Converging to higher capital requirements: the adjustment strategy and lending impact 1

Transcription

1 Converging to higher capital requirements: the adjustment strategy and lending impact 1 Kirsten van Nimwegen & Allard Bruinshoofd March 216 Forthcoming as Rabobank Special Report Basel III aims to improve the banking sector s shock resilience by increasing minimum capital requirements and enhancing the risk coverage of the capital framework Despite the long phase-in period up until 219 for full compliance, the banking sector has front-loaded the adjustment as expected Eurozone banks improved capitalization by adding capital, as well as de-risking the balance sheet and selling assets banks, in contrast, though operating under different accounting standards, adjusted entirely via capital additions, whilst expanding their balance sheets and adding risk Banks adjustment strategies correlate with bank lending: capital base expansion positively, and balance sheet reduction negatively Summary In a response to the global crisis, the finance ministers from the G2 countries agreed on stricter regulation of the banking system, with stronger capital requirements, both on a risk-weighted basis and on a non-weighted basis (the leverage ratio), and augmented liquidity standards, better known as the Basel III Accord (BIS, 211). Upon announcement of the Accord, central banks have suggested that its negative effects on credit supply should be limited due to a spread-out adjustment period up until full implementation in 219. However, the data show that a short-run race, which the banking sector always emphasized would happen, occurred instead. Our analysis shows that the majority of banks have already increased their 1 We thank Wim Boonstra, Johan van Donkersgoed, Berry van Nieuwkerk and Bouke de Vries for useful comments and various colleagues from our bank analysis team for discussions on initial results. All remaining errors are our own. The opinions presented in this report are not necessarily those of Rabobank. 1

2 leverage and (risk-weighted) capital ratios to full Basel III-compliant levels in their 214 annual reports. The impact on credit supply depends on banks adjustment strategy. Roughly said, there are three ways for banks to increase their capital ratios: i) increase the capital base; ii) decrease the riskiness of the balance sheet; and / or iii) selling assets. Adjustment via capital base expansion correlates positively with banks credit supply, whereas balance sheet reduction correlates negatively. Our analysis does not distinguish between supply and demand effects that jointly drive these results; particularly in countries that exhibited sharp adjustments in activity, it is likely that credit demand at least contributed to balance sheet reduction. The difference in adjustment between the and the Eurozone is striking: between 28 and 214 our sample banks credit supply increased by 2.5% in the, whereas it contracted by 4.% in the Eurozone. banks ability to increase their capital ratios entirely by capital base expansion, is likely facilitated by early stage public capital injections into the banking sector. These injections prevented banks from being forced to reduce balance sheets and allowed early hikes in risk weights. It is plausible that this also prevented constraints on credit supply, and thereby enhanced investment growth and economic recovery. In contrast, the Eurozone is still coping with very sluggish investment activity, which in part, looks attributable to de-risking the bank balance sheets. Meanwhile, prolonged economic stagnation itself and several bouts of Euro crisis may also have contributed to shorter bank balance sheets in the Eurozone by curbing (bank) credit demand. Why you have to read this study With economic growth across Europe being disappointingly slow amid abundantly available signals that the credit channel across large parts of the continent is clogged with non-performing loans, we want to know where banks stand today in their adjustment to the Basel III requirements and how they have made their adjustments so far. Our focus will be on loans provided by banks, which results from the interaction of supply and demand. Both supply constraints as well as demand shortfalls may 2

3 drive lower bank lending levels and growth rates. We explore the connection between the adjustment of capital ratios and bank lending to get a better feel of the relative importance of supply effects. Why a new Basel accord? The Great Recession, which was sparked by the subprime crisis of 27 that preceded the collapse of Lehman Brothers in 28, demonstrated that existing capital and (particularly) liquidity regulations in the banking sector were insufficient. Banks were exposed with high (and sometimes excessive) leverage, large amounts of high risk assets, over-stretched liquidity transformation and low levels of high-quality capital, making the sector as a whole vulnerable to a liquidity squeeze and exacerbating the economic contraction. Most of the rescued banks were compliant with the capital requirements in place at the time (Basel II), but nevertheless faced liquidity squeezes that threatened their continuity (Demirguc-Kunt et al., 213). In response to the crisis, the finance ministers from the G2 countries agreed on stricter regulation of the banking system, with stronger capital requirements and augmented liquidity standards (Basel III), in order to significantly reduce the probability and severity of future banking crises. What are the Basel III minimum requirements? The main objective of the Basel III accord is to enhance economic stability by improving the banking sector s ability to absorb shocks arising from financial and economic stress (BIS, 211). As a highly desirable by-product, this would also lower the probability that governments need to bail-out banks with public means. The Basel III framework aims to achieve this by both increasing capital, qualitatively (by raising tier 1 capital minimum requirements, the strongest form of risk-absorbing capital) as well as quantitatively (by increasing total capital requirements), and enhancing the risk coverage of the capital framework (BIS, 211). 3

