Page 1 of 18

European Banking Federation aisbl – 56 Avenue des Arts, B-1000 Brussels Phone: +32 2 508 37 11 – Website: www.ebf-fbe.eu - EU Transparency register ID number 4722660838-23

EBF_017047F

1 October 2015

EBF Draft Response to the EBA Discussion Paper on the SME Supporting Factor

The EBF welcomes the carefulness that EU regulators have shown during the regulatory reform in the assessment of policies affecting the SME lending asset class. In particular, the introduction of the SME Supporting Factor (SF) by co-legislators in the text of the Capital Requirements Regulation (CRR) reflected the reduced risk involved.

It is also important to note that the overall impact of the regulatory reform may affect asset classes differently. The use of capital has become a critical variable in banks management and the regulation imposes new constraints to the decisions on asset allocation. The combined effect of increased capital requirements and the liquidity ratios put certain asset classes, like SME lending, at a disadvantage. Notwithstanding, SME lending was not at the core of the financial crisis that gave rise to the regulatory reform. The SME SF came to alleviate the unwarranted overall pressure on SME lending.

Against this background, the CRR included a mandate to the EBA for an analysis of the lending trends and conditions for SMEs, their riskiness during an economic cycle and the consistency of own funds requirements with the outcome of that analysis. We note that the EBA has already done research and collected valuable information on overall SME performance and lending trends. Many of the questions raised in the discussion paper refer to the experience of individual banks. The EBF consolidated response gives indications received from member banks but needs to be complemented with the specific feedback from individual banks. An annex with research by the Italian Banking Association is included in this response. It sheds light on the trends of SME lending after the implementation of the SME SF using available data from various sources including the ECB and the Bank of Italy.

The ultimate question of the discussion paper is to ascertain if the SME SF has supported SME lending. Given the variety of economic factors involved it is, in our view, virtually impossible to isolate the effect of a single factor like the SME SF. In this regard, there are a number of factors influencing the SME lending that should be taken into account such as: a reduced demand for credit for investment purposes during economic downturn; the increased banks’ financing costs, reflected in credit spreads negatively affected the demand for loans; the increasingly enhanced credit guarantee schemes used by Governments as policy tool to improve the access to finance by SMEs; the EIB support provided to SMEs reduced the financing constraints. Hence drawing outright conclusions would not be plausible. The EBA rightly acknowledges this feature in the discussion

The European Banking Federation is the voice of the European banking sector, uniting 32 national banking associations in Europe that together represent some 4,500 banks - large and small, wholesale and retail, local and international - employing about 2.5 million people. EBF members represent banks that make available loans to the European economy in excess of €20 trillion and that securely handle more than 300 million payment transactions per day. Launched in 1960, the EBF is committed to creating a single market for financial services in the European Union and to supporting policies that foster economic growth. Website: www.ebf-fbe.eu

Page 2 of 18

paper. However some hints can be stated from the observations available. In the opinion of the EBF these observed facts should be highlighted in the EBA assessment:

1. In the absence of conclusive empirical evidence estimates take on special significance. A strong point of reference is the estimated impact of capital requirements on the volume of lending. The EBA paper cites some relevant studies including the global one coordinated by the BCBS in 2010 and a more recent one by Méssonier and Monks using EU data from the EBA. Both studies point towards a decline of more than 1% in lending volumes when the capital requirement is lifted by 100 basis points.

2. A simple and straight comparison of data collected in the EBA paper indicates that banks have roughly maintained the level of non-defaulted SME lending in the years after the crisis. Indeed, the decline in lending volumes is equivalent to the increase in defaulted assets. Presumably, risk managers cut credit lines to defaulted borrowers and this roughly explains the gap in a situation of limited new credit demand.

3. The fact that there is not a liquid market of SME distressed loans explains the seemingly high non-performing rates compared to those of large corporates where the availability of capital has been simultaneously better. In a protracted economic crisis, it is no wonder that the stock of SME defaulted assets piles up on the balance sheet of banks giving the impression of higher riskiness in the SME segment. For a consistent comparison, it would be necessary to account for the volume of write-offs and bad assets sold which are presumably lower in the SME segment.

The EBA assessment paper should start from this premise and, particularly, by acknowledging the inverse link between capital requirement and credit volume as noted in point 1 above.

