Understanding the Components of Bank Failure Resolution Costs

• Date: 01 December 2015
• Published date: 01 December 2015
• Author: Rosalind L. Bennett,Haluk Unal
• DOI: http://doi.org/10.1111/fmii.12031

Understanding the Components of Bank Failure

Resolution Costs

BYROSALIND L. BENNETT AND HALUK UNAL

In this paper, we demonstrate how the resolution costs associated with over 1,000 bank

failures from 1986 to 2007 are distributed across the method of resolution, bank size,

regulatory periods, and the existence of fraud. In addition, we document the time spent

in the resolution by the resolution method and legislative period. Finally, we show how

various classes of claimants against the failed banks bear the costs of the failure.

Keywords: bank failures, bank resolution costs, FDIC receivership, fire sales, banking

crises.

JEL Classification: G21, G28, G33.

I. INTRODUCTION

In the United States 2,427 depository institutions failed from 1986 to 2007. Out of

these 1,244 of them with $216.4 billion in assets were placed into a Federal Deposit

Insurance Corporation (FDIC) receivership for resolution.1As of year-end 2013,

the FDIC estimates that the total cost to the deposit insurance funds of resolving

these BIF- and DIF-insured failed banks is around $30 billion. We provide an

analysis of this extended and costly experience using classes of the population

of bank failures that were resolved and terminated by the FDIC during the 1986

to 2007 period. This analysis is important because understanding the different

dimensions of these costs is a critical input for effective resolution planning.

The standard definition of resolution cost is the difference between the lia-

bilities of the failed bank and the market value of its assets net of expenses. We

decompose the resolution cost of bank failures into three major categories—losses

incurred on the disposition of the assets of the failed bank, direct expenses, and

indirect expenses incurred by the FDIC to resolve these failures. We demonstrate

how resolution costs are distributed across the method of resolution, bank size,

regulatory periods, and the existence of fraud. In addition, we document the time

spent in the resolution by the resolution method and legislative period. Finally,we

show how various classes of claimants against the failed banks bear the costs of

the failure.

Our sample consists of 1,213 of the 1,244 banks that failed during the 1986 to

2007 period. We consolidate the individual bank failures under their respective

Corresponding author: Haluk Unal, Smith School of Business, University of Maryland. E-mail: hunal@

rhsmith.umd.edu.

1The remainder of the failures is thrift institutions, which were insured by Federal Savings and Loan

Insurance Corporation (FSLIC) or Saving Association Insurance Fund (SAIF) or resolved by the

Resolution Trust Corporation (RTC).We focus on Bank Insurance Fund (BIF) and Deposit Insurance

Fund (DIF)-insured institutions. Throughout the paper we refer to them as banks.

C2015 New YorkUniversity Salomon Center and Wiley Periodicals, Inc.

350 Rosalind L. Bennett and Haluk Unal

holding company name. This consolidated sample has 1,092 failures. Wefind that

the banks have an average ratio of the book value of equity to assets on the last

Call Report of negative 1.36% in the consolidated sample. The averagediscounted

loss on the disposition of assets as a percent of total assets is 23.38%. The mean

ratio of discounted receivership expense to assets is 12.02%, of which 3.49% is the

average ratio of discounted direct receivership expenses to assets. As a result, the

average discounted total resolution cost to asset ratio is 33.18%. In other words,

an average failed bank during 1986-2007 has a negative market value of equity

that is about one-third of the book value of its assets.

There is considerable time-variation in these ratios. For example, the average

ratio of discounted total resolution costs to assets is 39.82% for the failed in-

stitutions prior to the enactment of the Financial Institutions Reform, Recovery

and Enforcement Act (FIRREA) in 1989. In contrast, this ratio significantly de-

clines to 22.92% during the period that follows the passage of the Federal Deposit

Insurance Corporation Improvement Act (FDICIA) of 1991.

The FDIC uses a number of methods to resolve failed banks including deposit

payoffs, insured-deposit transfers, purchase and assumption (P&A) agreements,

whole-bank transactions, and open-bank assistance. The primary difference

between the methods is whether the FDIC assumes and liquidates the failed-bank

assets (deposit payoffs) or leaves most or all of the failed-bank assets in the private

sector (P&A agreements, whole bank transactions, and open-bank assistance).

