TROUBLESHOOTING BASEL II: THE ISSUE OF
Contents
1. 025 28000 fF 4 0 02 26000 fF 4 0 015 24000 F 4 0 01 22000 F 20000 L i 4 1 4 0 005 2002 2004 2006 2008 2010 Figure 1 PIT Capital Requirements Vs GDP growth Source Own calculations 467 0 035 T T T T T 34000 hpt_GDP right hpt_TTC_K left hpt_PIT_K left 0 03 32000 30000 0 025 28000 0 02 26000 0 015 24000 0 01 22000 0 005 L gt L E 20000 2002 2004 2006 2008 2010 Figure 2 PIT Capital Requirements Vs TTC Capital Requirements and GDP growth Source Own calculations 1 8e 006 T T T T T K_est_GDP hpt_K 1 6 006 fF 1 4e 006 F 1 2e 006 F 1e 006 F 800000 T T 600000 400000 200000 T 2002 2004 2006 2008 2010 Figure 3 GDP adjusted Capital Requirements Vs GDP growth Source Own calculations 4682. 4 medium and 4 small companies whose total exposures equal the aggregate credit portfolio to companies in the respective period Our model economy is explicitly characterised by Romanian macroeconomic data regarding the specified period To compute how regulatory capital levels would evolve over the business cycle we estimate a logistic model of the one month ahead PDs of Romanian firms The dependent variable DEFAULT is a binary variable that takes value 1 when a firm defaults in the course of a year on its outstanding loans at the end of the previous year and zero otherwise The explanatory variables comprise characteristics of the firm characteristics of its loans and macroeconomic variables A borrower is considered to have defaulted if it is 90 days overdue failing to meet his financial obligations on a certain loan or if with high probability it is considered to be unable to meet its obligations The explanatory variables used in the model dated in month f are firm specific variables and Romanian macroeconomic controls COL represents the proportion of guarantees in a firm s borrowing proxying the amount of collateral Jim nez Salas and Saurina 2006 show that banks ask for collateral to those firms that they denote as being riskier AGE captures the age of each firm with the idea that younger firms are more prone to default than older ones FSIZE proxies the size of a firm it is calculated via deflating the EAD growth of a firm
3. A possible solution for mitigating the cyclical effects of regulatory capital is the use of through the cycle TTC capital requirements To estimate DEFAULT by the TTC approach we follow the idea of Saurina and Trucharte 2007 that is replacing the current values of macroeconomic variables by their average values over the sample period We then compute the monthly capital requirements with the Basel II foundation IRB approach The results of the re estimation of the logit model are found in Table 2 Comparing TTC capital requirements with the PIT values the cyclical variability declines significantly as it can be observed in Figure 2 The peak low deviation in the TTC series is of 3 03 significantly better then in the PIT capital requirements The business cycle multiplier Adjusting the output of the Basel II formula The second approach for adjusting the Basel II capital requirements is to smooth the output of the formula Basically we adjust the PIT capital requirements series obtained from Table 1 with a business cycle multiplier as it can be seen in the following formula k_adj Uy Ke The multiplier can be of various forms but we use a simple and conventional approach ee It 9G_avg Me 2N a g In the equation k denotes the original PIT capital series and k_adj the adjusted series Regarding the multiplier equation g is the growth rate of one of the macroeconomic variables g_avg its average over the sample period
4. G18 G17 I Introduction The financial crisis proved that the capital requirements system which is based on risk weights defined by Basel II can t adapt to the new economic prosperity The recession is aggravated by the fact that the banks were forced to squeeze their credits and this caused a delay in the economy s availability to grow Several financial analysts pointed out the errors in the system therefore the Basel Committee on Banking Supervision must find a solution in order to develop the model 461 In our paper we compute an in depth analysis on the issue of the procyclicality comparing two possible methods that can be used to mitigate the cyclical effects of Basel II regulatory capital The procyclical effect can be observed in historical data from the banking sector since the capital regulation based on risk weights is very prosperity sensitive in recession the credit losses consume the bank s capital while venture capital increases If banks aren t able to gain enough capital for credit losses in a short period of time then credit crunch may occur This kind of process characterizes perfectly the financial crisis This leads to the fact that banks can t even provide loans for customers with high ratings because of the inadequate level of capital The structure of our paper is the following in the first part we compute a brief review of the literature regarding the issue of procyclicality This will be followed by an a
