ISSN No. 2454 - 1427

CDE Sept 2018

Public Private Partnerships Vs. Traditional Roads Project Delivery Time, Costs and Quality

Ram Singh Email: ramsingh@econdse.org Department of Economics Delhi School of Economics

Working Paper No. 290 http://www.cdedse.org/pdf/work290.pdf

CENTRE FOR DEVELOPMENT ECONOMICS DELHI SCHOOL OF ECONOMICS

DELHI 110007

0

Public Private Partnerships Vs. Traditional Roads

Project Delivery Time, Costs and Quality

Ram Singh

Centre for Development Economics

Delhi School of Economics

University of Delhi

Email: ramsingh@econdse.org

ABSTRACT

The Public Private Partnerships (PPPs) have become a mainstay of plans of the Centre and the

State governments towards infrastructure development. In this article, we discuss the various

considerations behind the rampant use of PPPs for infrastructure services. Next, we empirically

examine a widely held belief that PPPs are better than the traditional approach towards

infrastructure in that they can deliver superior quality infrastructure at a faster rate and lower

costs. Using a dataset of 313 national highways projects, we compare the performance of the

PPP and the traditional government (non-PPP) road projects. We show that the project delays

are relatively short for PPPs, but the cost overruns are significantly higher for PPP roads than

for the government managed road projects. The data and other available evidence analysed in

the paper suggests that the quality of PPP roads is superior to the government roads. However,

the overall quality of road services under PPPs is deficient on several counts.

ACKNOWLEDGMENT: I am thankful to J V Meenakshi for valuable comments and

suggestions. I have also benefitted from several suggestions provided by the research team of

the International Growth Centre (IGC). The institutional support by the Centre for

Development Economics at the Delhi School of Economics, and funding support by the IGC

are gratefully acknowledged. Kriti Jain provided excellent research assistance.

1

1. Introduction

During the last two decades, the successive governments at the centre have used private funding

to develop infrastructure. 1 The present government at the centre has also used PPPs with an

aim to develop a wide and extensive networks of Roads, Ports, and Airports.2 Private funding

has also become a mainstay of plans of the state governments towards infrastructure

development. The private funding is tapped through what are called the Public Private

Partnerships (PPPs).

In a typical PPP for an infrastructure project, the government provides land and is responsible

for regulatory clearances for the project. The private partners take responsibility to build the

project assets such as roads, railways, ports, airports, etc. Generally, the private partners are

required to fund the project upfront, but the government retains the ownership of the assets

developed through PPP. Moreover, the private partners are responsible for the upkeep of the

assets during operation and maintenance phase of the project. In any case, in a PPP the control

and managements rights over the project assets are assigned to the private partner during the

terms of the contract. Many a times the private partners are also granted ‘concessions’ (rights)

to charge user-fees. All these rights are returned back to the government at the end of the

contract period.

In contrast, under the traditional contracts used for infrastructure projects, the government is

responsible for the funding as well as maintenance of assets. Besides, it retains all the control

and ownership rights over the entire life cycle of the project.

In other words, compared to the traditional approach, the PPP arrangements are substantially

different in terms of the responsibility and rights of the private sector towards financing,

building, and maintaining of the assets needed for infrastructure services. In terms of

nomenclature, PPPs can be of several types, namely BOT (build, operate and transfer), DBOT

(design, build, operate and transfer), BOOT (build, own, operate and transfer), etc. These forms

differ mainly in terms of the decision and control rights delegated to the private investors during

the contract period.3

One of the several questions addressed in this article is: Why have the PPPs become so

important for upgradation of the infrastructure? There can be several reasons behind the use of

PPPs such as the political considerations of the government or lobbying by the ‘interest groups’

to use PPPs for real-estate and other private projects. Financial constraints faced by the centre

and the state governments is another possible reason. These reasons are analysed in the next

section.

However, the official case for the use of PPPs depends on their perceived economic superiority

over the traditional approach, that is, the belief that the PPPs can deliver public goods at

relatively low life cycle costs. The draft of the National Public Private Partnership Policy

(NPPPP) clearly states two of these objectives as:

1 See GoI (2006 a, 2006b and 2007) 2 For example, the vision 2020 document for the Indian Railways says: “Railways would have to make heavy

investments for the expansion of the network, modernization and upgradation of the technology and for

providing world class facilities to the customers in the coming years. ... we have to rely on the private sector

heavily .... have started many PPP schemes.” Indeed, the private funding has become the mainstay of policies of

the centre and the state governments toward infrastructural development. 3 For more on types of PPPs, see Yescombe (2007), and Grimsey and Lewis (2004).

2

“Harness private sector efficiencies in asset creation, maintenance and service delivery;

Provide focus on life cycle approach for development of a project, involving asset

creation and maintenance over its life cycle;…”4

Indeed, the policymakers seem to believe that private sector participation through the PPPs will

ensure a speedier delivery of infrastructure as it will reduce the life cycle project costs.

Moreover, it is believed that the PPPs have the potential to incentivize the private sector to

deliver a superior quality infrastructure services. At least officially, these beliefs seem to have

been the driving force behind the advent and expansion of the PPP programme. The following

quote amply illustrates the official belief in superiority of PPPs:5 ... it was agreed that for ensuring provision of better road services i.e. higher quality of

construction and maintenance of roads and completion of projects without cost and time

overrun, contracts based on BOT model are inherently superior to the traditional EPC

contracts....

Are the above mentioned official beliefs in the superiority of the PPPs justified? There seems

to be no empirical study to answer the question. In this paper, we attempt to address the

question using a dataset of 313 projects. In Section 3 we compare the potential sources of

efficiency of PPP contracts vis-à-vis the traditional contracts. The focus of this section is on

comparison of different types of contracts in terms of allocation of project risks and control

rights. We also examine and compare the incentive structure for the contractor under the PPPs

and the traditional procurement contracts.

Several sectors such as ports, energy, tourism and urban development have attracted a fairly

large number of PPPs. However, in these sectors the PPP and the non-PPP projects have been

implemented by different agencies under very different policy schemes; majority of PPPs have

been formed by the state governments. Moreover, most of the projects are still under

implementation. Therefore, these sectors are not amenable to statistical analysis comparing

performance of PPPs with the non-PPPs.

The largest number of PPPs have been formed in the road sector. The state as well as the central

government has used PPPs for building and expanding road networks under their respective

jurisdiction.6 While the state governments have formed PPP for the state highways and

expressways, the centre has sponsored PPPs for national highways under the National Highway

Development Programme. The PPPs sponsored by the state governments differ across the

states. Besides, different states have applied different engineering standards based on the local

needs. Again, it is hard to make a plausible comparison.

In contrast, PPPs for national highways projects have been implemented by only one agency,

namely the National Highways Authority of India (NHAI). In fact, the NHAI accounts for the

largest number of PPPs executed by any one agency in the country. So, in order to compare the

comparable, PPP and the non-PPP road projects implemented by the NHAI is a plausible choice

for empirical analysis. We use a dataset of 313 national highways projects to compare

performance of the PPPs with the traditionally procured highway projects.

4See National Public Private Partnership Policy, Ministry of Finance, GoI, Page 8. 5 This is an excerpt of a decision made on 15 March 2005 in a meeting (chaired by the Prime Minister)

regarding financing of the National Highways Development Project (NHDP). See GoI (2006a). 6 As per the official sources, between December, 2005 -September, 2017 the Public Private Partnership

Appraisal Committee (PPPAC) approved as many as 257 in number with total project cost (In crore)302,388.00

in value. See https://www.pppinindia.gov.in/

3

Dataset and summary statistics are described in Section 5. It shows that the project delivery

time is shorter for PPPs but the construction costs are relatively high. Section 6 discusses the

possible reasons behind the high construction costs for PPPs. In Section 7 we empirically

examine the outcomes under PPPs and the traditional contracts, in terms of the project delivery

time and the construction costs. Section 8 presents conclusions and policy implications.

2. Why PPPs?

The term PPPs is used to describe a wide range of contractual arrangements between the

government and the private sector for the provision of public goods.7 Therefore, government

agencies can use PPPs guided by several considerations. First of all, PPPs can be used by the

‘interest groups’ to get easy access to land. Land is a crucial factor for many private projects,

e.g., housing projects of developers, private schools and hospitals. At times it is difficult for

the project developers to buy the required land directly from the owners. Otherwise also, the

developers can find it cheaper to use the government to acquire the land since the compensation

cost under land acquisition law can be well below the market value of land.8 This was especially

the case in the past, so there was general temptation to induce the government acquire the land.