4 Higher capital requirements Capital requirements reflect the amount of capital banks have to hold in relation to their total (riskweighted) assets. Under Basel II a minimum risk-weighted capital ratio of 8.% was in place, meaning that banks were obliged to hold a capital base worth 8.% of their risk-weighted assets. 2 The Basel III agreement aims to improve the banking sector s ability to absorb shocks arising from financial and economic stress by increasing this capital base, both quantitatively and qualitatively (Fig. 1). Moreover, it aims to reduce the procyclical effects of Basel II with the introduction of a countercyclical buffer (Caruana, 21). Figure 1: Basel III minimum capital requirements Source: Basel Committee on Banking Supervision A greater focus on high-quality capital is reflected in minimum common equity tier 1 (CET1, the highest quality capital around, consisting of shares and retained earnings) and total tier 1 (CET1 plus other high-quality capital, which enables banks to absorb losses while remain going concern) capital ratios of 4.5% and 6.%, respectively, as of January 215 (see BIS, 211, pages 12-19). Although the absolute minimum level of the capital ratio is in theory still 8.%, several additional capital buffers are gradually phased in between 216 and 219, effectuating higher minimum capital ratios in practice. 2 Note that the addition of revaluation reserves to the capital base made capital ratios pro-cyclical in this framework; an issue that the higher quality capital requirements in Basle III amongst other strove to resolve. 4

5 Additional capital buffers In addition to the minimum capital requirement of 8.%, a capital conservation buffer of 2.5% will be required. This buffer obliges banks to hold additional high-quality capital. In times of economic distress, this buffer can be used to absorb unexpected losses. When a bank does not maintain this buffer, it is not allowed to pay dividends or bonuses. Therefore the minimum capital ratio under Basel III is effectively 1.5%. A countercyclical capital buffer of up to 2.5% allows national supervisors to oblige banks to hold even more capital in periods of excess credit growth. Finally, three different buffers for systemic risk are introduced. First, globally-systemically important banks (G-SIBs) must hold an additional amount of risk absorbing buffers, given the importance of these banks for the stability of the global banking system. This so-called G-SIB-buffer varies between 1.% and 3.5%. National systemically important banks (other systemically important institution; O- SII) should also hold additional capital of up to 2.%. Finally, a systemic risk buffer is introduced, which member states may set between.% and 5.%. These three systemic risk buffers are not additive; whichever is highest determines the additional amount of tier 1 capital a bank has to hold. Note that all these ratios are based on the banks risk weighted assets. Non-risk weighted leverage ratio A minimum non-risk weighted leverage ratio of 3.% is effective as from 1 January 219, expressing the tier 1 capital base as a percentage of total exposure (total assets + off-balance sheet items). This ratio serves as a back-stop against the risk-based capital requirements, in order to prevent excessive / deliberate optimization (gaming) of risk-weighted assets (ECB, 215a). How can banks adjust? 5

6 In order to comply with the Basel III capital requirements, banks have to increase their (tier 1) riskweighted capital ratios. There are three ways for a bank to do this: i) increase its capital base; ii) derisking its balance sheet (substituting low-risk for high-risk exposures), and / or iii) reduce its total assets (Cohen, 213). Firstly, expanding the capital base can be achieved roughly by issuing new shares and / or retaining earnings, the latter for instance by cutting costs or shareholder dividend payout. Note that the pallet of possible adjustments may vary across banks, with cooperative banks or savings banks, for instance, typically unable to issue equity. Secondly, a bank can conserve on its equity need by reducing the average risk weight applicable to its activities, via substitution of risky activities with safer ones. Thirdly, a bank can reduce the length of its balance sheet, i.e. reduce total assets, for instance by selling assets or reducing the pace of origination of new loans (credit rationing). The debate on bank capital adjustment The implementation of new requirements for bank capital and liquidity (Basel III) sparked a heated debate, mostly between practitioners and supervisors, with academics spread out across the arena. Will it or won t it: ration credit? The single most hotly debated issue was whether the new, tougher capital requirements would curtail credit supply and choke of the nascent economic recovery. Cecchetti (21) on behalf of the Basel Committee on Banking Supervision that launched the Basel III requirements suggested that full implementation would generate (temporary) losses in GDP of about.2%. The Institute of International Finance (IIF, 21) on behalf of the global banking industry put the output loss instead at 3%, incurred over the course of five years of adjustment. With at least the suspicion of partiality clinging to both guesstimates, the debate raged on for a while in a quite uninformed manner. Essentially, the supervisory side continuously emphasized the spread-out period of adjustment envisioned, up until 219 for full implementation, whilst the industry kept on insisting 6

7 that under market pressure there would be a short-run race to full compliance with the new requirements, triggering balance sheet adjustment that would ration credit supply. Amid the turmoil, Smolders (211) dubbed the Basel III requirements as tough, but realistic. With implementation of the Basel III requirements in full swing, the academic debate continued as to whether Basel III would in fact be sufficient, or whether further requirements of yet higher quality and / or quantity of capital would be desirable. Admati and Hellwig (213) have been the banner men of this position, with their book also firing up the debate within Dutch academic and policy circles. See for instance Fransman (213) hitting at the weak spots of the Admati and Hellwig argumentation, and Jakobs (214) embracing it. This study aims to contribute to this debate by analyzing how far banks are in adjusting to Basel III, what adjustment strategies banks used, and whether this relates to credit supply. Looking into the nature of banks capital adjustment Cohen (213) was the first contribution to our knowledge to actually use available empirical data to provide some guidance to the various debates raging in the Netherlands and abroad on the nature of adjustment of banks to the new requirements. He analyzes the adjustment of 82 banks globally to the new capital requirements and concludes that, the bulk of the adjustment has taken place through the accumulation of retained earnings, rather than through sharp adjustments in lending or asset growth (p. 25), although also noting that banks that came out of the crisis with relatively low levels of capital were more likely to pursue adjustment strategies involving slow asset growth (p. 26). While this provided clarity of some sort, it did not exactly remove the onus from banks in terms of them providing enough credit to the private sector to facilitate economic recovery. In the Dutch setting at least, it was felt that the banking sector that had to be saved during the Great Financial Crisis at least had a moral obligation to continue a healthy flow of credit to the private sector. It wasn t until the recent analysis by Duchi and Elbourne (216) that the decomposition of supply and 7