5. h its standard deviation over the sample period N x is the standard normal cumulative distribution function and a is a positive constant parameter The key features of the business cycle multiplier are it is continuous and increasing in the proxy for the business cycle g so capital requirements are increased in favourable periods and lowered during downturns or recessions Also u is bounded so capital requirements do not increase without bound or become negative Parameter a is defined as 0 1 however we tested various other values The purpose of a is to minimize the root mean square deviation RMSD of the adjusted series and also to obtain a reasonable amount of capital adjustment Summing up we choose the value of a that is best in terms of smoothing the cyclical component of the pit capital requirements series Figure 3 shows the p adjusted capital series when g GDP together with 464 the PIT series Note that the adjusted series has been applied a HP filter of lambda 500 so that the cyclical smoothing can be easily observed IV Results Taking into account the HP filter fitted values we compare the different smoothing procedures by computing the RMSD of each adjusted capital requirements series The values obtained are shown in Table 3 The output formula adjusted values show a much smaller deviation as the TTC series stands out with a RMSD of 0 16 Regarding the output smoothing procedure a choice has to be made in order to sp
6. stock markets in any period Their conclusion was that in the case of Basel II Capital Requirements capital reserves constituted by banks are higher during an economic boom than during recessions Secondly we observed that Pederzoli Toricelli and Tsomocos 2009 analyzed the problem of procyclicality with comparing two rating systems They built up a general equilibrium model which contains 2 heterogeneous banks 2 companies and 1 household They found that the cyclical rating system results higher default rates and lower profit in the case of banks in recession times Kashyap and Stein 2004 argue that if the shadow value of bank capital is low in expansions and high in recessions optimal capital charges for each type of risk should depend on the state of the business cycle Without such adjustments capital requirements would be too low in expansions when bank capital is relatively plentiful and has a low shadow value and too high in recessions when the shadow value of bank capital goes up leading to the amplification of business cycle fluctuations Greenspan 2002 noted that the supervisory leg of Basel II is being structured to supplement market pressures in urging banks to build capital considerably over minimum levels in expansions as a buffer that can be drawn down in adversity and still maintain adequate capital Lastly Caprio 2009 studied the counter cyclical capital requirements definition and Repullo Saurina and Trucharte 20
7. 10 analyzed and compared the procedures which mitigate the effects of the procyclical capital The foundation of Caprio s research was a macroeconomic data analysis from Spain and Columbia His conclusion was that the risk based capital regulatory system s rectification results only short term adjustments On the other side Repullo Saurina and Trucharte used data from Spanish companies and estimated a PD model His results from this 462 research were very complex Adjusting the output side of the Basel II formula with a credit growth multiplier or a yield multiplier we don t receive the optimal level of capital from the procyclical point of view They found two possible solutions for handling the procyclicality the first one is the input side smooth tuning and the second one is on the output side which can be received with the help of a multiplier based on GDP growth III Cyclical adjustments of Basel II capital requirements The recent financial crisis with its boom and bust lending cycle has brought to the forefront the need to address the potential procyclical effects of risk sensitive bank capital regulation To see how Basel II capital requirements evolve over the business cycle we construct a model economy that is composed of one commercial bank and ten firms The model is practically a simulation of the Romanian economy also its banking and financial sector over the period 2000 2010 The firms sector is composed of 2 corporate