The question is not so much to demonstrate that SME lending has increased thanks to the SME SF, but to figure out how much SME lending would have shrunk without it.

Specific EBA questions Q1. Do you have systems in place to track the reduction in capital due to the application of the SME Supporting Factor (capital relief)? Yes/No. Please explain and provide evidence. Though the question seems to be addressed to individual institutions, the EBF can presume that banks at large are able to calculate the impact of the SME Supporting Factor (SF). The capital relief due to the application of the SME Supporting Factor can be monitored through the COREP reporting templates (EU Regulation 680/2014 of 16 April 2014). These reports include a breakdown of “Exposures subject to SME Supporting Factor” under the Total Exposure caption. Therefore, based on this breakdown, which provides as well the Risk-Weighted Asset (RWA) information pre and post the application of the supporting factor, it would be easy to calculate the capital relief by simply adding the amounts relieved in each portfolio. In addition to these COREP reports, banks usually follow the variations on credit exposure by asset class for internal credit risk management purposes and thus they can control the impact of the SME SF.

Page 3 of 18

Q2. In your experience, is the reduction in capital requirements due to the application of the SME Supporting Factor (capital relief) being used to support lending to SMEs? Yes/No. Please explain and provide evidence. The perception of EU banks is that the SF has positively influenced SME lending patterns. In particular:

- The SF keeps SME capital requirements closer to those of Basel II. The EBF has long argued that SME lending was not at the core of the crisis and that imposing overly conservative requirements on SME finance would unwarrantedly increase SME credit spreads.

- More importantly, the allocation of capital among asset classes has become a critical factor in banks management after the financial crisis. Without the SF, SME lending would be put at a disadvantage compared to other lending alternatives. Banks’ lending business is driven by margins. Therefore any capital relief has a direct business-boosting impact. Specifically, the SME factor has an impact on the pricing as the cost of capital is reduced, whereas the credit decision is still very much linked to the risk assessment of the SME. However, two features of the SME supporting factor have made full use of it challenging. Firstly, the design of the SME supporting factor results in cliff effects. When the customer breaches the exposure or turnover cap an increase in capital requirement amounting to the SME supporting factor is imposed.

Q3. Is your internal definition of SMEs in line with the definition of SME exposures subject to the SME Supporting Factor? Yes/No. If no, how are you reconciling the internal definition of SMEs with the definition of SMEs subject to Supporting Factor? Please explain and provide specific examples. Our members are fully aware of the definition of SMEs as set out in the Commission’s recommendation 2003/361/EC of 6 May 2003 as well as the clarifying consolidated guidance set out by the EBA in Annex 3 of its Discussion Paper. This SMEs definition is the one used for the purpose of identifying those exposures which benefit from the SF criteria set out in the CRR. However, it should be noticed that there are multiple other definitions used for other purposes. No reconciliation of the internal definitions of SMEs to the definition used for the SF is performed because there is no business imperative or benefit from such reconciliation. Q4. In monitoring the total amount owed to you, your parent and subsidiary undertakings, including exposures in default, by the borrower and its group of connected clients (as defined in CRR Article 4(1)(39)), what reasonable steps do you take to ensure that amount does not exceed EUR 1.5 million in accordance with Article 501(2)(c)? Banks will be responding to EBA on an individual basis should they choose to do so. Nevertheless, we would like to note that the exposure limit of €1.5 million for eligibility for the SF seems too low. The range of business size covered should be more ambitious and include a wider range of SMEs.

Page 4 of 18

Q5. Do you see merits in having a harmonised definition of SMEs for reporting purposes? Yes/No. Please explain and provide specific examples. Yes. Harmonised definitions can be beneficial as they would ensure a uniform application across all EU member states. In addition, harmonisation promotes transparency and simplicity of the regulatory framework. Moreover, taking into account the next steps in relation with the Capital Markets Union a harmonised definition of SMEs will become a matter of level playing field. There are a number of different SME definitions used for internal, risk and regulatory reporting. For example the COREP reports also include splits for Retail SME and Corporate SME asset classes. Having one consistent definition would simplify reporting and reduce the potential for misunderstanding. An issue that EU policy makers will have to resolve is the relative size of SME in different EU countries. The country’s GDP, the number and concentration of SMEs, the productivity and other factors should be considered when fixing a threshold based on size. Q6. Do you agree with the proposed measures of SME riskiness? Yes/No. Are some of these measures more relevant than others? Yes/No. Whilst the five factors identified may very well be suitable proxies of SME riskiness, it should be taken into account some key issues highlighted by our members:

- These risk indicators, although useful from a credit management perspective, are largely cyclical and backward looking by design which hinder any risk assessment system from predicting SME risk of insolvency and allowing issues to be resolved and measures to be put in place before risks materialise. How to assess them objectively. Many of the proposed risk drivers are based on accounting conventions that are not yet globally harmonised. Depending on the country jurisdiction there are small companies which are not required to publicly file any accounts, and some other countries that although having the obligation to comply, this information may be prepared up to nine months after the financial year end. Therefore, in order to be able to comply with these measures of SME riskiness on a current value basis it would be necessary to change the small business accounting conventions. However, this might not be enough to identify and anticipate the SMEs risk-profiles. The main difficulty faced by banks when lending to SMEs is the accuracy, volatility, and availability of the financial information provided by SMEs.

- A “one size fits all” approach is too imprecise to do justice to the heterogeneity of the SME sector. A differentiated analysis making a distinction between sectors or clusters of sectors (i.e. trade, manufacturing industry, services) and countries is therefore required.

Q7. Are other aspects relevant in your assessment of the creditworthiness/riskiness of potential SME borrowers? Yes/No. If yes, please provide a list of those aspects and explain how you measure SME riskiness. Yes. Other key risk indicators used to assess the riskiness of SME borrowers include:

- Debt Service Coverage Ratio (EBITDA / Capital Repayments + Interest)

- Debt / EBITDA

- Overall size of the business

- Life stage of the business

Page 5 of 18

- Consolidation risk in the case of lending to finance mergers and acquisitions

- Ownership and financial position of natural person(s) behind the company

- Quality of management (competences, professional experience and perspectives about the succession)

- Quality of financial information

- Competitive position of the customer in the specific market; diversification of customer base; diversification of supplier base; number of competitors in market.

- Maturity of customers products in terms of life cycle and diversification / concentration of customers products

- Value of collateral offered

- Sensitivity to economic environment and interest rate changes

- Payment behavioural information

- Financial debt and behaviour in the financial system

- The stable relationship with the institution, measured, for example, by the costumers’ history or by the percentage of the business activity performed with the entities

- Indicators should be estimated both at a given time and with a past and forward-looking perspective

- Some specific indicators for individual entrepreneurs and self-employed, as for example: o Turnover (except if it is included in the overall size of the business indicator) o Taxes

VAT Income tax

o Financial Assets Statements As a conclusion, several studies show that a close bank-firm relationship has a positive effect on bank lending. Q8. In your experience, are SMEs as cyclical or more/less cyclical than large enterprises? Generally SMEs tend to be more cyclical than large enterprises, but this largely depends on the sectors, companies and their risk. Different industries and sectors may be impacted to a different degree, and with a different timing, through an economic cycle. The default rates and increases in the PDs produced by the models during the financial crisis align with this view; and with the view of the Discussion Paper.

In addition, it would be important to note that the analysis presented in the Discussion Paper does not take into full consideration the following aspects:

- By definition, the SME group includes the new businesses (start-ups) which are more prone to failure due to the less stable and robust source of revenues which penalise the access to credit. So, the SME data on default may be biased by the higher weight of new business, which are more likely to default than the universe of large corporate. In this sense, a kind of “vintage” analysis of SME defaults making distinction between new SMEs vs. established SMEs could be useful for the purpose of the discussion paper;

- SME usually have shorter loan maturities hence being more prone to refinance risk at times of prolonged crisis;

Page 6 of 18

- Further consideration should be given to the conclusion presented on paragraph 31, by pointing out, as a major factor for riskiness, the importance of the type of activity performed instead of firm size;

- Though the data focus on riskiness of SMEs no analysis is performed on the riskiness of portfolio composed of Large Corporates and SMEs. In the end that is what matters for capital requirements as stated on paragraphs 32 to 35 of the discussion paper “asset correlation increase with firm size”. So, it seems appropriate to have in place a SME supporting factor in order to incentivise the more diversified banking portfolios that may turn out to be more robust in a crisis scenario.

