Basel III: no Achilles spear

Laurent Quignon
n Greek mythology, Achilles spear had the virtue of healing the wounds that it had inflicted. The macroeconomic impact studies carried out on behalf of the Bank for International Settlements (BIS) suggest that the same will be true of the Basel III prudential reforms that will apply to banks. In other words, the effects on new constraints on financing for the economy could be neutralised in the long term by the benefits of improved financial stability. But some factors in this analysis invite a degree of scepticism. The financial crisis laid bare the shortcomings of the existing prudential framework and made a thorough overhaul an overriding necessity. The G20 approved the new Basel III solvency and liquidity rules at its Seoul summit in November 2010. In December 2010 and January 2011, the Basel Committee on Banking Supervision published its latest recommendations on bank solvency and liquidity 1 . Reflecting the orders of magnitude involved, joint quantitative impact studies carried out by the Basel Committee and the Committee of European Banking Supervisors (CEBS, now the EBA2) and published in December 2010 highlight the unparalleled distortions in bank balance sheets and in the structure of financial savings and funding that would follow from the application of the new standards. Aside from differences that have more to do with form than substance, the various macroeconomic impact studies are reasonably consistent on the impact negative of the new rules on the economy. The extent of that impact is more controversial. It seems that only the assumed improvement to financial stability, with its (questionable) benefit for economic growth, has led the BIS to its favourable conclusions.

From Basel II to Basel III: considerable quantitative impacts

Basel III does not represent all the changes in banks prudential rules since the first version of Basel II; the Basel Committee significantly altered the framework for trading business in the meantime (Basel 2.5). Quantitative impact studies suggest that the new standards will make such a big difference to bank balance sheets that they will significantly affect the structure and volumes of financial savings and funding. Minimum common equity requirements multiplied by a factor of five The new Basel rules for calculating the solvency ratios numerator may be grouped under three headings: improving the quality of instruments eligible as prudential capital, increasing the regulatory deductions laid down in Basel II and raising the minimum solvency ratios. On top of that, a leverage ratio will be introduced based on the balance sheet and certain off-balance sheet items without any risk weighting. A tighter numerator The crisis revealed that certain Tier 1 capital instruments classed, after all, as core capital were unable to absorb losses3. This was particularly true of preferred shares in Anglophone countries. The Basel Committee has therefore tightened its definitions of regulatory capital.

May-June 2011

Conjoncture

Phase in arrangements
(shading indicates transition periods all dates are as of 1 January) 2011 2012 2013 2014 2015 2016 2017 2018 Migration to Pillar 1 4.5 % 1.25 % 5.75 % 80 % 6.0 % 8.0 % 9.25 % 4.5 % 1.875 % 6.375 % 100 % 6.0 % 8.0 % 9.875 % 4.5 % 2.50 % 7.0 % 100 % 6.0 % 8.0 % 10.5 % As of 1 January 2019

Leverage Ratio Minimum Common Equity Capital Ratio Capital Conservation Buffer Minimum common equity plus Capital Conservation Buffer Phase-in of deductions from CET1 (including amounts exceeding the limit for DTAs, MSRs and financials) (1) Minimum Tier 1 Capital Minimum Total Capital Minimum Total Capital plus Conservation Buffer Capital instruments that no longer qualify as non-core Tier 1 Capital or Tier 2 Capital

Supervisory monitoring

Parallel run 1 January 2013 1 January 2017 Disclosure starts 1 January 2015 3.5% 4.0 % 4.5 % 4.5 % 0.625 % 3.5 % 4.0 % 20 % 4.5 % 8.0 % 8.0 % 5.5 % 8.0 % 8.0 % 4.5 % 40 % 6.0 % 8.0 % 8.0 % 5.125 % 60 % 6.0 % 8.0 % 8.625 %

Phased out over 10 year horizon beginning 2013 Observation period begins Observation period begins Introduce minimum standard Introduce minimum standard Sources: Basel Committee, BIS

Liquidity coverage Ratio

Net stable funding Ratio Table 1 (1) Mortgage servicing rights

Tier 1 capital will now be comprised predominantly of common shares and retained earnings. Basel III abandons the core Tier 1 concept in favour of the stricter common equity Tier 1. Up until now, deductions have been imputed essentially to total regulatory capital. They will now apply to the common equity component of Tier 1 capital. The initial proposals submitted in December 2009, which excluded subsidiaries minority interests, certain deferred tax assets and mortgage servicing rights, have been amended and these elements will be taken into account with caution. Either way, the switch to Basel III will mean a considerable increase in deductions. They will extend to a subsidiarys excess capital corresponding to minority interests, for example, which can legally cover losses only at the subsidiary and not at the holding company or other group entities. The same goes for equity stakes in other banks 4, insurance companies and financial companies that were only 50% deductible under Basel II and to which a higher percentage will apply once the sum of these stakes exceeds 10% of the owner banks common
May-June 2011

equity. As the question of equity stakes in insurance companies has already been dealt with in Europe by the financial conglomerates directive, Basel III will not mean any change on this point. The three components levelled off above the 10% threshold (apart from the financial equity stakes just mentioned, mortgage servicing rights and deferred tax assets) will also be subject to a combined ceiling set at 15% of the common equity component. The amounts exceeding this limit will be progressively deducted from the common equity component of Tier 1 capital between 1 January 2014 (20%) and 1 January 2018 (100%, see Table 1). Lastly, instruments that will no longer be eligible either as Tier 1 capital (excluding its ordinary shares component) or as Tier 2 will be eliminated by successive 10% tranches every year between 2013 and 2023. Based on accounting data to 31 December 2009, the CEBS and Basel Committee quantitative impact studies suggest that the new deductions could eventually amount to between 25% and 40% of the common equity component of Tier 1 capital (Chart 1).
Conjoncture 4

Large banks will be penalised by the magnitude of goodwill and deferred tax assets, both in the EU and worldwide. Mid-sized European banks could also suffer from the new rules on financial sector equity stakes.
From Basel II to Basel III: the impact of deductions on regulatory capital
Common equity Tier one before deductions, 100 = Basel II
100 95 90 85 80 75 70 65 60 55 50
57.9 59.9 66.6 19.8 19 75.3 4.8 1.8 0* 6.3 5 4.4 2.4 2.4 0.4* 7 4.3 4.7 1.40* 2.9 3.7 1 0* 2.8 5.5

In July 20095, less than two years after Basel II went into force in the EU (2008), the Basel Committee complemented its Basel II rules on trading books based on value at risk (VaR) 6 calculated with a 99% probability over ten days with the following: - an incremental risk charge aimed at covering risks of default and ratings downgrade, based on maximum VaR calculated with a 99.9% probability over one year; - an additional capital requirement to cover stressed VaR calculated over a one-year period, featuring significant losses. The idea was to reduce the procyclical nature of minimum requirements relative to market risks; - the application of banking book capital requirements to ABS held in the trading book, in order to avoid regulatory arbitrage between the two types of portfolio; - a comprehensive risk charge subject to qualitative requirements and based on stress tests, intended to cover correlation risk between financial institutions; - increased weights for complex securitisation transactions (ABS CDOs) and certain exposures to off-balance sheet structures. In the EU, all these recommendations have been transposed into regulatory law via CRD 3 and to some extent CRD 2, which are supposed to result in national application measures before the end of 2011 (application measures for the pay aspects took effect in January 2011) 7. According to the Basel Committee, the switch from the initial Basel II framework to the revised Basel 2.5 rules implies the multiplication of minimum capital requirements for trading books by an average factor of between three and four. Credit valuation adjustments In terms of the calculation of risk-weighted assets, Basel IIIs main innovation lies in the capital that will be required for credit value adjustments (CVAs) on OTC derivatives. The latter measure the difference between the value of a portfolio without counterparty risk and the value of that same portfolio once the probability of counterparty default is taken into account. This means that losses in market value stemming from a higher probability of counterparty default ignored under Basel II
Conjoncture 5

Dductions diverses . Misc. Deductions Ecrtement 15% . 15% cap

-41.3

8.9 3.1 12.4

-33.4

-24.7

2.3 9.4

Charges administratives sur . Mortgage servicing rights* crances hypothcaires*

-42.1

4.6

. Deferred tax diffrs Actifs d'imptscredits . Equity stakes in financial Participations financires companies . Intangible fixed assets Immobilisations incorporelles
Goodwill

. Goodwill . Common equity after

Major banks Major banks Other banks Autres banks Other Grandes Grandes Autres
banques (QIS CEBS) banques (QIS BCBS) banques (QIS CEBS) banques (QIS BCBS)

Common Equity aprs deductions Basel dductions Ble III III

Chart 1

Sources: BCBS, CEBS

The new definition of regulatory capital will automatically affect risk-weighted assets, as exposures no longer deducted from capital will be reintegrated in weighted assets. In the quantitative impact studies, this impact is assessed in the section relating to capital, for which an equivalent amount has been recalculated. A bigger range of risk-weighted assets With the exception of a few activities, notably market-related activities, Basel II will continue to apply to most risk measures. But the quantitative impact of the Basel Committees subsequent recommendations are significant, to say the least. Having already been extended considerably via new Basel Committee rules in July 2009 on the weights applied to trading books (Basel 2.5), risk-weighted assets will increase further under Basel III because counterparty risk on OTC derivatives will be taken into account. Taken together, these recommendations will be applied from end-2011 onwards. Basel 2.5 The idea of a capital requirement to cover issuer risks in trading books was being considered by certain national regulators even before Basel II was introduced. It started to materialise in July 2005, when an agreement between the Basel Committee and the International Organisation of Securities Commissions (IOSCO) made provision for an incremental default risk charge, but this methods coverage of default risk proved inadequate during the crisis.
May-June 2011

A credit valuation adjustment (CVA) measures the difference between the present value of a derivative without counterparty default risk and the value when this risk is taken into account. The general formula can be written:

CVA

T t 1

CF t (1 r ) t

T t 1

CF t (1 r )t

where T is the maturity of the contract, CFt the cash flow expected from the contract at date t, r the risk-free interest rate, and the market risk premium (spread) for the counterparty. increases in risk-weighted assets (before mitigating effects), depending on the structure of their business. The banks affected most by the new calculation method include UBS (risk-weighted assets up 95%), Crdit Suisse (up 75%) and to a far lesser extent Barclays (up 37%). Logically enough, banks with the smallest trading books relative to their total assets will be the least affected. A higher minimum requirement According to the quantitative impact studies based on data for late 2009, the combination of a tighter perimeter for regulatory capital the ratios numerator and a broader denominator will reduce core Tier 1 (to become common equity Tier 1) ratios by an average 3-6 points (Chart 3). The minimum requirement will be raised from 2% to 7% (i.e. a minimum 4.5% plus a 2.5% conservation buffer, Chart 4) over a timeline extending from 2013 to 2019 (Table 1).
The impact of redefinitions on solvency ratios
. Incremental Incremental risk charge Risk Charge

and that turned out to be massive during the crisis will be taken into account under Basel III. Trades cleared with a central counterparty (CCP) and securities financing transactions (SFTs) such as repos are not concerned. We also note that the Basel Committee slightly altered the prudential treatment of CVAs on 1 June 2010 by reducing the weight for CCC rated counterparties from 18% to 10% under the standard approach 8 . The Basel Committee believes that the introduction of a layer of capital for CVAs effectively doubles the requirements relating to counterparty risk established under Basel II (i.e. when only the risk of the materialisation of a risk of default was taken into account)9. Because of their heavy involvement in marketrelated activities, and as shown in quantitative impact studies, the major banks weighted assets will increase very significantly (Chart 2). Apart from the redefinition of capital, the main sources of this jump will be CVAs, the incremental risk charge and stressed VaR (cf. supra).
From Basel II to Basel III: a broader definition of risk-weighted assets
Weighted assets, 100 = Basel II
130 125 120 115 110 105 100 95 90
Grandes Major banks Autres banques Other banks banques (QIS BCBS) (QIS BCBS) Grandes Major banksAutres banques Other banks banques (QIS CEBS) (QIS CEBS) World (20 countries) EU (21 countries)

% of RWAs ; old (Basel II) and new ("Basel 2.5" and III) definitions 12 10 8 -5.8 -5.4 7.1 5.7 4.9 Common Ratio equity Common ratio, Equity BaselIII Ble III Major banks Autres banques Autres banques Other banks Other banks Grandes
banques (QIS BCBS) (QIS CEBS) (QIS BCBS)

10.7

11.1

11.1 -4.0

10.7 -2.9 7.8

Actions -. Equities Trading book trading book +24.5 +23 +4.1 +4.0 VaR stresse . Stressed VaR

Core Tier Ratio Core 1 ratio, Tier One Basel Ble II II

6
Titrisations trading book Banking book Risque de adjustments contrepartie (CVAs) (CVA) . Redefinition Redfinition of capital du capital
. Basel Ble RWA II RWAII . Credit value . Securities -

4 2 0 Major banks Grandes

banques (QIS CEBS)

Chart 2

Sources: BCBS, CEBS

Chart 3

Sources: BCBS, CEBS

Initial indications from the major European banks also suggest that the successive application of CRD 3 (Basel 2.5) and Basel III will result in very diverse
May-June 2011

As the minimum Tier 1 ratio will rise from 4% to 6% of weighted assets (8.5% with the conservation buffer), instruments eligible for Tier 1 other than common equity
Conjoncture 6

may account for only 1.5% of weighted assets (i.e. one quarter of the 6% Tier 1 ratio). The minimum capital ratio will stay at 8% of weighted assets. But as Tier 1 must be at least 6% of weighted assets, Tier 2 which, incidentally, will no longer be divided into upper and lower will be limited to 2% of weighted assets, i.e. a quarter of total regulatory capital. The Tier 3 category will disappear. Apart from the conservation buffer of 2.5% of common equity already mentioned, this requirement will be topped up by a countercyclical buffer also made up of common equity and that will vary between 0% and 2.5% depending on the economic cycle and the discretion of national regulators (Chart 4).
Capital: Basel III requirements

a factor of five (mid-sized banks) or even seven (large banks) with full Basel III application. Naturally, the adjustment effort will be rather less for banks with solvency ratios well in excess of the Basel II minima, and we note that many institutions have already started major recapitalisation programmes following the crisis and in anticipation of Basel III. At the end of 2009, according to the Basel Committees study, the capital needed to be raised by banks with a common equity ratio of below 7% in order to reach that level amounted to 602 billion, of which 577 billion for group 1 banks and 25 billion for group 2 banks. In the CEBS study, the capital needed to be raised was 291 billion, of which 263 billion by group 1 banks and 28 billion by group 2 banks (Charts 5 and 6).
Mojor banks' capital requirements

14 12 10 8 2,0 % 6 4 2 0
Chart 4 Source: Prudential Control Autority
Chart 5
Total Tier 1 6,0 %

Countercyclical capital cushion Conservation buffer

0-2,5%

2,0 % 1,5 % 0-2,5 %

Tier 2 Tier 1 CET or equivalent

700 600 500 400 300

euros bn (CET1: Common Equity Tier One)

CET1 -> 4.5% CET1 -> 7% Rsultat net 2009 Net income 2009
263

577

2,5%
Tier 2 Non-core Tier 1

1,5 % 4,5 %

7,0 %
Common equity Tier 1 - CET

Common equity Tier 1

209

200 100 0
EU (21 countries) Union Europenne (21 pays)

165 53 84

World (20 countries) Monde (20 pays) Sources: BCBS, CEBS

The BIS impact study shows that the 74 group 1 banks 10 that supplied the BIS with detailed enough information for solvency ratio calculation purposes would have had an average common equity Tier 1 ratio of 5.7% under Basel III on 31 December 2009 11 , assuming full application of the new rules. Their common equity ratio was 11.1% under Basel II at the same date 12 . 133 mid-sized banks would have seen these same ratios ease from 10.7% to 7.8%, suggesting a much greater impact for large institutions. The 33 major European banks that participated in the solvency part of an exercise conducted by the CEBS had a slightly lower common equity Tier 1 ratio under Basel III definitions (4.9%) and a common equity ratio of 10.7% under Basel II. Echoing the BIS impact study, 157 mid-sized banks suffer a less pronounced decline in their ratio under the new rules, from 11.1% to 7.1%. The tripling of the regulatory minimum from 2% to 7% understates the change in the requirement expressed as an amount of capital. Assuming constant balance sheets, the latter would actually be multiplied by

Major banks' capital requirements

30 25
euros bn (CET1: Common Equity Tier One) 28

CET1 -> 4.5% CET1 -> 7%

20 15 10 5 0 Union EU (21 countries) pays) Europenne (21
Chart 6 Net income 2009 Rsultat net 2009

25 20

12 9 8

World (19 countries) Monde (19 pays) Sources: BCBS, CEBS

In the light of these figures, announcements from several banks that they will be compliant with the new requirements as early as 2013 or 2014 by combining mitigation effects and retained earnings, and without needing to raise fresh capital, can be somewhat surprising. They give the impression that the new solvency rules do not represent much of a constraint.

May-June 2011

Conjoncture

Writedowns & losses/total assets (%)

The catch is that the time horizon implied in these announcements is considerably shorter than that used in the quantitative impact studies. And because of the Basel Committees timetable, deductions affecting capital ratios still have an only limited effect on the former but are taken fully into account in the latter (Table 1). By simulating the application of the new solvency ratio to 62 international banks, Otcker-Robe and Pazarbasioglu (2010) showed that all these institutions would indeed meet the new requirements in 2013 and 2014 but more and more would fail to do so as and when the regulatory minima and deductions applicable to common equity Tier 1 capital are increased between 2015 and 2019, even taking retained earnings into account. Assuming an average dividend payout of 40% of net profits between 2013 and 2019, 16% of these banks (10/62) could fall short of the 7% capital ratio during the period. Were profits to be zero or fully paid out, that proportion would rise to 77% (48/62) by 2019. Leverage On top of the solvency ratio, a leverage ratio between capital and a denominator made up of balance sheet and off-balance sheet items could be integrated into the first pillar on 1 January 2018.
Capital adequacy ratio Tier 1 (1) vs. writedowns & losses/total assets (2)
Writedowns & losses/total assets (%)

that the regulators themselves do not assess certain risks properly. Hence the losses suffered by some US and Swiss banks, among others, that had Basel II weighted solvency ratios that were well above not only the minimum requirements but also the solvency ratios of their international competitors, most of whom weathered the crisis rather better (Chart 7). But although the range of losses and write-offs suffered during the crisis appears to have been wider with initially high leverage (i.e. low capital to asset ratios, Chart 8), its limitation would unjustly penalise the banking systems with the least risky portfolios (France and Japan).
Common equity/assets (leverage ratio) (1) vs. writedowns & losses/assets (2)
2.5 2 1.5 1 0.5 0 1 1.5 2 2.5 3 3.5 4 4.5 Common equity/total assets (%) 5 5.5 6 Switzerland
(1) 2006-08 end-of-year data, 2007-08 for Japan Tier 1 ratio (2) 2007- mid 2009

Germany Belgium United Kingdom Canada Australia France Japan Ireland Norway Spain Italy

Chart 8

Sources: Bloomberg, Datastream, Worldscope, OECD

2.5 2 United Kingdom 1.5 Australia 1 Italy 0.5 0 6 7 8 9 Tier 1 ratio Ireland Spain France Canada Germany

Switzerland

Belgium

Norway Japan

(1) 2006-08 end-of-year data, 2007-08 for Japan Tier 1 ratio (2) 2007- mid 2009

10

11

12

Chart 7

Sources: Bloomberg, Datastream, Worldscope, OECD

The main theoretical justification for the leverage ratio lies in the fact that risk-based ratios cannot completely prevent the undervaluation of certain risks in the denominator. According to Blum (2008), the information advantage that banks have over the regulator can lead them to understate their regulatory risks in order to save on capital. Based on this somewhat forthright assertion, a regulatory leverage ratio could reduce banking risk. Another possibility is

