Annexes to COM(2010)327 - Effects of Directives 2006/48/EC and 2006/49/EC on the economic cycle SEC(2010)754 - Main contents
Please note
This page contains a limited version of this dossier in the EU Monitor.
| dossier | COM(2010)327 - Effects of Directives 2006/48/EC and 2006/49/EC on the economic cycle SEC(2010)754. |
|---|---|
| document | COM(2010)327  |
| date | June 23, 2010 |
Masschelein, N. (2007), "Monitoring pro-cyclicality under the capital requirements directive: preliminary concepts for developing a framework", NBB Working Paper Document No. 120
[1] Comprising Directive 2006/48/EC of the European Parliament and of the Council relating to the taking up and pursuit of the business of credit institutions and Directive 2006/49/EC of the European Parliament and of the Council on the capital adequacy of investment firms and credit institutions
[2] The mandate of the TFICF is to monitor the level and volatility of banks’ minimum capital requirements as defined by the CRD and analyze their impact on bank lending and the economic cycle.
[3] A questionnaire to the EU businesses ran from August 12 until September 28, 2009. Via the European Business Test Panel, the business community was consulted on their perceptions regarding the developments in availability and conditions of bank credit since October 2008. 429 members of the EBTP from 28 EU/EEA countries filled in the questionnaire. As regards the company size of the respondents, 59% of them were small enterprises, 20% medium enterprises and 21% large enterprises.
[4] A questionnaire to the EU banking industry ran from May 21 until August 31, 2009. Nineteen responses were received from IT, SI, BE, AT, DE, UK and NL. Respondents included banks whose majority ownership is private (publicly traded), national cooperative bank networks and publicly owned banks. Some replies were submitted by national associations representing a consolidated position of their members, effectively increasing the number of individual credit institutions covered by the questionnaire.
[5] Directive 2009/111/EC of the European Parliament and of the Council of 16 September 2009 amending Directives 2006/48/EC, 2006/49/EC and 2007/64/EC as regards banks affiliated to central institutions, certain own funds items, large exposures, supervisory arrangements, and crisis management
[6] For a thorough description of the main sources of pro-cyclicality and the role of capital adequacy rules therein see BIS (2008) and Masschelein (2007)
[7] In particular, financial institutions have difficulties in assessing absolute level of risk (while they fare better at assessing relative risk), especially over a prolonged period, and so rarely identify booms with consequences for systemic risk. Measures of risk may be quite low as vulnerabilities and risk build up during the expansion phase but spike once tensions arise, for example, the market risk embedded in banks' trading book can be easily underestimated if measured over short holding periods. Such limitations to perception of risk are in part attributable to the paucity of information regarding the dynamics of systemic risk and are explained by certain theories of behavioural finance such as disaster myopia and cognitive dissonance. For a comprehensive review, see Borio et al. (2001)
[8] When economic conditions are depressed and collateral values decline, information asymmetries with respect to the quality of clients’ balance sheets can imply that even borrowers with profitable projects find it difficult to obtain funding. When economic conditions improve and collateral values rise, the opposite situation may occur. This reasoning suggests that pro-cyclical effects may be more pertinent to borrowers which are more prone to asymmetric information, including small and medium-sized enterprises not subject to external ratings and extensive disclosure requirements.
[9] In some cases, responses are explained by short-term bias of remuneration structures or herding behaviour (tendency of market participants to conform their behaviour with that of their peers). Remuneration policies in financial institutions may have an enhancing pro-cyclical effect where they entail (possibly disproportionate) rewards on the upside and insufficient penalties on the downside, e.g., bonuses based on short-term profits that are paid immediately, with no risk adjustment or deferred payment to take account of future performance of the business unit or institution as a whole. For a comprehensive review, see Borio et al. (2001)
[10] Conversely, in an economic downturn, market developments may lead to a decrease in the fair value of assets as well as to increased provisioning requirements. These downward adjustments in asset valuations and increasing loan losses could have an effect on banks to tighten lending standards and cut back lending.
