Areas Where Further Work Will Improve the Methodology
Category: Risk Management in Banking
The banking industry is clearly evolving to a higher level of risk management techniques and approaches than had been in place in the past. Yet, as this review indicates, there is significant room for improvement. Before the areas of potential value added are enumerated, however, it is worthwhile to reiterate an earlier point. The risk management techniques reviewed here are not the average, but the techniques used by firms at the higher end of the market. The risk management approaches at smaller institutions, as well as larger but relatively less sophisticated ones, are less precise and significantly less analytic. In some cases they would need substantial upgrading to reach the level of those reported here. Accordingly, our review should be viewed as a glimpse at best practice, not average practices.
Nonetheless, the techniques employed by those that define the industry standard could use some improvement. By category, recommended areas where additional analytic work would be desirable are listed below.
A. CREDIT RISK
The evaluation of credit rating continues to be an imprecise process. Over time, this approach needs to be standardized across institutions and across borrowers. In addition, its rating procedures need to be made compatible with rating systems elsewhere in the capital market.
Credit losses, currently vaguely related to credit rating, need to be closely tracked. As in the bond market, credit pricing, credit rating and expected loss ought to be demonstrably closer. However, the industry currently does not have a sufficiently broad data base on which to perform the migration analysis that has been studied in the bond market.
The issue of optimal credit portfolio structure warrants further study. In short, analysis is needed to evaluate the diversification gains associated with careful portfolio design. At this time, banks appear to be too concentrated in idiosyncratic areas, and not sufficiently managing their credit concentrations by either industrial or geographic areas.
B. INTEREST RATE RISK
While simulation studies have substantially improved upon gap management, the use of book value accounting measures and cash flow losses continues to be problematic. Movements to improve this methodology will require increased emphasis on market-based accounting. However, such a reporting mechanism must be employed on both sides of the balance sheet, not just the asset portfolio.
The simulations also need to incorporate the advances in dynamic hedging that are used in complex fixed income pricing models. As it stands, these simulations tend to be rather simplistic, and scenario testing rather limited.
C. FOREIGN EXCHANGE RISK
The VaR approach to market risk is a superior tool. Yet, much of the banking industry continues to use rather ad hoc approaches in setting foreign exchange and other trading limits. This approach can and should be used to a greater degree than it is currently.
D. LIQUIDITY RISK
Crisis models need to be better linked to operational details. In addition, the usefulness of such exercises is limited by the realism of the environment considered.
If liquidity risk is to be managed, the price of illiquidity must be defined and built into illiquid positions. While this logic has been adopted by some institutions, this pricing of liquidity is not commonplace.
E. OTHER RISKS
As banks move more off balance sheet, the implied risk of these activities must be better integrated into overall risk management and strategic decision making. Currently, they are ignored when bank risk management is considered.
F. AGGREGATION OF RISKS
There has been much discussion of the RAROC and VaR methodologies as an approach to capture total risk management. Yet, frequently, the decisions to accept risk and the pricing of the risky position are separated from risk analysis. If aggregate risk is to be controlled, these parts of the process need to be integrated better within the banking firm.
Both aggregate risk methodologies presume that the time dimensions of all risks can be viewed as equivalent. A trading risk is similar to a credit risk, for example. This appears problematic when market prices are not readily available for some assets and the time dimensions of different risks are dissimilar. Yet, thus far no one firm has tried to address this issue adequately.
Finally, operating such a complex management system requires a significant knowledge of the risks considered and the approaches used to measure them. It is inconceivable that Boards of Directors and even most senior managers have the level of expertise necessary to operate the evolving system. Yet government regulators seem to have no idea of the level of complexity, and attempt to increase accountability even as the requisite knowledge to control various parts of the firm increases.