Accounting Discretion of Banks During a Financial Crisis

WP/09/207




Accounting Discretion of Banks During a
Financial Crisis

Harry Huizinga and Luc Laeven

---

© 2009 International Monetary Fund
WP/09/207


IMF Working Paper



Research Department

Accounting discretion of banks during a financial crisis

Prepared by Harry Huizinga and Luc Laeven1

Authorized for distribution by Stijn Claessens

September 2009

Abstract
This Working Paper should not be reported as representing the views of the IMF.
The views expressed in this Working Paper are those of the author(s) and do not necessarily represent
those of the IMF or IMF policy. Working Papers describe research in progress by the author(s) and are
published to elicit comments and to further debate.

This paper shows that banks use accounting discretion to overstate the value of distressed assets.
Banks’ balance sheets overvalue real estate-related assets compared to the market value of these
assets, especially during the U.S. mortgage crisis. Share prices of banks with large exposure to
mortgage-backed securities also react favorably to recent changes in accounting rules that relax
fair-value accounting, and these banks provision less for bad loans. Furthermore, distressed
banks use discretion in the classification of mortgage-backed securities to inflate their books.
Our results indicate that banks’ balance sheets offer a distorted view of the financial health of
the banks.

JEL Classification Numbers: A10; A11


Keywords: Key bank regulation, accounting standards, fair value accounting, corporate
disclosure, financial crisis


Author’s E-Mail Address: H.P.Huizinga@uvt.nl; LLaeven@imf.org

1 Huizinga is Professor of Economics at Center, Tilburg University, and Research Fellow at CEPR; Laeven is
Senior Economist at the International Monetary Fund, and Research Fellow at CEPR. We would like to thank
Rocco Huang, Christian Leuz, Joe Mason, Lev Ratnovski, and Wolf Wagner for comments or suggestions, and
Mattia Landoni for excellent research assistance. The findings, interpretations, and conclusions expressed in this
paper are entirely those of the authors. They should not be attributed to the IMF. Contact information: Harry
Huizinga: H.P.Huizinga@uvt.nl; Luc Laeven: llaeven@imf.org.

---

2
Contents Page
I. Introduction ............................................................................................................................3
II. Tobin’s q Value and Market Discounts.................................................................................7
III. The Data...............................................................................................................................8
IV. Market Discounts and Valuation Effects of Real Estate Related Assets...........................12
A. Empirical Evidence on Market Discounts ..............................................................12
B. Banks’ Stock Price Reaction to Amendments of Fair Value Accounting Rules ....16
V. Accounting Discretion on Impaired Assets and Asset Classification.................................18
A. Accounting Discretion on Accounting for Bad Loans............................................18
B. Classification of Mortgage-Backed Securities........................................................20
VI. Conclusions........................................................................................................................21

References................................................................................................................................24

Appendix
Variable Definitions and Data Sources....................................................................................29

Tables
1. Summary Statistics for 2008, Quarterly Data...................................................................... 30
2. Tobin’s q and Real Estate Related Assets in 2008...............................................................31
3. Tobin’s q and Real Estate Related Assets in 2001-2007......................................................32
4. Tobin's q Real Estate Related Assets and Asset Size...........................................................33
5. Tobin’s q and Additional Balance Sheet and Off-balance Sheet Items ...............................34
6. Event Study of New FASB Rules on Fair Value Accounting for
Illiquid Assets (FAS 157), Announced on October 10, 2008............................................35
7. Event Study of FASB Amendments to Fair Value Accounting of
Hard-to-Value Assets, Announced on April 9, 2009........................................................36
8. Loan Loss Provisions and Net Loan Charge-offs in 2008....................................................37
9. Share of Mortgage-backed Securities that is Held-to-Maturity in 2008...............................38
10. Share of Non-Guaranteed Mortgage-backed Securities that is Held-to-Maturity
in 2001-2007....................................................................................................................39