8. TROUBLESHOOTING BASEL II THE ISSUE OF PROCYCLICALITY Benyovszki Annamaria Babe Bolyai University Faculty of Economics and Business Administration Bordas Eszter Babe Bolyai University Faculty of Economics and Business Administration Kiirti L szl d m Babe Bolyai University Faculty of Economics and Business Administration Szodorai Melinda Babe Bolyai University Faculty of Economics and Business Administration A widespread concern about Basel II capital requirements is that it might amplify business cycle fluctuations forcing banks to restrict their lending when the economy goes into recession Under the IRB approach of Basel II capital requirements are increasing functions of the probability of default PD loss given default LGD and exposure at default EAD parameters estimated for each borrower and these inputs are likely to rise in economic downturns In this paper we compare two alternative procedures that are designed to somehow moderate the procyclical effects induced by Basel II type capital regulation The starting points of our analysis consist Jokivuolla Kiema and Vesala 2009 and Repullo and Suarez 2009 who both examined the impact of regulatory capital s procyclical effects It s vital to note remarks of Caprio 2009 that is making regulatory capital levels countercyclical could worsen the state of an economy during a recession As we do not have access to the Romanian Central Credit Register datab
9. Treat the Disease Without Killing the Patient Journal of Financial Intermediation 15 2006 395 417 3 Greenspan Alan Cyclicality and Banking Regulation Conference on Bank Structure and Competition Federal Reserve Bank of Chicago May 10 2002 4 Jimenez Gabriel Salas Vicente and Saurina Jesus Determinants of collateral Journal of Financial Economics 81 2006 255 291 5 Jokivuolle Esa Kiema Ilkka and Vesala Timo Credit allocation capital requirements and procyclicality Bank of Finland Research Discussion Papers 23 2009 6 Kashyap Anil and Stein John Cyclical Implications of Basel II Capital Standards Federal Reserve Bank of Chicago Economic Perspectives 1st Quarter 2004 18 31 7 Pederzoli Chiara Toricelli Costanza and Tsomocos Dimitrios Rating systems procyclicality and Basel II an evaluation in a general equilibrium framework Ann Finance 2009 33 49 8 Repullo Rafael and Suarez Javier The procyclical effects of bank capital regulation European Banking Center Discussion Paper No 2010 05S Center Discussion Paper No 2010 29S 2009 9 Repullo Rafael Saurina Jes s and Trucharte Carlos Mitigating the pro cyclicality of Basel Il Documentos de Trabajo 1028 2010 10 Saurina Jesus and Trucharte Carlos An assessment of Basel II procylicality in mortgage portfolios Journal of Financial Services Research 32 1 2007 81 101 11 Vasicek Oldrich Prob
10. ability of loss on loan portfolio KMV Corporation 1987 12Witzany Jiri Estimating LGD correlation Working Paper Series IES 21 2009 466 VIII Appendix Table 1 Initial estimation PIT PDs Table 2 Adjusted estimation TTC PDs Model 45 Logit using 1220 observations Model 3 Logit using 1220 observations Dependent variable DEFAULT Dependent variable DEFAULT coefficient std error coefficient std error const 12 0615 4 72765 const 12 6307 4 76106 COL 1 5 69615e 97 2 25951e 07 COL 1 5 70913e 07 2 28224e 07 FSIZE 1 33542 0 715938 FSIZE 1 29090 9 722443 HISTDEF 0 642915 0 140766 HISTDEF 9 639954 9 140788 HISTDEL 0 0886016 0 0502533 HISTDEL 0 0948422 090 0499572 UTIL _3 1 11269 2 71241 UTIL 3 1 99794 2 57275 CREDIT_1i 4 49947 8 14027 GDP TIC 11 4930 24 1395 BET_4 85 8620 32 9714 CREDIT TTC 1 11 4815 18 3083 MATURITY _4 090 0644921 0 774427 BET TTC 4 28 4766 97 8715 GDP 0 372822 1 77291 MATURITY TT_4 9 103762 9 893742 Mean dependent var 0 037705 Mean dependent var 0 037705 McFadden R squared 0 108445 McFadden R squared 9 089853 Log likelihood 174 6628 Log likelihood 178 3052 3 A ri Lf Source Authors calculations Source Authors calculations Table 3 Results of output adjustment Table 4 Overall Results Adjustment Type of adjustment Souce Authors calculations Souce Authors calculations 34000 T T T T T 0 035 hpt_PIT_K right hpt_GDP left 32000 deraa 30000 fF 4 0