The PPPs came in handy in this exercise. Ostensibly, PPPs were formed to harness private

funds to provide public goods. However, many of these partnerships were used as a disguise

by the project developers and the state governments to acquire land for private projects. The

land was acquired in the name of public purpose but a substantial part of it was used for real-

estate and other commercial purposes not directly related to the public good in question.

Housing projects under PPPs for the Taj and the Ganga expressway projects, and hospitality

projects clubbed with Delhi and Mumbai airports are some of the leading examples of such

practices.

Perceived inefficiency of the public sector organisations can be another important reason

behind the use of PPPs. Such perception has contributed to privatisation of several services like

distribution of water and electricity. However, for several public goods and infrastructure

facilities, an outright privatisation can trigger public outrage. In such a scenario, PPPs provide

a politically expedient middle path. They allow transfer of operation and maintenance rights

over the public goods to the private sector. Since the government retains the ownership of assets

so PPPs shield the government from protests against privatisation.

It is also argued that a developing country like India can benefit from participation of foreign

firms in infrastructure. Foreign firms are expected to possess superior construction equipment

and techniques. The spill overs from advanced construction technology can reduce the overall

cost of building the infrastructure. Moreover, the foreign firms can can help boost infrastructure

investment.

Financial constraints faced by the governments can be yet another important reason. The use

of PPPs enables the government to tap private funds which can be used to expeditiously

upgrade the infrastructure facilities. Otherwise, it is argued, the society will have to wait for

longer period before the government can provide funding using its own resources.

However, this argument does not stand scrutiny since on average at least three-fourth of the

funds invested by private partners in PPPs is borrowed from the market. This option is available

7 See Netter and Megginson (2001) for a review of the literature on the subject. 8 For details see Singh (2012).

4

to the government as well. Besides, compared to the private sector, the cost of borrowing is

always less for the government agencies. In other words, relative to the government funding of

infrastructure, the direct cost of funds is higher under the PPP route. Moreover, if the bidding

is not competitive, the government may not be able to get best value for money from the private

sector participants leading to higher cost for the tax-payer in the long run.

At the same time, statutory obligations, say under FRBMA do restrict the scope of market

borrowings by the government.9 Under such constraints, the PPP offers a channel to the

government for a quicker supply of public goods and services. Public goods such as

infrastructure serves as an input for the other sectors of the economy. An improvement in the

provisions of these services lead to a reduction in cost for economic activities in the rest of the

economy. So, the use of private sector resources to provide a better and efficient infrastructure

can boost the growth potential and increase competitiveness of the economy. In the absence of

PPPs, the economy will have to wait for longer to get to the same production frontier and the

associated gains in terms of higher growth and employment.

The flip side is that strategic considerations on the part of the government can lead to excessive

reliance on the PPPs route since it enables the government to move infrastructure projects off

the budget sheet. This especially is the case with those projects that allow charging of user-fee

by the private partner. Since the private partners are willing to fund these projects in lieu of the

user-fee, the social costs of such projects do not show up in governments accounts. For other

projects, PPPs allow a government to stagger its obligations over the longer operational and

maintenance (O&M) phase of the project.

Therefore PPPs help ease the government’s fiscal burden. So, when government is hard pressed

for funds, there is likely to be a strong temptation to use the PPP route even though the overall

costs of doing so is greater than that of raising funds through taxation or collection of user-fee

by the government. This concern is particularly relevant today since the cost of private finance

has increased due to the recent credit crisis.

Nonetheless the officially stated objective for the use of PPPs is to harness efficiency and

expertise of the private sector. There is scope for an efficiency improving partnerships between

the public and private sectors. Next, we turn to this issue.

3. PPP versus Traditional Contracts

3.1 Salient Distinctions In this section, we examine the allocation of risk and control rights

under the PPPs and the traditional approach for procurement in infrastructure. Any

infrastructure service requires building of assets. For example, road services require building

of road lanes, railway services require laying down of tracks and building of stations, etc. The

PPPs and the traditional approach (non-PPP) differ in that they employ different types of

contracts for building of assets. We compare the incentive structure for the contractors under

the two approaches.

Under a PPP, the focus is on the output features of the assets or the facility to be built by the

contractor. For example for a road project the output features are described in terms of the

location and length of the road, the number of traffic lanes, under and overpasses, etc. The

9 The Fiscal Responsibility and Budget Management Act, 2003

5

decision about the technical and engineering specifications needed to build the assets are left

to the contractor. The idea is to leave some room for innovation and technology transfer by the

private partner. In contrast, under the traditional approach the government department decides

the output features and also the technical and engineering specifications for building the project

assets. In other words, under a PPP the contractor is told ‘what’ is to be done. Under the

traditional approach, he is told ‘what’ is to be done and also ‘how’ it is be done.

Moreover, PPP uses longer term contracts. For the ease of illustration, see Figure 1 that shows

the life cycle of an infrastructure project. A typical project has four phases: i) Project planning

by the government; ii) tendering and awarding of contract(s); iii) building/construction phase;

and iv) operation-and-maintenance phase.

t=4

Figure 1: Four phases of a project

While the first two stages essentially belong to the domain of the public sector, the contract is

used for the remaining two stages. Under traditional contract, the contractual relation starts at

date t=2 and ends at date t=3. The contractor is paid for the actual works done at contractually

agreed rates. Under a PPP contract, the contractual relation between the parties starts at date 2

and lasts till the end of the O&M phase, i.e., till t=4. Box 1 lists the main differences between

contracts used under PPP and those under the traditional approach.

3.2 Relative Advantages It is argued that a suitable allocation of the risk and other contractual

obligations can make the outcome under PPPs better than the traditional contracts. Here we

discuss the possible sources of economic advantages of PPPs over the non-PPP approach.

The first source can be allocation of risk between the public and the private sector partners. An

optimal allocation of the risk has potential to lower the overall costs for the society. Indeed,

there is a scope for a mutually gainful risk sharing between the public and private sector in the

construction and the O&M phases of a project. For example, the government is in a better

position to bear the risk associated with land acquisition, shifting of utilities, and regulatory

clearances. These risks can be assigned to the government. In contrast, the risks such as those

related to construction costs and time can be assigned to the contractor/concessionaire firm.

The private construction firms have greater flexibility in adjusting its resources (personnel,

Bidding and

Contracting;

Choice of

Contract –PPP

or not

Operation and

Maintenance-

User fee,

investment is

recouped

t = 0 t = 1 t = 2

Project Planning–

Initial estimate of

project cost and time,

and mode of funding

Building/Constructi

- Construction costs

are incurred

Phase I

t = 3

Phase II Phase III Phase IV

6

equipment and materials) to constantly changing circumstances so they can bear the

construction costs related risks more efficiently. For similar reasons during the O&M phase,

risks related to maintenance tasks can gainfully be assigned to the private partners.

Box 1: Difference between PPP and Traditional Procurement Attributes PPP Traditional Contracts

Funding of

building/construction

and maintenance

expense

Privately funded, though

government may contribute

Full funding by the government

Responsibility of

contractor(s)

Construction as well as

maintenance of assets. Both

tasks are performed by the same

contractor.

Only construction of assets.

Maintenance is done in house by the

government or by another

contractor/entity hired at a later stage

Payments to the

contractor

Contractor charges user-fee, as

under BOT toll contracts;

Or, annuity payments by the

government

Payment by the government - for

construction works

Risk-sharing

Most of construction and

maintenance cost risks are

transferred to the contractor.

Commercial risks can also be

passed on to the contractor, e.g.,

in toll contracts.

Only some of the construction risks are

transferred to the contractors.

Maintenance and commercial risks stay

with the government

Control and decision

rights – design,

operation and

maintenance

Most rights are delegated to the

contractor

Mostly stay with the government.

