Using Probability of Default on the GCC Banks: A tool for Monitoring Financial Stability. Mahmoud Haddad and Sam Hakim

Transcription

1 Using Probability of Default on the GCC Banks: A tool for Monitoring Financial Stability Mahmoud Haddad and Sam Hakim Abstract Our research investigates the role of Probability of Default (PD), market and idiosyncratic GCC banks equity variables in explaining the shape of the term structure of PD and its spreads. We investigated the level of the PD during the following events 1- The Lehman Brothers Default of 09/2008, 2- The Arab Spring 12/18/2010 and 2011, 3- War in Yemen 17/5/2016, 4- Embargo on Qatar 6/5/2017 and 5- Arrest of Emirs in Saudi Arabia 11/4/2017 to determine if these events impact the credit risk of the investigated banks. By analyzing PD across GCC banks which represent the financial sector we find that Abu Dhabi Commercial bank to have the highest risk of default among the five investigated banks. Therefore, we conclude that the banks PD represent important determinants not only of the level, but also of the slope and curvature of probability of default term structures. A closer inspection of the credit spread slope also reveals that it contains important information about the impact of the economic and geopolitical events and the PD provides useful insights into the theoretical predictions of the Merton (1974) model. 1. Introduction A primary objective of bank regulatory policy is to preserve the integrity of banks core transactions and savings deposits in the event of insolvency while allowing losses to accrue to hybrid forms of bank capital. The distress that many banks experienced during the recent financial and geopolitical crisis has brought renewed emphasis on the importance of their financial stability. Assessment of bank probability of default risk is important not only from an investor's viewpoint, but also for regulators gauging the risk of bank failure. Accurate modeling of bank default risk is also required for valuing the benefits that banks derive from implicit and explicit government guarantees. Since the financial crisis, numerous central banks have started publishing financial stability reports; and the International Monetary Fund conduct periodic financial sector stability assessments of their member countries1. Among the quantitative tools used for assessing financial stability, central banks and international financial institutions are increasingly relying on market-based risk measures, as demonstrated by Kupiec (2006), Bundell-Wignall and Roulet (2012), and Saldias (2012). In many applications of this kind, structural models of default risk 1 For example: Turkey : Financial Sector Assessment Program-Detailed Assessment of Observance of the Basel Core Principles for Effective Banking Supervision, IMF, Monetary and Capital Markets Department, February 8, P age

2 are used in which equity and debt are viewed as contingent claims on the assets of a bank following the seminal work of Merton (1974). The options embedded in the bank's equity and debt can then be valued using traditional option pricing techniques. Recent examples of bank default risk analysis based on the Merton model include Acharya, Anginer, and Warburton (2014) and Schweikhard, Tsesmelidakis, and Merton (2014), who study the value of implicit (too-big-to-fail) government guarantees. We propose to apply this methodology on the leading publically held banks in Dubai. The goal is to develop an indicator of financial health based on Merton s contingent claims framework referred as the distance-to-default (PD). Using this insolvency bank indicator, we will then examine its ability to explain the variation in a currently used market-based default probability variable implied from credit default swaps (CDS). While there is a large literature on explaining bond yield spreads to measure default premium, using CDS data to assess default probabilities is relatively new. Comparison between the two probability measures, and with similarly capitalized banks elsewhere will enable a discussion whether the banks in Dubai are being assessed an unreasonably high default premium to ensure against their failure. 2. Model and Data The DTD measure initially proposed by Merton and later commercialized by KMV now Moody s begins with the assumption that the total value of a banking firm is assumed to follow geometric Brownian motion: dv = μv dt + σv V dw where V is the total value of the bank, μ is the expected continuously compounded return on V, σ V is the volatility of firm value and dw is a standard Weiner process. The second critical assumption of the Merton model is that the bank has issued just one discount bond maturing in T periods. Under these assumptions, the equity of the bank is a call option on the underlying value of the bank with a strike price equal to the face value of the bank s debt and a time-tomaturity of T. The distance to default can be calculated as: PD = ln(v/f) + μ 0. 5σ V 2 σ V T T where F is the face value of the bank s debt, r is the instantaneous risk-free rate, μ is an estimate of the expected annual return of the bank s assets. The corresponding implied probability of default, is given by: P = N( DTD) Where N(.) is the cumulative normal distribution. The KMV-Merton model basically translates the value and volatility of a bank equity into an implied probability of default. These probabilities are then compared with the default probabilities implied from credit default swaps (CDS) for a particular bank. The CDS allow market participants to trade credit risks with 2 P age

