Recent Items

Fitch Repeats Error that Led to Bad Subprime Calls

Posted on by

Saying bad things about rating agencies may seem like beating a dead horse, except this horse is still very much alive. Even though by some measures Fitch has comported itself better than Moody’s and S&P (it downgraded Ambac to AA and was early to point out excesses in commercial real estate lending), it has shown itself to be very much like its peers in other ways.

This little piece today from Reserach Recap was a reminder of why ratings are held in less than high esteem:

The U.S. unemployment rate would need to quadruple to 20% to have a major negative ratings impact on AAA-rated securities backed by credit card and auto debt, according to Fitch Research.

In a new report, Fitch analyzes the historical relationship between changes in the unemployment rate and changes in U.S. consumer ABS loss rates. Fitch’s forecast is for unemployment to increase steadily to the 5.5% range by year end.

According to analysis, increases in the unemployment rate are expected to cause auto loan and credit card loss rates to increase proportionally with subprime assets experiencing the highest proportional rate. For example, prime credit card chargeoffs should increase on a 1:1 basis. Accordingly, a 100% increase in the current unemployment rate, from 4.8% to 9.6%, would lead up to a 100% increase in current prime credit card chargeoffs, from 5.6% to 11.2%. For subprime credit cards and auto loans the proportional increase is closer to 1.3:1 meaning a 100% increase in unemployment would lead up to a 130% increase in chargeoffs or losses.

Based on this analysis, unemployment in isolation would have to increase, from 4.8% to more than 20% in order to cause a first dollar loss to typical ‘AAA’ rated credit card and auto transactions. Typical ‘BBB’ securities could withstand an increase in the unemployment rate to more than 9% for auto transactions and close to 11% for credit card transactions.

Consumer ABS transactions rated ‘AAA’ can withstand unemployment stresses to levels not seen since the Great Depression.

Fitch stated in February that it expects collateral performance to decline steadily throughout 2008 in the credit card receivable and auto loan assets. However, Fitch believes that the credit enhancement and structural features of the transactions are sufficient to stave off widespread negative rating actions in those sectors.

Now Fitch may turn out to be right (in part because unemployment levels will hopefully not even get close to the 20% level to test their thesis), but this so-called analysis is pretty dubious. I’ll admit I did not read the entire report, but it is only 656 words, so there imuch to it.

Let’s look at the methodology: it looks at the “historical relationship” between unemployment and ABS losses, and then claims that loss rates on auto loans, credit cards, and subprime should all move proportionally. That is, if subprime loss rates double, credit card loss rates will double (remember credit card loss rates are lower than subprime loss rates).

Now that may turn out to be a good assumption, but to call it anything more than that is intellectually dishonest. Why? Remember, recent subprimes bear no resemblance to historical subprimes; that’s how the industry botched up so badly in the first place. Subprime pre 2000 were for the purchase of largely manufactures housing; as probably every reader knows, the more recent subprimes got dodgier and dodgies, with a sharp deterioration in issuance standards starting in 2005.

In addition, the old subprimes have only endured one business downturn (the tech wreck slowdown) and this one is almost certain to be more severe.

Similarly, the credit card industry has been generous in extending credit to weak borrowers, particularly with the use of low interest rates and balance transfers. Thus, the level of unsecured credit (which can be discharged in a Chapter 7 bankruptcy) could now be proportionally greater than in past cycles.

Again, the conclusion may be correct, but naively projecting out trend lines is exactly what created the housing market mess. The rating agencies seem either incapable of or unwilling to learn from their mistakes.

Print Friendly
Twitter0DiggReddit0StumbleUpon0Facebook12LinkedIn0Google+0bufferEmail

5 comments

  1. Anonymous

    re: the analysis of these geniuses–on the farm where I was raised, we’d call it used hay

  2. Anonymous

    I think a bigger problem is that they are comparing unemployment rates of years ago to the current farce of stats. As an example the last unemployment report the unemployment rate dropped while jobs deceased by 63,000. They are quick to label people not looking for work compared to prior years.

  3. Anonymous

    I like how a complex financial organization uses a linear relationship to define the relationship between default and unemployment. Good thinking guys. Spreadsheets anyone?

  4. Anonymous

    Fitch is doing what it is suppposed to do (rememeber that Fitch, S%P and Moody are called Nationally Recognized Statistical Rating Org.). What do you expect them to say: for the past 50 years it has been this way, but now for some reason isn’t, and it will be this other way and believe us because we say so? That is not simply their job. The last part is
    the investor’s job.

  5. Anonymous

    Revolving credit card debt has also gone through adverse selection in the last cycle. Those without HELOC access dominate today’s credit card revolvers. While this group does not have to worry about falling home prices, they will be hit hard in a recession.

Comments are closed.