Also, it is argued that private firms can manage the demand and supply curves in a more

rigorous way than the public sector. They also have stronger motivation to earn good returns

on the investment. So, private investors can potentially use toll-fee more efficiently to recoup

their money. This means that when project can be financed through direct user-fee, there can

be a case for allocation of funding and commercial risks to the private partners. At the same

time, it must be noted that the private participants are risk averse. Moreover, as discussed

earlier, the cost of raising funds from the market is relatively high for the private sector. Due

to these two reasons, private firms are likely to demand high risk premium especially for

commercially risky projects. So, there are trade-offs. Depending on the riskiness of a project,

an optimal sharing of financial risks during the O&M phase may either require allocation of

most of the revenue risks to private sector or it may require the public sector to assume all of

these risks.

For example, road projects near big cities have high demand from the road users and therefore

carry lower commercial risks. It is for this reason that the private investors have been keen to

invest only in the projects near big cities.10 For such projects, funding and commercial risk can

be delegated to the private sector. In contrast, projects away from the urban centres and those

connecting the rural hinterland to nearest city have low demand and relatively high commercial

10 See Anant and Singh (2009)

7

risk. Private firms will invest in such projects only if they are paid very high premium, bringing

down the value for money for the public sector.

Therefore, for projects with medium commercial risk, the optimal risk sharing might entail

revenue sharing arrangements between the public and the private sectors. For projects with low

commercial viability, the public funding is a must.

The next source of efficiency is the potential relative efficiency of the private sector. Unlike

government departments, private firms have greater contractual options available to them. For

example, a private firm can promise a contractor to award future works, if he delivers good

quality of the existing works. However, a government department cannot make such promise

as the future contracts will also be awarded through competitive bidding. As a result, compared

to traditional government funded infrastructure projects, instances as well as magnitudes of

project delivery time and total project costs can be reduced with suitable private sector

participation.

The third source of efficiency is an appropriate bundling of responsibilities for the contractor

firm. There is a case for bundling of the responsibility to build/construct with the task of

maintenance of project assets during O&M phase. The quality of construction has a bearing on

how frequently the project assets would need repair during the maintenance phase. For

example, a poor quality road would need repair after every rainfall. The higher is the quality

of construction, the lower will be the required frequency of maintenance; the lower is the

construction quality, the higher will be the required frequency of repair.

Under traditional contract, the contractor is required only to construct the project facilities; he

is not responsible for maintaining it during the O&M phase which implies that the tasks of

construction and maintenance are not bundled. Since quality comes at the expense of higher

construction costs, a contractor who is not responsible for the maintenance of the project assets

will have no incentive to provide construction quality. His incentive is to build road of

minimum acceptable quality, regardless of the consequences of poor quality during the O&M

phase.

In contrast, under a PPP contract, the contractor firm is required not only to construct the project

facilities but also to maintain it during the O&M phase. In other words, the tasks of construction

and maintenance are bundled together and assigned to one and the same firm. Thus, a PPP

contract creates incentive for the partner firm to consider the life-cycle cost of the project. Also,

a good quality of assets implies lower maintenance expenses, and vice-versa. That is, a poor

construction quality comes at the expense of high maintenance costs. So, a PPP contractor does

not have strong incentive to dilute construction quality. In other words, a suitable bundling of

these tasks has the potential to encourage the contractor firm to maintain quality and yet

minimize the total project costs - the sum of construction plus the maintenance expenses.

The above discussion lead us to the following conclusion about PPP versus non-PPP projects:

Ceteris paribus, the life-cycle costs are expected to be relatively low the PPP projects, and the

quality of assets is also expected to be relatively high.

3.3 PPPs vs. Traditional Highways Contracts: Relative Efficiency

As mentioned earlier, our empirical analysis focuses on the national highways projects

implemented by the NHAI, the government body responsible for building and expanding the

network of (interstate) highways in the country. The NHAI has been using PPPs and traditional

8

contracts to build the highway projects as a part of the National Highways Development

Programme. As far as the PPP contracts used on the National Highways are concerned, there

are three types of PPPs – BOT Toll, BOT-Annuity and Special Purpose Vehicles (SPVs). The

PPP contracts differ in terms of the duration and the commercial risks borne by the contractor.

As the names indicate, while a BOT Toll contract entitles the contractor to charge toll from

road users, the BOT-Annuity and SPV contracts do not. Under BOT-Annuity contracts, the

contractor receives contractually agreed biannual payment from the government. Under the

SPVs, the ‘contractor’ also happens to be the main user of the service. For example, a port trust

can form SPV in partnership with the NHAI and the state government to improve its

connectivity by building a road. In that case, the port trust will invest in building and

maintenance of the project to get the right to use the road.

Box 2: PPP versus IR Contracts: Major Differences

Attributes PPP Contracts Item-rate Contracts

Funding Largely Privately funded Public Funding

Tasks assigned to

the Contractor

Construction and

Maintenance tasks. Under

BOT Toll, user-fee

collection is allowed.

Only the construction task is performed by the

contractor. The maintenance and user-fee

collection (if any) tasks are performed by some

other contractor/entity

Duration Long term; 12-18 years. Short term; lasts only for duration of the

construction phase only

Risk-allocated to

the contractor

Most of construction and

maintenance cost risks.

Under BOT Toll contracts,

commercial risks are also

allocated to the contractor.

Either some or all of the construction risks.

In-built incentive

for faster delivery

of project.

The revenue receipt for the

contractor starts only after

construction is over.

No such incentive

One of the important differences between the PPP and traditional contracts pertains to the

project delivery time. Under a PPP contract, the time period of construction (generally assumed

to be two-three years) is included in the concession period itself. An earlier completion of

project enables the concessionaire to increase the total toll revenue from the project. In case of

annuity contract, the concessionaire receives a bonus for an earlier completion. In contrast,

under an IR there is no such provision. Moreover, if there is any delay in the completion of

the project, contractor is penalized in the form of reduced annuity payments. Due to these

differences, one would expect that under PPP a contractor has stronger incentive to complete

the project sooner and avoid time overrun. Therefore, other factors held fixed, relative to the

IR based projects, the project delays are expected to be shorter for PPP projects.

Moreover, all types of partnerships possess the following common features: One, the tasks of

construction of project facility and its maintenance are performed by the same contractor (or

the same consortium of contractors); Two, most of the construction-related risks and all of the

maintenance risks are borne by the contractor; Three, the project designing, building, financing,

and its operation and maintenance are the responsibilities of the contractor. That is, PPP are

based on the Design, Build, Finance, Operate and Maintain (DBFO&M) contracts.

The Model Concession Agreement (MCA) for PPPs on National Highways contains the

standard contractual terms for these contracts. The non-PPP projects, on the other hand, are

implemented by using what are called the ‘item-rate’ (IR) contracts. These contracts are

popularly known as ‘cash contracts’. Under an IR, the contractor is responsible only for

9

construction of project assets or facilities; maintenance of the facility is not his responsibility.

Therefore, the contractual relation between the parties ends with the construction phase, i.e., at

t=3. Moreover, under IR contract, the contractor shares construction costs related risks with the

government, especially those arising due to variations in quantities of work-items. The Box 2

summarizes the major differences between the PPP contracts, on one hand, and IR contracts,

on the other hand.

To sum up, the partnership contracts used by the NHAI have all of the salient features of PPPs.

The IR contracts, on the other hand, are a type of the traditional contracts as described above.

Therefore, as discussed in the previous subsection, PPPs incentivize the contractors to

minimize the total project costs while maintaining quality of the construction. Due to lack of

bundling, such an incentive is missing under the traditional contracts. As a result, for the PPP

road projects, the life-cycle costs are expected to be relatively low, and the quality of roads is

expected to be relatively high.

There are no data available to compare the life-cycle costs of PPP versus non-PPP projects.

Similarly, a comparison of road quality is not feasible due to lack of data. However, we have

data on construction costs. Therefore, the questions is: What can we predict about the

comparative construction costs under the PPP and the same sector non-PPP projects?

As noted above, under traditional contracts, the contractor is responsible only for the

construction works. Predictably, he is interested in minimizing the construction costs within

the letter of the contract; the consequences for construction cost minimization for the O&M

phase is not his concern. So, a non-PPP contractor can afford to reduce construction cost by

reducing the quality of construction. In contrast, if a PPP contractor tries to minimize the

construction cost by diluting quality of construction, it will lead to disproportionate increase in

the O&M cost. Consequently, the total project costs will be excessive which is not in the

interest of the contactor. Therefore, we have the following testable conjecture about the

construction costs.

