Description
We investigate the determinants of bank representatives’ responses to the United States
Financial Accounting Standard Board’s 2010 Exposure Draft that proposes fair value measurement
for most financial instruments. Over 85% of the 2971 comment letters were
received from bank representatives, with most bank-affiliated letters addressing—and
opposing—one issue: fair value measurement of loans. The Exposure Draft proposes that
companies report both fair value and amortized cost measures for loans; thus, the proposal
should result in increased levels of loan-related information and improved financial reporting
transparency. We investigate three reasons for bank representatives’ resistance. First,
fair value measurement should result in less accounting slack than the current incurredloss
model for loan impairments; therefore, we propose that representatives from banks
that historically utilized that slack will resist fair value measurement for loans. Second,
we propose that agency problems are an important motivating factor because bank representatives
reaping more private benefits from their franchises have less incentive to support
increases in financial reporting transparency
Agency problems, accounting slack, and banks’ response
to proposed reporting of loan fair values
q
Leslie D. Hodder
?
, Patrick E. Hopkins
Kelley School of Business, Indiana University, Bloomington, IN 47405-1701, USA
a b s t r a c t
We investigate the determinants of bank representatives’ responses to the United States
Financial Accounting Standard Board’s 2010 Exposure Draft that proposes fair value mea-
surement for most ?nancial instruments. Over 85% of the 2971 comment letters were
received from bank representatives, with most bank-af?liated letters addressing—and
opposing—one issue: fair value measurement of loans. The Exposure Draft proposes that
companies report both fair value and amortized cost measures for loans; thus, the proposal
should result in increased levels of loan-related information and improved ?nancial report-
ing transparency. We investigate three reasons for bank representatives’ resistance. First,
fair value measurement should result in less accounting slack than the current incurred-
loss model for loan impairments; therefore, we propose that representatives from banks
that historically utilized that slack will resist fair value measurement for loans. Second,
we propose that agency problems are an important motivating factor because bank repre-
sentatives reaping more private bene?ts from their franchises have less incentive to sup-
port increases in ?nancial reporting transparency. Third, we test whether the most
common reasons for opposition included in the comment letters are associated with neg-
ative letter writing. Our analyses support the ?rst two determinants of bank representa-
tives’ resistance to the Exposure Draft. Speci?cally, accounting slack and lower demand
for accounting transparency are strongly associated with resistance to the standard. How-
ever, we ?nd that stated reasons for resistance are not associated with letter writing. Spe-
ci?cally, representatives at ?rms with dif?cult to value loans and ?rms that mostly hold
loans to maturity are no more likely to resist the standard than others. The narrow scope
of bank representatives’ comments and our empirical ?ndings suggest that bankers’
responses to the Exposure Draft may be more driven by concerns over reduced availability
of accounting slack and accompanying de facto regulatory forbearance than by the concep-
tual arguments they offer. Our results have implications for standard setters, who must
navigate special interests as they attempt to promulgate high quality accounting standards,
and for users of ?nancial statements who must consider how political forces shape gener-
ally accepted accounting principles.
Ó 2014 Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.aos.2013.10.002
0361-3682/Ó 2014 Elsevier Ltd. All rights reserved.
q
We thank Rob Bloom?eld, Mark Evans, Patrick Finnegan, Jim Leisenring, Tom Linsmeier, Chris Yust, an anonymous reviewer, workshop participants
from the University of Florida, Georgetown University, Michigan State University, the University of Toronto, and the University of Texas Centennial Research
Conference for their helpful suggestions, and Connor Egan, Mitch Harrison and Brett Jerasa for assisting with coding the comment letters. Professor Hodder
thanks EY, Professor Hopkins thanks the Deloitte Foundation, and both authors thank the Kelley School of Business for their ?nancial support.
?
Corresponding author. Tel.: +1 812 855 9951.
E-mail address: [email protected] (L.D. Hodder).
Accounting, Organizations and Society 39 (2014) 117–133
Contents lists available at ScienceDirect
Accounting, Organizations and Society
j our nal homepage: www. el sevi er. com/ l ocat e/ aos
Introduction
In May 2010, the United States (US) Financial Account-
ing Standards Board (FASB) issued an Exposure Draft (ED)
that proposes greatly expanding fair value recognition for
most ?nancial instruments, including long-term receiv-
ables, such as bank loans (FASB, 2010). Responses received
by the FASB during the ED’s comment period were over-
whelmingly negative and particularly concentrated within
the banking industry; speci?cally, the FASB received 2971
comment letters in response to the ED, with over 85% from
bank representatives and banking trade organizations.
Through the end of 2011, this is one of the highest com-
ment-letter volumes in response to any single FASB-issued
public-comment discussion document. Because of the
large volume of negative letters received from bank repre-
sentatives, the FASB withdrew the proposal in January
2011 (Rapoport, 2011). This study seeks to understand
the factors systematically associated with bank represen-
tatives’ decisions to submit comment letters.
The choice to submit an accounting-standards-related
comment letter (and the speci?c issues discussed in the
letter) should be a function of the perceived economic con-
sequences of the proposed accounting changes (Watts &
Zimmerman, 1978). Despite the many changes proposed
in the ED, the vast majority of letters submitted by com-
mercial-bank representatives addressed only one issue:
opposition to the proposed reporting of loans at fair value.
1
Given the large, uniform, and narrowly focused response
that was concentrated in the banking industry, this suggests
that responding bank representatives perceived an eco-
nomic threat from the FASB’s ED and its proposal to change
current accounting for loans.
Under current US generally accepted accounting princi-
ples (GAAP), companies report loans at amortized cost, ad-
justed for loan impairments determined via an incurred-
loss measurement model. Standard setters, regulators
and practitioners note substantial disagreement about fac-
tors that bank managers should consider in determining
loan-loss provisions under this model (Dugan, 2009; FASB,
2009). In addition, empirical research suggests that banks
use discretion available in loan loss reporting to opportu-
nistically manage income (e.g., Liu & Ryan, 2006) and cap-
ital (e.g., Beatty & Liao, 2011), and that this discretion
reduces transparency and diminishes the ability of outsid-
ers and regulators to monitor banks’ risk-taking behavior
(Bushman & Williams, 2012). Taken together, the practitio-
ners’ criticisms and empirical research ?ndings suggest
that the incurred-loss model for loans is a source of
accounting slack available to bank managers.
The accounting proposed in the ED should result in a
greater amount of relevant information about loans
because the ED requires loans to be reported at fair value
while retaining much of the information currently re-
ported for loans.
2
Speci?cally, the ED requires (1) that amor-
tized cost information for loans be presented on the face of
the statement of ?nancial position (parenthetically, and rec-
onciled to fair value) (FASB, 2010, para. 86), and (2) that esti-
mated credit losses be charged against net income with any
remaining change in the fair value of loans (i.e., that is not
related to counterparty credit quality) recognized in other
comprehensive income (i.e., outside of reported net income)
(FASB, 2010, para. 91).
3
Because only estimated credit losses
will be recognized in net income, and because currently rec-
ognized amortized-cost information will continue to be
prominently displayed on the face of the balance sheet,
the proposed changes should increase the amount of loan-
related information available to outside investors and cred-
itors while retaining key measurement principles embedded
in current accounting standards.
4
However, the changes
should also reduce the amount of accounting slack available
to bank managers and improve the transparency of banks’
loan reporting because loans will be recognized at fair value,
which, by construction, should re?ect more timely recogni-
tion of expected future losses and market-related opportu-
nity costs (Financial Crisis Advisory Group, 2009, p. 7).
Given the proposal’s potential informational bene?ts,
we investigate why representatives from more than 1000
unique commercial banks submitted comment letters
resisting only the loan-related provisions of the ED, while
the vast majority of commercial bank representatives did
not submit letters. We investigate three signi?cant sources
of incentives for resistance to the proposed fair value mea-
surement of loans. First, because fair value measurement of
loans will likely reduce the accounting slack available to
bank managers (i.e., when compared to the current in-
curred-loss model), representatives from banks that have
a history of exploiting the slack available under current
GAAP will have higher incentives to resist a change to fair
1
The FASB’s ED proposes a comprehensive set of changes related to fair
value recognition for most ?nancial instruments (including liabilities),
equity method accounting, derivatives, and hedging. The ED also proposes
that companies continue to report balance sheet and income statement
metrics using amortized-cost basis-measurement. Re?ecting its wide-
ranging scope, the ED contains 214-pages and requests that comment
letters potentially address 71 separate questions (i.e., 36 questions for all
respondents, 13 questions for preparers and auditors, 22 questions for
users).
2
Although current US GAAP requires banks to provide disclosure of loan
fair values in the notes to the ?nancial statements (i.e. FASB Accounting
Standards Codi?cation (ASC) 825-10-55-3), many small- and medium-sized
US banks avail themselves of the ‘‘practicability exception’’ that allows
them to disclose loan value amounts based on simpli?ed cash ?ow models
(i.e. FASB ASC 825-10-50-16 through -19). These reported values often
capture measurement attributes other than fair value (e.g., entry price). As
noted by Tschirhart, O’Brien, Moise, and Yang (2007), these valuation
methodologies need to be substantially modi?ed to arrive at measurements
that meet the de?nition of fair value.
3
Thus, the ED cannot be characterized as a ‘‘fair value versus historical
cost’’ proposal because it retains amortized cost measurement on the face
of the ?nancial statements, while complementing that information with
reconciliation between amortized cost and fair value bases. For these
reasons our analysis cannot address whether either amortized cost
measurement or fair value measurement provides better information in
the absence of the other. Instead, in this study we propose that ?nancial
statement recognition of accounts measured at fair value with supplemen-
tal prominent display of amortized cost information can improve trans-
parency by revealing additional information that is not apparent in
?nancial statements measured solely under amortized cost.
4
Any effect of the ED on net income would necessarily derive from
differences between managers’ estimates of incurred losses and expected
losses arising from non-collection of contractual cash ?ows. The incurred
loss model impounds less information, and is arguably less conservative
than the expected loss model. These terms are discussed in more detail in
the next section.
118 L.D. Hodder, P.E. Hopkins / Accounting, Organizations and Society 39 (2014) 117–133
value measurement for loans and will be more likely to
submit negative comment letters.
5
Second, we propose that
agency problems are an important motivating factor for
responding bank representatives because the incentives for
transparent reporting are decreasing in agency problems
and increasing in agency costs (Jensen & Meckling, 1976);
thus, bankers reaping more private bene?ts from their bank-
ing franchises have incentives to resist increases in ?nancial
reporting transparency.
6
Third, we investigate whether the
likelihood of banks submitting negative comment letters is
associated with the most-commonly cited reasons included
in those letters (i.e., measurement dif?culty and incongru-
ence with banks’ business models).
We use logistic regression analysis to test the determi-
nants of differential bank comment-letter writing. We col-
lect all 2971 comment-period letters submitted to the
FASB in response to the ED, read each of the letters, and
identify those written by commercial-bank representa-
tives. We identify 2472 ?nancial-institution-af?liated let-
ters, which, after adjusting for non-commercial-bank
submissions (e.g., credit unions and savings and loans),
multiple submissions per bank, and data availability, com-
prises 1047 unique commercial banks. This sample repre-
sents approximately 20% of US commercial banks that
have suf?cient ?nancial data available to conduct our anal-
yses (i.e., 1047 out of 5289 banks). Because the majority of
banks are not publicly traded, we collect much of our data
from regulatory ?lings and disclosures.
Our results suggest that bank representatives’ docu-
mented resistance to the loan-accounting provisions in the
ED is positively associated with (1) the historical use of the
accounting slack available through the current incurred-
loss method of accounting for loan losses and (2) proxies
for lower demand for transparency in the responding banks.
Our proxies for reduced demand for reporting transparency
include the choice to purchase additional, temporary, ?at-
rate, excess deposit insurance offeredby the Federal Deposit
Insurance Corporation (FDIC), the choice to acquire unin-
sured debt ?nancing, and the choice to forego independent
audits. The choice to purchase insurance re?ects a trade-
off between the ?xed rate premiums and bene?ts gained
by reducing investor uncertainty. Because we control for
?nancial condition, we interpret the choice to purchase
insurance as a substitute for ?nancial reporting transpar-
ency. The positive association between excess deposit-
insurance purchase and letter writing is consistent with
the ED’s opposition being driven by representatives of less
transparent ?rms. Consistent with this explanation, the
presence of uninsured debt and audited ?nancial
statements were also negatively associated with the likeli-
hood of bank representatives opposing the standard.
Interestingly, we ?nd no evidence that the factors most
commonly asserted by bank representatives are associated
with increased propensity to lobby against fair value mea-
surement for loans. Speci?cally, banks with more dif?cult-
to-value commercial loans appear no more likely to oppose
fair value reporting for loans than those with easier-to-va-
lue loans. In addition, opposite of the relation suggested by
the letters, those banks using fair value for internal deci-
sion-making purposes (i.e., as evidenced by more frequent
loans sales) are signi?cantly more likely to oppose fair va-
lue reporting for loans. Taken together, our ?ndings sug-
gest that preparer comment letters were more in?uenced
by concerns about restriction of accounting slack and by
agency-related manager demand for reporting opacity
than the concerns articulated in the comment letters.
Our results should be considered in the context of due
process for setting accounting standards and for legislating
?nancial regulation. Our study suggests that bank repre-
sentatives writing negative comment letters have motives
to resist accounting standards that result in reduced
accounting slack and increased transparency. As one would
expect, these motives and patterns of ?nancial-reporting
behavior are not mentioned in the comment letters. In-
stead, the negative letters submitted by responding bank
representatives cite conceptual reasons for resisting pro-
posals intended to increase ?nancial reporting transpar-
ency (e.g., insuf?cient precision for fair value
measurements, and irrelevancy of fair values for deci-
sion-making). Our results suggest that the processing of
comment-letters should explicitly include consideration
of letter-writer motivations, and weigh carefully the stated
reasons for support or opposition against letter-writer
incentives and potentially divergent facts.
Issues surrounding due process are complicated, in
practice, by the fact that feedback on proposed accounting
standards is less likely to be received from some constitu-
ent groups. For example, existing research shows that
?nancial statement preparers are much more active in lob-
bying than users (Weetman, Davie, & Collins, 1996; Ryan,
Dunstan, & Stanley, 2000; Sutton, 1984; Tutticci, Dunstan,
& Holmes, 1994). However, users’ lack of lobbying activity
is predicted by theory and may be dif?cult to overcome. To
the extent that high quality ?nancial reporting results in
societal rather than private bene?ts, the relative lack of
user lobbying occurs because private individuals are unli-
kely to voluntarily incur costs to promote public goods (Ol-
son, 1965; Sutton, 1984). Focus groups and individual
interviews may be more effective in eliciting users’ assess-
ment of proposed standards (Weetman et al., 1996).