8 demand frictions to credit supply suggested that supply disruptions had faded as a drag on Dutch economic activity since late 212. Our addition to the debate We return to the decomposition of the adjustment of the banking sector to the Basel III capital requirements. With economic growth across Europe disappointingly slow amid abundantly available signals that the credit channel across large parts of the continent still malfunctions, we first of all want to know where banks stand today in their adjustment to the new capital requirements and how they have made their adjustments so far. Secondly, we re interested in the differences in adjustment across the Eurozone in particular, as the incidence of the economic crisis has left vastly different scars in different countries. Thirdly, we want to better understand the connection between adjustment of capital ratios and credit supply. To that end, we relate the type of capital adjustment (adding fresh capital versus cutting assets and / or risky exposures) to credit supply. We operationalize this by regressing bank lending on the magnitude of the different means of adjustment available to banks. Note that with our unit of analysis being the banks, we limit our search for the impact on credit supply to the impact of our sample banks lending activities. This implies that a fraction of the observed cuts in assets could be the result of reductions in foreign activities of banks located in specific countries. Whilst we try to circumvent the resulting mismeasurement by grouping the Eurozone banks together, cuts in activities may still be felt outside aggregate credit supply because they relate to non-eurozone activities. Also, where we find significant impact on lending by our sample banks, we do not observe whether and to what extent other suppliers of credit (non-sample banks, markets, alternative financers) managed to step in. How far are banks in adjusting to Basel III? 8

9 Capital and leverage ratios were calculated and compared to the Basel III minimum requirements, for a sample of 81 banks: 52 banks in the Eurozone-12 countries, 5 banks in the United Kingdom, and 24 banks in the United States. The sample covers 28% of total bank assets worldwide and 61% of total bank assets in the Eurozone. The G-SIBs as defined by the Financial Stability Board (FSB, 215) and significant credit institutions as defined by the European Central Bank (ECB, 215a) for which sufficient data were available for the period were included. More information on the sample is displayed in table 1. A list of the banks in our sample is available in Appendix 1. Data were obtained from Bankscope on the highest level of consolidation and in the banks own currencies. Table 1: Descriptive statistics of the sample of included banks Country Number of banks Currency Minimum bank size (x 1,) Maximum bank size (x 1,) Austria 4 31,29,6 196,287,3 Belgium 2 194,47,2 245,174, Finland 1 11,427, 11,427, France 3 1,38,17, 2,77,759, Germany 1 35,126,5 1,78,73, Greece 3 75,518, 115,464, Ireland 3 36,293, 129,8, Italy 12 35,618,8 844,217,4 Netherlands 5 124,86, 992,856, Portugal 4 42,628,9 1,152, Spain 5 161,456,5 1,266,296, United Kingdom 5 725,914, 2,634,139, United States 24 $ 38,551, 2,573,126, Source: Bankscope, Rabobank Note that our data do not harmonize accounting principles between the Generally Accepted Accounting Principles ( GAAP) and the International Financial Reporting Standards (IFRS) applied in Europe. This impacts differences in descriptive statistics between the continents 3 and may also translate into differences in observed adjustment strategies. Secondly, differences in institutional and fiscal arrangements may also play their part. In the for instance, substantial volumes of mortgage loans are transferred to government sponsored agencies (Freddie Mac and Fannie Mae), 3 For instance due to the allowance of netting of derivatives positions under GAAP, but not under IFRS, which artificially shortens banks balance sheets relative to European banks balance sheets. 9

10 strongly reducing bank balance sheet length and shifting risks directly to the public sector. In the Netherlands, the mortgage interest rate tax reduction stimulates higher levels of gross mortgage lending, which to date predominantly sit on banks balance sheets. Thirdly, banks skipped the transition to Basel II rules and only now make the transition to Basel III. Such data limitations should be borne in mind when interpreting the empirical results. Individual countries The tier 1 capital ratios, total capital ratios and leverage ratios of the individual Eurozone countries, the UK, and the, are shown in figures 2a-c 4. For each ratio the minimum requirements are depicted as well. Additionally, the (fully phased in) capital conservation and (maximum) countercyclical buffers are displayed. Note that the tier 1 and total capital ratio requirements are already effective as from 215; the leverage ratio will be introduced in January 219; and the additional capital buffers will be phased in between 216 and 219 (figure 1). The sample banks in all individual countries, on average, show increases in tier 1 capital ratios and total capital ratios between 28 and 214. The minimum requirements regarding the tier 1 (6.%), and total capital ratio (8.%) are, on average, achieved in all countries. When the capital conservation (2.5%) and countercyclical ( 2.5%) buffers are fully taken into account, all countries except Portugal (12.%), on average, already meet the fully phased in Basel III requirements. Leverage ratio requirements of 3.% are achieved by all countries. GIIPS versus the other Eurozone countries There are meaningful differences between the GIIPS countries and other Eurozone countries. Greece, Italy, Portugal and Spain have relatively low capital ratios compared to the other Eurozone countries. At the same time, leverage ratios are relatively high in Greece, Ireland and Portugal. This indicates relatively high risk-weights in the GIIPS countries, associated with lower quality assets, which is 4 Luxemburg is excluded because of insufficient data. 1

