A few months ago, Timothy Ryan, director of the Office of Thrift Supervision, declared that we were "in the eighth inning" of the bailout process.
If he was right, then we are now going into extra innings.
In my opinion, the OTS and the Resolution Trust Corp. have been providing overly optimistic assessments of the condition of the thrift industry and the progress of its cleanup.
As far back as an American Banker comment on Oct. 3, 1990, I pointed to underreporting of the losses of thrifts placed in receivership and those sold before the quarterly reporting dates for the thrift industry.
Significant Understatement
The most recent and perhaps disturbing of OTS and RTC pronouncements has been an apparently significant understatement of the number and size of problem thrifts.
Based on my firm's analysis-- which the thrift regulators arc bound to dispute - the OTS significantly understated the number and asset size of the Group IV thrifts (those likely to be transferred to the RTC) in both the third and fourth quarters of 1991. The undercounts may be as great as 50%.
There has been considerable speculation about how the thrift crisis was handled during the election year of 1988. The basic allegation is that the true nature of the crisis was understated so that it would not become a major campaign issue.
When President Bush took ofrice in January 1989, the thrift industry cleanup became his No. 1 domestic issue.
Early Warning
A charge of cover-up, for lack of a better word, during or before 1988 is extremely difficult to prove. If there ever was one, it occurred at the beginning of the crisis in the early 1980s.
David Stockman, former director of the Office of Management and Budget, said in March 1981 that "any honest evaluation of the savings and loan industry would show that their equity has been wiped out."
In 20/20 hindsight, if We had acted to clean up the thrift problems in the early 1980s, we would have significantly reduced the scope and cost of the present bailout.
Now we face the fact that the number of commercial and savings banks on the Federal Deposit Insurance Corp.'s problem list remained constant at about 1,045 for the quarters ended December 1990 and March 1991.
After dropping slightly to 1,033 for the June 1991 quarter, the number increased by 5% to 1,081 by the end of the third quarter, but fell back by just 12, or 1%, to 1,069 at yearend.
Assets Swell at Sick Banks
In a sign that big banks are increasingly endangered, the assets of problem banks rose by more than 20% in both the third and fourth quarters. By contrast, the OTS reported significant drops in both the number and total assets of the lowest-ranked thrifts during the same time frame.
The OTS puts in its Group IV the thrifts most likely to require federal assistance, but the classification is somewhat subjective. Some in the group may have Macro ratings, similar to the Camel rating that applies to bank condition, of 3, but most are rated 4 or 5.
Associations in Group III, troubled but not on the verge of seizure, usually have Macro ratings of 3 or 4, but may sometimes be 2 or 5.
We, not the OTS, prefer to use the term "problem thrift" to refer to those in Group IV. OTS representatives maintain that the number of troubled thrifts may exceed that in Group IV, because not all troubled thrifts will, in the OTS's judgment, require federal assistance.
Some troubled thrifts with capital deficiencies have capital plans that do not call for government assistance, especially with an improving economy and low interest rates.
Clearer Criteria for Banks
The FDIC appears to use more objective criteria in defining a problem bank. It defines problem banks as those with financial, operational, or managerial weaknesses that threaten financial viability. In the absence of corrective measures, these banks could require disbursement of insurance funds.
We learned from the FDIC that all banks with a 4 or 5 rating would be on the problem list - a much less subjective definition than the OTS uses.
While FDIC problem banks increased 5% in the third quarter of last year, the OTS reported that the number of Group IV thrifts dropped 33%, to 79 from 118. Of the net change of 39, 37 left Group IV for the RTC and two were acquired. The number of downgrades into and upgrades out of Group IV canceled each other out.
And while FDIC problem bank assets were gaining 21% in the 1991 third quarter, Group IV thrift assets fell by $17 billion. or 21%..
OTS Group IV downgrades and upgrades also offset each other in last year's second quarter, but the-ratio of downgrades to upgrades was 4 to 1 in the first quarter.
Thus, it can be argued that the Group IV undercount began even earlier last year. With the exception. of 1990's second quarter, this ratio was 2.5 to 1 in 1990.
Stark Contrasts
The OTS reported that Group IV thrifts declined a net of 14 more, or 18%, during the fourth quarter of 1991. There were 14 RTC takeovers, two acquisitions, 12 downgrades into Group IV, and 10 upgrades out of the group. Group IV thrift assets fell by $14 billion, or 22%.