11. andle a lot of problems regarding the credit rating agencies In this consultation document they discuss in detail the possible payment models and new measures which prevent the potential risk of a rating arbitrage The supervision of the rating agencies will be done by the European Securities and Markets Authority The second regarding the correct estimation of PD LGD correlation is a vital in building up a safe and sound banking system The model standing behind the Basel II formula is the standard one factor model developed by Vasicek 1987 Many studies argue that this method cannot capture well the correlation between PD and LGD on a large asymptotic 465 portfolio The innovative approach of Witzany 2009 proposes a two factor model Results of testing the model prove that it is able to estimate more punctually and as realistic as possible the correlation between the two parameters implemented on real banking data The lack of reliable information however on public LGDs in the Romanian banking sector makes further research difficult VI Notes 1 Overdue loans are the ones that have been paid before the 90 day threshold 2 Long term exposures are the ones that exhibit 5 years VII References 1 Caprio Gerard Safe and Sound Banking A Role for Countercyclical Regulatory Requirements Paolo Baffi Centre Research Paper Series No 76 2010 2 Gordy Michael and Howells Bradely Pro cyclicality in Basel II Can We
12. ase we compute a model economy that stands as a proxy for the Romanian firms sector Our simulated Romanian economy can be characterised by all Romania specific macroeconomic controls Then we estimate a model of PDs during the period 2000 2010 and based on the estimated probabilities of default we compute the corresponding series of Basel II capital requirements After the diagnosis of procyclicality we analyze two procedures that try to mitigate the cyclical effects of capital regulation smoothing the output of the Basel II formula and smoothing the input by construction of through the cycle ITC PDs The comparison of the different procedures is based on the criterion of minimizing the root mean square deviations of each adjusted series Our results show that the best ways to moderate procyclicality are either to smooth the input of the Basel II formula by using through the cycle PDs or to smooth the output with a multiplier based on GDP growth We conclude that the GDP based smoothing may be more efficient than the use of TTC PDs in terms of simplicity and transparency In terms of the GDP adjustment regulatory capital levels should increase with approx 1 31 during an economic growth period and decrease with 4 03 during a recession in order to mitigate the cyclical effects induced by Basel II type capital regulation Keywords Basel II procyclicality regulatory capital probability of default credit crunch JEL classification
13. by the consumer price index FSIZE enters the model in logarithmic terms HISTDEF is considered to be the main risk profile variable that captures whether a certain borrower defaulted in the past In each observed default event the variable value is increased by 1 Similar to HISTDEF we use HISTDEL that stands for the borrowers record of overdue loans 1 UTIL is the ratio between the amount of credit drawn by a borrower and the credit line The macroeconomic explanatory variables are GDP that is the rate growth of the gross domestic product CREDIT the rate growth of non financial commercial and industrial loans over the one month period BET the monthly average return of the Romanian stock market and MATURITY that is the ratio between long term exposures 2 and the total exposures in the economy Our database contains a total number of 126 monthly observations over the last 10 years Table 1 in the Appendix presents the results of the estimation of the model all coefficients are Statistically significant at the 10 level It s interesting to note that some variables impacts on the default condition especially macroeconomic variables are significant after up to 3 or 4 lags The results show that firms that post collateral when granted a loan have higher probabilities of default Also firms whose exposures show a bigger growth rate than the average have bigger probability of default Yet vital to notice that the coefficient
14. ding the procyclicality issue However we emphasize that the TTC approach should not be written off as it is clearly a simple and effective way to make a quick and basic level fine tunement to regulatory capital levels We stick to our statement knowing that the use of TTC PDs has been criticized by Gordy and Howells 2006 who underline the fact that changes in a bank s capital requirements over time would be only weakly correlated with changes in its economic capital and there would be no means to infer economic capital from regulatory capital Our results are similar to those of Repullo Saurina and Trucharte 2010 agreeing that GDP growth based output adjustment of the Basel II formula is the way to go in terms of simplicity transparency low cost of implementation and even consistency with the idea of a single aggregate risk factor that underlies the capital requirements of Basel II Two major issues are still the purpose of our further research regarding Basel II The first one regulatory arbitrage is mostly threatening safe and sound banking in the European Union where its hazard is higher than in the countries outside the EU To reduce this hazard the specialists started to create a supervisory convergence From this step they expect that the discretionary assets will become reduced The European Commission started a public consultation in 2010 in association with the fact that the 2009th Decree concerning the CRAs does not h