11 characteristic for economies faced with prolonged, deep recessions. It is also in line with the impressive number of non-performing loans in the GIIPS countries. Several Eurozone countries show exactly the opposite. Belgium, Germany and the Netherlands face high capital ratios, but low leverage ratios, indicating relatively low risk weights / higher asset quality. Looking at the Eurozone, the GIIPS countries, and the Eurozone excluding GIIPS as a whole, all Basel III capital requirements are, on average, achieved. Tier 1 capital ratios, total capital ratios and leverage ratios for the Eurozone are displayed in figure 3a-c. Figure 2a: Tier 1 capital ratios of the individual Eurozone countries, the UK, and the * % % AT BE DE ES FI FR GR IE IT NL PT UK Tier 1 capitalratio 28 Tier 1 capitalratio 214 Minimum Tier 1 capital ratio Source: Bankscope, Rabobank Figure 2b: Total capital ratios of the individual Eurozone countries, the UK, and the * % % AT BE DE ES FI FR GR IE IT NL PT UK Capital ratio 28 Capital ratio 214 Minimum capital ratio Capital conservation buffer Countercyclical buffer Source: Bankscope, Rabobank Figure 2c: Leverage ratios of the individual Eurozone countries, the UK, and the * 11

12 % % AT BE DE ES FI FR GR IE IT NL PT UK Leverage ratio 28 Leverage ratio 214 Minimum leverage ratio * Differences between the and the Eurozone may arise from different accounting principles ( Generally Accepted Accounting Principles versus the International Financial Reporting Standards, which are applied in the Eurozone). Furthermore, differences may arise from differences in financial system structure, for instance the fact that substantial volumes of mortgage loans are transferred to government sponsored agencies. Source: Bankscope, Rabobank Figure 3a: Tier 1 capital ratios in the entire Eurozone, GIIPS countries and other Eurozone countries % % Eurozone Eurozone GIIPS ex. GIIPS Tier 1 capital ratio 28 Tier 1 capital ratio 214 Minimum Source: Bankscope, Rabobank Figure 3b: Total capital ratios in the entire Eurozone, GIIPS countries and other Eurozone countries % % Eurozone Eurozone GIIPS ex. GIIPS Total capital ratio 28 Total capital ratio 214 Minimum capital ratio Capital conservation buffer Countercyclical buffer Source: Bankscope, Rabobank 12

13 Figure 3c: Leverage ratios in the entire Eurozone, GIIPS countries and other Eurozone countries % % Eurozone Eurozone ex. GIIPS GIIPS Leverage ratio 28 Leverage ratio 214 Minimum leverage ratio Source: Bankscope, Rabobank Globally-systemically important banks G-SIBs are subjected to an additional capital buffer, varying between +1.% and +2.5%. All G-SIBs in the sample individually achieve the minimum capital ratio of 8.% increased with their specific G-SIB buffer (+1.% to +2.5%). 5 However, when the fully phased in levels of the capital conservation and countercyclical buffers are also taken into account, only 67% of the G-SIBs individually meet the minimum requirements in 214. Table 2 shows the minimum requirements and actual capital ratios of the G-SIBs in our sample. Banks are allocated to buckets, corresponding to the required level of additional loss absorption, as proposed by the Financial Stability Board (FSB, 215). On average, minimum capital requirements are achieved by banks in the lower buckets (1, 2, and 3). G-SIBs that are allocated to bucket 4, on average, do not yet meet minimum capital ratio requirements. Table 2: Minimum capital requirements for G-SIBs Bucket G-SIB buffer Number of banks Mean capital ratio (214) Minimum capital ratio including GSIB buffer % n.a. n.a. n.a. Minimum capital ratio including capital conservation and countercyclical buffer 5 Note that the additional buffer may be higher if the applicable O-SII or systemic risk buffers exceed the G-SIB requirement. 13

14 4 +2.5% % % % Source: Financial Stability Board, Rabobank How have banks adjusted? The capital ratio consists of three elements: capital, total assets, and the average risk-weights for these assets: Capital ratio = Capital TTTTTTTTTT aaaaaaaaaaaa aaaaaaaaaaaaaa rrrrrrrr wwwwwwwwhtt (1) Banks can generically adjust their capital ratio by adjusting either one of these three factors or a combination of them. In order to examine how banks have adjusted, the change in the capital ratio is decomposed to isolate these factors, see equation 2. CCCC tt CCCC tt 1 = (1+ CC tt ) CC tt 1 RRRRRR ( tt /TTTT tt ) ( TTTT tt ) RRRRRR tt 1 /TTTT tt 1 TTTT tt 1 (2) where CR is the relevant capital ratio, C is the amount of relevant capital, RWA is the value of riskweighted assets, and TA represents total assets. 6 For empirical operationalization, equation 2 is transformed, so that the individual factors can be expressed as additive components, adding up to the total change in the capital ratio. To that end, equation 3 is expressed in logarithms, and a common factor F, which captures the impact of a linear transformation that allows us to cast the equation in terms of percentage point changes in the respective capital ratio. This results in the following decomposition (cf. Cohen, 213): 6 Strictly speaking, a change in risk weighted assets can be driven by either a shift in exposures between asset classes with different risk weights, or by an adjustment in individual risk weights. We do not have the data available to drive the decomposition to that level of detail, however. 14