The considerable extension of the scope of risks taken into account under Basel 2.5 and Basel III removes much of the theoretical justification of a leverage ratio in Basel II conditions. Such a ratio could also weigh most on banks with large balance sheets and limited regulatory risks, i.e. retail banks. Banks constrained by a leverage ratio would have an incentive to reduce the size of their balance sheets via the sale or non-renewal of the assets with the lowest weights (the least risky in Basel terms), which would result in an increase in average regulatory risk. Frenkel and Rudolf (2010) emphasise that a leverage constraint could encourage the transfer of assets on banks balance sheets to the market or unregulated sector, limit hedging via derivatives and reduce financing. Note also that a leverage ratio would affect a higher proportion of European banks, which have less risky portfolios in Basel terms (Chart 9). Considering that a rigorous study of these knock-on effects is indispensable, the European Commissioner for the Internal Market told the European Parliaments Committee on Economic and Monetary Affairs in May that he had no intention at this stage to give the

May-June 2011

Conjoncture

A more restrictive leverage ratio for major and/or European banks

Proportion of banks contrained by a 3% ratio
70 60 50 40 30 20 10 0
2.4 2.6 2.8 3 3.2 3.4 3.6 3.8 4

High-quality liquid assets (the numerator) HQLAs are broken down into two categories. Level 1 assets comprise cash, reserves held at the central bank and government bonds or similar rated at least AA13. Level 2 assets, which may not represent more than 40% of HQLAs after a 15% haircut, comprise government bonds or similar rated between A- and A+14, corporate and covered bonds with an external rating of at least AA- or an internal rating at least as high, with the exclusion of securities issued by financial institutions or, in the case of covered bonds, by the bank and its subsidiaries. Net cash outflows (the denominator)

Major EU banks (21 countries)

Leverage ratio

Major banks, rest of world (20 countries)

Other EU banks (21 countries)

Other banks, rest of world (19 countries)

Common equity/assets
Chart 9 Sources: BCBS, CEBS

leverage ratio proposed in the draft CRD 4 any legal force. CRD 4 is to be submitted before the end of July. International liquidity standards: unknown territory In contrast with international bank solvency standards, which appeared with Basel I in 1988, and following the failure of negotiations throughout the 1980s, bank liquidity standards have so far escaped international harmonisation. Where they exist, national prudential rules on liquidity differ greatly from country to country. They are qualitative in nature in the USA, Spain, Italy and Japan and more quantitative in Germany, France and the UK, for example. Arrangements in Italy and France consist of limiting maturity mismatches, and in Germany and the Netherlands this is combined with a minimum liquid asset requirement. The new Basel standards provide for a short-term and a long-term liquidity ratio. The BIS and CEBS quantitative impact studies published in December 2010 suggest that the new liquidity coverage ratio (LCR) and the net stable funding ratio (NSFR) will be far more exacting than existing national requirements, irrespective of the country. The LCR: a short-term liquidity ratio The liquidity coverage ratio is aimed at ensuring that banks have enough high-quality liquid assets (HQLAs) to cover 30 days net outflows following a short-term liquidity crisis, and on the basis of a cashflow scenario defined by the regulator. HQLAs will have to amount to at least 100% of net outflows.

Cash outflows are made up of runs on deposits and the non-renewal of other short-term resources at rates defined by the regulator and expressed as a percentage of outstandings. Inflows correspond to contractual repayments and interest on performing assets over the next 30 days. In order to take account of loan renewals, the Basel Committee recommends a 50% reduction on contractual inflows relating to loans extended to retail banking clients and non-financial companies. The reduction to be applied to inflows related to reverse repos is all the bigger for the fact that the quality of the securities on which such operations are based and the probability of renewal are high. It is set at 100%, 85% and zero, respectively, for level 1 assets, level 2 assets and other assets falling into neither category. Lastly, the inflow amount used in the calculation is limited to 75% of total outflows calculated according to the regulatory parameters (theoretical run or nonrenewal rates). In other words, banks will have to maintain a proportion of liquid assets representing 25% of outflows as simulated during crisis periods. European banks trillion-euro liquid asset deficit One of the main findings of the BIS-CEBS impact studies is that the major banks are in a particularly tight corner as far as the LCR is concerned (Table 2). According to the BIS impact study, the average LCR for group 1 banks at end-2009 was 83%, while that for group 2 banks was 98%. The CEBS impact study came up with 67% and 87%, respectively. We also note that only 43% of the 166 banks of all sizes that participated in the BIS study were compliant with the new standard, i.e. a LCR higher than 100%. The CEBS study did not have any figure for comparison, but some analysts

May-June 2011

Conjoncture

The results of quantitative impact studies on the new liquidity requirements

BCBS QIS (World) % of 30-day outflows Group 1 Number of banks (QIS) Number of banks (LCR) Total outflows (%) Inflows excl. 75% cap Inflows with cap (%) Cap effect (%) Liquidity Coverage Ratio (LCR) (%) LCR deficit LCR ( bn) % banks with LCR >= 100% Number of banks (LCR) Net Stable Funding Ratio (NSFR) (%) NSRF deficit ( bn) % banks with NSFR >= 100% % banks with NSFR >= 85% Table 2 93 2,890 43 67 Sources : Comprehensive Impact Studies, Basel Committee on Banking Supervision (BCBS), Committee of European Banking Supervisors (CEBS), December 2010 100 29.2 22.2 7.0 83 1,730 46 166 103 91 1,800 94 2 1,090 9 94 169 100 55.9 40.5 15.4 98 -26.7 -18.3 Group 2 Group 2 169 50 48 100 33.4 28.8 4.6 67 1,000 196 182 100 56.6 38.1 18.5 87 -23.2 -9.3 -4,2 -6,6 2,4 16 730 -0,7 2,4 -3,1 11 Group 1 Group 1 Group 2 Group 2 CEBS QIS (European Union) Group 1 Group 1 Group 2 BCBS - CEBS

estimates suggest that between a third and two-thirds of major European banks are already LCR compliant. As of end-2009, the additional liquid assets needed for all banks to meet the 100% requirement amounted to 1,730 billion for banks in the BIS universe15 and 1,000 billion for EU banks. The LCR from a theoretical viewpoint To some extent at least, the LCR may be viewed as a mandatory reserve requirement, where commercial banks have to constitute reserves at the central bank in proportion to their deposits, and on which academic literature abounds. The first effect of an increase in the reserve requirement is an arithmetic decline in the credit (or money) multiplier. While this multiplier is not directly affected by an increase in the proportion of liquid assets in the LCR numerator that correspond to counterparts of money supply (lending to the economy, and notably government bonds), the components of money supply

will be distorted in favour of the latter, and to the detriment of credit to the private sector. The diversity of assets eligible for the LCR numerator contrasts with application of a simple mandatory reserve requirement and makes the ex ante assessment of the ratios impact on the monetary aggregates singularly complicated. After all, the latter are sensitive to the breakdown of liquid assets and the asset portfolios interest-rate elasticity. The impact probably falls within two extremes. The minimum impact on money supply in its narrowest sense (i.e. restricted to components that can be used immediately as a means of payment) would be in the situation where the increase in liquid assets takes the exclusive form of purchases of non-government securities, although the probability of this hypothesis is limited by the fact that level 2 assets cannot exceed 40% of the HQLA total. An intermediate situation would stem from the purchase of government securities, and the maximum impact would stem from increases in reserves that commercial banks hold at the central

May-June 2011

Conjoncture

10

bank. In the latter case, the increase in reserves would directly influence demand for the monetary base and thereby reduce overall liquidity in the economy. The second effect of an increase in the reserve requirement, which is undoubtedly even more pertinent for the LCR, is connected with the higher opportunity cost associated with the ownership of assets yielding nothing or very little. This phenomenon puts banks at a competitive disadvantage vis--vis market financing or the unregulated financial sector. Empirical studies demonstrate the positive effects of mandatory reserve requirements on the cost of credit (Chart 10a) and their negative effects on money supply (Chart 10b) and domestic credit outstandings (Chart 10c). Although credit institutions generally remain subject to central bank reserve requirements (2% of deposits outstanding amounts maturing in less than two years in the euro zone, for example), the negative externalities with which they are associated have led to their gradual abandonment as a monetary policy tool. The authorities now prefer open market operations instead. Even so, and in the current definition arising from the Basel Committees December 2010 proposals, the LCR may force banks to tie up part of their assets in the form of lowyield instruments, and to a far greater extent than mandatory reserve requirements do. The amount of mandatory reserves required from euro zone credit institutions has been remarkably stable over time and was precisely 208.3 billion for the reserve maintenance period ending on 10 May, for example. This is only one fifth of the 1 trillion shortfall in liquid assets that EU banks would have had under Basel III at end-2009, according to the CEBS quantitative impact study. We should emphasise that the nature of eligible assets debt securities would tend to limit the unit opportunity cost relative to that of a mandatory reserve requirement. The outcome depends very largely on whether banks seek to respect the LCR by amassing the liquidity required without altering the profile of their balance sheets (apart from substituting for the liquid assets). In an admittedly very different context, Wagner (2004) showed that an increase in the liquidity of one part of a portfolio can also encourage banks to make up for the resulting opportunity cost by reducing the liquidity of the other part in the hope of a higher yield. Banks may reason in terms of average portfolio liquidity, in other words. The most likely outcome is that banks reduce the net cash outflows forming the LCRs denominator relative to the cushion of liquid assets (the numerator) by extending the maturity of liabilities and/or shortening the maturity of assets.
May-June 2011

Reserves requirements and interest margins

14
Reserve requirement ratio

2006, %, logarithmic scale BR CN US SA SG KR CZ XM JP 12 2 32 HK TH ID PH VE TR CL PE HU ZA PL MX 7 92 UK 12 112 IN AR