Hypothesis 1: Ceteris paribus, the construction costs are expected to be higher for PPP roads

than for non-PPP road projects.

Even though under both types of PPPs the contractors care about the quality of construction to

reduce maintenance costs. However, in case of toll PPPs the quality can further help them

attract more traffic to their project thereby raising their profits. This logic does not apply to

annuity projects. Moreover, the duration of toll contracts is longer than that of the annuity ones.

On this count also the toll contractors care more for the quality. If these arguments about

construction quality and costs are valid, we can expect that the construction costs for toll-PPPs

will be higher than for non-toll-PPPs. Specifically, we have the following conjecture.

Hypothesis 2: The difference between construction costs for toll-PPPs versus non-PPPs will

be even higher than the corresponding difference between the construction costs of the non-

toll-PPPs versus the non-PPP roads.

4. Bids and Contracts

Once the government decides to develop a highway project, the NHAI decides on the main

features of the project such as the number of traffic lanes, under and overpasses, etc. Thereafter

the authority commissions a project feasibility report. It hires a private firm for the purpose

through competitive bidding. The selected firm prepares the project feasibility report which

provides estimate of engineering works needed to be performed for the project concerned, the

10

estimates of construction time and costs. The report also provides estimates of traffic. The next

step is to invite bids.

Between 1997 and 2001 the government had offered only a few projects to investors on

experimental basis. However, since 2002 most projects in our dataset have been offered to

investors, except those for which the estimated traffic was rather low. If the feasibility report

indicated high or medium traffic, the project was first offered to investors as toll PPP. If there

were no takers for a project on toll basis, it was offered on annuity basis. If a project did not

succeed as PPP - neither as toll nor as annuity PPP - public funding was provided.

For PPPs, the bid documents provide all the information from the feasibility report along with

the contract period. The duration of contract period is 12 years for annuity based projects. For

toll contracts, it is generally between 12-18 years depending on the traffic. The contract period

is decided before inviting the bid. For toll PPPs, it is the expected number of years (in

government’s assessment) needed for the investors to recoup their investment. The contract

period is revised upward (downward) according to a common and publicly known formula; if

the actual traffic turns out be less than (greater than) the expected traffic according to feasibility

reports.

The bidders are required to submit one-dimensional bid. For toll PPPs, the bid is the amount

of money the contractor wants from the government to design, build and maintain road with

pre-specified output features. This amount is over and above the toll revenue expected by the

bidder during the operational phase of the project at hand. It is worth mentioning that the toll

rates and the revision rules are based on a uniform and publicly known criterion. These rules

are the same for the entire country, regardless of whether project is PPP or not - government

also charges toll for many projects. Besides, the contractor shares a fraction of toll revenue

with government, especially in the later years of the contract.

For lucrative (high traffic) projects, the bids can be negative. A negative bid means that in order

to get the right to charge toll, the bidder is willing to pay money to the government in addition

to taking responsibility for incurring cost of project design, construction and maintenance. For

less attractive projects, bidders don’t expect full recovery cost from toll. For such projects they

submit positive bids. That is, in addition to the right to charge toll, the bidders demand money

from the government. In policy documents, the positive bid for toll PPPs is known as Viability

Gap Funding (VGF). The upper limit for VGF is either 40 % of the expected project costs as

defined above or 40 % of the cost in the financial agreement with the lenders, whichever is

less. The VGF/bid-amount is submitted by the bidders. Lowest bidder wins the project.

Under annuity-PPP, the contractor is paid biannual annuity by the government - he cannot

charge toll. Therefore, the bid is the annuity that the bidder is willing to accept to carryout

project design, construction and maintenance tasks with pre-specified output features. Lowest

bid wins the contract. There is no provision of VGF in these PPPs.

When a project does not succeed as PPP and public funding is to be provided, the bids are

invited only for construction work. Bid documents provide the above information mentioned,

along with the engineering design prepared by the government. The bidders are required to

submit bids (per-unit asking price) for each of the work item mentioned in the design.

11

5. Data and Comparative Statistics

Our dataset consists of all the highways projects started and completed during December 1997

and August 2015 under the supervision of the NHAI. During this period, a total of 313 national

highway projects started and completed under the aegis of NHAI.11 The start date of these

projects falls in the period December 1997- August 2012. Out of the 313 fully completed

projects, the information needed for the empirical analysis was missing for 6 toll PPPs and 3

non-toll PPPs.

Figure 2a gives the year-wise number of PPP and non-PPP projects in our dataset. The first

project was completed in February 2000 and the last one was in August 2015. Figure 2b gives

the year-wise number of projects completed in this period for each category.

As is apparent from the graphs, the most of PPPs are recent. Most of the projects started and

completed during late 1990s and early 2000s were non-PPPs, in the later years the PPPs

account for a substantial proportion of the projects.

Figure 2a: Year-wise number of PPPs and non-PPPs projects started during 1997 - 2012

Source: Calculations based on the NHAI data

Figure 2b: Year-wise number of PPPs and non-PPPs projects completed during 2000- 2015

Source: Calculations based on the NHAI data

11 During this period a total of 376 national highway projects were completed. The NHAI does not provide data

for 60 odd projects most of which were implemented by the Ministry of Road Transport and Highways.

0 1 2 13

14

2 2 2

21

1613

2

10

30

1 3

24

16

61

9

23

55

33

85

0 0 01

1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012

PPP Non PPP

1 12 2

4

8

2 2

4

25

11 1110

6

87

1

10

1516

18

26

10

12

21

1716

23

10

8

4

2

2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015

PPP Non PPP

12

Figure 3: Delays – PPPs versus non-PPPs12

As predicted, the project delivery time is

faster for PPPs. Figure 3 shows the

percentage delays (the actual time as a

proportion of the estimated construction

time) for PPPs versus non-PPPs. The

average delay is 43 percent, i.e., for the

entire data the average actual construction

time is 43 percent longer than the estimated

time. However, PPPs projects have taken

only 12 percent extra time. In contrast, non-

PPPs have taken 57 percent extra time in

completion.

As far as the project costs are concerned, as discussed above, we do not have data on project

life-cycle costs. Therefore, we turn to the other conjectures about the project construction

costs. We do have data on construction costs. However, now the question is: How do we go

about comparing the construction cost for PPPs vis-à-vis the non-PPP projects?

One approach to compare the construction costs of PPPs relative to non-PPP projects is to

compare the per-kilometre per-lane costs of the two project types. However, the projects in our

dataset are located across the length and breadth of the country. One consequence of this is that

the per-kilometre costs vary across regions, due to variations in surface and underground

conditions, and also due to the choice of the material used in road construction. Moreover, road

projects differ from each other in terms of the number and design of over and underpasses,

service-lanes, and their proximity to urban centres, all of which have a significant bearing on

the per-km construction costs. Consequently, even within a region, the per-km construction

costs can vary greatly from project to project. Therefore, a comparison in terms of per-

kilometre per-lane costs does not seem to be plausible.

The alternative is to compare the following ratio for PPPs relative to non-PPP projects:

aCCO /EC ,

where aC is the actual construction cost and

EC is the expected construction cost of the

project. EC is assessed by private firms hired by NHAI through competitive bidding. The hired

firm prepares a project feasibility report, lists major works and assess the expected project costs

and the expected construction time. With these estimates, the NHAI takes the project to

investors. If the project attracts private investment, it becomes PPP. The point is that EC is

assessed ex-ante to the choice of contract, i.e., before the PPP-non-PPP status of the project is

determined. Therefore, it does not depend on the type of the contract used for the project. On

the other hand, aC , the actual construction costs become known when the construction gets

completed, i.e., ex-post to the choice of contract.

However, both EC and

aC are about one and the same project. Plausibly, EC factors-in

considerations like surface and underground conditions, the material used in road construction,

the number over and underpasses, service-lanes, etc. This is indeed the case with the aC .

Therefore, it seems plausible to compare the CO for PPPs with non-PPP projects.

12 Source: Calculations based on the NHAI data

1.122

1.5701.431

PPP Non PPP Average (forall projects)

13

For similar reasons, for a plausible comparison of the project delivery time for the two types

of contracts, we use the following ratio: TO= aT / ET , where aT is the actual time (in months)

taken by construction and ET is the initially expected construction time.