Figures
1. Tobin’s q and Share of Zombie Banks.................................................................................40
2. Real Estate Loans and Mortgage-backed Securities.............................................................40
3. Share of Mortgage-backed Securities that is Held-to-Maturity............................................40
4. Fair Value of Mortgage-backed Securities Relative to Amortized Cost..............................41
5. Tier 1 Capital Ratio and Share of Tier 1 Capital in Total Capital........................................41

---

3
I. INTRODUCTION
The current financial crisis has reinvigorated a debate on the effectiveness of the existing
accounting and regulatory frameworks for banks. Questions abound, ranging from adequate
capitalization levels of banks to the boundaries of financial regulation (see Financial Stability
Forum, 2008). Part of the debate on financial reform centers around required information on
banks for effective market discipline and supervisory action. This includes not only thinking on
the required level of detail on disclosure of bank assets and liabilities but also on their valuation
techniques and the appropriateness of current accounting rules more generally (see Laux and
Leuz, 2009, for a survey).

Part of this debate centers around the pros and cons of fair value accounting, where fair
value is meant to indicate the price at which an asset could be bought or sold in a current
transaction between willing parties, other than in a liquidation. Accounting standards stipulate
that as a guiding principle, the quoted market price in an active market should be used as the
basis for the measurement of the fair value of an asset. The problem is that such a price is not
always available, for example, in illiquid markets. In such cases, fair values need to be estimated
based on available information. A related concern is the potential procyclical nature of fair value
accounting, which could magnify fluctuations in bank lending and economic activity (see IMF,
2009, and Heaton et al., 2009). A broader concern is that the current “mixed attribute” model of
accounting, in which some financial instruments are measured based on historical cost and some
at fair value, together with discretion over how financial instruments are measured, gives rise to
accounting arbitrage.2

Despite difficulties of determining fair values in illiquid markets, advocates of fair value
accounting maintain that fair value is the most relevant measure for financial instruments.3 They
argue that financial assets, even complex instruments, tend to trade continuously in markets and
it should therefore be possible to use information embedded in market prices to compute fair
values of financial assets.

Faced with massive write-downs and expected losses, banks in contrast have used the
momentum to lobby against the use of fair value accounting. They claim that most of their assets
are currently not impaired, that they intend to hold them to maturity anyway, and that market
prices reflect distressed sales into an illiquid market. Potential buyers of such assets, however,
are unlikely to value them at origination value but at prices well below book value. Banks may
ignore such signals to avoid recognizing a loss, claiming that unusual market conditions, not an
actual decline in value, cause low market pricing.


2 As emphasized by Jackson M. Day, Deputy Chief Accountant, U.S. Securities and Exchange Commission, in his
year 2000 remarks “Fair Value Accounting–Let's Work Together and Get It Done!” at the 28th Annual National
Conference on Current SEC Developments.
3 See, for example, Kaplan, Robert, Robert Merton and Scott Richard, 2009, “Disclose the fair value of complex
securities”, Financial Times, August 17, 2009.

---

4
Accounting techniques do not generally generate large differences between the book and
market value of bank assets. At times of financial crisis when asset markets become distressed,
however, large differences between book and market values of assets may arise, especially when
assets are carried at values based on historical cost. Such differences may give rise to incentives
for banks to use accounting discretion to preserve the book value of the bank, for example, by
using advantageous asset classifications or valuation techniques. As a consequence, discretion in
accounting rules enables banks to understate underlying balance sheet stresses. Overstated book
values of bank assets may further give rise to undue regulatory forbearance.4

During the ongoing financial crisis, large differences have indeed arisen between market
and book values of U.S. banks as their market values have sharply eroded on the expectation of
major writedowns and losses on real estate related assets. By end-2008, more than 60% of U.S.
bank holding companies had a market-to-book value of assets of less than one, compared to only
8% of banks at end-2001. At the same time, the average ratio of Tier 1 capital to bank assets has
stayed constant at about 11% throughout this period. The market value of bank equity thus has
dropped precipitously against a backdrop of virtually constant book capital. This raises doubts
about the relevance and reliability of banks’ accounting information, the two main criteria on the
basis of which accounting systems are evaluated, at a time of financial crisis.