12 12 10
4 8 MY IL

CO

6
2 4

2 0

3 4 52 72 Net interest margin (3)

Chart 10a

Sources: Mohanty & Turner (2008), IMF, Bankscope, BIS

Reserve requirements and money supply

14
Reserve requirement ratio

12 12 10
4 8

6
2 4

2006, %, logarithmic scale BR AR PH VE ID CL PE TR US IN CO HU SA ZA CZ PL MX 12 35 32 55 52 XM

CN IL MY KR SG TH

2 0

JP UK 150 92 250 112

HK

90 72 M2/GDP

Chart 10b

Sources: Mohanty & Turner (2008), IMF, Bankscope, BIS

Reserves requirements and credit

2006, %, logarithmic scale BR AR VE 12 12 PH CN ID 10 TR CL IN HU PE 4 8 MY CO US SG SA IL KR 6 CZ ZA 2 4 PL TH XM 2 MX JP HK UK

14

Reserve requirement ratio

12
Chart 10c

12 32

20 35 55 90 52 72 92 112 Domestic bank credit (2)/GDP

150 132

Sources: Mohanty & Turner (2008), IMF, Bankscope, BIS

AR=Argentina; BR=Brazil; CL=Chile; CN=China; CO=Colombia; CZ=Czech Republic; HK=Hong Kong; HU=Hungary; ID=Indonesia; IL=Israel; IN=India; JP=Japan; KR=Korea; MX=Mexico; MY=Malaysia; PE=Peru; PH=Philippines; PL=Poland; SA=Saudi Arabia; SG=Singapore; TH=Thailand ; TR=Turkey; UK=United Kingdom; US=United States; VE=Venezuela; XM=euro area; ZA=South Africa. (1) Midpoint in range for Agentina, Brazil, Chile, the euro area, Indonesia, Israel, Japan, Korea, Poland, Saudi Arabia and the United States. (2) Domestic bank credit to private sector. (3) Net interest revenue/average earning assets.

Conjoncture

11

The NSFR, a medium- to long-term liquidity ratio The idea of the net stable funding ratio (NSFR) is to protect banks from the consequences of a liquidity crisis by prohibiting the funding of long-term assets (loans or securities) maturing in more than a year with resources that mature within a shorter time and may not be renewed. The NSFR requires that weighted assets maturing in more than a year and certain off-balance sheet commitments are 100% covered by long-term, stable funding. The regulators intentions are certainly good, but the NSFR in its present form would dangerously limit banks maturity transformation. The deficits in additional long or stable funding resources indicated in the impact studies are so large that they make it more likely that banks will limit assets maturing in more than a year than raise the funding they need to keep assets as they are. Available stable resources (the numerator) Stable resources are composed of various types of liabilities. Financial resources with a residual maturity of more than a year equity capital, debt, term deposits are taken into account for their full balance sheet amounts, but resources maturing in less than a year are weighted according to their supposed stability: 50% when they emanate from non-financial corporates, the public sector or central banks, 80% for less stable retail deposits16 and 90% for stable retail deposits17. Required stable funding (the denominator) The stable funding requirement is based on the sum of balance sheet and some off-balance sheet items. The weights for balance sheet items reflect their theoretical one-year liquidity. Most assets with a residual maturity of one year are weighted at 100%, unless they are extremely liquid. The latter would include corporate bonds rated better than AA- (20%), non-financial companies listed shares (50%) and bonds eligible for refinancing operations not issued by financial institutions (50%)18. Off-balance sheet undrawn credit or swingline facilities have a stable funding requirement of 5% of their nominal amount. The NSFR and monetary equilibrium In terms of the economys monetary equilibrium, the NSFR could lead to a faster increase in the stable resources (savings accounts, bonds and net equity) that are subtracted from traditional counterparties (external, debt on the economy) in money supply calculations. We
May-June 2011

may consider a simple dynamic process to show how such an adjustment could eventually produce much larger requirements than the impact studies static approach suggests. If we assume that other balance sheet items are unchanged, the restrictive impact of the NSFR on M3 money supply is amplified. Each bank will seek individually to comply with the NSFR while considering, ex ante, that the balance sheet items that are not subject to any adjustment are exogenous. But from a macroeconomic point of view, money supply equals lending to the economy minus financial and monetary institutions non-monetary resources, assuming a closed economy 19 . For the banking system as a whole and in the medium term, and without the creation of money (because the creation of money by commercial banks would worsen their NSFRs 20 ), bond issues from banks would eventually lead to a contraction in other balance sheet resources, monetary or otherwise. Suppose, for example, that the banking system issues bonds worth 100 billion in period 1 to cover an initial stable resources deficit. Without the creation of money, that will mean eventually a reduction in other balance sheet resources by the same amount, and the size of the banking systems balance sheet will not change. But if these resources are attributed a 50% weight in the NSFR numerator (e.g. a mix of short term market resources and more or less stable deposits), the banks will gain 100 billion in resources but lose 50 billion of that in Basel terms (100 billion x 0.5), and their stable resources will increase by only 50 billion, far from the 100 billion desired. Continuing with this reasoning ad infinitum is hardly realistic, as we ignore the knock-on effects of an increase in the average cost of resources and of intermediation rates on loan volumes, but it does at least highlight the iterative nature of the process. If banks then issue 50 billion in period 2 (100 billion - 50 billion), they increase their stable resources by only 25 billion in Basel terms. So they have to issue 25 billion in period 3, and so on. Mathematically, the additional amount to be issued each period converges on zero and banks have to raise a total n times the initial stable resources deficit in order to increase the NSFR denominator by the required amount. The multiple n actually equals 1/(1- ), where is the average weight in the numerator of the resources for which the newly created stable resources are substituted. In the above example, where the average weight of resources is 50%, this multiple is 2.
Conjoncture 12

In the real world, the need for additional stable resources would probably decline by more in each period because of the contraction in the NSFR denominator. More expensive bank resources (liabilities are distorted in favour of longer maturities) and the need to switch some resources to uses yielding little or nothing would put pressure on margins. Banks would seek to compensate for that by increasing yields on their assets. More expensive financing would have a negative impact on credit volumes and the NSFR denominator. The evaluation of these two effects and their interactions is no easy matter. For the sake of argument, we assume that additional requirements for long or stable financing correspond to the stable resources deficit highlighted in the CEBS quantitative impact study. European banks: a 1.73 trillion long-term stable funding deficit at end-2009 Basel Committee and CEBS impact studies claim a long-term stable funding deficit of 2,890 billion for the 166 banks providing enough information to calculate their NSFRs at end-2009, or 1,800 billion for the European banks participating in the CEBS study (Table 2). Changes in NSFRs for banks in the main economic areas between 2005 and 2009 calculated by IMF researchers (IMF, 2011a) show that European banks are in a relatively unfavourable situation, while American and Asian banks are relatively comfortable (Chart 11). As with the LCR, quantitative impact studies have shown that mid-sized banks are better positioned than large banks in NSFR terms, with a bigger size-related differential in the BIS universe (103% vs. 93%) than in the EU (94% vs. 91%). The same calculations carried out by type of bank show that NSFRs declined more during the crisis among investment and commercial banks than they did among universal banks (Chart 12). NSFRs picked up again in 2009, reflecting the normalisation of funding structures. Investment banks tend to amplify these trends because they are more dependent on market financing and have more flexible business profiles. A very strict standard The two new liquidity ratios follow a largely microprudential line of reasoning 21 . It is therefore tempting to assess their ability to prevent or predict banking failures, just as the IMF has recently done for the NSFR22.

Average net stable funding ratio by region

115 %, NSFR 110 105 100 95 90 85 80 2005
Chart 11

North America

Asia

All Europe

2006

2007

2008

2009

Sources: Bankscope, IMF staff calculations

Average net stable funding ratio by business model

110 %, NSFR 105 100 95 90 85 80 2005
Chart 12

All Commercial

Universal

Investment

2006

2007

2008

2009

Sources: Bankscope, IMF staff calculations

IMF economists simulated NSFRs for a sample of around 60 banks (34 European, 14 American and 15 Asian) at end-2006. According to their calculations, the banks that failed during the financial crisis had NSFRs distributed fairly evenly between 80% and 107%, and only seven of them were at under 100% (including one with a level well below the bottom of this range, Chart 13). As inadequate data made several assumptions necessary, and the sample of banks was very small
Net stable funding ratio estimates for an expanded set including failed banks, 2006
2.5 %, NSFR 2 1.5 1 0.5 0
Chart 13

*failed banks are depicted in white

Banks*
Sources: Bankscope, IMF staff calculations

May-June 2011

Conjoncture

13

(particularly for NSFRs below 80%), the significance of these findings may not be that great. But they suggest nonetheless that a 10-point decline in the minimum requirement to 90% would make very little difference to the NSFRs effectiveness as a microprudential tool (Chart 14). Yet it would reduce its economic impact significantly.
Probability of bank failure and the NSFR
% of banks defaulting during the crisis below NFSR threshold (x axis)

100% 80% 60% 40% 20% 0%

43% Chart 14

% of banks with NSFR ratio below threshold (x axis)

78%

83%

87%

91%

94% 100% 103% 108% 116% 136% NSFR ratio

Sources: Bankscope, IMF calculations, BNP Paribas

Economic impacts: from the damaging to the expected

Put in the context of the euro zones financing circuit, the shortfall in EU banks long-term resources appears all the more daunting. The respective roles of the various categories of financial intermediary also indicates that apart from the matter of bank liabilities, NSFR compliance implies a distortion in the structure of savings among non-financial agents and/or the rest of the world. Extending the maturity of banks liabilities is inseparable from longer savings The impact of the NSFR on the euro zones financing circuit may be examined with reference to the financial accounts published by the ECB. The top part of the diagram describes financial intermediation in its broad sense, with insurance companies, pension funds, other financial intermediaries such as general category mutual funds and monetary and financial institutions (MFIs, a category that includes money market funds, credit institutions and central banks). Financial intermediaries collect around 53,000 billion in resources from resident non-financial agents, the rest of the world and other financial intermediaries.
May-June 2011