15. ecify the best smoothing macro component As the GDP and CREDIT deviations are more or less the same we consider the amount of adjustment made during economic booms and downturns for the respective series Results are shown in Table 4 In terms of the GDP adjustment regulatory capital levels should increase with approx 1 31 during an economic growth period and decrease with 4 03 during a recession Having a look at the CREDIT adjustment results say that capital levels should increase with 0 86 in case of an upward trend and decrease with 0 88 in case of a downturn in the economy As the CREDIT variable adjustment makes no significant reasonable changes regarding Basel II regulatory capital we consider the output adjustment based on GDP growth to be the best smoothing procedure As mentioned earlier Figure 3 shows the GDP smoothed series together with the real GDP Note that this result is not due to the fact that GDP growth is one of the explanatory variables in our logit model V Conclusions In this paper we focused on finding the optimal method for mitigating the procyclical effects of Basel II capital regulation We analysed two major approaches regarding the issue which is estimating TTC PDs and fine tuning the output of the Basel II formula By building up a model economy that simulates the Romanian banking and finance sector we observed that a GDP growth based smoothing of Basel II capital requirements would be a good solution regar
16. nalysis regarding the methods of reducing the procyclical effects by the use of a logistic model containing a one month ahead probability of default PD Our goal is to find an answer to the fact that the cyclical effects can be effectively mitigated by fine tuning of the PD indicators or by the change of the capital requirements In the conclusion we seek for other problems that arise when allocating regulatory capital II Literature Review We analyzed a series of different papers about procyclicality and its issues Firstly we noticed that Jokivuolla Kiema and Vesala 2009 and Repullo and Suarez 2009 both examined the impact of the regulatory capital s procyclical effects Contrary to Repullo and Suarez Jokivuolla Kiema and Vesala created a comparison between the regulatory capital requirements of the Basel I and Basel II Their main question was that whether the risk based or the constant weights based regulatory capital requirement shows less procyclical impacts on the credit market They used a simplified model which interprets 3 types of market participants low risk profile investors high risk profile investors and risk free investors Their conclusion was that the optimal risk based capital is the least procyclical They added that the present Basel II s necessary direction for further development is the implementation of a higher venture capital Repullo and Suarez used a dynamic equilibrium model in which the banks can t access
17. s of GDP and MATURITY_4 are negative meaning that as the growth rate of real GDP and proportion of long term exposures increase in the model economy the PD decreases The coefficient of variable UTIL_3 shows that the higher the utilization of credit lines the higher the PD so liquidity constraints also seem to 463 play a role in a firm s default Summing up the analysis of the logit model we can say that firm s defaults increase during downturns and decrease during and economic upward trend PIT capital requirements Based on the results in Table 1 we compute the point in time PIT capital requirements Ki for each borrower and month using the formula k PD LGD EAD M the estimated probability of default PD and assuming a loss given default LGD of 45 as in the foundation IRB approach of Basel II The PIT capital requirements per unit of loans for each month is calculated via ky mg Li lit where 1 denotes the value of the loans to firm i at the end of the month t Figure 1 shows how PIT capital requirements evolve among the GDP in the observed period The cyclical effects can be easily captured by applying the Hodrick Prescott HP trend to the series with a lambda value of 500 Regarding the HP smoothed PIT capital series a significant cyclical variation can be observed with a gap of 5 82 between the peak and the worst point of the business cycle TTC capital requirements Adjusting the input of Basel II formula
Download Pdf Manuals
Related Search