We infer that regulatory concerns also appear to be an
important determinant of the banking response to the
2010 ED because of the heavy concentration of letter writ-
ing by nonpublic commercial-bank representatives. Many
of these banks are not required to issue general purpose
?nancial statements and are not required to obtain ?nan-
cial statement audits. Beatty, Ke, and Petroni (2002) argue
that private US banks have fewer incentives to manage re-
ported ?nancial information because ownership concen-
trations among manager-shareholders reduce agency
5
We de?ne accounting slack as the acceptability of multiple alternatives
within-GAAP accounting methods and/or numerical estimates for a given
economic transaction, event, right or obligation.
6
According to Jensen and Meckling (1976), ceteris paribus, managers
have a higher likelihood of private bene?t extraction when agency
problems are high and agency costs are low. Agency problems describe
the general condition where an agent with delegated authority has the
opportunity to consume perquisites without the approval of the principal.
Agency costs imposed on the agent mitigate the likelihood that he will
consume perquisites. Thus, the demand for transparency decreases when
(1) the value of perquisite consumption increases or (2) when agency costs
decrease.
L.D. Hodder, P.E. Hopkins / Accounting, Organizations and Society 39 (2014) 117–133 119
costs of equity. Therefore, private-bank representatives’
resistance to the reduction of available accounting slack
likely is driven by the reliance of US government regulators
on US GAAP ?nancial statements.
7
Our ?nding that regula-
tory concerns play a prominent role in bankers’ resistance to
slack-reducing transparency complements the ?ndings of
other studies to suggest that the marriage of Regulatory
Accounting Principles (RAP) and Generally Accepted
Accounting Principles (GAAP) can lead to value-decreasing
behavior by banks (e.g., Hodder, Kohlbeck, & McAnally,
2002), and can decrease the effectiveness of regulatory over-
sight (Bushman & Williams, 2012).
Our study also contributes to the literature on manage-
ments’ reporting incentives and the decisions to lobby for
or against various accounting standards (e.g., Watts & Zim-
merman, 1978). These studies usually are based on the as-
sumed effects that a proposed rule will have on the
primary ?nancial statements of public corporations. In par-
ticular, studies predict increases in lobbying behavior
when proposed standards are anticipated to reduce com-
panies’ levels of income. For example public ?rms resist
expensing the fair value of stock options because doing
so decreases reported net income (Dechow, Hutton, &
Sloan, 1996), and resist including foreign exchange transla-
tion adjustments in net income because these adjustments
increase net income volatility (Kelly, 1985). Our investiga-
tion extends these studies by including a large sample of
public and private companies, and by analyzing lobbying
in the context of agency problems within a regulated
industry where private bene?t extraction is a concern.
Our study is related to Hochberg, Sapienza, and Vissing-
Jørgensen (2009), who show that US public companies
with greater agency problems lobbied more intensely
against Sarbanes–Oxley Act of 2002. Our study is distinct
from Hochberg et al. (2009) because we address the inter-
section of accounting slack, agency problems, ?nancial
reporting standards, and regulatory accounting.
We organize the paper as follows: In Current and
proposed accounting for loans, we discuss current and pro-
posed accounting for loans; in Hypotheses, we propose our
hypotheses; in Sample and research design, we describe
our sample and research design; in Results, we discuss
our research ?ndings; and we provide concluding remarks
in Summary and conclusions.
Current and proposed accounting for loans
Accounting slack inherent in current GAAP for loan losses
Current US GAAP for loan losses is based on an incurred
loss model of accounting for counterparty credit risk.
8
Losses are recognized when probable, based on past events
and economic conditions existing as of the reporting date.
Even in the absence of subjectivity, incurred losses differ
from losses embedded in expected values because expected
values take into account unbiased expectations about future
economic conditions that would give rise to future incurred
losses. Moreover, unlike continuous measurement of losses
via expected values, the incurred loss method in current US
GAAP results in recognition of losses only when the probabil-
ity of loss exceeds the ‘‘probable’’ threshold. For these rea-
sons, unbiased loss provisions recognized using the
incurred loss method will generally result in cumulative loan
loss provisions that are less than cumulative expected losses
re?ected in unbiased expected values or unbiased fair values.
The incurred loss method also appears to have wide-
spread variation in its application (Dugan, 2009; FASB,
2009). At one extreme, required reliance on objective evi-
dence of impairment leads some to argue that losses
should not be recognized until loans become delinquent
or otherwise default. At the other extreme, risk-averse reg-
ulatory ?eld auditors and conservative managers may
cause a bank’s loan loss allowance to converge to or exceed
an unbiased estimate of expected losses. This divergence
causes great variability in the extent to which loan loss
provisions recognized under current GAAP are associated
with nonperforming loans and realized losses in subse-
quent periods (e.g., Beatty & Liao, 2011; Nichols, Wahlen,
& Wieland, 2009; Bushman & Williams, 2013).
In addition to potentially non-strategically divergent
implementation of the existing loan-loss standard, slack
associated with the incurred loss model is actively used to
manage both earnings and capital (Moyer, 1990; Collins,
Shackelford, & Wahlen, 1995; Beatty, Chamberlain, & Magl-
iolo, 1995). Although research generally is mixed on the
consequences of earnings management, evidence on loan-
loss-provision-related earnings and capital management
in the banking industry suggests negative consequences.
For example, Bushman and Williams (2012) ?nd that using
the loan loss provision to smooth earnings reduces trans-
parency, making more dif?cult ef?cient regulatory moni-
toring and undermining market discipline that would
restrain investor losses due to agency problems. In addi-
tion, Beatty and Liao (2011) ?nd that banks delaying loss
recognition suffer from an incremental decrease in lending
capacity during economic downturns that exacerbates the
business cycle. Taken as a whole, prior research suggests
that the incurred loss model of loan-loss recognition under
current GAAP is an important source of accounting slack
available to bank managers and that the use of such slack
can be socially and economically detrimental.
FASB’s 2010 ?nancial instruments exposure draft
In May 2010, the FASB issued an ED for a Proposed
Accounting Standards Update titled, Accounting for Financial
Instruments and Revisions to the Accounting for Derivative
Instruments and Hedging Activities (FASB, 2010). This proposal
includes a comprehensive set of changes relatedto classi?ca-
tion and measurement of ?nancial instruments, impairment
of ?nancial instruments, and hedge accounting. Of the many
changes proposed in the ED, the most controversial—at least
according to responding banks—is the FASB’s proposal that
most ?nancial instruments, including loans, be measured
7
As we describe more fully in the next section, US federal law requires
banking regulators either to accept ?nancial information prepared in
conformity with GAAP or on a basis of accounting that is ‘‘no less stringent’’
than GAAP.
8
Under FASB ASC 825-10, companies have an option to measure most
?nancial instruments (including loans) at fair value. The vast majority of
commercial banks do not exercise this option.
120 L.D. Hodder, P.E. Hopkins / Accounting, Organizations and Society 39 (2014) 117–133
and reported on the balance sheet at fair value. Loans com-
prise thesinglemost signi?cant class of assets oncommercial
banks’ balance sheets, andbyconstruction, recognitionat fair
value should re?ect more timely recognition of expected fu-
ture losses and market-related opportunity costs (Financial
Crisis Advisory Group, 2009, p. 7).
9
Although the ED proposes that most ?nancial instru-
ments should be carried at fair value with the periodic
change in fair value reported in net income (FV-NI), the
ED does allow for a subset of debt instruments held as as-
sets, including loans, to have the non-credit portion of the
change in fair value recognized in other comprehensive in-
come (FV-OCI). Consistent with current GAAP the ED pro-
poses that credit-related impairment of loan value should
be recognized as a component of net income in the period
in which credit-related decline in loan value occurs. Thus,
the most signi?cant effect of the proposal for measurement
of income is the recognition of expected, rather than in-
curred credit losses. Because the ED mandates continuing
?nancial statement reporting of important amortized-cost
amounts, it should result in a net increase in relevant loan-
related information, even if loan fair values include signif-
icant measurement error, as suggested by critics. Speci?-
cally, the ED requires that amortized cost be presented
parenthetically (and reconciled to fair value) on the face
of the statement of ?nancial position.
Bank regulatory accounting
Because bank regulators are legally authorized to di-
rectly collect and audit all necessary ?nancial information
for prudent monitoring, GAAP’s relevance for regulation is
not immediately clear. However, in the case of commercial
banks, US federal law requires that regulatory accounting
principles must be based on GAAP or a basis of accounting
that is ‘‘no less stringent’’ than GAAP (i.e., ‘‘RAP-GAAP con-
formity’’).
10
RAP-GAP conformity was made law in 1991 as
part of the FDIC Improvement Act (FDICIA) in response to
perceived de?ciencies in US bank regulation that contrib-
uted to the S&L crisis. Prior to FDICIA, regulators could en-
gage in case-by-case forbearance; for example, by allowing
an undercapitalized bank to continue to operate while
attempting to raise capital. In addition, regulators had dis-
cretion to exercise ‘‘wholesale’’ forbearance by selectively
relaxing accounting principles to indirectly ease capital
standards.
11
FDICIA also attempted to limit case-by-case dis-
cretion by adopting an objective capital-ratio based system
of prompt corrective action. As reported bank capital ratios
decline, FDICIA prescribes mandatory intervention. To fore-
stall wholesale forbearance, FDICIA also placed limits on reg-
ulators’ ability to selectively override GAAP.
Within the banking industry, RAP-GAAP conformity has
caused GAAP implementation ?exibility to become a source
of regulatory slack. For example, Hodder et al. (2002) docu-
ment that, as a direct result of RAP-GAAP conformity, banks’
initial ‘‘intent-based’’ classi?cation choices for investment
securities were driven more by concerns about potential ef-
fects of classi?cation choice on regulatory capital than by
?rms’ intended holding periods for securities. Speci?cally,
Hodder et al. (2002) show that when regulators relaxed
RAP-GAAP conformity by excluding fromregulatory capital
unrealized gains and losses on securities classi?ed as Avail-
able for Sale, banks responded by reclassifying the majority
of their investment securities from Held to Maturity (i.e.,
recognized at amortized cost) to Available for Sale (i.e., rec-
ognized at fair value). This choice to recognize investment
securities at fair value contrasts markedly with the indus-
try’s near unanimous opposition to fair value recognition
of investment securities that was expressedincomment let-
ters on the FASB’s Exposure Draft of Statement of Financial
Accounting Standards No. 115, Accounting for Certain Invest-
ments in Debt and Equity Securities (Schulz &Hollister, 2003),
suggesting divergence betweencompanies’ statedpositions
in comment letters and their ?nancial reporting behaviors
after implementation of accounting standards.
We propose that the disproportionately large private-
bank-af?liated response to the fair value provisions of
the 2010 ED indicates that RAP-GAAP conformity contin-
ues to heavily in?uence banks’ stated positions against fair
value measurement. In general, private-company repre-
sentatives do not submit high volumes of comment letters
in response to FASB standard-setting proposals because,
unlike public companies, private companies are not statu-
torily required to issue general-purpose GAAP ?nancial
statements. Private companies are more likely to be ?-
nanced with debt capital and may customize ?nancial
information to suit the requirements of their capital pro-
viders (e.g., lenders).
However, FDICIA’s de facto RAP-GAAP conformity has re-
sulted in unprecedented pressure on ?nancial accounting
standard setters to consider the effects of accounting stan-
dards on private companies not otherwise bound by GAAP,
as well as the concerns of bank regulators who are not con-
strained in their ability to collect private ?nancial informa-
tion from regulated enterprises. Cognizant that regulatory
objectives may diverge fromthe US Securities and Exchange
Commission’s (SEC’s) joint objectives of investor protection,
enhancedtransparency, andrelevancy of ?nancial reporting
for outside capital providers, the SEC has resisted the notion
that the content of general purpose ?nancial statements
should be dictated by regulatory concerns (Johnson, 2009).
The Financial Crisis Advisory Group (2009) also noted
that general-purpose ?nancial reporting and ‘‘prudential
regulation’’ have different reporting objectives. However,
despite concerns about divergent regulatory and investor
objectives, the US banking industry continues to press for
oversight over the ?nancial accounting standard setting
process (American Bankers Association, 2009), and the
AICPA has proposed that bank regulators should be in-
cluded in an International Accounting Standards Board
Monitoring Group (AICPA, 2008). Our paper suggests that
pressure to modify measurement principles to achieve de-
sired bank capital ratios places accounting standard setters
in a direct role of bank policy-making and can subvert their
9
Untabulated descriptive statistics reveal that, as a percentage of total
assets, loans for our median sample bank equal 66% of total assets, with the
?rst quartile equal to 54% of total assets.
10
Title 12, Chapter 16 §1831n(a)(2)(A) and (B)).
11
For example, prior to the S&L Crisis, regulators allowed certain losses to
be deferred as assets and amortized instead of immediately recognized in
income. Such changes were inconsistent with GAAP at the time.
L.D. Hodder, P.E. Hopkins / Accounting, Organizations and Society 39 (2014) 117–133 121
designated role of setting standards that provide unbiased,
decision-useful information.
12
Hypotheses
While US RAP-GAAP conformity can explain the heavy
concentration of public and private banks responding to
the ED (compared to other industries), it does not explain
why representatives from 1047 commercial banks submit-
ted comment letters while representatives from the
remaining 4242 commercial banks (for which we have suf-
?cient data to conduct our analyses) did not. US RAP-GAAP
conformity also does not explain why most representatives
from responding banks commented on only one issue re-
lated to one class of assets (i.e., fair value recognition of
loans) in an ED that spanned 214 pages and that requested
responses to 71 separate questions.
13
We investigate three potential determinants of responding
banks’ decisions to submit comment letters in response to the
ED. First, as we describe in the previous section, current US
GAAP for loans and related impairments includes signi?cant
opportunities for bank managers to use accounting slack. In
the banking industry, regulator-imposed RAP-GAAP confor-
mity creates de facto regulatory forbearance that is derived
from accounting slack within GAAP. By retaining important
elements of the current amortized-cost system while also
requiring fair value recognition for loans, the FASB’s ED is
likely to increase the amount and quality of relevant informa-
tion about loans. Further, the proposed rules should constrain
managers’ ability to opportunistically time their recognition
of losses to meet capital and earnings expectations. For these
reasons, we predict that individuals af?liated with banks that
historically have used the reporting ?exibility available in the
incurredloss model to manage net income andcapital will va-
lue that ?exibility and will lobby against accounting stan-
dards that threaten to undermine that ?exibility. This leads
to our ?rst alternative-form hypothesis:
H1. Banks that historically used the ?exibility available
under the incurred loss approach for loan losses are more
likely to lobby standard setters against a change to fair
value measurement of loans.
Second, we propose that because the additional disclo-
sures required by the standard will increase transparency,
private bene?t extraction facilitated by high agency prob-
lems and lowagency costs is an important motivating factor
for responding bank representatives. Incentives for trans-
parent reporting are decreasing in net bene?t extraction
(Jensen & Meckling, 1976). This means that bank represen-
tatives obtaining more private bene?ts relative to agency
costs imposed by outside capital providers will have less
incentive to improve banks’ ?nancial reporting transpar-
ency. This leads to our second alternative-form hypothesis:
H2. Banks with lower demand for transparent reporting
are more likely to lobby standard setters against fair value
measurement of loans.
