ENV/WKP(2010)2.

4. DEMAND-SIDE POLICIES TO SUPPORT INNOVATION: TRENDS AND CHALLENGES – 83


Chapter 4

Demand-side policies to support innovation: Trends and challenges

This chapter reviews and assesses countries’ policies and programmes to stimulate demand for innovation. It presents the motivation, rationale and scope of demand-side policies and provides examples of different national strategies and approaches. The policy instruments reviewed include innovation-based public procurement, technology-oriented regulations, product standards and the European Union’s Lead Market Initiative.

84 – 4. DEMAND-SIDE POLICIES TO SUPPORT INNOVATION: TRENDS AND CHALLENGES


Evidence and policy trends

From the United Kingdom to Finland, OECD countries, as well as the European Commission, have made explicit policy statements on the need to give greater emphasis to demand-side innovation policies (see Box 4.1). While there is no single definition of demand-oriented innovation policy, it can be understood as a set of public measures to increase the demand for innovations, to improve the conditions for the uptake of innovations or to improve the articulation of demand in order to spur innovations and facilitate their diffusion (Edler, 2007).

Box 4.1. General action plans for demand-driven innovation: the cases of Finland and the United Kingdom

Finland has a general plan for demand-led innovation: The Finnish Policy Framework and Action Plan for Demand and Use-driven Innovation Policy (2010)* for which the Ministry of Employment and the Economy is responsible. This identifies four key areas of intervention: 1) competence development; 2) regulatory reform; 3) public sector operating model; and 4) developing incentives for grass roots initiatives.

The United Kingdom’s plan Innovation Nation (2008)** prepared by the Department for Innovation, Universities & Skills has introduced demand-side initiatives affecting procurement and regulation. As regards innovative procurement, each government department is expected to produce an Innovation Procurement Plan aligned with its commercial strategy, setting out how it will drive innovation through procurement. Secondment of experts from the private sector to the public sector is also being stressed for the purpose of mentoring on pro-innovation procurement. BIS and the Better Regulation Executive are to collaborate with the Business Council for Britain and other bodies to identify how regulation can promote or hinder innovation. * www.tem.fi/index.phtml?l=en&s=2382

** www.dius.gov.uk/reports_and_publications/innovation_nation

The growing interest in demand-side policies has emerged in part because of greater awareness of the importance of feed-back linkages in the innovation process between supply and demand. Demand-oriented innovation policies are thus part of an evolution from a linear model of policy, usually focused on R&D, to a more broad-based approach that considers the full innovation cycle.1 The focus on demand-side policies also reflects a generalised perception that traditional supply-side policies – despite many refinements in their design over past decades – have not been sufficient to bring about the desired level of improvements in innovation performance and productivity. Furthermore, current pressures on governments’ discretionary spending also create incentives to explore how innovation might be fostered



without new programme spending. With pressing societal needs in such areas as health, security, population ageing and the environment, consideration is being given as to how government spending might foster innovations that could lower the costs of meeting these needs. Japan for instance, has recently re-oriented its innovation strategy towards a series of demand-oriented national goals such as the transition into a low-carbon economy and tackling the challenges of an aging society.

Demand-oriented policies take a variety of forms. Innovation-oriented public procurement, innovation-related regulations and standards are principle instruments. Tax policies aimed at raising investment demand are also relevant. With the exception of experiences in the United Kingdom, Finland and the European Union, demand-side innovation policies typically target specific sectors. For instance, in the United States, defense-related R&D procurement schemes have operated for decades. In the energy sector, demand oriented policies have included, guaranteed tariffs (for renewable), and specific power purchase agreements with local utilities. Targeting consumers, governments might offer rebates, for instance on energy efficient products, as has happened in many countries with Compact Fluorescent Lamps. Governments might also promote comparison labelling (to inform consumers on the relative efficiency of products) or endorsement labelling (e.g. “CFC-free”). Framework policies that give primacy to market signals and which open markets for competition are also demand-oriented in an important respect (but are not considered in this study in detail).

Demand-oriented policies respond to situations where markets for innovative products may be insufficiently developed (e.g. certain renewable energy technologies) and/or where public sources of demand afford opportunities to stimulate innovation. Demand-oriented policies focus on the end stage of the innovation cycle. Consequently, they often involve a broad range of stakeholders, which can create specific challenges in both their implementation and their evaluation.

Responses from OECD member countries to the OECD Science, Technology and Industry Outlook 2010 questionnaire (Table 4.1) indicate that demand-side innovation policy is still considered a low priority compared to supply-side policy approaches.



Table 4.1. Level of priority accorded to demand-side policies

Level of priority Country

High priority(8) Finland, Spain

Medium-high priority (6-7) Austria, Korea, Japan, Norway, Slovenia

Medium priority (4-5) Canada, Germany, Hungary, Netherlands, Sweden

Medium-low priority (1-3) Denmark, France, Israel, New Zealand, United States

Note: On a scale of 1 to 8 (a score of 0 suggests the policy is not important, while a score of 8 indicates the policy is very important in the new national STI strategies).

The statistical data for Israel are supplied by and under the responsibility of the relevant Israeli authorities. The use of such data by the OECD is without prejudice to the status of the Golan Heights, East Jerusalem and Israeli settlements in the West Bank under the terms of international law.

Source: Country responses to the 2010 STI Outlook questionnaire.

Nevertheless, despite this relatively low priority ranking, some demand-side policies to strengthen innovation are widely used. These include public procurement, regulations and standards. Table 4.2 sets out the main features of these policy mechanisms. The subsections thereafter consider the rationale for and modus operandi of each instrument.

Table 4.2. Key features of demand-side policy instruments

Demand-side policy

Procurement Regulation Standards

Objective New product or service

Market uptake, increased compe-tition, social goals

Market uptake, interoperability, transparency

Input Money, performance requirements, Skills

Legal process, need to co-ordinate

Standards agencies, need to co-ordinate

Participatory incentive

Sales, preferential treatment (e.g.SMEs)

Mandatory Voluntary

Main player Government Government Industry

Effects on success Improved public services and stimu-lation of innovation

Reducing market risk Reduce market risk

Possible risks Insufficient skills in the public sector, idiosyncratic demand

Conflicting goals, lengths of the process

Technology lock-in

Source: OECD based on Aschhoff and Sofka, 2008.



Public procurement

The rationale for using procurement as a policy tool is perhaps four-fold:

Because of their purchasing power, governments can shape innova-tion directly and indirectly. They can foster innovative activities within firms: firms benefit because procurement can help them recuperate the sunk costs of large and sometimes risky investments over a pre-determined period of time. And by creating a signalling effect as a lead user they can also influence the diffusion of innovation (the expectation of course is that an advance in innovation caused by procurement policies will translate into benefits for the domestic economy, rather than for overseas suppliers of innovative goods or services).

The delivery of essential public services can become more cost-effective as targeted innovation proceeds (governments can also develop innovations of their own in order to improve the efficiency of their processes and enhance the quality and availability of public service delivery – see Table 4.3 for a typology of public procurement).

Particularly for procurement from small firms, public sector demand may help to counter problems in access to finance that can parti-cularly affect this business segment. Depending on their design, they may also help offset problems of bias against small firms in the public tendering market. The provision of a market entailed in the awarding of a contract, and the fact that a public agency has evaluated information on the firm awarded the tender, might also serve to attract additional finance from private sources for innovative activities.

More generally, governments may need to create a market for a new technology in order to meet a policy challenge that is time-bound. The search for commercial-scale low carbon emission technologies could be considered case in point.

88 –

4. D

EMA

ND

-SID

E PO

LIC

IES

TO S

UPP

OR

T IN

NO

VA

TIO

N: T

REN

DS

AN

D C

HA

LLEN

GES BU

SIN

ESS

INN

OV

ATI

ON

PO

LIC

IES:

SEL

ECTE

D C

OU

NTR

Y C

OM

PAR

ISO

NS

– ©

OEC

D 2

011

Tabl

e 4.

3. P

rocu

rem

ent t

ypes

and

pos

sibl

e ef

fect

s of

pub

lic s

ecto

r int

erve

ntio

ns o

n in

nova

tion

Rol

e of

the

publ

ic s

ecto

r M

ain

mot

ivat

ion

of

proc

urem

ent o

r aw

ard

Pote

ntia

l in

nova

tion

type

In

nova

tion-

rela

ted

risks

on

the

supp

ly-s

ide

Geo

grap

hy o

f pr

ocur

emen

t

Effic

ient

pr

ocur

emen

t La

rge

effic

ienc

y-dr

iven

use

r

Best

val

ue fo

r mon

ey

Incr

emen

tal

Ove

rdep

ende

nce

on p

ublic

m

arke

ts, r

isk

of

obso

lesc

ence

Cen

tralis

ed

spec

ifica

tions

(s

tand

ard)

Adap

ted

proc

urem

ent

Nic

he u

ser

The

best

ada

pted

sol

utio

n M

arke

t nic

he

Mar

ket u

ncer

tain

ty

Reg

iona

l sp

ecifi

catio

ns,

regi

onal

pro

cure

men

t

Tech

nolo

gica

l pr

ocur

emen

t La

rge

(sop

hist

icat

ed)

cust

omer

The

best

ava

ilabl

e so

lutio

n Ar

chite

ctur

al

Insu

ffici

ently

relia

ble

dem

and

to ju

stify

in

vest

men

t

Cen

tralis

ed

spec

ifica

tions

, na

tiona

l pro

cure

men

t

Exp

erim

enta

l pr

ocur

emen

t E

xper

imen

tal

(lead

) use

r Th

e m

ost i

nnov

ativ

e so

lutio

n R

adic

al

Mar

ket u

ncer

tain

ty, d

iffic

ult

user

-pro

duce

r com

mun

ica-

tion,

insu

ffici

ent i

ncen

tives

(e

.g. I

P pr

otec

tion)

.

Reg

iona

l sp

ecifi

catio

ns,

natio

nal p

rocu

rem

ent

Not

e: A

dapt

ed p

rocu

rem

ent:

proc

urem

ent a

ddre

ssin

g sp

ecifi

c de

man

d ni

ches

, but

em

ploy

ing

know

n pr

oduc

tion

met

hods

and

pra

ctic

es (n

ew o

r mor

e co

mpl

ex r

equi

rem

ents

e.g

. cus

tom

ised

sof

twar

e);

Tech

nolo

gica

l pro

cure

men

t: pr

ocur

emen

t en

cour

agin

g ne

w t

echn

ical

sol

utio

ns t

o m

eet

a ge

neric

ne

ed (e

.g. w

aste

man

agem

ent);

Exp

erim

enta

l pro

cure

men

t: pr

ocur

emen

t with

ada

pted

tech

nolo

gica

l sol

utio

ns (e

.g. s

peci

alis

ed te

chni

cal e

quip

men

t).E

ffici

ent p

rocu

rem

ent:

proc

urem

ent o

f sta

ndar

dise

d pr

oduc

ts s

ervi

ng a

gen

eric

mar

ket (

com

mon

pre

fere

nces

, lar

ge n

umbe

r of p

urch

aser

s, li

ttle

need

fo

r var

iety

from

end

use

rs, e

.g. o

ffice

sup

plie

s).

Sou

rce:

OE

CD

, ada

pted

from

Uya

rra a

nd F

lana

gan

(201

0).



The concept of fostering innovation through procurement is not new and some countries have pursued active technology procurement policies for decades. Public procurement has for instance been a key determinant of the emergence of a number of high-tech sectors in the United States, Japan and France (where public procurement has been used, for instance, to develop high-speed rail technology and to ensure a competitive advantage in nuclear energy technologies). However, the potential of public procurement has received renewed impetus. In recent years, OECD countries such as Australia, Finland, Germany, Sweden and the United Kingdom as well as the European Commission have stressed public procurement as a key means to drive innovation and meet societal goals. New approaches to public procurement are developing. For instance, the United Kingdom has actively sought to integrate procurement for innovation across government since 2003 (it is estimated that the annual value of public procurement contracts is around ten times the amount devoted to the R&D tax credit in the United Kingdom). And Germany has created a new Agreement on Public Procurement of Innovation by which six federal ministries (interior, economics, defence, transport, environment and research) will promote innovative procurement. All six ministries will publish long-run demand forecasts, engage in continu-ous market analysis to identify potential new solutions, offer professional training on the legal options to promote innovation, and foster a strategic dialogue and exchange of experiences between procuring agencies, end-users, and industry and procurement agencies on all state levels.

Two levels of public procurement can be distinguished, but are not usually treated separately in the literature. First, there is government procure-ment in general, which can be organised so as to be more conducive to innovation. In this case innovation-related criteria are incorporated in the tender specifications and in the assessment of tender documents. Innovation-oriented public procurement can be implemented in this way for a vast number of products and services purchased by public authorities, from construction, transport, energy and catering services, to health products and equipment.

In general, this type of procurement operates in several stages: defining the subject matter of the contract, drawing up the technical specifications and the contractual parameters for products/services, and determining the best bid. This general structure is similar to procurement procedures in the private sector, although public authorities obviously have a set of additional criteria to apply – they have the responsibility to get the best value for tax-payers’ money and must ensure that all competitors have an equal opportu-nity to compete for the contract.



Secondly, public procurement can be strategic, when targeted demand is deliberately created for certain technologies or services. Strategic procure-ment is associated with sectoral policy and therefore generally neither initiated nor co-ordinated by the ministries responsible for innovation. Public technology procurement involves purchasing a not-yet-existing product or systems. This requires novel technological development work on the part of the companies or institutions responding to the call for tender (Edquist and Hommen, 2000). Ideally, functional requirements of demanded products are predefined by government. Cases in point could include a public health administration seeking equipment to perform novel surgical procedures, or a defence contractor requiring equipment with new functionalities.

The use of public procurement to stimulate innovation involves a number of challenges. Public procurement must be designed to be efficient and to not distort competition. Thus, the design of pro-innovation procure-ment mechanisms, as with traditional procurement, must avoid the risk of capture by large firms and/or other anti-competitive effects. An additional challenge for governments is that procurement itself is often highly fragmented across local, regional and national governments agencies. Sub-national governments are particularly important players, as they account on average for 64% of public investment in OECD countries. Innovative public procurement is far more challenging for OECD countries with decentralised procurement systems. Furthermore, many agencies with responsibilities for public procurement operate separately from line ministries or government agencies that have a remit to foster innovation. Specialised procurement agencies are mainly responsible for the efficiency of purchasing, and expertise in the respective fields of innovation may be lacking. Some OECD countries have issued guidelines for innovation-oriented procurement (e.g.United Kingdom), while others (e.g. Finland) have even introduced funding instruments to encourage government agencies to undertake innovation-oriented procurement. Also in Germany, the method of “best available technology” in Green Public Procurement (GPP) has long been used to facilitate innovation in environmental sectors and has helped make German companies a world-leader in this sector (Blind et al., 2004).

A further consideration is that the procurement of innovation entails a number of distinct risks (above and beyond those entailed in all procurement procedures). European Commission (2010) identifies major risks associated with the procurement of innovation, including the following:

Technological risk – that is, non-completion risk stemming from technical features of the good or service to be procured. One mitigation option is contract design, for instance using cost-reimbursement or incentive contracts. As compared with procurement of standard off-the-shelf items, uncertainties inherent to innovative items create diffi-



culties in writing contracts that frame incentives to reduce or eliminate risk. For instance, the expected quality of a wholly new item might not be verifiable beforehand. Another mitigation strategy is to use framework agreements or multi-stage procurement processes. The latter effectively restrict the degree of competition in the final stage of the process, while giving opportunities to screen out more risky bids during early stages of the procurement. The report also recommends involving potential users in the process, although difficulties can arise however with respect to the permissible extent or timing of any pre-contract interaction with suppliers.

Organisational and societal risks – that is, risks stemming from within the procuring organisation and/or those related to uptake of the good or service by users. The former can stem from such issues as inadequate absorptive capacities in procuring institutions or incompatibilities with existing technologies or routines. Such risks can be mitigated through joint foresight exercises with public and private lead users as well as early user involvement in the procurement process. Transparency of procurement goals should also be main-tained, and caution should be exercised if procurement involves prompt introduction of significantly new technologies to an institution.

Market risks – these risks exist on the side of both supply and demand. On the demand side, risks are greatest for wholly novel items. Public bodies might mitigate such risk by implementing additional demand-side measures, such as user training schemes or using demand aggregation, in particular by bundling public demand. However, possible downsides of aggregating procurement contracts – such as limiting the opportunities for SME participation – may also need to be countered. On the supply side, the main risk is that suppliers do not respond to the tender. To mitigate this risk, market intelligence capacities should exist, developed for instance through structured exchanges with internal or external experts.

General public procurement of innovative goods and services Making public procurement in general more innovation-oriented is

being attempted in several of the countries reviewed in this study. The following paragraphs describe a range of government initiatives aimed at incorporating an innovation dimension into general public procurement.