This paper shows that banks use accounting discretion to systematically understate the
impairment of their real estate related assets, especially following the onset of the current
financial crisis, in an effort to preserve book capital. We provide the first evaluation of such
behavior and offer three pieces of compelling evidence to support our thesis that banks use
accounting discretion to overstate the book value of capital.

First, we estimate large market discounts on real estate related assets, including mortgage
loans and mortgage-backed securities (MBS). To estimate implicit market discounts on bank
assets, we empirically relate Tobin’s q, computed as the market-to-book value of assets, to
banks’ asset exposures using quarterly accounting data on U.S. bank holding companies for the
period 2001 to 2008. Our primary focus is on real estate related assets, as these assets constitute
a large fraction of the total assets of the average bank, and as recent declines in U.S. real estate
prices have raised doubts about the underlying value of these assets. However, we also apply our
methodology to other on- and off-balance sheet items. We estimate significant discounts on
banks’ real estate loans (relative to other loans) starting in 2005, averaging about 10% in 2008.
As the average bank holding company in 2008 holds about 54% of its assets in the form of real
estate loans, the implicit discount in loan values goes a long way toward explaining the current
depressed state of bank share prices. We further find that investors started discounting banks’
holdings of MBS in 2008. For that year, we find an average discount on these assets of 24%
(relative to other securities), while the average MBS exposure amounted to 10% of assets. The
market discount on MBS that are available-for-sale (and carried at fair value) is estimated to be
23%, against a discount of 32% for MBS that are held-to-maturity (and carried at values based

4 For evidence of regulatory forbearance and the political economy of bank intervention, see Kane (1989), Kroszner
and Strahan (1996), Barth et al. (2006), and Brown and Dinc (2005, 2009).

---

5
on historical cost). Thus, even MBS that are carried at fair value appear to be overvalued on the
balance sheets of banks.

Second, using an event study methodology we find that banks with large exposure to
MBS experienced relatively large excess returns when rules regarding fair value accounting were
relaxed. Pressures arose during the summer of 2008 to provide banks with more leniency to
determine the fair value of illiquid assets such as thinly traded MBS to prevent these fair values
from reflecting ‘fire sale’ prices.5 Correspondingly, on October 10, 2008 the Financial
Accounting Standards Board (FASB) clarified the allowable use of non-market information for
determining the fair value of financial assets when the market for that asset is not active.
Subsequently, on April 9, 2009, the FASB announced a related decision to provide banks greater
discretion in the use of non-market information in determining the fair value of hard-to-value
assets. As expected, the stock market on both occasions cheered the banks’ enhanced ability to
maintain accounting solvency in an environment of low transaction prices for MBS. Using an
event study methodology, we find that banks with large exposure to MBS experienced relatively
large excess returns around both announcement dates, indicating that these banks in particular
are expected to benefit from the expanded accounting discretion.

Third, we show that banks use accounting discretion regarding the realization of loan
losses and the classification of assets to preserve book capital. Banks have considerable
discretion in the timing of their loan loss provisioning for bad loans and in the realization of loan
losses in the form of charge-offs. Thus, banks with large exposure to MBS and related losses can
attempt to compensate by reducing the provisioning for bad debt in an effort to preserve book
capital. We indeed find that banks with large portfolios of MBS report relatively low rates of
loan loss provisioning and loan charge-offs.