Indeed, the literature on cost overruns uses the ratio CO to measure and analyse the cost

overruns. A project is said to suffer from positive [negative] cost overruns if and only if the

CO>1 [CO<1].13 Similarly the ratio TO is used to define the time overrun. For a project with

positive delay (time overrun), we will have TO>1; the time overrun is negative if TO<1.

Coming back to the cost comparisons, in view of the Hypothesis 1, ceteris paribus, CO is

expected to be higher for PPP projects. However, for the purpose of data analyses an

explanation is in order. The official data on PPPs does not provides figures for the EC . It

provides figures on a head called the ‘total project cost’ (TPC). For PPP projects, the TPC is

defined as the sum of the estimated construction costs plus 25 % of these costs. That is,

TPC = EC (1+0.25)

The second term (25 % of EC ) is added on account of the cost of interest during construction

(IDC) incurred by the PPP contractor. From the above definition, we get: EC =0.80 TPC. That

is, to retrieve theEC , we need to discount the official figures on TPC by 20 %. In Table 1, 20

% discounting is precisely on that count. Figures with 10% and 15% discounting correspond

to low discounting and therefore are an under-estimate of cost ratio for PPPs. In interest of

completeness, we have also provided figures for CO for PPPs without any discounting at all.

Figure 4: Cost overrun – PPPs versus non-PPPs

Source: Calculations based on the NHAI data

For non-PPPs, TPC=EC . For non-PPPs the government provides full funding therefore the

issue of IDC does not arise. Figure 4 shows the cost ratio for different subsets of projects. Table

1 provides more details.

13 See Singh 2011 for literature review.

1.3091.455

1.5401.636

1.126 1.126 1.126 1.1261.187 1.235 1.263 1.295

NO DISC 10% DISC 15% DISC 20% DISC

PPP Non PPP Average (including all projects)

14

As is clear from the figure and the table, the ratio of the actual construction costs to the

estimated construction costs is much higher for PPP roads. It is only for 1.13 for government

roads but is as much as 1.54 for PPP roads. From another perspective, we see that the PPP

projects have experienced cost overruns much higher than for the non-PPPs, even though both

types of projects have experienced positive cost overruns. Within PPPs, the construction cost

ratio is much higher for the BOT toll roads than the non-Toll PPPs. There are no significant

differences between the two types of PPPs as far as the delays are concerned.

Table 1: Summary Statistics – PPPs Vs non-PPPs

PPP Non-PPP Average (for all projects)

Number 104 209 313

No. of projects with positive TO 58 167 225

Average TO (for all projects) 1.12 1.57 1.43

No. of projects with positive CO 80 129 209

Average CO (for all projects) 1.31 1.13 1.19

No. of projects with positive CO at 10% disc 86 129 215

Average CO at 10% disc (for all projects) 1.45 1.13 1.23

No. of projects with positive CO at 15% disc 91 129 220

Average CO at 15% disc (for all projects) 1.54 1.13 1.26

No. of projects with positive CO at 20% disc 94 129 223

Average CO at 20% disc (for all projects) 1.64 1.13 1.29

Average Expected Cost 411.16 203.00 272.16

Average Actual Cost 555.57 225.71 335.31

Average Project Length 58.08 37.15 44.10

Source: Calculations based on the NHAI data

Table 3: PPPs – BOT Toll Vs. Non-Toll PPPs BOT Toll-PPPs Non-Toll PPPs

Number of projects 66 38

No. of projects with positive CO at 10% disc 58 28

Average CO at 10% disc 1.59 1.22

No. of projects with positive CO at 15% disc 59 31

Average CO at 15% disc 1.68 1.31

No. of projects with positive CO at 20% disc 60 34

Average CO at 20% disc 1.79 1.37

Source: Calculations based on the NHAI data

15

6. Higher Costs for PPPs: Possible Reasons

Due to several reasons, the ex-post actual costs are generally different from their initial

estimates. For instance, due to imperfect estimation techniques, the expected and the actual

project costs generally turn out to be different. Moreover, extensive literature on delays and

cost overruns shows that imperfect estimation techniques along with incompleteness of the

project design leads to time and cost overruns.14 However, these reasons are equally applicable

to traditional as well as PPP projects and therefore cannot account for the above observed

differences in the cost ratios. Therefore, the question remains: Why are the construction cost

ratios significantly higher for PPPs?

As explained in Section 3.2 the PPPs and non-PPP roads have been built using very different

types of contracts. Due to bundling, a PPP contractor has incentive to maintain quality of

construction to save on O/M costs. That is, the other factors held fixed, we expect the cost ratio

to be higher for PPPs than for non-PPPs. However, there are other factors that can lead to a

relatively high cost ratio for PPP roads. We discuss them below.

6.1 Purposeful underestimation of EC for PPP projects: In order to make the project seem

attractive to investors, the NHAI might deliberately understate the expected cost of PPP

projects. If so, the understated expected cost will show up has a higher cost ratio and cost

overruns for the PPPs. However, this does not seem to be the case with projects under study.

First of all, the project time and costs are not estimated by the NHAI. For all of the projects,

PPPs or not, these estimates are made by private firms hired for the purpose through

competitive bidding. Presumably, these firms have no incentive to systematically understate

the costs for a subset of projects. Moreover, the detailed project report that has details of the

factors used by a firm to arrive at estimates of construction time and cost are shared with the

bidders. This enables the bidders to check if there is an attempt to mislead them by understating

the expected project costs. Besides, investors also arrive at their own estimates of time and

costs. Therefore, it does not seem plausible that the government can make a project seem

attractive by understating the costs. Even otherwise, the estimates of time and costs are arrived

at before the PPP-non-PPP status of a projects gets known. Therefore, it is not obvious how the

costs can be understated only for PPPs.

6.2 Ex-post changes in project works: If ex-post to the signing of the PPP contract the

government demands more additions of works for PPP projects, it will lead to a higher

construction cost than the actual and therefore higher cost ratio for PPPs. However, a PPP

contract explicitly requires that the total cost of ex-post changes cannot exceed 15% of the

initially estimated costs (EC ). In contrast, non-PPP contract does not have such an upper limit.

In fact, the contract explicitly says that the contractor will have to perform all the work

demanded by the government. Moreover, since the government has to compensate the

contractor for such demands, therefore, it cannot gain by asking for additional works from PPP

contractors. In other words, ex-post works are more of a concern for non-PPPs than PPPs.

6.3. Purposeful escalation of aC by PPP Contractors: It is tempting to think that the PPP

contractors may gain by inflating the construction costs. This will indeed be the case, if the

contractor can use cost escalations to reduce tax liability or renegotiate the contract to extract

more concessions from the government. The PPP contactors in India enjoy ‘tax holiday’ for as

14 For details see Singh (2010a, and 2011). Also see Arvan and Leite. (1990), Flyvbjerg, Holm and Buhl (2003,

and 2004), Morris (1990 and 2003), among others.

16

much as ten years of their choice. So, inflating construction cost will not help them save tax

payments. Besides, as per the MCA15, construction costs cannot be a basis for any contract

renegotiation. As a matter of fact also, no PPP contract has been renegotiated so far for any of

the completed projects in our dataset. Moreover, as per a standard contract term, the maximum

permissible limit for debt financing is 70 percent of project costs – the investor contractor has

to make a direct equity contribution to the tune of at least 30 percent of project costs. In view

of these terms and an oversight by the lender banks, it seems unlikely that the PPP contractors

can pad up the aC . It is pertinent to add that for PPPs the cost estimates in the financial

documents prepared by the PPP contractors and the lenders are significantly higher than the EC since the forward-looking contractors know that their costs will be higher than the

estimated ones and hence, factor it in while raising the funds from banks.16

However, one issue needs further investigating. As explained above for less lucrative toll

projects, the bid can be positive. In that case, the government provides VGF. The VGF cannot

be increased after signing the contractor. However, it can be reduced proportionately if the

project costs mentioned in the financial agreement or the actual costs turn out to be less than

the TPC. This means that in principle there may be temptation to keep the costs in the financial

agreement and the actual expenditure higher than the expected costs. Since there is no VGF for

annuity PPPs, so this problem does not arise. Moreover, this problem does not arise for projects

with negative winning bids.