The bottom part of the diagram describes financial savings and the funding of non-financial agents and the rest of the world, whether intermediated (assets appearing as liabilities on financial intermediaries books) or not (banknotes, direct holdings of securities not issued by financial intermediaries). Each amount is then broken down by maturity (short-term and long-term). The application of a proportional rule based on the respective sizes of the banking systems falling within the scope of the CEBS impact study and euro zone offers a reasonable approximation, under acceptable hypotheses, of euro zone banks additional long-term stable funding requirements. On the basis of this simple calculation, the euro zones banking system needs another 1,400 billion or so of these resources, the equivalent of 16% of MFI long-term financial resources (long-term debt securities and equities) or 22% of outstanding long-term debt securities. The term structure of financial assets and liabilities tells us how each institutional category transforms maturities. Among financial agents, the only category that actually extends maturities is MFIs, whose longterm financial resources (8,583 billion as of 31 December 2010) are much smaller than their long-term uses (19,320 billion). In contrast, insurers, pension funds and other financial intermediaries tend to slightly reduce maturities. The upshot of all this is that an extension of the maturity of MFI financial liabilities necessarily implies a distortion in the structure of financial saving among nonfinancial agents in the long term, as non-MFI financial intermediaries seem incapable to extending the maturity of the short-term resources collected from non-financial agents and reinvesting them in bank balance sheets. All other things being equal, it would take a substantial increase in long-term interest rates to produce such a change in the term structure of financial savings, as that would encourage non-financial agents to forego liquidity. Economic impact studies: similar in principle, divergence over methodology and form Since the summer of 2010, the Institute of International Finance (IIF), the BIS and most recently the OECD have published reports on Basel IIIs economic impact. At first sight their conclusions appear to differ. But they more or less agree on the transmission mechanisms involved which is the key part even though their methodologies diverge considerably.
Conjoncture 14

Banks, and banks only (in Euro zone) approximately create EUR 10,700 bn of long-term funding ressource
6 670
ST 1 019 LT 5 651 ST 155 LT 6 680 ST 1 633

Insurances and Funds of pension

Savings
6 835

Financing

Euro area, Outstanding Amounts EUR bn, 2010 Q4

14 905 OFIs (1)
ST 3 828 LT 11 077 LT 12 771 14 404

ST 18 059
MFIs (2)
ST 23 046 LT 8 583

LT 36 048
32 532
ST 13 212 LT 19 320

31 629

May-June 2011

54 107

52 868

1 672 5 739
+56 317
CT 15 330 LT 40 987

ST 3 424 LT 13 874

17 298 26 655 6 627

18 864

ST 6 676

Of which deposits from non-financial ST 24 834 agents 8 839 LT 28 034 (RSF conversion factor Savings comprised between intermediation 0.5 and 1)
LT 12 188 ST 360 LT 6 267

Conjoncture

(3) NFC ST 3 596 -9 357

Households +12 237

LT 23 059

-14 537

15

428
3 777 -57 134
CT 8 154 LT 48 980

ST 832

GG (4) 8 957

LT 2 945

-5 180

ST 1 316

LT 7 641

3 489 (o.w. 1000 from Non-banks) +13 720 16 378 ST 4 398 Rest of LT 11 980 the world ST 2 882 + 1 483 14 895 LT 12 013

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

Other financial intermediaries Monetary and financial institutions Non financial corporations General government

Sources: ECB, BNP PARIBAS

At the BIS, research on the macroeconomic impact of Basel III has been conducted by two separate working parties. The Macroeconomic Assessment Group (MAG), which brings together experts from 15 countries central banks and regulators as well as from international institutions like the IMF, BIS and FSB, was asked to evaluate the costs associated with the new prudential framework during the transition period. On the other hand, the Long-term Economic Impact Working Group was asked to examine the long-term benefits of prudential reform, assumed to be a smaller probability of banking crises and associated GDP losses as well as more limited fluctuations in GDP outside crisis periods. Any assessment has to start with a calculation of the increase in the average cost in banking resources implied in increases in capital and the extension of the maturity of bank liabilities stemming from Basel III. This step implicitly raises the question of whether the Modigliani-Miller theorem (cf. infra) can be applied to banks. The following steps consist of quantifying the impact of more expensive bank resources on bank loan rates and lending volumes, either successively (the IIF) or simultaneously (MAG). This impact is then translated into economic growth. While this result is what the IIF and OECD are aiming at, it is another point of departure for the BIS, which considers the effects of (assumed) greater financial stability on growth. In contrast, the methodological approaches used to identify these macroeconomic effects have very little in common. The IIF has simulated the incidence of the new rules on aggregate bank balance sheets for each major economic area, adjusting for the different components of balance sheets and income statements, and used elasticities generated by its own econometric estimates. The BIS largely dispensed with this intermediate stage regarding changes in bank balance sheet structure, relating the increase in solvency ratios directly to the cost of credit via median elasticities derived from 97 economic models concerning different geographical areas. For each intermediate step, the BIS experts used the median elasticity obtained from an array of empirical models testing a comparable relationship. The macroeconomic impact study is based on transmission mechanisms similar to those used by the BIS. Tougher capital requirements and new liquidity rules make banking resources more expensive and lead to an increase in the cost of bank financing then a decline in its volume. The IIF also assumes that banks seek to compensate for the opportunity cost linked to
May-June 2011

holdings of liquid assets for LCR purposes by increasing their lending spreads. And that results in a slide to the bottom end of the economys growth track. While the IIF study turns on the impact of the new rules on the banking systems balance sheet, the BIS looks at these effects via a series of models that generates a wide variety of results. It then retains the median of these effects. Can Modigliani-Miller be applied to banks? How much will the distortion of banking assets inherent in Basel III affect the average cost of banking resources? This question immediately raises the issue of whether the Modigliani-Miller theorem applies to banks at all. According to Modigliani and Miller (1963), a companys economic value and the weighted average cost of its capital are independent of its financing method (equity or debt). In other words, there is no optimal financial structure that can minimise the weighted average cost of financial resources. If we assume that the company resorts to lower gearing, the savings on the cost of capital resulting from the substitution of a more costly resource (shares) for a less costly resource (debt) will be exactly offset by lower unit costs of debt and equity, whose yield requirements decline with the leverage. This theory, which appears to have offered at least some inspiration to regulators in the context of capital requirements, applies only under the assumption of perfect markets, with an absence of taxes, transaction costs or default. The authors themselves acknowledged these shortcomings in a corrected version of the theorem (Modigliani and Miller, 1963). At the Bank of England, Miles et al (2011) have developed the idea of tax neutrality, arguing that additional taxation on banks does not reduce social wellbeing in the final analysis because it helps finance the governments budget. This argument takes no account of the distortions created by taxation, however, or of the impact of more expensive banking resources on activity. In any case, Merton H. Miller (1995) has himself explained that the theory applies only to financial structures ex ante. It states that any given company, if it had a different balance sheet structure, would bear an identical average cost of resources. That does not mean that a change to an initial financial structure would be neutral, especially if the company has to raise fresh capital from the market.
Conjoncture 16

In a brief article that specifically addressed the applicability of the Modigliani-Miller theorem to banks, Miller (1995) did not come down on one side or the other but made several helpful points. The Nobel prizewinner explained that banking activities partly avoided market mechanisms because of distortions introduced by deposits insurance and prudential regulation. If the cost of deposits is lower than it would be without a guarantee, banks will seek to minimise their capital as far as the regulator or the market will accept. This is a corner solution of the type described in the corrected version of ModiglianiMiller (1963). In a more general vein, and acknowledging a debt to the Austrian economist and philosopher Friedrich Hayek (1899-1992), Miller emphasised that a uniform rule applied to all banks, excessively severe for some and not enough for others, could not give rise to monitoring as efficient as private contracts would, and that it would always be a source of inefficiency and friction. The same goes (although to a lesser extent) for more sophisticated rules that can never evaluate a banks intrinsic risk as well as agents in the ideal Modigliani-Miller efficient market framework. But this is also one of the arguments which the crisis appears to have weakened the most. Lastly, Miller observed that business and risk profiles were sometimes very diverse and difficult to explain from one bank to another. This made it impossible to approve or refute the neutral financial structure theorem outright. For their part, the authors of the macroeconomic impact studies are highly circumspect on the issue. The IIF makes the explicit assumption that the Modigliani-Miller theorem was to some extent true. More specifically, it argues that any one-point increase (reduction) in differential between the actual solvency ratio and the regulatory target reduces (increases) the yield required by the market by half a percentage point. The MAG cites several empirical studies that point to a positive impact of higher solvency ratios on the cost of credit, implying that the financial structure is not neutral. Even so, the working partys experts agree with Kashyap (2010) 23 that the impact of capital and liquidity requirements is linked to effects not incorporated into the idealised Modigliani-Miller framework, notably the tax advantage of debt and the premium that banks have to pay to investors to ensure that they are prepared to hold longer-dated and less liquid debt.
May-June 2011

Basel III still open to interpretation The impact of higher solvency ratios may well be an indispensable part of any macroeconomic impact study, but the actual adjustments in those ratios and the amounts of capital involved are still moot points. In its June 2010 study, the IIF assumed that the regulatory Tier 1 ratio would rise from 4% to 6% in 2012 and buffers would be added at national authorities discretion. The latter could be interpreted as the safety margin that banks feel they need or as the gap between the regulatory minima and implicit market requirements. The IIF assessed the additional capital to be raised in response to these measures at 210 billion (including redefinition effects) for euro zone banks between 2009 and 2014. In its impact study, the MAG considered that every point on the capital ratio means an extra 114 billion in additional capital requirements. The 1.3-point gap between EU banks ratios under the 2019 Basel III rules as of end-2009 (5.7%, according to the quantitative impact studies) and the regulatory minimum (7%) is therefore the equivalent of 148 billion in capital at constant balance sheets. For its part, the OCDE believes that euro zone banks will have to raise their common equity ratios by 3.8 points (3.7 points for G3 banks) but does not venture a capital amount. In contrast with the new solvency rules, the coming liquidity requirements are either ignored (OECD) or taken into account to a lesser (MAG) or greater (IIF) extent. The June 2010 IIF research is based on the notion that the LCR and NSFR will come into force in 2012, and that will now not be the case. But for what it is worth, the NSFR was assumed to be virtually respected on average in the USA but not in the euro zone, where the amounts of longdated resources that need to be raised run up against the bond markets capacity to absorb supply. After a peak at 75.5% in 2013, the NSFR will gradually decline to 67% in 2020, identical to its simulated 2011 level. In the BIS research, the proxy used to simulate the new liquidity requirements does not always mean the same thing. In the first MAG macroeconomic impact study of August 2010 (2010a), 25- and 50-point increases in the ratio of liquid assets to total bank assets defined the range of equivalent effects of combined LCR and NSFR application. But later work at the BIS (cf. King, 2010) then suggested that that range referred to the effects of the NSFR alone, adjusted in accordance with different interaction hypotheses with the solvency ratio. It seems from these latest developments that BIS impact studies have ignored the LCRs macroeconomic impact.