3 each other. We focus on default probabilities rather than credit spreads because (i) they are not affected by additional factors such as liquidity, tax differences, and recovery rates; and (ii) prediction of the relative likelihood of default is often stated as the objective of bond ratings. The use of CDS data for the purpose of default prediction is not new and has been employed by Longstaff, Mithal and Neis (2004) and Duffie et al (2007). We use the information in credit default swap premia to extract a direct measure of default probabilities and compare it with the estimates obtained from our methods. The data for this study will be collected from Bloomberg and the banks quarterly reports. We focus on the top 5 banks in the GCC region, namely 1- Qatat National Bank, 2- National Bank of Kwait, 3- Al Rajihi Bank, 4- between 01/02/2006 and 21/04/2017, and will calculate their DTD for the past 3 month, 1 year and 5 years. We use the PD and 5Year-1Year spread, 5year- 3month spread and the 1Year- 3mmonth spread the 5-3time spread to compute the sterm structure of probability of default. 3. Results Insert Table 1: Table 1 represents the probability default statistics distribution of variables used in analysis of probability of default and probability of default 5 year 1 year, 5 years 3month and 1 year -3 month spreads for all 5 banks and Coefficient of Variation ranking. We also computed the coefficient of variation (SD/AV= standard deviation of probability of default divided by the average probability of default for category and the DS/DF as a measure of risk on the difference between the maximum and the minimum probability of Default. Using the average probability of default, we find the Samba to have the highest average followed by National bank of Kuwait, Qatar National Bank Al Rajihi Bank and Abu Dahbi Commercial Bank to have the lowest average probability of default. Using the coefficient of variation (SD/AV) the data showed Samba to have the highest level for variation per one unit of probability of default followed by Abu Dahbi Commercial Bank, Al Rajihi Bank, Qatar National Bank and National bank of Kuwait, to have the lowest coefficient of variation ratio. Table one also shows the 5year- y1year, the 5year- 3month and the 1year -3month spreads of the probability of default. Using the coefficient of variation on the spreads we find the Abu Dhabi Commercial bank 1year 3month spread to have the highest variation followed by the National bank of Kuwait on the 5year - 3month spread, the Al Rajihi Bank on the 1yera- 3month spread, the Qatar National Bank on the 1year- 3 month spread and the Samba bank on the 1year 3month spread. 3 P age

4 The results indicate that four of the five bank variability of probability of default is generated by short term probability of default and only the National bank of Kuwait was found to be impact by the long-term probability of default. Insert Figure 1a and 1b.: Figure 1a : Represents the 3month, 1Year and five years term structure of the probability of default for Qatar National bank. The graph shows insignificant 3-month short term and the 1 year medium term probability of default. The 5-year probability of default shows a significant variation and especially during the Lehman Brothers Default of 09/2008, the Arab Spring 12/18/2010 and 2011, War in Yemen 17/5/2016, and the Embargo on Qatar 6/5/2017. Figure 1b : Represents the 5Year -1Year probability of default spread, the 5year- 3month probability of default spread and the 1Year-3month probability of default spread for the Qatar National Bank. The graph shows insignificant 1year -3month short term spread which is view here as the cost of the probability of default for the short-term duration. It also showed significant size and variation of the cost during the medium and long term probability of default. It is noticeable here that during change in the economic and political events the spread is widened. Insert Figure 1a and 2b.: Figure 2a: Represents the 3month, 1Year and five years term structure of the probability of default for National Bank of Kuwait. Again, the graph shows insignificant 3-month short term and the 1 year medium term probability of default. The 5-year probability of default shows a significant variation and especially during the Lehman Brothers Default of 09/2008, the Arab Spring 12/18/2010 and 2011, War in Yemen 17/5/2016, and the Embargo on Qatar 6/5/2017. Figure 2b: Represents the 5Year -1Year probability of default spread, the 5year- 3month probability of default spread and the 1Year-3month probability of default spread for National Bank of Kuwait. The graph shows insignificant 1year -3month short term and 5year - 3month long tern spread which is view here as the cost of the probability of default for the short-term and long-term duration. However, it showed significant size and variation of the cost during the 5year -1year medium term probability of default. It is noticeable here that during change in the economic and political events the spread is widened to reflect the impact of these event on the probability of default. Insert Figure 3a and 3b.: Figure 3a: Represents the 3month, 1Year and five years term structure of the probability of default for Al Rajihi Bank. The graph shows insignificant 3-month short term and the 1 year medium term probability of default. The 5-year probability of default shows a 4 P age