Finland Finland’s broad-based Innovation Strategy2, adopted in 2008, emphasises

the role of the public sector in developing, applying and introducing inno-vations. Demand and user-driven innovation policy is one of four key areas in the national innovation strategy. Annual public procurement in Finland amounts to some EUR 23 billion (USD 32 billion), offering considerable purchasing power with which to promote and encourage innovation. The 2010 Action Plan for the implementation of demand and user-driven innovationpolicy includes several proposals for enhancing demand for innovations through public procurement. These consider: the development of central and local government procurement procedures and methods, strengthening the role of actors supporting public procurement and examining different incentive and risk management models. In addition, the Government public procure-ment strategy3 was revised in 2009 and includes guidelines for promoting innovation in government procurement (e.g. by encouraging the search for innovative solutions together with suppliers).

Under the management responsibility of Tekes, a procurement funding instrument was launched in June 2009 to provide incentives for promoting innovation through public procurement. Public procurement units and public utilities (at central and local level) can apply for funding for public procure-ment of innovations. Tekes funds can be used both for the planning and R&D stages. External advisors can be utilised in the planning stage (e.g. in legal, commercial and technological as well as user experience issues) in order to support the procurement process.

During the first year of operation of the funding instrument, Tekes has focused on areas such as energy, environment, construction and health sectors, as these are considered important to meet future demand and address societal challenges. However, activities in other areas are also eligible for funding. To date, 12 projects have been accepted for funding. Projects mainly focus on local authority services, especially in the social and healthcare sectors. Sustainable development and energy efficiency are objectives in a few of the projects.

Preliminary surveys show that interest in the funding instrument has emerged more slowly than expected, in part because the target group of the funding represents a new group of customers for Tekes. Finally, the criteria for obtaining funding from Tekes are stringent: they require that goods or services procured must either be entirely innovative (not available on the market) or that the procurement results in new forms of public service delivery.



Australia – public procurement to support innovative firms Released in 2008, Australia’s Commonwealth Procurement Guidelines4

accepted that Agencies “should seek to ensure that wherever possible their processes allow for suppliers to provide innovative solutions to their require-ments”. In 2009, Australia introduced its 10 year innovation agenda, Powering Ideas5, where the role of public procurement was highlighted as an important means to support innovative Australian firms. In 2009 government procurement stood at around AUD 26.3 billion (USD 24.5 bil-lion) per annum.

The Australian Industry Participation (AIP) National Framework (2001) aims at supporting Australian industry to innovate, develop and enhance competitive capabilities, and take advantage of investment opportunities. In its 2009 Procurement Statement, the Australian government announced a series of measures to extend and strengthen the AIP framework, notably to:

Apply the AIP framework to large Commonwealth tenders (above AUD 20 million; USD 16 million) and Commonwealth infrastructure projects;

Emphasise the connection of Australian suppliers to Commonwealth-funded infrastructure.

The main source of information about the Commonwealth Government procurement market is AusTender, a Government website that provides information on contracts across the Commonwealth. It includes information on: i) all Commonwealth procurement contracts and agency agreements above AUD 10 000, including standing offers; ii) Agency Annual Procure-ment Plans, which draw suppliers’ early attention to potential procurement opportunities; and iii) all new tendering opportunities. By registering on AusTender suppliers can “opt in” to be informed electronically about govern-ment tenders.

Government Agencies are required to manage contracts from the tender stage through to the delivery of the goods or services. Because of the large number of agencies that participate in procurement processes, the extent to which agencies adhere to the procurement framework is unclear. This also makes it difficult for the Government to obtain an overall picture of the status of implementation of new whole-of-government procurement policies. 6



Netherlands – making government procurement more innovation-oriented

Government purchasing in the Netherlands amounts to about EUR 60 bil-lion (USD 83 billion) per year. Around one thousand contracting authorities put more than ten thousand tenders on the market yearly. Policy measures have been introduced to make government procurement more innovation oriented, notably through the Public Innovation Procurement (PIP) pro-gramme. The PIP is managed by the Dutch Public Procurement and Expertise centre (PIANOo7). It is intended to provide public authorities with innovative solutions to solving societal problems in priority areas (water, energy, security, health, education, mobility, construction and agriculture and fishing). Currently, a monitoring process is underway (using a key performance indicator in the 2010 budget of the Ministry of Economic Affairs) to assess the number of procurement cases aimed at finding innova-tive solutions. PIANOo is currently supporting pilot projects on innovation in procurement, through the provision of finance and advice. The experiences of these pilots are shared in the network as part of the government effort to collect, analyse and highlight best practice provide guidance to public agencies.

United Kingdom – Innovation-oriented procurement policies and guidance

The United Kingdom has instituted several innovation-based procurement-related policies. It issued a procurement guidance in 2007 – Finding and Procuring Innovative Solutions – and introduced an Innovation Procurement Plan in 2009, making innovation a key requirement in large facilities and capital programmes.

Forward Commitment Procurement (FCP) is a procurement model that is used to deliver cost effective products and services to the public sector and to help to create market conditions for innovative goods and services. This scheme was initially developed in partnership with the Office of Government Commerce to address market barriers faced by environmental innovations. It is now a process to encourage industry to bring forward innovative solutions in markets relating to sustainable development, health-care and construction, subject to needs being met with products at specified quality and performance levels and price. FCP provides the market with information on unmet needs, and early engagement takes place to develop specifications and appropriate procurement routes. Currently nine projects are being supported.



Public procurement of R&D: Small Business Innovation Research-type programmes

Innovative small firms often face difficulties in attracting investors to support their innovation projects – especially at the seed stage. This has incited governments to play a role in funding the development of new technologies in small companies through R&D contracts. From the govern-ments’ perspective, SBIR-type programmes have a double aim: to stimulate technological innovation, while at the same time providing government agencies with new cost-effective solutions to their needs. In some countries an integral and advantageous feature of such programmes is their facilitation of small-firm access to public R&D contracts. Allowing recipients to retain rights to any resulting intellectual property is another feature that can make such contractual arrangements attractive to firms.

The Small Business Innovation Research (SBIR) programme was intro-duced in the United States in 1982. Some evaluative work has shown that that SBIR funding has led to increased growth and employment creation and a greater likelihood of attracting venture financing (Lerner, 1999; NRC, 2000), although other analyses have cast doubt on the additionality of SBIR impacts (Wallsten, 2000). The perceived success of the programme has inspired similar programmes in other OECD countries, notably Japan, Australia, the United Kingdom and the Netherlands.

Risks, however, are also associated with SBIR-type programmes. A key concern is that government funds might simply crowd out privately-financed R&D. Schemes should finance proposals not likely to receive funds from private sources if additionality is to be maximised (Wallsten, 2000, 1998) (see Chapter 5). Another criticism is that SBIR-like initiatives tend to develop a technology to a certain level of readiness, while most major commerciali-sation successes would require substantial subsequent funding.

The United Kingdom – Small Business Research Initiative Introduced in 2001, the United Kingdom’s Small Business Research

Initiative (SBRI) earmarks a share of the government’s procurement budget (about 11% of the budget in financial year 2007-08) to be assigned to SMEs through competitive R&D contracts. The SBRI has been reformed several times to increase its performance, reach and impact. The last reform was launched in 2009. The new SBRI involves a pre-commercial procurement process. The Technology Strategy Board is the agency in charge of the programme. Funding operates in two phases: 1) a feasibility phase (GBP 100 000; USD 156 000); and 2) a development phase (GBP 250 000-GBP 1 million (USD 390 000-USD 1.6 million). There are currently 370 contracts



in the areas of defence, health and construction with a total value of GBP 25 million (USD 39 million).

The SBRI programme has been evaluated by the former Department for Innovation Universities and Skills (DIUS). Problems encountered in the early days of the programme were linked to: a lack of participation from government departments, the low total value of the contracts going to small firms and the fact that these were rarely linked to technical development (this led to the reform of the SBRI in 2009). Some studies still point to insufficient participation across government, and also state that awards are highly skewed towards a number of very small phase I demonstration projects (Connell and Probert, 2010).

The Netherlands – SBIR The government launched a small scale SBIR on several different

themes: agriculture, energy, transport, water management and defense. SBIR is managed by the Dutch Agency for Sustainability and Innovation (SenterNovem). It incorporates the basic features of the United States’ SBIR programme, providing funds to SMEs on a procurement basis to develop innovations that could contribute to solving societal challenges. SBIR projects are procured via tenders and funds are granted in two phases, a feasibility phase (EUR 50 000; USD 69 000), and a development phase (EUR 450 000; USD 625 000). SBIR’s budget was EUR15 million (USD 21 million) in 2008. An independent committee evaluates proposals and makes a ranking, which the Minister uses in the choice of candidate projects.

Using data from 88 firms, a first evaluation of the SBIR pilot pro-gramme in 2007 showed that SBIR brought in companies that were new to the procurement market, that companies receiving funds are small (less than 100 employees) and that they co-operate more with other companies and research institutes than firms that did not receive a contract.8

Australia – The Victorian State Government Smart SMEs Market Validation Program (MVP)

The Market Validation Program (MVP) was introduced by the Victoria State Government in 2008 as part of the Boosting Highly Innovative SMEs (BHIS)9 programme (Victoria is one of eight state/territory governments in Australia). The MVP commits AUD 28 million (USD 26 million) over four years and is administered by the Victorian Department of Innovation, Industry and Regional Development (DIIRD).



The aim of the MVP is to help SMEs commercialise new intellectual property and develop globally competitive technologies, products and services. The MVP seeks to yield R&D proposals that deliver a solution to a public sector technological requirement.

Structurally, MVP is a demand-led programme. It uses a three-stage approach: specification of technology requirements by public sector entities, feasibility study and proof of concept. The MVP engages two stakeholder groups – public sector entities and SMEs. The MVP is broadly modelled on the United States’ SBIR programme. As with the SBIR it is a tendering and contractual scheme and is based on solicitations using a description of the problem rather than pre-determined solution specifications. It follows a milestone funding model, along venture capital lines, which allows for “fast fail” decisions and systematic evaluation. SMEs own intellectual property developed under the programme.

One important difference between SBIR and MVP is that MVP aims to encourage participation from public sector entities by providing funding through a central and independent agency (DIIRD). DIIRD also undertakes extensive administrative work to support the public sector entities through-out. Thus, participating agencies are not required to exclusively use their own human resources to manage the programme but they are required to commit to the in-kind contribution of a Project Manager who oversees the technology project through feasibility study and (if successful) through proof of concept.

The MVP is open to over 300 public sector agencies and organisations in Victoria. It is a four-year pilot programme comprising two funding rounds. Evaluation of the MVP includes establishing measurable performance indicators through a benchmarking and evaluation framework, an interim evaluation report at the midway point of the MVP and a final evaluation and review at both the project and programme levels.

Canada – Canadian Innovation Commercialisation Program (CICP). CICP is a CAD 40 million pilot programme launched in 2010. It has been

created to bolster innovation in Canada’s business sector and to help companies bridge the pre-commercialisation gap for their innovative products and services by awarding contracts to entrepreneurs with pre-commercial innovations (i.e. first sales of technology) through an open, transparent, competitive and fair procurement process. This allows government to:



Test and provide feedback to entrepreneurs on the performance of their products or services (many entrepreneurs struggle to find buyers due to the risk associated with untested products).

Provide innovators with the opportunity to enter the marketplace with a successful application of their new products and services.

The CICP will target innovations in four priority areas: environment, security and safety, health and enabling technologies. The programme will include four calls for proposals with a full description of the priority areas and specific criteria for selection.

Innovation-oriented regulations and standards

Regulations and standards play important roles in structuring markets for goods and services. This section considers these roles, the rationale for government action, and selected policy initiatives in the countries under study. There is often complementarity between regulations and standards. Regulations set the essential levels of safety, environmental or health protection and are frequently complemented by harmonised consensus-based standards-setting on technical specifications. This allows other economic actors to collaborate with public authorities to design the most appropriate implementation standards and to regularly update them to take stock of evolving needs and technical progress.

For some time, much of the policy attention given to regulation has not been concerned with innovation. Rather, the focus has been on the ways in which regulations contribute to overall framework conditions, in particular their effects on burdens for doing business and on the primacy of market signals. This subject is not treated in detail here, save to say that competition is clearly central to innovative activity and that relatively liberal product and labour market conditions can enhance the adaptability of firms and lower the chances of becoming locked in to given technologies.10 Lower administra-tive burdens can also facilitate business creation, an important seedbed for innovation. The focus in this section however is on regulations that have some sector- and/or innovation-specific intention or effect.

Regulations Regulation refers to the implementation of rules by public authorities

and governmental bodies to influence the behaviours of private actors in the economy. Regulation influences innovation indirectly, since it affects the framework conditions for firms and involves no direct outlay of public funds (Geroski, 1990).



Regulations can affect the performance (quality, compatibility) or consequences (health, safety, the environment) of products or services (e.g. labelling, recycling regulations, emission standards etc.), thus having a direct impact on demand for innovative goods and services. Metcalfe and James (2001) note the importance of regulation in the area of medical devices, where public policy has been critical in shaping innovation processes in Europe and in the United States. The Promotion of Renewable Energies Heat Act (2009) in Germany is an example of regulation that promotes the diffusion of innovation. The Act stipulates that owners of newly-constructed buildings must use renewable energies. Moreover, building owners who use particularly efficient innovative technologies, or that have low emissions figures, will receive money from the state. And in Japan, METI’s Top Runner programme adopts a dynamic process of setting and revising performance standards by taking the current highest energy efficiency rate of products as a benchmark in 23 product groups. This flexible setting of benchmarks creates positive incentives and competition among manufacturers to quickly improve their product performance, without calling on public financial support. In addition, positive innovation-related effects of regulations for firms can stem from the increased acceptance of new products by consumers.

However, the effects of economic regulation on innovation are far from straightforward, and can be ambiguous a priori. Mahdi et al. (2002) review the impact of health, safety and environmental regulation on the chemical industry in Europe. This study was spurred by concerns that the more stringent regulatory conditions in Europe would retard innovation relative to competitors in the United States. But their findings indicate that rates of new chemicals notification between Europe and the United States had experienced convergence over the previous decade. Their review of the literature suggests that in most cases regulation both inhibits and stimulates innovation. They conclude that “despite a long tradition of research on the question of how regulation influences innovation in different industries and in different countries, it is far from clear where the balance between these two effects falls”.

The impacts of regulation on innovation are likely to be highly tech-nology and industry specific. As noted in Chapter 5, evidence has been gathered demonstrating that anticipation of regulatory change in some sectors has induced innovation. Studies of asbestos product development (Ashford et al., 1985) and SO2 removal technologies (Taylor et al., 2005) are cases in point. But Nemet (2009) examined wind-power technologies and found that an array of demand-inducing policies in California had not spurred significant innovation, in part because a dominant industry technology



had already been identified (this study is described in greater detail in Chapter 5).

To assess the appropriateness of regulatory policy targeted at a specific sector, analysts also need to ask whether the market would introduce the right level technology in the absence of the regulation. For instance, with respect to regulation on fuel efficient vehicle, if the market were efficient in terms of fuel economy technologies the regulation could be redundant. Whether the market is efficient or not will likely have industry-specific considerations.11

The precise form that the regulation takes will also affect its impact on innovation. For example, uncertainty in the duration of a regulation could reduce the strength of influence on demand conditions. And in the United States the Corporate Average Fuel Economy (CAFE) regulation introduced in 1978 was framed in such a way that increases in average vehicle fuel efficiencies could be achieved through manufacturers changing relative car prices so as to sell fewer large cars and more small cars. Regulations in the United States enacted in the 1970s and governing energy efficiency in refrigerators served to increase efficiencies over time, but only up to levels already existing in equivalent appliances in Europe. No technological innovation was seen initially. And in the environmental sphere, the empirical evidence suggests that market-based instruments such as tradable emissions permits are more likely to stimulate innovation than direct regulations such as technology-based standards (OECD 2010a).

OECD (2010a) suggests a further nuance to thinking about the role of regulation in encouraging innovation. This work, based on patents data, considers the characteristics of environmental policies – including direct regulation – that are likely to induce innovation. The authors observe that, when considering environmental impacts, it is most helpful to take account of the specific design characteristics of different instruments (market-based or regulation-based). They outline a minimal set of essential design charac-teristics to take account of. These are:

Stringency – that is, how ambitious is the environmental policy target?

Predictability – that is, what effect does the policy have on investor uncertainty?

Flexibility – that is, does the innovator identify the best way to meet the objective?

Incidence – that is, does the policy target the externality directly, or is the point of incidence a “proxy” for the pollutant?, and



Depth – that is, do incentives exist to innovate through a range of potentially ascending objectives?