Q9. Do you agree with the proposed methodology to assess the own funds requirements in relation to SME riskiness? Yes/No. If no, please provide alternative methodologies or indicators, if available. We do not believe there is sufficient information to evaluate the proposed methodology on own funds requirements robustly. EBF members look forward to the publication of the results of the EBA empirical project. Q10. Did the arrears and loss experience in 2009/2010/2011 exceed an (internal) limit? Yes/No. Were (expected/unexpected) losses adequately covered by loan loss provisions? Yes/No. Please explain and provide specific figures. Banks will be responding to EBA on an individual basis should they choose to do so. Q11. Do you agree with the above interpretation of statistical data on lending trends and conditions? Yes/No. If no, please explain. In relation to the statistical data on lending trends and conditions analysis our members would like to point out the next comments:

- The EBA should cover all EU countries rather than only the euro area countries.

- The definition of “New Lending” as ‘loans other than revolving loans and overdrafts…’ is a flawed definition which undermines the statistical data. These products are a key source of funding for SMEs. Indeed, for many businesses whose activities are confined to trading, such as those involved in wholesale or retail distribution overdrafts and revolving exposures are frequently the most significant source of funding. This impacts the analysis contained in the section entitled “Lending trends”. By contrary Paragraph 3 in section 4.1 incorporates both overdrafts and loans within the definition of “banking financing”.

- Finally a necessary caveat is that SMEs trends are proxied by loans < 1M€. Q12. Since 1 January 2014, have you changed your SME credit lending and assessment policies and procedures, specifically as a result of the introduction of the Supporting Factor? Yes/No. If yes, please explain and provide specific examples. Since January 2014, banks do not seem to have changed significantly their SME credit lending assessment policies and procedures. The risk assessment rather than the cost of funding the transaction in focus is crucial when the decision to grant a loan is taken. The cost of capital does, on the other hand, has an impact on pricing.

Page 7 of 18

It can be argued that banks could have changed their policies to the detriment of SME lending if the SF had not been introduced. The SF must have partially offset the disadvantaged position where SME lending was left as a result of the wider regulatory reform Q13. Have changes to your SME credit lending and assessment policies and procedures been driven by other factors (e.g. competition from alternative sources of SME financing as described in section 4.1)? Yes/No. Please explain and provide specific examples. No comment at aggregate level. Q14. In your experience, is there an impact of the SME supporting factor on the volume of SME lending compared to other loans? Yes/No. Please explain and provide evidence.

Although there is a wide range of factors influencing lending conditions in the market place, the cost of capital is a decisive factor. Therefore, to a certain extent, the SME SF reduces the cost of SME lending. Otherwise, banks could arguably be forced to deleverage their SME portfolios.

The overall impact of the regulatory reform leaves SME lending worse-off than other asset classes in spite of the fact that SME lending was not at the core of the crisis. In particular, the combined effect of enhanced capital requirements and liquidity rules put SME lending at a disadvantaged position compared to other assets. The stricter capital requirements lead to deleveraging whilst the liquidity rules force banks to invest a considerable part of their resources in qualifying assets, mainly government bonds. The stringent scenario is compounded by the overly restrictive conditions for SME loan securitisation.

The original intention was to maintain the capital requirements of Basel II as SME lending was not at the core of the assets that gave rise to the financial crisis.

Please refer to the annex for further evidence on the effect of the SME SF.

Page 8 of 18

Q15. In your experience, is there an impact of the SME supporting factor on the pricing and overall conditions of SME lending compared to other loans? Yes/No. Please explain and provide evidence. Although this question could be answered with previous response to Question 14, we could complement it by adding that the SME SF is included in calculation of lending rates via the cost of capital. Banks usually set minimum margins for credit transactions and the application of the SME SF allows for a wider range of SME transactions to meet the minimum margin target. Q16. Do you consider SMEs are a consistent group when it comes to access to credit or should a distinction be made between different types of SMEs (e.g. micro, small and medium ones)? Yes/No. Please explain and provide specific examples. Given the diversity among SMEs, namely on size, sector and market, we acknowledge that further regulatory differentiation might be challenging considering the complexity involved. However, we believe that given the diversity among the SMEs, it is necessary that regulators establish the right balance between complexity and risk sensitivity. In order to adequately reflect the SME risk profiles, we consider that is necessary a different treatment according to the type of product, geography, etc. SMEs show a clear different behaviour depending on their jurisdiction. SMEs treatment should be calibrated in order to reflect the specificities of the sector and geography where these companies are located. These two components are critical for this segment, highly influenced by the local business culture and the local rules. For instance, some SMEs are not required to publish financial statements according to the local accounting standards.

