Real impact of Basel II on emerging economies

Transcription

1 Real impact of Basel II on emerging economies Will Basel II discourage exposures to emerging economies and exacerbate their cyclicality? Master Thesis by Gerbert van der Kamp Supervisor: Dr. Van Roij

2 Table of contents Table of contents 2 Index of tables and figures 5 Foreword 6 Abstract 7 Introduction 8 1. Basel I and II capital Accords Basel I The current capital framework Background and objectives of Basel I Basel I framework Advantages of Basel I Disadvantages of Basel I Basel II The revised framework Background and objectives of Basel II Basel II framework Pillar Basic credit risk measurement technique 16 1) The standardised approach 16 2) The IRB approach Credit risk mitigation Securitization framework 21 1) The standardised approach 21 2) The ratings-based approach Operational risk 22 1) The basic indicator approach 22 2) The standardised approach 22 3) The advanced measurement approach Pillar 2 Supervisory review 23 1) Credit concentration risk 24 2) Interest rate rate risk in the banking book Pillar 3 Market discipline Advantages Basel II 25 2

3 Disadvantages Basel II Conclusionary remarks Procyclicality of Basel II Risk sensitive capital requirements and the credit crunch Measurement of the risk components and procyclicality PD and the business cycle The correlation between PDs across borrowers over the business cycle LGD and the business cycle Estimated procyclicality of capital requirements Empirical studies Overview of estimations Simulation by Kashyap & Stein Accuracy of these estimations 36 1) Neglecting the advanced IRB approach 36 2) Lucas critique Reasons why changes in capital requirements will not lead to procyclicality The buffer stock channel Alternative sources of credit Bank lending is also demand determined Conclusionary remarks Borrowing position of emerging economies and Basel II Basel II in emerging economies Implementation of Basel II in emerging economies The standardised or IRB approach? What will emerging economies do? Dependence on credit from foreign banks Will lending to emerging economies be discouraged? IRB capital requirements for emerging economy exposures Effects on interest rates Assumptions on the required return on capital Capital requirements binding or not binding? 49 1) Does Basel I bind? 50 2) Will Basel II-IRB bind? Diversion to other sources of credit Competitive position of emerging economy banking sector Procyclical effect Basel II in emerging economies Business cycle amplitude and proneness to crisis of emerging economies 53 3

4 Basel II and the acceleration of emerging economy crisis External rating volatility Internal rating volatility Treatment of short-term debt Conclusionary remarks Policy options Softening capital requirements for emerging economy exposures Flattening the IRB curve for emerging economy exposures Including diversification effects Measures against procyclicality Adjusting capital requirements to the state of the business cycle Smoothing fluctuations in capital requirements Forward-looking/ dynamic provisioning Conclusionary remarks Summary and conclusion 65 References 68 4

5 Index of tables and figures Table 1 Risk weight per asset category 10 Table 2 Risk weights under the standardised approach 16 Table 3 Risk components within the IRB approach 17 Table 4 Summary of research on volatility of capital requirements 34 Table 5 Volatility of capital requirements 35 Table 6 Comparing the IRB with the standardised approach 46 Figure 1 Choices under pillar 1 15 Figure 2 Risk weights for corporates 19 Figure 3 Procyclical effect of a PD-LGD correlation 32 Figure 4 IRB capital requirements for private sector exposures in Latin America 47 Figure 5 Average volatility by region 54 Figure 6 The IRB curve flattened 59 Figure 7 Impact of different capital formulae on capital requirements for SME exposures 59 Figure 8 How the IRB capital formula can be flattened as soon as recession hits 61 5

6 Foreword I would like to express my gratitude to dr. Van Roij, without whose kind help I would never have succeeded in writing this thesis. Further, I would like to acknowledge friends, family, and colleague students for providing me with advice and cheerful company. Gerbert van der Kamp 14 th of August

7 Abstract This paper tries to examine whether the switch from Basel I to Basel II as planned on the 1 st of January 2007 will have a real impact on emerging economies. Namely due to that the new framework installs a highly risk-sensitive capital regulation, it has been suggested that it will strongly work to the detriment of emerging economies borrowing positions. Two problems will be of central importance in my thesis: first of all that exposures to emerging economies may be discouraged and second; that cyclicality of emerging economies in specific may be exacerbated. Although it is difficult to find a consensus on how serious these problems will be, the literature has brought forth several ideas for how to mitigate them. My recommendation is that the range of policy options should be under careful consideration on an international level. 7

8 Introduction Before giving attention to the position of emerging economies I will first present both Basel capital frameworks in chapter 1, and after that in chapter 2 the literature on procyclicality that has arisen in connection to Basel II. Clear from these chapters will be how the new highly risk-sensitive capital regulation is going to work and how it could lead to stronger business cycles by either stimulating or destimulating credit growth. Dependent on various countervailing factors it seems that there is some potential for a procyclical effect. Then, chapter 3 contains the heart of my thesis as it examines the threats to emerging economies borrowing positions. First of all, it is examined whether exposures to high-risk emerging economies will cause strong increases in the required level of capital with the switch to Basel II, and whether these economies will have to face increased interest rates while borrowing. Second, it is examined whether the already more volatile emerging economies will have a higher likelihood of facing increased procyclicality under Basel II. Finally, in chapter 4, several policy options are presented in order to curb potential negative effects of Basel II on emerging economies. Two of which concentrate on adapting the new capital Accord to the specific benefit of these countries. Three other measures focus on countering procyclicality, either by adapting the Basel II regulation to make capital requirements less volatile over the business cycle, or by offsetting part of the capital requirement volatility by more forward-looking provisioning. Emerging economies clarified Most interesting from my point of analysis is to consider those countries that display a high level of riskiness and volatility, which are characteristics usually associated with less-developed countries. Therefore I tend to focus on those countries whose sovereigns are assigned an external credit rating not any better than speculative grade (BBB- and lower). On the other hand however, they should be on a level of development that their situation can be influenced by Basel II through at least one of the following possibilities: 1) local banks adopt the new Accord 2) foreign banks on which these economies depend adopt the new Accord. Note hereby that these possibilities are of no relevance if a country both has barely adopted Basel I and has not opened up yet to credit from foreign banks. Due to the literature which on this issue doesn t always clearly distinguishes between developing countries and emerging economies, I will use both terms alternatively throughout my thesis. 8

9 1. The Basel capital Accords In this chapter the current framework of Basel I will first be presented in section 1.1. After that in section 1.2 the new framework of Basel II is presented. Finally in section 1.3 some conclusionary remarks are given Basel I The current capital framework 1 In this section the current capital framework, Basel I, is discussed. First, in subsection the background, objectives and application of Basel I will be discussed. Then subsection will give a brief account of how Basel I sets capital requirements. After that in and 1.1.4, the advantages and disadvantages of this Accord will respectively be discussed Background, objectives, and application The capital regulation currently most widespread is that of the Basel I Accord, which was installed in The framework has been endowed with two fundamental objectives, as stated by the Basel Committee of Banking Supervision (BCBS): 2 Two fundamental objectives lie at the heart of the Committee's work on regulatory convergence. These are, firstly, that the new framework should serve to strengthen the soundness and stability of the international banking system; and secondly that the framework should be fair and have a high degree of consistency in its application to banks in different countries with a view to diminishing an existing source of competitive inequality among international banks. 3 The first objective, shortly to serve bank strength and soundness, is given shape by installing risk sensitive capital requirements. Wherein the required capital depends on the assumed riskiness of the category of asset, such as residential mortgages or loans to companies. With their second objective the Basel I Accord has aimed at establishing competitive equality between international banks. The problem in the mid eighties namely was that there were major differences in capital requirements between countries, leading to an unfair competitive advantage for those international banks whose regulations were relatively lax. 1 See BIS (1988) 2 The Basel Committee of Banking Supervision is an institution founded in 1974 by the G10 countries that concerns itself with the improvement of banking supervision on an international level. Its current members are Belgium, Canada, France, Germany, Italy, Japan, Luxembourg, the Netherlands, Spain, Sweden, Switzerland, United Kingdom and United States. For further information see: 3 BIS (1988; Page 1, paragraph 3) 9

10 On the issue of application of the capital requirements within international banks the Committee has chosen for the principle of consolidation. Which is to say that the totality of assets and capital of a banking group is to be added up in a way that that crosses out mutual assets and liabilities inside of the banking group e.g. the parent group holding equity in its subsidiary. The reason behind this principle is that default of any entity within the banking group foreign affiliates included can impair the overall capital position of the bank, and thus can threaten solvability of the bank s balance sheet The Basel I framework 4 In the capital requirement calculation of Basel I it is not simply the totality of assets that counts, it is about how risky the assets are. By actually ascribing a risk-weighting according to asset category, riskiness of each particular asset is roughly taken into account. The Accord treats certain categories of assets as more risky and others as less risky. For example, loans to OECD governments were perceived by the Committee as being riskless and therefore are assigned a risk weight of 0%. Which means that a bank entirely focussing on lending to this group is not required to hold any capital at all. Loans to certain riskier groups of borrowers such as companies or non-oecd governments would bear full (100%) risk weight. As one can see in the table below, every kind of asset is assigned a risk weight likewise. Table 1 Risk weight per asset category Risk Weight under Basel I Loans to 0% OECD-government Non-OECD governments locally funded Cash 20% Banks and securities firms in OECD countries Short-term loans to non-oecd banks 50% Residential mortgages 100% Non-OECD governments Source:: UAB (2003; Page 6) Companies, and all other types By multiplying the size of a certain exposure with the proper risk weight, any asset could be transformed to a risk-weighted asset, as is shown in the formula below. Risk weighted asset = asset * risk weight 4 See BIS (1988) 10

11 Against the sum of risk weighted assets 8% of capital must be held a percentage also denoted as the core capital asset ratio. This required amount of capital can be met by holding a combination of tier 1 and tier 2 capital. Tier 1 capital consists of high quality equity capital such as share capital and disclosed reserves, whereas tier 2 includes lower quality forms of capital such as undisclosed reserves, general provisions and subordinated debt. Taking in mind the quality difference, the Committee decided that at least half of the capital requirement which a bank faces must be met with tier 1 capital (i.e. 4% of risk weighted assets). Next to the exposures on the balance sheet, off-balance sheet exposures are also taken into account in the calculation of capital requirements. The procedure herein is to first translate it into a credit risk equivalent through the use of a credit conversion factor. This conversion factor ranges from 0 to 1 depending on the type of off-balance sheet exposure. For example commitments that can be unconditionally cancelled receive 0 and full loan guarantees receive 1. The credit equivalent will then be acquired by multiplying this number with the the underlying exposure. Its result is finally to be risk weighted according to the category of counterparty just as is represented in table 1. The requirements presented above concern credit risk held in the banking book. The banking book is, as opposed to the trading book, the book that deals with the traditional banking business of keeping nonspeculative assets. The largest risk faced in the banking book is credit risk. The trading book on the other hand contains activities of speculative intention and can be characterized as having short duration. Assets in the trading book typically have high liquidity, although they do carry the risk of rapidly losing their value on the market (i.e. market risk). Market risk is defined by the Committee as the risk of losses in trading positions when prices move adversely. 5 In order to cover this element under Basel I the Committee added a treatment of market risk in This amendment obligated banks to hold capital against their market risk exposure. In determining the size of this exposure, banks where for the first time allowed to use their own measurement systems. With the venue of Basel II, treatment of market risk will remain largely unchanged. I will not profoundly discuss market risk since it only is responsible for a small part of total capital requirements, and is of little importance to the issues that I try to deal with in my paper. Nonetheless, together with credit risk market risk is one of the two risks treated under the current capital Accord, whose main metric can be summed up in the following formula: CapitalRat io TotalCapital = CreditRisk + MarketRisk 8% 5 BIS (2001-A; Page 11) 11