15 CC tt RRRRRR tt CC tt 1 RRRRRR tt 1 = FF ln 1 + CC tt CC tt 1 FF (ln RRRRRR tt TTTT tt ln RRRRRR tt 1 TTTT tt 1 ) FF ln TTTT tt TTTT tt 1 (3) where common factor F equals (C t / RWA t C t-1 / RWA t-1 ) / (ln(c t /RWA t ) ln(c t-1 / RWA t-1 )). Eurozone versus United States and United Kingdom Eurozone banks adjusted to the Basel III capital requirements by expanding their capital base (+1.8 percentage point), as well as reducing the average risk weights (+.9 percentage point), and balance sheet length (+1.1 percentage point) (Fig. 4a). On average, capital ratios in the Eurozone increased considerably, by 3.8 percentage points. In the United States, on the other hand, capital ratios were already high to begin with, and on average, they increased only slightly (+.8 percentage point). This increase is entirely driven by the acquisition of capital (+3. percentage point). In fact, the effect of this capital base expansion allowed for a concomitant increase in total assets and average risk weights (Fig. 4b). Several elements may help explain the difference in adjustment strategy between and Eurozone banks. First, the government injected considerable amounts of capital into the banking sector, already in an early stage of the crisis (29). These injections and the major cleanup of bank balance sheets through the Troubled Asset Relief Plan (TARP) prevented banks from being forced to reduce balance sheets and risk weights in the ensuing years; the adjustment was very strongly front-loaded, and orchestrated by the government. In Europe, bank balance sheets where (and still are) not relieved of their troubled assets and reportedly now clog the credit channel (Beck et al., 213; Oliver Wyman, 215). Second, from 28 onwards, the Federal Reserve embarked on quantitative easing, i.e. it bought government bonds on a large scale. This fed into extremely ample liquidity conditions and low interest rates. Combined with an accommodative fiscal stance, these measures likely prevented constraints on credit supply, with positive effects also on economic growth. The economic recovery in the Eurozone is very slow when compared to the. Third, the Euro crisis may 15

16 have added insult to injury in protracting the recession in Europe and facilitating total asset reduction via lower demand for (bank) credit. In the UK, where economic recovery has been stronger than in continental Europe, though weaker than in the, an intermediate adjustment strategy was seen (Fig. 4c). Capital ratios increased considerably (+3.9 percentage points). This was achieved by a large expansion of the capital base (+3. percentage points, at least partly powered by infusion of public capital into various UK banks) as well as a reduction in total assets (+2.5 percentage points). A pickup in average risk weights, probably partly caused by the quantitative easing program of the Bank of England, counteracted these effects on the capital ratios (-2.1 percentage points effect on the capital ratio). Figure 4a: Adjustment strategies to Basel III in the Eurozone % % CR 28 dcapital drwa dta CR 214 Source: Bankscope, Rabobank Figure 4b: Adjustment strategies to Basel III of UK banks 16

17 % % CR 28 dcapital drwa dta CR 214 Source: Bankscope, Rabobank Figure 4c: Adjustment strategies to Basel III of banks % % CR 28 dcapital drwa dta CR 214 CR 28 = total capital ratio 28; dcapital = change in capital ratio due to capital base expansion; drwa = change in capital ratio due to changes in average risk weights; dta = change in capital ratio due to change in total assets; CR 214 = total capital ratio 214. Source: Bankscope, Rabobank Individual Eurozone countries There is considerable variation in adjustment strategies between individual countries (Table 3). Two thirds of the Eurozone countries increased their capital ratios by acquiring new capital, varying from +.9 (Italy) to +4. (Greece) percentage points. In other countries (Belgium, Ireland, Portugal) banks managed to improve their risk-weighted capital ratios despite seeing absolute levels of capital erode. Absorption of losses on troubled assets were more than offset by de-risking of the balance sheet (in Ireland offloading of troubled assets off bank balance sheets into the national bad bank, see box 1) and selling of activities. All Eurozone countries excepting Germany decreased the riskiness of their balance sheets, thereby increasing their capital ratios, varying from +.2 (the Netherlands) to

18 (Ireland) percentage points. The contribution of total assets to the capital ratio was positive (i.e. balance sheets were reduced) in about two thirds of the Eurozone countries, with an effect on capital ratios ranging between +.3 (Austria) and +6.1 (Ireland) percentage points. Table 3: Strategies of adjustment to the Basel III capital requirements in percentage points Country / region Change in capital ratio Change in capital base Change in RWA Change in total assets Eurozone GIIPS Eurozone excluding GIIPS Austria Belgium France Germany Greece Ireland Italy Netherlands Portugal Spain United Kingdom Unites States * Luxemburg and Finland are left out of table because of insufficient data. Source: Bankscope, Rabobank The results of our analysis do not match the results of an earlier study by Cohen (213). While he concludes that Eurozone banks continued to expand lending activity, our study shows that most Eurozone banks gross loans and balance sheets decreased in the process of making capital ratios meet the Basel III minimum capital requirements. This difference may result from the difference in time horizon applied. While Cohen studies the period between 29 and 212, we extended this time horizon to This means we included the situation before the crisis hit hard. Box 1 provides a detailed description of capital adjustments in the individual countries. Box 1: Notes on decomposition per country 18

