3. Failed bank resolutions and lending activity
To examine the impact of bank failures on lending, it is useful to examine the resolution techniques employed by the FDIC and the location of the failures. Panel A of table I provides some descriptive evidence on resolutions over the 1987-1991 period, while Panel B lists the 10 states with the greatest number of failures. During the late 1980s and early 1990s, the FDIC resolved the majority of bank failures through purchase and assumption (P&A) agreements. As shown in Panel A, P&As accounted for 73%, or 675 of the 927 total resolutions. Panel B shows that, although failures occurred across the nation, they were concentrated in the Southwest. For example, Texas and Louisiana had the greatest number of failures, experiencing 446 and 58 failures, respectively.
The use of the P&A approach is important when examining the impact of bank failures on lending activity, because the P&A approach involves a transfer of assets from a failed bank to a healthy institution, while other approaches, such as an insured deposit transfer or a depositor payout, do not. Under a typical P&A, a healthy bank purchases some assets and liabilities of the failed bank, with the remaining assets left under FDIC receivership.10 The FDIC typically sells the assets acquired by the healthy institution at a discount from book value, with the discount refecting the losses incurred by the failed institution on those assets. Given that the value of the acquired assets typically is less than the value of the assumed liabilities, the FDIC transfers cash to the acquiring bank in an amount equal to the difference between the two. In addition, the FDIC often includes a put option that enables the acquiring bank to return the asset to the FDIC at a later date (up to three years).11
Resolution of bank failures through P&As is widely believed to be the least disruptive to bank borrowers and depositors, since the P&A approach assumes that the failed bank customers will maintain their banking business with the acquiring institution.12 However, only anecdotal evidence exists about whether, in fact, these business relationships are maintained after resolution. This paper seeks to address this gap in the literature by investigating how the acquiring bank's business lending is affected by the acquisition.13
A failed bank acquisition may impede lending activity through two possible channels. First, the acquisition of failed-bank assets may impair the lending behavior of the acquiring institution through its impact on bank capital, a process Baer and McElravey (1993) term the recycled asset hypothesis. According to Baer and McElravey (1993), to acquire failed bank assets, the acquiring bank must have suffcient capital to assume the assets without becoming undercapitalized. However, the acquisition of these assets raises the acquiring bank's total assets, thereby increasing its regulatory capital requirement. If issuing new equity is costly, then the acquiring bank will not immediately replace the capital absorbed by the acquired assets. Instead, it may slow or reduce lending, since the acquiring bank has less capital available to support new lending. Jacques and Nigro (1995) provide descriptive support for the recycled asset hypothesis and surmise that business lending declines because C&I loans often are short-term and relatively easy to terminate.
Second, a loss of information about borrowers occurs as part of the overall financial turmoil associated with the acquisition of assets, which also may affect lending. For example, since business lending often is relationship based, certain borrower-specifc information will not be easily transferable. Diamond (1984), in fact, argues that financial intermediaries exist due to their ability to reduce information costs. The literature on