12 Since its gradual implementation from the end of the 80s to the beginning of the 90s, Basel I has formed the international standard for capital adequacy up to this day. It has been adopted by around 130 countries Advantages of Basel I The Committee points to two fundamental merits of the Basel I Accord: The Committee believes the 1988 Accord and subsequent additions and amendments have helped to strengthen the soundness and stability of the international banking system and have enhanced competitive equality among internationally active banks. 7 Herewith they conclude that both of the Accord s objectives have been accomplished from its 1988 inception onwards. They namely argue for this as follows: The Accord was followed by substantial increases, primarily during the transitional period between , in the capital ratios of nearly all internationally active banks. This trend generally has continued, particularly since pressure from the market on banks to maintain strong capital ratios has increased. The widespread adoption of the Accord in many countries has contributed to achievement of the objective of competitive equality Disadvantages of Basel I Despite the Accord s accomplishments, the Committee argues that it also has its weaknesses: The financial world has developed and evolved significantly during the past ten years, to the point where a bank s capital ratio, calculated using the current Accord, may not always be a good indicator of its financial condition. The current risk weighting of assets results, at best, in a crude measure of economic risk, primarily because the degrees of credit risk exposure are not sufficiently calibrated as to adequately differentiate between borrowers differing default risks. Another related and increasing problem with the existing Accord is the ability of banks to arbitrage their regulatory capital requirement and exploit divergences between true economic risk and risk measured under the Accord. Regulatory capital arbitrage can occur in several ways, for example, through some forms of securitisation, and can lead to a shift in banks portfolio concentrations to lower quality assets. 6 Federal Ministry for Economic Cooperation and Development Division of Development Education and Information Sectoral policy paper on financial system development January BIS (1999; Page 8 9) 8 BIS (1999; Page 8 9) 12

13 Finally, for some types of transactions, the Accord does not provide the proper incentives for risk mitigation techniques. For example, there is only minimal capital relief for collateral, and in some cases, the Accord s structure discourages the use of credit risk mitigation techniques. Clearly, the large discrepancy in risk sensitivity with banks own risk methods, the incentives for capital arbitrage, and the poor treatment of credit risk mitigation, all give scope for a considerable improvement Basel II - The revised framework This section presents the revised framework of Basel II. First in subsection 1.2.1, the background and objectives of Basel II will be discussed. Then I will present the three pillar framework in subsection 1.2.2, followed by its advantages and disadvantages in and respectively Background and objectives of Basel II As mentioned in the previous section Basel I can be thought of in several respects as obsolete. In acknowledgement of this, the Committee proposed a revised capital adequacy framework in 1999 called Basel II. This second Basel capital Accord was first planned to start in 2003, but because of critique from the financial community it has been postponed. Until now the Committee issued three consultative documents and conducted an equal number of Quantitative Impact Studies (QIS). These papers respectively served the purpose of engaging in a dialogue with the banking industry and assessing the impact on the aggregate level of capital requirements. In June 2004 a final version of the Accord has been scheduled to be implemented by the end of 2006 by the BCBS members. The Committee has stated the following on the Accord s objective: The fundamental objective of the Committee s work to revise the 1988 Accord has been to develop a framework that would further strengthen the soundness and stability of the international banking system while maintaining sufficient consistency that capital adequacy regulation will not be a significant source of competitive inequality among internationally active banks. 9 Extra soundness and stability is to be achieved by Basel II s creation of a framework with much more risksensitive capital requirements. Accompanying that is the notable change made by including the risk assessments coming from banks internal systems. Development of these systems as part of a step toward stronger risk management is actually promoted now, in contrast to the situation under Basel I. The 9 BIS (2005; Page 2, paragraph 4) 13

14 Committee considers the adoption of stronger risk management practices beneficial. Something which clearly shows from the attention drawn to this point in the framework of Basel II The Basel II framework The Basel II framework is built on three pillars. The first pillar covers minimum capital requirements, the second pillar concerns the supervisory review process and the third pillar concerns market discipline. The idea is that these pillars will be mutually enforcing, so that the new Accord will lead to soundness and stability Pillar 1 Pillar 1 is the pillar that sets the capital requirements. The risk categories that are already treated under Basel I, will again be included in the capital calculation. In addition, there will be a new type of risk, called operational risk. Which the Committee defines as follows: the risk of direct or indirect loss resulting from inadequate or failed internal processes, people and systems, or from external events. 10 After taking into account the new risk category of operational risk, the main metric for the calculation of required capital isn t much different from that of the current Accord: CapitalRat io TotalCapital = CreditRisk + OperationalRisk + MarketRisk 8% It is clear that it is the same as under Basel I except that operational risk entered the denominator. For an average economy the aggregate level of capital requirements doesn t have to change much: there is no change in the definition of capital from Basel I and the capital requirements for the three risks summed up is targetted to be similar in magnitude to those of the current framework. 11 With the Committee s aim, therefore, of keeping the aggregate level of capital the same, operational risk will replace credit risk a little. As the determination of market risk, already known from Basel I, remains practically unchanged, I will not give specific attention to this risk. Both in determining credit risk and operational risk Pillar 1 offers different alternatives to be chosen from. To be more precise: three in basic credit risk measurement, two in credit risk mitigation, two in securitisation risk measurement and three in operational risk measurement. 10 BIS (2001-B; Page 2, paragraph 6) 11 BIS (2005; Page 4, paragraph 14) 14

15 The following organigram gives an overview of alternatives in each of the four areas. In this figure the approaches which are more advanced are placed closer to the right. Figure 1 Choices under pillar 1 Pillar Credit risk measurement Approach Standardised Foundation IRB Advanced IRB Credit risk mitigation Simple Comprehensive Advanced IRB Securitisation risk Standardised Ratings based Operational risk Basic indicator Standardised Advanced measurement Next to the alternatives in the organigram there can also be opted for the so-called simplified standardised approach which is especially designed for unadvanced banks. It more or less corresponds to having combined the adoption of the simplest approaches in each of the four areas except that it includes some complementary restrictions. More generally it holds that banks are likely to choose an advanced approach in one area if also chosen in others. In the four following paragraphs the four areas mentioned in the organigram together with each of its alternatives will be discussed. 15

16 Basic credit risk measurement technique The Committee envisaged to improve treatment of credit risk by introducing a standardised approach and an internal ratings based approach into the new approach. Of which the latter can be split up into foundation and advanced. The first main approach (the standardised) relies on external credit assessment, whereas the second approach (the IRB) uses the internal credit risk assessments of banks. In the following paragraphs the two different approaches will be discussed. 1) Standardised approach 12 In this system credit ratings, usually assigned by external rating agencies, play a crucial role. In the table below we find how credit ratings of borrowers map into concrete risk weights. A high credit rating stands for low-risk and herein would result in a low risk weight and vice versa for low-rated credit. In order to arrive at the required amount of capital, these risk weights would have to be multiplied with 8%. The capital charge for example for a $100 exposure that carries a 50% risk weight, would then be 8% * 50% * $100 = $4 Table 2 Risk weights under the standardised approach Credit AAA to BBB+ to BB+ to A+ to A- rating AA- BBB- BB- B+ to B- Below B- Unrated Sovereign 0% 20% 50% 100% 100% 150% 100% Bank Option 1 20% 50% 100% 100% 100% 150% 100% Bank Option 2 20% 50% 50% 100% 100% 150% 50% Bank Option 2 (maturity 3 20% 20% 20% 50% 50% 150% 20% months) Corporates 20% 50% 100% 100% 150% 150% 100% Residential mortgage 35% (if 90 days past due 100%) Retail exposure 75% (if 90 days past due 100% or 150%) Source: BIS (2005; Page 15 20) As can be seen in the table, the sovereign group is treated slightly more favorable than the other groups. An exposure to an AAA rated sovereign would for example carry a risk weight that is 0%, whereas that to an 12 See BIS (2005; Page 15 47) 16

17 AAA rated corporate is 20%. Also banks are treated slightly more favourable than corporates, as exposures to banks can be treated according to three different options. The first of which assigns a risk weighting one notch higher than the rating of the country they are settled. Which could be beneficial to some extent to a low rated bank in a high rated country. The second option for exposures to banks does use the banks own credit rating. The third option is for exposures to banks that are of a maturity lower than three months, which prescribes risk weights that are relatively favourable. Another thing to notice is that the standardised approach has a special treatment of loans supported by residential mortgage, which receives a weight of merely 35%. And also retail exposures, which are exposures of under $1 million, are assigned a special risk weight of 75%. As these categories of borrowers are usually too small to acquire a credit rating, they are not assigned a risk-sensitive risk weight. 2) IRB approach 13 Under the internal ratings based approach banks have to rely on their own internal ratings systems to rate their exposures for risk weighting purporses, instead of referring to external credit ratings. In order to internally rate their exposures, banks will be relying on their own risk measurement models. Before being allowed to the IRB approach banks internal systems must be qualified as adequate by supervisors (even according to a detailed set of minimum requirements 14 ). The ratings that they produce then have to subsequently be mapped into risk components ready to enter a specified IRB capital formula. Four different risk components are to be assigned to each asset: Table 3 Risk components within the IRB approach PD - Probability - The likelihood that the borrower will of default default over a given time horizon LGD - Loss given - The proportion of the exposure that will default be lost if a default occurs EAD - Exposure at - Book value of the asset less effects from default credit-risk mitigation M - Maturity - Remaining economic maturity of the Source: Illing, Paulin (2004; Page 16) exposure in years 13 See BIS (2005; Page ) 14 BIS (2005; Page 84, paragraph 387) 17

18 Within the IRB approach, the choice can be made between advanced and foundation, each of which represents a different level of complexity. The key difference is that under the foundation IRB approach banks are only responsible for determining one risk component, namely probability of default (PD), whereas the other three are set by the regulator. Under the advanced IRB approach banks have to be able to determine all four risk components. Note that before being allowed to these approaches, the regulator must find banks capable of adequately performing the necessary risk assessments. If an asset has been assigned its risk components, the IRB risk charge can be determined. The formula used in calculating capital requirements is similar for different types of assets, therefore I will only show the formula used for corporates exposures: K IRB N = LGD * N 1 ( ) 1/ 2 1 PD + p N ( q) 1 p * EAD ( PD * LGD * EAD) Herewith, an instrument with some nominal value, as designated by EAD, can be assigned its risk charge. In this formula both risk components PD and LGD positively affect the capital requirement. Crucially, the part representing expected losses (PD*LGD*EAD) is substracted. Which reflects the idea that the provisions should cover these expected losses, whereas capital requirements are only set to cover unexpected losses. If provisions fall short of expected losses then the shortfall will be substracted from available capital. 17 Due to that the formula prescribes capital per single instrument it doesn t deal with correlation risk in the same manner as a portfolio credit model. However, in order to correct for this the parameter p is included which represents the assumed correlation between different assets. This correlation factor is set by the supervisor according to a formula depending on PD (correlation is assumed somewhat lower as PD increases). The higher the correlation factor the more capital is required, because high correlation would imply interdependence among borrowers and therefore high riskiness. What the formula aims at is attaining a certain confidence level represented by q. Which is a parameter that stands for the probability the bank has of not defaulting next year, when complying with the requirements determined by the formula. Under Basel II this parameter is set at which would mean that the formula guarantees a confidence level of 99.9%. As Majnoni and Powell (2004; Page.10) put it: then a bank is only expected to use up its capital in one year with a probability of 0.1% or once every 1000 years. 15 Illing, Paulin (2004; Page 13) 16 Gordy, Howells (2004; Page 4) 17 See subsection for a deeper discussion 18

19 To get a picture of what risk weights the IRB formula prescribes I plotted in the graph beneath how risk weight increases with PD if LGD and M are assumed fixed at 45% and M 2.5 years respectively. 300 Figure 2: Risk weights for corporates IRB risk weight Probability of Default (PD) Source: calculation is from BIS (2005; Page 230), with LGD = 45% and M = 2.5 Note hereby that to obtain the actual corresponding capital requirement it is necessary to multiply the risk weight with 8%. For example an $100 exposure that is assigned a PD of 5%, which according to the graph corresponds to a IRB risk weight of 150%, would lead to a capital requirement of $100 * 150% * 8% = 12%. Clear is that the IRB approach is more risk sensitive than the standardised approach. Under the IRB approach the risk weight can sometimes reach 250% whereas the standardised approach doesn t allow it go to further than 150%. Further it must be mentioned that the curve is less steep in case of a lower maturity of the exposure short term loans namely are considered less risky Credit risk mitigation In the area of credit risk mitigation Basel II is more extensive than Basel I. Important changes include the following: Under the new Accord the recognition of collateral is not limited anymore to cash or securities issued by OECD- governments and multilateral development banks; also that of non-oecd 19