19 Austrian banks have been exposed to elevated risks from their large exposures in Central and Eastern Europe. This has on the one hand led to capital losses, for instance due to Eurodenominated Hungarian mortgages being reconverted into Hungarian Forints. On the other hand, it has led them to have become more selective in adding risks, likely resulting in the positive contribution of the change (reduction) in average risk weights since 28. Belgian banks show a decrease in in risk-weights due to repositioning of assets to the domestic and eastern European market. German banks have seen risks increase, predominantly on shipping exposures, where activities have also been scaled back, translating into upward pressure on average risk weights and downward pressure on total assets. There have been asset write-downs in German banks as well, outright and through revaluation reserves, limiting the net upward strength of capital additions. Spanish banks have been capitalized by capital injections (indirectly) from the European Union. Moreover, bad assets have been sold, and government bonds acquired, resulting in much lower risk weights. Total assets increased as a result of international acquisitions. French banks have been de-risking, resulting in a positive contribution to the capital ratio. There have been claim emissions, but also the purchase of Fortis bank by BNP which is considered to be a cash cow. A strong capitalization is seen, which is primarily due to retained earnings. Greek banks show a large capitalization, which is due to capital injections in Greek banks via the Hellenic Financial Stability Fund, which is powered by the Troika support package. Irish banks are a special case, as massive interventions have taken place in the Irish banking sector. There have been huge asset write downs (also reflected in the government s budget deficit in 21/11 when it plugged the holes in Irish banks). Also, banks have off-loaded massive amounts of bad loans into the National Asset Management Agency (NAMA, the national bad bank ), resulting in lower risk weights on the fewer assets remaining in the bank books. Italian bank have added capital to their balance sheets, although there have also been numerous write-downs due to the prolonged economic downturn. Reduction in risk weights is probably due, at least for a large part, to ECB actions, stimulating Italian banks to invest in Italian government bonds. Dutch banks show increased capitalization, as a result of recapitalization of two of the largest Dutch banks by the Dutch government (which ING has fully repaid already and in the case of ABN Amro is being partly recovered by the phased sale of publicly held share capital), as well as retained earnings. As to the limited positive contribution from the average risk weights: mortgage and SME loan origination (both carrying relatively low risk weights) has been low for quite a part of the period analyzed. The positive contribution from (the decrease in) total assets is likely driven by the restructuring of the international operations of the banks, following state intervention and European Commission directives for change. Portuguese banks have written down, and offloaded large amounts of non-performing loans off their balance sheets. Together with the acquisition of government bonds, this resulted in lower risk weights. Additionally, banks have been more selective in loan origination. The UK central bank bought government bonds from the British banks on a large scale, thereby limiting government bond absorption in banks balance sheets and increasing the average riskweights of the British banks. 19

20 banks have been faster in cleaning up their balance sheets after the financial crisis. Early stage capital injections by the government contributed to their fast capitalizations. Compared to the other Eurozone countries, GIIPS countries banks show lower capital ratios, and capitalization lagging behind. Adjustment mainly takes place by lowering risk weights, caused amongst other factors by acquisition of government bonds, which have a regulatory risk weight of zero (Fig. 5a). The other Eurozone banks considerably adjusted by balance sheet reductions, e.g. by disposing assets (Fig. 5b). For the troubled GIIPS banks, selling assets to improve capitalization was much more difficult. Additionally, total lending (and hence assets) of GIIIPS banks may have shown resilience due to the cut downs in cross-border lending to the GIIPS region from the other Eurozone countries. Figure 5a: Capital adjustment strategies of GIIPS banks 2

21 CR 28 = total capital ratio 28; dcapital = change in capital ratio due to capital base expansion; drwa = change in capital ratio due to changes in average risk weights; dta = change in capital ratio due to change in total assets; CR 214 = total capital ratio 214. Note: the results do not change meaningfully when Irish banks are excluded from the GIIPS sample. Source: Bankscope, Rabobank Figure 5b: Capital adjustment strategies of non-giips Eurozone banks CR 28 = total capital ratio 28; dcapital = change in capital ratio due to capital base expansion; drwa = change in capital ratio due to changes in average risk weights; dta = change in capital ratio due to change in total assets; CR 214 = total capital ratio 214. Note: the results do not change meaningfully when Irish banks are excluded from the GIIPS sample. Source: Bankscope, Rabobank The impact of bank s adjustment strategies on credit supply In the light of the debate on whether the tougher capital requirements would curtail credit supply, we relate the change in total bank credit supply to the three elements of capital ratio adjustment. This allows us to examine whether the adjustment mix is related to bank credit rationing effects. In our empirical equation, the change in total gross bank loans between 28 and 214 is our explanatory variable. The changes in capital, average risk-weights and total assets between 28 and 214 are used as independent variables, as per equation (3). This regression was done on both the micro-economic (bank credit supply) and macro-economic (credit supply at the country level) level of 21