The authors note that to induce innovation, the ideal policy instrument will be one which is: sufficiently stringent to encourage an optimal level of innovation; stable enough to give investors adequate planning horizons for risky investments; flexible enough to encourage innovators to create genuinely novel solutions; be closely targeted on the policy goal, so as to avoid misallocation of innovative effort; and provide incentives for continu-ous innovation. The potential innovation stimulus delivered by market-based and regulation-based instruments needs to be assessed against these criteria. As this work makes clear, there is no automatic correspondence between the type of instrument and the critical design attributes. For instance, different environment-related taxes can have different combinations of these design attributes, and a regulatory standard might have more in common with a tax than a technology-based standard.

A further critical consideration is that even in cases where regulation spurs innovation, regulation-based policy might be cost-ineffective overall. Kleit (2004) provides a detailed economic cost benefit analysis of the aforementioned vehicle efficiency regulations in the United States. The analysis shows that a small increase in the gasoline tax would deliver equivalent savings in fuel consumption but at a much lower cost to society (in part because the regulation lowers the marginal cost of driving and thus induces more driving, with concomitant increases in pollutant emissions, accidents and congestion).

The time period over which policy yields impact might also vary from one regulation to another, again reflecting industry specificities. Greenberg et al. (1979) found a six year lead period in the ammonia industry and were unable to identify specific regulatory effects using an econometric model.

It can also be relatively difficult to isolate the specific effects of regulation from other influences. This reflects the inherent complexity of the pathways by which regulation might shape innovation, the possibility of long lead times between a regulatory stimulus and an industry response, the simultaneous impacts of an array of supply-side factors, as well as inherent uncertainties in the dynamics of innovation (including exhaustion of the research frontier).

StandardsStandards are documents based on various degrees of consensus

(industry-wide, national, regional or international) which lay out rules, practices, metrics or conventions used in technology, trade and society at large. They range from proprietary standards (e.g. exploited by a company and



based on patented technologies) to formal international consensus-based standard (e.g. those produced by the International Organisation for Standardi-sation, ISO). The scope of standards cuts across all areas of economic, environmental and social issues. They can for instance specify terms and definitions, codes, dimensions, physical interoperability, product and service safety and quality (Bryden, 2010).

The economic benefit of standards has become clearer to policy makers in recent years. A study in the United Kingdom by the Department of Industry and British Standards Institution (BSI) has estimated that standards make an annual contribution of GBP 2.5 billion (USD 3.9 billion) to the national economy. Standards were also found to be at the origin of 13% of labour productivity improvement during the period 1948-2002 (Department of Trade and Industry, 2005). Similar studies in the United States, Australia or Canada corroborate the benefits of standards.

There are multiple routes through which standards can affect innovation and other economic outcomes. Swann (2000) provides a comprehensive review of the literature, which includes evidence that successful standardi-sation can achieve some or all of the following:

Standardisation drives innovation because innovation requires com-petition and competition requires interoperability. Successful standards facilitate that interoperability.

Standardisation can increase trade.

Standardisation codifies and diffuses information on technology and best practice. By setting ground rules, common terminology, develop-ment methods and measurement techniques, standards enable the diffusion of innovation. Publicly available standards in particular have a potentially powerful effect on the dissemination of information about technology, from both domestic and international sources. In some markets – such as automotive parts supplies, flat screen TVs and mobile telephony – quality certification and consumer safety rules are important in shaping demand and in the diffusion of innovations.

Standardisation reduces risks for producers and consumers. For instance, standardisation of measurement helps producers of innova-tions demonstrate innovative traits to consumers. And standards lower the risk of investing in a redundant technology.

Standardisation reduces transactions costs between producers and between producers and consumers.

Standardisation may protect against situations in which high-quality producers are driven out of the market by low-quality producers



because information is not fully available to consumers on the quality content of their output (Gresham’s law).

Standardisation may efficiently reduce unnecessary variety among products (in construction, for instance, there might be no need for production of a continuous variety of steel girder widths).

Standardisation helps create critical mass in the formative stages of a given market. Standards can focus demand for innovations that might otherwise be spread over multiple technical solutions. Standards are especially important in network industries, such as ICTs, in that they can facilitate a critical mass of users. In this connection, standards ease the emergence of technological platforms – independently supplied yet interoperable compo-nents with shared technical standards. Many successful platforms, such as the Internet and the cellular telephone, are based on open standards.

The development of standards is likely to experience some degree of market failure. By itself, the market may provide too few standards. Creating standards entails fixed costs, while the gains may not be appro-priable by the individual firm. But problems of regulatory capture might also threaten publicly generated standards.

There is a clear trend towards standardisation work being conducted at the international level because, in a globalised economy compatibility and interface across borders are important. Competitive disadvantages could arise if a country were to free ride on the standards setting work of others. Countries and firms that play primary roles in setting international standards can enjoy advantages from doing so, to the extent that the new standards align with their own national standards and/or features of their productive base. Broader participation of stakeholders is expected to lead to better quality standards, but takes longer to effect (it takes about three years on average to produce an international standard).

The public sector’s role largely involves measures to include under-represented groups in the process of developing standards, and subsidisation of teams drafting international standards. Unlike regulation, the setting of standards is mainly the responsibility of industry bodies – with government acting as facilitator or co-ordinator. This has practical implications: pro-cedures in standards bodies can be slow and bureaucratic and can be held up by large players, which raises the important issue of timing. If standardi-sation is brought into effect too early, it could preclude and shut out better technologies. But if standardisation occurs too late then the costs of transi-tion to the new standard could be high enough to slow or prevent diffusion. If product life-cycles are shortening, the issue of timing is likely to increase in importance.



Box 4.2. The case of biometrics standardisation in the United Kingdom A biometric system is a system for the automated recognition of individuals based on

their behavioural and/or biological characteristics. Wherever there is a need to identify or verify a person there is a potential application for biometrics. This includes entry control to buildings and secure areas, as well as access control to resources such as bank accounts and entitlement services.

The United Kingdom government decided to support standardisation in the area of biometrics, and technical standards supporting interchangeability and interoperability. The objective was for standards to reduce the risk for the procurer, system integrator and the end user, because they simplify integration and enable vendor substitution, technology enhancement and development.

The Government’s support of biometrics standardisation in the United Kingdom had several aims:

To open public procurement contracts to competitive tender through reference to standards, in turn facilitating access for smaller companies and potentially saving public money.

To create confidence that the United Kingdom’s view of biometrics systems development is aligned with international advances in technology (the British Standards Institution provides the United Kingdom’s input to the international biometrics subcommittee and its working groups).

To facilitate information exchange with other national authorities. An independent 2009 review of standardisation and innovation programmes in the

United Kingdom found that funding in the area of biometrics had facilitated the diffusion of technology in the marketplace, made procurement more cost-effective and eased access for SMEs to the procurement market:

Open-systems based standards had saved the United Kingdom Government considerable sums by enabling competition on identity card contracts.

The use of standards had accelerated progress on biometrics programmes, such as that run by the Identity and Passport Service, and had future-proofed the technology.

Standards had enabled United Kingdom-based system integrators to operate in a fair and open market and had prevented domination by a small number of overseas companies.

There are however challenges linked to standardisation and biometric technologies. These relate in part to the changing global security situation, which may create new demands for work and new calls on public resources. Furthermore, the typical timeframes for publication of standards – at around three years – could conflict with the shorter funding horizons typical of government. Source: OECD (2010), Demand-side Innovation Policies.



Box 4.3. The European Union Lead Market Initiative Promoting the creation of lead markets is a new EU policy initiative focusing on

creating demand for innovations. It aims to improve the European market environment to provide better conditions for the creation and growth of new markets for innovative products and to support the development of worldwide operations by pioneering companies operating in Europe. It is held that the fragmented nature of the internal market and innovation system slows the creation of lead markets and a lead market position in the European Union. Following intense stakeholder consultations, the EU Lead Market Initiative has been launched in six sectors in 2008: eHealth, protective textiles, sustainable construction, recycling, bio-based products and renewable energies. These markets have been selected because they are highly innovative, address broader strategic, societal, environmental and economic challenges, have a strong technological and industrial base in Europe and depend more than other markets on the creation of favorable framework conditions through public policy measures. The European Commission, Member States and industry work together to carry out action plans for the next 3-5 years in order to facilitate the emergence of new products or services in these six Lead Markets. A combination of policy instruments will be used including, regulation, public procurement and standardisation.

The United Kingdom In his report, Race to the Top (2007), Lord Sainsbury recognised that

interaction between standards and innovation is key to stimulating research, establishing communication networks and encouraging industrial develop-ment – all prerequisite steps to the commercialisation and widespread uptake of new technologies. He recommended greater collaboration within the United Kingdom standardisation infrastructure in order to better co-ordinate support for emerging industries.

Since the publication of this report, the United Kingdom has provided GBP 2.5 million (USD 3.9 million) in direct support for standards develop-ment in emerging technology areas through the Department for Business, Innovation & Skills (BIS) (see Box 4.2). Biometrics, nanotechnology and regenerative medicine were jointly identified in 2005 by BIS’ predecessors as areas to benefit from funding for standardisation. A preliminary study of the impact and effectiveness of these key emerging technologies was carried out by Ernst and Young on behalf of BIS. The findings of the study indicate that this support is appropriate and beneficial and that Government should develop the model and apply it to other emerging technologies as appropriate.



Lead markets

The European Commission has recently launched the Lead Market Initiative (LMI) that has identified six markets (e-health, protective textiles, sustainable construction, recycling, bio-based products and renewable energies) in which a combination of policies will be used to strengthen the competiveness of leading firms in these markets (see Box 4.3).

Conclusions

Although demand-side policies have received increased attention in recent years, their role in the full portfolio of government policies on innovation remains relatively marginal. The rationale for these policies is linked in general terms to the need to stimulate innovation in areas where societal needs are pressing (e.g. health, environment) and where government action can complement market mechanisms with, ideally, minimal financial outlays. But specific rationales pertain to individual demand-side instru-ments. For example, innovation-oriented public procurement can be designed to help counter gaps in the supply of risk finance for small early-stage ventures. Such procurement can also be structured to help offset biases against SMEs in public procurement. And procurement processes can help accelerate the emergence of technologies for which there is an urgent time-bound societal need. By contrast, as discussed in the text, the rationale for government action in the area of technical standards is somewhat different, corresponding to public-good characteristics possessed by such standards. To be efficient, demand-side policies should clearly target specific market or systemic failures and their objectives, and impacts should be carefully evaluated.

Implementation of each of the key demand-side policies poses challenges. Some demand-side policies arguably carry the risk of too strong a government intervention by comparison with policies to stimulate the supply of R&D and foster knowledge spillovers. Errors in the setting of key regulations for instance could have far-reaching economic consequences. This is complicated by the fact that the effects of economic regulation on innovation – and the timing of these effects – can be complex and ambiguous a priori. Moreover, the effects of regulation on innovation are likely to be highly technology and industry specific, which implies a need for significant industry-specific intelligence among policy makers when framing innovation-oriented regula-tions. Such intelligence also relates to the need to assess the appropriateness of regulatory policy in terms of whether the market would introduce the right level technology in the absence of the regulation (the case of automotive engine fuel efficiency is described, which illustrates the complexity of



determining the degree of efficiency in that specific market). Indeed, all demand-side policies should consider whether the action undertaken is efficient from a market perspective and whether it improves social welfare. The text likewise highlights that the precise form that regulation takes will also shape its impact on innovation. The example is given of environmental regulations, showing that consideration should be given to such policy design features such as stringency, predictability, flexibility, incidence and depth. A further critical consideration is that even in cases where regulation spurs innovation, regulation-based policy might be cost-ineffective overall. Other less costly means might be employed to achieve the same goal, and conflicting regulatory goals might also exist. This underscores the importance of performing cost benefit analysis on key regulatory decisions. Indeed, innovation is usually not among the key objectives of regulators. The policy focus has typically been on avoiding regulatory burdens, rather than on the targeted use of regulation to allow the emergence of new technologies.

Demand-side policies are systemic in nature and this implies that policy measures need to be closely articulated with supply-side incentives, as supply restrictions could potentially jeopardise some demand-led measures. Matching supply with demand requires building bridges along the value chain and this takes time.

Demand-side policies imply a lead role for the public sector. However, the public sector is not always best placed to support the innovation process, and new capacities may need to be developed. For instance, as regards public procurement, despite considerable policy interest, the traditional focus on value-for-money as well as the problem of fragmentation of public demand (often between different levels of government) can limit potential scale effects for innovative procurement. In many countries numerous sub-national units of government play important roles in the public procurement market. This in turn creates challenges in terms of governance, co-ordination and strategic planning. Furthermore, many agencies with responsibilities for public procurement operate separately from line ministries or government agencies with a remit to foster innovation. Technical expertise in the respective fields of innovation may be lacking in the procurement bodies. As many firms do not see public procurement as a relevant source of business, the scope of policy can be limited. And the text discusses capacities that may need to be enhanced in order to mitigate risks associated with procure-ment of innovation, such as technological risks, organisational and societal risks, and specific market risks. Furthermore, innovation goals must be balanced against the need for competition, transparency and accountability in public procurement. OECD countries should adhere to national competi-tion and public procurement rules as well as related international standards



and obligations (e.g. the WTO Government Procurement Agreement). Large player dominance is also possible and in particular occurs in areas where the potential for learning by doing is high. Governments should take measures to ensure this does not occur, sourcing competitively from different firms to prevent dominance.

The public sector’s role with respect to standards largely involves measures to include under-represented groups in the process of developing standards, and subsidisation of teams drafting international standards. Unlike regulation, the setting of standards is mainly the responsibility of industry bodies – with government acting as facilitator or co-ordinator. However, procedures in standards bodies can be slow and bureaucratic and can be held up by large players, which raises the important issue of timing. If standardisation is brought into effect too early it could shut out better technologies. But if standardisation occurs too late then the costs of transition to the new standard could be high enough to slow or prevent diffusion. If product life-cycles are shortening, the issue of timing is likely to increase in importance. Policy makers might also consider monitoring the number and age distribution of the national stock of standards, with a focus on standards developed by national standards bodies (which are open and carry additional credibility). Policy action could be called for if the number, rate of creation and age distribution of the stock of public standards were seen to diverge greatly from norms in other advanced economies. Govern-ment also plays a role in shaping the behaviour of consumers. And consumer policy regimes and consumer education can play a role in providing an arena for innovation in meeting new market demands.



Notes

1. This recognition of the essential interaction between factors of demand-pull and supply-push is also reflected in the broader academic literature. For example, Mowery and Rosenberg (1979) conclude that neither supply nor demand factors are sufficient for innovation. Both must exist simultaneously. And Freeman (1974) surveyed a set of 40 innovations, showing that successful cases were able to link technical and market opportunities.

2. www.tem.fi/index.phtml?l=en&s=2411.

3. www.vm.fi/vm/en/04_publications_and_documents/01_publications/ 08_other_publications/20091008Govern/name.jsp.

4. www.finance.gov.au/publications/fmg-series/procurement-guidelines/index.html

5. www.innovation.gov.au/Section/Innovation/Pages/PoweringIdeas An InnovationAgendaforthe21stCentury.aspx

6. Australian Government Procurement Statement, July 2009. www.financeminister.gov.au/media/2009/docs/Australian_Government_ Procurement_Statement.pdf.

7. PIANOo offers a platform for all contracting authorities to enhance their profes-sionalisation and knowledge through the sharing of information on problems encountered and the discussion of applied solutions.

8. www.senternovem.nl/SBIR

9. There are two sub-programmes in BHIS: the Smart SMEs Market Validation Programme (MVP) (AUD 28 million) and the complementary Smart SMEs Innovation Commercialisation Program (ICP) (AUD 12 million). See Chapter 3.

10. Aghion et al. (2002) find evidence that the degree of product market competition bears an inverted U-shaped relationship to innovation.



11. With respect to fuel efficient vehicle technologies Kleit (2004) reviews the arguments, which are paraphrased here. He observes that many engineering studies suggest that a wide range of technological possibilities exist to improve new vehicle fuel efficiency and which could more than pay for themselves in terms of fuel savings over vehicle lifetimes. Several hypotheses have been proposed to explain why vehicle manufacturers may not adopt technologies that pay for themselves. For instance, consumers might undervalue future savings in gasoline purchases because they lack information or are uncertain about future fuel prices. Oligopolistic profit-maximising manufacturers could also under-supply vehicle attributes even when potential buyers value them. But others argue that manufacturers have incentives to provide improvements in fuel economy that consumers will pay for, and that buyers are reasonably well informed about fuel economy, thanks in part to fuel efficiency labeling schemes. It is also observed that engineering studies alone may give mistaken estimates of the total costs of mandated increases in fuel economy. For instance, they may not capture important costs of implementing a new technology such as marketing and retraining of mechanics. Moreover, auto manufacturers have for decades devoted their technological efforts mainly on improvements to vehicle performance (e.g. acceleration and towing capacity) rather than fuel economy. Foregone performance enhancements of this sort would be part of the real cost of devoting technological development to improving fuel economy. Another potential market failure could arise from non-appropriability of R&D into vehicle fuel economy. If the social return to such R&D exceeds the private return then, to the extent that CAFE standards stimulate innovation, additional welfare gains could be had.