Page 9 of 18

Annex1 1. Section 4.3 Riskiness of SMEs in the European Union

Questions 6 to 10 o the EBA document look at the risk profile of SMEs and, therefore, the adequacy of the current risk weightings for loans to this category. As is known, the application of a discount to the RW of loans to SMEs is justified by the lower correlation among the individual PDs of enterprises in this size class, presumably because the characteristics of SME default events are more diversified than those of larger enterprises. In prudential terms, this means that despite a higher average probability of default, over time a portfolio of SME loans may generate a lower unexpected loss than that deriving from a portfolio of loans to large firms. In Basel terms, this results in a lower capital requirement for loans to the former than to the latter. Estimation of this correlation is somewhat complex, partly due to the limited availability of microeconomic data for a broad range of loan portfolios. In order to estimate this important parameter, in the past (Finance and economics “Towards Basel 3. Asset correlation and SMEs: evidence from estimates using macro data”, July 2012) we proposed a macro approach, using Bank of Italy quarterly data on the rate of new defaults on loans to the productive sector (enterprises and family businesses), to estimate the default correlation (and, consequently, the asset correlation) for small and medium-sized enterprises and large firms. The estimates obtained at the time indicated that

the asset correlations for portfolios of SME loans were systematically lower than those

for large firms.

Basel 2 uses asset correlation thresholds which require significantly higher capital

absorption than that justified by the estimated risk, with the corresponding implication

for policy, as translated into the SME SF, to lower the capital requirement for loans to

SMEs.

The past estimates were based on a sample time interval extending from the first quarter of 1990 to the last quarter of 2010. In this paper, we start from the above methodology to determine if the adverse macroeconomic conditions experienced over the past three years confirm the expected lower pro-cyclical nature of the risk associated with a portfolio of SME loans with respect to another comprising loans to large firms. The time series of new default rates is currently updated to the first quarter of 2015. As a proxy for the riskiness of loans to SMEs we have referred to the new default rates of family businesses and enterprises with credit facilities of up to 500 thousand euro, while the new default rates of enterprises with credit facilities in excess of 500 thousand euro reflect the riskiness of loans to large firms. Chart 1 shows the cyclical trend in the risk associated with the two categories of enterprise, considering the new default rate based both on the loan amount and on the "loan numbers" (summation of amounts outstanding x days outstanding). The chart clearly shows the more cyclic nature of the riskiness of loans to large firms. When measuring risk based on the loan amounts, the difference between the max and min values of

1 Annex developed by the Italian Banking Association with a few refinements and broad support from other EBF members.

Page 10 of 18

3.9 percentage points for large firms compares with 2.1 percentage points for smaller enterprises. Considering the risk based on loan numbers, the difference between the max and min values is 3.8 points for large firms and only 1.5 points for smaller enterprises. In addition, looking at the chart for loan amounts, the risk cycle is seen to be symmetrically wider for large firms, while the risk indicator based on loan numbers shows that the more cyclic nature of the riskiness of loans to large firms mainly relates to phases during which the risk is increasing.

Chart 1. New default rates (moving annual rate) Loan amounts Loan numbers

It is also important to note the behaviour of risk during the recent cyclic downturn. Starting from the end of 2007, the riskiness of large firms rose by 3.3 points in loan amount terms, compared with an increase of 2.1 points for smaller enterprises. The imbalance is even clearer from the measurements based on loan numbers: in this case, the increase in risk of 3.7 points for large firms compares with just 1.4 points for smaller enterprises. So, with reference to the recent cycle, it seems fair to conclude that the hypothesised greater pro-cyclic nature of the riskiness of loans to large firms has been confirmed once again. This fact is more clearly shown by analysing the distribution over the above time interval of these two risk indicators (Chart 2).

