Economic models used to calculate market risk contain a fundamental flaw that could cause banks to underestimate their capital requirements.
Federal Reserve Bank of Chicago economist Subu Venkataraman finds that almost all risk management models produce a value-at-risk figure, which tells traders how much they are likely to lose. For instance, the model for a $1 million securities portfolio may produce a VAR of $100,000 with a 95% probability. This means that 5% of the time the trader will lose more than $100,000; the rest of the time, he will either lose less or make money.
These models work by assuming banks are equally likely to gain or lose money. This produces the famous bell-shaped curve at the heart of modern statistics.
Mr. Venkataraman, however, says this assumption is wrong because extreme price fluctuations occur much more frequently in real life than the models predict. For example, most models would predict that the 1987 stock market crash could occur once every 5,900 years.
Banks can fix this statistical problem if they use a new technique, known as a mixture-of-normals approach, to calculate value-at-risk. This technique assumes large losses occur frequently in the market.
To support his conclusion, Mr. Venkataraman used the new approach to calculate the VAR from 1978 to 1996 for a foreign currency portfolio. He then compared the results to actual price swings. The new technique did a better job than the standard model of predicting large price fluctuations and, thus, prudent capital levels. For a copy of "Value at Risk for a Normal Distribution," call 312-322-5111.
Newly chartered banks need an average of nine years before they earn profits as efficiently as their more established rivals, according to a recent study by the Office of the Comptroller of the Currency.
Agency economist Robert DeYoung and New Jersey Institute of Technology Prof. Iftekhar Hasan find new banks initially are four times less efficient at making money than established banks, but most of the discrepancy disappears after three years. The study is based on data from 1988 to 1994. For a copy of "Economic Working Paper 97-3," call 202-874-5000.
Regulatory restrictions, not market forces, are behind the recent consolidation of banks with less than $100 million of assets, according to two New England economists.
A study of mergers by Federal Reserve Bank of Boston economist Eric S. Rosengren and Boston College Prof. Joe Peek concludes that borrower concentration limits are preventing small banks from competing for loans above $1 million. To overcome this regulatory obstacle, they write, small banks are combining to create larger institutions. For a copy of "Have Borrower Concentration Limits Encouraged Bank Consolidation?" call 617-973- 3000.
High debt burdens will not cause consumers to cut back on credit card spending, says another study.
Jonathan McCarthy, an economist at the Federal Reserve Bank of New York, examined the historical relationship between household debt and spending.
"If spending slows down, the cause is more likely to be weak income growth than high household debt," he notes. And higher delinquency rates may indirectly affect lending by causing banks to restrict credit. For a copy of "Debt, Delinquencies, and Consumer Spending," call 212-720-6134.