References

Aghion, P., N. Bloom, R. Blundell, R. Griffith and P. Howitt (2002), “Competition and Innovation: An Inverted U Relationship”, Working Paper 9269, www.nber.org/papers/w9269

Aschhoff, B. and W. Sofka (2008), “Innovation on Demand – Can Public Procurement Drive Market Success of Innovations”, ZEW Discussion Paper No. 08-052, Centre for European Economic Research.

Ashford, N.A., C. Ayers and R.F. Stone (1985), “Using Regulation to Change the Market for Innovation”, Harvard Environmental Law Review 9(2),pp. 419-466.

Blind, K. et al., Fraunhofer Institute for Systems Research (2004), “New Products and Services: Analysis of Regulations Shaping New Markets”, European Commission.

Bryden, A. (2010), “Standards Are Boring? Think Twice…”, article published by Paris Tech Review.

Connell, D. and Probert, J. (2010), “Exploding the Myths of UK Innovation Policy: How ‘Soft Companies’ and R&D Contracts for Customers Drive the Growth of the Hi-Tech Economy”, Centre for Business Research, University of Cambridge, Cambridge, United Kingdom.

Department of Trade and Industry (DTI), (2005), R&D Intensive Businesses in the UK, DTI Economics Paper No.11, March 2005, United Kingdom.

Edler, J. (2007), “Demand-based Innovation Policy”, Manchester Business School Working Paper no. 529.

Edquist, C. and L. Hommen, “Public Technology Procurement and Innovation Theory”, in C. Edquist, L. Hommen and L. Tsipouri, (eds.), Public Technology Procurement and Innovation, pp. 5-70, Springer.

European Comission (2010), “Risk management in the procurement of innovation: Concepts and empirical evidence in the European Union”, Expert group report, Directorate-General for Research, Brussels.

Freeman, C., (1974), The Economics of Industrial Innovation, MIT Press, Cambridge, MA.



Geroski, P.A. (1990), “Procurement Policy as a Tool of Industrial Policy”, International Review of Applied Economics 4 (2), 182-198.

Greenberg E., C.T. Hill and D.J. Newburger (1979), Regulation, Market Prices and Process Innovation – The Case of the Ammonia Industry, Westview Press, Boulder, Colorado.

Kleit, A.N. (2004), “Impacts of Long-Range Increases in the Corporate Average Fuel Economy (CAFE) Standard”, Economic Inquiry 42, pp 279-294.

Lerner, J. (1999), “The Government as Venture Capitalist: The Long-run Impact of the SBIR Program”, Journal of Business, 72.

Mahdi, S., P. Nightingale and F. Berkhout (2002), “A Review of the Impact of Regulation on the Chemical Industry, Final Report to the Royal Commission on Environmental Pollution”, SPRU Science and Technology Policy Research, University of Sussex.

Metcalfe, S. and A. James (2001), “Emergent Innovation Systems and the Delivery of Clinical Services: The Case of Intro-ocular Lenses”, ESSY Working Paper, Centre for Research on Innovation and Competition (CRIC) and School of Economic Studies, University of Manchester.

Mowery, D. and N. Rosenberg (1979), “The Influence of Market Demand upon Innovation: A Critical Review of Some Recent Empirical Studies”, Research Policy, Vol. 8, No. 2, pp. 102-153.

National Research Council (2000), The Small Business Innovation Research Program: An Assessment of the Department of Defense Fast Track Initiative, Charles W. Wessner (ed.), National Academy Press, Washington, DC.

Nemet, G.F. (2009), “Demand-pull, Technology Push, and Government-led Incentives for Non-incremental Technical Change”, Research Policy,Vol. 38, pp. 700-709.

OECD (2010a), “Environmental Policy Design Characteristics and Technological Innovation: Evidence from Patent Data”, ENV/WKP(2010)2.

OECD (2010b), Demand-side Innovation Policies, OECD Publishing, Paris.

Swann, P.G.M. (2000), The Economics of Standardisation (Final Report for Standards and Technical Regulations Directorate), Department of Trade and Industry, United Kingdom.

Taylor, M.R., E.S. Rubin and D.A. Hounshell (2005), “Control of SO2 Emissions from Power Plants: A Case of Induced Technological Innovation in the U.S.”, Technological Forecasting and Social Change, Vol. 72 (6), pp. 697-718.



Uyarra, E. and K. Flanagan (2010), “Understanding the Innovation Impacts of Public Procurement”, European Planning Studies, Vol. 18, No.1.

Wallsten, S. (1998), Re-thinking the small business innovation research program. In: Branscomb, L. and Kelle, J, (eds.), Investing in Innovation. MIT Press, Cambridge (MA).

Wallsten, S. (2000), The effects of government-industry R&D programmes on private R&D: the case of the Small Business Innovation Research Program. Rand Journal of Economics, Vol. 31, N°1 (Spring 2000).

5. EVALUATING PUBLIC SUPPORT FOR INNOVATION IN BUSINESS: METHODOLOGIES AND METRICS – 115


Chapter 5

Evaluating public support for innovation in business: Methodologies and metrics

This chapter considers a number of conceptual, practical and institutional issues in the evaluation of direct and indirect support for private sector research and innovation. Challenges and possible approaches to the evaluation of demand-side innovation policy are also examined. The emphasis is placed on methodological challenges, rather than the content of findings from evaluation studies. Examples are drawn from the seven countries considered in this report, although work from other countries is also used where the evaluation record is instructive. The chapter concludes with a series of selected lessons learned.

116 – 5. EVALUATING PUBLIC SUPPORT FOR INNOVATION IN BUSINESS: METHODOLOGIES AND METRICS


Introduction

Effective evaluation of policies and programmes to stimulate research, development and innovation has become increasingly important for policy makers (see Box 5.1). This increased emphasis on evaluation is driven by constraints on discretionary public spending, a greater focus on account-ability and transparency in policy, and the desire to minimise distortions arising from government actions while maximising their impact.

Drawing on evaluations from the countries covered in this study, as well as material from other countries, this chapter reviews methodological and practical challenges in evaluating direct and indirect public support for private sector R&D and innovation. Consideration is also given to the under-explored issue of how to evaluate demand-side interventions. The emphasis is on evaluation methods, processes and policy learning, rather than the content of evaluation findings (although these are referred to at some points). A number of suggestions are offered regarding good practice.

The aims of evaluation can be summarised as follows:

To assess the effectiveness, value for money, efficiency and appropri-ateness of policy and programme interventions ex post, with a view to both shaping and justifying future interventions. Effectiveness refers to whether the programme’s objectives were achieved, and whether their achievement was sufficient to change the original problem situation. Value for money gauges the extent to which benefits exceed costs. Efficiency concerns whether objectives were achieved at the lowest possible cost.1 Appropriateness refers to whether a programme was suitable to the problem situation. For instance, an inappropriate scheme might be found to compete with other initiatives financed from the same budget.

To offer opportunities to a mix of stakeholders – from programme sponsors and managers to beneficiaries – to reflect upon the policies and programmes being evaluated and to make suggestions for improvements.

To provide one of the bases for holding policy makers and pro-gramme managers to account.



Box 5.1. Policy evaluation and monitoring: Examples of recent initiatives

Recent initiatives to strengthen policy evaluation and monitoring in the reviewed countries include the following:

The Danish Ministry of Science, Technology and Innovation has drawn up a framework for research evaluation. The framework addresses a number of overarching questions concerning research evaluations, including, for example, what can be evaluated, how evaluations are to be organised and the principles that the evaluations should follow. To create the highest possible degree of transparency, the Danish Agency for Science, Technology and Innovation has drawn up guidelines that contain a detailed description of the evaluation process, including an overview of when the different stakeholders are involved.

Finland’s Science and Technology Policy Council has initiated an effort to develop a commonly accepted Impact Framework and Indicators for Science, Technology and Innovation (VINDI). Within the framework, the impacts of science, technology and innovation policies are examined in relation to four key themes: i) economic impacts; ii) the accumulation of knowledge, skilled labour and networks of experts; iii) objective and subjective factors of well-being, such as health and social relations; and iv) environmental challenges.

The Norwegian Ministry of Education and Research has begun the process of developing a set of indicators for national research and innovation goals. As a part of this process an independent expert committee – Fagerbergutvalget – has been created. The committee has been given the task of assessing the achievements of publicly funded research. The committee is one of several efforts announced in a new White Paper on research to facilitate a more efficient use of resources in the Norwegian system. The committee´s final report is expected in May 2011. Indicators to measure the goal achievement of the Research Council are also under development and will be set out in the Research Council’s annual report for 2010.

NESTA in the United Kingdom is developing an Innovation Index to improve how the government measures investment in innovation and its impacts. The Pilot Index was launched in November 2009 and is expected to be ready in final form in autumn 2010. The Pilot Index has three components: i) a measure of the amount of investment in innovation in the United Kingdom and its effect on productivity and growth; ii) a tool to understand innovation at firm and sectoral levels; and iii) a set of metrics to monitor the climate for innovation in the United Kingdom. A parallel work stream to measure innovation in the public sector is also underway.



Evaluation addresses fundamental questions regarding the types of programmes that do or do not work, the fundamental design features of programmes that affect performance, and how well programmes fare on a range of efficiency criteria (such as the cost to the public purse of each unit of private R&D spending generated). As such, the distinction needs to be drawn with monitoring and audit approaches. These are forms of process control that examine issues essential for programme management such as whether intended expenditures have been made, whether activities have been performed on time, etc. Studies of this sort are important, as deficient operational practice can shape programme performance. However, process control and other forms of audit should not substitute for economic impact assessment.

Standard methodological challenges to policy evaluation

This section briefly reviews standard challenges to the evaluation of public policy. This serves to introduce the generic methodologies available to meet these challenges, which in turn provides a point of reference and understanding for the subsequent review of how a range of pro-innovation programmes have been, or could be, evaluated.

Evaluations of pro-innovation programmes – from R&D subsidies and tax credits to manufacturing extension and advisory services – face a standard set of challenges. The ultimate aim is to attribute changes in a target group (usually firms, but also institutions and sometimes individuals) to the effect of a given programme. But simply comparing the situation of target groups before and after a programme is insufficient. The programme may be only one among a number of causes of any observed change of state.

The observed change in state in the target group is a result of three sets of influences:

a) The impact of the programme.

b) Unrelated factors – among many possibilities, influences unrelated to the programme might include changes in the policy context, trend changes in the target population (for instance, if firms in a given sector are operating on a learning curve, leading to improvements in performance over time), and fluctuations in the business cycle (in the current economic crisis, for instance, firms might face particular constraints on increasing their R&D spending).2

c) How the programme is observed – observations of a programme might focus on output or outcome variables that are inappropriate, for instance concentrating on employment outcomes rather than more suitable measures of innovation performance. Observations



might be made too early or too late. Furthermore, biases in the way respondents reply to surveys can be problematic. For instance, survey respondents might exaggerate the benefits of a moderately beneficial scheme if they believe that this will increase the likeli-hood of future assistance. And large firms are more likely to respond to surveys than small firms. Observations might also fail to capture the full scope of programme benefits. This might easily be done, for instance, with programmes to stimulate R&D partnerships, where hard-to-quantify changes can occur in the behaviour of participants.

Estimating the counterfactual: what would have happened without the programme?

Identifying changes in target groups and attributing these to the effects of programmes requires knowledge of what would have happened without the programme. Without information on what might have happened to target groups in the absence of a programme, evaluators cannot know that it is the programme that was responsible for any observed changes in state (such as increased outlays on R&D). In other words, they cannot know what was truly “additional” in the observed changes in the target group. Evaluations therefore need to identify genuine programme impacts by discounting changes coming from unrelated factors – such as credit constraints in the wake of the global financial crisis – as well as other changes in the target groups that would have occurred without the programme.

The beneficiaries of pro-innovation programmes usually undergo some form of selection procedure. Accounting for so-called selection bias is a key part of understanding what would have happened to the target group without the programme. Selection bias refers to the possibility that there are some unobserved characteristics of target group members that cause them to be selected and which, at the same time, affect how they will perform in the programme. For instance, researchers with the best ideas might be most likely to receive government support for R&D spending (i.e. to be selected into a support programme). But they might also be more likely to perform well as regards various dimensions of R&D output: because of their good ideas they could have a better chance of attracting other sources of funding; they could have higher incentives to invest their own funds in their good ideas; and they might be more likely to generate good ideas in future.

There are two forms of selection bias. The first is self-selection. For instance, entrepreneurs who seek assistance in the use of new technologies might be inherently forward-looking, open to change and accepting of risks associated with new investments. Such traits could be associated with superior enterprise performance, irrespective of the quality of public policy.



The second form of selection bias stems from administrative selection. Programme administrators may try to maximise the impact of a programme by selecting “good” participants. When this occurs, programme participants might outperform non-participants, regardless of the merits of the pro-gramme. Research has shown that selection bias can be of considerable quantitative importance.

Evaluations essentially take three generic forms. Each has strengths and weaknesses in being able to identify unrelated effects, minimise selection bias, and accurately identify programme impact. The appropriateness of each method should be fully appreciated before proceeding with an evaluation, and in some cases will require in-depth understanding of programme data and econometric method. The different approaches are as follows:3

Experiments Experiments generally represent the highest evaluation standard.

Comparisons are made between firms or persons assigned randomly to two categories: those that receive the programme (the target group) and those that do not (the control group). Random assignment can sift out the influence of extraneous factors because these affect the target and control groups equally. Random assignment also minimises selection bias because selection occurs randomly: observed and unobserved characteristics linked to participation are equally present in the target and control groups. In the example of self-selection cited above, the pro-active forward-looking, risk-friendly entrepreneur is just as likely to end up in the target group as in the control group.

One of the strengths of random assignment is its intuitive appeal: policy makers and the public more broadly are generally familiar with the concept of randomisation, in part from knowledge of medical trials. However, a drawback to random assignment is that some potential programme partici-pants must be overtly excluded from assistance. The method of random assignment might therefore be most acceptable in situations where there is excess demand for support.

Quasi-experiments Quasi-experiments compare the assisted firms or persons and a similar

non-assisted group. However, unlike experiments, the process does not entail random assignment. The identification of a control group occurs after the programme has commenced. As described below, quasi-experimental evaluations have been performed on a number of pro-innovation programmes, including R&D tax credits and manufacturing extension services. The critical issues are how to control for selection bias and how to statistically identify a



control group that is similar to the target group in terms of the mean, range and distribution of characteristics that affect programme outcomes.

Quasi-experiments take many forms. The choice of statistical model must fit the type of data being analysed. Statistically equated controls use multivariate regression techniques to ensure that the target and control groups are as similar as possible with respect to outcome-related charac-teristics.4 Regression discontinuity offers analytic insight by exploiting differences in the probability of selection to the target group. For example, applicants for an R&D grant might be ranked in terms of the quality of their submitted proposals. Assuming this can be done in a consistent way, a quality threshold could be set. Proposals below the threshold are rejected. The difference in outcomes between applicants with accepted proposals and those with rejected proposals is compared, statistically adjusting for the selection variable (proposal quality). Less sophisticated approaches use generic controls, for instance comparing a target group of firms with the entire population of firms in the sector concerned, or matched controls,identifying members of the control group based on their possessing key shared characteristics with the members of the treatment group (such as firm size, sector of operation, owner characteristics, enterprise age, geographic location, etc.). One form of quasi-experiment only considers the target group but relies on members of the target group receiving different intensities, or dosages, of treatment. For example, the hours of participation in a pro-gramme providing technical advice might vary significantly among partici-pants. This variation can form the basis for impact assessment, even if data on non-participants are unavailable.

Participant opinion and expert review This approach involves members of the target group being asked for

their assessment of how the programme has affected performance. Partici-pant opinion can often be an unreliable source of information with which to evaluate programme impacts. To yield valid impact estimates, respondents would have to be able to correctly estimate what would have happened in the absence of the policy or programme.5 But unravelling the impact of a government programme from the many other determinants of innovative performance may be beyond the respondent’s cognitive capacities. The estimation of benefits over time can be particularly problematic. For instance, evaluations of technical extension programmes in the United States show a tendency for firms to overestimate benefits immediately after assis-tance by comparison with views expressed later (Shapira, 2003). Another drawback to reliance on participant opinion is that participants may have reasons for responding strategically. For instance, entrepreneurs might reply positively to surveys that will determine whether a soft-loan scheme is



continued, regardless of the intended use of the funds. Respondents might also feel uncomfortable telling an assessor that a programme was ineffectual. And if innovation is part of a corporate marketing image, respondents might be reluctant to declare that public support is a driver of the company’s innovation effort.