Chart 2. Distribution of new default rates (quarterly deseasonalised 1996q1 - 2015q1)

Loan amounts Loan numbers

ABI analysis of Bank of Italy data

ABI analysis of Bank of Italy data

Page 11 of 18

In both cases, the distribution of the new default rate for large firms has a heavier tail. In particular with regard to the loan amounts, the risk distribution for large firms has heavier tails both to the left (smaller amounts) and to the right (larger amounts): as a summary indicator of variability, the difference between the 95th and the 5th percentiles is 0.9 points (3.6 in annualised terms) for large firms and 0.5 (2.0 annualised) for SMEs. On the other hand, considering a summary indicator of unexpected risk, the difference between the 95th percentile and the median is 0.6 points (2.2 in annualised terms) for large firms and 0.3 (1.2 annualised) for SMEs. With regard to the distribution of the new default rate, determined based on loan numbers, the left tails are the same but the right tail (higher levels of risk) for large firms is very much heavier: the summary indicator of variability is 0.9 points (3.6 in annualised terms) for large firms and 0.3 (1.2 annualised) for SMEs. In this case, the summary indicator of unexpected risk is 0.7 points (2.7 in annualised terms) for large firms and just 0.2 (0.9 annualised) for SMEs. Taken together, the macro evidence obtained from analysing the new default rates for the productive sector seems to confirm the hypothesised lower asset correlation for the risk distributions of smaller enterprises with respect to large firms. The greater diversification of risk inherent in a portfolio of loans to SMEs appears to have strengthened during the recent downturn, during which the riskiness of large firms grew significantly more than that of smaller enterprises. This conclusion applies regardless of whether the measurement is based on loan amounts or loan numbers, although the difference in behaviour in the latter case is much more significant.

2. Section 4.4 SME lending trends and conditions

Questions 11 to 16 address the effects to date of applying the SME SF on the availability of loans and the related cost. In summary, the EBA discussion paper asks two principal questions and seeks to provide initial responses:

1) Following the SME SF, has lending to SMEs started to increase? If yes, is this increase

faster than the rise in lending to large firms?

2) Interest rates: following the SME SF, has the cost of borrowing for SMEs fallen with

respect to that for large firms?

Two types of empirical work are presented in relation to these questions:

the first type, based on quantitative data and emulating the analyses presented by the

EBA, seeks to highlight better the differences in the trends pre and post SME SF;

the second type, based on qualitative data and considering solely the experience in Italy,

seeks to obtain the opinions of firms about the willingness of banks to grant finance.

Quantitative analysis With regard to the analysis of quantitative data, we refer to the EBA analysis of new lending by size of firm. Based on the data available, the EBA discussion paper considers the loan amount granted to be a proxy for firm size, so:

1. loans of less than 1 million euro are treated as loans to SMEs

Page 12 of 18

2. loans of more than 1 million euro are treated as loans to large firms

the data used by the EBA was taken from the ECB database of harmonised interest rates and refers to the volume of new loans granted to firms. We used the same database and have tried to focus attention on the differences before and after introduction of the SME SF. For this reason, Chart 3 shows the cumulative flow of new lending over a 19-month period, being the period from introduction of the SME SF (January 2014) to the latest available data (July 2015).

Chart 3. Flow of new lending by size of firm (19-month moving total; December 2013 = 1)

Given the accumulation period, the period-end data shown in the chart indicates the growth in new lending to firms from the start of the SME SF, with respect to a period of equal length prior to the start of the SME SF (June 2013 - December 2014). In addition, the trend in this moving total over time gives an idea of the change in the availability of finance to firms. Just considering the change in lending during the 19 months from introduction of the SME SF, the results are very clear: lending to SMEs has increased by 2%, while lending to large firms has decreased by 7%. Further, notice how the downward trend in lending to SMEs changed direction rather quickly after introduction of the SME SF, while the documented decline in lending to larger firms continued until the end of 2014, before showing signs of recovery. The change in the flow of new loans therefore appears to indicate that, following introduction of the SME SF, lending to SMEs has risen by more than in the past and by more than the rise in lending to large firms. The above evidence refers to the Euro area as a whole, while Table 1 shows the same data at national level, in order to check if the improvement in lending to SMEs following introduction of the SME SF is common to all member States, or the result of dynamics found only in certain countries. This check appears to indicate that the above result is spread somewhat uniformly among the members of the EMU. Calculating the difference between the growth in lending to small and

<<<-SME SF->>>

ABI analysis of ECB data

Page 13 of 18

large firms (last column in the table), it emerges that there is a positive growth differential of 9 p.p. at area level and that this positive differential exists for as many as 8 of the 12 countries considered; also, the greater growth in lending to SMEs is confirmed in all 4 principal EMU countries, where the positive growth differential in lending to SMEs is almost 12 percentage points. Table 1. Flow of new lending by size of firm (data in millions of euro)