20 entities is recognized. Furthermore, the better securities issued by normal banks are also recognized as collateral, just as equities listed on a main index. 18 Different approaches can be chosen from in the treatment of collateral. Next to the simple approach, which replaces the credit rating of the underlying exposure with that of the collateral, most banks can also choose for a comprehensive approach. This approach effectively reduces the underlying exposure by substracting the collateral from it. Banks adopting the standardised approach for credit risk usually treat collateral under the simple approach, whereas banks under the foundation IRB have to use the comprehensive approach. Banks under the advanced IRB approach can use their own internal estimates to adjust the loss given default from their exposures. Paragraph. 19 Guarantees and credit derivatives receive wider recognition than under Basel I. Under the new Accord any sovereign entity, public sector entity or bank can give credit protection, provided that they have a higher credit rating than the underlying exposure. Also entities rated A- or better (like non-bank companies) are now allowed to do the same. 20 Treatment of guarantees and credit derivatives will now depend on the credit risk measurement approach adopted. In the standardised approach the credit rating of the underlying exposure is substituted by that of the guarantee or credit derivative. 21 In the foundation IRB the LGD is adjusted for these instruments according to supervisory values, whereas under the advanced IRB approach banks use their own internal ratings for LGD. 22 Netting arrangements are recognised in the new Accord as a credit risk mitigation technique. In these arrangements in which banks agree to offset cash flows or other obligations against each other capital requirements can be calculated on the basis of net credit exposures BIS (2005; Page 31, paragraph 145) 19 BIS (2005; Page 28, paragraph 121) 20 BIS (2005; Page 44, paragraph 188) 21 BIS (2005; Page 30 31, paragraph 141) 22 BIS (2005; Page 66, paragraph 300) 23 BIS (2005; Page 41, paragraph 188) 20

21 Securitization framework 24 The securitization framework is the part of Basel II that extensively covers the treatment of securitization exposures. These exposures include those coming from the investment in a securitised instrument as well as the risk retained from the origination of an asset securitization. Securitisation exposures treated include the following: investing in asset-backed securities or mortgage-backed securities, providing credit enhancements or liquidity facilities. The prudent treatment of securitization is something which the Committee sees as very important: The Committee believes that it is essential for the New Accord to include a robust treatment of securitisation. Otherwise the new framework would remain vulnerable to capital arbitrage, as some securitisations have enabled banks under the current Accord to avoid maintaining capital commensurate with the risks to which they are exposed. To address this concern, Basel II requires banks to look to the economic substance of a securitisation transaction when determining the appropriate capital requirement in both the standardised and IRB treatments. 25 So in order to deal with the fact that securitized assets can carry considerable risk, the new Accord includes a more risk-sensitive approach especially to deal with the capital arbitrage incentives under the current Accord. Therefore, much alike the treatment of credit risk, the framework introduces risk-sensitivity through the design of a standardised approach and ratings based approach. 1)The standardised approach Banks using the standardised approach for credit risk measurement also have to adopt the standardised approach in the securitisation framework. In this approach securitisation exposures are given a risk weight in accordance to ratings from credit rating agencies. The risk-sensitivity is quite strong. For example, between BB+ and BB- the risk weight is as high as 350% and the Accord prescribes a deduction from capital for positions with long-term ratings of B+ and below. 2)The ratings based approach If banks measure their credit risk under the IRB approach, then they have to do likewise in the securitisation framework. Namely by adopting the ratings based approach (RBA). Under this approach banks have to either risk weight their exposures according to external ratings or (if not available) infer the rating from another securitisation which can reasonably act as a reference. These risk weights that the Committee has prescribed in the RBA are even more risk-sensitive than those under the standardised approach: varying from 7% for AAA to 650% for BB-. A design which must give an incentive towards creating or investing in securitisation structures of high credit quality. If no external rating can be found 24 BIS (2005; Page , paragraph ) 25 BIS (2003; Page 7, paragraph 35) 21

22 and no rating can be inferred, then instead of adopting the RBA a supervisory formula can be used for assigning capital requirements Operational risk 26 Bankruptcies occur due to reasons as unexpected legal costs or administrational errors. Both of which are examples of operational risk, a risk to be separately treated under Basel II. With activities of financial institutions becoming more complex (think of financial innovation, globalisation, new technologies) the necessity of more accurately addressing this risk has grown. On average operational risk provisions constitute 20% of internal capital of many major banks. 27 As had been mentioned in figure 1, the new Accord has three approaches to operational risk, namely: the basic indicator-, standardised- and advanced measurement approach or AMA. Each of which I will shortly describe. 1)The basic indicator approach The basic indicator approach is the least advanced approach designed for banks that are less sophisticated. This option is available for banks that adopt the standardised approach to credit risk, or the easy-access simplified standardised approach. The capital required under the basic indicator approach is simply 15% of the bank s gross income. 2)The standardised approach The standardised approach to operational risk is a bit more advanced and is meant for the slightly more sophisticated banks, specifically those that also adopt the standardised approach for credit risk. In calculating capital requirements it makes a difference between business lines, each represented by a different beta which is some sort of risk weight. The capital required for each business line can be calculated by multiplying its beta by its gross income. In total there are 8 business lines, some of them receiving a beta of 18%, like corporate finance, and others receiving a beta of 15% or even 12%, like retail banking. 3)The advanced measurement approach As soon as a bank has transgressed to the IRB approach in credit risk measurement, it is expected to adopt the advanced measurement approach (or AMA). In the AMA capital required is calculated by the use of banks own risk measurement systems. As an input to these systems use is to be made of data on past internal losses. Because of this the Committee has set as one of the prerequisites for advancing to the AMA 26 See BIS (2005; Page ) 27 BIS (2001-A; Page 4) 22

23 the availability of sufficient data resources. The Committee however does acknowledge the difficulty of measuring operational risk, since the topic is still rather unexplored in the literature Pillar 2 Supervisory review 29 The second pillar covers the supervisory review process, which aims to provide supervisors with instructions on how to best oppose inadequate risk management by banks. The second pillar proposes four key principles that should be incorporated into supervisory practices. These are the following: 30 Principle 1: Banks should have a process for assessing their overall capital adequacy in relation to their risk profile and a strategy for maintaining their capital adequacy Under this principle is advised that banks should be required to demonstrate to their supervisors that they hold enough capital in proportion to the risk on (and off) their balance sheet. Also they should be expected to undertake stress tests in order to maintain capital adequacy even in the face of deteriorating business conditions. Principle 2: Supervisors should review and evaluate banks internal capital adequacy assessments and strategies, as well as their ability to monitor and ensure their compliance with regulatory capital ratios. Supervisors should take appropriate supervisory action if they are not satisfied with the result of this process In this principle is stressed that supervisors should check on whether banks employ the right risk assessments methods for setting their capital. Certainly in relation to the fact that the new Accord allows banks to use internal ratings systems for calculating required capital, it is especially important that supervisors evaluate whether banks assess and manage their risk prudently. Action undertaken to correct imprudent risk measurement can hereby be an invaluable tool. Principle 3: Supervisors should expect banks to operate above the minimum regulatory capital ratios and should have the ability to require banks to hold capital in excess of the minimum So in addition to the pillar 1 capital requirements which are rather mechanical in nature, banks should also be expected to hold buffers above the minimum capital requirements. These buffers can then better take into account bank-specific uncertainties. 28 BIS (2001-A; Page 4) 29 See BIS (2005; Page ) 30 See BIS (2005; Page ) 23

24 Principle 4: Supervisors should seek to intervene at an early stage to prevent capital from falling below the minimum levels required to support the risk characteristics of a particular bank and should require rapid remedial action if capital is not maintained or restored Under this principle the supervisor is advised to intervene by demanding higher amounts of capital as soon as risk mismanagement turns out to be too poor. The Committee stresses however that such an intervention usually doesn t solve the problem, as addressing the underlying problem of poor risk management is more important for the long term. Further, in addition to adhering to these four principles, banks and supervisors should specifically take into account two risks that remain largely untreated under the first pillar. These are: 1) credit concentration risk and; 2) interest rate risk in the banking book. Both of which I present in the following two paragraphs 1) Credit concentration risk The danger of having a concentrated set of exposures is clear: the default probability of one big exposure is larger than the probability of a simultaneous default of many smaller ones. The Committee says the following on credit concentration risk: A risk concentration is any single exposure or group of exposures with the potential to produce losses large enough to threaten a bank s health or its core operation. Risk concentrations are arguably the single most important cause of major problems in banks. 31 Due to its importance, the Committee wants to address this type of risk under pillar 2. It advises supervisors to make banks perform stress tests, especially with the aim of discovering whether credit concentrations could cause vulnerabilities in a downturn. If banks don t address this risk prudently supervisors are urged to demand a higher capital level. 2) Interest rate rate risk in the banking book Interest rate risk is risk arising from changing interest rates. A well-known way through which such vulnerability to interest rate can arise is spread risk. In a situation of lending against a fixed interest rate and financing those loans at a flexible rate, spread risk is the risk of having declining spreads due to rising interest rates. Although the Committee recognizes this risk to be important it has not chosen to treat it under pillar 1. As they say, a major difficulty in prescribing a set of uniform minimum requirements for interest rate risk is the fact that the risk is by its nature perceived to be very heterogenous (i.e. sensitivity to interest 31 BIS (2005; Page 172, paragraph 770) 24

25 rates is very different across banks). The Committee however does see an option for countries in which interest rate risk is homogenous to establish a set of minimum capital requirements covering this risk. But foremostly supervisors should expect banks to maintain internal systems that are capable of accurate interest rate risk assessment. If banks don t manage exposure to this risk prudently, then supervisors should intervene by demanding a higher capital level Pillar 3 Market discipline 32 The idea of market discipline is that market participants are induced to protect their money and therefore demand higher risk premiums from banks that are handling it imprudently. If this mechanism works then banks can be disciplined for imprudent risk management. However for it to work well, banks must disclose adequately the information on their practices. For this reason the Committee has created the third pillar, which sets disclosure requirements. This intention is clear from the following statement: The Committee aims to encourage market discipline by developing a set of disclosure requirements which will allow market participants to assess key pieces of information on the scope of application, capital, risk exposures, risk assessment processes, and hence the capital adequacy of the institution 33 Clearly, if market participants get a better picture of risk-taking by banks, they will be more capable of assessing capital adequacy, and therefore better able to engage into market discipline. Effectively this implies that if bank disclosure indicates a poor risk management and poor capital adequacy, it is likely that depositors start demanding a relatively high risk premium. This market discipline could coerce banks to keep higher capital levels even considerably above required minimum capital in order to increase solvability. By stimulating market discipline, pillar 3 is meant to be complementary to the other two pillars. Especially regarding the large extent of freedom granted to banks under the internal rating approaches, scrutiny by market participants is thought to be crucial. The Committee mentions that market discipline is significantly enhanced due to the fact that the disclosures required of banks will now be according to the common structure of capital requirements of pillar 1. The idea is that such an arrangement that would make disclosures of different banks easier to compare and understand for market participants. 32 See BIS (2005; Page ) 33 BIS (2005; Page 184, paragraph 809) 25

26 Advantages Basel II The great advantage of Basel II is that it greatly alleviates the weaknesses of Basel I. First of all the discrepancy between economic capital and regulatory capital is reduced significantly, due to that the regulatory requirements will rely on banks own risk methods. Second, the scope for capital arbitrage will be greatly diminished as the new Accord is risk sensitive, including the treatment of securitization exposures. Third Basel gives wider recognition of credit risk mitigation. Other advantages are the guidance offered for supervisory review under pillar 2 and the encouragement of market discipline under pillar Disadvantages Basel II Next to underwriting the well known advantages mentioned in the previous section, the biggest part of the literature on the new Accord is about its potential disadvantages. Critique on Basel II is therefore widespread and easy to find, a more challenging task however is seeking out which of the concerns are most serious. Important concerns that relate to my paper are the following: There is a scope for enhanced procyclicality, as the risk sensitive capital requirements of Basel II can substantially vary over the cycle. 35 The new Accord is complex and therefore demanding for supervisors, and unsophisticated banks. 36 Strong risk differentiation in the new Accord can adversely affect the borrowing position of risky borrowers. 37 Each of which are especially relevant for the borrowing position of emerging economies for the following reasons respectively: Emerging economies already have large business cycles, volatile capital flows, and carry risk of financial crisis. If Basel II works in a procyclical manner, this may especially impact these countries. 38 Supervisors and banks in these countries tend to be somewhat unsophisticated, making it difficult and unwise for them to fully implement it Dierick et al (2005; Page 8) 35 Dierick et al (2005; Page 29) 36 Majnoni, Powell (2004; Page 20 21) 37 Tanaka (2003; Page 224) 38 Claessens et al (2003) 39 Majnoni, Powell (2004; Page 20 21) 26