We also examine banks’ choices regarding the classification of MBS as either held-to-
maturity or available-for-sale. We consider this categorization separately for MBS that are
covered or issued by a government agency. In 2008, the fair value of especially non-guaranteed
MBS tended to be less than their amortized cost. This implies that banks could augment the book
value of assets by classifying MBS as held-to-maturity. Indeed, we show that the share of non-
guaranteed MBS that are held-to-maturity increased substantially in 2008. Classification of this
kind is particularly advantageous for banks whose share price is depressed on account of large
real estate related exposures. Consistent with this, we find that the share of MBS kept as held-to-
maturity is significantly related to both real estate loan and MBS exposures. Moreover, these
relationships are stronger for low-valuation banks.

Taken together, the evidence of this paper shows that banks use considerable accounting
discretion regarding the categorization of assets, valuation techniques, and the treatment of loan
losses. Accounting discretion appears to be used to soften the impact of the crisis on the book

5 The primary concern was one of maintaining solvency at affected banks. There was also a concern that losses
induced by fire sales could spread to other financial institutions. Allen and Carletti (2008) and Plantin et al. (2008)
offer theoretical models investigating potential contagion effects among banks if fair value accounting forces banks
to value their securities according to observed ‘fire sale’ prices.

---

6
valuation of assets. While some accounting discretion is unavoidable as accounting systems in
part are mechanisms for firms to reveal asymmetric information to investors and other outside
parties6, accounting discretion entails the risk of generating highly inaccurate accounting
information at a time of great turmoil, such as the present financial crisis. Inaccurate accounting
information in the case of banks can be especially harmful, as it may lead to regulatory
forbearance with concomitant risks for tax payers. In the present crisis, the financial statements
of banks appear to overstate the book value of assets to the point of becoming misleading guides
to investors and regulators alike.7 Thus, the present crisis can be seen as a ‘stress test’ of the
accounting framework that reveals that book valuation need not always reflect the best estimate
of asset value, especially at a time of sharp declines in market values. Accounting reforms
announced so far and discussed in this paper, however, seem to go in the direction of increasing
the gap between book and market values. This may be testimony that bank interests weigh
heavily in this debate.

Our paper relates to a large literature in accounting and finance on how accounting
principles and systems affect corporate behavior and that of banks in particular (see, e.g., Collins
et al., 1995, Shackelford et al., 2008, and Leuz and Wysocki, 2008). Much of this work analyzes
the cost and benefits of earnings management of firms (see, e.g., Leuz et al., 2003, and Hutton et
al., 2008). There is also work on the costs and benefits of enhanced corporate disclosure and
accounting transparency (see Leuz and Wysocki, 2008, for a review). For example, Karpoff et al.
(2008) using firm-level information on legal enforcement actions show that financial
misrepresentation has reputational consequences for firms and depresses firm valuation.

A related literature reviewed by Barth et al. (2001) and Holthausen and Watts (2001) asks
whether accounting information is value relevant in the sense that it conforms to the information
that bank shareholders use to price bank shares. Barth et al. (1996) and Eccher et al. (1996) find
that fair value estimates of loan portfolios and securities help to explain bank share prices
beyond amortized cost. There is also recent work on the market pricing of bank assets reported
under different fair valuation techniques (e.g., Kolev, 2009, Goh et al., 2009, and Song et al.,
2009). Bongini et al. (2002) further find that measures of bank fragility based on market
information are a better predictor of bank failures than measures of bank fragility based on
accounting information.

Our paper is part of an emerging literature on the causes and effects of the 2007 U.S.
financial crisis. This work shows that house price appreciation (e.g., Demyanyk and Van Hemert,
2008) and asset securitization (e.g., Keys et al., 2008; Mian and Sufi, 2008; Loutskina and
Strahan, 2009), combined with a more general deterioration of lending standards by banks (e.g.,

6 A theoretical literature outlines that managers of firms may have incentives to smooth reported accounting incomes
either to smooth their own compensation, to increase their job security, or to increase firm valuation by investors
(see, e.g., Trueman and Titman, 1988, Fudenberg and Tirole, 1995, and Sankar and Subramanyam, 2000).
7 The outcomes of stress tests of major U.S. banks conducted by the U.S. Treasury in 2009, which calculated capital
shortfalls at several major banks, are testimony to the fact that publicly available accounting information at the time
provided an inadequate picture of the health of the concerned banks.