Although banks’ demand for decreased ?nancial report-
ing transparency is unobservable, this demand should be
positively related to other observable choices that are indi-
cators that the value of private-bene?t extraction exceeds
the potential value of reduced agency costs that would
be achieved via greater reporting transparency. Speci?-
cally, we predict that banks submitting comment letters
opposing the loan-reporting provisions of the ED are (1)
more likely to elect to participate in the new optional, sup-
plemental, ?xed-price deposit insurance and debt guaran-
tee programs offered by the FDIC, (2) less likely to have
uninsured outside debt and (3) less likely to voluntarily
purchase ?nancial statement audits of their non-regula-
tory GAAP ?nancial statements.
When government-sponsored insurance is provided for
bank liabilities, total agency costs decline because insured
depositors do not demand a risk premium and opportuni-
ties for private bene?t extraction increase because regula-
tors do not have incentives to provide the same level of
monitoring provided by private creditors (Forssbaeck,
2011). Therefore, banks purchasing additional ?at rate
insurance have lower agency costs, greater opportunities
for private bene?t extraction, and less incentive to increase
?nancial reporting transparency. Similarly, and consistent
with Jensen and Meckling (1976), ?rms choosing to ?nance
with uninsured debt are more likely to bene?t from a com-
mitment to transparent reporting, and ?rms that value
transparent reporting are more likely to voluntarily pur-
chase ?nancial statement audits.
The third reason bank representatives may have sub-
mitted comment letters is that they believe the standard
would negatively affect the relevance and reliability of
their ?rms’ ?nancial reports. For example, many of the
comment letters speci?cally mention the belief that fair
values are dif?cult to measure and the belief that fair value
is not relevant for loans because loans are generally held to
maturity. Thus, we also investigate banks’ most commonly
stated reasons for opposition to the ED. This leads to our
third alternative-form hypothesis:
H3. Banks decision to lobby standard setters against fair
value measurement of loans is related to the reasons for
opposition included in the comment letters.
Sample and research design
Sample selection
Our primary sample consists of all private and public
commercial banks with necessary ?nancial data in the
12
Regulators’ unwillingness to expand disclosure in the middle of the
credit crisis might be explained by political incentives. Kane (1989)
proposes that information suppression provides plausible deniability for
regulators whose short horizons allow problems to be passed off to
successors.
13
The ED includes other proposals that would affect banks’ regulatory
reporting under RAP-GAAP conformity. For example, as compared to
today’s mostly amortized-cost accounting, the ED proposes that companies
initially recognize and subsequently remeasure most liabilities at fair value
in the balance sheet. Despite this potentially major change in liability
measurement, the comment letters were almost entirely silent about the
liability proposal. We note that, compared to incurred-loss accounting for
loans, amortized-cost accounting for liabilities does not represent a
signi?cant source of accounting slack for most banks; thus, bank repre-
sentatives do not have a strong incentive to lobby against it.
122 L.D. Hodder, P.E. Hopkins / Accounting, Organizations and Society 39 (2014) 117–133
periods covered by our tests. We collect ?nancial state-
ment data from publicly available regulatory ?lings. Our
proxy for banks’ historical use of discretion under the in-
curred loss model requires quarterly data, so we also in-
clude quarterly data spanning 2001Q1 through 2011Q4.
Our sample includes 5,289 unique banks.
From the FASB’s web site, we individually collect all
2,971 comment letters submitted in response to the ED.
14
Two accounting graduate students analyzed each letter
and coded its content. For each letter we collect any identi-
fying information provided by the letter writer, including
name, state, and company af?liation. For letters not provid-
ing any company af?liation, we collect implicit company
af?liation by cross-referencing the letter-writer name, email
address domain, street address, and phone number to
company pro?les. If the letter writer has a documented asso-
ciation with a company (e.g. of?cer, director, or employee)
and takes an industry-related position on the issues, we
attribute company af?liation to the letter writer for statisti-
cal purposes. We classify the letters as either supporting or
opposing various issues mentioned in the ED, including fair
value measurement for loans, (more generally) fair value
measurement for assets and liabilities, equity method
accounting, hedge accounting, convergence with Interna-
tional Financial Reporting Standards (IFRS), and other
miscellaneous issues. The coders did not have knowledge
of our hypotheses, or our proxies for banks’ prior use of
accounting slack and agency problems.
To avoid double-counting of entities, we do not include
bank holding companies in our sample together with their
consolidated subsidiaries. Consistent with this convention,
letters written by personnel of holding companies are
attributed to the subsidiary banks, and the subsidiary
banks are included in the sample, but not the holding com-
pany. We attribute letters to preparers, users, or others
based on company af?liation. Because stated af?liation
may differ from implied af?liation, we use the following
classi?cation hierarchy to resolve contradictory identifying
information: (1) letterhead, (2) identi?able as of?cer or
director of ?rm, (3) identi?able as ?rm representative with
phone number, mailing address, of ?rm, or email domain
of a ?rm, and (4) self-referenced af?liation provided in
the text of the correspondence.
Panel A of Fig. 1 provides an example of a letter with
contradictory identifying information. Although the writer
self-identi?es as a ‘‘bank investor’’ in the text and as an
‘‘investor’’ in the signature line, we classify the letter as
preparer-originated because, according to regulatory ?l-
ings, the writer is the Chief Executive Of?cer of a commer-
cial bank. Correspondingly, we classify the industry as
‘‘Financial Firm.’’
Model speci?cation
To test our hypotheses, we estimate the following
model:
PrðLetterÞ
t
¼ a þ b
1
TIMELINESS
t
þ b
2
DEP INS
t
þ b
3
CREDITOR
t
þ b
4
AUDITED
t
þ b
5
COMM RBC
t
þ b
6
SOLD
t
þ b
7
NPA PCT
t
þ b
8
CAPITAL
t
þ b
9
ENFORCEMENT
t
þ b
10
SMALL
t
þ b
11
SIZE
t
þe
t
ð1Þ
where TIMELINESS – (H1) The increase in explanatory
power for the loan loss provision of future non-performing
assets over past non-performing assets. (Refer to the
Appendix for the exact speci?cation of this variable.); DE-
P_INS – (H2) An indicator variable equal to 1 if the com-
pany elected to purchase additional insurance from the
FDIC during 2008 and 2009; CREDITOR – (H2) Indicator
variable equal to 1 if the ?rm has outstanding uninsured
long-term debt, and 0 otherwise; AUDITED – (H2) Indicator
variable equal to 1 if the ?rm engages an independent
auditor to conduct an audit of ?nancial statements pre-
pared in conformity with GAAP, and 0 otherwise;
COMM_RBC – (H3) Total commercial and industrial loans
(including real estate) divided by risk-based capital at
December 31, 2009; SOLD – (H3) Indicator variable equal
to 1 if the ?rm sold loans during the two years ending
December 31, 2009, and 0 otherwise; NPA_PCT – (Control)
Average nonperforming assets divided by the average bal-
ance in loans computed over the year ended December 31,
2009; CAPITAL – (Control) Tier 1 regulatory capital as of
December 31, 2009 divided by average assets over the year
ended December 31, 2009; ENFORCEMENT – (Control) Indi-
cator variable equal to 1 if the bank was under an active
enforcement order in the 5 years preceding the beginning
of the comment period, and 0 otherwise; SMALL – (Control)
Indicator variable set to 1 if the average balance of total as-
sets for the year ended December 31, 2009 is less than or
equal to $10 billion, and 0 otherwise; SIZE – (Control) Nat-
ural log of average assets computed over the year ended
December 31, 2009.
All continuous variables are winsorized at the top and
bottom 1%. TIMELINESS is included to test H1, and is a
proxy for banks’ historical use of slack available in current
GAAP for loan losses for which we predict a negative coef-
?cient. The Appendix provides detailed speci?cation of
TIMELINESS. High levels of TIMELINESS are associated with
lower historical use of accounting slack. TIMELINESS is
measured as the increase in explanatory power for banks’
loan loss provision of future non-performing assets as
compared to past non-performing assets.
15
Because we re-
quire at least 5 degrees of freedom in the ?rm-speci?c esti-
mation, we cannot compute TIMELINESS for all ?rms in the
sample.
14
In January 2014, the letters were available athttp://www.fasb.org/jsp/
FASB/CommentLetter_C/CommentLetterPage&cid=1218220137090&pro-
ject_id=1810-100&page_number=1.
15
Greater timeliness is not equivalent to conservatism because timeliness
is symmetric and captures the extent to which the current provision
captures both positive and negative future changes in non-performing
assets. For example, a provision would be timely if future expected
decreases in nonperforming loans were recognized as a reduction in the
current provision; however, the timely recognition of expected future
decreases in nonperforming assets would probably not be considered
conservative.
L.D. Hodder, P.E. Hopkins / Accounting, Organizations and Society 39 (2014) 117–133 123
Fig. 1. Examples of Fair Value Related Letters. Panels A and B of Fig. 1 provide examples of comment letters opposing and supporting fair value
measurement for banks, respectively. The letter in Panel A also provides an example of a bank manager representing himself as an investor and failing to
disclose that, at the time he submitted the letter, he was also the CEO and Chairman of the Board of Directors of a bank holding company. We were able to
con?rm the letter writer’s identity by matching his name and home address to current bank regulatory reports and SEC ?lings for a predecessor bank at
which he was an executive manager.
124 L.D. Hodder, P.E. Hopkins / Accounting, Organizations and Society 39 (2014) 117–133
For our tests of H2, we use three proxies for the demand
for transparency in banks: DEP_INS, CREDITOR, and
AUDITED, all obtained from SNL, Inc. DEP_INS is based on
banks’ decisions to purchase supplemental deposit
insurance through the temporary Transaction Account
Guarantee Program (TAGP) or Debt Guarantee Program
(DGP). The one-time supplemental insurance was offered
by the FDIC to forestall potential depositor runs at a time
Fig. 1 (continued)
L.D. Hodder, P.E. Hopkins / Accounting, Organizations and Society 39 (2014) 117–133 125
when con?dence in ?nancial institutions was at a nadir.
Because premiums were ?xed per unit coverage, supple-
mental insurance provided the most bene?t to ?rms with
?nancial conditions about which investors were most
uncertain. To the extent that greater uncertainty is associ-
ated with lower transparency, we expect DEP_INS to be
positively associated with banks’ decisions to lobby against
the proposed exposure draft.
Because banks with outside creditors have greater
incentives to reduce agency costs by providing transparent
?nancial information, we expect a negative coef?cient on
CREDITOR. For similar reasons, we include an indicator
variable for AUDITED, which is equal to 1 if the ?rm elects
to engage an independent auditor to attest that its general
purpose ?nancial statements are prepared in accordance
with GAAP. Firms distributing audited GAAP ?nancial
statements have made a commitment to more transparent
disclosure and, by engaging an auditor, have constrained
their own discretion. We expect a negative coef?cient on
AUDITED.
In H3, we investigate whether bank representatives’
decisions to submit comment letters are associated with
factors explicitly mentioned in the comment letters. We
include two variables designed to capture reasons most
commonly included in the comment letters. First, bank
representatives expressed concern over measurement of
hard-to-value loans. Speci?cally, they noted that a move
from current GAAP to the ED’s fair value measurement
for loans could cause banks with higher concentrations of
harder-to-value assets to incur signi?cant direct and indi-
rect costs in the preparation of periodic ?nancial state-
ments and regulatory reports. To control for variation in
the relative dif?culty of valuing loans across banks, we
include COMM_RBC, which is equal to total commercial
and industrial loans (including real estate) divided by
risk-based capital at December 31, 2009. These loan types
are most frequently mentioned as examples of hard-to-va-
lue loans. If negative comment letters are more likely from
banks with these types of loans, the coef?cient on
COMM_RBC will be positive.
A second and even more ubiquitous justi?cation in-
cluded in bank representatives’ negative comment letters
is that the respondents’ banks do not use fair values for
internal decision-making because fair values are irrelevant
to managers and investors. To control for this stated belief,
we identify a context in which fair values are clearly rele-
vant: the decision to sell certain loans in the portfolio. We
argue that banks making informed judgments to sell sub-
sets of loans for purposes that include liquidity, credit risk
management, and interest rate risk management, must
understand values, liquidity risks, and market risks inher-
ent in all eligible loan assets.
16
In contrast, representatives
from banks not engaging in loan sales might be motivated
to submit negative comment letters because they believe
fair value measurement is irrelevant given their business
models. To provide an additional text of H3, we include
SOLD as an indicator variable equal to 1 if the ?rm sold
any of its loans during the two years ending December 31,
2009, and 0 otherwise. If bankers’ assertions are accurate,
then we expect a negative coef?cient on SOLD.
We also include several control variables designed to
capture banks’ propensity to lobby. To control for the po-
tential effects of banks’ loan quality on the likelihood of
negative comment letter submission, we include NPA_PCT,
which is the average balance of nonperforming assets di-
vided by the average balance in loans computed over the
year ended December 31, 2009. The coef?cient on
NPA_PCT will be positive if banks with poor loan quality
were more likely to oppose the standard.
Dhaliwal (1982) proposes that capital structure will
in?uence the choice to lobby. Speci?cally, if a proposed
accounting standard presents a threat to capital levels or
is believed to increase capital volatility, then capital levels
should be negatively related to the choice to lobby against
the standard. This is particularly true in the banking indus-
try because regulators impose minimum capital standards.
For these reasons we expect a negative sign on the control
variable, CAPITAL.
An implicit assumption underlying banks’ resistance to
proposed fair value standards is that RAP-GAAP conformity
has caused the incurred loss model to become an impor-
tant source of regulatory slack that equates to forbearance
for troubled banks. If this is true, then those banks not eli-
gible for forbearance should be less likely to lobby for
accounting slack. Speci?cally, banks identi?ed by regula-
tors as troubled are given very little slack and are subjected
to enhanced monitoring, additional reporting, and the
imposition of higher minimum capital levels. For these
banks, regulatory constraints are already binding and
accounting slack is less valuable. To incorporate the effects
of active enforcement, we include an indicator variable,
ENFORCEMENT, which is equal to 1 if the ?rm has been
subject to a regulatory order in the ?ve years preceding
the beginning of the comment period, and 0 otherwise.
We predict a negative sign on the control variable,
ENFORCEMENT.
Two factors lead us to believe that an unusually high
proportion of small banks will respond to the ED. First,
RAP-GAAP conformity causes even the smallest banks to
be subject to the ED’s mandated fair value measurement
of loans in their regulatory ?lings, and loans are—by far—
the single largest class of assets on commercial banks’ bal-
ance sheets. Second, the American Bankers Association
(ABA) and state banking associations mounted an intense
lobbying campaign targeted speci?cally at smaller regional
and community banks (Ciesielski, 2010, p. 1). As part of
this campaign, these associations made available resources
that facilitated comment letter submission by less-sophis-
ticated banks. To control for this, we include SMALL, which
is equal to 1 if AVG_ASSETS (i.e., average daily balance of
total assets for the year ended December 31, 2009 reported
on the bank’s call report) is less than or equal to $10 billion,
and 0 otherwise. We predict a positive sign on the control
variable, SMALL.