Despite their limitations in attributing impact, interviews with and surveys of participants are essential evaluation tools. Qualitative techniques can help shed light on less tangible programme benefits, such as changes in company research strategy. The qualitative information they yield can be useful in explaining why a programme has or has not worked, beyond quantitative measures of impact.6 In Norway for instance, Godoe and Nygaard (2006) point to system-wide factors that have impeded development of fuel cell technologies, beyond the impacts of a number of individual support programmes.

Furthermore, even within the same country, many pro-innovation programmes are neither standardised nor focused on the same outcomes. And some initiatives, such as those that encourage collaborative research or research networks, have multiple objectives, some of which are not easily quantified.7 Such breadth and heterogeneity pose challenges for quantitative assessment – and benchmarking – and underscore the importance of in-depth qualitative research.

When preparing surveys, methods should be employed to maximise the response rate and minimise response bias. For instance, response bias might be reduced if there is anonymity among respondents. Oldsman (2000) also recommends that: the population to be surveyed should be clearly defined and an appropriate sampling strategy adopted. Survey questions should be worded clearly and address the entire range of programme effects, including immediate, intermediate and final outcomes. Explicit reference should be made to the counterfactual, with multiple questions asked for the same concept using different scales. Responses should be grounded in empirical data, while respondents should not be asked to make calculations. It is also important that the correct respondent is approached during a survey: the business unit that pays the fee to participate in a programme might not be the same unit that receives programme services.

The three generic evaluation approaches described above were presented in descending order of strength of causal inference, generalisability of results and cost. These are important considerations that have to be balanced. Public authorities should use the most rigorous evaluation technique that is practical. But budget constraints may preclude experimental approaches. The high cost of experimental methods means that they are most appropriate when the programme to be evaluated is large, or is new with the possibility of being



significantly expanded. In practice, more than one evaluation method might be used to evaluate a given programme. Indeed, multi-method approaches can afford otherwise unattainable insight and credibility.

Steps in undertaking an evaluation

In practice, evaluation should follow a number of steps. An initial step is to define the scope of the evaluation, its purpose, coverage and the assess-ment criteria to be used. This scoping is best done in consultation with a wide variety of stakeholders. In some instances, however, the scope might be mandated by law and predetermined.

A map or model of the programme logic can be created that links the relevant goals, inputs, activities, outputs and expected outcomes. This should help to define the evaluation’s coverage and assessment criteria, and should provide a description of expected impacts, ordered according to the timing of their anticipated occurrence (e.g. immediate, intermediate, ultimate). Ideally, a programme logic model would have been developed during the programme’s conception. Figure 5.1 shows a highly stylised programme logic model for an R&D tax credit.8 The tax credit is intended to lower the cost of spending on R&D and stimulate additional R&D. This, in turn, is expected to lead to the development and commercialisation of innovative products that meet customer needs (other outcomes could also be expected, relating for instance to process innovation or increases in enterprise productivity. It could also be assumed that companies in receipt of the R&D tax credit will hire additional research workers, and/or other workers needed to meet an expanding demand for the new products). It is further expected that these products will generate increased sales and profitability, as well as creating benefits for consumers.

Figure 5.1. Logic model for an R&D tax credit policy

The programme logic model provides a concise description of how the programme will affect the target group, and thereby presents hypotheses that could be tested in the evaluation. For example, the case of the R&D tax credit involves the fundamental hypothesis that the credit generates new investment in R&D.

and Profitability

Improvedeconomic and

social conditions

Additionalinvestment in R&D

Development ofinnovative productsR&D tax credit Growth in sales



A programme logic model should not preclude opportunities to capture the unintended consequences (positive or negative) of a programme or policy that fall outside the intervention logic. These might include, for instance, changes in outputs, such as the development of new skills within the target firms, or changes in behaviour, for example if firms engage in newcollaborative practices.

Institutions in many countries have introduced systems of project cycle management. Such systems – of which there are many broadly similar permutations – can assist the collection and organisation of information necessary for monitoring and evaluation. They can elicit information that helps to clarify programme objectives and the scope, nature and number of outputs and outcomes. For example, in its support of innovation, from 1986 to 1999, the United Kingdom’s Department of Trade and Industry applied the ROAME system – Rationale, Objectives, Appraisal, Monitoring and Evaluation. After 1999 the evaluation system was modified, in part to incorporate evaluation into a comprehensive system of performance indica-tors, although the underlying ROAME principles remained (Barber, 2009).The proper application of project cycle management systems can also help avoid bad programme designs, such as when project objectives are diffuse, too numerous or contradictory.

A next step is to identify indicators to be used as outcome measures. Measures chosen must be relevant to the initiative being assessed. That is, they must accurately reflect the concept being measured. They must be reliable, subject to as little measurement error as possible. The practicality of data requirements must also be considered. For instance, many pro-innovation programmes aim to enhance productivity. There are a number of ways to measure productivity. These include output per employee, value-added per worker hour and total factor productivity. While total factor productivity is the most complete measure of productivity, its calculation requires data that are usually difficult to obtain. However, data on output per employee are easier to find (even if this measure of productivity could mislead, as a firm’s productivity could appear to rise when it increases out-sourcing). The choice of outcome measure will often entail tradeoffs between relevance, reliability and practicality.

The variety of outcomes of pro-innovation programmes Evaluations of public support for innovation have generally sought to

identify incremental or additional changes in three types of outcome:

Changes in inputs: for instance, participating firms could increase the volume of their own resources dedicated to R&D or other innovation-related activities.



Changes in outputs: innovation outputs might include patents or prototypes, an increase in skills, increased capacity for scientific or technological problem-solving, and changes in firm performance such as productivity, profitability, market share, exports, etc.

Changes in behaviour: firms in receipt of support might undertake changes in collaborative practices, in other research strategies, in the management of R&D, in information search procedures, etc.

With a focus on how government funding of R&D affects company behaviour, OECD (2006a) brings together a set of evaluative studies carried out in ten OECD member countries, plus the European Union, during 2004 and 2005. This work shows that behavioural changes occurred both during and after programme implementation. During project implementation some firms altered decisions regarding when innovation-related projects should begin and how long they should last. In some cases the scale and scope of projects, as well as the difficulty of the research challenge, were enlarged. After programme completion companies sometimes undertook more collaboration. In some instances public support did not achieve stated goals, but capacities were created within firms that were then used to establish research partnerships not originally foreseen, long after the programme had formally ended. Changes in company management routines were also evident.

When to evaluate? Deciding when to perform an evaluation is also important. Different

programmes have different gestation periods. The impact of some pro-grammes might be significant initially, but decline over time. Some initia-tives offer the possibility of almost immediate benefits, such as schemes offering near-to-market funding solutions. Others require months or even years before change is evident, as can occur with collaborative R&D pr-grammes. Programmes can also have long-term effects that go beyond the immediate objectives of the schemes. For example, in the United Kingdom in 1982 the Department of Trade and Industry provided a grant of GBP 2 million to a company named Acorn Computers to develop a computer for use in schools. Decades later, descendents of the technology created by Acorn continued to be developed, including ARM chips used in most mobile phones, a browser used in PlayStations sold outside of the United States and Japan and the protocols for transmitting television over the Internet (Barber, 2009). Research teams in companies assisted by governent may also need to move several times before they end up in a business fully able to exploit the results of their work. Indeed, a high-cost programme with long-term positive effects could easily fail a cost-benefit test if evaluated soon after programme implementation. Accordingly, the time period over which an evaluation is performed might need to exceed the



duration of participation in the scheme by months or even years, depending on the type of programme being examined.

Early evaluation has the advantages that programme participants are easily reached, data and information can be more readily available and details of the programme more easily recalled. Firms that applied but were not accepted to a programme might be more easily traced. Participating firms expecting future benefits might also be more responsive to surveys or other data requests (Barber, 2009). By contrast, late evaluation can encounter difficulties in data collection and in tracing key individuals. Late evaluations may also suffer from “survivor” bias. Part of the solution is to ensure that good systems of data collection are built into programmes from their inception. If well designed, such systems need not impose significant reporting burdens on programme beneficiaries. Another response is to conduct an interim evaluation based on monitoring returns and/or a pre-liminary survey, with a more considered evaluation implemented later. It may be appropriate for instance to show that a programme has achieved certain well chosen intermediate objectives or results which are known from academic research or case studies to be likely to lead to certain economic and social benefits. Whether an interim evaluation is appropriate could depend on a variety of considerations, including the policy cycle – noting too that time horizons can differ across stakeholders, including levels of government –, the resources available for evaluation, the size and signi-ficance of the programme and the burden which evaluation places on programme participants.

When the timeframe of expected impacts is greater than the period covered by the evaluation the evaluator may need to project the estimated benefits outside the period of the data. Smith (2004) notes that while theory or evidence from other evaluations of similar programmes can help to guide this decision, the best course will probably be to make cost-benefit estimates that assume multiple plausible durations for programme benefits.9

Drawbacks to using performance standards as a measure of programme impact

Administrative performance standards typically consist of quantitative measures of programme outputs (such as the number of firms receiving technology-related advice) or outcomes (such as an increase in enterprises’ market share). Their use has become more prevalent since the 1990s (being particularly important in the United States government as a consequence of the Government Performance and Results Act (GPRA) of 1993).



Performance standards have many uses. They can for instance provide managers with essential information on the extent to which a programme is meeting its institutional remit (for instance, to respond to information requests from firms within a given time period). However, reliance on performance standards as a proxy for impact estimates requires evidence of a systematic relationship between the two. But there is no reason why such a relationship should exist (see Smith, 2004). This can be illustrated by considering a hypothetical programme that aims to help young technology-based firms obtain risk finance. Assume that the programme uses as its performance standard the percentage of firms able to sign a financing agreement with a venture capital fund within 24 weeks of participation in the programme. For this performance standard to serve as a valid proxy for the programme’s impact it must be positively correlated with programme impact. But the reverse may be true. To see why, consider the extreme case of easy-to-serve firms who would always find finance on their own and hard-to-serve firms who never would, but do so half the time when they participate in the programme. Assume too that in a given year the programme serves half of each type of firm. The programme’s rate of “signing a financing agreement” is 0.75, because the easy-to-serve all obtain finance, as do half of the hard-to-serve, while its impact is 0.25 (reflecting the fact that the programme only creates benefits for the hard-to-serve clients). But in a second year the programme might only cater to hard-to-serve firms. Then, its rate of “signing a financing agreement” is 0.50 but its impact is also 0.50, because none of the hard-to-serve firms would have found finance without the programme. In other words, achievement against the performance standard has fallen, but the impact has risen. For this pro-gramme, the performance standard provides a poor substitute for programme impact (and a strong incentive for programme managers to choose to work only with easy-to-serve firms).

Similar issues arise in government programmes that subsidise private R&D, which are often judged against the fraction of projects that are commercialised. Such a performance standard provides an incentive for programme managers to pick projects that would have been funded anyway without the subsidy. To meet their public policy remit, such schemes should be financing projects with (potentially) large spillovers but low private benefits. In a now well-known evaluation, Wallsten (2000) examines the United States’ Small Business Innovation Research (SBIR) program and suggests that each dollar of programme funding crowds out a dollar of private research funds, with no impact on the total amount of research undertaken. This follows from a focus on choosing projects likely to succeed in the market, the programme’s performance standard.



The above discussion also underlines that the choice of performance standards for pro-innovation programmes will create incentives that shape the behaviour of programme providers. When the continued funding of programmes depends on performance targets being met, inappropriate yardsticks can have sometimes unforeseen effects on implementation and impact. For example, if the performance of a scheme to link firms with research centres is assessed against the number of firms entering the programme an incentive exists to increase client throughput. Firms might be encouraged into the scheme – or be accepted when they should be dissuaded – regardless of probable outcomes. However, careful selection of performance standards can help focus programmes in order to achieve overarching goals. The performance standards used need to be sufficiently varied to account for the different effects and tradeoffs that a given programme involves. And weightings could be assigned to different performance metrics, based on the utility of the incentives created. For instance, to raise the additionality of programmes, higher weightings might be assigned to outcomes affecting hard-to-serve groups such as smaller or geographically remote firms.

Limitations of ex ante evaluation Evaluation is mostly concerned with what has happened at some point in

the past. However, the evaluation function must have a future orientation, because a key reason for doing evaluation is to improve the quality of future policies. Evaluators should be encouraged to consider the implications of evaluation findings for various possible courses of policy. For instance, “prospective” evaluations might be concerned with the diminishing returns found in some programmes when they increase in scale. Evaluators should clearly articulate such insights for decision-makers.

While policy should build on evaluation findings – among other inputs - evaluative evidence can require careful interpretation. Serious pitfalls could accompany attempts to forecast programme outcomes ex ante, based on past evaluation results. For instance, depending on the nature of the programme, a positive evaluation finding under one set of circumstances may not be replicable under others. Even evaluations using random assignment only produce estimates that are internally valid: that is, they do not apply to the same programme operating at a different scale or in significantly dissimilar conditions. Indicative of the caution that needs to be exercised in reading evaluations, Lentile and Mairesse (2009) review factors that have led to major differences in the findings of studies of R&D tax credits. These include:



Differences in policy efficiency.

Differences in study methodology.

Differences in key country characteristics, for instance in corporate tax policy, industry structure, and regulatory contexts.

The time period over which the study has been performed, which might be important if economic conditions or institutions have under-gone change over time, and which also affects the likelihood that long-term programme effects have been captured.

Publication biases, with articles that achieve publication more likely to show some programme impacts rather than no impacts.

The existence of divergent evaluation findings across similar programmes urges against mechanical interpretation of future programme results based on past studies. An additional example of this concerns the longstanding issue of whether public financing of R&D is complementary to or crowds out private spending on R&D. Various studies have produced results that are inconclusive. However, David and Hall (2000) suggest that the answer may depend in part on a country’s starting point, with the relationship being more complementary when public sector R&D accounts for a large share of all R&D, when the supply of qualified scientific labour is relatively abundant, and when public support is skewed towards grants rather than public contracting. Policy analysts must therefore be aware of the parameters that shape reported evaluation findings, and the sensitivity of outcomes to changes in those parameters.

Examples of evaluations of different forms of innovation support

Evaluating the impacts of research in universities and public research organisations

Cozzens (1994) examined the methods used to evaluate public invest-ment in research and grouped these into three categories:

Retrospective tracing of the knowledge inputs to a particular innova-tion: such approaches can shed light on the relative contribution to an innovation of different sources of knowledge as well as the role of important sources of institutional or other support. However, this approach has various drawbacks: subjective judgement is called upon to assess the importance of different inputs to the innovation process; the method is costly, especially if the evaluators seek a number of cases sufficient to permit statistical inference; and the scope of the impacts from R&D is underestimated. This restriction of



scope occurs because single products of R&D are traced backwards to their multiple inputs, instead of R&D inputs being traced forward to their multiple outcomes, which can also include research dead ends that in themselves provide valuable information for innovators (Georghiou and Roessner, 2000).

Research outputs measured in patent counts, citations, bibliometrics, etc.: a large literature examines the usefulness of such output measures and shows that each measure has strengths and weaknesses. For instance, biases in patent counts arise from: differences across countries in the economic costs and benefits of patents (stemming for example from variations in the rigor of the patent examination, the size of the applicable market and subject matter coverage); differences across technologies and sectors in the importance of patents as a source of protection against imitation; and differences across firms in the propensity to patent, particularly in the patenting of unimportant innovations (Pavitt, 1988). Despite drawbacks to the use of patent counts, citations and bibliometrics, a set of such indicators taken together can provide valuable information on programme effectiveness and efficiency. Difficulties exist however in relating this set of output measures to economic and market outcomes (although patents that are more frequently cited in other patents tend to be more valuable [Harhoff et al., 1999]). It is these outcomes that are of greatest interest to policy makers concerned with the allocation of public resources.

Changes in productivity, national income or social welfare. A number of relatively recent studies indicate that public R&D spending is positively related to growth in output and productivity. Typically, a production function approach is used, where public R&D is one of a number of inputs in equations explaining productivity or growth. Guellec and van Pottelsberghe (2004) offers an example of this approach, and distinguishes the effects of private sector, public and foreign R&D on productivity and growth in 16 countries between 1980 and 1998. However, not all support for R&D aims to foster economic growth (support for medical research, for instance, does not generally have this goal). Furthermore, the relative contributions of basic and applied research – which is important for policy analysis – can be difficult to disentangle.

Linking R&D to policy, Parsons and Phillips (2007) review eight multi-country studies of the relationship between direct support and privately funded or performed R&D. Four studies conclude that the relationship is positive, with three finding that the relationship is positive only in some countries or under certain specifications of the models used.



Various studies attempt to estimate a social rate of return to investments in R&D. These measure social benefits as the sum of consumer and producer surpluses.10 However, calculating these surpluses requires information on supply and demand curves, which is problematic when entirely new or radical innovations are generated (Georghiou and Roessner, 2000). Furthermore, being based on individual case studies, such studies are costly and hard to generalise from. Externalities from the R&D are also difficult to capture. And uncertainties exist over the correct discount rate to use, given that benefits to R&D might accrue over long periods, the duration of which might be unknowable a priori.