22 credit supply. Note that four banks in our sample had to be excluded from the micro-economic regression analysis due to insufficient data on credit supply at the bank-level. Impact on bank credit supply Between 28 and 214, our sample banks on average reported an increase in lending volumes by 2.5%, whilst the comparative statistics amount to -6.9% and -4.% for the UK and Eurozone sample banks, respectively. To determine the impact of the adjustment strategies of banks on their credit supply, we applied regression analysis. Table 4 shows the results of this analysis. Adjustment via capital base expansion is positively related to credit supply. An additional 1%-point improvement of the capital ratio via additions to lossabsorbing capital relates to a 2.3% higher level of bank lending. This finding seems plausible, as healthier banks are better able to either attract new capital, and originate loans, compared to less healthy banks (Admati and Hellwig, 213). These findings are in line with the findings of Cohen (213). Adjustment of the capital ratios by means of changes in total assets correlates negatively with credit supply. As a reduction in total assets leads to an increase of the capital ratio, this finding indicates that banks that adjusted their capital ratio by shedding assets and curtailing loan origination (i.e rationing credit), decreased credit supply and vice versa. A 1%-point improvement of the capital ratio via balance sheet reduction relates to a 5.8% lower level of bank lending. Adjustment via risk-weighted assets compared to total assets, showed a negative correlation with credit supply, suggesting that banks that lowered the riskiness of their loan portfolio also decreased credit supply. This correlation is not significant, however. Table 4: Impact of changes in capital, RWA, and TA on banks credit supply Coefficient Standard error t-value P-value Capital base adjustment * RWA adjustment TA adjustment ** 22

23 The dependent variable is bank credit supply (gross loans). *, ** denotes statistical significance at the 5% and 1% level, respectively. Country dummies included, but coefficients not reported. Source: Bankscope, authors own calculations The same analysis was done for the sub regions, and shows varying impacts of adjustment strategies on credit supply in the different regions (Table 5). Though varying in the degree of statistical significance, capital additions correlate positively with bank credit supply across the board, whilst total asset reductions consistently correlate negatively with it. Table 5: Impact of changes in capital, RWA, and TA on banks credit supply per region Adjustment strategy Region Number of banks Capital RWA TA Eurozone * ** GIIPS ** Eurozone minus GIIPS * UK * -11. ** The dependent variable is bank credit supply (gross loans). All regressions contain a constant term and country-specific dummy variables. *, ** denotes statistical significance at the 5% and 1% level, respectively. Source: Bankscope, authors own calculations Impact on macro credit supply The results above indicate that banks adjustments to the Basel III capital requirements involve bank credit rationing in the Eurozone and UK. This does not automatically mean that credit supply on the country level is also rationed, as besides our sample banks there are non-sample banks, but also nonbank credit supplying institutions and markets that may fill the credit gap. We ran regressions using macro-level credit supply as well, but this yielded no meaningful relationship with banks adjustment strategies. Although this may suggest that macro-level credit developments have been independent of micro-level adjustment of banks to new capital ratios, a more likely explanation lies in the fact that our bank lending data do not allow us to discriminate between domestic and foreign activities; where domestic banks curtail international exposures to conserve on capital, the credit rationing effects spill over into foreign markets and vice versa. We also note that between 28 and 214, macro-level 23

24 credit supply in the Eurozone and the UK decreased considerably more (-12.6% and -16.3%, respectively) than our sample banks loan volumes. Implications of the Basel III capital requirements Under pressure from markets, banks have hurried to achieve the Basel III capital requirements, and in their 214 annual figures, banks across the Eurozone countries, UK, and, on average, already comply with fully phased in Basel III capital requirements. Only when all additional buffers are fully taken into account, Portuguese banks on average do not yet comply with the Basel III capital requirements. The G-SIB buffer increases the minimum capital requirements for G-SIBs with an additional 1.% to 2.5%, and only two thirds of G-SIBs already meet their minimum capital requirements. When comparing the Eurozone to the, Eurozone banks were more dependent on balance sheet and risk-weight reductions than banks, which capitalized strongly and quickly. While our Eurozone sample-banks increased their capital ratios much more than their counterparts, their credit supply decreased with 4.%, whilst that of our sample-banks increased by +2.5%. banks show a strong capitalization and continued to increase total assets and add risk to their balance sheets. This strong capitalization of banks is a result of early-stage capital injections by the government, accommodating fiscal policy and early monetary interventions by the Fed. These measures likely prevented credit rationing, thereby enhancing economic growth in the. While Eurozone GDP levels are anno 215 (Q2) still below pre-crisis levels (28 Q1), the economy exhibited a relatively quick recovery in GDP volume from 29 onwards. Measures taken in the were not possible in the Eurozone on such a short term, due to the fragmented and complex political governance of the process and the emphasis on meeting the budget criteria of the Stability and Growth Pact. On the one hand, the Eurozone as a whole knew a political barrier for collective measures, as all countries had to agree on possible capital injections in individual countries banking sectors. On the other hand, local measures were not always possible, as the countries that suffered 24

25 most from the crisis, also struggled with elevated public debt levels whilst the others failed to stimulate growth. As a result, governments were not able to inject capital into their banking sectors. Bank lending is not the whole story, however. Firstly, decreased bank credit supply can be replaced by non-bank credit supply. Nevertheless, besides bank lending, total credit supply in the Eurozone also decreased considerably (-12.6%), although we have not been able to establish meaningful relationships with the type of capital ratio adjustment in the banking sectors. Secondly, various surveys indicate that besides credit supply restrictions, also credit demand falls have contributed to the observed lower lending volumes. In conclusion, banks seem far advanced in the process of adjusting capital ratios to the Basel III requirements a on average, banks in almost all Eurozone countries, the, and the UK, already comply with the these requirements, even when the additional buffers are taken into account. This confirms the short-run race for full compliance with the new requirements, which was flagged in advance by the industry. Although we establish a relationship between banks adjustment strategies and credit supply, we cannot conclude that adjusting to the Basel III capital requirements itself has rationed credit, as non-eurozone banks have managed to adjust with limited impact on credit supply. Faster adjustment by means of public injections of capital, á la the approach, might have limited the impact on bank credit supply and supported the economic recovery. Directions for further research Although this study provides insight into the connection between the adjustments to the Basel III capital requirements and bank lending, further research is desirable. It would be interesting to determine via what channels balance sheet reduction took place. Banks might have achieved this for example by selling subsidiaries or by reducing credit supply, either foreign or domestic. If banks cut in their loan portfolios it would be interesting to know whether or not the so-created credit gaps are filled by non-bank credit supplying institutions and markets. If not, one would want to know whether this might be related with a decreasing credit demand. 25