By separately controlling for small banks, we are able to
also control for bank size within the small versus large par-
tition of banks. To accomplish this we also include SIZE,
which is the natural log of average assets computed over
16
Even if banks choose only to sell subsets of loan assets, all loan assets
are eligible for sale. Markets exist for most loan assets, even those with
unique terms (Ryan 2007).
126 L.D. Hodder, P.E. Hopkins / Accounting, Organizations and Society 39 (2014) 117–133
the year ended December 31, 2009. After controlling for
SMALL in the regression equation, we expect larger banks
within each of the small and large partitions to submit
comment letters. Therefore, we predict a positive sign on
the control variable, SIZE.
17
Results
Hypotheses tests
Panel A of Table 1 presents descriptive statistics for the
2971 comment letters obtained from the FASB in response
to the 2010 Exposure Draft. Consistent with prior litera-
ture, we ?nd that preparers are much more likely than
users to submit letters, with preparers accounting for
91.7% of all letters compared to 5.3% of letters attributable
to individual investors and investment advisors (Ryan
et al., 2000; Sutton, 1984; Tutticci et al., 1994; Weetman
et al., 1996).
18
As we previously noted, ?nancial statement
users (e.g., current and prospective investors) may be reluc-
tant to invest in accounting-standards-related lobbying be-
cause the cost of disclosure is borne by the disclosing
parties (i.e., it is only indirectly shared by current investors),
and the resulting bene?ts of enhanced disclosure is a public
good (Olson, 1965; Sutton, 1984).
Representatives from ?nancial ?rms, which include
banks, credit unions, specialty ?nance, and insurance com-
panies, prepared the vast majority of comment letters,
accounting for 2525 (i.e., 85%) out of 2971 separate pieces
of correspondence. The content of the correspondence ad-
dresses a variety of concerns, including fair value measure-
ment, hedging, convergence with IFRS, and the equity
method of accounting. However, the vast majority of re-
sponses was submitted by commercial-bank representa-
tives, and the subject of this correspondence is generally
limited to criticism of the proposed expansion of fair value
measurement for loans.
For purposes of analysis, we focus on comments
addressing fair value measurement and code the comments
Table 1
Descriptive statistics comment letters.
Af?liation subgroup Number Overall percent (%) Positive Negative Mixed Subgroup percent negative (%)
Panel A: Comment letters categorized by source and overall letter position
Preparers
Financial ?rms 2525 85.0 0 2521 4 99.8
Bank Trade Assoc. 50 1.7 0 50 0 100.0
Other ?rms 150 5.0 7 116 27 77.3
Total preparers 2725 91.7 7 2687 31 98.6
Users
Investors/advisors 131 4.4 12 114 5 87.0
Other/individuals 27 0.9 3 18 6 66.7
Total users 158 5.3 15 132 11 83.5
Other
Public accountants 35 1.2 0 33 2 94.3
Regulators 18 0.6 0 15 3 83.3
Other 35 1.2 3 24 8 68.6
Total other 88 3.0 3 72 13 81.8
Overall total 2971 100.0 25 2891 55
Overall total percent 0.8% 97.3% 1.9%
Number
Panel B: Reconciliation of total comment letters to sample of commercial banks submitting letters
Total letters 2971
Non-?nancial ?rms 446
Duplicate letters from individual ?rms 828
Potential ?nancial ?rm sample 1697
Missing required data ?elds 650
Final sample 1047
It reports descriptive statistics for all 2971 comment letters received by the Financial Accounting Standards Board (FASB) in response to its Exposure Draft
titled, Accounting for Financial Instruments and Revisions to the Accounting for Derivative Instruments and Hedging Activities (FASB, 2010). Panel A categorizes
the comment letters according to their source. Panel B provides a reconciliation of the total number of letters received by the FASB to the ?nal unique ?rm
sample used in our statistical analyses of commercial banks.
17
Size has been used in prior research as a proxy for political costs (Watts
& Zimmerman, 1978) because theory suggests that large ?rms are more
likely to lobby for income-decreasing accounting standards that reduce
political visibility. Further, scale economies are likely important both for
the probability of engaging in lobbying activity and for the potential gains
from in?uence (Hill, Kelly, Lockhart, & Van Ness, 2011). Given the relative
large number of small commercial banks that submitted comment letters,
these theoretical relations do not appear to apply in the present setting. The
results of our analyses are qualitatively similar (i.e., coef?cient direction,
magnitude and signi?cance levels for accounting-slack and agency-related
independent variables) if we omit SMALL, SIZE or both variables from our
speci?cation.
18
Our tabulation may understate the number of users, because some user
groups, including Financial Executives International surveyed their mem-
bers and conveyed the tabulated results in a single letter. Extrapolation of
survey results to letter counts would have resulted greater user participa-
tion and a higher ratio of favorable responses to the ED.
L.D. Hodder, P.E. Hopkins / Accounting, Organizations and Society 39 (2014) 117–133 127
as ‘‘Positive,’’ ‘‘Negative,’’ or ‘‘Mixed.’’
19
Panel A of Fig. 1 pro-
vides an example of a comment letter which we coded ‘‘Neg-
ative’’ with respect to fair value. Panel B provides an example
of a ‘‘Positive’’ response. Responses that were neither clearly
‘‘Positive’’ nor ‘‘Negative’’ were coded ‘‘Mixed.’’ Our coding
reveals that ?nancial ?rms are almost uniformly negative
about fair value (99.8%), while non-?nancial ?rms are also
negative, but signi?cantly less so (77.3%, X
2
p
= 0.000). Those
classi?ed as users are less negative about fair value than pre-
parers (X
2
p
= 0.000). Because no commercial banks take a
favorable position with respect to fair value of loans, we treat
the entire commercial bank sample as ‘‘Negative’’ for the
remaining analyses.
Panel B of Table 1 provides a reconciliation of the num-
ber of letters to the ?nal unique-?rm sample used for the
remaining statistical analyses. Non-?nancial ?rms
accounted for 446 letters. Many ?nancial ?rms were
represented multiple times, leading to deletion of 828
pieces of correspondence to arrive at our unique ?rm sam-
ple. The maximum number of letters sent by individuals
associated with a single ?rm was 46 (untabulated). Credit
unions, savings banks, specialty ?nance, and commercial
banks missing necessary quarterly data items account for
another 650 pieces of omitted correspondence. The ?nal
sample comprises 1047 unique commercial banks.
Table 2 provides descriptive statistics for our regression
variables. Except as otherwise noted, regression variables
are measured as of December 31, 2009. Statistical compar-
isons reported in Table 2 suggest that lobbying banks differ
from non-lobbying banks on several dimensions.
20
Related
to our proxy for banks’ propensity to use prior accounting
slack, TIMELINESS, measured as the change in R
2
between
Table 2
Descriptive statistics for lobbying and non-lobbying banks.
Lobbying banks (n = 1047) Non-lobbying banks (n = 4242)
Mean Q1 Median Q3 Mean Q1 Median Q3 t Test Nonpara-metric
H1: Proxy for prior use of accounting slack
TIMELINESS 0.103 0.030 0.069 0.138 0.113 0.033 0.078 0.156 ÃÃÃ ###
H2: Proxies for the demand for transparency
DEP_INS 0.854 0.817 + ###
CREDITOR 0.023 0.031 +
AUDITED 0.452 0.482 + #
H3: Proxies for reasons provided in comment letters
COMM_RBC 3.351 1.907 3.205 4.515 3.359 1.652 3.085 4.742
SOLD 0.105 0.085 + ##
Control and other variables
NPA_PCT 0.021 0.004 0.012 0.026 0.024 0.003 0.012 0.029 ÃÃÃ ##
CAPITAL 9.780 8.260 9.160 10.700 10.167 8.140 9.300 11.220 ÃÃÃ ###
ENFORCEMENT 0.325 0.376 + ###
AVG_ASSETS 0.344 0.080 0.163 0.333 1.477 0.064 0.131 0.283 ÃÃÃ
SMALL 0.939 0.929 +
SIZE 12.046 11.291 11.999 12.717 11.936 11.069 11.782 12.553 ÃÃ
R
2
Eq. (A2) 0.298 0.150 0.258 0.413 0.322 0.160 0.279 0.451 ÃÃÃ ###
ALLOW/LOANS 0.016 0.012 0.015 0.019 0.018 0.012 0.015 0.021 ÃÃÃ ##
Notes: The ?nal sample includes 5289 banks, which is comprised of 1047 unique lobbying banks from which the FASB received comment letters (i.e.,
described in Table 1) and 4242 non-lobbying banks from which the FASB did not receive comment letters and for which we are able to obtain necessary
?nancial data. LETTER is an indicator variable equal to 1 if the bank submitted a comment letter in response to the ED (i.e., left half of Panel A), and 0
otherwise (i.e., right half of Panel A); TIMELINESS is the ?rm-speci?c difference in adjusted R
2
from estimating Eq. (A2) compared to that derived from
estimating Eq. (A1) described in Appendix A; DEP_INS is an indicator variable equal to 1 if the ?rm elected to participate in the Transaction Account
Guarantee Program or the Debt Guarantee Program when these programs were offered by the Federal Deposit Insurance Corporation, and 0 otherwise;
CREDITOR is an indicator variable equal to 1 if the ?rm has outstanding non-deposit long term debt on December 31, 2009, and 0 otherwise; AUDITED is an
indicator variable equal to 1 if the ?rm engages an independent auditor, and 0 otherwise; COMM_RBC is total commercial and industrial loans (including
real estate) divided by risk-based capital at December 31, 2009; SOLD is an indicator variable equal to 1 if the ?rm sold any of its loans during the two years
ending December 31, 2009, and 0 otherwise; NPA_PCT is the average nonperforming assets divided by the average balance in loans computed over the year
ended December 31, 2009; CAPITAL is Tier 1 regulatory capital as of December 31, 2009 divided by average assets over the year ended December 31, 2009;
ENFORCEMENT is an indicator variable equal to 1 if the bank was under an active enforcement order in the 5 years preceding the beginning of the comment
period, and 0 otherwise; AVG_ASSETS is the average daily balance of total assets for the year ended December 31, 2009 reported on the bank’s call report;
SMALL is a an indicator variable set to 1 if AVG_ASSETS is less than or equal to $10 billion, and 0 otherwise; SIZE is the natural log of AVG_ASSETS; R
2
Eq.
(A2) is the adjusted R
2
from the ?rm-speci?c estimation of Eq. (A2) described in Appendix A; ALLOW/LOANS is the ratio of the allowance for loan losses
divided by loans at December 31, 2009.
The t-test column reports two-sample pooled t-test signi?cance levels with Satterthwaite approximations of degrees of freedom in cases of unequal sample
variance. ÃÃÃ, ÃÃ, and à denote parametric statistical signi?cance at the 1%, 5%, and 10% (two-tail) levels, respectively. For variables measured on a ratio
scale, we report Savage two-sample nonparametric signi?cance levels. Indicator variables are denoted ‘‘+’’ in the t-test signi?cance column, and are only
tested using nonparametric X
2
test statistics (i.e., we do not report parametric signi?cance levels for indicator variables). ###, ##, and # denote non-
parametric statistical signi?cance at the 1%, 5%, and 10% (two-tail) levels, respectively.
19
Responses were coded twice by independent research assistants. There
were no cases of disagreement.
20
Table 2 reports two-sample pooled t-test signi?cance levels with
Satterthwaite approximations of degrees of freedom in cases of unequal
sample variance. For variables measured on a ratio scale, we report Savage
two-sample nonparametric signi?cance levels. Indicator variables are only
tested using nonparametric X
2
test statistics.
128 L.D. Hodder, P.E. Hopkins / Accounting, Organizations and Society 39 (2014) 117–133
Eqs. (A2) and (A1), is larger for non-lobbying ?rms (0.113
compared to 0.103; p < 0.01). This is consistent with non-
lobbying ?rms’ provisions incorporating more timely for-
ward-looking information in the years preceding the com-
ment period.
Turning to our H2 proxies for the demand for transpar-
ency, banks receiving additional deposit insurance are
more likely to lobby against fair value (85.4% compared
to 81.7%; p < 0.01). In addition, banks choosing to provide
audited ?nancial statements prepared in accordance with
GAAP appear to be less likely to submit comment letters,
with 45.2% of lobbying banks purchasing ?nancial state-
ment audits, compared to 48.2% of non-lobbying banks
(p < .10). In contrast, the simple univariate comparisons
of CREDITOR reveal that lobbying and non-lobbying banks
are not statistically different in their propensity to have
non-deposit, long-term debt. However, overall, univariate
results are consistent with our hypotheses.
Table 2 reveals an interesting pattern in our H3 proxies
for reasons provided by responding banks in their com-
ment letters. First, there appears to be no statistical differ-
ence in the relative proportions of hard to value
commercial and industrial loans across lobbying and
non-lobbying banks. Second, banks opposing the ED actu-
ally had higher relative amounts of loans sold in the period
leading to the issuance of the FASB’s ED (0.105 versus
0.085, p < 0.05). Thus, there is no univariate evidence to
suggest that the reasons for opposition to the ED stated
in the comment letters drive the propensity to oppose
the ED.
We now turn to our control and other variables. Coun-
ter to the argument that banks with higher levels of non-
performing assets will resist a move to fair value
measurement, the relative proportion of NPA_PCT is higher
in non-lobbying banks than in banks opposing the ED
(0.024 compared to 0.021, p < 0.05). The average value of
CAPITAL of non-lobbying banks is signi?cantly higher than
that of lobbying banks (10.167% compared to 9.780%;
p < 0.01). A smaller percentage of banks with open
ENFORCEMENT actions in the 5 years preceding the com-
ment period choose to lobby (32.5% compared to 37.6%;
p < 0.01), consistent with these banks (1) reaping lower
bene?ts from accounting slack and (2) having decreased
desire for visibility given the extreme level of regulatory
scrutiny accompanying ENFORCEMENT.
The size distribution AVG_ASSETS is highly skewed for
both groups, as suggested by the fact that medians are
much smaller than the means. The mean value of AVG_AS-
SETS for non-lobbying ?rms is larger than lobbying ?rms
(average assets of $1.477 billion compared to $0.344 bil-
lion); however, the median values do not statistically dif-
fer. Given the skewed distribution, we use the natural log
of average assets in our regressions to capture SIZE. We
?nd that the mean value for SIZE is signi?cantly larger
for lobbying ?rms (12.046) than that of non-lobbying ?rms
(11.936) (p < 0.05), but the medians are, again, not statisti-
cally different. We include the variable SMALL because,
compared to companies’ responses to other FASB propos-
als, an unusually large number of small, private banks sub-
mitted comment letters. Interestingly, the proportion of
small banks submitting comment letters appears to
represent the overall proportion of small banks in the en-
tire sample because the difference in SMALL between lob-
bying (93.9%) and non-lobbying banks (92.9%) is not
statistically different.