Conjoncture

17

While the impact of the solvency rules was reassessed between the August 2010 report and the final report in December 2010, the MAG (2010b) left its analysis of the liquidity ratios impact unchanged, even though most of the banks concern is focused in this area. It cites the existence of observation periods in defence of this decision (the LCR will be introduced in 2015 and the NSFR in 2018). From banking margins to credit volumes The impacts on interest margin and credit volumes depend on hypotheses that differ from one study to another and that make comparisons delicate. That said, the MAG distinction between the impact of tougher solvency rules and the introduction of liquidity ratios permits comparisons with the OECD study, which is dedicated exclusively to the impact of the Basel III solvency ratios. The MAG results cannot be directly compared with IIF results, as the latter attempt to capture the overall impact of the prudential reform without any distinction between liquidity and solvency. The MAG starts by examining the impact of a onepoint rise in solvency ratios on loan volumes and on spreads. The results published in the final report are the unweighted median of 38 different models in the case of loan volumes and that from 53 models for loan spreads. The impacts obtained are a 1.4% contraction in loan outstandings after around nine years24 relative to where they would be without the reform, and assuming that the new requirements are spread over eight years. Similarly, loan spreads are 15.5 basis points higher to the same horizon. While the geographical match is not exact, the MAG results are consistent with those obtained by the OECD. The latter found that the increase in bank spreads following a 1 point increase in solvency ratios would be an average 16.1 basis points25, once weighted by G3 countries GDP26. The MAG analysis bears on its entire universe, and no geographical breakdown is possible. The conclusions drawn from Fed and ECB models are supplied for illustrative purposes only. According to the ECB, a one-point rise in solvency ratios would raise bank loan spreads by 28 basis points. The ECBs model is based on the 20 largest banking groups in the euro zone, and this impact is larger than that obtained by Slovik and Cournde (2011) or by Cosimano and Hakura (2011) in the three euro zone countries they cover27. The unitary median impact given by these models was then put in the context of the quantitative impact studies in the final report in December 2010. As we have seen, the MAG experts
May-June 2011

believe that starting with an average initial level of 5.7% in December 2009 under the new rules, the world banking system will have to raise its common equity Tier 1 ratio by 1.3 point to get to the 7% requirement. Assuming that the impacts obtained are linear, loan outstandings would decline by 1.82% and loan spreads would rise by 20.15 basis points as a result of the Basel III solvency rules alone. At the same time, the MAG summarised the liquidity requirements as a 25% hike in the ratio of liquid to total assets 28 . The impact is evaluated initially in terms of credit volumes (3.2% lower after eight years) and then in spreads (up 14.1 basis points after eight years, assuming application spread over four years), retaining the median scores of numerous empirical models. A 50 basis-point rise in the ratio of liquid to total assets (the NSFR without interaction) would result in a 25 basis-point increase in spreads and a durable 0.15% loss of GDP. Consideration of synergies29 between the capital requirement and the NSFR could limit the equivalent increase in the ratio of liquid to banking assets to 25 basis points, the rise in credit spread to 14 basis points and the consequent loss of GDP to 0.08%. Although the various impacts obtained by the BIS for each type of ratio are not intended to be aggregated, their sum does at least offer an approximation that can be compared with the cumulative impact obtained by the IIF. If we refer back to the ECB analysis, a 2-point increase in the solvency ratios, a 25% rise in the ratio of liquid to total assets and the introduction of the NSFR would widen bank lending spreads by 56 30 , 15 and between 57 and 71 basis points, respectively. The midpoint of the resulting total impact range (135 basis points) is remarkably close to that obtained by the IIF on real bank lending rates (137 basis points) five years after the beginning of the reforms introduction (i.e. 2015 in the provisional June 2010 report). Impacts on GDP: major discrepancies While the impacts of increases in solvency ratios all on their own and of the whole reform in under standardised hypotheses on bank lending spreads may be more or less reconciled with the various macroeconomic impact studies (Table 3), estimations of the impact of Basel III on GDP differ greatly. The many sources of these discrepancies are significant. The BIS limited itself to evaluating the impacts of the solvency and liquidity parts of Basel III independently of each other. Because of the interactions between these rules, it suggests no overall impact. Even with its
Conjoncture 18

Impact of Basel III on bank interest margins and interest rates in the euro zone
IIF (June 2010) Cause Effect Rise in risk-free rate (a): 0 MAG/BRI (August 2010) ECB model Effect on interest Cause margin 2-point rise in solvency ratios + 56 basis points OECD, Slovik and Cournde (2011) Effect on interest Cause margin 1-point rise in solvency ratios + 14.3 basis points

Cumulative impact to 2015:

- of a 2.3 point rise in the average solvency Interest margin (b): + 25 point rise in liquid asset ratio 101 basis points ratio - of an LCR of 79.8% (vs. 32.9% in the reference scenario) - of an NSFR of 73% (vs. 57.5% in the reference scenario)

+ 15 basis points

Rise in solvency ratios required to 2015 + 18.6 basis points (+ 1.3 point)

GDP deflator (c) : 36 basis points Bank loan rates (a + b + c): +137 basis points

NSFR (1)

Between + 57 basis Rise in solvency ratios required to 2019 points + 54.3 basis points and 71 basis points (+ 3.8 points) Between + 128 basis points and 142 basis points

Sum of effects

(1) The impact of the NSFR is assessed by the ECB but not in the MAG study

Table 3

Sources: BIS, IIF, OECD

obvious drawbacks, the addition of the various estimates has the merit of producing an order of magnitude from the MAG study. Thus with a comparable time horizon (32 vs. 35 quarters), the sum of these impacts suggests a loss of around 5% in loan outstandings (-1.8% and -3.2%), a 34.2 basis-point rise in spreads and a 0.35% loss in GDP. That compares with a cumulative impact of 3.1% for the IIF, nine years after the reforms are introduced. The heterogeneity of the basic hypotheses underlying estimates of capital adjustments and the partial integration or not of the new liquidity rules need not be dwelt on any further. Moreover, the application timetable that the IIF used was unofficial and turns out to be shorter than that finally recommended by the Basel Committee. That may have led to an over-estimate of the economic impact. Above all, the methodologies used to measure the impact of wider lending spreads and/or lower loan volumes on GDP are radically different. While the MAG re-injects the impacts on loan volumes and bank spreads into 97 dynamic stochastic general equilibrium and semi-structural models with varying specifications and where countries are represented inaccurately with respect to their economic weight, then retains the median result, the IIF uses models specific to each

major economic bloc (the USA, euro zone and Japan) but with strictly identical relationships and structures. If that were not enough, the geographical spread of the OECD and IIF studies are the same G3 countries but the 97 models on which the BIS bases its final report cover nine countries which do not fall within the scope of the first two studies31. This means that the IIF universe represents only 63% of the BIS universe. The euro zone taken as a whole and five Member States account for 44 of the 97 models, while the USA accounts for 11 and Japan six. Lastly, the BIS incorporates monetary policy via a Taylor rule. This tends to compensate for the restrictive impact of Basel III on the creation of money and significantly reduces the impact on economic activity. A controversial view of the benefits of greater financial stability Apart from an assessment of the economic costs of the reform during the transition period, attributed to the MAG, the Basel Committee asked a second working party to examine the long-term economic impact of the changes. The Long-term Economic Impact (LEI) Group used a cost-benefit type approach that drew on the MAG results and assessed the benefits of the reform in terms of a smaller mathematical expectation of a loss of

May-June 2011

Conjoncture

19

GDP associated with financial crises. The group started by estimating the decline in the probability of financial crises stemming from Basel III to around 3%. Thus the probability of crisis drops from 4.6% to 1.4% (with solvency requirement), or even to 1.2% (plus NSFR compliance). The product of the fall in the annual probability of crisis and the median long-term cost of a financial crisis (a cumulative 63% of GDP) gives the mathematical expectation of the annual GDP loss avoided. Controversially, the LEI group then assimilates the latter to supplementary points of GDP growth. All in all, for a solvency ratio (common equity that increases from 7% to 10%), the LEI experts put the benefits of the reform at between 1.98% and 2.10% of GDP per year under their baseline hypotheses 32 , depending on whether liquidity requirements are respected or not. At the same time, the same prudential rules are supposed to reduce the annual growth rate by only 0.27% to 0.35%. The net long-term benefit given by these figures is between 1.71% and 1.75%. While we do not doubt that greater financial stability has inherent benefits, the assessment of the contribution of prudential reform to financial stability raises serious methodological issues. In the empirical literature, the most frequently used approach involves logit-type equations 33 that link the frequency of financial crises within a sample of countries over a given period to bank leverage and liquidity ratios. The main criticism of these studies lies in the diversity of crises taken into consideration, and without reference to their initial cause. And we do not believe that tougher prudential controls reduce the probability of exchange rate crises, for example, or crises that start with inappropriate monetary policies and undermine the banking sector. Moreover, that sort of approach completely contradicts the hypothesis that monetary policy will tend to compensate for the depressive impact of prudential regulation, as retained by the MAG in the BIS impact study relating to the transition period. According to Laeven and Valencia (2008, 2010), the cost of a crisis between its beginning and end is often far greater for emerging countries. Their inclusion greatly increases the median GDP loss and therefore tends to exaggerate the benefits associated with a lower probability of crisis for developed countries. The authors of the LEI report also put forward the notion that the average cost of financial crises is reduced by the introduction of appropriate prudential rules. But the fragility of the empirical results34 led them to reject this hypothesis and argue along with most of the others that the cost of a crisis is exogenous.
May-June 2011