The methodological constraints noted above suggest that multi-method approaches – combining case studies and the various forms of econometric research – are required to best estimate the returns to public support for R&D.

Evaluating support for collaborative R&D A characteristic of many schemes to support collaborative R&D is that

they entail multiple objectives. These can also be wide-ranging11 and difficult to measure (such as the disposition to collaborate). The evaluative evidence of benefits to collaborate R&D has relied primarily on case studies. If policy failures are less likely to be published than policy successes, selection bias among these case studies may be a problem. Three examples of evaluations of collaborative research initiatives are reviewed here.

Georghiou and Roessner (2000) describe an evaluation of the National Science Foundation’s Engineering Research Center (ERC) programme in the United States. The programme seeks to encourage industry-university research collaboration. The evaluation employed case studies, a focus group made up of beneficiary companies, a survey of around 500 companies based on case study and focus-group findings, and follow-up phone interviews. Some notable results relevant to survey-based evaluations of programmes for collaborative R&D were that:

Often, only a small number of persons in each firm were in a posi-tion to assess the benefits of participation. Consequently, the nature of each respondent’s involvement with the programme needs to be established at the outset;

The business unit that pays for programme participation was not necessarily in a position to judge programme impacts;



Most companies had not tried to place a financial value on the benefits of participation (a finding common to evaluations of other innovation support programmes). They evaluated participation as a dynamic process giving rise to an array of possible benefits that changed over time.

An evaluation of the aforementioned CARAD programme in the United Kingdom describes a diverse set of benefits. The programme provided support for collaborative applied research in aerospace, with programme expenditure of over GBP 270 million (USD 415 million) between 1990 and 2002. Thirty-nine organisations each received over GBP 500 000 (USD 779 000). The context was one in which aerospace manufacturers developed knowledge and expertise of a spectrum of new technologies, each of which might or might not be incorporated into future aircraft and aero-engine platforms. This made it difficult for evaluators to assess the programme on the basis of new products (Barber, 2009). The rationale for the programme responded to what was described as “the high degree of technical and financial risk over timescales of 10-15 years or more in researching and developing new civil aircraft and associated equipment. A contributory factor was the reluctance of capital markets to invest in research where they lack understanding of high technology, the risk and the business operations” (BERR, 2008a). The evaluation combined in-depth interviews and a survey and found that, among key benefits:

Participating companies adopted a wider perspective on the range of possible technologies to cover, exceeding those that even the largest firms could research on their own. Financial support encouraged the aerospace companies to increase the amount and variety of their research. Being able to research a wider choice of technologies would improve their ability to meet future market needs and regula-tory standards.

Financial support enabled smaller specialist companies to become suppliers to the aerospace sector, thereby increasing their techno-logical capabilities, which in turn were used to help the innovation efforts of their non-aerospace customers.

However the evaluation was unable to identify CARAD’s contribution to all of the programme’s qualitative objectives. Nor was it able to assess whether projects funded by CARAD produced more benefits than other forms of R&D support.

Also in the United Kingdom, the LINK programme, established in 1982, has for long been the main government programme encouraging collabora-tive research between companies and research organisations. Malik et al.(2006) undertook an evaluation based on interviews, a questionnaire and



case studies. An issue highlighted by this study, and relevant to others, was the problem of having to isolate the effects of the programme from those of other government initiatives. Many of the companies interviewed had received numerous government grants and other types of innovation support over time. The value of face-to-face interviews was also highlighted, as they revealed details of complex programme effects that were unlikely to have been brought to light by a survey instrument alone.

Evaluating R&D tax credits R&D tax credits have been evaluated in many countries using diverse

methods. Useful overviews of the literature are provided by Lentile and Mairesse (2009) – on which this subsection draws heavily – and Hall and van Reenan (2000).

For the purpose of policy assessment, firms cannot legally be excluded from a tax incentive to which they are entitled. This removes the possibility of evaluating R&D tax credits by constructing a control group using randomisation. Evaluations have therefore been based on the following approaches:

Surveys.

Quasi-natural experiments.

Techniques using statistically constructed control groups.

Structural econometric modelling.

The evaluations using surveys suffer from the limitations of survey methods noted earlier. Entrepreneurs or managers might be unable to accurately assess the genuine impacts of the scheme, distinguishing these from many other possible determinants of R&D spending. Long-run effects might be ignored, especially if the survey is administered shortly after the commencement of the policy. And respondents could also have strategic reasons for overstating or understating programme impacts. Surveys, moreover, have typically been based on small sample sizes.

Another set of evaluations of R&D tax credits uses econometric techniques that exploit discontinuities in the process of administrative selection of target firms. If for instance there is a ceiling on the application or rate of the tax credit, then firms operating above that ceiling could be used as a control group. A comparison over time can be made between the change in R&D spending among firms below the ceiling – following introduction of the scheme – and the change in R&D spending over the same period among firms above the ceiling (the so-called “difference in differences” method). An evaluation of Norway’s R&D tax credit followed



this approach (Haegeland and Moen, 2007). That evaluation concluded that the scheme, introduced in 2002, increases private spending on R&D.

The third evaluation approach statistically constructs a control group. Analysts have recorded observable characteristics of firms that have used the tax credit, and then identified firms that match those characteristics but did not use the tax credit. The difference in R&D spending in the two groups is attributed to the tax credit. However, even using a comprehensive set of matching criteria – such as firm turnover, age, sector of operation, geographic location, etc. – perfect comparability between the two groups is not possible. There may always be some unobserved differences between beneficiaries and non-beneficiaries that also affects the policy outcome. For instance, in all observable respects two firms may have the same probability of receiving a tax credit, but the abilities of their respective managements may differ. This might lead the managerially able firm to use the tax credit, and the managerially less able firm not to. The managerial abilities in question might also be associated with a higher propensity to innovate. Evaluations that use matching-methods also require that there be a sufficient number of firms that qualify for the tax credit but fail to apply.12

A fourth approach to evaluating R&D tax credits has been structural econometric modelling. Developed by the United States’ Government Accounting Office, this approach uses models of R&D investment behaviour and assumes that R&D spending is a function of the cost to the firm of the capital used.13 The modelling first seeks to estimate the sensitivity of the cost of capital used to the R&D tax credit. In a second step, estimates are made of the sensitivity of firms’ R&D spending to changes in the user cost of capital. Lokshin and Mohnen (2009) use structural econometric modelling to assess the R&D tax credit in the Netherlands (WBSO). Their results indicate that for SMEs, each euro of the tax credit generates EUR 0.2 in additional R&D. The figure is considerably lower in large firms, at around EUR 0.07.

Mairesse and Lentile (2009) point out that R&D tax credits also produce outcomes other than increased R&D. These include: decisions to begin investing in R&D for the first time; changes in the productivity of R&D; changes in the wages of researchers; and social welfare improvements (taking into consideration all direct and indirect economic effects of the policy). However, these outcomes are rarely assessed in policy evaluations. Nevertheless, with respect to the countries covered in this report, Mairesse and Lentile identify a number of evaluations that do consider the less-frequently assessed outcomes: in Norway, Haegaland and Moen (2007) found that after the introduction of the R&D tax credit, firms that had not previously invested in R&D experienced a 7% increase in their probability of doing so. In Canada, Czarnitzki et al. (2005) found evidence both for and



against the proposition that R&D tax credits generate increased productivity or profitability. And in both Norway and the Netherlands, evaluations found a positive but small effect on the wages of research workers.

For their literature review, Mairesse and Lentile (2009) found only one full-blown cost benefit analysis of an R&D tax credit: a study of the scheme in Canada by Parsons and Phillips (2007). This evaluation sought to separately quantify and then combine five effects of policy. The study indicates a median increase in social welfare of 11 cents for each dollar of tax credit. However, the authors observe that variations in assumptions underlying the estimates of each policy effect can lead to net outcomes that are either positive or negative, leading to the conditional conclusion that the “tax credit likely generates positive net economic benefits under a reasonable range of assumptions.”

Despite the large number of evaluations, comparability across studies is hazardous. In part this reflects variations in method. Problems of compara-bility also arise because of variation in the design of tax incentive schemes: some are based on R&D volume and others on incremental R&D spending. The generosity of schemes also varies, and can vary within the same country for different segments of the enterprise population (in the Netherlands and Norway, for example, the tax credit rate is higher for small firms). The scheme in the Netherlands also targets the wages of R&D workers as well as social insurance liabilities. A variety of additional eligibility criteria can also apply. These might favour co-operative research, or the hiring of young researchers. Programme objectives also vary, from increasing the overall volume of R&D spending, to encouraging firms to begin investing in R&D, to stimulating the recruitment of R&D workers and attracting/retaining direct investment by foreign firms engaged in R&D. This diversity of designs and goals underscores the importance of establishing the right evaluation metrics.

Evaluating other pro-innovation policies Chapter 4 outlined a variety of other forms of public support for

innovation. These range from schemes to promote equity financing for investment in innovation, to technical extension services, programmes to facilitate access to research-related advice and information, and technology incubators.



Programmes supporting access to finance Cowling et al. (2008) undertook an econometric evaluation of the United

Kingdom’s Enterprise Investment Scheme (EIS) and the Venture Capital Trust (VCT). The EIS and VCT were introduced as a response to perceived failures in the market for equity finance for small and young firms with high growth potential. Within predetermined threshold amounts, the EIS provides tax relief of different sorts (on income, capital gains and share disposal) for individuals who invest in shares of certain types of small high-risk unquoted companies. By contrast, the VCT encourages investment in young enter-prises indirectly. It does this by giving income tax relief on investment in venture capital trusts (a retail venture capital product for private investors, similar to an investment trust). This evaluation is interesting for a number of reasons. Evaluations of policy instruments to facilitate small firms’ access to credit have been relatively numerous. But assessments of tax-based initia-tives in the market for risk capital are relatively few. Moreover, the empirical basis for the evaluation was far-reaching: a pooled set of panel data, covering up to 90 000 firms, sampled annually for up to 11 years (with an average of 4-6 years). The evaluation used a matched control group of non-recipient companies. In addition, outcome measures drew on admini-strative data held by HM Revenue and Customs, rather than subjective survey-based assessments. The outcome measures included trading performance (profitability and sales), enterprises’ capital structure, labour productivity and business survival.14

Technical extension programmes Extension services – sometimes referred to as industrial modernisation

programmes – have been designed with a variety of objectives, from facilita-ting adoption of state-of-the-art technologies to upgrading know-how in day-to-day business management, marketing or training. Extension programmes cater to firms presenting a great variety of constraints and opportunities. Most programmes focus on the needs of SMEs.

The March 1996 issue of Research Policy focused on the evaluation of industrial extension programmes. Shapira (2003) also affords an insightful review of evaluation practices pertaining to manufacturing extension. These studies mainly focus on programmes in the United States, where schemes operate in all States and in some cases have existed for over forty years. Owing in part to an evaluation mandate associated with federal co-financing, the record of evaluation of extension programmes is particularly strong in the United States.



Evaluations of extension programmes have taken most of the possible methodological forms, with the exception of random assignment experi-ments. The studies by Jarmin (1999) and Oldsman (1996) provide examples of quasi-experimental evaluations that clearly describe the steps taken to outline programme logic and select appropriate output and outcome metrics.

The evaluations by Jarmin (1999) and Shapira and Youtie (1998) also illustrate different approaches to gathering information on control-groups. The former study exploits data on non-client firms available from the United States Census Bureau’s Longitudinal Research Database. By contrast, the latter study used a survey instrument sent to all manufacturers with more than ten employees in the State where the programme operated (with responses then weighted to reflect the principal features of the state’s enterprise demography). Client-firm productivity was the outcome measure in Jarmin’s evaluation, whereas Shapira and Youtie estimated a range of public and private outcomes using multiple methods. These outcomes included private benefits such as growth of sales, reductions of inventory and energy savings. Estimated public benefits comprised federal, state and local taxes. Displacement effects were also calculated, on the basis that some part of the increases in sales among client firms could come at the expense of non-client firms. All benefit and cost data were combined to estimate an overall cost-benefit ratio.

Feller et al. (1996) observe that, as concerns extension programmes, unexploited opportunities exist to employ randomised experiments. For instance, schemes could vary the set of services they offer to clients in a randomised way. Or, to assess dosage effects, client firms could be randomly assigned to groups with different limits on the duration of the services received. Shapira et al. (1996) considered that, overall, the evaluation effort should better identify the impacts of specific services, among the many offered, rather than assess the performance of entire programmes. This could facilitate resource allocation within programmes.

Programmes facilitating access to research-related information or expertise

Programmes that seek to link research centres and firms often have more than one goal. For instance, a programme might aim to facilitate commerciali-sation of university research while also augmenting firms’ access to university-based expertise. Given multiple programme objectives, evaluators need to isolate specific outputs and outcomes and select proper metrics for both. But whether quantitative techniques are best used will depend on the scope and quality of available data. Where a programme is small, structured case studies are likely to be the most suitable evaluation technique.



Many evaluations of such schemes exist. For example, the study by Roessner et al. (1996) evaluates the Center for Advanced Technology Development using multiple methods and data types. It included a survey of clients, case studies and a benchmarking exercise using three methods. The evaluation highlighted the limitations to using benchmarking across schemes that are barely comparable owing to the diversity of programme designs.

Technology incubation Technology incubators are widely used by sub-national governments to

promote high-tech businesses and local economic development. Policy makers have turned to business incubation to meet a wide range of policy goals – from raising enterprise birth rates, to commercialising university research, to expanding the supply of infrastructure. Technology incubators typically possess features that differentiate them from mixed-use incubators. They are frequently affiliated with a University and often have selective entry criteria focusing on businesses with high growth potential. They can also be more expensive to establish and run on account of specialised facilities and higher staffing costs. Their service offering may also include a greater emphasis on services related to intellectual property. Nevertheless, in both mixed-use and technology-oriented incubators job creation is a frequent goal of publicly supported schemes.

Rigorous studies of the impacts of technology incubation are scarce. Many evaluations serve a promotional function for incubator sponsors or associations. In preparing this review, no examples were found of evaluations that used control groups. Challenges exist however in identifying member of a possible control group. For example, many databases that contain information on non-tenant firms do not cover enterprises that fail early on. Most firms that fail early, for instance, will not have registered with Chambers of Commerce. The following approaches have been suggested to finding a set of firms that might comprise a control group:

When accepting a tenant into the incubator, incubator managers could seek to identify a similar local non-tenant firm at the same time. Assisted firms may themselves be in a position to help identify comparable enterprises.

Firms that have applied to enter an incubator might have different characteristics from firms that operate in the same sector but which have not sought external assistance. Incubator managers could there-fore seek to monitor firms that are accepted into the incubator but which for administrative or other non-substantive reasons failed to take up a place.



Control groups might also be identified from among the clients of enterprise support agencies located close to the incubator – if agencies maintain appropriate data. Such agencies may have worked with early-stage firms similar to those assisted by incubators.

An incubator containing firms with affiliates might use data collected from the affiliate enterprises if these share significant common charac-teristics (and are sufficiently numerous to form a control group).

There is a need to critically consider the outcome measures commonly used in evaluations of incubation schemes. The focus of technology incuba-tion should be on different dimensions of enterprise and technology develop-ment, recording such things as: the time that enterprises need to establish market niches or develop new products; the adoption of advanced manage-ment practices; the use of new or superior technologies; the number of patents registered; cost reduction resulting from technology developed through the incubator; the number of research projects transformed into business opportunities; the volume of royalties obtained by the incubator, university or research centre, as a result of projects supported by the incubator; etc.. Programmes can also have long-run indirect effects that are difficult to measure. For example, technology-oriented incubators can sensitise aca-demics to the problems of industry, or provide demonstration effects to university staff regarding the commercialisation of research.

Many sponsors of incubation programmes focus on employment impacts. This measure however is often not appropriate. Employment growth will follow successful commercial outcomes. Incubator managers should work with performance metrics, and thereby management incentives, closely tied to commercial and technological outcomes. Moreover, anecdotal evidence suggests that most job creation occurs after tenant firms graduate from their incubators (most incubators operate graduation schedules that begin three to five years after entry).

Business survival is also sometimes used as a measure of impact. But survival rates will not be a suitable yardstick if many of the surviving firms are marginal survivors. In addition, many entrepreneurs whose ventures fail learn from the experience and establish successful firms at a later date, while business terminations with losses to creditors are often only a fraction of total exits. This means that closure cannot always be equated with failure. Furthermore, survival probabilities tend to correlate negatively with the age of a business. So an incubator with low failure rates among older tenant firms might not be performing as well as an incubator with higher failure rates among younger firms. Above all, it is difficult to gauge the signifi-cance of improved survival rates, and indeed other outcomes, if firms enter the scheme after a selection process, as is typical. In this case, the observed



success of incubated companies may be due to intrinsic characteristics rather than the effects of incubation.

