Making Unique Observations in a Very Cluttered World

Monday, 3 August 2009

1 Month, 24 Bank Failures: Random Event or Wilting Economy?

Reading - 1 Month, 24 Bank Failures: Random Event or Wilting Economy? http://tinyurl.com/xiam007-9

In what has become as much a staple of Friday nights as drunkenness itself, the FDIC on Friday conducted "seizures" of the following five banks:

  1. Mutual Bank (Harvey, IL)
  2. First BankAmericano (Elizabeth, NJ)
  3. Peoples Community Bank (West Chester, OH)
  4. Integrity Bank (Jupiter, FL)
  5. First State Bank of Altus (Altus, OK)

Despite the fact that JP Morgan (JPM) and Goldman Sachs (GS) continue to earn (?) healthy profits, the plight of the nation's small community banks appears substantially more imperiled.

The familiar Fridays that used to feature a single bank failure are now characterized by a new tradition which dictates that at least four - and possibly five banks must fail each Friday. The rate and magnitude of recent bank failures does not correspond well with the popular belief of the day; being that we have "turned a corner", "stopped the tailspin", "stabilized" or any other number of foolish statements currently in vogue.

To take a look at some data that slaps the recovery hypothesis right in its face, let's take a look at the past thirteen months' bank failures, by the number that have occurred each month (source: FDIC: Failed Bank List):

Month / Number of Bank Failures

  • July 2008 - 3
  • August 2008 - 3
  • September 2008 - 4
  • October 2008 - 4
  • November 2008 - 5
  • December 2008 - 3
  • January 2009 - 6
  • February 2009 - 10
  • March 2009 - 5
  • April 2009 - 8
  • May 2009 - 7
  • June 2009 - 9
  • July 2009 - 24

It's immediately apparent that July's number of failures represents a severe departure from normal.

The question is, did we just experience a run of the mill anomaly that occurred independently of economic conditions, or do the numbers indicate something more troublesome? To say that 24 bank failures in one month alone is NOT an indication of economic deterioration, one would have to suppose that there is an element of randomness to the number of failures that occur each month.

If that is so, we should be able to apply some elementary statistics to the situation to discern just how probable it is that July would randomly produce 24 bank failings.

Running the numbers on the 12 months ended June 2009, we calculate that the mean number of bank failures for a single month is 5.58, and the population standard deviation is 2.33. For the uninitiated, this means that - assuming the Failures are distributed normally - the number of financial institutions collapsing in a single month will be between 3.2 and 7.9, approximately 68% of the time.

Taking the exercise a step further, we can state that 97.5% of the time, a given month will produce 10.23 or fewer Friday night FDIC raids. To calculate the probability of the 13th month in the series (July 2009) bearing witness to a full 24 FDIC seizures, one must first arm himself with spreadsheet software capable of displaying results to the Quadrillionth (Google's spreadsheet program is just barely able to do so).

The probability-value associated with such an event is 1.18755X10^-15, or 0.0000000000000018755.

Based upon the monthly rate of bank closings during the most recent 13 periods, it's apparent that the random forces of nature are, in all likelihood, not responsible for the FDIC's hectic July schedule. We must then conclude that the responsible party is none other than a deteriorating economy.

To be more specific, the recent spike in bank failures is a reflection of the rapidly deteriorating economy on Main Street a.k.a those who did not receive a bailout and continue to either fear or experience job losses; consequently, straining the ability of these individuals to honor the obligations made to many smaller, community oriented banks.