Conclusion
In contrast with the Basel I and Basel II recommendations, the Basel III prudential reform is a response to a financial crisis that was even worse than that of 1929. Naturally, its endogenous nature carried the risk of an overreaction from the G20 and the regulators. The results of quantitative impact studies suggest that wide-ranging application of the new rules would produce a more stable banking system. That said, the economic benefits claimed for the BIS reform are open to doubt, and banking system stability does not mean financial system stability when the regulations do not prevent arbitrage between the regulated and unregulated sectors. The main area of concern is the economic cost of the reform. The results of macroeconomic impact studies should be seen as minima, as a great many factors have not been taken into account (bank taxes, additional requirements for systemically important banks). Moreover, the absence of an internationally agreed liquidity requirement rules out any complacency on the reforms impact. According to the figures used by the BIS, the median gap between two crises is 20 years. The observation period could therefore be used to deepen our understanding of the interactions between solvency and liquidity requirements. As the two can be substituted for one another to some extent, less onerous liquidity rules could help preserve the economys financing mechanisms and the flexibility of banking models without jeopardising financial stability.

10 June 2011 laurent.quignon@bnpparibas.com

Conjoncture

20

BIBLIOGRAPHY Atkinson P., Blundell-Wignall A. (2010), Thinking beyond Basel III Necessary solutions for capital and liquidity, OECD Market Trends No. 98 Volume 2010/1 BCBS (2010a), Basel III: A global regulatory framework for more resilient banks and banking systems, December (rev. June 2011) BCBS (2010b), Basel III: International framework for liquidity risk measurement, standard and monitoring, December BCBS (2010c), Results of the comprehensive quantitative impact study, Bank for International Settlements, December BCBS (2011), Revisions to the Basel II market risk framework updated as of 31 December 2010, February BIS (2011), Basel III: Long-term impact on economic performance and fluctuations, Working paper No 338, February Blum J. (2008), Why Basel II may need a leverage ratio restriction, Journal of Banking and Finance, 32, 16991707 CEBS (2010), Results of the comprehensive quantitative impact study, 16 December Cosimano Thomas F., Hakura Dalia S (2011), Bank Behavior in Response to Basel III: A Cross-Country Analysis, IMF Working Paper WP/11/119, May Frenkel M., Rudolf M. (2010), The implications of introducing an additional regulatory constraint on banks business activities in the form of a leverage ratio, mimeo for German Banking Federation IMF (2011a), How to address the systemic part of liquidity risk, Global Financial Stability Report, chapter 2, April Institute of International Finance (2010), Interim Report on the Cumulative Impact on the Global Economy of Proposed Changes in the Banking regulatory Framework, June Kashyap, A, S Hanson and J Stein (2010): An analysis of the impact of substantially heightened capital requirements on large financial institutions, Working paper, May King, M.R. (2010), Mapping capital and liquidity requirements to bank lending spreads , BIS Working Paper, No. 324, November Laeven L. and Valencia F.(2008), Systemic Banking Crises: A New Database, IMF Working Paper 08/224. Laeven L. and Valencia F. (2010) Resolution of Banking Crises: The Good, the Bad, and the Ugly IMF Working Paper 10/146. LEI (2010), Long-term Economic Impact working group (BIS), An assessment of the long-term economic impact of stronger capital and liquidity requirements, August

May-June 2011

Conjoncture

21

MAG (2010a), Macroeconomic Assessment Group established by the FSB and the BCBS, Assessing the macroeconomic impact of the transition to stronger capital and liquidity requirements, Interim Report, August MAG (2010b), Assessing the macroeconomic impact of the transition to stronger capital and liquidity requirements, Final Report, December Miles D., Yang J. and Marcheggiano G. (2011) Optimal Bank Capital, Discussion Paper n31, Bank of England, January Miller M. (1995), Do the M&M propositions apply to banks?, Journal of Banking & Finance 19 (1995) 483-489 Modigliani, F.; Miller, M. (1958). The Cost of Capital, Corporation Finance and the Theory of Investment. American Economic Review 48 (3): 261297 Modigliani, F.; Miller, M. (1963). Corporate income taxes and the cost of capital: a correction. American Economic Review 53 (3): 433443 Mohanty, M. S. and Philip Turner (2008), Monetary policy transmission in emerging market economies: what is new? , Bank for International Settlements Papers no. 35, pp. 1-59, January tker-Robe I. and Pazarbasioglu C., (2010) Impact of Regulatory Reforms on Large and Complex Financial Institutions, International Monetary Fund, Staff Position Note, 3 November, SPN/10/16$ Slovik P., Cournde B. (2011), Macroeconomic Impact of Basel III, OECD Economic Departement Working Papers, No 844, OECD Publishing Wagner W. (2004) The Liquidity of Bank Assets and Banking Stability, Cambridge University Working Paper

May-June 2011

Conjoncture

22

NOTES
The Basel III package comprises three Basel Committee documents dealing with new solvency requirements (BCBS 189, December 2010), liquidity rules (BCBS 188, December 2010) and the absorption of losses when the institution is no longer viable (Annex to BCBS 189, January 2011). 2 European Banking Authority 3 In order to be recognised as core capital, hybrid debt instruments have to meet criteria relating to loss absorption, payment flexibility and permanence. These criteria were approved at G10 level and announced in a press release in Sydney on 27 October 1998. http://www.bis.org/press/p981027.htm11 4 i.e. those outside the regulatory consolidation parameter. 5 The July 2009 recommendations altering the prudential treatment of market risks were revised in December 2010 to take account of changes announced by the Basel Committee in a press release on 18 June 2010. cf. BCBS (2011). 6 Value at Risk measures the risk of maximum ex ante loss for a given probability and time horizon. 7 31 December 2010 was the implementation date initially retained in the July 2009 document, but it has been put back one year. 8 cf. BCBS (2010a), June 2011 revised version, page 36. 9 cf. Basel Committee press release, 1 June 2011 (Basel Committee finalises capital treatment for bilateral counterparty credit risk). 10 The 94 group 1 banks (of which 91 supplied information) have excess Tier 1 capital of over 3 billion, are diversified and active internationally. All other banks are in group 2. 11 Common equity Tier 1 net of Basel III deductions, compared with weighted assets calculated according to the new rules. 12 Common equity Tier 1 before any deductions compared with weighted assets calculated according to the old rules. 13 The minimum rating required to benefit from a 0% weighting in the Basel II standard approach. 14 Ratings corresponding to a 20% weighting in the Basel II standard approach. 15 Unlike capital requirements, the requirements inherent in the new liquidity standards have not been broken down by size of bank (group 1, group 2). 16 A deposit is deemed less stable when it is not insured (a large amount exceeding the guaranteed amount) or when the national regulations class it as liable to be withdrawn quickly (via internet) or denominated in a foreign currency. 17 Under the Basel rules, a deposit cannot be deemed stable unless it is covered by a deposit guarantee scheme. Moreover, the depositor has to have other relationships with the bank that make a massive withdrawal unlikely or the deposit has to be a transaction account to which wages or salary are automatically paid (cf. BCBS, Basel III: international framework for liquidity risk, December 2010, paragraph 56, p.13). 18 The condition concerning the non-financial nature of the issuer does not apply to covered bonds. 19 In open economies, net overseas assets should be added to credit. 20 Credits certainly give rise to deposits, but when their expiry date is more than a year away the former increase the NSFR denominator by their full amount while deposits created in the banking system raise the numerator only by between 50% of their amount (e.g. deposits by non-financial corporates) and 90% of their amount (stable deposits from retail clients). The result is therefore a drop in the NSFR. 21 For example, These metrics () provide the cornerstone of information that aids supervisors in assessing the liquidity risk of a bank. BCBS 188, December 2010, paragraph 138, p. 31. 22 cf. IMF (2011a), box 2.1 page 9. 23 Cited p. 30 of the MAG transition study, August 2010. 24 35 quarters. 25 Weighted by their respective GDP. 26 Up 20.5 basis points in the USA, 14.3 basis points in the euro zone, up 8.4 basis points in Japan. 27 Slovik and Cournde (2011) evaluate the impact of a one-point rise in the average solvency ratio at 14.3 basis points in additional interest margin. According to Cosimano and Hakura (2011), the same rise would lift lending rates by an estimated 11.6 basis points in Germany, 8.1 basis points in Greece and 21.6 basis points in Ireland. 28 cf. Final Report, Assessing the macroeconomic impact of the transition to stronger capital and liquidity requirements, Macroeconomic Assessment Group, December 2010, p.2. 29 The increase in liquid assets to obtain a higher liquidity ratio, ceteris paribus, leads to the substitution of low-risk assets for risky assets in Basel terms, and therefore a decline in risk weighted assets. This makes it easier for banks to comply with the solvency ratio. 30 According to ECB research, the impact of a one-point increase in solvency ratios on loan spreads is 28 points. Under a linearity hypothesis, the impact would be 56 basis points for a two-point increase in solvency ratios. 31 The number of models is as follows: Australia 3, Brazil 5, Canada 8, China 2, India 2, South Korea 6, Mexico 3, Russia 2, UK 5. 32 The baseline hypothesis is a cumulative median loss in long-term GDP amounting to 63% of its pre-crisis level (moderate permanent effects). The extremes are 19% (no permanent effect) and 158% (major permanent effects). 33 A logit model is a regression of the probability that an event materialises. 34 cf. LEI (2010), page 17.
1