5 significant variation and especially during the Lehman Brothers Default of 09/2008, the Arab Spring 12/18/2010 and 2011, War in Yemen 17/5/2016, and the Embargo on Qatar 6/5/2017. Figure 3b: Represents the 5Year -1Year probability of default spread, the 5year- 3month probability of default spread and the 1Year-3month probability of default spread for Al Rajihi Bank. The graph shows insignificant 1year -3month short term spread which is view here as the cost of the probability of default for the short-term duration. It also showed significant size and variation of the cost during the medium and long-term probability of default. It is noticeable here that during change in the economic and political events the spread is widened. Insert Figure 4a and 4b.: Figure 4a: Represents the 3month, 1Year and five years term structure of the probability of default for Samba. The graph shows insignificant 3-month short term and the 1 year medium term probability of default. The 5-year probability of default show a significant variation and especially during the Lehman Brothers Default of 09/2008, the Arab Spring 12/18/2010 and 2011, War in Yemen 17/5/2016, and the Embargo on Qatar 6/5/2017. Figure 4b: Represents the 5Year -1Year probability of default spread, the 5year- 3month probability of default spread and the 1Year-3month probability of default spread. For the Samba. The graph shows insignificant 1year -3month short term spread which is view here as the cost of the probability of default for the short-term duration. It also showed significant size and variation of the cost during the medium and long-term probability of default. It is noticeable here that during change in the economic and political events the spread is widened. Insert Figure 5a and 5b.: Figure 5a: Represents the 3month, 1Year and five years term structure of the probability of default for Abu Dhabi Commercial bank. The graph shows insignificant 3-month short term and the 1 year medium term probability of default. The 5-year probability of default show a significant variation and especially during the Lehman Brothers Default of 09/2008, the Arab Spring 12/18/2010 and 2011, War in Yemen 17/5/2016, and the Embargo on Qatar 6/5/2017. Figure 5b: Represents the 5Year -1Year probability of default spread, the 5year- 3month probability of default spread and the 1Year-3month probability of default spread. For the Abu Dhabi Commercial bank. The graph shows insignificant 1year -3month short term spread which is view here as the cost of the probability of default for the shortterm duration. It also showed significant size and variation of the cost during the medium and long-term probability of default. It is noticeable here that only during the 5 P age

6 Lehman Brothers Default of 09/2008 that the bank spread has widened significantly and higher the other investigated banks in the GCC region. 4. Summary and Implications The probability of Default can shed insights on the stability and instability of the GCC financial system. Four banks namely Qatar National Bank, National Bank of Kuwait, Al Rajihi Bank and Samba Bank exhibited similar reaction to the economic and political events that took place between 2006 and However, Abu Dhabi Commercial Bank reacted only to the financial crisis. We noted that Risks and the cost of risk are large during the global financial crises. It also showed a lag effect on the GCC banks. Having said that, authorities need to be aware of risks and they to work proactively to contain these risks. In general, regulators have a strong incentive to intervene well ahead of a bank s default. Bank insolvencies have substantial welfare costs, as reported in Hoelscher and Quintyn (2003). The fiscal costs associated with banking crises can range from 3 percent of GDP, as experienced in the United States, to as high as 50 percent of GDP, as experienced in Chile and Indonesia. As a result, many countries have in place prompt-corrective-action frameworks aimed at preventing bank failure. 4. References Acharya, Viral V, Deniz Anginer, and A Joseph Warburton, 2014, The end of market discipline? investor expectations of implicit state guarantees, Working paper, New York University. Blundell-Wignall A and C Roulet (2012), Business Models of Banks, Leverage and the Distance- To-Default, OECD Journal: Financial Market Trends, 2012/2, No 103. Duffie, Darrell, Leandro Saita, and Ke Wang, Multi-period corporate default prediction with stochastic covariates, Journal of Financial Economics 83, 2007 Duffie, Darrell: Measuring Corporate Default Risk. Oxford University Press, Francis Longstaff, Sanjay Mithal and Eric Neis: Corporate Yield Spreads: Default Risk or Liquidity? New Evidence from the Credit Default Swap Market, Journal of Finance, 2005, vol. 60, issue 5. Hoelscher, David, and Mark Quintyn, 2003, Managing Systemic Banking Crises, Occasional Paper No. 224 (Washington, International Monetary Fund). 6 P age