Related to banks’ past use of the accounting slack avail-
able in the incurred loss model for loans, we ?nd that the
average R
2
corresponding to the ?rm-speci?c estimation
of Eq. (A2) is signi?cantly lower for lobbying ?rms
(0.298) than for non-lobbying ?rms (0.322), suggesting
that lobbying banks’ provisions are more likely to deviate
from amounts predicted by past and forward-looking
explanatory variables (p < 0.01). We also ?nd that, in abso-
lute terms, lobbying ?rms have signi?cantly lower allow-
ances for loan losses as a percentage of loans (0.016
compared to 0.018, p < .01).
Table 3 presents statistics from our estimations of Eq.
(1) with TIMELINESS serving as our H1 proxy for banks’
historical use of the slack available in the incurred-loss
model for loans; DEP_INS, CREDITOR and AUDITED serving
as H2 proxies for banks’ demand for decreased transpar-
ency, and COMM_RBC and SOLD serving as our H3 proxies
for reasons commonly included in banks’ comment letters.
Overall, the model explanatory power is signi?cant for Eq.
(1) (log-likelihood of 76.78, p < 0.000), with the overall
Table 3
Logistic regression of banks’ probability of lobbying on hypothesized
determinants.
Pred. Eq. (1)
Sign Coeff. X
2
Intercept ? À2.251 42.64
ÃÃÃ
H1: Prior use of accounting slack
TIMELINESS
t
À À0.489 6.96
ÃÃÃ
H2: Demand for transparency
DEP_INS
t
+ 0.141 6.17
ÃÃÃ
CREDITOR
t
À À0.268 3.99
ÃÃ
AUDITED
t
À À0.094 4.97
ÃÃ
H3: Reasons in comment letters
COMM_RBC
t
+ À0.001 2.28
SOLD
t
À 0.204 5.37
ÃÃ
Control variables
NPA_PCT
t
+ À1.605 6.00
ÃÃ
CAPITAL
t
À À0.016 6.69
ÃÃÃ
ENFORCEMENT
t
À À0.126 7.97
ÃÃÃ
SMALL
t
+ 0.433 13.89
ÃÃÃ
SIZE
t
+ 0.106 19.72
ÃÃÃ
Model Log-likelihood 76.78
DF 11
p-Value 0.10), while repre-
sentatives from banks that sell loans were more likely to
submit negative comment letters (b
6
= 0.204, X
2
= 5.37,
p < 0.05).
Turning to our control variables, we ?nd that represen-
tatives from banks with higher levels of nonperforming
assets were actually less likely to submit negative com-
ment letters (b
7
= À1.605, X
2
= 6.00, p < 0.05). We also ?nd
that, as predicted, our control variables CAPITAL and
ENFORCEMENT are signi?cantly negatively associated with
lobbying (b
8
= À0.016, X
2
= 6.69, p < 0.01 and b
9
= À0.126,
X
2
= 7.97, p < 0.01 respectively). The results for these vari-
ables are also consistent with the signi?cance of regulatory
factors in banks’ decisions to lobby against fair value
accounting for loans. Low CAPITAL levels immediately
preceding the comment period in?uenced ?rms to lobby
for accounting slack that would be bene?cial in maintain-
ing compliance with minimum regulatory capital
standards, while the presence of ENFORCEMENT damp-
ened banks’ propensity to lobby, consistent with bankers
perceiving lower bene?ts of accounting slack when regula-
tory scrutiny is high.
The signi?cantly positive coef?cient on SMALL
(b
10
= 0.433, X
2
= 13.89, p < 0.01) con?rms our observation
that a disproportionally large number of small banks re-
sponded to the FASB’s ED, while the positive and signi?-
cant coef?cient on SIZE (b
11
= 0.106, X
2
= 19.72, p < 0.01)
is consistent with the notion that after controlling for
SMALL (and other factors), the larger banks within the
small/large partitions have a greater propensity to lobby.
21
Taken together, the results reported in Table 3 suggest that
our proxies for historical use of accounting slack (H1) and
lower demand for transparency (H2) are robust determi-
nants of bank representatives’ choices to submit comment
letters critical of the FASB’s ED. In contrast, our proxies for
the reasons for opposition most commonly included in com-
ment letters were either not signi?cant or opposite of the
predicted relation.
Robustness
As described in our discussion of model speci?cation,
we use include TIMELINESS (Beatty & Liao, 2011) as our
proxy for banks’ historical use of slack available in current
GAAP for loan losses. To mitigate concerns that our results
are contingent on the speci?c proxy we use for the histor-
ical use of accounting slack, we also conduct our analyses
using three other measures of historical slack usage by
banks: LLCONSERV, LLACCRUAL and STD_ROA. LLCONSERV
is based on the relative size of the allowance for loan losses
controlling for non-performing assets. This is consistent
with Watts’ (2003) notion of conservatism, which he
describes as an intentional downward bias in reporting
the value of net assets. We compute LLCONSERV as the
residual from a regression of ALLOWANCE/LOANS on
non-performing assets (NPA) as of December 31, 2009.
Higher (lower) values of LLCONSERV indicate that banks
are using accounting slack to report relatively higher (low-
er) loan-loss allowances, which means they are reporting
relatively lower (higher) net asset balances for loans after
controlling for the amount of problem assets.
We calculate LLACCRUAL as the (signed) ?rm-speci?c
residuals from a cross-sectional estimation of Eq. (A2)
(speci?ed in the Appendix) over the period 2010Q1
through 2011Q3. We structure this ?rm-speci?c time ser-
ies measure to re?ect short-term earnings management,
which is distinct from our measure of TIMELINESS. LLAC-
CRUAL is estimated in cross-section, and captures the
residual short-term discretion in the provision after con-
trolling for the variables that in?uence TIMELINESS. Firms
with more negative (positive) realizations of LLACCRUAL
relatively underprovided (overprovided) for expected fu-
ture losses.
Our last alternative proxy, STD_ROA, is consistent with
Bushman and Williams’ (2012) ?nding that banks’ use of
loan loss discretion to report smooth earnings reduces
transparency and hinders the ability of outsiders to
21
Consistent with Beatty et al. (2002) we also consider whether a ?rm is
publicly traded or private and ?nd that publicly traded ?rms are less likely
to lobby against the ED. Inferences are unchanged by inclusion of PUBLIC.
However, AUDIT, becomes insigni?cant. AUDIT is correlated with PUBLIC
because all public ?rms must be audited. When the sample is restricted to
private ?rms, AUDIT is signi?cantly negative, and all inferences remain
unchanged.
130 L.D. Hodder, P.E. Hopkins / Accounting, Organizations and Society 39 (2014) 117–133
monitor. We calculate STD_ROA as the standard deviation
of quarterly net income divided by average assets for the
quarter, computed over the period 2001–2011.
In untabulated regression analyses, we ?nd that each of
these alternative proxies loads signi?cantly (and in the
predicted direction) when they are substituted for TIMELI-
NESS in equation (1) (lowest t statistic = 7.24, p < .001).
In addition, when all four proxies are simultaneously
included in our regression analysis, only LLACCRUAL
becomes statistically insigni?cant, with all three agency-
related proxies retaining statistical signi?cance. Taken
together, these supplemental analyses suggest that TIME-
LINESS is a robust proxy for banks’ historical use of
accounting slack in current GAAP for loan losses, and that
the inferences from our hypotheses are robust to
alternative measures of accounting slack.
Summary and conclusions
In 2010, the ABA and state banking associations
engaged in an intensive campaign to convince representa-
tives from member banks to lobby the FASB against the
ED’s proposal to measure and report loans at fair value
(Ciesielski, 2010, p. 1). Despite the fact that the FAF
(1977) states that ‘‘it would not be appropriate to establish
a standard based solely on a canvass of the constituents’’
(p. 19), it appears that the FASB decided to scuttle fair
value recognition for loans because of the ‘‘overwhelm-
ingly negative reaction to its proposal from companies
and investors’’ (Rapoport, 2011). Given the importance of
due process and the in?uence of comment letters, we seek
to better understand why representatives from 1047
unique public and private commercial banks submitted
comment letters opposing the FASB’s proposal, while rep-
resentatives from 4242 commercial banks (for which we
have suf?cient data to conduct our analyses) chose to
remain silent.
We provide evidence suggesting that bank representa-
tives’ objections to the ED are motivated less by stated con-
ceptual concerns and more by self-interest. The incurred
loss model for accounting for loan losses makes available
accounting slack that can be used to manage capital and
earnings. Our investigation reveals that representatives
lobbying the FASB against fair value measurement for
loans are more likely to be af?liated with a bank that his-
torically has used accounting slack inherent in the incurred
loss model to less-timely record loan losses. One objection
to fair values is that their use may increase measurement
discretion. Our ?nding that banks historically using
accounting slack are more likely to oppose the ED suggests
that the FASB’s fair value proposal is perceived to decrease
available accounting slack related to loans.
The fact that the ED retains amortized cost reporting on
the face of the ?nancial statements, and also requires the
credit-loss component of fair value changes to be esti-
mated and charged against net income, suggests that
?nancial statement users and regulators will have more
information under the proposal. Given that the demand
for transparent information is increasing in agency costs,
bank representatives obtaining more (unobservable)
private bene?ts relative to agency costs are more likely
to resist a movement to more transparent accounting stan-
dards. In our analysis of transparency-related motives, we
?nd that banks submitting comment letters opposing the
loan-reporting provisions of the ED are (1) more likely to
elect to participate in the new optional, supplemental,
?xed-price deposit insurance and debt guarantee programs
offered by the FDIC, (2) less likely to have nonguaranteed
outside creditors and (3) less likely to obtain ?nancial
statement audits of their non-regulatory GAAP ?nancial
statements. These results are consistent with bank manag-
ers and shareholders af?liated with lobbying banks bene-
?tting from the lower agency costs that accompany
government guarantees of indebtedness and the lower lev-
els of external monitoring that otherwise would be pro-
vided by unguaranteed debt holders and by auditors.
Our results are inconsistent with bankers’ arguments
that their opposition to fair value measurement is driven
primarily by concerns about costs, decreased transparency,
measurement dif?culty, or a mismatch with their existing
business model. Similarly sized banks using accounting
slack to delay the recognition of loan losses are unlikely
to face systematically higher implementation costs than
those not using such slack. Further, banks using accounting
slack (1) to manage earnings, (2) to less-timely recognize
anticipated non-performing assets, and (3) to present rela-
tively higher net asset balances than banks with similar as-
set quality, are also unlikely to be concerned with
potentially diminished transparency or lower reliability
associated with fair value estimation. Moreover, we ?nd
no evidence that representatives from banks with more-
dif?cult-to-value loans are more likely to lobby against fair
value measurement, and our results suggest that represen-
tatives from banks that use loan fair values for internal
decision-making are actually more likely to lobby against
fair value measurement.
Conceptually, fair value recognition should increase
transparency (Linsmeier, 2011). For example, Blankespoor,
Linsmeier, Petroni and Shakespeare (2013) ?nd that fair
value measurement of ?nancial instruments better de-
scribes banks’ credit risk. For public banks disclosing fair
values annually, Barth, Beaver and Landsman (1996) ?nd
that loan fair values are value-relevant. Further, Hodder,
Hopkins, and Wahlen (2006) suggest that income volatility
measured under full fair value accounting is signi?cantly
more informative about banks’ risks—and better re?ects
the risk priced in banks share prices and expected re-
turns—than volatility measured under the current US
reporting system. As compared to disclosure, ?nancial
statement recognition may increase the precision of fair
value estimates, leading to higher ?nancial statement
quality (Libby, Nelson, & Hunton, 2006).
Our ?ndings contribute to the continuing debates re-
lated to fair value measurement of ?nancial instruments,
the incurred-loss model for loan losses, RAP-GAAP confor-
mity and the impact of regulators on the promulgation of
general-purpose ?nancial accounting standards. Our re-
sults should also be considered in the context of due pro-
cess for setting accounting standards and for legislating
?nancial regulation. Our analysis reveals (1) that a nontriv-
ial number of individuals identi?ed as investors were, in
L.D. Hodder, P.E. Hopkins / Accounting, Organizations and Society 39 (2014) 117–133 131
reality, bank-af?liated individuals who concealed their
af?liation, (2) that the negative comment letters were con-
centrated among bank representatives, and (3) that com-
ment letters were submitted by representatives from
banks that had a lower demand for transparency and sys-
tematically utilized the accounting slack available in cur-
rent GAAP for loan losses.
To the extent comment letter submissions continue to
be part of regulatory and accounting-standard-setting
due process, regulators and standard setters may wish to
consider ways to effectively identify all letter writers and
their af?liations, and to explicitly consider the incentives
of all participants in the rule making process. Dispropor-
tionate preparer response can result in disproportionate
weighting of preparer concerns relative to potential public
bene?ts. Indeed, standard setters appear to be in?uenced
by comment letters, and research ?nds that standard set-
ters, when explaining standard setting results, are more
likely to mention preparer groups as exerting in?uence
through the comment letter process and other forms of
lobbying (Hope & Briggs 1982; Hope & Gray, 1982; Jupe,
2000; Ang, Gallery, & Sidhu, 2000). In the context of the
ED, some preparers represented themselves as users,
rather than preparers, perhaps to appear less self-serving
in their motivations. Our results suggest that dispropor-
tionate weighting of preparer feedback on proposed
accounting standards may have unintended consequences
with respect to public markets: successful lobbying that
results in less transparency may facilitate managers’ reten-
tion of private bene?ts while increasing public costs.
News accounts suggest that the FASB decided to drop
the ED because of the negative reaction to fair value mea-
surement for loans (Rapoport 2011).
22
This episode raises
important questions about the extent to which feedback
from vested interests should in?uence the standard setting
process. The Financial Accounting Foundation (FAF) notes
that standard setters must exercise judgment and consider
constituents’ comments in context: ‘‘(b)ecause standard set-
ting requires some perspective, it would not be appropriate
to establish a standard based solely on a canvass of the con-
stituents’’ (FAF, 1977, p. 19).
Some argue that accounting should be protected from
politically motivated or special interest lobbying (e.g. Ger-
both 1973; Horngren 1973). Because conceptually sound
accounting standards are likely to affect different compa-
nies in different ways, or favor one constituent group over
another, standard setting inherently involves decisions
that affect relative social welfare. As Solomons (1978)
notes, ‘‘[n]eutrality, in the sense in which the term is used
here does not imply that no one gets hurt’’ (p. 70). He sug-
gests that reactions to proposed accounting standards
must be considered in context with ultimate decisions
based on conceptual and social merits. The results of our
inquiry support this view.
Appendix A. Supplementary material
Supplementary data associated with this article can be
found, in the online version, athttp://dx.doi.org/10.1016/
j.aos.2013.10.002.
References
American Bankers Association (2009). Comment letter addressed to
members of the house ?nancial services committee concerning support
the Perlmutter-Lucas revised accounting amendment to H.R. 3996.
(November 18).
American Institute of Certi?ed Public Accountants (2008). Comment letter
Ms. Tamara Oyre concerning the review of the International Accounting
Standards Committee Foundation’s, Constitution (September 15).
Ang, N., Gallery, N., & Sidhu, B. K. (2000). The incentives of Australian
public companies lobbying against proposed superannuation
accounting standards. Abacus, 36(1), 40–70.
