The financial crisis has exposed serious weaknesses in risk measurement and management practices. This column argues for both market practitioners and their supervisors to make concerted efforts to achieve a more integrated measurement and management of different forms of risk.
The financial crisis that began in August 2007 has exposed, inter alia, serious weaknesses in risk measurement and management practices. One key weakness was the lack of “enterprise risk management”, by which we mean the integrated measurement and management of all relevant risks without regard to where they may lie in the organisation.
The Senior Supervisors Group (2008) found, for example, that banks that were more severely affected by the crisis tended to have difficulties integrating certain market and credit risks across business lines, whereas firms that were less affected practiced a more comprehensive approach to risk management. In 2006, well before the crisis erupted, the Research Task Force of the Basel Committee on Banking Supervision initiated a major project to study the interaction of market and credit risk in the context of an integrated approach to risk measurement and management. The project eventually involved 25 researchers from 12 countries based in 18 policy authorities.
The fruit of this project – published as Basel Committee on Banking Supervision (2009) – sheds light on this critical aspect of bank stability and regulation (for a more complete summary, see, the underlying research papers forthcoming in the Journal of Banking and Finance edited by Hartmann, Pritsker and Schuermann).
The covariance of market and credit risk matter: The compounding effect
A central result supported by several research papers is that the way in which banks aggregate, or sum up, market and credit risk is very important for the precision with which they capture their overall risk exposure. In particular, if the two types of risk are not considered in an integrated fashion “from the bottom up”, starting at the level of individual instruments and exposures, overall risk estimates can exhibit significant biases. “Top down” aggregation approaches, in which market and credit risk are calculated separately for different positions and combined only at a higher level – for example at the level of the bank as a whole – neglect how closely the two types of risks are related and that their relationships can be highly non-linear. As a consequence, either substantial underestimation or significant overestimation of total risk can occur (Alessandri and Drehmann, 2007; Breuer et al., 2008; and Kupiec, 2007).
For example, summing the separately calculated risks at a higher level – thereby assuming that they are perfectly correlated – may, in contrast to much conventional wisdom, not be conservative. A particularly clear illustration of how a large underestimation of risk can occur through “top down” aggregation is given by foreign currency loans (see Breuer et al., 2008), which constitute a sizable part of lending in some countries.
Consider a bank lending in foreign currency to domestic borrowers. These positions contain market risk (exchange rate risk) and credit risk (default risk of borrowers). Consider the two risks separately.
- When, for example, the domestic economy slows, ceteris paribus, the probability of domestic borrowers defaulting increases.
- When, for example, the domestic currency depreciates, ceteris paribus, the value of the loan in domestic currency increases as it is denominated in foreign currency.
Thus, on the surface a risk manager could believe that the two effects tend to offset each other. But this reasoning would neglect the strong relationship between exchange rate changes and default risk in this type of contract.
More specifically, the ability of a domestic borrower to repay a loan in foreign currency depends in a non-linear way on fluctuations in the exchange rate. Indeed, research suggests that home-currency depreciation has a particularly malign effect on the repayment amount (and therefore repayment probability) of a foreign currency loan by an unhedged domestic borrower – even stronger than the valuation effect mentioned above. As a result, a “compounding effect” emerges (overall risk is greater than the sum of the separately measured different risk components).
Estimates of the bias from the compounding effect
An analysis of foreign currency loans in Austria indicates that simply adding up the separately measured exchange rate and default risk components underestimates the actual level of risk by a factor of several times. For example, for a B+ rated obligor, the integrated risk measurement approach leads to an overall risk that is 1.5 to 7.5 times larger than the risk derived from a compartmentalised approach in which each risk is measured separately and then added up. This bias becomes more pronounced for portfolios with lower ratings. Importantly, foreign currency loans are not an isolated example for such possible compounding effects. Other examples include adjustable rate loans (including sub-prime mortgage loans) or matching long and short positions in OTC derivatives.