27 These economies usually have low sovereign ratings, just as the companies and banks residing within them. 40 Banks subject to the risk-sensitive Basel II Accord will therefore face higher capital requirements when lending to these countries, which can result in higher lending rates. 41 I will extensively present these problems in the following chapters Conclusionary remarks The Basel I capital Accord has with its widespread adoption established a well-recognized capital standard which is likely to have enhanced competitive equality between international banks, and improved bank solvability. However, clear disadvantages have been recognised which give considerable scope for improvement. In the near future Basel I is planned to be replaced by Basel II, which is an Accord that tries to overcome these disadvantages, making capital regulation more advanced. The new Accord with its three pillars is much more extensive and complex than the current Accord. A significant change is in the area of credit risk measurement, where, under the first pillar of Basel II, the IRB approach brings capital regulation closer to the risk management practices that banks have been using over the years. To support these changes, the important role of the supervisory review process as well as that of efficient market discipline have been addressed respectively under the second and third pillar. The new Accord also incorporates more advanced treatment of credit risk mitigation, securitization exposures and introduces a separate capital charge for operational risk. What has come clear is that the changes brought by the new Accord are welcomed on the one hand, but also heavily criticised on the other. The concerns that I deal with in my paper are all important for the borrowing position of emerging economies. In the following chapter I will start discussing the problem of procyclicality. 40 Tanaka (2003; Page 224) 41 Liebig et al (2004; p 1) 27

28 2. Procyclicality of Basel II The shift to Basel II is often suggested to bring about an increase of procyclicality. First section 2.1 discusses procyclicality in the financial system and how Basel II may increase this procyclicality. After that section 2.2 discusses how measurement of risk components can determine the volatility of IRB capital requirements. Section 2.3 discusses the empirical literature on estimated capital requirement volatility under Basel II. Finally, section 2.4 gives some conclusionary remarks Risk sensitive capital requirements and the credit crunch The financial system already is considered to be of procyclical nature due to a mechanism referred to as the financial accelerator, which focuses on information asymmetries between borrowers and lenders. 42 The idea namely is that as the economy goes into a downturn and prospects of credit risk increase, these information asymmetries become much larger with the result of inhibiting the lending process. Hereby this tendency is reinforced by the procyclical behaviour of asset prices, which provides good collateral in booms but makes sure that in downturns collateral loses its value. The procyclical result then of the financial accelerator effect is that economic activity is stimulated in a boom by easier lending and discouraged in a downturn as lending becomes relatively difficult. In addition, during recession the capability of banks to lend usually declines due to a lower interest income and an increased rate of loan defaults. Moreover, severe losses from defaulted loans can affect the bank s profitability, potentially even to the point that solvability is endangered with the result that capital reserves first have to be replenished before taking on new loans. 43 Nonetheless, a common subject in the literature is that Basel II may potentially lead to banks behaving in an extra procyclical manner. The idea namely is that if banks determine their capital requirements by their own assessments of risk, there is the potential for capital requirements to fluctuate over the cycle. This may be the case when estimated risk in a boom turns out to be relatively low and in a downturn relatively high, so that the risk-sensitive Basel II regulation pushes down capital requirements in a boom and upwards in a recession. The procyclical effect will then come from the fact that low capital requirements helps to expand the credit supply, and high capital requirements may decrease the credit supply. Jackson et al (1999; Page 16) explain this reaction in the credit supply as follows: During booms, banks will find it easy to raise equity capital and potential earnings retentions will be high. During downturns, with declines in loan demand and increased default risks, banks may prefer to cut back their loan base. So in a boom period banks will find less difficulty expanding their credit supply due to that extra capital can be easily raised and decreasing capital requirements release a part of the capital stock. On the other hand in a downturn poorly- 42 See for example Borio et al (2001) 43 See subsection for a further discussion on this point 28

29 capitalized banks are somewhat likely to have to contract their credit supply in order to meet the increasing capital requirements. These procyclical changes in the credit supply caused by the capital regulation will then bring extra reinforcement to the business cycle. Stimulating consumption and investment in the boom and discouraging it in the downturn. A well known danger is that such a procyclical effect of capital regulation can hit the economy especially hard on the downturn. Particularly if a large part of the economy is dependent on bank lending, a contraction in bank credit supply may lead to a so-called credit crunch. Which is a situation that can develop even in spite of expansionary monetary policy. Banks may namely be very willing to increase their credit supply but unable to do so because they are likely to be constrained by capital requirements. As lowering interest rates brings no relief, the usefulness of monetary policy to play a role in stimulating the economy in times of recession therewith seems to be much reduced. Bliss & Kaufman (2003; Page 9) for this reason argue: if, in recessions, banks cannot raise new capital at favorable prices, the only direct tool the Fed has to remove a binding capital constraint and encourage increases in bank credit and deposits is to lower the regulatory capital requirement. On the other hand, however, such a solution is no panacea as lowering required capital could decrease bank solvability. Since that the procyclicality of the Basel II capital requirements effectively depends on the way in which the risk components PD, LGD, LGD, M are set, it is interesting to look at the methods employed by banks to arrive at estimates for these variables. Hereby it is the crucial question how these components are influenced by the movement of the business cycle. For example it is likely to see a low PD at the peak of the cycle, and a higher one at the bottom. Nonetheless, it is also possible that banks specifically aim to set PD in such a manner that it doesn t have to fluctuate too much over the cycle, and therefore giving less fuel to procyclical capital requirements. Therefore in the following section I would like to discuss the most widespread methods used by banks for setting these risk components Measurement of the risk components and procyclicality In this section I discuss the factors that determine the capability of IRB capital requirements to be volatile over time. In the following subsections I will discuss the relationship between the building blocks of risk measurement and the business cycle. First in the PD and the business cycle, in the correlation between PDs across borrowers and the business cycle, and lastly in the LGD and the business cycle PD and the business cycle Lowe (2002; Page 5) broadly divides ratings systems for determining PD into two categories: 1) those that derive their ratings solely from market-based information about firms, usually equity prices; 2) those that 29

30 use a more ample set of information, particularly with the intention of recognizing the risk of deteriorating macroeconomic circumstances. If based on market-based information, the rating process goes as follows: for a given firm the PD (and thus implicitly the rating) is a decreasing function of the firm s equity price, and an increasing function of the volatility of the equity price and the firm s leverage. 44 An obvious downside of using market-based information is that as equity prices tend to be rather volatile and sometimes even out of line with fundamentals (either being overvalued or undervalued) it is not unreasonable to argue that the ratings based upon them will also fluctuate in an unexpected manner. 45 In contrast to the first type, the ratings set by the second type of rating systems tend not to heavily correlate with the business cycle. Instead, they are specifically designed to be stable throughout the cycle, which is generally thought of as being achieved by estimating PDs in case of a downturn scenario. Therefore, if a downturn happens a rating founded on this principle will not have to undergo a sudden decrease the benefit of which is that it is unlikely to bring about a procyclical effect. Due to that it is characterised by its stability, the Committee is strongly endorsing such a through the cycle approach, as it appears from the revised framework: A borrower rating must represent the bank s assessment of the borrower s ability and willingess to contractually perform despite adverse economic conditions or the occurrence of unexpected events. For example, a bank may base rating assessments on specific, appropriate stress scenarios. 46 Observe the endorsement of basing assessments on a stress scenario which is, as mentioned, an earmark of the through the cycle approach. According to Lowe (2002; Page 6) most banks neither entirely apply a market-based rating philosophy nor a through the cycle one, but rather something in between. What most banks have in common with the market-based philosophy is a 1 year rating horizon. However, a key difference is that internal rating models and internal analysts are depended upon, instead of just relying on equity prices. The procyclical tendency of which then crucially depends on the degree in which the bank s analysts take into account the potential build-up of financial imbalances and assess the future movement of the business cycle in a cautious manner. If not, the PD would have the capability of being internally rated much lower in the downturn than in the upturn, with the result that the risk-sensitive IRB approach pushes up capital requirements in a recession and downwards in a boom. 44 Lowe (2002; Page 5) 45 Lowe (2002; Page 5) 46 BIS (2005; Page 89, paragraph 415) 30

31 The correlation between PDs across borrowers over the business cycle While discussing 47 the IRB formula, I mentioned the fact that it includes a correlation factor. Which is a coefficient designed to correct for the correlation between probabilities of default across borrowers. A potential weak point however is that this correlation factor is static over the business cycle, whereas correlation across borrowers is actually known to vary over the business cycle in a procyclical manner. Or, as put by Lowe (2002): there is some evidence that during periods of financial stress asset correlations tend to increase. The consequence then of such an increase in correlation, is that actual risk becomes higher than the amount of risk accounted for by the IRB formula. To adjust for that, the correlation factor would have to rise in periods of financial stress, leading capital requirements to a temporarily higher level. The obvious drawback of course is that this would lead to higher capital requirement volatility LGD and the business cycle Although the IRB formula constantly assumes a LGD of 45% under the foundation approach for all unsecured loans, the advanced approach would allow banks to calculate required capital using its own estimations of LGD. The literature seems to suggest that LGD should to some degree be expected to correspond to the business cycle in a procyclical manner. Losses namely tend to be higher during recessionary periods when both asset prices and collateral are depressed, and lower in a boom when asset prices and collateral values tend to be amplified. 48 Moreover a substantial amount of empirical literature indicates that the PD and LGD of exposures are positively correlated. 49 This correlation is normally thought of as being the result of a common systematic factor that causes both the PD as well as the LGD to increase during a recession, and vice versa during a boom. The result of which is that credit risk during a recession gets extra large due to a cojoint increase of PD and LGD. To assess the procyclical effect that the correlation between PD and LGD may have, Altman et al (2002) simulate how lending behaviour would change if this correlation were to be included in the calculation of capital requirements see figure 3. The first scenario assumes away any correlation, as the LGD is fixed at 50% over the business cycle. In the second scenario correlation is incorporated by letting LGD fluctuate in conjunction with the default rate it will be closer to 60% in a high default year and closer to 40% in a low default year. For a simulated loan portfolio both scenarios are applied to calculate IRB capital requirements, which are then used to derive lending behaviour. The authors namely derive this behaviour by assuming that an increasing level of capital requirements leads to a decrease in lending, and vice versa. Also, they take into account changes in the capital ratio such as losses due to defaulted loans, the incurring 47 See paragraph Allen, Saunders (2003; Page 19) 49 See Altman et al (2002) 31

32 of which directly lowers the capital ratio so that lending needs to decrease. 50 The outcome of the simulation points to considerable relevance of the PD-LGD correlation: when we let LGDs free to fluctuate with default rates, the procyclicality effect increases significantly. An effect which is clearly shown in figure 3, where the pattern associated with the second scenario is much more variable than that of the first. Figure 3 Procyclical effect of a PD-LGD correlation Source: Altman et al (2003; Page 39) Although the authors express their concern about what these findings may imply for procyclicality of Basel II, they also mention the following: Of course, one might object that the bank in our simulation behaves in a somewhat myopic way, and that regulation should encourage advanced IRB systems to use long-term average recovery rates. Herewith is pointed out that banks may adopt a longer term perspective so that LGD will be relatively stable over the cycle, with the consequence that it is less likely to be a factor in increasing capital requirement volatility. An argument which can be underlined by the fact that under the advanced IRB approach banks are in fact discouraged to make estimates of LGD over time in a variable manner. The revised framework namely says the following: This LGD cannot be less than the long-run default-weighted average loss rate given default calculated based on the average economic loss of all observed defaults within the data source for that type of facility. So the Accord actually obliges banks to not go lower than the default-weighted average for calculating the LGD from historical data. And in addition to that, the Accord requires banks to set LGD up to a level that would be adequate even in case of severe downturn: In addition, a bank must take into account the potential for the LGD of the facility to be higher than the default-weighted average during a period when credit losses are substantially higher than average. 51 Clear herefrom is the intention that banks should not determine the LGDs relating to their 50 Clearly they don t see the possibility of raising new capital as a viable alternative 51 BIS (2005; Page 99, paragraph 468) 32