[11] As regards the relative impact of individual risk parameters of the IRB approach, PD is considered to be the main contributor to the cyclicality of the framework, while, to a lesser extent, the effects of LGD and EAD may also be relevant. The degree to which capital requirements oscillate may depend, among other factors, on whether a Point-in-Time (PIT) or a Through-the-Cycle (TTC) system is implemented by banks in their internal rating processes. Typically, TTC ratings do not change rapidly in response to fluctuations in the macroeconomic conditions, and thus are less influenced by the economic cycle momentum. Use of TTC rating systems is suggested by the Basel II and the CRD frameworks as a way to smooth the potential volatility of the capital requirements. However, due to banks' preferences to use recent default data and early warning systems as well as difficulties in obtaining a sufficient set of data, they primarily use PIT approaches, especially for exposures to SMEs and retail clients.
[12] Under the SA, requirements are expected to fluctuate less as banks use risk weights based on external ratings (which tend to be determined on the TTC basis) issued by external credit assessment institutions
[13] The FIRB risk-weighted assets for corporate portfolios are expected to be less responsive to business cycle than those under the AIRB, as under the latter banks use their own estimates of LGD and EAD. Internally estimated LGD and EAD are seen as more responsive to business cycles than the supervisory estimates used under the FIRB
[14] In addition to credit risk parameters, there is likely to be an impact of the use of products of credit risk transfer, such as securitisations, on the cyclicality of capital requirements. In favourable circumstances, securitisation markets support the transfer of credit risk from individual institutions, thus having a relieving impact on banks’ capital positions. Recent evidence revealed, however, that this risk transfer function might not necessarily be used effectively in a stressed market situation, which, in turn, may amplify the adverse effects on banks’ capital positions.
[15] June 2008, December 2008 and June 2009 for about 60-80 IRB banks from fourteen EU countries
[16] Without taking into account the transitional floors
[17] A banurteen EU countries
[18] Without taking into account the transitional floors
[19] A bank is considered a Group 1 bank if its Tier 1 capital is above €3 billion and it is well diversified and internationally active
[20] The impact of the CRD on capital requirements from the IRB approaches to credit risk may not yet fully be reflected in data disclosed by banks to the extent that banks are applying the partial use of the Standardised Approach. In December 2008, the share of banks' partial use exposures was significant for both Group 1 (31%) and Group 2 (40%) banks
[21] Indeed, MRC per exposure - a measure of riskiness - was either unchanged (for Group 1 banks) or slightly lower (for Group 2) in December versus June 2008. While Group 1 banks’ portfolio-level PDs increased on average, non-defaulted PDs were virtually unchanged indicating that the rise in average PDs was due to a higher number of defaults. To the extent that the rise in PDs reflects an increase in the share of defaulted exposures, the impact of rising average PD on RWA (and on MRC) is less straightforward since the risk weight on a defaulted exposure is zero (with some exceptions under AIRB). The risk is then accounted for by the expected loss, for which risk provisions should be made (in case of their shortfall, a deduction from capital is made). Furthermore, LGDs were lower for a number of exposures in December versus June 2008.
[22] MRC effectively summarises the numerator and the denominator of the regulatory capital ratio. It measures the capital required to cover (i) 8% of RWA; (ii) differences between total eligible provisions and the total expected loss amount; and (iii) other deductions.
[23] PD are estimated using the P&L data, which are affected by the business cycle with some delay and are released with a further time lag. For example, if revenues of a firm start decreasing at the beginning of a year, they are reflected in its year-end financial statements that are published only some months into the following year and might be incorporated in the rating systems some additional months later.
[24] Represented by lagged output gap and measured as a difference between actual and potential GDP
[25] Only for the corporate portfolio the coefficient is statistically significant (at 1%)
[26] Regression analyses on the relationship between the change in individual risk input parameters and the business cycle, conducted by the ECB, found that PDs of corporate and retail portfolios tend to increase more strongly in countries with low business activity, measured by the output gap lagged by one period. Similar regressions with respect to LGDs produced coefficients for all portfolios that are negative but statistically insignificant. This may indicate that the LGD values recorded in banks' internal models are somewhat stickier than PDs and thus may react less immediately to changes in the macroeconomic environment. Therefore, a longer sample period might be needed to provide firm evidence of the expected strong negative relationship between economic activity and LGDs. An additional explanation for failing to find any link between LGDs and the business cycle may be given by regulatory requirements that stipulate the use of stress LGDs.