6.4 Time-cost trade off: As explained above, the PPP contractors have incentives to shorten the

project delivery time. The revenue stream for the PPP contractors starts only when they charge

user-fee or receive the annuity payments from the government. However, the user-fee can be

levied only during the O&M phase, i.e., after construction is complete. Similar is the case with

the annuity payments. SPV contractors also want to expedite the work in order to use road

services. In such a scenario, the PPP contractor may be willing to incur extra cost to complete

construction sooner than they would do otherwise. This additional cost if incurred can increase

the total construction cost, leading to higher cost ratio for the PPPs.

6.5 Trends over time: There is time difference between award of PPP and non-PPP contracts.

The formers are more recent. Therefore, in principle, the cost ratio might behave differently

over time, say on account of developments in cost and time estimation technologies. If so, the

PPP and non-PPP projects might exhibit different cost ratios since most of the PPPs are

relatively new. See Figure 1a and 1b above.

7. Empirical Analysis and Results

Table 1 shows that as expected the PPP projects have experienced much shorter delays.

Compared to the non-PPPs, the time over ratio is as much as 45% less for PPP roads. Statistical

tests show that the difference in the delays between PPPs and the government roads is

significant at all levels of significance.

For our empirical analysis of the relative construction costs, we need to address the issues

discussed in 6.3-6.5 in the previous section. To factor in the time-cost trade-off, we have used

the time overrun ratio (TO) as a variable to account for the cost differences due to differences

15 Model Concession Agreement is the standard contract document for PPPs. 16 The CAG report 36 of 2014 shows that the project costs worked out by the concessionaires is generally higher

as compared to the EC based TPC worked out by the Authority.

17

in project delivery times. We have attempted to capture the effect of trends in the cost ratios by

introducing a variable ‘Timelapse’. It is defined as the time gap (in months) between the start

date of the project and the start date for the first project in our dataset (December 1997).

Moreover, we have run logit regressions to identify variables that affect the likelihood of a

project becoming PPP and therefore its cost ratio. This exercise shows that the factors like the

project size, density of population near the project site have bearing on probability of a project

becoming a PPP. Big projects are more likely to attract private investment and therefore

become a PPP (See Anant and Singh (2009)). Similarly, the larger the size of the population

around a project more likely it is to become a PPP. The projects offered to investors more

recently have a greater likelihood of becoming PPPs. See Table A1 in the Appendix. The details

about the variables, data sources and models used are provided in the Appendix.

The problem discussed in 6.3 can arise only for the toll projects with positive bids and the VGF.

In our dataset there are 35 such projects. We examine whether the bid amount has a bearing for

toll as well as all PPPs. To address this issue, we have used the following ratio as an explanatory

variable:

BP=Bid price/ TPC

Figure 5: Cost overrun versus bidding price

The Figure 5 plots bid price versus the

cost overruns, i.e., cost ratio CO as

defined above for all of the 66 toll PPPs.

There does not seem to be any relation

between the two. Moreover, our

empirical analysis shows that the bid

price does not have any statistically

significant bearing on the cost ratio.

Table A3.

In the main text, we report results for the

15% discounting of the costs for PPPs.

As explained in the previous section, it

corresponds to a low discounting and

therefore the reported results are an under-estimate of cost ratio for PPPs. Detailed results at

various levels of discounting are provided in the Appendix.

Our econometric analysis provides empirical support to the above hypotheses 1 and 2.

Specifically, controlling for the effect of several relevant factors and project characteristics, we

get the following result: Discounting for the interest costs, the construction cost ratio for PPPs

is 34% higher than the ratio for non-PPPs. The difference is large and statistically significant.

The cost ratios for PPP are significantly higher even if we adopt a rather conservative approach

on the interest costs. More recent projects and those with large local population also exhibit

high cost ratio. See Tables A4.

To further corroborate these claims, we exploit the key differences between the two types of

PPPs - the toll PPPs and the annuity PPPs. As explained above, between the toll and non-toll-

PPPs, the quality concerns and therefore the construction costs are expected to be higher for

toll-PPPs than for non-toll-PPPs. See Hypothesis 2. From Table A5 it can be seen that this

indeed is the case. For the non-toll-PPPs the cost ratio is 15% higher than for the non-PPPs.

0

0 . 5

1

1 . 5

2

2 . 5

3

- 1 . 2 - 0 . 7 - 0 . 2 0 . 3 0 . 8 1 .

C o s t

o v e r r

u n

B i d d i n g P r i c e

18

However, for the toll-PPPs the cost ratio is 52 % higher than the non-PPPs! Together these

results indicate towards a higher quality investment in PPP projects.

It is important to point out that our estimates may be biased on account of potential problem of

endogeneity. Since the contract assignment to projects - PPP versus non-PPP - is not random,

therefore the issue of potential endogeneity is relevant. It is possible that there are project

characteristics that are unobservable to us but are observable to the investors leading to

selection bias. The characteristics that make a project attractive to the investors may also be

correlated with greater susceptibility to cost escalations; for example, heavy traffic on projects

in or near big cities make them more lucrative for investors but can also be a cause of the cost

escalations. In such cases, the estimates of the differences in the construction cost of PPPs

versus the rest will be biased. However, for the approach adopted here to examine differences

in costs, the scope of the problem of endogeneity seems limited. Recall, our focus is on the cost

ratio rather than the absolute costs. Plausibly and hopefully the specialist firms hired to estimate

the project costs would have factored in the factors having bearing on the construction costs.

However, we cannot rule out the theoretical possibility of endogeneity.

8. Conclusions and Policy Recommendations

Our study shows that PPP roads have performed better than the government roads (non-PPPs)

in terms of project delivery time. Under a PPP, the contractor has strong interest in completing

the project as soon as possible. Consequently, PPP roads have experienced significantly shorter

delays than the government roads. In view of the fact that delays are one of leading causes

behind cost overruns, the cost overruns attributable to the delays are lower for PPPs than for

the non-PPPs roads. Besides, the economy gets to benefit from the road services sooner than

would be the case under non-PPP approach.

However, the cost overruns are significantly higher for PPP road projects than for the

government road projects. Compared to the government projects, for PPPs have shown 34-43

% higher cost overruns. Within PPPs, cost overruns for the BOT Toll roads are even higher –

52 % greater than the government roads.

The empirical analysis suggests that project life-cycle approach induced by the PPPs is a

leading factor behind the relatively high construction costs for PPPs. The PPP contractors want

to maintain the quality of construction during the road building phase, so as to reduce their

maintenance costs during the operational phase of the project. This enables them to minimize

the life-cycle costs. In contrast, under the contracts used for government projects, the

contractors are not responsible for the maintenance of roads build by them. So they have strong

incentive to minimize the construction costs at the expense of the quality. The result is low

construction costs for government roads17 but poor quality roads with high maintenance and

life-cycle costs. In other words, the PPP contracts induce quality of construction higher than

the level induced under traditional contracts. Alternatively put, there is a desirable absence of

quality-shading/cost-cutting incentives under the PPP contracts. The relatively high

construction costs for PPP roads seems to be due to the differences in quality of construction.

This argument is also strengthened by our finding that the construction costs are highest for the

toll based PPPs. As explained above, for the BOT toll contractors the quality investment is

beneficial on several counts.

17The construction cost cutting efforts by the IR contractors make them to bid aggressively, leading to lower

contracted costs for the government department, that is, the NHAI in this case.

19

These inferences are further corroborated by the CAG reports on the National Highways

Development Programme. In these reports most of the government projects (non-PPPs) were

found to be deficient in quality. The quality of most of the PPP projects has been found

satisfactory.18 Besides, we have searched for and examined the media reports on the issue of

quality of national highway projects. Most of the standard problems namely, potholes, poor

quality of road surface, deficient design, etc., are reported for the non-PPP projects.19

It is important to note that the project costs as defined above do not cover several aspects such

as the social costs. Therefore, our findings do not make an unambiguous case for PPPs. The

choice between the conventional approach and the PPPs involve several trade-offs. The overall

superiority of PPPs over the conventional approach towards infrastructure depend on all of the

costs as well as benefits for the road users and the society in general. For instance, the cost of

borrowing is higher for the private sector, than is the case with the public sector. So, the direct

cost of raising funds under PPPs is relatively high. This can be a concern if the upward pressure

on interest rates continues. Similarly, the benefit in terms of reduced delays under PPPs can be

misleading. Even though the construction delays are shorter under PPPs, these projects seem

to take much longer during the pre-construction phases. Awarding of PPP contracts is a

protracted process. There is suggestive evidence that the PPPs are shifting the delays from the

construction to the procurement phase.