33 exposures in a variable manner. If indeed LGDs remain relatively fixed over the cycle, it seems unlikely that any relationship between LGD and the business cycle would raise the volatility of capital requirements Estimated procyclicality of capital requirements As the previous section has indicated, there is some propensity for capital requirements to be determined in a procyclical manner. First, subsection will look at various empirical studies that simulate the movement of capital requirement over time. Each of which apply the IRB rule on historical credit data, therewith arriving at counterfactual capital requirements counterfactual since these capital requirements were never really there. These estimations can give a good picture of the potential procyclical impact of Basel II. However, in several reasons will be discussed why changes in capital requirements may not have to lead to procyclicality Empirical studies This section presents several of the empirical studies performed on procyclicality of Basel II. First, paragraph starts off with an overview of the empirical literature, as presented in Kashyap & Stein (2004). Second, paragraph will present a simulation done by Kashyap & Stein (2004) themselves. Third, in paragraph two reasons will be given why the various simulations of IRB capital requirements may miss some accuracy Overview of estimations As a reaction on the proposals for the new Accord a considerable amount of research has emerged on the procyclicality of capital requirements. Many of these studies have tried to simulate how capital requirements would have evolved over a recent time period if Basel II had been in place. The counterfactual capital requirements herein are being calculated by looking at external or internal ratings that correspond to a certain sample of loans that is usually fixed. In an attempt to give an overview of the different simulations Kashyap & Stein (2004) made table 4. In this table it is clear in how many ways these studies are different from each other for example they focus on different countries and different periods. Also, as is shown in the fourth column, they use different variants of the IRB formula for calculating capital requirements something which reflects the fact that the proposals for the new framework have been altered over the years. But even though the studies are quite different, it is still a bit surprising at first glance to see that the results are very dissimilar. The fifth column which depicts the maximum percentual change in simulated capital requirements over the sample period namely shows a very wide variance between the studies. Clearly, the 33

34 result by Jordan et al (2003) that IRB capital requirements can change with 280% seems to be a world apart from the result by Corcostegui et al (2002) which only arrive at a maximum change of 6.1%. In order to explain this diversity of results Kashyap & Stein (2004) point to differences in methodology and dataset. Table 4 Summary of research on volatility of capital requirements Source: Kashyap & Stein (2004; Page 27) Simulation by Kashyap & Stein Kashyap & Stein (2004) also do their own empirical study on volatility of IRB capital requirements. In this study they simulate how IRB capital capital requirements on corporate exposures would have evolved from 34

35 1998 until 2002, which is a period characterised by diminishing economic perspectives world-wide. 52 Two different datasets are used for calculating capital requirements: one consisting of external credit ratings from S&P, the other of ratings that are acquired with the KMV model, which is a model that is based on equity prices. The first type is thought of as a through the cycle philosophy, whereas the second is a market based methodology and therefore more likely to be volatile over the business cycle. 53 A problem largely left untreated in the previous literature, but extensively addressed by Kashyap & Stein (2004) is that of survivorship bias, which depicts the fact that usually in a sample of loans a proportion simply disappears over time due to unkown reasons. 54 Due to that disappearing loans usually are relatively weak and thus prone to worse-than-average performance, it is not unreasonable to believe that simply forgetting about these loans would produce an estimation of capital requirement volatility that is too low. The authors argue that the most reasonable way to deal with this is to consider how such a loan would have performed if it would not have disappeared. They do this by first looking at how the average loan from the same geographical area and the same rating class moves from the moment of disappearance onwards. Subsequently, these average movements are then used as an approximation of how a disappeared loan would have been likely to move, so that a reasonable estimate of the appropriate capital requirement can be made. The results of the simulations are shown in table 5, which feature the percentages by which IRB capital requirements on investment grade 55 exposures would have risen over the period The first column is for the S&P dataset and the second for the KMV dataset. Table 5 Volatility of capital requirements Rating class % change cap req S&P ratings KMV ratings Investment grade Source: Kashyap & Stein (2004; Page 25-26) Judging from the results it can be said that capital requirements based on S&P ratings can be quite volatile. The estimated change in capital requirements is namely around 40%, which according to the authors could mean that Basel II-IRB brings forth a level of procyclicality nearly twice that of under the current Accord to the extent that there have previously been concerns expressed about capital crunches during recessions, the change from Basel I to Basel II might be expected to very loosely speaking almost double the impetus behind these episodes. 56 Note herein that the volatility under the current Accord is seen as coming 52 Kashyap, Stein (Page 22) 53 See subsection They also experience this in their own S&P dataset in which 542 loans from the initial 3599 disappear over the sample period. 55 Investment grade is BBB- and higher 56 Kashyap & Stein (2004; Page 25) 35

36 from the effect of loan defaults on capital levels. Namely due to that losses from loan defaults are negatively correlated to the business cycle and can ultimately deplete capital levels, it may cause banks to become capital inadequate during recession also when the capital charge itself is constant as under Basel I. As table 5 shows, it is clear that a KMV rating system has an even stronger potential to lead to volatility of capital requirements, as the increase in counterfactual capital requirements over the period is 82.5%. A figure nearly twice as high as the one relating to the S&P dataset, which is 43.7%. Comparing the two differently-rated samples therefore, seems to suggest that a market-based approach such as the KMV system would lead to a substantially higher amount of volatility in ratings than a through the cycle approach, such as that employed by S&P. Something which seems to be broadly in line with the way both approaches can be characterised. 57 However, seen from the fact that the S&P dataset still produces capital requirements of considerable volatility, it could be argued that a through the cycle approach is unlikely to be a very strong guarantee for rating stability Accuracy of these estimations What the numerous simulations discussed sofar suggest is that the changeover to Basel II-IRB is likely to have serious implications for volatility of capital requirements. However, as pointed out in the literature, there are several drawbacks to these studies. In this section two of them will be identified and discussed, namely: 1) the lack of attention to the situation under the advanced IRB approach, 2) and the Lucas critique. I will treat both of them in the following paragraphs. 1) Neglecting the advanced IRB approach As the simulations discussed in the previous paragraphs have all been calculated using the foundation IRB approach it doesn t provide a good picture of what capital requirement volatility would be under the advanced IRB approach. The estimations for the advanced IRB approach could namely be very different, since this approach also incorporates banks assessments of LGD and M. The first of which, as discussed in subsection 2.2.3, tends to move in a procyclical manner over the business cycle. It therefore may have a reinforcing impact on capital requirement volatiltiy. 58 On the other hand, Gordy & Howells (2004; Page 7) argue, the impact of maturity may be the opposite as this variable tends to move in a countercyclical manner i.e. higher in boom and lower in a downturn. Note hereby that the IRB formula treats loans of a lower maturity more favorably. Although it seems that the literature is rather ambiguous on whether the advanced approach would lead to higher procyclicality than the foundation variant, the question is obviously highly relevant as the advanced 57 See subsection Kashyap & Stein (2004; Page 30) 36

37 IRB approach is expected to be implemented by many of the largest international banks. To illustrate, banks in the USA that are going to adopt the IRB approach are obliged to choose for advanced IRB. 59 2) Lucas critique Another point of controversy regarding the simulations in this chapter is that they are built on data gathered under the current Accord, while the situation under the new Accord may be very different also known as the Lucas critique. Lowe (2002; Page 1) says the following on this point: the evidence that does exist is subject to the Lucas critique, namely that structural changes that are likely to occur after the implementation of risk-based capital requirements mean that evidence from the current regime says little about the future. Herein is suggested that the Lucas critique is especially valid due to that the new Accord will likely bring about structural changes in risk measurement. Borio (2003; Page 15) gives two reasons why risk measurement should structurally improve as the regime changes. First he mentions the following: the Accord is helping to spread and hardwire an historic improvement in risk measurement and management culture. The level of the debate has risen immensely over the last couple of years. And awareness of the potential adverse implications of unduly procyclical risk assessments has risen pari passu, both among market participants and supervisors. More generally, better risk management means that problems can be identified and corrected earlier. Herewith the author points to the positive attention given by Basel II that risk is to be assessed and managed in a careful, non-procyclical, manner. The fact that a through the cycle methodology is endorsed 60 underlines this argument. In his second reason, the author further stresses how supervisors and markets can reinforce this prudency: Pillars 2 and 3 can underpin this shift. Greater disclosure means that markets may become less tolerant and more suspicious of risk assessments that move a lot over time and lead to substantial upgrades in good times. And supervisors, if they so wished, could rely on the strengthened supervisory review powers to induce greater prudence in risk assessments. So due to stronger supervisory powers (pillar 2) and greater disclosure (pillar 3) chances are increased that supervisors and markets will discipline banks early on for the build-up of financial imbalances or for a procyclical assessment of risk in general. If the introduction of Basel II will indeed cause risk to be treated in a less procyclical manner, it is reasonable to assume that the procyclicality of the simulated capital requirements in the previous paragraphs are somewhat overstated as they are based on ratings under the current Accord Reasons why changes in capital requirements will not lead to procyclicality Further it is important to realize that there are several points of criticism which challenge the idea that changing IRB capital requirements would make an impact on the real economy. Namely the following: 1) the fact that banks tend to operate with a capital buffer, 2) the fact that financial markets and unconstrained 59 Finansinspektionen (2006; Page 5) 60 See subsection

38 institutions can soften an impact on the credit supply and; 3) the fact that the economy is also demand determined instead of just supply determined. I will present these issues in the following three sections, before presenting the various empirical studies The buffer stock channel 61 The procyclical impact of the new Accord hinges on the assumption that a changing level of capital requirements affects bank lending. Which effectively implies that as soon as decreasing capital requirements releases capital, it would to some extent have to be used to expand its credit supply, and vice versa with increasing capital requirements. An important element that can weaken this link between capital requirements and bank lending is the fact that banks normally manage their actual economic capital level somewhere above the level of regulatory capital. The development of the economic capital buffer economic capital minus regulatory capital can namely be an important determinant of cyclicality in bank lending. As banks may have the willingness to keep sizable buffers during good periods and consume those buffers no earlier than in a recession, procyclicality of Basel II can be greatly softened. On the other hand this buffer stock channel may also amplify the impact of procyclical capital requirements, as banks can instead choose to increase its buffer exactly during downturns. The following example will clarify how both effects can take place. Example 2 The buffer stock channel A bank has a $100 loan portfolio, on which an IRB capital requirement must be held of $5 in boom periods and $10 in recession when the portfolio is being downgraded or faces default losses. There are three broad ways in which the economic capital buffer can be managed, as exemplified in the following scenarios: 1. It can remain neutral being of no impact on procyclicality by remaining of the same magnitude over the cycle. An example of which would be that in boom periods the economic capital buffer would be $4, and remaining that during recession. This would result in a total capital level of $9 ($5 + $4) during booms and $14 ($10 + $4) during recession. 2. It can dampen procyclicality by being larger in boom periods and partly or wholly consumed during recession. An example of which would be that in boom periods the economic capital buffer would be $5, and decreased to $3 during recession. Which would result in a total capital level of $10 ($5 + $5) during boom and $13 ($10 + $3) during recession. This decrease in economic capital buffer partly offsets the increase in regulatory capital, and therefore procyclicality is dampened. 3. It can increase procyclicality by being relatively small in boom periods and increasing during recessions. An example of which would be that in boom periods the economic capital buffer would be $3, and increased to $5 during recession. Which would result in a total capital level of $8 ($ See Jacques (2005) 38

39 $3) during boom and $15 ($10 + $5) during recession. This increase in economic capital buffer adds up to the increase in capital requirements, leading to extra procyclicality. Clearly, the effect of the buffer stock channel depends on how banks manage their economic capital buffer. As it is yet unknown how the buffer will be managed under Basel II it may be indicative to look at the experience under Basel I. Ayuso et al (2002) and Lindvquist (2003) are two studies that do this for the case of Spain and Norway respectively. What they find is that the economic capital buffer behaves over the cycle in a way that adds to the procyclicality of Basel I capital requirements. If we assume that this behaviour is widespread and that it will not improve under Basel II, the forthcoming situation could become much like that of the third scenario in the above example. Nevertheless, the evidence from under the current Accord seems to be of only minor value as the Lucas critique is applicable here. The new Accord may namely bring structural changes, which improve the way of managing economic capital. Especially considering the focus of the new Accord on holding an economic capital buffer large enough to be able to withstand downturn scenarios, it is perhaps more realistic to expect the capital buffer to be managed in a manner that counters procyclicality. 62 Altogether it seems that the literature has not reached a consensus on the future role played by the buffer stock channel. Therefore it is hard to get an exact picture whether actual capital levels will be more or less cyclical than the levels required by Basel II Alternative sources of credit Another reason why changing capital requirements may leave the real economy relatively unaffected, is due to the possibility that financial markets and other unconstrained institutions offset changes in bank lending. Jackson et al (1999; Page 27) namely say the following: Even if some banks do cut back lending because of capital constraints, for this to affect the real economy the reduction must, for some reason, not be fully offset through increased lending either by better capitalised banks or by other financial intermediaries or by credit markets. So if demand for credit can be easily met by lending from other sources, then the procyclical effect of Basel II on bank lending is likely to be of little influence on the total credit supply. However, the authors seriously doubt whether any dearth of bank lending can be easily offset as lending is usually characterised by asymmetric information, which is a problem that banks are better able to overcome with their expertise in assessing borrowers. Therefore, Jackson et al (1999; Page 27) argue: if information imperfections are important, then the inability of banks to perform these activities would disrupt the flow of credit. The result of asymmetric information would then be that financial markets are incapable of offsetting any decrease in bank lending, so that a contraction in the total credit supply will not be prevented as well as a possible impact on the real economy. 62 Borio (2003; Page 15). Also see paragraph

