Portfolio Management and Risk Limits
The need for credit portfolio management emanates from the necessity to optimize the benefits associated with diversification and reduce the potential adverse impact of concentration of exposures to a particular borrower, sector or industry. The conventional approach to credit portfolio management has been largely based on the counter party exposure limits, which largely provides the guideline for incremental asset/ exposure build-up. This “forward” or incremental approach to credit portfolio management is, to an extent, a reactive strategy and though it does guide the decision-making process, it has limited contribution for managing the existing credit portfolio of the bank.
The recent developments in the measurement and management of portfolio credit risk have been based on two key attributes: correlation and volatility. Consider two companies, one operates large capacities in the steel sector and the other a large player in the cement sector, promoted by two entirely unrelated promoters. Though these would classify as two separate counter parties, both of them may be highly sensitive to the Government’s expenditure in new projects/ investments.
Thus, reduction in Government investments could impact these two companies simultaneously (correlation), impacting the credit-quality of such a portfolio (volatility), even though from a regulatory or conventional perspective, the risk had been diversified (2 separate promoters, 2 separate industries). Thus, though the credit portfolio may be well diversified and fulfils the prescribed criteria for counter party exposure limits, the high correlation in potential performance between two counter parties may impact the portfolio quality (default levels) under stress conditions.
In addition to the widespread instances of high correlation and resultant volatility, the emergence of new techniques for managing a bank’s credit portfolio have actively contributed to the development and adoption of broader credit risk management practices. Specifically, the adoption and wider acceptance of securitisation of loan assets in the developed markets has permitted banks to pursue credit portfolio management on a proactive manner. These have usually been in the nature of collateralised loan obligations.
Though securitisation of loan receivables, mainly consumer and auto loans, has been prevalent in India, it has usually been deployed as an asset acquisition or a hive-off approach rather than active credit portfolio management. The steps taken to enhance the liquidity and depth of debt markets in India and simplify the process of securitisation are expected to improve the prospects of credit portfolio management in the near future.
The measurement of credit-portfolio concentration has been elaborated in detail in the regulatory prescriptions for counter-party exposures in India. The issue of credit portfolio correlation is discussed here in some detail. In statistical terms, credit portfolio correlation would mean the number of times companies/ counter-parties in a portfolio defaulted simultaneously. As evident, this analysis is impossible in practice and the number of such instances for developing a reasonable generalisation would be too few.
Some credit portfolio management techniques developed overseas estimate the correlation between defaults and bond-market spreads and generalise this for assessing the correlation in a given portfolio. Given the limited data on corporate bond market spreads and their statistical linkages with ratings in India, this approach may not be appropriate for Indian banks at this stage. Banks should, however, direct their data management efforts in this direction so that a beginning in active portfolio management can be made.
One possible technique for analysing credit-portfolio correlation is based on a macro-economic factor model. This approach involves projecting the performance (volatility) of a credit portfolio under altering macro-economic environments. In specific terms, this would involve a stress-test on the debt-servicing ability of a portfolio of borrowers under alternative scenarios. The input data would consist of the projected financial performance (income statement and balance sheet details) of each of these portfolio constituents.
In the appraisal system adopted by Indian banks in general, these are normally developed for individual borrowers for seeking credit approvals from the specific internal authorities. Such portfolio constituents could be relatively independent counterparties, spanning a relatively wide spectrum of region, industry, size of operation, adoption of technology and promoters. The key financial parameters of these counterparties (growth, profitability, access to funds, etc.) should be linked to the macro-economic parameters under consideration.
Some of the relevant macro-economic parameters could include overall growth rates, growth in exports/ industrial/ agricultural sectors, interest rates, exchange rates, import duties, equity market and liquidity conditions. By developing alternative scenarios for these parameters, the credit-portfolio’s aggregate performance (default rates and levels) can be assessed and possible correlation between a set of obligors (even though they constitute entirely separate counterparties) may be established.
For instance, under the assumptions of low overall economic growth, poor growth in agriculture sector and reduction in import duties, the assessment may give some correlation between the borrowers in the petrochemicals industry and the consumer-electronics industry. Though there may not be any “counter party” relationship in this set of borrowers, both of them are possibly susceptible to reduced import duties and low economic growth. These illustrations are relatively simplistic and the detailed analysis, as discussed above, may give critical inputs for minimising the credit risk of the given portfolio.
A possible advantage of starting with the macro-economic factor model is that it is amenable to the current levels of credit risk assessment practices in Indian banks and can be correspondingly adopted with relative ease. Though superior and more sophisticated tools have been developed, their findings may be limited due to the lack of representative data. Such options can be considered as Indian banks further enhance their internal systems and processes in credit risk management.