19-month moving total ended

Small firms Large firms Small-Large

dic-13 lug-15 Chng dic-13 lug-15 Chng Chng

Austria 21,728 20,235 -6.9% 150,193 142,541 -5.1% -1.8%

Belgium 128,816 123,230 -4.3% 400,890 322,839 -19.5% 15.1%

Germany 183,727 191,923 4.5% 920,659 915,721 -0.5% 5.0%

Spain 215,499 242,860 12.7% 447,524 352,407 -21.3% 34.0%

Finland 11,605 11,838 2.0% 49,164 45,593 -7.3% 9.3%

France 113,318 114,531 1.1% 259,643 237,587 -8.5% 9.6%

Ireland 5,638 5,768 2.3% 12,954 29,623 128.7% -126.4%

Italy 259,368 263,995 1.8% 410,064 398,136 -2.9% 4.7%

Netherlands 29,049 28,540 -1.8% 129,125 150,168 16.3% -18.0%

Portugal 30,025 29,557 -1.6% 45,777 31,620 -30.9% 29.4%

Slovenia 3,231 2,078 -35.7% 10,560 6,960 -34.1% -1.6%

Slovak Rep. 1,940 2,083 7.4% 18,382 19,302 5.0% 2.4%

Euro area 1,037,967 1,062,101 2.3% 3,038,016 2,830,864 -6.8% 9.1%

Big 4 771,912 813,309 5.4% 2,037,890 1,903,851 -6.6% 11.9%

Another quantitative factor to consider is the cost of borrowing. Here too, we have followed the indications given by the EBA and considered (Chart 4) the change in the spread between the rate charged on SME loans and that on loans to large firms over the 19-month periods pre and post SME SF. Once again, the distinction between small and large firms is made with reference to the size of the loan granted. Consistent with the previous evidence, the period following introduction of the SME SF appears to be marked by a relative improvement in the borrowing costs incurred by small firms: while the rate spread at period end was 3-tenths narrower than it was prior to introduction of the SME SF, it averaged 135 bp over the 19-month period considered, which was almost 20 bp less than the average for the 19 months prior to introduction of the SME SF. Once again, the trends in the various EMU countries were well aligned, with a reduction in the spread between the average rates for the two periods indicated in 8 of the 112 countries considered.

2 Out of the 12 countries considered previously, data for loans in excess of 1 million euro was not available for Belgium.

ABI analysis of ECB data

Page 14 of 18

Chart 4. Spread of rates on loans to small vs large firms

Overall, the quantitative evidence reported shows that there has been an improvement in both the availability and cost of SME loans following introduction of the SME SF: this improvement can be seen in comparison with both previous trends and the dynamics of larger firms. This fact suggests that the capital discount applied to SMEs has had the desired effect, by improving the conditions for their access to finance.

Qualitative analysis For completeness of information, it is useful to refer to the qualitative comments expressed by SMEs about their degree of satisfaction with the availability of bank finance. This type of analysis has certain advantages, with respect to the analysis of quantitative data, when studying the dynamics of lending to firms. In particular, the separation of supply-side from demand-side considerations enables the effectiveness of the measure under review to be assessed more clearly. Additionally, there is no need to use proxies to study the flow of lending by size of firm, since the loan details are available. The EBA discussion paper uses some of the replies to surveys of firms and lending that were carried out on a harmonised basis as part of survey work on the “Access to finance of enterprises”. For the purposes of this paper, we used the questions about loans contained in two surveys of Italian firms. In particular:

1. Survey by the Bank of Italy on the expectations for inflation and growth: this survey

addressed two classes of enterprise (medium and large) in the industrial and service

sectors; the results enabled us to calculate a credit restriction indicator;

2. Survey by Istat on the confidence of manufacturing firms: this survey considered three

classes of manufacturing enterprise (small, medium and large), enabling us to both

determine the opinion of firms about lending conditions and calculate the percentage

of firms that did not obtain the finance requested.