40 Bank lending is also demand determined As a third reason why changes in capital requirements are unlikely to strongly affect the real economy is the fact that fluctuations in bank lending are not only explained by supply factors, but also by demand factors. Which is a difference in perspective that is reflected in two different theories of monetary policy. 63 Namely on the one hand there is the bank lending channel theory that argues that monetary policy affects the economy via the supply of bank credit. This would be the channel responsible for a contraction in the credit supply that reduces bank lending as soon as capital requirements go up. On the other hand there is the balance sheet channel theory, which argues that changes in the net worth of firms (due to e.g. changing profits or changing asset prices) can affect the demand for credit and cause and/or amplify economic shocks. A study by Bikker & Hu (2001; Page 17) points to a greater importance of the balance sheet channel theory: These outcomes underline findings in the empirical literature that the influence of demand factors dominates the market and that barring exceptional circumstances it is fairly difficult to observe demonstrably supply-driven credit crunch effects. So due to that the effect of supply factors is relatively small in comparison to that of demand factors, changes in the credit supply, such as Basel II, will be less likely to cause a credit crunch. The authors therefore conclude that the threat of Basel II is rather small: Thus, while procyclicality of banks behaviour and the perceived increased procyclicality caused by the new Capital Accord are genuine problems, they are unlikely to have more than a minor effect on macroeconomic stability. 64 Which puts the worries about potential procyclicality of Basel II in a less alarming perspective 2.4. Conclusionary remarks As procyclicality can potentially increase after the switch to the more risk-sensitive Basel II Accord, it is relevant to look at whether risk components of the IRB formula will be set in a procyclical manner or not. In the measurement of the probability of default a market based rating methodology would lead to ratings of a higher volatility than a through the cycle methodology. Due to that banks tend to apply rating systems that are somewhere in between the two extremes, it is difficult to get a good idea of how stable their PD assessments will be under Basel II. Both the cyclically-changing correlation of PD across borrowers as well as the movement of LGD over the business cycle is generally not thought to have a procyclical impact under Basel II. Moreover, correlation is assumed fixed in the IRB formula, and the freedom to determine LGD in a variable manner is somewhat restricted. 63 See Bikker & Hu (2001) 64 Bikker & Hu (2001; Page 4) 40

41 The overall result coming from volatility simulations in the literature is that the level of procyclicality has the potential to significantly increase with the switchover to Basel II. Nonetheless, dataset and methodology can make a big difference. Having a through-the-cycle rating system is likely to lead to a much smaller amount of procyclicality than its market-based counterpart. Drawbacks to these simulations are that they don t give attention to the situation under the advanced IRB approach, and that they are vulnerable to the Lucas critique. Further, it is important to realize that there are several points of criticism which challenge the idea that changing IRB capital requirements would make an impact on the real economy. First of which is that maintenance and use of an economic capital buffer can reduce the effect of changing capital requirements. Second is that alternative sources of credit can somewhat offset a decrease in bank lending. Third is that demand effects on lending are thought to overshadow supply effects. Altogether, while taking into account these factors, it is very hard to predict whether a shift to Basel II will increase procyclicality of banking behaviour. 41

42 3. Borrowing position of emerging economies and Basel II In this chapter we try to get a picture of how the switch towards Basel II can affect the borrowing position of emerging economies. It has namely been suggested that the borrowing position of these countries could be damaged in two ways. First, it is possible that poorly-rated emerging economies will become less attractive to lend to as exposures to these countries will ask for higher capital requirements under Basel II- IRB. The result could be that emerging economies will be charged higher interest rates perhaps even with the risk of decreasing the flow of credit. Second, the potential procyclicality of Basel II can impact emerging economies, likely to a stronger degree than developed economies. The literature on both these problems will be discussed in section 3.2 and 3.3 respectively. First, however, in 3.1, it will be discussed how emerging economies will be subject to Basel II Basel II in emerging economies In order to find out how emerging economies may be affected by Basel II it is relevant to consider the following. First, how Basel II will be implemented in these countries, as discussed in section Second, as discussed in section 3.1.2, the dependence of emerging economies on credit from foreign banks Implementation of Basel II in emerging economies Emerging economies will have to make a choice about whether and how to adopt Basel II. Some adopt the pillar 2 and 3 requirements but prefer to stick to Basel I capital requirements, while most others plan to also include pillar 1 capital requirements. For those adopting pillar 1 there is a choice between implementing the IRB approach to credit risk measurement, or to just stick to the standardised approach. I will discuss this latter trade-off by looking at the difficulties related to each of these approaches in context of emerging economies. After doing so I will present the likely course emerging economies are going to take The standardised or IRB approach? One of the major options under consideration is to fully adopt Basel II with the exception that the approach to credit risk measurement is limited to standardised. An approach which is meant for less advanced banks, and which may be the logical choice for most banks in emerging economies. A serious problem however is that credit ratings are still very scarce, with the result being that most assets there will fall into the unrated category. A downside of which is that the 100% risk weight is the same for all unrated assets and thus flat, with the consequence being that the standardised approach will not deliver much risk sensitivity. 65 Majnoni 65 Powell (2005; Page 17) 42

43 & Powell (2004; Page 33-34) therefore argue: For countries adopting the SA, moving to Basel II will imply only a marginal correction to other problems in Basel I and such a move would not address the fundamental problems of Basel I that motivated the new Accord. On the other hand the implementation of the IRB approach could be even less successful. Majnoni & Powell (2004; Page 21) namely gives two reasons as to why implementing the internal rating based approach may be too demanding for emerging economies. The first reason concerns the inexperience of these countries banks with internal risk management: banks in emerging economies are in general less advanced in terms of developing and using a) internal rating methodologies, b) mappings of those ratings into default probabilities and c) portfolio models of credit risk. Many emerging economy banks can therewith considered to be not ready for the IRB approach. A second reason why countries might not be ready for implementing the IRB approach is the lack of supervisory resources and lack of disclosure practices: developing country supervisors tend to have significantly less resources and in many developing countries supervisory human capital, information systems and de jure legal and de facto real powers are sadly lacking. 66 This weakness of emerging economies is something that is demonstrated in the figures on compliance to the Basel Core Principles (BCPs) of banking supervision 67, which are pointing to a very poor performance. Particularly concerning, according to Majnoni & Powell (2004; Page 28), is the lack of (i) effective consolidated supervision, (ii) supervisory independence, resources and authority and (iii) effective prompt corrective action. As supervisors have the important function of monitoring the bank s implementation of the IRB approach under Basel II, a weak supervisory system can be problematic. Therefore Majnoni & Powell (2004; Page 28) suggest: If supervisors lack resources and the basics of effective bank supervision, correcting this should be the first priority and more complex rules on capital requirement (Basel II Pillar 1) may well be counterproductive. So for many emerging economies the first thing would be to improve compliance to pillar 2 (and pillar 3) standards. No sooner than that, should the implementation of IRB follow. However, as suggested by Ward (2002; Page 35), it is possible that countries due to signalling reasons have an incentive to implement the IRB approach, even though it may be too complex for them: if there is some reward for adopting the sophisticated approach, and the cost of implementing the sophisticated approach is lower for a sophisticated country than for an unsophisticated country (eg, it does less harm), then countries may be able to signal that they are sophisticated by implementing Basel 2 or a model-based approach within Basel 2, even if doing so achieves no prudential good. So by implementing the IRB approach, a country can be seen as more developed, and hope to benefit from that. Nevertheless, the result would be that the capital Accord were to be implemented in a way that is, as discussed, too demanding. 66 Majnoni & Powell (2004; Page 21) 67 Which is a set of 25 principles introduced by the BCBS in 1997 with the aim of listing the prerequisites for effective banking supervision. 43

44 What will emerging economies do? According to Powell (2005; Page 18) official indications point out that the grand part of countries are going to adopt Basel II, of which many will implement the standardised approach. In case of some of the larger emerging economies expectations are that a part of the banks adopt the IRB approach. If we follow the assumption that only a few large emerging economy banks adopt the IRB, then already it is likely that a big part of a nation s assets would fall under it the reason being the big size of the large banks. 68 If that is the case then it is reasonable to say that the calibration of the IRB is of importance for domestic lending from emerging economy banks. If the case however is that the standardised approach will be most prevalent for domestic lending in a country, then the capital requirements set by this approach will of course be more important Dependence on credit from foreign banks When looking at statistics it can be seen that lending from foreign banks is of strong importance for emerging economies. Namely according to this Powell (2005; Page 2) the sum of foreign bank claims on emerging economies is around 1.7 trillion 69, which represents 26% of total domestic bank lending in these countries. Interestingly, the foreign share in bank lending would be 69% for Latin American countries and 78% for the group of emerging economies in Europe. This dependence on credit from foreign banks has as a result that if implementation of Basel II in foreign banks is to have an adverse effect on lending, emerging economies could find themselves in a vulnerable position. 70 Relevant to mention is that the foreign (internationally-active) banks under consideration will likely be adopting the IRB approach of Basel II. 71 So therefore it could be argued that even if some emerging economies themselves choose not to implement Basel II, they can still be affected by the Accord as they strongly rely on credit from foreign banks under the IRB approach Will lending to emerging economies be discouraged? As capital requirements are made much more risk-sensitive particularly under the IRB approach of Basel II, they are thought to increase for exposures to poorly-rated emerging economies. Consequences of which can be serious. Meier-Ewert (2002; Page 17) for example predicts the following: It seems at least plausible that the steep increase in capital-requirements for loans to developing countries under the IRB-approach 68 Majnoni, Powell (2004; Page 9) 69 BIS statistics over September 2005 indicate a higher number, of $2.3 trillion (see BIS Quarterly review March 2006) 70 UAB (2003; Page 70) 71 Powel (2005; Page 18) 44

45 will effectively lead sophisticated banks to exit lending to this category of borrowers and to concentrate on those categories, for whom their IRB approach yields the greatest advantages. Herewith is argued that the higher capital requirements would lead banks to reduce their lending to these countries. However, a less radical possibility is that the impact will mainly cause a rise in interest rates rather than a reduction or reversal in lending flows. A major strand of empirical literature 72 namely tries to estimate the effect on interest rates. Another but related issue is that of the competitive position of foreign banks versus local banks in emerging economies.. Before discussing these issues in subsection and respectively, I will first take a close look at the potential increase in capital requirements that banks will face under Basel II when lending to emerging economies in subsection IRB capital requirements for emerging economy exposures As indicated before, the IRB approach is designed to require a high level of capital against risky exposures. Expectations are therefore that borrowers that are considered to be of high risk, such as many emerging economies, result in high capital requirements. Although banks tend to have their own ways for determining what amount of (economic) capital is adequate, it is still interesting to consider the changes in capital under the assumption that banks exactly hold the levels of capital that is required. 73 In the following table is shown how differently-rated sovereign exposures are associated with different default probabilities and thus lead to widely differing capital requirements. The comparison is made to the situation under the current Accord and the standardised approach. 72 For example Weder & Wedow (2002), Powell (2005), Liebig et al (2004) 73 Paragraph will examine the credibility of this assumption 45

46 Table 6 Comparing the IRB with the standardised approach 74 Ratings Regulatory capital Probability Risk Capital required approach of Default (%) weight (%) per $100 BBB Current Standardised IRB BB Current Standardised IRB BB- Current Standardised IRB B+ Current Standardised IRB B Current Standardised IRB CCC Current Standardised IRB Source: Standards & Poor (2001) ; Risk weights calculated by Weder, Wedow (2002; Page 12) with LGD=50% and M=3 Clear from the table is that the IRB approach is by far the most risk sensitive. While observing the table of risk weights one can suspect that sovereign lending to a large and influential economy such as Indonesia (B+) would under the IRB approach require an amount of capital nearly 50% higher than that under the current Accord. For lower rated sovereigns this difference is even higher e.g. for Uruguay (B) the difference is around 138% with $19.10 of required capital on a $100 exposure. 75 In comparison to sovereign lending, it is somewhat more difficult to assess how capital requirements will be on exposures to the private sector of emerging economies. One general difficulty with many of the borrowers in these countries is that they are not assigned an external credit rating. Which is due to the fact that rating agencies charge a fixed fee for each assessment so that so that average costs for small firms would get too high. The consequence of which is that small firms find acquiring an external credit rating too costly and choose to do without. 76 As external credit ratings are normally a good source of information for internal ratings assessments by banks 77, their absence would make applying the IRB approach to these firms also more difficult. Additionally, Griffith-Jones & Spratt (2003; Page 11) argue that the process of risk assessment may be further hindered by a serious lack of historical default data on companies in emerging economies. In recognition of this problem, the Committee has ordered banks to assign PDs to these exposures very conservatively. The implication of which could be as follows: Thus a bank operating rules under an IRB 74 S&P assessments are hereby used to convert S&P ratings into default probabilities Danielsson, Jonsson (2004; Page 10) 77 Danielsson, Jonsson (2004; Page 10) 46

