"Migration Analysis" of Banks, Part 1
Thu May 20, 2010If you are in the banking space, you are familiar with one of Capitalytics' features: migration analysis. Migration Analysis entails building a model that shows the likelihood of a borrower letting a current loan lapse, or a late loan defaulting. In mathematical terms, this analysis involves determining the conditional probabilities of a loan's state transitioning from one state of delinquency to another.
The following analysis, though, is based on the state of bank's capitalization. In other words, we started looking at how banks migrate from one capital category to another. Banks are classified into one of five bank categories: "well capitalized", "adequately capitalized", "under-capitalized", "significantly under-capitalized", and "critically under-capitalized". The definitions of these categories are provided here by the FDIC. (For the purposes of this article, we used the 3% threshold for the leverage ratio condition in determining whether a bank is "adequately capitalized" or "under-capitalized". We also did not account for any confidential written orders written by the FDIC and imposed on a bank.)
In this study, we looked at not only the evolution of banks across the county, but also the evolution of banks across each of Alabama (which might be seen as a state that has weathered the latest banking issues fairly well) and Georgia (which might be seen as a one of the states that has been affected the most by the recent banking issues). We collected at the 40 quarters of data across the 2000's and looked at how every bank's capital category evolved from quarter-to-quarter. For instance, in the state of Alabama, during the 40 quarters from 1Q2000 to 4Q2009, there were 52 instances during which a bank's capital category went from "well capitalized" to "adequately capitalized", and 82 instances during which a bank ceased doing business after being found to be "well capitalized" (due to, most likely, dissolution, merger, or acquisition).
The following tables show the transition probabilities for the US, Alabama, and Georgia, respectively. The rows in these matrices show the "previous quarter's" capital category, as compared to the columns (that show the "following quarter's" capital category).
US
| Well Capitalized | Adequately Capitalized | Under-Capitalized | Significantly Under-Capitalized | Critically Under-Capitalized | Not Found | |
|---|---|---|---|---|---|---|
| Well Capitalized | 99.2546% | 0.2643% | 0.0006% | 0.0000% | 0.0213% | 0.4592% |
| Adequately Capitalized | 46.1736% | 50.2746% | 0.0366% | 0.0000% | 1.8308% | 1.6844% |
| Under-Capitalized | 100.0000% | 0.0000% | 0.0000% | 0.0000% | 0.0000% | 0.0000% |
| Significantly Under-Capitalized | 33.3333% | 0.0000% | 0.0000% | 0.0000% | 33.3333% | 33.3333% |
| Critically Under-Capitalized | 14.2907% | 18.5850% | 0.0000% | 0.0000% | 60.1197% | 7.0046% |
AL (no banks ever found to be under-capitalized or significantly under-capitalized)
| Well Capitalized | Adequately Capitalized | Under-Capitalized | Significantly Under-Capitalized | Critically Under-Capitalized | Not Found | |
|---|---|---|---|---|---|---|
| Well Capitalized | 99.4664% | 0.2071% | 0.0000% | 0.0000% | 0.0000% | 0.3265% |
| Adequately Capitalized | 57.6271% | 33.8983% | 0.0000% | 0.0000% | 3.3898% | 5.0847% |
| Under-Capitalized | N/A | N/A | N/A | N/A | N/A | N/A |
| Significantly Under-Capitalized | N/A | N/A | N/A | N/A | N/A | N/A |
| Critically Under-Capitalized | 50.0000% | 12.5000% | 0.0000% | 0.0000% | 25.0000% | 12.5000% |
GA (no banks ever found to be significantly under-capitalized)
| Well Capitalized | Adequately Capitalized | Under-Capitalized | Significantly Under-Capitalized | Critically Under-Capitalized | Not Found | |
|---|---|---|---|---|---|---|
| Well Capitalized | 98.9516% | 0.4498% | 0.0076% | 0.0000% | 0.0305% | 0.5604% |
| Adequately Capitalized | 53.2495% | 42.3480% | 0.4193% | 0.0000% | 2.0964% | 1.8868% |
| Under-Capitalized | 100.0000% | 0.0000% | 0.0000% | 0.0000% | 0.0000% | 0.0000% |
| Significantly Under-Capitalized | N/A | N/A | N/A | N/A | N/A | N/A |
| Critically Under-Capitalized | 15.7205% | 36.6812% | 0.0000% | 0.0000% | 34.9345% | 12.6638% |
Before going any further, we must note that examining this system as a Markov chain entails certain assumptions, which are not even close to being met in this "real world" scenario. First, while randomness may appear be observed from the outside of the system, we can only hope that the management of the 10,000+ banks in the US are not actually making operational decisions such that their banks move through this system with truly random transitions. Second, changes in the economic climate will surely have an impact on the factors which govern these banks. (We will examine one time-varying aspect of this system in the following post.) Third, but not finally, we should keep in mind that the goal of this study will be to evaluate intuitively (not in any academically rigorous fashion) if Alabama's and Georgia's banks, respectively, are obviously traversing this system in a different manner than an arbitrary bank in the US, and, if there has been any change in these states' banks' behavior since the current economic crisis began.
What conclusions can be reached from the above data? Obviously, there are not enough instances of banks being classified as "Under-Capitalized" or "Significantly Under-Capitalized" in either Alabama or Georgia to be able to draw any conclusions. However, at least looking at the data superficially, a few things become obvious: first, Alabama's banks seem to avoid the "Critically Under-Capitalized" situation (which requires government intervention) much better than the general norm; gleaning that the threshold for this classification is driven based on the ratio of equity to asset values implies that Alabama's banks likely have tended to grow in a more controlled fashion than their counterparts.
Beyond this point, we can see that banks in both Alabama and Georgia tend to recover from being classified as "Adequately Capitalized" faster than banks across the country (based on the lower transition probability from the "Adequately Capitalized" state back to itself). Admittedly, in Alabama, it is more significantly more likely for an "Adequately Capitalized" bank slips into a "Critically Under-Capitalized" state than average (twice as likely, in fact), but it is also much more likely for this same "Adequately Capitalized" bank to recover by adjusting its operations in order to boost its key ratios. This last point (regarding the potential for a bank to recover) also applies to Georgia banks.
What else can be said about banks in these two states versus the norm? The main other point that stands out is the apparent disposition of Georgia banks once they reach a state of being a "Critically Under-Capitalized" bank; these banks appear to reach their final disposition (either recovery or, more likely, otherwise) more quickly than in other states, which is more a credit to the efforts of the regulatory agencies that operate in that state. The next questions to be addressed, however, is how much of these numbers have been tainted by the issues that have affected bank in the US, and both state, over the past few years. Our next post will address this question.