Barth, M., Beaver, W., & Landsman, W. (1996). Value relevance of banks’
fair value disclosures under SFAS No. 107. The Accounting Review, 71,
513–537.
Beatty, A., Chamberlain, S. L., & Magliolo, J. (1995). Managing ?nancial
reports of commercial banks: The in?uence of taxes, regulatory
capital, and earnings. Journal of Accounting Research, 33, 231–261.
Beatty, A. L., Ke, B., & Petroni, K. R. (2002). Earnings management to avoid
earnings declines across publicly and privately held banks. The
Accounting Review, 77, 547–570.
Beatty, A., & Liao, S. (2011). Do delays in expected loss recognition affect
banks’ willingness to lend? Journal of Accounting and Economics, 52,
1–20.
Blankespoor, E., Linsmeier, T., Petroni, K., & Shakespeare, C. (2013). Fair
value accounting for ?nancial instruments: Does it improve the
association between bank leverage and credit risk?. The Accounting
Review.http://dx.doi.org/10.2308/accr-50419 (forthcoming).
Bushman, R. M., & Williams, C. D. (2013). Delayed expected loss recognition
and the risk pro?le of banks (February 28). .
Bushman, R. M., & Williams, C. D. (2012). Accounting discretion, loan loss
provisioning, and discipline of Banks’ risk-taking. Journal of
Accounting and Economics, 54(1), 1–18.
Ciesielski, J. T. (2010). It’s déjà vu all over again: The fair value accounting
wars. The Analyst’s Accounting Observer, 19 (August 16).
Collins, J. T., Shackelford, D. A., & Wahlen, J. M. (1995). Bank differences in
the coordination of regulatory capital, earnings, and taxes. Journal of
Accounting Research, 33, 263–291.
Dechow, P. M., Hutton, A. P., & Sloan, R. G. (1996). Economic consequences
of accounting for stock-based compensation. Journal of Accounting
Research, 34(Suppl.), 1–20.
Dhaliwal, D. (1982). Some economic determinants of management
lobbying for alternative methods of accounting: Evidence from the
accounting for interest costs issue. Journal of Business Finance and
Accounting, 9, 255–265.
Dugan, J. C. (2009). The need to preserve uniform standards for national
banks. Speech delivered on September 24, 2009. .
Financial Accounting Foundation (1977). The structure of establishing
?nancial accounting standards. Report of the Structure Committee of
the FAF: Norwalk, CT.
Financial Accounting Standards Board (2009). Minutes of the March 18,
2009 Board meeting—Loan loss disclosures (April 28).
Financial Accounting Standards Board (2010). Accounting for ?nancial
instruments and revisions to the accounting for derivative instruments
and hedging activities. Exposure draft of proposed accounting
standards update of topic 825 and topic 815 (May 26).
Financial Crisis Advisory Group (2009). Report of the Financial Crisis
Advisory Group (July 28). .
Forssbaeck, J. (2011). Ownership structure, market discipline, and banks
risk-taking incentives under deposit insurance. Journal of Banking &
Finance, 35(10), 2666.
Gerboth, D. (1973). Research, intuition and politics in accounting theory.
Accounting Review, 48(3), 475–482.
22
The comment period for the ED ended on September 30, 2010. The
FASB announced, in January 2011, that it would not consider measuring
and reporting loans at fair value as part of a ?nancial instruments standard.
In interviews, FASB Chair Leslie Seidman characterized these actions as a
response to feedback on the proposal (Rapoport, 2011). In April 2013, the
FASB issued a new Exposure Draft, Financial Instruments—Overall (Subtopic
825-10): Recognition and Measurement of Financial Assets and Financial
Liabilities—Proposed Amendments to the FASB Accounting Standards Codi?ca-
tion
Ò
, that largely excludes loans from fair value measurement.
132 L.D. Hodder, P.E. Hopkins / Accounting, Organizations and Society 39 (2014) 117–133
Hill, M. D., Kelly, G. W., Lockhart, G. B., & Van Ness, R. A. (2011).
Determinants and effects of corporate lobbying (January 19). or .
Hochberg, Y. V., Sapienza, P., & Vissing-Jørgensen, A. (2009). A lobbying
approach to evaluating the Sarbanes-Oxley Act of 2002. Journal of
Accounting Research, 47, 519–583.
Hodder, L. D., Hopkins, P. E., & Wahlen, J. M. (2006). Risk-relevance of fair-
value income measures for commercial banks. The Accounting Review,
81, 337–375.
Hodder, L., Kohlbeck, M., & McAnally, M. L. (2002). Accounting choices
and risk management: SFAS No. 115 and U.S. bank holding
companies. Contemporary Accounting Research, 19, 225–270.
Hope, T., & Briggs, J. (1982). Accounting policy making – Some lessons
from deferred taxation debate. Accounting and Business Research,
12(46), 83–96.
Hope, T., & Gray, R. (1982). Power and policy making: the development of
an R&D standard. Journal of Business Finance and Accounting, 9(4),
531–558.
Horngren, C. (1973). The marketing of accounting standards. Journal of
Accountancy, 137(4), 61–66.
Jensen, M. C., & Meckling, W. H. (1976). Theory of the ?rm: Managerial
behavior, agency costs and ownership structure. Journal of Financial
Economics, 3, 305–360.
Johnson, S. (2009). Keep bank regulators way from FASB: SEC Commissioner.
CFO.com (November 17). .
Jupe, R. (2000). Self-referential lobbing of accounting standards board:
The case of Financial Reporting Standard No. 1. Critical Perspectives on
Accounting, 11(3), 337–359.
Kane, E. J. (1989). How incentive-incompatible deposit-insurance funds fail.
National Bureau of Economic Research Working Paper No. 2836.
.
Kelly, L. (1985). Corporate management lobbying on FAS No. 8: Some
further evidence. Journal of Accounting Research, 23, 619–632.
Libby, R., Nelson, M., & Hunton, J. (2006). Recognition v. disclosure:
Auditor tolerance for misstatement, and the reliability of stock-
compensation and lease information. Journal of Accounting Research,
44, 533–560.
Linsmeier, T. J. (2011). Financial reporting and ?nancial crises: The case
for measuring ?nancial instruments at fair value in the ?nancial
statements. Accounting Horizons, 25, 409–417.
Liu, C., & Ryan, S. G. (2006). Income smoothing over the business cycle:
Changes in banks’ coordinated management of provisions for loan
losses and loan charge-offs from the pre-1990 bust to the 1990s
boom. The Accounting Review, 81, 421–441.
Moyer, S. (1990). Capital adequacy ratio regulations and accounting
choices in commercial banks. Journal of Accounting and Economics, 13,
123–154.
Nichols, D. C., Wahlen, J. M., & Wieland, M. M. (2009). Publicly-traded vs.
privately-held: implications for bank pro?tability, growth risk, and
accounting conservatism. Review of Accounting Studies, 14, 88–122.
Olson, M. (1965). The logic of collective action. Cambridge, MA: Harvard
University Press.
Rapoport, M. (2011). Banks force retreat on fair-value plan. The Wall Street
Journal (January 26), C2.
Ryan, S. G. (2007). Financial instruments and institutions: Accounting and
disclosure rules (2nd ed.). New York: John Wiley & Sons.
Ryan, C., Dunstan, K., & Stanley, T. (2000). Local government accounting
standard-setting in Australia: Did constituents participate? Financial
Accountability and Management, 16, 373–396.
Schulz, S. M., & Hollister, J. (2003). Lobbying FASB on accounting for
investments. The Journal of Applied Business Research, 19, 85–93.
Solomons, D. (1978). The politicization of accounting. Journal of
Accountancy, 146(5), 65–72.
Sutton, T. G. (1984). Lobbying of accounting standard-setting bodies in
the UK and the USA: A Downsian analysis. Accounting, Organizations
and Society, 9, 81–95.
Tschirhart, J., O’Brien, J., Moise, M., & Yang, E. (2007). Bank commercial loan
fair value practices. Washington, DC: Federal Reserve Board. .
Tutticci, I., Dunstan, K., & Holmes, S. (1994). Respondent lobbying in the
Australian Accounting standard-setting process: ED49. Accounting,
Auditing & Accountability Journal, 7, 86–104.
Watts, R. L. (2003). Conservatism in accounting Part I: Explanations and
implications. Accounting Horizons, 17, 207–221.
Watts, R. L., & Zimmerman, J. L. (1978). Towards a positive theory of the
determination of accounting standards. The Accounting Review, 53,
112–134.
Weetman, P., Davie, E. S., & Collins, W. (1996). Lobbying on accounting
issues: Preparer/user imbalance in the case of the operating and
?nancial review? Accounting, Auditing & Accountability Journal, 9,
59–76.
L.D. Hodder, P.E. Hopkins / Accounting, Organizations and Society 39 (2014) 117–133 133
doc_213708268.pdf
We investigate the determinants of bank representatives’ responses to the United States
Financial Accounting Standard Board’s 2010 Exposure Draft that proposes fair value measurement
for most financial instruments. Over 85% of the 2971 comment letters were
received from bank representatives, with most bank-affiliated letters addressing—and
opposing—one issue: fair value measurement of loans. The Exposure Draft proposes that
companies report both fair value and amortized cost measures for loans; thus, the proposal
should result in increased levels of loan-related information and improved financial reporting
transparency. We investigate three reasons for bank representatives’ resistance. First,
fair value measurement should result in less accounting slack than the current incurredloss
model for loan impairments; therefore, we propose that representatives from banks
that historically utilized that slack will resist fair value measurement for loans. Second,
we propose that agency problems are an important motivating factor because bank representatives
reaping more private benefits from their franchises have less incentive to support
increases in financial reporting transparency
Agency problems, accounting slack, and banks’ response
to proposed reporting of loan fair values
q
Leslie D. Hodder
?
, Patrick E. Hopkins
Kelley School of Business, Indiana University, Bloomington, IN 47405-1701, USA
a b s t r a c t
We investigate the determinants of bank representatives’ responses to the United States
Financial Accounting Standard Board’s 2010 Exposure Draft that proposes fair value mea-
surement for most ?nancial instruments. Over 85% of the 2971 comment letters were
received from bank representatives, with most bank-af?liated letters addressing—and
opposing—one issue: fair value measurement of loans. The Exposure Draft proposes that
companies report both fair value and amortized cost measures for loans; thus, the proposal
should result in increased levels of loan-related information and improved ?nancial report-
ing transparency. We investigate three reasons for bank representatives’ resistance. First,
fair value measurement should result in less accounting slack than the current incurred-
loss model for loan impairments; therefore, we propose that representatives from banks
that historically utilized that slack will resist fair value measurement for loans. Second,
we propose that agency problems are an important motivating factor because bank repre-
sentatives reaping more private bene?ts from their franchises have less incentive to sup-
port increases in ?nancial reporting transparency. Third, we test whether the most
common reasons for opposition included in the comment letters are associated with neg-
ative letter writing. Our analyses support the ?rst two determinants of bank representa-
tives’ resistance to the Exposure Draft. Speci?cally, accounting slack and lower demand
for accounting transparency are strongly associated with resistance to the standard. How-
ever, we ?nd that stated reasons for resistance are not associated with letter writing. Spe-
ci?cally, representatives at ?rms with dif?cult to value loans and ?rms that mostly hold
loans to maturity are no more likely to resist the standard than others. The narrow scope
of bank representatives’ comments and our empirical ?ndings suggest that bankers’
responses to the Exposure Draft may be more driven by concerns over reduced availability
of accounting slack and accompanying de facto regulatory forbearance than by the concep-
tual arguments they offer. Our results have implications for standard setters, who must
navigate special interests as they attempt to promulgate high quality accounting standards,
and for users of ?nancial statements who must consider how political forces shape gener-
ally accepted accounting principles.
Ó 2014 Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.aos.2013.10.002
0361-3682/Ó 2014 Elsevier Ltd. All rights reserved.
q
We thank Rob Bloom?eld, Mark Evans, Patrick Finnegan, Jim Leisenring, Tom Linsmeier, Chris Yust, an anonymous reviewer, workshop participants
from the University of Florida, Georgetown University, Michigan State University, the University of Toronto, and the University of Texas Centennial Research
Conference for their helpful suggestions, and Connor Egan, Mitch Harrison and Brett Jerasa for assisting with coding the comment letters. Professor Hodder
thanks EY, Professor Hopkins thanks the Deloitte Foundation, and both authors thank the Kelley School of Business for their ?nancial support.
?
Corresponding author. Tel.: +1 812 855 9951.
E-mail address: [email protected] (L.D. Hodder).
Accounting, Organizations and Society 39 (2014) 117–133
Contents lists available at ScienceDirect
Accounting, Organizations and Society
j our nal homepage: www. el sevi er. com/ l ocat e/ aos
Introduction
In May 2010, the United States (US) Financial Account-
ing Standards Board (FASB) issued an Exposure Draft (ED)
that proposes greatly expanding fair value recognition for
most ?nancial instruments, including long-term receiv-
ables, such as bank loans (FASB, 2010). Responses received
by the FASB during the ED’s comment period were over-
whelmingly negative and particularly concentrated within
the banking industry; speci?cally, the FASB received 2971
comment letters in response to the ED, with over 85% from
bank representatives and banking trade organizations.
Through the end of 2011, this is one of the highest com-
ment-letter volumes in response to any single FASB-issued
public-comment discussion document. Because of the
large volume of negative letters received from bank repre-
sentatives, the FASB withdrew the proposal in January
2011 (Rapoport, 2011). This study seeks to understand
the factors systematically associated with bank represen-
tatives’ decisions to submit comment letters.
The choice to submit an accounting-standards-related
comment letter (and the speci?c issues discussed in the
letter) should be a function of the perceived economic con-
sequences of the proposed accounting changes (Watts &
Zimmerman, 1978). Despite the many changes proposed
in the ED, the vast majority of letters submitted by com-
mercial-bank representatives addressed only one issue:
opposition to the proposed reporting of loans at fair value.
1
Given the large, uniform, and narrowly focused response
that was concentrated in the banking industry, this suggests
that responding bank representatives perceived an eco-
nomic threat from the FASB’s ED and its proposal to change
current accounting for loans.
Under current US generally accepted accounting princi-
ples (GAAP), companies report loans at amortized cost, ad-
justed for loan impairments determined via an incurred-
loss measurement model. Standard setters, regulators
and practitioners note substantial disagreement about fac-
tors that bank managers should consider in determining
loan-loss provisions under this model (Dugan, 2009; FASB,
2009). In addition, empirical research suggests that banks
use discretion available in loan loss reporting to opportu-
nistically manage income (e.g., Liu & Ryan, 2006) and cap-
ital (e.g., Beatty & Liao, 2011), and that this discretion
reduces transparency and diminishes the ability of outsid-
ers and regulators to monitor banks’ risk-taking behavior
(Bushman & Williams, 2012). Taken together, the practitio-
ners’ criticisms and empirical research ?ndings suggest
that the incurred-loss model for loans is a source of
accounting slack available to bank managers.