Even if the problem of creating a control group is surmounted, improved business performance across a variety of dimensions is only to be expected in a group of firms that receives assistance. For policy evaluation purposes, what matters is the relationship between the costs incurred and the benefits generated. Unfortunately, as regards technology incubation, such work appears to be lacking.

Evaluating demand-side programmes and policies

The evaluation of demand-side policies poses particular evaluation challenges: how for instance might the evaluator create a control group to assess the impacts of a technology-oriented regulation or standard? How can the influence of supply-side drivers of innovation be separated out when assessing the role of demand? Might the technology-oriented regulation have a positive effect on innovation, but be inefficient overall? And how can the secondary benefits of technology-oriented public procurement be properly captured, such as increases in private financial support for firms that win procurement contracts? This section briefly considers the evalua-tion record, and a series of methodological issues, relating to demand-side policies.

Technology-oriented regulations Various studies have assessed the technological effects of regulatory

policy. One of the most closely examined instances concerns regulations on minimum fuel economy standards for vehicles. Research in the United States indicates that Corporate Average Fuel Economy (CAFE) regulations introduced in 1978 have led to increased average automotive fuel efficiency (Kirby, 1995). Cost-benefit analysis of a similar policy in Japan has also been undertaken (Kainou, 2007). However, no equivalent research was found for the countries examined in this review.

A feature of the available studies of vehicle efficiency regulations is that they do not focus on innovation, but seek instead to assess overall costs and benefits of the regulations. This reflects the secondary nature of the innovation goal in such regulations, the primary objective being the provision of a public good (such as clean air).

The study by Kleit (2004) provides a detailed economic cost benefit analysis of the CAFE regulations. This study usefully illustrates the scope of the costs, benefits and changes in consumer behaviour that need to be accounted for. For instance, because increased vehicle efficiency lowers the



marginal cost of driving, the regulation induces more driving. Consequently, estimates of the policy’s costs need to include the effects of more pollutant emissions, accidents and congestion caused by increased driving.

This and other such studies underscore that regulatory policy of this sort, even if effective, need not induce innovation. In the CAFE scheme, increases in average vehicle fuel efficiencies could be achieved through manufacturers changing relative car prices so as to sell fewer large cars and more small cars. Depending on price elasticities of demand, automotive manufacturers might deploy more costly but already existing technologies that otherwise would be unprofitable. Indeed, to take an example from the domestic electrical appliances sector, in the United States an average standard size refrigerator/freezer consumed 2000kWh of electricity per year in 1972. Following the introduction of energy efficiency standards the equivalent refrigerator consumed 690 kWh per year in 1993. At least initially, this increase in efficiency occurred with no technological innovation, relying instead on rationalisation using existing technology (Colombier and Menanteau, 1997).

Even in cases where regulation spurs innovation, regulation-based policy might be cost-ineffective overall. Alternative policies might achieve the overarching public policy goal more cost effectively. For example, Kleit (2004) finds that a long-run increase of three miles per gallon in the efficiency standard set by CAFE creates welfare losses of USD5.6 billion per year and saves 5.1 billion gallons of gasoline per year. An 11 cent per gallon increase in the gasoline tax would save the same volume of fuel at a welfare cost of just USD 275 million per year. Clearly, conflicts could arise between a regulation-based policy that accelerates technological innovation, and the most efficient means of meeting the broader policy goal.

A related aspect of the evaluation challenge is to isolate the influence of a regulation from other determinants of innovation. Changes in consumer preference, for instance, might cause manufacturers to invest in improved vehicle efficiency. The problem of establishing causality can be made more difficult by the possibility of long lead times between a regulatory stimulus and an industry response. For example, Greenberg et al. (1979) found a six year lead period in the ammonia industry and were unable to isolate regulatory effects using an econometric model. Establishing causality is further complicated by the possibility of a technical hiatus on innovation – such as the current slowdown in the rate of discovery of antibiotics – that may be difficult to overcome irrespective of the regulatory context. Further-more, the fact that an innovation plateau could affect all firms in the regulated sector simultaneously raises the question of how to compare the effects of the regulation against a realistic alternative basis (other than asking firms what they think would have happened without the regulation).



During this review, no studies were encountered that attempted a control-group based evaluation of innovation-related regulatory policy.15

Case study evidence has been gathered demonstrating that anticipation of regulatory change can induce innovation. Case studies of asbestos product development (Ashford et al., 1985) and SO2 removal technologies (Taylor et al., 2005) indicate that innovation took place before specific regulations came into effect but after their passage into law had become probable.

Technical standards Various macro-economic studies have examined the impacts of

standards on trade and growth (Department of Trade and Industry [2005a] was cited earlier in this report). These studies relate changes in the incidence of standards over time and across sectors to changes in economic performance. However, such studies do little to illuminate the relative importance of the many causal routes through which standardisation might affect economic growth. Another set of studies explore the role of standards in developing markets and facilitating competition. And numerous case studies examine the effects of standards in different industries (see Swann [2000], section 1.6, for an extensive list of references). What such research generally does not do however is to evaluate the effects of government support for standardisation. This may be because the public sector’s role is in fact rather limited. While important, this role largely involves measures to include under-represented groups in the process of developing standards, and subsidisation of teams drafting international standards.

Policy makers might consider monitoring the number and age distribution of the national stock of standards, with a focus on standards developed by national standards bodies. Public standards – as distinct from proprietary standards created by individual firms – have the important features of openness and credibility. They are open in that they are equally available to all. And they are credible because government sponsorship enhances confidence that their use will become widespread. Policy action could be called for if the number, rate of creation and age distribution of the stock of public standards were seen to diverge greatly from norms in other advanced economies.

Technology-oriented public procurement Evaluations of technology-oriented public procurement are scarce. Edler

et al. (2009) cite an evaluation of policies in Sweden intended to augment the demand for and supply of energy efficient products. This work was based on interviews and case studies. Econometric evaluation using actual



or constructed control groups would be challenging, as the policy stimulus affects all firms in the sector concerned, and the stimulus is also not neatly divisible in ways that might permit assessment of dosage effects. The findings of various qualitative assessments of technology procurement programmes illuminate the importance of certain administrative features, including for instance whether procurers develop in-house technology-related competencies or have legal authority to contact potential suppliers to learn about technological possibilities (Edquist et al., 2000).

While not an evaluation of a specific policy, Aschhoff and Sofka (2008) sought to quantify the effects of public procurement on innovation, and to compare these effects with other determinants of innovation. Innovation was measured as the share of turnover achieved with products possessing market novelties. The study examined general rather than technology-oriented procurement. A survey of 1 100 innovative firms in Germany was used, with effects differentiated by firm size, industry and geographic location. The survey data were self-reported and subjective, raising problems in connection with possible response biases as well as the accuracy of responses (see discussion above). However, the methodology used was the same as that of the Community Innovation Survey, which has been pre-tested and piloted in a number of countries. The validity and reliability of the responses was therefore relatively well understood. A comprehensive non-response analysis was also undertaken of over 4 000 firms. This showed no systematic differences between responding and non-responding firms with respect to innovation activities. The model used could not rule out selection bias however.

SBIR-type programmes Bound and Puttick (2010) examine whether the United Kingdom’s

Small Business Research Initiative (SBRI) is helping to stimulate innovation. This scheme has two aims: to help alleviate a financing gap for early stage high-tech ventures; and to facilitate small-firm access to public contracts for procurement of pre-commercial R&D. The SBRI is a model process that involves identification of a public policy problem, an open competition awarding R&D contracts to promising solutions offered by small firms, applications for further prototype development for proposals that passed the feasibility stage, and final public procurement, market commercialisation (or both). Roll out of the current SBRI model began in early 2009. The assessment does not claim to afford an impact assessment, but rather to provide qualitative insights on performance that could also be evaluated quantitatively later. The study method entailed 30 interviews. The evaluation found that government departments have been able to widen the search for solutions. For instance, in response to a need in the National



Health Service for better detection of drug-resistant pathogens and improved hand-cleaning among staff, a small company was able to utilise technology developed in the food processing industry. It was also seen that the SBRI offers credibility for potential follow-on investments from the private sector. The study highlights qualitative factors affecting performance of the scheme, such as the importance of expeditious decision-making in awards procedures, and notes the potential value of developing user networks to facilitate public sector use of the programme. To facilitate genuine economic impact assessment, the study also calls for an open data policy, including the collection of data on applicants who did not win contracts.

As noted, the SBRI seeks to alleviate a financing constraint on a specific class of small firms and facilitate their access to public procurement markets. Accordingly, in this and similar programmes in other countries, it would seem appropriate to develop evaluation metrics that explore both of these outcomes.

Detailed information might be gathered for instance on the value of contracts awarded to target firms before and after the programme’s inception in the sectors concerned. Data on patent awards among beneficiaries – or responses to previously tested innovation survey questions – might be compared across beneficiaries and non-beneficiaries to try and gauge innovation-related impacts. If the SBRI also aims to offer credibility for follow-on investors, patterns of debt and equity financing might be compared among grant recipients and non-recipients (venture funding is one of the criteria used to assess the United States’ SBIR programme). Creative steps might also be taken to measure how the procurement process itself has become more efficient, possibly relying on structured qualitative informa-tion from procurement agencies.

Other demand-side policies In specific markets a range of measures are open to governments to

stimulate demand. In the energy sector these could include guaranteed tariffs and specific power purchase agreements with local utilities. Targeting consumers, governments might offer rebates, for instance on energy efficient products, as has happened in many countries with Compact Fluorescent Lamps. Governments might also promote comparison labelling (to inform consumers on the relative efficiency of products) or endorsement labelling (e.g. “CFC-free”). The general research literature has encountered methodo-logical difficulties in establishing the relative importance of technology-push and demand-pull influences on innovation (in part due to inadequate panel data and to the sensitivity of studies to assumptions regarding the rate of depreciation R&D as a stock of knowledge). Additional challenges exist in attributing innovations to changes in demand conditions deliberately stimu-



lated by policy. Detailed industry-specific case studies are likely to represent the most persuasive evaluation tool. Considering wind power technologies, Nemet (2009) provides a useful example such a case study, and attempts to elucidate the role of multiple policy drivers on demand for product innovations. The choice of study location, California, was strategic, as that State represented more than 90% of the global market for wind power in the study period (late 1970s and 1980s). This makes it reasonable to link product innovations to demand-based policies instituted by State authorities. The plausibility of the link is heightened by the fact that international efforts at technology development in this sector were minimal at the time.

Interestingly, Nemet finds that the number of highly cited patents declined even as the market rapidly increased in scale. He proposes three possible explanations. First, the sector experienced convergence on a dominant windmill design – three blades mounted upwind on a vertical axis – with subsequent innovations occurring within this configuration. With hundreds of millions of dollars being invested in dominant-design turbines, the attention of wind-farm developers appeared to focus on maximising the return on the stock of dominant-design machines. Secondly, uncertainty in the longevity of policy reduced the strength of demand-side factors. And thirdly, a decline occurred in federal spending on wind energy R&D. This decline was found to correlate closely with a fall in highly-cited wind energy patents.

Evaluation of policy systems

While much of the policy discussion on evaluation has focused on applying quantitative methods and tools to assess impacts of discrete policy interventions or instruments, evaluation also concerns entire research portfolios or the overall research and innovation system. For example, Finland has a long tradition of assessing policy instruments and science and technology organisations. This evaluation culture started to emerge in the latter half of the 1980s and has strengthened in the past decade. Finland’s Science and Technology Policy Council (STPC) was established to handle important matters concerning research, technology and their utilisation. The STPC also addresses issues relating to the development and exploitation of evaluation and impact analysis. The STPC is chaired by the Prime Minister and includes seven other ministers with remits relevant to science and technology policy. The STPC also commissions system-level assessments. Since the early 1990s, all the stakeholders of the Finnish innovation system – ministries, funding organisations, universities, government R&D institutes, public and private intermediaries etc. – have been evaluated, some more than once. However, policy decisions are almost never based only on



the results of evaluations. But evaluation often provides a basis for debates on development needs and other issues.

More generally, in a number of countries, international peer review of institutions is increasingly being used for assessment of entire innovation systems.

Selected policy lessons

This section seeks to summarise a number of generic lessons learned on the practice of policy and programme evaluation.

Overt recognition of the importance of evaluation by senior policy makers is essential

Governments often draw lessons from social science research and policy evaluations in a selective way (Feller, 2000). Indeed, institutional resistance to evaluation is not uncommon. There are a number of reasons for this. Perhaps the principal concern among programme managers and implementing agencies is that support will be withdrawn if programmes receive a negative assessment. But the direct and indirect costs of evaluation can also produce aversion to assessment. In addition, the more systematic evaluations have often been useful in judging whether a policy has worked or not, but have been weak in describing how the policy might be improved. This may have diminished the perceived usefulness of evaluation among programme administrators.

Developing a culture of open policy learning is essential. To facilitate this, regular government reporting of performance should draw extensively from evaluation results. Human and financial resources for evaluation are more likely to become available once recognition of the critical role of evaluation is evident. In Australia, all public policy proposals require the inclusion of an evaluation strategy.

Encourage greater openness to evaluation by making clear that the aim of evaluation is to improve the quality of public policy. Evaluations that reveal problems should not provide automatic grounds for termination of a programme. Rather, evaluation should be viewed as a tool to provide a basis for improved policy.

Take a strategic approach to evaluation A small number of long-term well-designed evaluations and sectoral

studies using controlled samples of beneficiaries and non-beneficiaries is likely to yield more generalisable policy-relevant results (with lower response burdens on beneficiaries) than multiple evaluations of varied



quality employing disparate methods. Shapira et al. (1996) noted that, in the United States, variety in the quality of evaluations of manufacturing extension programmes in part reflected the practice of developing evaluation procedures without referring to existing best practice. There is a need to standardise evaluation practices around good practice, such that generalisations can be more easily derived from the mass of evaluative work done.

Evaluations should seek to explain causality and rule out alternative explanations. Being strategic about evaluation could facilitate this, for instance by testing the replicability of results in similar settings.

Another feature of a strategic approach to evaluation would be to have grant agencies establish data collection and programme design features that facilitate impact assessment (for instance exploiting possible discontinuities in selection probabilities discussed above).

Evaluation should be timed to coincide with the need for information in decision making cycles. Moreover, the involvement of multiple stakeholders who are addressed in the evaluation’s findings and recommendations can also contribute to its success.

An additional strategic consideration is that frequent changes of policy direction will hinder evaluation. Data will be hard to accumulate, the responsiveness of programme managers may be diminished, and firms or other potential beneficiaries can become confused.

Plan evaluation at the inception of programmes All significant programmes should be required to develop a formal

evaluation strategy. Preparing for and undertaking evaluation is in itself a pedagogical exercise that has the potential to improve thinking about programme logic. Planning evaluation before beginning a programme will allow data strategies to be put in place. But thought should also be given to: specific policy questions that may need to be addressed; how findings should be communicated; and procedural issues having to do with evalua-tion timelines, budgets and the possible role of outside contractors.

Consider situations when evaluation might not be merited In some circumstances evaluation might be a poor public investment.

These include the following:

Some programmes, such as extension-type schemes, have been evaluated in rigorous ways in a number of countries. Other schemes are notably under-evaluated, including for instance the array of demand-side initiatives and schemes to encourage access to equity for



small innovative firms. As noted in this chapter, careful reading of the specific parameters of evaluation studies and their findings is needed to gauge ex ante applicability to new initiatives. However, for schemes in which the evaluation record is rich, new evaluations of similar programmes may not represent a good public investment.

Evaluation may not represent a good investment when the sample sizes of target and non-target groups are too small for statistical inference. Statistics provides formal methods for determining the number of observations needed to identify impacts of a given magni-tude with a certain degree of confidence under a range of assumptions about the amount of variation in the outcome metric. These methods should be applied before deciding to invest in evaluation.

Investments in evaluation may be wasted if data quality is inadequate. For example, caution is required if an evaluation must rely on survey data with low response rates, because the non-responses may be selective. In such situations, the evaluation effort may need to invest first in generating better quality or alternative data.

And, obviously, there is little sense to investing significant resources in evaluating programmes that will in any event be discontinued.

Develop competencies necessary to undertake and interpret statistically sophisticated evaluations

In general, evaluation should not be done by “in-house” staff alone. Programmes should also be evaluated by independent external experts (ideally, the body that implements the evaluation would work with pro-gramme managers but would not be dependent on continued contracts from the sponsor of the programme). In-house skills are essential however in writing proper terms of reference for evaluations, in interpreting results and in presenting these to policy makers in a technically amenable and policy-relevant way.

Choose the evaluation techniques in the light of the size and nature of the programme concerned

Assessments of outcomes using experimental evaluation approaches are expensive. As such, they should be used to assess either the largest pro-grammes, or pilot programmes that might be significantly enlarged.