Moreover, the above mentioned media reports reveal several problems with the way PPPs are

being executed. Among the frequently reported problems are the defective designs and poor

maintenance of the service lanes meant for the local traffic in urban areas. Similarly, at most

of the toll plazas, the waiting time is unreasonably long. This is also a result of the cost cutting

and profit maximising behaviour on the part of the private partners. The toll plazas are located

strategically to capture maximum traffic but the number of booths is kept low. While such

decisions are good for the contractor firm, it imposes huge cost in terms of time and fuel wasted

by the waiting road users at these plazas. Another set of problems include inadequate capacity

at entry and exit points, inadequate provisioning of pedestrian on bridges and under-passes,

poor lighting facilities and other safety measures. This shows that the PPPs are putting

excessive emphasis on cost cutting measures at the expense of the welfare of the road users.

To conclude the analysis, we turn to the question: How can PPPs be made to deliver a better

outcome? First of all, there seems to be a case for reducing the cost of direct funding for PPPs

while maintaining the bundling, a key feature of PPPs. In principle, the construction and

maintenance tasks can be bundled together to be performed by the same contractors, yet the

contract may not require the contractor to provide full funding for the project - the government

can provide major funding with substantial equity stakes from the private partners.

Besides, the preconstruction delays for PPPs need to be reduced. For the success of the

ambitious infrastructure development programme it is crucial to make a time bound decision

on funding and also to obtain regulatory clearances from the ministries involved.

Finally, there is need to restrict the scope for the private partners to maximise their profit at

expense of the social welfare. PPPs delegate most of decision rights related to project features,

18See, for example, CAG Report No. 7 of 2005 (PSUs), CAG Report No. PA 16 of 2008, and CAG report 36 of

2014. Also see Lok Sabha (2006). The CAG report 36 of 2014 does point out to poor road conditions in case of

6 PPPs. Unlike the PPPs in our data set for these PPPs the contractors were allowed to charge toll even prior to

completing the project. This and the related terms of contracts diluted the key features of PPPs. Also see Singh

(2010b) 19 For this exercise, we did google search using key words ‘national highway, potholes, poor quality road, road

users, toll plaza, commuters, road users, BOT, PPP’. We found 60 odd print media reports in English. The

summary of these reports is available at http://econdse.org/ram-research/

20

design, and its operation to the private partners who are misusing them to make money at the

expense of safety and welfare of the road users. Rather than delegating all the decision rights,

the government agency concerned should remain closely involved in all key decisions related

to the project design as well as operation and maintenance of the project.

APPENDIX

Variables: Description and Data Sources

For each project in our data set, we have compiled to derive project specific value for the

following variables and controls:

1. Expected Costs EC : It is the construction costs as assessed before inviting the bids. It

is derived from the NHAI data on the total project cost (TPC), according to the official

rules described in the Section 5.

2. Actual Construction Costs, aC : These are the total actual construction costs of the

project when the construction gets over. In the NHAI files, the construction costs are

recorded as ‘cumulative expenditure’. Note for PPPs ‘cumulative expenditure does not

include IDC. For PPPs, as a condition in the financial package, no re-payment of

interest or principal can start until the second year after completion of the construction

phase. Therefore, the contractors do not actually pay/incur IDC during the construction

phase.

3. Cost overrun or Construction Cost Ratio (CO): It is the ratio of the actual construction

costs and the expected construction costs. That is, aCCO / EC , where aC and EC

are as defined above.

4. Expected construction timeET or the Implementation Phase (impphase): The duration

(in months) in which a project is planned to be completed, i.e., ET . Specifically, it is

the time difference between the date of signing of the contract, and the expected

construction completion date at the time of signing of contract.

5. Timeoverrun Ratio, TO: It is defined as aT /

ET , where aT is the actual time (in

months) taken by construction, and ET is as defined above. Data source is the NHAI

files.

6. Length (lenghtkm): It is the length of the project (in kms).

7. Timelapse: It is the time difference (in months) between the date of contract signing for

the project at hand, and the date of start of the contract for the first project in the dataset

(December 1997).

8. Local Population (LPopulation): It is the population of Class 1 towns (according to

2011 census classification) around the project. For each project, we have identified all

class 1 town/cities located within 10 kms of the project. The idea is to see if extent of

urbanization has any bearing of probability of a project becoming PPP and also on

construction costs ratio.

21

9. Distance: It is the distance of the project from the nearest megacity (million plus city).

For each project, we have identified the nearest megacity and calculated its distance

from the project. Various studies show that megacities are hubs of economic and

commercial activities, and therefore act as growth center for nearby areas. The idea is

to see if proximity to a mega city has any bearing on probability of a project becoming

PPP and also on the construction costs.

10. Bid Price: We have used the following three definitions of the bid price (BP): BP1=

Awarded-Cost / TPC (for all PPPs); BP2 = Awarded-Cost / TPC (only for Toll PPPs),

and BP3 = bp2 (only for Toll PPPs with positive bids i.e. bp2>0).

Data source for variables 1, 2, 4, 6 and 10 is NHAI. These variables are used to compute values

for variables 3, 5, 7 and 9. Variable 8 is computed using census 2011 data.

Table A1: Logit Regression Results

For the results reported in

Table A1, the chi-sq

statistic is 100.57. So, the

probability that there is no

effect of the independent

variables on the choice of

projects captured by the

binary variable “PPP” is

zero. Since the p-value

obtained here is less than

the critical value of 0.05 or

0.01, the model is

statistically significant.

Odds ratio is expressed as

Prob(PPP=1)/

Prob(PPP=0).

As is clear from the table, as length increases, the odd ratio for a project (becoming PPP)

increases. As the distance from the nearest megacity increases, the odd ratio decreases.

Multinomial logit regression results are similar except that the effect of project length on

relative log ratio are seemingly more pronounced for toll-PPPs, and the effect of distance from

the nearest mega city has more pronounced bearing on the relative log ratio for non-toll PPPs

than for the toll PPPs.

(1) (2)

PPP (odds ratio)

Length in km 0.0254*** 1.026081

(4.27) -

Timelapse since Dec 1997 0.0318*** 1.033996

(5.81) -

Distance from nearest mega city -0.00144 0.9986427

(-0.64) -

Local population around project 0.000000233* 1

(2.21) -

Constant -4.465*** 0.0102209

(-7.67) -

Observations 304 -

Pseudo R-sq 0.2289 -

Log likelihood -148.71206 -

t statistics in parentheses. * p < 0.05, ** p < 0.01, *** p < 0.001

22

Table A2: Multinomial Logit Table

(1) (2)

Dummy Variable (Relative risk ratio)

0 (Non-PPPs) (Base Outcome)

1 (Toll PPPs)

Length in km 0.0346*** 1.035168

(4.76) -

Timelapse since Dec 1997 0.0429*** 1.043833

(6.05) -

Distance from nearest mega city -0.00235 .9976527

(-0.84) -

Local population around project 0.000000342** 1

(2.83) -

Constant -6.454*** .0015748

(-8.17) -

2 (Non-Toll PPPs)

Length in km 0.0154* 1.015507

(2.03) -

Timelapse since Dec 1997 0.0233*** 1.023583

(3.54) -

Distance from nearest mega city -0.000432 .9995679

(-0.15) -

Local population around project 0.000000121 1

(0.83) -

Constant -4.058*** .0172792

(-6.05) -

Observations 304 -

Pseudo R-sq 0.2226 -

Log likelihood -207.77064 -

t statistics in parentheses * p < 0.05, ** p < 0.01, *** p < 0.001

To examine the effect of bid price on the construction costs, we have used the following

specification for the cost ratio, CO:

Ca

Ce=a0 +a1BP+a2Length+a3Timelapse+a4LPopulatoin+a5Dis tance +a6TO+e

Since the bid prices for PPPs are not comparable with the non-PPPs, so the analysis is restricted

only to the set of PPPs. As the following Table A3 shows the bid price does not have significant

effect on the cost ratio. These OLS regression results are based on 15 % discounting, as

explained above. The results are very similar for the other levels of discounting as well.