47 approach faces two options, in relation to lending to developing countries; 1) withdraw from lending, which would reduce supply of loans or 2) adopt a conservative approach to assigning borrowers to PD banks, which would increase cost, as banks will assume the worst about those borrowers creditworthiness. 78 So because of data deficiency borrowers in these countries may be given an internal rating that is disproportionally poor. A possibility conjectured by the Spanish Bank Association (2003; Page 19) is that unrated small and medium sized enterprises in emerging economies will be internally rated according to the rating of their respective sovereigns. The difference, as they argue, between the sovereign rating on the one hand and the rating given to a company on the other, would then be in the order of 2 or 3 brackets. Which would imply that lending to such companies in countries with a sovereign rating of B and lower, would face an IRB capital requirement far beyond 20%. 79 Another illustration of what capital requirements to expect while lending to the private sector of emerging economies is given by research from Powell (2004; Page 10). This author calculated the IRB capital requirements for 302 Latin American banks and corporations that did have an external credit rating. The results (in the graph below) indicate that for a country like Venezuela, which carries a BB- sovereign rating, loans to most of its banks and corporates would carry a capital requirement of between 11.5% and 20.3%. 80 Figure 4 IRB capital requirements for private sector exposures in Latin America Source: Powell (2004; Page 10) After having looked at capital requirements for individual exposures it is interesting to get a picture of what the new Accord will mean taking into account real bank portfolios. A series of Quantitative Impact Studies 78 Griffith-Jones&Spratt (2003; Page 11) 79 Powell (2005; Page 18)

48 (QIS) have been initiated by the Committee exactly with that in mind. The third of which was most widespread, involving 365 banks from 43 countries. Its results, among other things, indicated that large internationally-active banks would see an increase in capital requirements for their sovereign exposures of 47% if they choose the foundation IRB approach and 28% if they choose the advanced variant. While commenting on these estimations Liebig et al (2004; Page 2) point out the danger when actual capital levels move in the same manner: Given the prominent role of Group 1 81 banks in lending to emerging markets, this rise in requirements might potentially lead to large adjustments in international bank lending to emerging markets. Next subsection presents several assessments of this effect Effects on interest rates If capital levels have to rise against certain exposures, the costs of holding the extra capital can be passed on to the borrowers causing it. As seen in the previous section, high-risk emerging economy borrowers are most likely to cause an increase in capital requirements. Therefore, these countries are by some expected to face higher interest rates. In this subsection this effect of Basel II on interest rates will be discussed. It must be noticed however that the literature focuses on the borrowing position of emerging economy sovereigns and doesn t make estimations of how the private sector may face changes in interest rates. According to Weder & Wedow (2002) the magnitude of the effect of Basel II-IRB on lending behaviour depends heavily on the following three factors: Assumptions on the required return on capital Whether regulatory capital is binding or not Effects of diversion to other sources of credit I will explain each of them in the following three paragraphs Assumptions on the required return on capital If the switch to Basel II obliges a bank to keep more capital against its loan portfolio, it would have to bear the costs of holding more capital. These extra costs may then be passed on to borrowers by higher interest rates. Therefore it is important to have a realistic vision of required return on capital, as this variable would determine what these extra costs would be. The following example will illustrate. 81 Group 1 is the group of internationally active large banks. 48

49 Example 2 Importance of required return on capital On a $100 loan portfolio, a bank currently has to keep (and is keeping) $8 whereas under Basel II this will double to $16. In effect, $8 of extra capital will have to be held. If required cost of capital is 12.5%, then the costs of holding extra capital will be 1% (=12.5%* $8). Which, if passed on to borrowers, leads to a full percentage point increase in interest rates. However, if required return on capital is twice as high with 25%, then interest rates may increase with two percentage points (=25%*$8). With an assumed rate of return of 18%, Ward (2002; Page 18) calculates that the spread increase on loans to B-rated sovereigns needs to be 1.71% and on loans to CCC-rated sovereigns 2.8% Capital requirements binding or not binding? Clearly, the estimations in the last paragraph are under the assumption that capital requirements always bind. However, this is probably incorrect due to that many banks determine their capital levels according to their own risk management systems which, as mentioned before, is called economic capital. If this economic capital namely is higher than the regulatory capital, then the level of regulatory capital does not have an influence since it doesn t form a binding constraint. In the following example I will illustrate how the economic capital buffer can dampen or absorb an increase in capital requirements following the changeover to Basel II. Example 3 Here we take again the situation of a $100 exposure, against which a bank faces a Basel I risk charge of $8 and a future Basel II risk charge of $16. With the help of three hypothetical scenarios I will show how implications of this increase in capital charge may depend on the level of economic capital. The three scenarios are the following: 1. Economic capital is bigger than $16 so that both Basel I and II don t bind. The bank will not have to increase its capital level. 2. Economic capital is somewhere between $16 and $8 so that Basel I doesn t bind but Basel II does. If we assume economic capital is at $12, then the bank will have to increase its capital level by $4 under Basel II. 3. Economic capital is at $8 or below so that both Basel I and II do bind. The bank will have to increase its capital level by $8 (from $8 to $16). 49

50 In the first scenario, economic capital completely buffers the effect of an increase in regulatory capital. In the second scenario, economic capital partly buffers the increase half of it to be precise. Only in the third scenario economic capital makes no difference. Therefore the consequence of this buffer effect is that higher capital requirements under Basel II will not necessarily cause banks to actually change their capital levels to the full extent or at all. As illustrated by the example, to find out the influence of economic capital it is important to first find out 1) whether Basel I binds, and 2) whether Basel II will bind. I will address both questions. 1) Does Basel I bind? Basel I is by Weder & Wedow (2002) considered unlikely to be binding as they find that the capital requirements that would have been prescribed by the IRB approach can already explain the lending flows that have taken place to emerging economy sovereigns over the previous decade. Herewith they conclude that the IRB approach and measures of economic capital that are currently used under Basel I are similar so that the switch to Basel II will not affect the borrowing position of emerging economies. 2) Will Basel II-IRB bind? If economic capital is also likely to exceed the capital requirements prescribed by the Basel II IRB approach, then an increase in capital requirements is even more likely to be completely dampened pointing to the first scenario of example 3. So the important question is: will Basel II-IRB bind? Liebig et al (2004), and Powell (2005) both delve into the question of whether economic capital standards maintained by banks tend to be higher or lower than those of the IRB approach of Basel II. Liebig et al (2004) argue that it is unlikely that the capital prescribed by the IRB approach will be a constraining factor in emerging economy lending by German banks, as they find that the models for setting capital currently used by these banks already achieve higher levels of confidence than the IRB formula would the level normally targetted by banks of 99.5% would namely be superior. Powell (2005) looks at whether the current spreads on sovereign bonds are already large enough to make room for a level of economic capital higher than that of regulatory capital under Basel II. Herein the author concludes that for most sovereign exposures both capital Accords are unlikely to be binding so that the Basel II won t have any effect. However, for countries with a low sovereign rating (B and lower) the estimations are that interest rates will still have to rise in the order of at least 1 to 3 % These estimations are somewhat higher when they make use of alternative figures to turn S&P ratings into default probabilities 50

51 The evidence from both Liebig et al (2004) as well as from Powell (2005) suggests that economic capital could considerably soften the impact on emerging economies. Nevertheless, other research indicates that the use of economic capital is not sufficiently widespread as to prevent Basel II from having an impact on lending. 83 And even though economic capital at places is an integrated practice, regulatory requirements may still be binding, as is noticed by Pricewaterhouse Coopers: The PricewaterhouseCoopers analysis assumes that the changes in capital requirements from Basel I to Basel II will lead to corresponding changes in capital assigned to the type of lending in question. This is due to the fact that even if a bank is a sophisticated user of economic capital, regulatory capital will still be a constraint. While this may not always be considered at business unit or transaction levels it will ultimately have an influence on the institution s overall level of capital. 84 Herewith is argued that changes in the overall level of regulatory capital will not leave economic capital unaffected. Something which suggests that the cushion between economic capital and regulatory capital will continue to be of a similar magnitude. 85 As a result, the dampening effect of economic capital would be minimal so that an increase in capital requirements under the switch to Basel II may still have a strong impact Diversion to other sources of credit The analysis sofar of the effects that Basel II-IRB may have on interest rates has neglected the possible dampening effect of other sources of credit. As already identified in chapter 2 lending also takes place through financial markets which can offset changes in bank lending. Weder & Wedow (2002; Page 20) therefore argue the following: we prefer to state qualitatively that competition from other sources of finance may lead to some diversion and thus dampen the effect of higher capital costs for emerging markets. So a potential deterioration in the borrowing position of emerging economies can be alleviated by credit from financial markets. However, it is doubtful whether emerging economy borrowers will have easy access to financial markets, as argued by Claessens et al (2003; Page 33): This would reduce the impact of B-II; of course, the access to capital markets and other financing may be more limited for precisely these same lower rated countries, thus negating this effect. Therefore for emerging economies, diversion to financial markets is less likely to be a way in which to soften an adverse impact of Basel II. Another source of credit that can soften a potentially negative impact of Basel II-IRB on emerging economies, is that of banks under the standardised approach. Weder & Wedow (2002; Page 20) argue the following: From the point of view of low-rated emerging markets, the regulatory arbitrage of financing 83 Griffith-Jones, Segoviano, Spratt (2003; Page 8) 84 PriceWaterhouse Coopers (2004; Page 6) 85 An idea also forwarded by Griffith-Jones et al (2004; Page 1) 51

52 towards standardised banks may dampen the effect on prices and flows. Herewith is suggested that regulatory arbitrage will take place as bank lending to low-rated borrowers may be taken over by banks that have adopted the standardised approach. Namely due to that the standardised approach prescribes lower capital requirements against these borrowers than the IRB approach does, interest rates will likely be affected to a lesser extent as more banks choose for the standardised approach. Weder & Wedow (2002; Page 20) argue that this alleviation could be considerable as major offshore centres and emerging economies are not obliged to introduce the IRB variant Competitive position of emerging economy banking sector In general the expectation is that banks adopting the IRB approach will have a competitive advantage over the banks adopting the standardised approach. Namely as most borrowers lead to lower capital requirements under the IRB approach than under the standardised approach, banks under the IRB approach have the ability to benefit by holding less capital. The consequence may be that the more advanced foreign banks that are most capable of adopting the IRB approach start to dominate the banking sector. 86 As argued by Griffith Jones, Spratt (2001; Page 1): [G]reater competition from internationally active banks could see banks from the developing world being taken over by foreign banks, at a pace even quicker than has occurred in recent years. Note however that for as far as borrowers in emerging economies are low rated, banks under the IRB approach will not have an advantage. In this regard Powell (2005) brings to the fore a serious concern about the competitive position of foreign-owned banks. Since it is expected of them to be consistent in the application of the IRB approach in all their affiliates worldwide they will tend to face higher capital requirements than banks under the standardised approach in low-rated economies. The consequence then could be as follows: The implication is that global banking may then retreat towards international banking where foreign banks operating under a lead regulator essentially only lend to higher rated local banks and selected corporates in emerging economies. 87 Considering the fact that affiliates of foreign banks are important providers of funding for these countries, their decline could have a serious impact on lending. However, Powell (2005) does not mention to what extent any possible gap caused by it could be filled by banks under the standardised approach Procyclical effect of Basel II in emerging economies As discussed in chapter 2, the switch to Basel II may increase procyclicality of the banking system, which during recession periods could help produce a credit crunch. For emerging economies however which are countries that already suffer from a higher business cycle amplitude and banking instability the 86 Griffith Jones, Spratt (2001; Page 1 2) 87 Powell (2005; Page 19) 52