---

7
Dell’Ariccia et al., 2008), helped fuel a crisis in U.S. mortgage markets, with bank capital being
eroded as the asset price bubble in real estate markets burst starting in 2007.

The paper continues as follows. Section II sets out the relationship between Tobin’s q and
market discounts on bank assets. Section 3 discusses the data. Section IV first presents empirical
evidence on market discounts of real estate related assets relative to book values. Subsequently,
it provides evidence on the stock market response to the announcements of more lenient rules for
accounting for illiquid assets. Section V examines the use of bank discretion regarding loan loss
provisioning, loan charge-offs, and the classification of MBS into different accounting
categories. Section VI concludes.

II. TOBIN’S Q VALUE AND MARKET DISCOUNTS

In this section, we describe how observations of Tobin’s q can be used to infer discounts
on bank assets implicit in the stock market.8 Let MV be the market value of the bank. At the
same time, let Ai be the accounting value of asset i and let Li be the accounting value of liability
i. Assuming there are operating markets for a bank’s assets and liabilities, we can state a bank’s
market value as follows:

a
l
MV ? ? v A ? ? v L





(1)
i
i
i
i
i
i
where a
v is the market value of asset i and l
v is the market value of liability i.9
i
i

We can define q as the market value of the equity of the bank plus the book value of all
liabilities divided by the book value of all assets as follows:

MV ? ? Li
q ?
i
?

Ai
i
Substituting for MV from (1) into the expression for q, we get:

8 In similar fashion, Sachs and Huizinga (1987) estimate discounts on third world debt on the books of U.S.
commercial banks at the time of the international debt crisis of the 1980s. A related literature, starting with Lang and
Stulz (1994) and including Laeven and Levine (2007), has studied discounts in Tobin’s q arising from corporate
diversification. In that literature, discounts are computed for each business unit of a conglomerate with respect to the
value of comparable stand-alone firms, while here we compute discounts for different assets and liabilities of the
same bank.
9 In eq. (1), we ignore that market value may depend on the co-existence of certain assets and liabilities as discussed
in, for instance, DeYoung and Yom (2008).

---

8
a
l
q ? 1? ? d a ? ? d l





(2)
i
i
i i
i
i
A
L
where a
a
d ? 1? v , l
l
d ? 1? v , a ?
i
and l ?
i
. Note that a
d and l
d are the discounts
i
i
i
i
i
? A
i
? A
i
i
i
i
i
i
implicit in the bank’s stock price of a bank’s assets and liabilities relative to book values. At the
same time, a and l are the accounting values of particular assets and liabilities relative to the
i
i
book value of all assets.

From eq. (2), we see that if all assets and liabilities of the bank are valued at market value
in the bank’s balance sheet, then q equals 1. Alternatively, a deviation of q from 1 implies that
the market valuation of at least one balance sheet items differs from its accounting value.10

III. THE DATA
In this study, we consider U.S. bank holding companies that are stock exchange listed.
These companies report a range of accounting data to the Federal Reserve System by way of the
Report on condition and income (Call report). We are using quarterly data from these Call
reports from the final quarter of 2001 till the end of 2008. This covers a full business cycle as
defined by the National Bureau of Economic Research (NBER) from the previous recession
which ended in November 2001 until the current ongoing recession which started in December
2007. Our focus is on the year 2008, one year into the recession and what is generally considered
the start of the U.S. mortgage default crisis (see for example Dell’Ariccia et al., 2008, and Mian
and Sufi, 2008), when delinquencies on mortgage loans increased sharply.

Using stock market data from Datastream, we use the market value of common equity
plus the book value of preferred equity and liabilities as a proxy for the market value of a bank’s
assets. Tobin’s q is then constructed as the ratio of this proxy for the market value of bank assets
and the book value of assets. Figure 1 reports the average Tobin’s q per quarter over our sample
period. The mean value of q has declined from 1.064 in the final quarter of 2001 to 0.998 in the
final quarter of 2008. This suggests that over this period, the market value of bank assets has
declined more than its book value.