For comparisons of construction costs for PPPs relative to non-PPPs, we estimate several

versions of the following model:

Ca

Ce=a0 +a1DPPP+a2Length+a3Timelapse+a4LPopulatoin+a5Dis tance+a6TO+e

23

Table A3: CO and Role of Bidding Price for PPPs

(1) (2) (3)

discCO15 discCO15 discCO15

Length in km 0.00218 -0.00177 -0.00191

(1.29) (-0.92) (-0.82)

Timelapse 0.00753*** 0.0135*** 0.0152***

(3.79) (5.09) (4.72)

TO 0.0801 0.380 0.355

(1.68) (1.90) (1.49)

Local population 5.11e-08 2.67e-08 3.18e-08

(1.94) (0.87) (0.75)

Distance -0.00103 -0.000752 -0.00249*

(-1.36) (-0.73) (-2.08)

bp1 -0.0906 - -

(-0.54) - -

bp2 - 0.108 -

- (0.53) -

bp3 - - 0.0708

- - (0.14)

Constant 0.720** 0.207 0.173

(3.36) (0.63) (0.39)

Observation 91 56 35

R-sq 0.3021 0.4264 0.5573

Adj R-sq 0.2522 0.3562 0.4625 t statistics in parentheses. * p < 0.05, ** p < 0.01, *** p < 0.001

Table A4: Higher Costs for PPPs Vs. non-PPPs

(1) (2) (3) (4)

CO discCO10 discCO15 discCO20

PPP 0.125* 0.263*** 0.344*** 0.435***

(2.28) (4.54) (5.76) (7.05)

Length in km 0.0000573 0.000185 0.000260 0.000345

(0.07) (0.22) (0.30) (0.38)

Timelapse 0.00314*** 0.00334*** 0.00346*** 0.00359***

(4.24) (4.30) (4.32) (4.34)

TO 0.147* 0.158* 0.181* 0.202*

(1.98) (1.96) (1.99) (2.03)

Local population 3.13e-08* 3.43e-08* 3.61e-08* 3.81e-08*

(2.33) (2.43) (2.48) (2.53)

Distance 0.000181 0.000161 0.000149 0.000135

(0.55) (0.47) (0.42) (0.37)

Constant 0.820*** 0.804*** 0.794*** 0.783***

(10.14) (9.46) (9.07) (8.65)

Observations 304 304 304 304

R-sq 0.1502 0.2339 0.2855 0.3409

Adj R-sq 0.1330 0.2814 0.2710 0.3276 t statistics in parentheses. * p < 0.05, ** p < 0.01, *** p < 0.001

24

The OLS regression results for construction costs ratios are presented in the tables A4 and A5.

As can be seen from Tables A4 controlling for the factors, the average cost ratio for PPPs as

group is 30-40 % greater than the non-PPPs. Next, we examine the difference effect of toll-

PPPs and non-toll-PPP relatives to the non-PPPs. For this purpose, we define two dummy

variables;

DV1 = {10

𝑓𝑜𝑟 𝑇𝑜𝑙𝑙 𝑃𝑃𝑃𝑠

𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒 and DV2 = {

10

𝑓𝑜𝑟 𝑁𝑜𝑛 − 𝑡𝑜𝑙𝑙 − 𝑃𝑃𝑃𝑠

𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒

Table A5 shows the difference between the average cost ratio for toll-PPPs versus the non-

PPPs (DV1) and the cost ratio for the non-toll-PPPs versus non-PPPs. As expected the

difference in cost ratio is higher for the toll-PPPs than for the non-toll-PPPs vis-à-vis the non-

PPPs. Moreover, t-test based analysis shows that the differential impact of the toll-PPPs

(relative to non-PPPs) versus the non-toll-PPPs (relative to non-PPPs) is statistically significant

at all levels of significance.

Table A5: Higher Costs for toll PPPs Vs. non-toll-PPPs

(1) (2) (3) (4)

CO discCO10 discCO15 discCO20

DV1 0.278*** 0.432*** 0.523*** 0.624***

(4.26) (6.31) (7.43) (8.60)

DV2 -0.0375 0.0826* 0.153** 0.233**

(-0.56) (2.18) (3.13) (4.14)

Length in km 0.000353 0.000269 0.000220 0.000164

(0.45) (0.33) (0.26) (0.19)

Time lapse 0.00267*** 0.00282*** 0.00291*** 0.00302***

(3.66) (3.69) (3.71) (3.72)

TO 0.0759* 0.0774* 0.0783* 0.0793*

(2.01) (1.99) (1.98) (1.99)

Local population 2.48e-08 2.71e-08 2.85e-08* 3.00e-08*

(1.88) (1.96) (2.00) (2.05)

Distance 0.000240 0.000225 0.000217 0.000208

(0.75) (0.67) (0.63) (0.58)

Constant 0.829*** 0.814*** 0.805*** 0.794***

(10.52) (9.85) (9.47) (9.06)

Observation 304 304 304 304

R-sq 0.1961 0.2799 0.3306 0.3847

Adj R-sq 0.1771 0.2628 0.3148 0.3702

25

REFERENCES

Anant, TCA and Singh, Ram (2009), “Public Private Partnership Contracts in Roads: Key

Legal and Economic Determinants,” CDDRL, Stanford University, Working Paper

Number 100.

Arvan, L. and Leite. (1990), 'Cost Overruns in Long-Term Project,' International Journal of

Industrial Organization, 8(3), 443-67.

Flyvbjerg, B., Holm , M.K.S. and Buhl, S.L (2003), “How common and how large are cost

overruns in Transport infrastructure projects?,” Transport Reviews, 23, 71-88.

Flyvbjerg, B., Holm , M.K.S. and Buhl, S.L (2004), “What Causes Cost Overrun in Transport

Infrastructure Projects?” Transport Reviews, 24 (1), 3-18

GoI (2006 a), “Financing of the National Highway Development Programme”, Report of the

Core Group, Planning Commission, Government of India.

GoI (2006 b), “Guidelines on Financing Support to Public Private Partnerships in

Infrastructure, Planning Commission.

GoI (2007), “The report of The Committee on Infrastructure Financing” (Deepak Parekh

Committee) available at www.pppinindia.com

Grimsey, Darrin, and Mervyn K. Lewis. 2004. ‘Public–Private Partnerships: The Worldwide

Revolution in Infrastructure Provision and Project Finance.’ Cheltenham, UK: Edward

Elgar.

MCA (2006) “Model Concession Agreement (MCA) for Public Private Partnerships in

National Highways,” , published by Planning Commission, New Delhi.

MoF (2011) “National Public Private Partnership Policy”, Department of Economic Affairs,

Ministry of Finance, Government of India

Morris, Sebastian (1990), “Cost and Time Overruns in Public Sector Projects”, Economic and

Political Weekly, Vol. XXV, No.47, pp.M-154-M-168.

Morris, Sebastian (2003), “Public Investment in India: Delays and Cost Overruns”, India

Infrastructure Report 2003, Chapter 5.1.

Netter, Jeffrey, and William Megginson. 2001. ‘From State to Market: A Survey of Empirical

Study on Privatization’. Journal of Economic Literature 39(2): 321–89.

Singh, Ram (2012). Inefficiency and abuse of compulsory land acquisition: an enquiry into the

way forward. Economic and Political Weekly, 47(19):46-53.

Singh, Ram. (2011). Determinants of Cost Overruns in Public Procurement of Infrastructure:

Roads and Railways” India Policy Forum, Vol 7, pp 97-158.

Singh, Ram. (2010a). “Delays and Cost Overruns in Infrastructure Projects: Extents, Causes

and Remedies” Economic and Political Weekly, Vol XLV No 21, pages 43-54

Singh, Ram. (2010b). “A High-handed Approach to National Highways” in Economic and

Political Weekly February 2010, Vol XLV (No 8), pp 19-21

Yescombe, E. R. 2007. Public–Private Partnerships: Principles of Policy and Finance.

Burlington, MA: Butterworth-Heinemann/Elsevier.

Data/Information Sources:

CAG: Various reports

MOSPI: various quarterly reports and other publications.

NHAI: various reports and other publications.

MOF website for PPPs https://www.pppinindia.gov.in/