7 Merton R (1974), On the Pricing of Corporate Debt: The risk structure of interest rates, The Journal of Finance, Vol (2), pp Tudela M and G Young (2003), A Merton-Model Approach to Assessing Default Risk in UK Public Companies, Bank of England Working Paper, No 194. Kupiec, Paul H: Basel II: A Case for Recalibration. Division of Insurance and Research, Federal Deposit Insurance Corporation, Washington, DC. October, Saldias M (2012), Systemic Risk Analysis Using Forward-Looking Distance-To-Default Series, Federal Reserve Bank of Cleveland Working Paper, No Schweikhard, Frederic A, Zoe Tsesmelidakis, and Robert C Merton: The value of implicit guarantees, Working paper, Oxford University, P age

8 Table 1: Distribution of variables used in analysis of probability of default for all 5 banks and Coefficient of Variation ranking Qatar National Bank Qatar National Bank 3M_Df_PR 1YR_Df_PR5Y_Df_PR Rank 5y-1y 5y-3m 1y-3m Rank Ave 2.92E Var 1.29E E E E E E-07 SD 3.59E MIN 1.09E E E-05 MAX DF SD/AV SD/DF National Bank of Kuwait National Bank of Kuwait Ave 3.55E Var 1.63E E E E E E-07 SD 4.03E MIN 6.55E MAX DF SD/AV SD/DF Al Rajihi Bank Al Rajihi Bank Ave 2.19E Var 7.2E E E E E E-08 SD 2.68E MIN 5.96E E E-05 MAX DF SD/AV SD/DF Samba Samba Ave 4E Var 1.62E E E E E E-07 SD 4.03E MIN 3.21E MAX DF SD/AV SD/DF Abu Dhabi Commercial Bank Abu Dhabi Commercial Bank Ave Var 4.15E E E E E E-06 SD MIN 7.35E MAX DF SD/AV SD/DF P age

9 Figure 1a Qatar National Bank Term Structure of Default P age

10 Figure 1b Qatar National Bank Term Structure of Default of Spread P age

11 Figure 2a National Bank of Kuwait Term Structure of Default P age

12 Figure 2b National Bank of Kuwait Term Structure of Default of Spread P age

13 Figure 3a AL Rajihi Bank Term Structure of Default /31/06 1/31/07 1/31/08 1/31/09 1/31/10 1/31/11 1/31/12 1/31/13 1/31/14 1/31/15 1/31/16 1/31/17 13 P age

14 Figure 3b AL Rajihi Bank Term Structure of Default of Spread /31/06 1/31/07 1/31/08 1/31/09 1/31/10 1/31/11 1/31/12 1/31/13 1/31/14 1/31/15 1/31/16 1/31/17 14 P age

15 Figure 4a SAMBA Term Structure of Default /31/06 1/31/07 1/31/08 1/31/09 1/31/10 1/31/11 1/31/12 1/31/13 1/31/14 1/31/15 1/31/16 1/31/17 15 P age

16 Figure 4b SAMBA Term Structure of the Spread /31/06 1/31/07 1/31/08 1/31/09 1/31/10 1/31/11 1/31/12 1/31/13 1/31/14 1/31/15 1/31/16 1/31/17 16 P age

17 Figure 5a Abu Dhabi Commercial Bank Term Structure of Default /31/06 1/31/07 1/31/08 1/31/09 1/31/10 1/31/11 1/31/12 1/31/13 1/31/14 1/31/15 1/31/16 1/31/17 17 P age

18 Figure 5b Abu Dhabi Commercial Bank Panal 1;Term Structure of Default of Spread 1/31/06 1/31/07 1/31/08 1/31/09 1/31/10 1/31/11 1/31/12 1/31/13 1/31/14 1/31/15 1/31/16 1/31/17 18 P age