Meanwhile, the FDIC had a mere 1% drop in the number of problem banks but a 26% asset increase.
Thus there were significant trend differences in the FDIC and OTS problem classifications during the third and fourth quarters last year. While there clearly are differences between banks and thrifts and their problems, such a stark contrast is not expected, and is therefore worthy of further examination.
To gauge the extent of what we believe is an undercount in the number and assets of problem thrifts, we used a proxy to measure Group IV thrifts.
This proxy, which has proved quite consistent over the past several years, is the lowest, Rank I thrift group as defined quarterly by IDC Financial Publishing Inc. in its S&L-Savings Bank Financial Quarterly.
Thrifts with a rank of 1 have the highest probability of failure, and the IDC ranking has predicrive value: Since 1985, 92% of thrifts that failed had a rank of 1 before they failed.
A full 96% of the thrifts taken over by the RTC during 1989 and 1990 had a rank of 1. More recently, 80% of thrifts sent to the RTC in 1991 and 87% in the first quarter of 1992 had a rank of 1.
The IDC data, published with a roughly one-quarter lag, don't provide numerical ratings for thrifts that fail during the quarter of publication. Thus, the number of Rank 1 thrifts, which almost always includes these failures, is usually less than the number of OTS Group IV thrifts at the end of any quarter.
For example, none of the 15 thrifts (13 of which were ranked 1 ) that failed in the first quarter this year had an IDC Rank 1 as of Dec. 31, 1991, but all 15 were probably OTS Group IV thrifts at that time.
The significant drop in the number and asset size of Group IV thrifts during the third quarter of last year significantly contrasts with the IDC increase in both Of these categories. Both sources reported a drop in the number of categorized thrifts during the fourth quarter, but differed in their aggregate assets.
A Rise, Not a Fall
How much might the number and assets of Group IV thrifts have been understated for the last two quarters of 1991 ?
Based on the the IDC predictor. the number of Group IV thrifts for the third quarter of 1991 would have not decreased but rather increased to 149, with a total asset size of $90 billion. OTS Group IV thrifts as reported were 47% below the predicted number, and their assets 30% below.
Using the same approach for the fourth quarter of 1991, we concluded that the number of Group IV thrifts would have dropped from the predicted level of 149 to 125 by the end of that quarter.
The actual OTS number is therefore 48% below the predicted level. The predicted asset size would again be $90 billion, leaving the OTS number 46% too low.
Another way to look at the discrepancy is that Group IVthrifts, rather than accounting for 5% of private-sector industry assets, would have had 10%.
Convincing Evidence
To the extent that the IDC Rank 1 is a good proxy for the OTS Group IV, then there appears to have been a significant understatement in both the number and asset size of Group IV thrifts by the OTS during the third and fourth quarters of last year.
To the extent that there would be an expected relationship between the number and asset-size trends of problem banks and thrifts regulated by the FDIC and the OTS, then trends in the FDIC data for the last two quarters of 1991 would provide additional support for this claim.
We believe the best support, however, is found by looking at the implications for a state we know well, namely Florida, which has consistently ranked No. 1 or 2 in number of Group IV thrifts.
An analysis by my firm has led me to conclude there is no question that the OTS understated by at least 50% the number of Group IV thrifts in Florida for both the third and fourth quarters of 1991.
By our reading of OTS data, all Group IV thrifts in Florida apparently should have been closed, and the thrift crisis there would be over, which is certainly not the case.
A Breach of Faith?
Who cares if the number or asset size of problem thrifts has been understated?
First, it raises the question of the reliability of OTS quarterly industry releases and the administration's progress in resolving the thrift crisis.
Second, many people, especially in Congress, use problemthrift data as a measure of the thrift industry cleanup and the extent to which additional resources should be directed there.
Putting aside other public policy concerns, we find particularly troubling the several overly optimistic assessments of the thrift industry and the progress of its cleanup over the past year or so.
The OTS and RTC will flatly deny any such criticisms, particularly in an election year; An OTS staffer who works on the Group IV classifications claims that they are made in a totally independent manner, without outside interference.
Because the OTS is the sole keeper of the data, we would have no way to rebut their criticism of us, other than by standing by what we believe to be valid calculations using data that are available.
Mr. Thomas is president of K.H. Thomas Associates, a Miami-based consulting firm, and a lecturer in finance at the Wharton School of the University of Pennsylvania.
