Readers’ Questions Answered on Credit Risk

Q: There is a lack of standardized credit information on counterparties particularly in developing countries. What suggestions do you have for taking decisions on this?

A (Mat Newman, Sungard): The difficulty in finding good reliable credit data on counterparties in developing countries is common to many banks. When a bank does not have a strong regional presence and an internal history of credit data, and the major ratings agencies do not provide sufficient coverage, you must proceed with caution. Simply using a default model built using data from one economic region (e.g. the US or the EU) on counterparties in another region will not work. At a minimum, differences in accounting standards and practices must be reflected in the model’s inputs. Where there is a lack of data the qualitative side of credit assessment becomes more important, and intimate knowledge of prospective credits must be obtained before lending decisions are undertaken.

A (Veena Gundavelli, Emagia): As global markets continue to grow in areas like China, India, Europe and Latin America, a pressing need arises to manage credit risk. This is often a difficult task because of the limited amount of standardized credit information available for many companies based in developing countries. What companies should do to address this is to look at multiple aspects of the potential business relationship – past payment history (if available), trade references, bank references, payment profiles from similar companies in the region, and get input from in-country personnel who are in close contact with the customers (read Mitigating Corporate Credit Risk).

A (Sara Statman, Statman Consulting): In 1975, Standard & Poor’s rated seven sovereign bonds; by 1990 they rated 31 sovereign bonds, nine of which were from developing countries. Historical default data is even sparser and analysis is complicated further by a lack of common accounting standards, poor transparency, corruption, taxes and governance. A growing trend to overcome these limitations appears to be the use of structured finance. CVRD, Brazil’s iron ore exporter, experienced no direct relationship to the sovereign in terms of price and risk issues and in 2002 issued unsecured debt wrapped with a liquidity facility to mitigate transfer and convertibility risks. Political risk insurance, partial guarantees or deals wrapped with full financial guarantees are also being employed, as they provide investors with assurance and the cost for issuers is lower than for unwrapped deals.

For non-structured deals a traditional approach can be used based on available data, to create a scoring model through the financial analysis of: income statements; balance sheets; capital structure and liquidity; cash flows; peer group comparison; industry outlook; management experience; risk management controls; market share; size and geographic diversity. Adjustments will need to be made to standardize data inputs.

For issuers with publicly traded bonds or equities, current market prices are available and market price methods may be applied. Emerging market bond indices can be used as a benchmark to measure the general credit environment. Current prices should reflect any credit deterioration. Credit events are harder to account for as they tend to occur too rapidly for any corrective action to be taken.

Q: Many treasures make decisions based on ‘gut feeling’ or ‘past experience’ rather than any scientific model. This appears to be working well in some cases. What are your opinions on credit decision taking – is it a science or an art?

A (Mat Newman, Sungard): There will always be a trade-off between the quantitative side of credit ratings and the qualitative side. Where good, reliable, historic data exists the emphasis should be placed on quantitative models over ‘gut feeling’. Of course the credit process should allow for deviation from the models, but only with sufficient justification and an explanation for why the model score is not appropriate. In the absence of compelling evidence to the contrary a credit score should only be able to be adjusted one notch up or down for purely ‘gut feeling’ reasons.

A (Sara Statman, Statman Consulting): A degree of qualitative input will always be important. Models are imperfect as they often make simplifying assumptions and are subject to modelling error. Qualified, experienced professionals should treat model outputs as a starting point for decision-making. However, as financial products and structures have increased in complexity, so too have the tools and models with which the associated risks are analyzed. To assign the appropriate qualitative oversight, end users will have to become far more quantitative in order to understand the credit exposures at hand (e.g. credit exposure to currency swaps is far greater than to interest-rate swaps) as well as the benefits and limitations of the models in use. Credit decision-making is and always will be both an art and a science.

Q: How can I determine effective hedging on credit risk? How can I measure credit risk and account for credit risk?

A (Mat Newman, Sungard): To effectively hedge credit risk you must first build a reliable model for measuring the credit risk. Typically this will involve building a Monte Carlo simulation engine which will simulate many thousands of possible future market scenarios, re-value the banks positions under these scenarios, and determine the likelihood of each counterparty defaulting on their obligations. The resulting distribution of possible future exposures and losses is then used to compute economic capital, and to highlight where credit risk mitigation may be most effective through the use of hedging tools such as collateral or credit derivatives.

Also read Sara Statman’s article, Guide to Credit Risk Measurement


Related reading