26 As a related topic, it would be interesting to see to what extent the sale of domestic versus foreign assets in balance sheet reduction is driven by political pressure to continue the flow of credit to domestic credit markets. In addition to political pressure, regulatory incentives may also have stimulated reduction of foreign versus domestic activities as a side effect, for instance by reducing the solvency relief achieved in the past by diversifying activities across borders. References Admati, A. and M. Hellwig (213), The Bankers' New Clothes: What's Wrong with Banking and What to Do about It, Princeton University Press, Princeton. Beck, R., P. Jakubik, and A. Piloiu (213), Non-performaing loans. What matters in addition to the economic cycle? ECB Working Paper Series 1515, European Central Bank, February. BIS (211), Basel III: A global regulatory framework for more resilient banks and banking systems, Basel Committee on Banking Supervision. Caruana, J. (21), Basel III: towards a safer financial system, speech delivered at the 3rd Santander International Banking Conference, Madrid, 15 September. Cecchetti, S.G. (21), Strengthening the financial system: comparing costs and benefits, Remarks prepared for the Korea-FSB Financial Reform Conference, 2 September. Cohen, B. (213), How have banks adjusted to higher capital requirements? BIS Quarterly Review, September. Demirguc-Kunt, A., E. Detragiache, and O. Merrouche (213), Bank Capital: Lessons from the Financial Crisis, Journal of Money Credit and Banking 45(6), Duchi, F. and A. Elbourne (216), Credit supply shocks in the Netherlands, CPB Discussion Paper 32, 12 January. ECB (215a), The impact of the Basel III leverage ratio on risk-taking and bank stability, Financial Stability Review, November. ECB (215b), List of supervised entities, European Central Bank, 3 December. Fransman, R. (213), The blind eye of Admati and Hellwig. FSB (215), 215 update of list of global systemically important banks (G-SIBs), Financial Stability Board, 3 November. IIF (21), Interim Report on the Cumulative Impact on the Global Economy of Proposed Changes in the Banking Regulatory Framework, Institute of International Finance, June. Jacobs, B. (214), Waarom banken veel meer kapitaal moeten hebben, Economie.nl blog, 6 February. Oliver Wyman (215), European retail and SME credit. Recovery time? Oliver Wyman & Intrum Justitia. Smolders, N. (211), Basel III: Tough but realistic, Rabobank Special Report 211-1, January. Appendix 1: List of sample banks Bankname Erste Group Bank Ag Raiffeisen Landesbanken Holding Raiffeisen-Holding Niederosterreich-Wien Reg.Gen.Mbh Raiffeisenlandesbank Oberosterreich Ag Belfius Banque Country AT AT AT AT BE 26

27 Kbc Groep Nv/ Kbc Groupe Sa-Kbc Group Bayerische Landesbank Commerzbank Ag Deutsche Apotheker- Und Aerztebank Eg Deutsche Bank Ag Dz Bank Ag-Deutsche Zentral-Genossenschaftsbank Hsh Nordbank Ag Landesbank Baden-Wuerttemberg Landeskreditbank Baden-Wuerttemberg - Förderbank-L-Bank Landwirtschaftliche Rentenbank Norddeutsche Landesbank Girozentrale Nord/Lb Banco Bilbao Vizcaya Argentaria Sa Banco De Sabadell Sa Banco Popular Espanol Sa Banco Santander Sa Bfa Tenedora De Acciones Sau Op Financial Group Bnp Paribas Credit Agricole-Credit Agricole Group Societe Generale Sa Eurobank Ergasias Sa National Bank Of Greece Sa* Piraeus Bank Sa* Allied Irish Banks Plc Bank Of Ireland-Governor And Company Of The Bank Of Ireland Permanent Tsb Group Holdings P.L.C Banca Carige Spa Banca Monte Dei Paschi Di Siena Spa-Gruppo Monte Dei Paschi Di Siena Banca Popolare Dell'emilia Romagna Banca Popolare Di Milano Scarl Banca Popolare Di Sondrio Societa Cooperativa Per Azioni Banca Popolare Di Vicenza Societa Cooperativa Per Azioni Iccrea Holding Spa Intesa Sanpaolo Mediobanca Spa-Mediobanca - Banca Di Credito Finanziario Societ Per Azioni Unicredit Spa Unione Di Banche Italiane Scpa-Ubi Banca Veneto Banca Scpa Abn Amro Group N.V. Bank Nederlandse Gemeenten Nv, Bng Cooperatieve Centrale Raiffeisen-Boerenleenbank B.A-Rabobank Nederland Ing Groep Nv Sns Reaal Nv Banco Bpi Sa Banco Comercial Português, Sa-Millennium Bcp BE DE DE DE DE DE DE DE DE DE DE ES ES ES ES ES FI FR FR FR GR GR GR IE IE IE IT IT IT IT IT IT IT IT IT IT IT IT NL NL NL NL NL PT PT 27