<<<-SME SF->>>

1.35 1.54

ABI analysis of ECB data

Page 15 of 18

As before, we used this information to assess whether or not, following introduction of the SME SF, the target group's (SMEs) access to finance has improved with respect to both earlier trends and that of large firms. The first analysis developed a credit restriction indicator from the Bank of Italy's survey of expectations about inflation. This survey requested firms to express an opinion on the conditions for access to finance, making it easy to calculate a credit restriction indicator as the net of adverse responses (“conditions worsened”) and positive responses (“conditions improved”). The trend in this indicator over time is shown in Chart 5.

Chart 5. Degree of credit restriction (Italy: Industrial and service-sector enterprises)

Only two classes of enterprise were considered: those with between 50 and 200 employees and those with more than 200 employees. In this case, the comparison is between medium-sized and large enterprises, rather than between SMEs and large firms. The chart appears to tell us that overall financing conditions have improved for all firms; this said, the data in Table 2 indicates that the rate of improvement for medium-sized enterprises has accelerated following introduction of the SME SF (-24 points, compared with -15 in the 6 quarters prior to introduction of the capital discount) and, also, that the reduction was greater than that obtained by large firms.

Table 2. Change in credit restriction and SME SF

Size class

50-200 >200

June2012-Dec2013 -14.6 -19.7 Dec2013-June2015 -24.1 -20.4

ABI analysis of Bank of Italy data

Page 16 of 18

The results are even more evident considering the Istat survey of firms in the manufacturing sector (Chart 6). Considering both the chart and the summary provided in Table 3, the different magnitude of the improvement in the conditions for access to finance is very clear: an increase of 25-27 points for SMEs, compared with 11 points for large firms.

Chart 6. Access to finance (Italy: manufacturing firms)

Additionally, the change in the trajectory of the SME improvement following introduction of the capital discount is very obvious, while this change is not identifiable from the information for large firms: in the 20 months prior to introduction of the SME SF the access conditions for SMEs improved by 4-9 points, while in the following 20 months there was an improvement of 25-27 points; for large firms, on the other hand, the improvement was about 10 percentage points both before and after.

Table 3. Change in the conditions for access to finance

Size class

<50 50-250 >250

20 months before 4 9 10

20 months after 27 25 11

This differential effect benefiting SMEs is also suggested in relation to the rationing measures identifiable from the survey replies. In particular, the percentage of firms that did not receive a requested loan can be calculated, as shown for the three classes of enterprise in Chart 7 and summarised in Table 4.

ABI analysis of Istat data

Page 17 of 18

Chart 7. Percentage of firms that did not receive a requested loan

As can be seen, the probability of not obtaining finance has decreased considerably for both small and medium-sized enterprises post-SME SF, while it is essentially unchanged at a fairly low level for large firms. However the change in the probability gradient before and after introduction of the SME SF is of even greater interest. In the 20 months before, both small and medium-sized enterprises complained about an increase in the probability of not obtaining finance, while it was essentially stationary for large firms. As already described, this probability declined significantly for SMEs in the following 20 months and, at least in the case of small enterprises, this inversion occurred essentially at the same time as the start of the SME SF.

Table 4. Change in the probability of not obtaining finance

Size class

<50 50-250 >250

20 months before 4.2 6.1 -1.1

20 months after -5.7 -5.6 -0.1

The evidence gathered from these sample surveys appears to indicate that application of the SME SF in Italy has had a significant positive effect on the financing conditions available to SMEs. Following introduction of this measure, the opinions of these firms about lending conditions have improved significantly, with respect to both the past and those of large firms. The same also seems true for the probability of not obtaining finance, which has contracted significantly for SMEs following introduction of the SME SF. Accordingly, it appears reasonable to conclude that application of the SME SF in Italy has significantly and positively influenced the strategies of banks in their lending to SMEs.

ABI analysis of Istat data

Page 18 of 18

3. Conclusions

Taken together, the evidence supports a position that a) requests continuation of the SME SF

b) confirms, contrary to fears expressed by the EBA and consistent with the position

already taken by ABI, that the measure has not impeded the desired growth in

capitalisation ratios, having an effect of about 20 basis points on a CET1 that has

risen significantly

c) highlights how, in the still brief application period, the positive effects have

mitigated the adverse consequences of the deep and prolonged recession

d) confirms that the macroeconomic reasons supporting this measure combine with

the structural reasons, being the lower riskiness of SME loan portfolios compared

with large firm portfolios, due to a diversification effect that makes the default rate

on SME loan portfolios less volatile.