The accounting proposed in the ED should result in a
greater amount of relevant information about loans
because the ED requires loans to be reported at fair value
while retaining much of the information currently re-
ported for loans.
2
Speci?cally, the ED requires (1) that amor-
tized cost information for loans be presented on the face of
the statement of ?nancial position (parenthetically, and rec-
onciled to fair value) (FASB, 2010, para. 86), and (2) that esti-
mated credit losses be charged against net income with any
remaining change in the fair value of loans (i.e., that is not
related to counterparty credit quality) recognized in other
comprehensive income (i.e., outside of reported net income)
(FASB, 2010, para. 91).
3
Because only estimated credit losses
will be recognized in net income, and because currently rec-
ognized amortized-cost information will continue to be
prominently displayed on the face of the balance sheet,
the proposed changes should increase the amount of loan-
related information available to outside investors and cred-
itors while retaining key measurement principles embedded
in current accounting standards.
4
However, the changes
should also reduce the amount of accounting slack available
to bank managers and improve the transparency of banks’
loan reporting because loans will be recognized at fair value,
which, by construction, should re?ect more timely recogni-
tion of expected future losses and market-related opportu-
nity costs (Financial Crisis Advisory Group, 2009, p. 7).
Given the proposal’s potential informational bene?ts,
we investigate why representatives from more than 1000
unique commercial banks submitted comment letters
resisting only the loan-related provisions of the ED, while
the vast majority of commercial bank representatives did
not submit letters. We investigate three signi?cant sources
of incentives for resistance to the proposed fair value mea-
surement of loans. First, because fair value measurement of
loans will likely reduce the accounting slack available to
bank managers (i.e., when compared to the current in-
curred-loss model), representatives from banks that have
a history of exploiting the slack available under current
GAAP will have higher incentives to resist a change to fair
1
The FASB’s ED proposes a comprehensive set of changes related to fair
value recognition for most ?nancial instruments (including liabilities),
equity method accounting, derivatives, and hedging. The ED also proposes
that companies continue to report balance sheet and income statement
metrics using amortized-cost basis-measurement. Re?ecting its wide-
ranging scope, the ED contains 214-pages and requests that comment
letters potentially address 71 separate questions (i.e., 36 questions for all
respondents, 13 questions for preparers and auditors, 22 questions for
users).
2
Although current US GAAP requires banks to provide disclosure of loan
fair values in the notes to the ?nancial statements (i.e. FASB Accounting
Standards Codi?cation (ASC) 825-10-55-3), many small- and medium-sized
US banks avail themselves of the ‘‘practicability exception’’ that allows
them to disclose loan value amounts based on simpli?ed cash ?ow models
(i.e. FASB ASC 825-10-50-16 through -19). These reported values often
capture measurement attributes other than fair value (e.g., entry price). As
noted by Tschirhart, O’Brien, Moise, and Yang (2007), these valuation
methodologies need to be substantially modi?ed to arrive at measurements
that meet the de?nition of fair value.
3
Thus, the ED cannot be characterized as a ‘‘fair value versus historical
cost’’ proposal because it retains amortized cost measurement on the face
of the ?nancial statements, while complementing that information with
reconciliation between amortized cost and fair value bases. For these
reasons our analysis cannot address whether either amortized cost
measurement or fair value measurement provides better information in
the absence of the other. Instead, in this study we propose that ?nancial
statement recognition of accounts measured at fair value with supplemen-
tal prominent display of amortized cost information can improve trans-
parency by revealing additional information that is not apparent in
?nancial statements measured solely under amortized cost.
4
Any effect of the ED on net income would necessarily derive from
differences between managers’ estimates of incurred losses and expected
losses arising from non-collection of contractual cash ?ows. The incurred
loss model impounds less information, and is arguably less conservative
than the expected loss model. These terms are discussed in more detail in
the next section.
118 L.D. Hodder, P.E. Hopkins / Accounting, Organizations and Society 39 (2014) 117–133
value measurement for loans and will be more likely to
submit negative comment letters.
5
Second, we propose that
agency problems are an important motivating factor for
responding bank representatives because the incentives for
transparent reporting are decreasing in agency problems
and increasing in agency costs (Jensen & Meckling, 1976);
thus, bankers reaping more private bene?ts from their bank-
ing franchises have incentives to resist increases in ?nancial
reporting transparency.
6
Third, we investigate whether the
likelihood of banks submitting negative comment letters is
associated with the most-commonly cited reasons included
in those letters (i.e., measurement dif?culty and incongru-
ence with banks’ business models).
We use logistic regression analysis to test the determi-
nants of differential bank comment-letter writing. We col-
lect all 2971 comment-period letters submitted to the
FASB in response to the ED, read each of the letters, and
identify those written by commercial-bank representa-
tives. We identify 2472 ?nancial-institution-af?liated let-
ters, which, after adjusting for non-commercial-bank
submissions (e.g., credit unions and savings and loans),
multiple submissions per bank, and data availability, com-
prises 1047 unique commercial banks. This sample repre-
sents approximately 20% of US commercial banks that
have suf?cient ?nancial data available to conduct our anal-
yses (i.e., 1047 out of 5289 banks). Because the majority of
banks are not publicly traded, we collect much of our data
from regulatory ?lings and disclosures.
Our results suggest that bank representatives’ docu-
mented resistance to the loan-accounting provisions in the
ED is positively associated with (1) the historical use of the
accounting slack available through the current incurred-
loss method of accounting for loan losses and (2) proxies
for lower demand for transparency in the responding banks.
Our proxies for reduced demand for reporting transparency
include the choice to purchase additional, temporary, ?at-
rate, excess deposit insurance offeredby the Federal Deposit
Insurance Corporation (FDIC), the choice to acquire unin-
sured debt ?nancing, and the choice to forego independent
audits. The choice to purchase insurance re?ects a trade-
off between the ?xed rate premiums and bene?ts gained
by reducing investor uncertainty. Because we control for
?nancial condition, we interpret the choice to purchase
insurance as a substitute for ?nancial reporting transpar-
ency. The positive association between excess deposit-
insurance purchase and letter writing is consistent with
the ED’s opposition being driven by representatives of less
transparent ?rms. Consistent with this explanation, the
presence of uninsured debt and audited ?nancial
statements were also negatively associated with the likeli-
hood of bank representatives opposing the standard.
Interestingly, we ?nd no evidence that the factors most
commonly asserted by bank representatives are associated
with increased propensity to lobby against fair value mea-
surement for loans. Speci?cally, banks with more dif?cult-
to-value commercial loans appear no more likely to oppose
fair value reporting for loans than those with easier-to-va-
lue loans. In addition, opposite of the relation suggested by
the letters, those banks using fair value for internal deci-
sion-making purposes (i.e., as evidenced by more frequent
loans sales) are signi?cantly more likely to oppose fair va-
lue reporting for loans. Taken together, our ?ndings sug-
gest that preparer comment letters were more in?uenced
by concerns about restriction of accounting slack and by
agency-related manager demand for reporting opacity
than the concerns articulated in the comment letters.
Our results should be considered in the context of due
process for setting accounting standards and for legislating
?nancial regulation. Our study suggests that bank repre-
sentatives writing negative comment letters have motives
to resist accounting standards that result in reduced
accounting slack and increased transparency. As one would
expect, these motives and patterns of ?nancial-reporting
behavior are not mentioned in the comment letters. In-
stead, the negative letters submitted by responding bank
representatives cite conceptual reasons for resisting pro-
posals intended to increase ?nancial reporting transpar-
ency (e.g., insuf?cient precision for fair value
measurements, and irrelevancy of fair values for deci-
sion-making). Our results suggest that the processing of
comment-letters should explicitly include consideration
of letter-writer motivations, and weigh carefully the stated
reasons for support or opposition against letter-writer
incentives and potentially divergent facts.
Issues surrounding due process are complicated, in
practice, by the fact that feedback on proposed accounting
standards is less likely to be received from some constitu-
ent groups. For example, existing research shows that
?nancial statement preparers are much more active in lob-
bying than users (Weetman, Davie, & Collins, 1996; Ryan,
Dunstan, & Stanley, 2000; Sutton, 1984; Tutticci, Dunstan,
& Holmes, 1994). However, users’ lack of lobbying activity
is predicted by theory and may be dif?cult to overcome. To
the extent that high quality ?nancial reporting results in
societal rather than private bene?ts, the relative lack of
user lobbying occurs because private individuals are unli-
kely to voluntarily incur costs to promote public goods (Ol-
son, 1965; Sutton, 1984). Focus groups and individual
interviews may be more effective in eliciting users’ assess-
ment of proposed standards (Weetman et al., 1996).
We infer that regulatory concerns also appear to be an
important determinant of the banking response to the
2010 ED because of the heavy concentration of letter writ-
ing by nonpublic commercial-bank representatives. Many
of these banks are not required to issue general purpose
?nancial statements and are not required to obtain ?nan-
cial statement audits. Beatty, Ke, and Petroni (2002) argue
that private US banks have fewer incentives to manage re-
ported ?nancial information because ownership concen-
trations among manager-shareholders reduce agency
5
We de?ne accounting slack as the acceptability of multiple alternatives
within-GAAP accounting methods and/or numerical estimates for a given
economic transaction, event, right or obligation.
6
According to Jensen and Meckling (1976), ceteris paribus, managers
have a higher likelihood of private bene?t extraction when agency
problems are high and agency costs are low. Agency problems describe
the general condition where an agent with delegated authority has the
opportunity to consume perquisites without the approval of the principal.
Agency costs imposed on the agent mitigate the likelihood that he will
consume perquisites. Thus, the demand for transparency decreases when
(1) the value of perquisite consumption increases or (2) when agency costs
decrease.
L.D. Hodder, P.E. Hopkins / Accounting, Organizations and Society 39 (2014) 117–133 119
costs of equity. Therefore, private-bank representatives’
resistance to the reduction of available accounting slack
likely is driven by the reliance of US government regulators
on US GAAP ?nancial statements.
7
Our ?nding that regula-
tory concerns play a prominent role in bankers’ resistance to
slack-reducing transparency complements the ?ndings of
other studies to suggest that the marriage of Regulatory
Accounting Principles (RAP) and Generally Accepted
Accounting Principles (GAAP) can lead to value-decreasing
behavior by banks (e.g., Hodder, Kohlbeck, & McAnally,
2002), and can decrease the effectiveness of regulatory over-
sight (Bushman & Williams, 2012).
Our study also contributes to the literature on manage-
ments’ reporting incentives and the decisions to lobby for
or against various accounting standards (e.g., Watts & Zim-
merman, 1978). These studies usually are based on the as-
sumed effects that a proposed rule will have on the
primary ?nancial statements of public corporations. In par-
ticular, studies predict increases in lobbying behavior
when proposed standards are anticipated to reduce com-
panies’ levels of income. For example public ?rms resist
expensing the fair value of stock options because doing
so decreases reported net income (Dechow, Hutton, &
Sloan, 1996), and resist including foreign exchange transla-
tion adjustments in net income because these adjustments
increase net income volatility (Kelly, 1985). Our investiga-
tion extends these studies by including a large sample of
public and private companies, and by analyzing lobbying
in the context of agency problems within a regulated
industry where private bene?t extraction is a concern.
Our study is related to Hochberg, Sapienza, and Vissing-
Jørgensen (2009), who show that US public companies
with greater agency problems lobbied more intensely
against Sarbanes–Oxley Act of 2002. Our study is distinct
from Hochberg et al. (2009) because we address the inter-
section of accounting slack, agency problems, ?nancial
reporting standards, and regulatory accounting.
We organize the paper as follows: In Current and
proposed accounting for loans, we discuss current and pro-
posed accounting for loans; in Hypotheses, we propose our
hypotheses; in Sample and research design, we describe
our sample and research design; in Results, we discuss
our research ?ndings; and we provide concluding remarks
in Summary and conclusions.
Current and proposed accounting for loans
Accounting slack inherent in current GAAP for loan losses
Current US GAAP for loan losses is based on an incurred
loss model of accounting for counterparty credit risk.
8
Losses are recognized when probable, based on past events
and economic conditions existing as of the reporting date.
Even in the absence of subjectivity, incurred losses differ
from losses embedded in expected values because expected
values take into account unbiased expectations about future
economic conditions that would give rise to future incurred
losses. Moreover, unlike continuous measurement of losses
via expected values, the incurred loss method in current US
GAAP results in recognition of losses only when the probabil-
ity of loss exceeds the ‘‘probable’’ threshold. For these rea-
sons, unbiased loss provisions recognized using the
incurred loss method will generally result in cumulative loan
loss provisions that are less than cumulative expected losses
re?ected in unbiased expected values or unbiased fair values.
The incurred loss method also appears to have wide-
spread variation in its application (Dugan, 2009; FASB,
2009). At one extreme, required reliance on objective evi-
dence of impairment leads some to argue that losses
should not be recognized until loans become delinquent
or otherwise default. At the other extreme, risk-averse reg-
ulatory ?eld auditors and conservative managers may
cause a bank’s loan loss allowance to converge to or exceed
an unbiased estimate of expected losses. This divergence
causes great variability in the extent to which loan loss
provisions recognized under current GAAP are associated
with nonperforming loans and realized losses in subse-
quent periods (e.g., Beatty & Liao, 2011; Nichols, Wahlen,
& Wieland, 2009; Bushman & Williams, 2013).
In addition to potentially non-strategically divergent
implementation of the existing loan-loss standard, slack
associated with the incurred loss model is actively used to
manage both earnings and capital (Moyer, 1990; Collins,
Shackelford, & Wahlen, 1995; Beatty, Chamberlain, & Magl-
iolo, 1995). Although research generally is mixed on the
consequences of earnings management, evidence on loan-
loss-provision-related earnings and capital management
in the banking industry suggests negative consequences.
For example, Bushman and Williams (2012) ?nd that using
the loan loss provision to smooth earnings reduces trans-
parency, making more dif?cult ef?cient regulatory moni-
toring and undermining market discipline that would
restrain investor losses due to agency problems. In addi-
tion, Beatty and Liao (2011) ?nd that banks delaying loss
recognition suffer from an incremental decrease in lending
capacity during economic downturns that exacerbates the
business cycle. Taken as a whole, prior research suggests
that the incurred loss model of loan-loss recognition under
current GAAP is an important source of accounting slack
available to bank managers and that the use of such slack
can be socially and economically detrimental.
FASB’s 2010 ?nancial instruments exposure draft
In May 2010, the FASB issued an ED for a Proposed
Accounting Standards Update titled, Accounting for Financial
Instruments and Revisions to the Accounting for Derivative
Instruments and Hedging Activities (FASB, 2010). This proposal
includes a comprehensive set of changes relatedto classi?ca-
tion and measurement of ?nancial instruments, impairment
of ?nancial instruments, and hedge accounting. Of the many
changes proposed in the ED, the most controversial—at least
according to responding banks—is the FASB’s proposal that
most ?nancial instruments, including loans, be measured
7
As we describe more fully in the next section, US federal law requires
banking regulators either to accept ?nancial information prepared in
conformity with GAAP or on a basis of accounting that is ‘‘no less stringent’’
than GAAP.
