The recent study by Kenny S&P Evaluation Services entitled "Municipal Bond Defaults, The 1980's: A Decade in Review," which you commented on in your May 14 issue, is yet another attempt to define the default risk in municipal bonds. The press release for this impressive-looking document goes right to the bottom line; namely, that the default rate for the municipals is only 0.27% and for nonrated municipals it is only 1.93%. The Kenny conclusion on unrated defaults is therefore in line with the 2% rate found by the Public Securities Association in its Jan. 8, 1993, study and, thereby, helps to cement this number in the minds or investors as a true measure of default risk.
Nothing could be further from the truth.
The calculation technique used by both the PSA and Kenny studies was successfully used by Michael Milken in the 1980s as part of a campaign to convince investors that junk bonds were safe. Curiously enough, Milken also came up with a 2% default rate by selectively counting default amounts and backloading the issue volume denominator with recent issues that had not had sufficient time to show their true default rate.
In fact, Kenny enhances this technique by cutting off its counting of defaults at yearend 1991, thereby leaving out the $2 billion of 1992 defaults on bonds issued during Kenny's study period (most notably the $635 million in issues backed by Mutual Benefit Life).
Aside from a skewed yardstick for measuring default rates, the study is fatally flawed in other respects:
* Kenny's data base of defaults is too incomplete to make a default percentage calculation. For the time period in question, they used $3.3 billion in rated and $2.9 billion in nonrated defaults. Using their selection criteria, we counted $5.3 billion and $4.2 billion, respectively, in defaults for the same time period. These differences directly understate the default numerators by 23.8% and 77.3%, respectively.
* The statement that Washington Public Power Supply System, LTV Corp., and Executive Life Insurance Co. accounted for $4.68 billion, or 54.10%, of the rated and nonrated defaults is repeated several times in an apparent attempt to have the reader conclude that without these extraordinary events, the modest default rates would be only half as much. This is wholly misleading. The WPPSS and LTV bonds that total $2.8 billion were almost all issued before 1980 and almost all were rated. As such, these bonds were excluded from the default rate calculation for rated bonds and had no effect on the calculation of the unrated default rate.
* Kenny chose to ignore defaults where a third-party guarantor had to suffer the default loss. There would have been nothing wrong with such an approach had they also ignored the issue volume from the default equation numerator. This is not an inconsequential oversight since such bonds make up 50% of the municipal bond market. Therefore. by definition, one can double their reported default rate for rated bonds based on this discrepancy alone.
The one-sided bias of the discrepancies noted in the Kenny study is indicative of someone who sets out to prove a preconceived default rate rather than make an objective determination. Certainly, whenever one divides defaults by issue volume, one comes up with a percentage they can then call a default rate. Under the Kenny methodology, however, one must conclude that Kenny's default rate bears no relationship to the real world.
Richard Lehmann President Bond Investors Association
As the authors of the special report "Municipal Bond Defaults, The 1980's: A Decade in Review." we would like to commend The Bond Buyer for its extensive coverage of our findings in your May 14, 1993, edition. We believe this topic to be of vital interest to the industry. Getting message out to the general public that the majority of municipal bonds are unlikely to default is of paramount importance.
However, we feel we must take exception to a misstatement of the facts as presented in an argument attributed to Richard Lehmann. Mr. Lehmann apparently contends that our study "is ~fatally flawed' because it includes bonds backed by third parties, including insured bonds, in total issuance volume. But the study excludes defaults in such "bonds, making the calculated default rates ~an apples to oranges comparison.'"
Mr. Lehmann is wrong in his assumptions on several counts. We used a widely accepted industry standard in determining what constitutes a monetary default -- namely, a loss of money to a bondholder. Only those bonds that missed a payment to the bondholder (whether it be principal or interest) are in true default and should be counted as such. Therefore, although third-party "enhanced" bonds are counted in our original base of bonds (as well they should be, seeing that they meet the criteria of being issued), they are not and should not be counted in the default ratio simply because the payments to bondholders were never missed! The insurer, or enhancement provider, made the bondholders whole. Thus, a real default did not occur.
If we had decided not to count insured/enhanced bonds in the denominator of our calculation, the results would have been seriously skewed. This negative (but inherently incorrect) emphasis presumably would have been acceptable to Mr. Lehmann but it would have been contrary to the proper, application of accepted principles.
We are proud of our contribution to this very important topic of tax-exempt debt that defaults. We drew upon a data base of over 1 million Cusips representing over 150,000 issues and we stand by our analysis and the accuracy of our conclusions.
Donald K. Cirillo Director, Securities Analysis
Thomas F. Jessop Senior Research Analyst Kenny S&P Evaluation Services