We define a zombie bank as a bank with a q of less than one.11 The decline of the average
q has been accompanied by an increase of the share of banks that are zombie banks. As presented

10 Current book values of, say, real estate loans could already reflect some loan loss provisioning. Estimated
discounts on bank assets then reflect the difference between market perception of asset impairment and the
recognition of this impairment through reported loan loss provisioning (rather than the difference between market
value and origination value). Put differently, the estimated discount reflects the difference between market
perception of any asset impairment and the accounting treatment of this impairment.
11 The term zombie bank has frequently been used in the context of Japan during the 1990’s banking crisis when
Japanese banks continued to lend to unprofitable borrowers (e.g., Caballero et al., 2008).

---

9
in Figure 1, the share of zombie banks has increased from 8.2% at the end of 2001 to 60.4% at
the end of 2008. During this period, the share of zombie banks has tended to be smaller than in
2001 and 2008 reflecting an upswing of the business cycle. In fact, the share of zombie banks
reached a low of 0.3% during the second quarter of 2004.


U.S. banks are exposed to the real estate market in two important ways. First, they have
significant portfolios of real estate loans. As an index of this exposure, we construct the ratio of
real estate loans to overall assets. From 2001 to 2008 this share of real estate loans has increased
substantially from 45.2% to 53.3% for the average bank holding company as reflected in Figure
2. Thus, about half of the average bank’s assets consist of real estate loans by 2008. In addition,
banks are exposed to the real estate market through their holdings of MBS. Interestingly, the
average ratio of the book value of MBS to the book value of all assets has increased only slightly
from 10.0% in 2001 to 10.2% at the end of 2008.

While there has been a move towards fair value accounting of bank assets, most assets of
the average bank, including mortgage loans held for investment, are still reported based on
historical cost.12 The book value of MBS reflects different accounting conventions depending on
whether these securities are held-to-maturity or available-for-sale. MBS classified as held-to-
maturity are carried at amortized cost. This amortized cost may be adjusted periodically for
capitalized interest and may also reflect previous loan loss provisioning. However, these
adjustments to amortized cost are likely to be relatively small so that amortized cost is relatively
close to origination values. Alternatively, MBS can be available-for-sale. In this case, these
securities are to be carried at fair value.

Fair value is meant to reflect observed market values (of either the underlying asset –
level 1 assets – or a comparable asset – level 2 assets) or otherwise reflect the outcome of a
bank’s own valuation models (level 3 assets).13 Again, banks’ assessments of fair value may
differ across banking institutions as the determination of fair value in practice leaves banks with
significant discretion.14 At any rate, at a time of declining asset values, one expects fair values to
be less than amortized cost.


12 The majority of (real estate) loans are carried at historical cost, as loans held for sale, that are reported at the lower
of historical cost and fair value, constitute only a small fraction of less than 1% of total assets for the average bank.
13 A breakdown of fair value assets by valuation technique (level 1 to 3) is in principle available from Schedule HC-
Q of the Call report. We do not use this information in our analysis, because, unlike securities that are reported at
both amortized cost and fair value, these assets are reported for only one of the three fair valuation techniques,
making it difficult to draw any inference based on a direct comparison of the amount of assets reported in each
category. Furthermore, the level 1 to 3 assets are not broken down separately for real-estate related assets, which are
the primary focus of our study, and are reported only for a subset of banks that have elected to report such assets
under a fair value option. Moreover, the majority of these assets are valued as level 2 assets (about 90 percent of fair
value assets in 2008), so there is not much variation in fair valuation technique.
14 Indeed, work by Kolev (2009), Goh et al. (2009), and Song et al. (2009) shows that market discounts differ for
level 1, level 2, and level 3 assets.

---