Diversification effects are also possible
It is important to understand that the bias in assessing market and credit risk is not necessarily one-sided. Diversification gains are clearly possible, and whether risk diversification or risk compounding effects dominate is case-specific.
An example where calculating market and credit risk separately and then summing the risks may be conservative – overestimating total risk – is provided by Alessandri and Drehmann (2007). Using UK data, these authors examine interactions between interest rate and credit risk in the banking book, including both on- and off-balance sheet items. For example, when interest rates rise, more borrowers default and margins are initially compressed as short-term borrowing rates react faster than long-term lending rates.
Over time, however, banks may succeed in passing higher interest rates as compensation for increased credit risk on to borrowers and thus margins recover. The authors’ results suggest that significant diversification benefits between interest rate and credit risk in the banking book may be achieved through this mechanism. Put differently, their data and model indicate that simply adding up the separately calculated market and credit risk components can imply a large upward bias in the estimation of overall risk over longer horizons.
Practical challenges
While the message that integrated risk measurement and management of market and credit risk from the bottom up is superior to compartmentalised top-down approaches is clear in principle, the practical application of this principle faces significant challenges.
- First, perhaps the most important obstacle to the further integration of market and credit risk assessments is reconciling the different time horizons over which risks are measured.
Market risk horizons tend to range between daily and two weeks, whereas credit risk horizons tend to be around one year.
- A second obstacle is that all gains and losses emerging from both types of risks must be measured in a consistent fashion (Kupiec, 2007).
One aspect of this problem is that both types of risks must be captured in the same metric. Currently, with market risk is frequently measured using a volatility metric and credit risk is measured using a currency loss metric. Another requirement is that all relevant cash flows are included. For example, credit risk measurement should include both losses on bad loans and revenues on performing loans, and market risk measurement should also include funding costs. Today, such cash flows are generally not included in banks’ risk measurement and management systems.
- Last, and certainly not least, it must be understood that integrated risk measurement and management make very significant demands on data and technological infrastructure, demands that are likely to be quite expensive to meet.¹
Conclusion
In conclusion, while properly integrating market and credit risk is indeed quite challenging and there are still important unknowns, the experiences of the current crisis and the results of extant research strongly indicate the on-going need for both market practitioners and their supervisors to make concerted efforts to achieve a more integrated measurement and management of different forms of risk.
References
•Alessandri, P., and M. Drehmann (2007), An economic capital model integrating credit and interest rate risk, mimeo., Bank of England.
•Basel Committee on Banking Supervision (2009), Findings on the interaction of market and credit risk, Working Paper, no. 16, May.
•Breuer, T., M. Jandacka, K. Rheinberger and M. Summer (2008), Regulatory capital for market and credit interaction: Is current regulation always conservative?, Deutsche Bundesbank Discussion Paper Series 2, no 14.
•Hartmann, P., M. Pritsker and T. Schuermann (2009, eds.), Interaction of market and credit risk, Special Issue of the Journal of Banking and Finance, forthcoming.
•Kobayashi, S. (2007), Relationship between liquidity and credit risk from a viewpoint of searching and bargaining, working paper, Proceedings of the 27th Japanese Association of Financial Econometrics and Engineering (JAFEE) conference.
•Kupiec, P. (2007), An integrated structural model for portfolio market and credit risk, working paper, Federal Deposit Insurance Corporation.
•Masschelein, N., and K. Tsatsaronis (2008), Measuring default risk in the trading book, National Bank of Belgium Financial Stability Review, 163–172.
•Rosenberg, J., and T. Schuermann (2006), A general approach to integrated risk management with skewed, fat-tailed risks, Journal of Financial Economics, 79, 569–614.
•Sawyer, N., and R. Davidson (2009), The shock of the interaction, Risk, June, 60-62.
•Senior Supervisors Group (2008), Observations on risk management practices during the recent market turbulence, 6 March.
1 The working group also addressed the role of liquidity in the interaction of market and credit risk, which is not covered in this article for reasons of space. See Basel Committee on Banking Supervision (2009) and, for example, Kobayashi (2007).
Leave a Reply