Technical choices regarding the econometric evaluation methods to use should be driven by the nature of the available data. Matching methods briefly referred to earlier in this chapter require rich data sets on participants and non-participants. Longitudinal methods are obviously inappropriate to



cross-sectional data. They are also unsuited to situations where the selection of participants into a programme depends on transitory conditions (such as firms deciding to enter an R&D grant competition because they have been experiencing particularly good – or bad – recent performance).

Smith (2004) notes that, where appropriate, results should be reported presenting multiple estimates that make explicit any necessary range of assumptions used. He observes that doing so allows readers with different views to assess the results under their own preferred assumptions. Reporting multiple estimates will also oblige policy makers to confront uncertainty about programme performance and provide a sense of the robustness of any recommendations concerning the policy.

Studies of major programmes should ideally use a variety of methods, from random assignment to interviews with beneficiaries and other relevant stakeholders. The approaches described in the paper are not mutually exclusive. Similar results from different methods can add to the credibility of findings. Moreover, qualitative research should be used to complement quantitative techniques, providing insights into the specific causal mechanisms that come into play in generating outcomes.

Make evaluation data public Placing evaluation data in the public realm, possibly in anonymised

form, could facilitate (cost-free) academic analysis and serve as a form of evaluation quality control. Furthermore, in the pluralistic policy systems that characterise OECD countries, evaluation should be expected to serve the knowledge needs of a wider set of actors beyond programme managers and public policy officials.



Notes

1. Programmes can have economies of scale and scope. So efficiency might be enhanced by operating fewer general purpose schemes. But given that firms, sectors and their innovation requirements are highly diverse, a larger number of specialised programmes may be required. Evaluation should help to identify a suitable balance (Barber, 2009).

2. If that were to occur, and crisis-induced budgetary restrictions also caused governments to restrict the generosity of R&D tax credits, then a spurious correlation could arise. The appearance could be given that the tax credit is a determinant of private R&D outlays.

3. This overview draws on Oldsman and Hallberg (2004).

4. Selection bias is accounted for using two-stage regression, instrumental variable techniques or propensity score matching (Heckman and Robb, 1985).

5. This may be possible in some restrictive cases where causal links in the programme’s operation are direct and self-evident – for instance the introduction of a new computer-controlled machine on an assembly line leading to faster throughput.

6. Quantitative techniques could also help elucidate why programmes have worked if sufficient data were available on important and varying features of programme design.

7. For instance, Wixted and Holbrook (2008) review approaches to evaluating formal research networks such as Australia’s Co-operative Research Centres or Canada’s Networks of Centres of Excellence. They conclude that assessment criteria must include network governance – encompassing resource generation and the use of progress markets, among other features – and the extent of varied forms of network connection to research communities (depending on the nature of the network).

8. See Oldmsan and Hallberg (2004).

9. Smith (2004) also observes that the cost-benefit literature has produced a variety of estimates regarding the rate at which future programme benefits should be



discounted. Accordingly, reporting based on the use of multiple plausible rates is advised.

10. The consumer surplus is the benefit experienced by consumers from being able to purchase a product at a price less than the maximum that they would be willing to pay. The producer surplus is the benefit had by producers from selling at a price above the minimum that they would be willing to sell at.

11. For example, the United Kingdom’s Civil Aeronautics and Technology Demonstration (CARAD) programme provided support for collaborative applied research in Aerospace and had the following broad and diverse goals: “To improve UK industry competitiveness and market share by developing world-class technology from which to launch successful ventures in world civil aerospace markets; To make UK industry an attractive partner in civil collaborative projects in Europe and internationally; To encourage longer term industrial investment in innovation without early prospect of commercial return; To help maximise spin-off to civil aviation from military research and demonstration support; To increase collaboration with higher education institutions and promote maximum “pull through” of academic research; To help meet global standards e.g. for reduced emissions (aircraft & aeroengines); To study the scope and recommend ways of disseminating CARAD technology to non-aerospace sectors in UK, and other R&D programmes (1990-2000); To promote the participation of small and medium sized enterprises (SMEs) in collaborative ventures (1990-2000).” (BERR, 2008a).

12. Lentile and Mairesse (2009) note that such a situation might arise if firms are unaware of the policy, their R&D spending does not qualify for the tax credit, applying for the tax credit is excessively complex or costly, or they fear that applying for the tax credit could increase the probability of a tax audit.

13. This cost includes the interest foregone on the money used, economic depreciation, gains or losses stemming from inflation, corporate taxes and the specific fiscal rules governing amortisation of R&D, as well as the tax credit (Hall and Jorgenson, 1971).

14. The findings of the study indicated that EIS and VCT investments have had a positive effect on capacity building and expansion in sales in recipient companies. However, in absolute terms these effects were small. There was some limited evidence of profit-enhancement. In both schemes, differential impacts were found depending on the size, age and sector of recipient firms.

15. This task might be facilitated however in countries where regulations are enacted differently across subnational jurisdictions, as might occur in some federal states.



References

Aschhoff, B. and W. Sofka (2008), “Innovation on Demand – Can Public Procurement Drive Market Success of Innovations”, ZEW Discussion Paper No. 08-052, Centre for European Economic Research.

Ashford, N.A., C. Ayers and R.F. Stone (1985), “Using Regulation to Change the Market for Innovation”, Harvard Environmental Law Review 9(2), pp. 419-466.

Barber, J.M. (2009), “What Lessons Can Evaluation of Support for Innovation under ERDF Learn from Experience of National Innovation Policies?”, http://ec.europa.eu/regional_policy/conferences/evaluation2009/abstracts/ barber.doc

BERR (2008), Evaluation of Regional Selective Assistance (RSA) and its successor, Selective Finance for Investment in England (SFIE), March 2008 www.bis.gov.uk/files/file45548.pdf

BERR (Business Enterprise and Regulatory Reform) (2008a), Evaluation of the Civil Aeronautics and Technology Demonstration (CARAD) Programme, www.bis.gov.uk/files/file46284.pdf

Bound, K. and R. Puttick (2010), “Buying Power? Is the Small Business Research Initiative for Procuring R&D Driving Innovation in the UK?”, research report, National Endowment for Science, Technology and the Arts, United Kingdom, June.

Colombier, M. and P. Menanteau (1997), “From Energy Labelling to Performance Standards: Some Methods of Stimulating Technical Change to Obtain Greater Energy Efficiency”, Energy Policy, Vol.25, No. 4, pp 425-434.

Cozzens, S.E., (1994), “Methods for Evaluating Fundamental Science”, DRU-875/2-CTI, Critical Technologies Institute Report for the White House Office of Science and Technology Policy, Washington, DC.

Cowling, M., P. Bates, N. Jagger and G. Murray (2008), “Study of the Impact of Enterprise Investment Scheme (EIS) and Venture Capital Trusts (VCT) on Company Performance”, HM Revenue & Customs, Research Report 44, www.employment-studies.co.uk/pdflibrary/hmrc44.pdf

Czarnitzki, D., P. Hanel and J.M. Rosa (2005), “The Impact of R&D Tax Credits on Innovation: An Empirical Analysis Based on a Survey on Innovation in Canadian Manufacturing”, Innovation Analysis Bulletin 7(2), 4-6.

David, P.A. and B. Hall (2000), “Heart of Darkness: Public-Private Interactions Inside the R&D Black Box”, Research Policy, Vol. 29, pp. 1165-1183.



Department of Trade and Industry (DTI), (2005), The Empirical Economics of Standards, DTI Economics Paper No.12, June 2005, United Kingdom.

Edler, J., L. Georghiou, E. Uyarra and K. Blind (2009), “Monitoring and Evaluation Methodology for the EU Lead Market Initiative: A Concept Development”, University of Manchester/Fraunhofer ISI, Technical University, Berlin.

Edquist, C., L. Hommen and L. Tsipuri (2000), “Policy Implications”, in C. Edquist, L. Hommen and L. Tsipuri (eds.), Public Technology Procurement and Innovation, pp. 301-311, Springer.

Feller, I. (2000), “The Academic Policy Analyst as Reporter: The Who, What and How of Evaluating Science and Technology Programs”, in P. Shapira and S. Kuhlmann, Proceedings from the 2000 U.S.-European Workshop ‘Learning from Science and Technology Policy Evaluation’, Georgia Institute of Technology, Fraunhofer Institute, www.cspo.org/library/presentations/?action=getfile&file=217...lib

Feller, I., A. Glasmeier and M. Mark (1996), “Issues and Perspectives on Evaluating Manufacturing Modernization Programs”, Research Policy 25,309-319.

Georghiou, L. and D. Roessner (2000), “Evaluating Technology Programs: Tools and Methods”, Research Policy, Vol. 29, pp. 657-78.

Godoe, H. and S. Nygaard (2006), “System Failure, Innovation Policy and Patents: Fuel Cells and Related Hydrogen Technology in Norway 1990-2002”, Energy Policy, Vol. 34, pp. 1697-1708.

Guellec, D. and B. van Pottelsberghe (2004), “From R&D to Productivity Growth: Do Institutional Settings and the Source of Funds Matter?”, Oxford Bulletin of Economics and Statistics, Vol. 66, pp. 353-378.

Haegeland, T. and J. Moen (2007), “Input Additionality in the Norwegian R&D Tax Credit Scheme”, report 2007/47, Statistics Norway.

Hall, B.H. and J. van Reenen (2000), “How Effective are Fiscal Incentives for R&D ? A Review of the Evidence”, Research Policy, Vol. 29, pp. 449-469.

Hall, R. and D. Jorgenson (1971), “Application of the Theory of Optimum Capital Accumulation”, in G. Fromm, Tax Incentives and Capital Spending,Brookings Institution, Washington.

Harhoff, D., F. Narin, F. Scherer and K. Vopel (1999), “Citation Frequency and the Value of Patented Inventions”, The Review of Economics and Statistics81 (3), pp. 511-515.



Heckman, J. and R. Robb (1985), “Alternative Methods of Evaluating the Impact of Interventions: An Overview”, Journal of Econometrics, Vol. 30, pp. 239-267.

Jarmin, R.S. (1999), “Evaluating the Impact of Manufacturing Extension on Productivity Growth”, Journal of Policy Analysis and Management,Vol. 18, No. 1, pp. 99-119.

Kainou, K. (2007), “Quantitative Policy Evaluation of the Top Runner Method Regulations for Fuel Consumption Standards for Passenger Cars in Japan by Cost-benefit Analysis”, Research Institute of Economy, Trade and Industry, www.rieti.go.jp/en/publications/summary/07030011.html

Kirby, G.E. (1995), “An Evaluation of the Effectiveness of US CAFÉ policy”, Energy Policy, Vol. 23, No. 2, pp.107-109.

Kleit, A.N. (2004), “Impacts of Long-Range Increases in the Corporate Average Fuel Economy (CAFE) Standard”, Economic Inquiry 42, pp. 279-294.

Lentile, D. and J. Mairesse (2009), A Policy to Boost R&D: Does the R&D Tax Credit Work ?”, European Investment Bank Papers, Vol.14, No. 1.

Lokshin, B. and P. Mohnen (2009), “How Effective Are Level-based R&D Tax Credits? Evidence from the Netherlands”, Updated UNU-MERIT Working paper no.29, University of Maastricht.

Malik, K., L. Georghiou and H. Cameron (2006), “Behavioural Additionality of the UK SMART and LINK Schemes”, in Government R&D Funding and Company Behaviour: Measuring Behavioural Additionality, OECD, Paris.

Nemet, G.F. (2009), “Demand-pull, Technology Push and Government-led Incentives for Non-incremental Technical Change”, Research Policy,Vol. 38, pp. 700-709.

OECD (2006a), Government R&D Funding and Company Behaviour: Measuring Behavioural Additionality, OECD Publishing, Paris.

Oldsman, E. (1996), “Does Manufacturing Extension Matter? An evaluation of the Industrial Technology Extension Service in New York”, Research Policy Vol. 25, pp. 215-232.

Oldsman, E. (2000), “Measuring Employment Impact of Sustainable Enterprise Development”, presentation to ILO meeting, Geneva, 22 May.

Oldsman, E. and K. Hallberg (2004), “Evaluation Business Assistance Programs”, in OECD, Evaluating Local Economic and Employment Development: How to Assess What Works Among Programmes and Policies, OECD, Paris.

Parsons, M. and N. Phillips (2007), “An Evaluation of the Federal Tax Credit for Scientific Research and Experimental Development”, Department of Finance, Government of Canada, September.



Pavitt, K. (1988), “Uses and Abuses of Patent Statistics”, in A.F.J. van Raan, Handbook of Quantitative Studies of Science and Technology, Elsevier Science Publishers, Amsterdam.

Roessner, D., Y. Lee, P. Shapira and B. Bozeman (1996), Evaluation of Iowa State University’s Center for Advanced Technology Development, Volume I: Assessments of Benefits and Costs Associated with CATD, School of Public Policy, Georgia Institute of Technology, Atlanta, June. Summary available at: www.cherry.gatech.edu/mod/pubs/catdx.pdf.

Shapira, P. (2003), Evaluating Manufacturing Extension Services in the United States: Experiences and Insights, in P. Shapira and S. Kuhlmann, “Learning from Science and Technology Policy Evaluation”, Edward Elgar, Northampton, Massachusetts and Cheltenham, United Kingdom.

Shapira, P., J. Youtie and D. Roessner (1996), “Current practices in the evaluation of US industrial modernisation programs”, Research Policy 25, pp. 185-214.

Shapira, P. and Youtie, J. (1998), “Evaluation Industrial Moderization: Methods, Results and Insights from the Georgia Manufacturing Alliance”, Journal of Technology Transfer, 23(1), Spring, 17-28.

Smith, J. (2004), Evaluating Local Economic Development Policies: Theory and Practice, in OECD (2004), Evaluating Local Economic and Employment Development: How to Assess What Works among Programmes and Policies, OECD Publishing, Paris.

Swann, P.G.M. (2000), “The Economics of Standardisation”, Final Report for Standards and Technical Regulations Directorate, Department of Trade and Industry, United Kingdom.

Taylor, M.R., E.S. Rubin and D.A. Hounshell (2005), “Control of SO2 Emissions from Power Plants: A Case of Induced Technological Innovation in the U.S.”, Technological Forecasting and Social Change, Volume 72 (6), pp. 697-718.

Wallsten, S. (2000), “The Effects of Government-industry R&D Programmes on Private R&D: The Case of the Small Business Innovation Research Program”, Rand Journal of Economics, Vol. 31, No.1, Spring.

Wixted, B. and A.J. Holbrook (2008), “Conceptual Issues in the Evaluation of Formal Research Networks”, CPROST Report 2008-01, Simon Fraser University.

ORGANISATION FOR ECONOMIC CO-OPERATION AND DEVELOPMENT

The OECD is a unique forum where governments work together to address theeconomic, social and environmental challenges of globalisation. The OECD is also at theforefront of efforts to understand and to help governments respond to new developmentsand concerns, such as corporate governance, the information economy and the challenges ofan ageing population. The Organisation provides a setting where governments can comparepolicy experiences, seek answers to common problems, identify good practice and work toco-ordinate domestic and international policies.

The OECD member countries are: Australia, Austria, Belgium, Canada, Chile, theCzech Republic, Denmark, Estonia, Finland, France, Germany, Greece, Hungary, Iceland, Ireland,Israel, Italy, Japan, Korea, Luxembourg, Mexico, the Netherlands, New Zealand, Norway, Poland,Portugal, the Slovak Republic, Slovenia, Spain, Sweden, Switzerland, Turkey, the United Kingdomand the United States. The European Union takes part in the work of the OECD.

OECD Publishing disseminates widely the results of the Organisation’s statistics gatheringand research on economic, social and environmental issues, as well as the conventions,guidelines and standards agreed by its members.

OECD PUBLISHING, 2, rue André-Pascal, 75775 PARIS CEDEX 16

(92 2011 08 1 P) ISBN 978-92-64-11565-1 – No. 59663 2011

www.oecd.org/publishing

Please cite this publication as:

OECD (2011), Business Innovation Policies: Selected Country Comparisons, OECD Publishing. http://dx.doi.org/10.1787/9789264115668-en

This work is published on the OECD iLibrary, which gathers all OECD books, periodicals and statistical databases. Visit www.oecd-ilibrary.org, and do not hesitate to contact us for more information.

-:HSTCQE=VVZ[ZV:ISBN 978-92-64-11565-192 2011 08 1 P


This study is concerned with trends in and key features of policies and programmes used by governments to support innovation in the business sector. In addition to identifying good practices across a range of programme types, it compares business innovation policies across several countries, with a particular focus on Canada.

Contents

Chapter 1. Business innovation policies: Findings and policy conclusions

Chapter 2. Public support for business R&D

Chapter 3. Non R&D-based public support for business innovation

Chapter 4. Demand-side policies to support innovation: Trends and challenges

Chapter 5. Evaluating public support for innovation in business: Methodologies and metrics

Business Inno

vation P

olicies S

ELE

CT

ED

CO

UN

TR

Y C

OM

PAR

ISO

NS