May-June 2011

Conjoncture

23

Economic Research Department

Philippe d'ARVISENET Chief Economist OECD countries Jean-Luc PROUTAT Head of OECD countries Caroline NEWHOUSE Country economics Alexandra ESTIOT Structural issues, European Financial Integration UNITED STATES, CANADA 33 1.43.16.95.58 philippe.darvisenet@bnpparibas.com

33 1.58.16.73.32 33 1.43.16.95.50

jean-luc.proutat@bnpparibas.com

caroline.newhouse@bnpparibas.com

33 1.58.16.81.69

alexandra.estiot@bnpparibas.com jean-marc.lucas@bnpparibas.com

Jean-Marc LUCAS
JAPAN, AUSTRALIA, NEW ZEALAND, BENELUX, PENSIONS, LONG TERM FORECASTS

33 1.43.16.95.53

Raymond VAN DER PUTTEN

EURO ZONE, ITALY, EUROZONE LABOUR MARKET

33 1.42.98.53.99 33 1.42.98.27.62 33 1.43.16.95.52 33 1.55.77.71.89 33 1.57.43.02.91 33.1.43.16.95.50

raymond.vanderputten@bnpparibas.com clemente.delucia@bnpparibas.com frederique.cerisier@bnpparibas.com catherine.stephan@bnpparibas.com thibault.mercier@bnpparibas.com caroline.newhouse@bnpparibas.com

Clemente De LUCIA
FRANCE, EURO ZONE PUBLIC FINANCES

Frdrique CERISIER Catherine STEPHAN

SPAIN, PORTUGAL, GREECE, IRELAND

GERMANY, AUSTRIA, SWITZERLAND, EU ENLARGEMENT

Thibault MERCIER
UNITED KINGDOM, NORDIC COUNTRIES

Caroline NEWHOUSE
BANKING ECONOMICS Laurent QUIGNON Head

33 1.42.98.56.54 33 1.43.16.95.41 33 1.43.16.95.54 33 1.42.98.44.24

laurent.quignon@bnpparibas.com delphine.cavalier@bnpparibas.com

Delphine CAVALIER Cline CHOULET Laurent NAHMIAS COUNTRY RISKS

Guy LONGUEVILLE Head Franois FAURE Deputy Head

celine.choulet@bnpparibas.com laurent.nahmias@bnpparibas.com

33 1.43.16.95.40 33 1.42 98 79 82

guy.longueville@bnpparibas.com

francois.faure@bnpparibas.com

Capital flows to emerging markets, Turkey

ASIA

Hlne DROUOT Christine PELTIER

LATIN AMERICA

33 1.42.98.33.00 33 1.42.98.56.27 33 1.42.98.26.77 33 1.42.98.74.26 33 1.42.98.02.04 33 1.42.98.43.86

helene.drouot@bnpparibas.com christine.peltier@bnpparibas.com sylvain.bellefontaine@bnpparibas.com valerie.perracino@bnpparibas.com stephane.alby@bnpparibas.com jeanloic.guieze@bnpparibas.com

Sylvain BELLEFONTAINE
Valrie PERRACINO-GUERIN AFRICA

Stphane ALBY Jean-Loc GUIEZE

EASTERN EUROPE

Central Europe, Baltic countries, Balkan countries Alexandre VINCENT

RUSSIA AND OTHER CIS COUNTRIES

33 1.43.16.95.44 33 1.42.98.48.45 33 1.43.16.95.51

alexandre.vincent@bnpparibas.com anna.dorbec@bnpparibas.com pascal.devaux@bnpparibas.com

Anna DORBEC
MIDDLE EAST SCORING

Pascal DEVAUX

economic-research.bnpparibas.com

Our publications
Conjoncture focuses each month both on the main economic issues and structural problems. Economic Market Monitor provides a detailed follow-up of the economic situation whilst analysing interest and exchange rate developments in OECD countries (8 issues per year). EcoWeek focuses on specific and current economic issues (every Friday). EcoFlash comments and analyses the main economic events (data releases, economic policy decisions) in the hours following their release. EcoTV the monthly broadcast programme from BNP Paribas economists. Each month, Philippe d'Arvisenet and his teams will help you to make sense of the economic and financial news in both French and English. These interviews are available on our website. EcoTV Week the weekly broadcast programme from BNP Paribas economists. In a succinct two-minute format, the new show, presented by one of the Groups economists, covers the economic and financial events of the previous week as well as those that are expected in the coming days. These interviews are available on our website, in both French and English. To receive directly our publications, please subscribe on our website
BNP Paribas is regulated by the FSA for the conduct of its designated investment business in the UK and is a member of the London Stock Exchange. The information and opinions contained in this report have been obtained from public sources believed to be reliable, but no representation or warranty, express or implied, is made that such information is accurate or complete and it should not be relied upon as such. This report does not constitute a prospectus or other offering document or an offer or solicitation to buy any securities or other investment. Information and opinions contained in the report are published for the assistance of recipients, but are not to be relied upon as authoritative or taken in substitution for the exercise of judgement by any recipient, they are subject to change without notice and not intended to provide the sole basis of any evaluation of the instruments discussed herein. Any reference to past performance should not be taken as an indication of future performance. No BNP Paribas Group Company accepts any liability whatsoever for any direct or consequential loss arising from any use of material contained in this report. All estimates and opinions included in this report constitute our judgements as of the date of this report. BNP Paribas and their affiliates ("collectively "BNP Paribas") may make a market in, or may, as principal or agent, buy or sell securities of the issuers mentioned in this report or derivatives thereon. BNP Paribas may have a financial interest in the issuers mentioned in this report, including a long or short position in their securities, and or options, futures or other derivative instruments based thereon. BNP Paribas, including its officers and employees may serve or have served as an officer, director or in an advisory capacity for any issuer mentioned in this report. BNP Paribas may, from time to time, solicit, perform or have performed investment banking, underwriting or other services (including acting as adviser, manager, underwriter or lender) within the last 12 months for any issuer referred to in this report. BNP Paribas, may to the extent permitted by law, have acted upon or used the information contained herein, or the research or analysis on which it was based, before its publication. BNP Paribas may receive or intend to seek compensation for investment banking services in the next three months from an issuer mentioned in this report. Any issuer mentioned in this report may have been provided with sections of this report prior to its publication in order to verify its factual accuracy. This report was produced by a BNP Paribas Group Company. This report is for the use of intended recipients and may not be reproduced (in whole or in part) or delivered or transmitted to any other person without the prior written consent of BNP Paribas. By accepting this document you agree to be bound by the foregoing limitations. Analyst Certification Each analyst responsible for the preparation of this report certifies that (i) all views expressed in this report accurately reflect the analyst's personal views about any and all of the issuers and securities named in this report, and (ii) no part of the analyst's compensation was, is, or will be, directly or indirectly, related to the specific recommendations or views expressed herein. United States: This report is being distributed to US persons by BNP Paribas Securities Corp., or by a subsidiary or affiliate of BNP Paribas that is not registered as a US broker-dealer, to US major institutional investors only. BNP Paribas Securities Corp., a subsidiary of BNP Paribas, is a broker-dealer registered with the Securities and Exchange Commission and is a member of the National Association of Securities Dealers, Inc. BNP Paribas Securities Corp. accepts responsibility for the content of a report prepared by another non-US affiliate only when distributed to US persons by BNP Paribas Securities Corp. United Kingdom: This report has been approved for publication in the United Kingdom by BNP Paribas London Branch, a branch of BNP Paribas whose head office is in Paris, France. BNP Paribas London Branch is regulated by the Financial Services Authority ("FSA") for the conduct of its designated investment business in the United Kingdom and is a member of the London Stock Exchange. This report is prepared for professional investors and is not intended for Private Customers in the United Kingdom as defined in FSA rules and should not be passed on to any such persons. Japan: This report is being distributed to Japanese based firms by BNP Paribas Securities (Japan) Limited, Tokyo Branch, or by a subsidiary or affiliate of BNP Paribas not registered as a financial instruments firm in Japan, to certain financial institutions permitted by regulation. BNP Paribas Securities (Japan) Limited, Tokyo Branch, a subsidiary of BNP Paribas, is a financial instruments firm registered according to the Financial Instruments and Exchange Law of Japan and a member of the Japan Securities Dealers Association. BNP Paribas Securities (Japan) Limited, Tokyo Branch accepts responsibility for the content of a report prepared by another non-Japan affiliate only when distributed to Japanese based firms by BNP Paribas Securities (Japan) Limited, Tokyo Branch. Hong Kong: This report is being distributed in Hong Kong by BNP Paribas Hong Kong Branch, a branch of BNP Paribas whose head office is in Paris, France. BNP Paribas Hong Kong Branch is regulated as a Licensed Bank by the Hong Kong Monetary Authority and is deemed as a Registered Institution by the Securities and Futures Commission for the conduct of Advising on Securities [Regulated Activity Type 4] under the Securities and Futures Ordinance Transitional Arrangements. Singapore: This report is being distributed in Singapore by BNP Paribas Singapore Branch, a branch of BNP Paribas whose head office is in Paris, France. BNP Paribas Singapore is a licensed bank regulated by the Monetary Authority of Singapore is exempted from holding the required licenses to conduct regulated activities and provide financial advisory services under the Securities and Futures Act and the Financial Advisors Act. BNP Paribas (2011). All rights reserved. Prepared by Economic Research BNP PARIBAS Registered Office: 16 boulevard des Italiens 75009 PARIS Tl : +33 (0) 1.42.98.12.34 Internet : www.bnpparibas.com Publisher: Michel Pbereau Printed in France by: Ateliers J. Hiver SA Dpt lgal :May-June 2011 ISSN 0224-3288 Copyright BNP Paribas

economic-research.bnpparibas.com