Univariate tests show that the average ratio of discounted resolution costs to

assets for deposit payoffs, 36.74%, is statistically different than that for P&A

methods, 34.63%.

Our results show that large banks that fail have higher capital ratios at failure,

lower loss on assets, and lower receivership expenses than small banks. We find

no univariate evidence that failures caused by fraud have higher total resolution

costs. Finally, the average time of resolution is about five years, which is roughly

twice as long as a typical non-financial bankruptcy.

The remainder of this paper is organized as follows. Section 2 describes the

number and types of failures over the 1986 to 2007 period. Section 3 provides our

definition of resolution costs and Section 4 describes the components of resolution

costs. Section 5 describes the patterns of resolution costs across different resolu-

tion methods. Section 6 looks at resolution costs across different size categories.

Section 7 describes the patterns of resolution costs across different legislative

periods. Section 8 describes some measures of fraud for failed banks and shows

the pattern of resolution costs across these different measures of fraud. Section 9

evaluates how the time in receivership varies over different resolution types and

legislative periods. Finally,Section 10 shows the link between the total resolution

costs and losses to the FDIC and other claimants.

II. BANK FAILURES AND RESOLUTION TYPES

Banks can fail for a variety of reasons including undercapitalization, liquid-

ity, safety and soundness, and fraud. The chartering agency has the authority to

Understanding the Components of Bank Failure Resolution Costs 351

terminate the bank’s charter and appoint the FDIC to resolve the failure. The

FDIC establishes an independent legal entity called a “receivership” that over-

sees the orderly resolution of the failed bank. The following chartering agencies

have the authority to essentially close a bank: the Office of the Comptroller of

the Currency (OCC), Office of Thrift Supervision (OTS), and the state banking

authorities. For insured federal savings associations and national banks, the FDIC

must be appointed receiver. In the case of state-chartered banks that are members

of the Federal Reserve System, the state banking authority may appoint the FDIC

receiver. In 1991, Congress gave FDIC the power to appoint itself as receiver for

state chartered insured depository institutions.2

As part of the resolution process, the FDIC develops a marketing strategy that

includes determining the resolution structures that it offers to potential bidders

(Bovenzi and Muldoon (1990) and FDIC (1998b)). The FDIC then markets the

assets and liabilities of the failing bank and evaluates the bids it receives. One

option that the FDIC is required to consider is a deposit payoff, where the FDIC

pays the insured depositors and liquidates the assets. The FDIC can employ two

methods to pay off depositors. In a deposit payout, the FDIC pays off the insured

depositors in cash (by check). The uninsured depositors and general creditors file

claims against the receivership and they are paid their pro-rata share of their claim

if funds are available as the assets are liquidated. In this case, any deposit franchise

of the failed bank is destroyed. The other method is an insured-deposit transfer

where the FDIC transfers insured deposits and secured liabilities to a healthy

institution, along with a cash payment.3This cash payment is typically less than

the face value of the liabilities because the FDIC usually receives a premium from

the agent institution, and thus recovers some of the value of the deposit franchise.

In either method the FDIC does not cover uninsured deposits which are reimbursed

their pro-rata share as the assets are liquidated.

Alternatively, the FDIC can receive bids to purchase all or part of the assets

and assume all or part of the deposit liabilities. The FDIC compares the cost of

these bids to the cost if the FDIC liquidates the assets. Prior to the passage of

the FDICIA, which required the FDIC to close institutions in a manner that is

least costly to the deposit insurance fund, a bid had to pass the cost test to be

acceptable.4The cost test required that the final resolution be less costly than a

deposit payoff, however it did not require that the accepted bid be the least costly

of all of the bids. For example, suppose the FDIC’s cost estimate for a deposit

payoff and liquidation of the assets is $450 and two bids exist, where the cost of

Bank A’ s bid is $300 and cost of Bank B’s bid is $400. The FDIC could choose

2The FDIC has elected to do so on one occasion—in the failure of Meriden Trust & Safe Deposit

Company in Meriden Connecticut on July 7, 1994.

3The healthy bank can also be viewed as a “paying agent” for the FDIC. Depositors of the failed

firm have access to their insured deposits and can choose to whether to move their account to a new

depository or accept the new relationship with the acquiring bank.

4The cost test was established in the Depository Institutions Deregulation and Monetary Control Act,

1980.