53 procyclical effect of Basel II could be even more dramatic. Griffith-Jones et al (2002; Page 8) for example argue as follows: For developing countries, an increase in pro-cyclicality of lending both international and domestic is very likely to contribute to further increase in the number and depth of banking and currency crises, which would be developmentally very negative. As crises can be very costly, an increased cyclicality of these economies by introducing Basel II would of course be a major downside. Due to that under Basel II capital requirements are sensitive to the development and perception of risk, the characteristics of the already strong businesss cycle in emerging economies will be the first point of interest in subsection After that, in subsection it is discussed how particularly for emerging economies the new Accord may cause bank lending to become less stable and therewith more procyclical, even to the extent of facilitating financial crises Business cycle amplitude and proneness to crisis of emerging economies Carstens (1998) mentions two noticeable features of emerging economies business cycles over time. First, business cycles of these countries tend to show a relatively close synchronization with industrialized countries cycles. 88 Elsewhere by Ward (2002; Page 23) even the term business cycle takers has been applied to indicate emerging economies. The implication of which is that emerging economies are, in addition to domestic shocks, also highly vulnerable to foreign shocks. Second, emerging economies are found to display considerably larger volatility. Something which can be clearly illustrated by looking at the typical depth of recession of these countries. As the author mentions: While in the period the typical recession in industrialized countries registered a reduction in GDP of approximately 2 percent, in Latin America the average fall in GDP during a recession amounted to 8 percent. 89 Also in comparing the standard deviation of various economic indicators of emerging economies with that of industrialized countries, a clear difference shows through. In the graph below such a comparison is made by region over the period 1970 to The figures point out several of the weak spots of emerging economies: namely the volatility in GDP growth, terms of trade, and inflation volatility. 90 Each of which can be thought of as rather detrimental, or for some even potentially reinforcing each other. For example, since that many emerging economies are dependent on revenues from export, a deterioration in terms of trade may affect their income considerably, and therewith lowering their GDP growth. 88 Carstens (1998; Page 1) 89 Carstens (1998; Page 1) 90 The first five regions on the left can broadly be earmarked as emerging economies, whereas several of the East Asian Miracle countries can be thought of as belonging to the group of industrialized countries e.g. Japan, Hongkong, and Singapore. 53

54 Figure 5 Average volatility by region Source: Ward (2002; Page 24), based on data from Caprio & Honohan (1999) Next to these factors, Carstens (1998) also adds another element to the mix: namely the role of capital flows. The author stresses the notion that emerging economies are highly dependent on foreign lending even though these flows generally tend not to be very stable. The consequence of which could therefore be that crises are aggravated or even caused by sudden capital outflows such as happened in Mexico and East Asia in the nineties. Especially if a fixed exchange rate regime has been set up often with the intention to curb inflation an economy is thought to be especially vulnerable. On the one hand such an installation may of course attract extra foreign capital due to the promised exchange rate stability. However on the other hand if the regime collapses as is known to happen the plummeting exchange rate may initiate a capital flight and drag the country into a financial crisis Basel II and the acceleration of emerging economy crisis This subsection discusses how Basel II can help accelerate emerging economy crises. First, as discussed in following paragraph such an acceleration has to do with the volatile behaviour of external sovereign ratings. Then in the second paragraph, it is discussed how internal rating behaviour in emerging economies would also lead to a destabilizing effect under Basel II. Finally in the third paragraph, the destabilizing effects of a preferential treatment of short-term debt is discussed. 54

55 External rating volatility A reason why the procyclical effect of Basel II may be extra marked for emerging economies in crisis period, can come from the use of external credit ratings for determining capital requirements under the Basel II standardised approach. Namely as indicated by Meier-Ewert (2002; Page 13): there are a number of difficulties with the use of External Ratings Agencies, chief among which is the strong indication that they would exacerbate fluctuations in the cost of external financing for developing countries, and enhance the procyclical effect of the New Accord. This extra procyclicality resides in the fact that external sovereign ratings have the tendency to excessively drop after 91 materialization of a shock. Particularly, due to that in emerging economies where external ratings are scarce external sovereign ratings are expected to be the basis for the ratings assigned to the private sector, the effect of large volatility in this variable can be extra damaging. In this situation namely, in which the private sector does not receive a rating any higher than that of its sovereign, an excessive ratings dive could severely affect banks and companies access to bank credit. 92 An additional problem with basing capital requirements on external sovereign ratings is that of potential circularity. Namely as soon as the sovereign rating plunges during perceived weakness a country will have more difficult access to international capital markets, which in turn reduces the ability to roll over debt. Then, due to the reduced ability to roll over debt, the country s creditworthiness would decrease even further which has an adverse effect on the sovereign rating, giving even more fuel to the process. Herewith a mere liquidity problem can deteriorate into a severe financial crisis Internal rating volatility The results of a study performed by Segoviano & Lowe (2002) indicate that internal ratings as well can behave in a manner that reinforces emerging economy crisis. These authors namely look at how internal ratings of Mexican banks have developed in the second half of the 1990s, and how these changes would have led to volatility of IRB capital requirements had Basel II been in place at the time. The timeframe picked is especially interesting as it is at the aftermath of a financial crisis, which happened in Mexico at the end of Segoviano & Lowe (2002; abstract) say the following on how this period of instability has been reflected in banks risk assessment, and therewith internal ratings: We find that measured risk increased after the crisis and then fell as the recovery took hold. So just as with the experience of external sovereign ratings, internal ratings tend to drop no sooner than after the crisis. Also, ratings start to improve only after the economy is picking up again, which suggests that the predictive ability of banks obviously has its limits. The authors then continue by arguing that this rating behaviour would have led to procyclical 91 Kraussl (2003; Page 32) 92 See Ferri et al (2000; Page 2) 93 Powell (2005; Page 10) 55

56 capital requirements if Basel II-IRB were to have been in effect: In turn, despite the limitations of the data, we find that the proposed internal ratings-based approach would have generated large swings in regulatory capital requirements of the second half of the 1990s, with required capital increasing significantly in the aftermath of the crisis, and then falling as the economy recovered. To illustrate, the simulation concludes that the additional impact of Basel-IRB would lead to a maximum change in capital requirements of 56.7% over the sample period. Which indicates that in case of an emerging economy such as Mexico, the IRB approach could lead to a considerable increase in procyclicality compared to the situation under Basel I. Therewith an amplification of the business cycle is not unthinkable: Looking forward, if movements in actual bank capital were to show this same cyclical variation, then business cycle fluctuations might be amplified by developments in the banking industry. So, then again, the procyclical effect of capital requirement volatility hinges on the condition that actual capital levels must move in the same way as required capital, instead of forming an economic capital buffer during boom periods Treatment of short-term debt Due to that Basel II includes a favorable treatment of short term loans, the new Accord has often been criticised for making emerging economies more vulnerable to financial crisis. Namely if international lending is largely of a short-term nature, the response to an adverse shock can be greater as banks have the possibility of costlessly withdrawing their money. 95 An effect which, according to Reisen (2002; Page 12), has left a strong mark on international finance: Short-term foreign debt, in relation to official foreign exchange reserves, has been identified as the single most important precursor of financial crises triggered by capital-flow reversals. Therefore, in interest of financial stability, one can argue that short term debt should be discouraged. On the other hand it has to be noted that due to higher liquidity, interest rates demanded on these loans will lie lower, making it cheaper for emerging economies to borrow. 96 Even though Basel II takes away some incentives towards short-term lending that were present under Basel I more so in the standardised approach than in the IRB approach 97 the new Accord is still considered to give preferential treatment to short term loans. Highly disputed for example under the standardised approach is the choice to change the definition of short term debt from one year to three months. An alteration that according to Felaban (2004; Page 2) would create an incentive to shorten the maturity of loans to emerging markets. 94 Something which is already discussed in paragraph Ward (2002; Page 17) 96 Ward (Page 17) 97 Reisen (2002; Page 14) 56

57 3.4. Conclusionary remarks From the implementation discussion it stands clear that emerging economies face a difficult choice between alternatives. On the one hand opting for the standardised approach may be uncalled for in case of low rating penetration. On the other hand banks also shouldn t adopt the IRB approach if BCP compliance is low, or if it is considered unfeasible or undesirable to increase capital levels. Nonetheless, the expectation is that banks in emerging economies will generally choose for standardised approach, except for several of the larger ones that adhere to the IRB approach. Figures also indicate that emerging economies, especially in Latin America and developing Europe, heavily depend on funds from the international banks. Due to the fact that most of these international bank will adopt the IRB approach one could argue that for the case of emerging economies this approach will at least be of the same relevance as the standardised approach. The first way in which the borrowing position of emerging economies can be damaged is due to that exposures to these countries will be discouraged under Basel II. When looking at potential increases in capital requirements it could be expected that poorly rated emerging economies may be charged higher interest rates, perhaps even to the point of decreasing the flow of credit. Especially the largely unrated private sector of emerging economies will cause strong increases in capital requirements. Whether and into what extent there will be such an effect on interest rates depends on three factors, namely: 1) the cost against which banks can acquire extra capital; 2) the economic capital buffer held by banks (now and under Basel II) and; 3) the potential for diversion to other sources of credit. A related concern is that of the competitive positions of both local and foreign banks. Unsophisticated local banks may namely lose market share to the more advanced foreign banks under the IRB approach. But also, due to that banks under the IRB approach have a relative disadvantage lending to low rated borrowers, it is possible that foreign banks will start to retreat from these parts of the market, ultimately leading to a decrease in emerging economy lending. Second, another threat to emerging economy borrowing positions is the potential that procyclicality of Basel II is particularly damaging for emerging economies. The already ample business cycle and crisis proneness of these countries may namely be increased due to the following destabilizing effects of Basel II: 1) that volatile external sovereign ratings may accelerate crises; 2) that the procyclical behaviour of internal ratings may also accelerate crises and; 3) that the preferential treatment of short-term debt can encourage rapid capital retreats during crisis. 57

58 4. Policy options Several policy options have been suggested in the literature that aim to curb potential negative effects of Basel II on emerging economies. Two of which concentrate on adapting the new capital Accord to the specific benefit of these countries. Three other measures focus on countering procyclicality, either by adapting the Basel II regulation to make capital requirements less volatile over the business cycle, or by offsetting part of the capital requirement volatility by a more forward-looking provisioning Softening capital requirements for emerging economy exposures Due to that the impact on lending to emerging economies is largely attributed to the proposed IRB approach, it is interesting to examine whether there is benefit to be made from amending it. Moreover, assuming that regulatory capital is the main determinant of the level of economic capital, Basel II-IRB can greatly affect the borrowing position of low-rated economies. 98 Two main options are forwarded in this respect by the Spanish Banking Association: Flattening the IRB curve for emerging economy exposures, by making the capital requirements less responsive to risk 2. Counting in beneficial diversification effects which originate in the fact that the correlation between emerging markets and developed markets tends to be low I will discuss each of them respectively in the next two subsections Flattening the IRB curve for emerging economy exposures As mentioned in chapter 1, the IRB formula tries to set an amount of capital requirements that corresponds to a target confidence level of 99.9%. But due to the fact that it may cause a huge increase in required capital for weaker borrowers, it is perhaps appropriate to target a less ambitious level of confidence. By for example setting it lower at around 99% still a higher level of prudence may be achieved than under Basel I or under the standardised approach, but the impact on capital requirements could be softened immensely. The figure below gives a simple illustration of this idea. 98 See paragraph Spanish banking association (2003; Annex Page 1) 58

59 Figure 6 The IRB curve flattened Confidence level: 99.9% Capital requirement Confidence level: 99% Probability of default One must note however that this amendment should be applied in high risk markets only, since it is generally undesirable and unnecessary to apply a lower confidence level in the low-risk developed markets. The Spanish banking association (2003) therefore proposes to make the IRB confidence level dependent on a country s sovereign rating. If the sovereign rating is as high as for example BB which is characterised by a PD of around 0.94%, then the confidence level of lending to this country is to be diminished by the same amount (i.e. it would supposedly come down to 99.06%). By following this rule capital requirements placed on lending to the private sector in risky emerging economies would not have to dramatically increase. In figure 7 their assessments for the category of SMEs in emerging economies are demonstrated if the risk formula is to be adjusted this way. The assumption made in this calculation is that SMEs carry a credit rating 3 brackets lower than their corresponding sovereign. As displayed in the graph it is clear that this adjusted IRB regulation is much milder. Figure 7 Impact of different capital formulae on capital requirements for SME exposures Source: Spanish banking association (AEB) (2003; Page 9) 59