[27] Banks' answers to the questionnaire conducted by the Commission, however, indicate that regulatory capital requirements are the most important approach to determine target capital level - for both the SA and the IRB banks
[28] BLS of July 2009, covering a sample of more than 100 euro area banks, included questions about the impact of the CRD on the banks’ lending policies (measured by their credit standards) during the period from Q1 2008 to Q2 2009
[29] After trading account profit (including unrealised gains and losses) of €82 billion in 2007, a sample of 105 EU banks incurred trading account loss of €94 billion in 2008; sources: Orbis, Bloomberg
[30] A survey conducted by the ECB in cooperation with the Commission between June 17 and July 23, 2009, covering 6,000 SMEs and large firms in the euro area
[31] The larger the firms, the more successful they were in obtaining a loan as only 5% of large firms had their loan applications rejected
[32] Feedback from 429 EU/EEA businesses showed that in the period from October 2008 to September 2009 13% of businesses who applied for bank credit were not successful in obtaining any funding in the form of bank credit, and this percentage was higher for SMEs (15%) than for large enterprises (5%). The replies also showed that for a sub-sample of businesses who said that they had been adversely affected by availability of bank credit, obtaining the funds needed from other sources was more difficult to SMEs relative to large enterprises.
[33] Cappiello et al. (2009) provide empirical evidence on the existence of bank lending channel of monetary policy transmission in the euro area. The authors find that changes in the supply of credit, in terms of both volumes and credit standards applied to loans to enterprises, have significant effects on real economic activity. Therefore, bank balance sheet constraints, such as binding regulatory capital requirements - to the extent that they lead banks to reduce their loan supply – may have negative repercussions on real economic growth.
Ciccarelli et al. (2009) use a panel VAR for the euro area with GDP, prices, short-term interest rates, loan demand and loan supply conditions. Using impulse response functions, the authors find that a tightening of credit standards lead to a significant decline in real GDP growth. Hence, to the extent that the CRD affects bank credit standards, this finding suggests that eventually economic activity would be also affected by the CRD.
[34] http://ec.europa.eu/internal_market/bank/regcapital/index_en.htm
[35] Banking regulators have not been in favour of allowing banks to utilise their own credit risk models to derive capital requirements, due to the uncertainty and lack of data with respect to asset correlations. As recent experience has shown, banks’ internal models have not performed particularly well
[36] Directive 2009/111/EC which amended the CRD requires full due diligence on the part of banks, including 5% retention of securitised assets, while the governance of rating agencies will be addressed by the Regulation on Rating Agencies
[37] However there are important factors, that can reduce the impact of the increases in capital requirements if ratings deteriorate, because: (i) rating downgrades were accompanied by accounting valuation losses. As EU banking rules allow banks to off-set accounting valuation losses against the capital requirement, the effect of rating downgrades will be far lower for those banks that adopted conservative valuation policies for their securitisation positions; (ii) banks that have benefited from government guarantees for their securitisation positions or that have sold them into the markets or to specialised government sponsored vehicles ("bad banks"), will also have had full or partial relief from the increases in capital requirements.
[38] At the end of June 2008, these averages stood at 2.9% for Group 1 and 4.5% for Group 2
[39] The IASB has undertaken a comprehensive review of the IAS 39 "financial instruments" including credit loss provisioning, and in November 2009 issued a separate Exposure Draft (ED). The ED contains five high level measurement principles (Expected Cash Flow Model, ECF) that reporting entities should use in combination with the principles-based application guidance to determine the amortised cost. Under the proposed method banks would:
a) initially estimate the expected cash flows for the (remaining) life of the financial instrument including the expected credit losses (taking into account the collateral);
b) calculate the “effective interest rate” (internal rate of return) on the basis of the expected cash flows. The effective interest rate would be lower than the contractual interest rate because estimated credit losses would be taken into account;
c) at each reporting date review the initial cash flow and credit loss expectations and revise them when necessary.
The ECF approach could improve provisioning for credit risks compared to the current incurred loss model of IAS 39 because it would no longer require a trigger event and would allow an earlier recognition of credit losses. The ECF approach would also achieve a better matching of contractual interest income with expected losses. However, the ED provides limited guidance on the use of data and models. The ED basically requires point-in-time estimates for cash flows and expected credit losses. As well as the Basle Committee and other stakeholders, the Commission is assessing the IASB proposal and will contribute to its improvement in accordance with the objectives set out by the G20.
[40] G-20 (2009)
[41] See para 5 for more on inappropriate responses of market participants to changes economic conditions
[42] COM (2009) 362 final
[43] Ibid
[44] See COM (2009) 499 final
[45] See para 5 for more on these issues
[46] ECOFIN (2009)