8
Under FASB ASC 825-10, companies have an option to measure most
?nancial instruments (including loans) at fair value. The vast majority of
commercial banks do not exercise this option.
120 L.D. Hodder, P.E. Hopkins / Accounting, Organizations and Society 39 (2014) 117–133
and reported on the balance sheet at fair value. Loans com-
prise thesinglemost signi?cant class of assets oncommercial
banks’ balance sheets, andbyconstruction, recognitionat fair
value should re?ect more timely recognition of expected fu-
ture losses and market-related opportunity costs (Financial
Crisis Advisory Group, 2009, p. 7).
9
Although the ED proposes that most ?nancial instru-
ments should be carried at fair value with the periodic
change in fair value reported in net income (FV-NI), the
ED does allow for a subset of debt instruments held as as-
sets, including loans, to have the non-credit portion of the
change in fair value recognized in other comprehensive in-
come (FV-OCI). Consistent with current GAAP the ED pro-
poses that credit-related impairment of loan value should
be recognized as a component of net income in the period
in which credit-related decline in loan value occurs. Thus,
the most signi?cant effect of the proposal for measurement
of income is the recognition of expected, rather than in-
curred credit losses. Because the ED mandates continuing
?nancial statement reporting of important amortized-cost
amounts, it should result in a net increase in relevant loan-
related information, even if loan fair values include signif-
icant measurement error, as suggested by critics. Speci?-
cally, the ED requires that amortized cost be presented
parenthetically (and reconciled to fair value) on the face
of the statement of ?nancial position.
Bank regulatory accounting
Because bank regulators are legally authorized to di-
rectly collect and audit all necessary ?nancial information
for prudent monitoring, GAAP’s relevance for regulation is
not immediately clear. However, in the case of commercial
banks, US federal law requires that regulatory accounting
principles must be based on GAAP or a basis of accounting
that is ‘‘no less stringent’’ than GAAP (i.e., ‘‘RAP-GAAP con-
formity’’).
10
RAP-GAP conformity was made law in 1991 as
part of the FDIC Improvement Act (FDICIA) in response to
perceived de?ciencies in US bank regulation that contrib-
uted to the S&L crisis. Prior to FDICIA, regulators could en-
gage in case-by-case forbearance; for example, by allowing
an undercapitalized bank to continue to operate while
attempting to raise capital. In addition, regulators had dis-
cretion to exercise ‘‘wholesale’’ forbearance by selectively
relaxing accounting principles to indirectly ease capital
standards.
11
FDICIA also attempted to limit case-by-case dis-
cretion by adopting an objective capital-ratio based system
of prompt corrective action. As reported bank capital ratios
decline, FDICIA prescribes mandatory intervention. To fore-
stall wholesale forbearance, FDICIA also placed limits on reg-
ulators’ ability to selectively override GAAP.
Within the banking industry, RAP-GAAP conformity has
caused GAAP implementation ?exibility to become a source
of regulatory slack. For example, Hodder et al. (2002) docu-
ment that, as a direct result of RAP-GAAP conformity, banks’
initial ‘‘intent-based’’ classi?cation choices for investment
securities were driven more by concerns about potential ef-
fects of classi?cation choice on regulatory capital than by
?rms’ intended holding periods for securities. Speci?cally,
Hodder et al. (2002) show that when regulators relaxed
RAP-GAAP conformity by excluding fromregulatory capital
unrealized gains and losses on securities classi?ed as Avail-
able for Sale, banks responded by reclassifying the majority
of their investment securities from Held to Maturity (i.e.,
recognized at amortized cost) to Available for Sale (i.e., rec-
ognized at fair value). This choice to recognize investment
securities at fair value contrasts markedly with the indus-
try’s near unanimous opposition to fair value recognition
of investment securities that was expressedincomment let-
ters on the FASB’s Exposure Draft of Statement of Financial
Accounting Standards No. 115, Accounting for Certain Invest-
ments in Debt and Equity Securities (Schulz &Hollister, 2003),
suggesting divergence betweencompanies’ statedpositions
in comment letters and their ?nancial reporting behaviors
after implementation of accounting standards.
We propose that the disproportionately large private-
bank-af?liated response to the fair value provisions of
the 2010 ED indicates that RAP-GAAP conformity contin-
ues to heavily in?uence banks’ stated positions against fair
value measurement. In general, private-company repre-
sentatives do not submit high volumes of comment letters
in response to FASB standard-setting proposals because,
unlike public companies, private companies are not statu-
torily required to issue general-purpose GAAP ?nancial
statements. Private companies are more likely to be ?-
nanced with debt capital and may customize ?nancial
information to suit the requirements of their capital pro-
viders (e.g., lenders).
However, FDICIA’s de facto RAP-GAAP conformity has re-
sulted in unprecedented pressure on ?nancial accounting
standard setters to consider the effects of accounting stan-
dards on private companies not otherwise bound by GAAP,
as well as the concerns of bank regulators who are not con-
strained in their ability to collect private ?nancial informa-
tion from regulated enterprises. Cognizant that regulatory
objectives may diverge fromthe US Securities and Exchange
Commission’s (SEC’s) joint objectives of investor protection,
enhancedtransparency, andrelevancy of ?nancial reporting
for outside capital providers, the SEC has resisted the notion
that the content of general purpose ?nancial statements
should be dictated by regulatory concerns (Johnson, 2009).
The Financial Crisis Advisory Group (2009) also noted
that general-purpose ?nancial reporting and ‘‘prudential
regulation’’ have different reporting objectives. However,
despite concerns about divergent regulatory and investor
objectives, the US banking industry continues to press for
oversight over the ?nancial accounting standard setting
process (American Bankers Association, 2009), and the
AICPA has proposed that bank regulators should be in-
cluded in an International Accounting Standards Board
Monitoring Group (AICPA, 2008). Our paper suggests that
pressure to modify measurement principles to achieve de-
sired bank capital ratios places accounting standard setters
in a direct role of bank policy-making and can subvert their
9
Untabulated descriptive statistics reveal that, as a percentage of total
assets, loans for our median sample bank equal 66% of total assets, with the
?rst quartile equal to 54% of total assets.
10
Title 12, Chapter 16 §1831n(a)(2)(A) and (B)).
11
For example, prior to the S&L Crisis, regulators allowed certain losses to
be deferred as assets and amortized instead of immediately recognized in
income. Such changes were inconsistent with GAAP at the time.
L.D. Hodder, P.E. Hopkins / Accounting, Organizations and Society 39 (2014) 117–133 121
designated role of setting standards that provide unbiased,
decision-useful information.
12
Hypotheses
While US RAP-GAAP conformity can explain the heavy
concentration of public and private banks responding to
the ED (compared to other industries), it does not explain
why representatives from 1047 commercial banks submit-
ted comment letters while representatives from the
remaining 4242 commercial banks (for which we have suf-
?cient data to conduct our analyses) did not. US RAP-GAAP
conformity also does not explain why most representatives
from responding banks commented on only one issue re-
lated to one class of assets (i.e., fair value recognition of
loans) in an ED that spanned 214 pages and that requested
responses to 71 separate questions.
13
We investigate three potential determinants of responding
banks’ decisions to submit comment letters in response to the
ED. First, as we describe in the previous section, current US
GAAP for loans and related impairments includes signi?cant
opportunities for bank managers to use accounting slack. In
the banking industry, regulator-imposed RAP-GAAP confor-
mity creates de facto regulatory forbearance that is derived
from accounting slack within GAAP. By retaining important
elements of the current amortized-cost system while also
requiring fair value recognition for loans, the FASB’s ED is
likely to increase the amount and quality of relevant informa-
tion about loans. Further, the proposed rules should constrain
managers’ ability to opportunistically time their recognition
of losses to meet capital and earnings expectations. For these
reasons, we predict that individuals af?liated with banks that
historically have used the reporting ?exibility available in the
incurredloss model to manage net income andcapital will va-
lue that ?exibility and will lobby against accounting stan-
dards that threaten to undermine that ?exibility. This leads
to our ?rst alternative-form hypothesis:
H1. Banks that historically used the ?exibility available
under the incurred loss approach for loan losses are more
likely to lobby standard setters against a change to fair
value measurement of loans.
Second, we propose that because the additional disclo-
sures required by the standard will increase transparency,
private bene?t extraction facilitated by high agency prob-
lems and lowagency costs is an important motivating factor
for responding bank representatives. Incentives for trans-
parent reporting are decreasing in net bene?t extraction
(Jensen & Meckling, 1976). This means that bank represen-
tatives obtaining more private bene?ts relative to agency
costs imposed by outside capital providers will have less
incentive to improve banks’ ?nancial reporting transpar-
ency. This leads to our second alternative-form hypothesis:
H2. Banks with lower demand for transparent reporting
are more likely to lobby standard setters against fair value
measurement of loans.
Although banks’ demand for decreased ?nancial report-
ing transparency is unobservable, this demand should be
positively related to other observable choices that are indi-
cators that the value of private-bene?t extraction exceeds
the potential value of reduced agency costs that would
be achieved via greater reporting transparency. Speci?-
cally, we predict that banks submitting comment letters
opposing the loan-reporting provisions of the ED are (1)
more likely to elect to participate in the new optional, sup-
plemental, ?xed-price deposit insurance and debt guaran-
tee programs offered by the FDIC, (2) less likely to have
uninsured outside debt and (3) less likely to voluntarily
purchase ?nancial statement audits of their non-regula-
tory GAAP ?nancial statements.
When government-sponsored insurance is provided for
bank liabilities, total agency costs decline because insured
depositors do not demand a risk premium and opportuni-
ties for private bene?t extraction increase because regula-
tors do not have incentives to provide the same level of
monitoring provided by private creditors (Forssbaeck,
2011). Therefore, banks purchasing additional ?at rate
insurance have lower agency costs, greater opportunities
for private bene?t extraction, and less incentive to increase
?nancial reporting transparency. Similarly, and consistent
with Jensen and Meckling (1976), ?rms choosing to ?nance
with uninsured debt are more likely to bene?t from a com-
mitment to transparent reporting, and ?rms that value
transparent reporting are more likely to voluntarily pur-
chase ?nancial statement audits.
The third reason bank representatives may have sub-
mitted comment letters is that they believe the standard
would negatively affect the relevance and reliability of
their ?rms’ ?nancial reports. For example, many of the
comment letters speci?cally mention the belief that fair
values are dif?cult to measure and the belief that fair value
is not relevant for loans because loans are generally held to
maturity. Thus, we also investigate banks’ most commonly
stated reasons for opposition to the ED. This leads to our
third alternative-form hypothesis:
H3. Banks decision to lobby standard setters against fair
value measurement of loans is related to the reasons for
opposition included in the comment letters.
Sample and research design
Sample selection
Our primary sample consists of all private and public
commercial banks with necessary ?nancial data in the
12
Regulators’ unwillingness to expand disclosure in the middle of the
credit crisis might be explained by political incentives. Kane (1989)
proposes that information suppression provides plausible deniability for
regulators whose short horizons allow problems to be passed off to
successors.
13
The ED includes other proposals that would affect banks’ regulatory
reporting under RAP-GAAP conformity. For example, as compared to
today’s mostly amortized-cost accounting, the ED proposes that companies
initially recognize and subsequently remeasure most liabilities at fair value
in the balance sheet. Despite this potentially major change in liability
measurement, the comment letters were almost entirely silent about the
liability proposal. We note that, compared to incurred-loss accounting for
loans, amortized-cost accounting for liabilities does not represent a
signi?cant source of accounting slack for most banks; thus, bank repre-
sentatives do not have a strong incentive to lobby against it.
122 L.D. Hodder, P.E. Hopkins / Accounting, Organizations and Society 39 (2014) 117–133
periods covered by our tests. We collect ?nancial state-
ment data from publicly available regulatory ?lings. Our
proxy for banks’ historical use of discretion under the in-
curred loss model requires quarterly data, so we also in-
clude quarterly data spanning 2001Q1 through 2011Q4.
Our sample includes 5,289 unique banks.
From the FASB’s web site, we individually collect all
2,971 comment letters submitted in response to the ED.
14
Two accounting graduate students analyzed each letter
and coded its content. For each letter we collect any identi-
fying information provided by the letter writer, including
name, state, and company af?liation. For letters not provid-
ing any company af?liation, we collect implicit company
af?liation by cross-referencing the letter-writer name, email
address domain, street address, and phone number to
company pro?les. If the letter writer has a documented asso-
ciation with a company (e.g. of?cer, director, or employee)
and takes an industry-related position on the issues, we
attribute company af?liation to the letter writer for statisti-
cal purposes. We classify the letters as either supporting or
opposing various issues mentioned in the ED, including fair
value measurement for loans, (more generally) fair value
measurement for assets and liabilities, equity method
accounting, hedge accounting, convergence with Interna-
tional Financial Reporting Standards (IFRS), and other
miscellaneous issues. The coders did not have knowledge
of our hypotheses, or our proxies for banks’ prior use of
accounting slack and agency problems.
To avoid double-counting of entities, we do not include
bank holding companies in our sample together with their
consolidated subsidiaries. Consistent with this convention,
letters written by personnel of holding companies are
attributed to the subsidiary banks, and the subsidiary
banks are included in the sample, but not the holding com-
pany. We attribute letters to preparers, users, or others
based on company af?liation. Because stated af?liation
may differ from implied af?liation, we use the following
classi?cation hierarchy to resolve contradictory identifying
information: (1) letterhead, (2) identi?able as of?cer or
director of ?rm, (3) identi?able as ?rm representative with
phone number, mailing address, of ?rm, or email domain
of a ?rm, and (4) self-referenced af?liation provided in
the text of the correspondence.
Panel A of Fig. 1 provides an example of a letter with
contradictory identifying information. Although the writer
self-identi?es as a ‘‘bank investor’’ in the text and as an
‘‘investor’’ in the signature line, we classify the letter as
preparer-originated because, according to regulatory ?l-
ings, the writer is the Chief Executive Of?cer of a commer-
cial bank. Correspondingly, we classify the industry as
‘‘Financial Firm.’’
Model speci?cation
To test our hypotheses, we estimate the following
model:
PrðLetterÞ
t
¼ a þ b
1
TIMELINESS
t
þ b
2
DEP INS
t
þ b
3
CREDITOR
t
þ b
4
AUDITED
t
þ b
5
COMM RBC
t
þ b
6
SOLD
t
þ b
7
NPA PCT
t
þ b
8
CAPITAL
t
þ b
9
ENFORCEMENT
t
þ b
10
SMALL
t
þ b
11
SIZE
t
þe
t
ð1Þ
where TIMELINESS – (H1) The increase in explanatory
power for the loan loss provision of future non-performing
assets over past non-performing assets. (Refer to the
Appendix for the exact speci?cation of this variable.); DE-
P_INS – (H2) An indicator variable equal to 1 if the com-
pany elected to purchase additional insurance from the
FDIC during 2008 and 2009; CREDITOR – (H2) Indicator
variable equal to 1 if the ?rm has outstanding uninsured
long-term debt, and 0 otherwise; AUDITED – (H2) Indicator
variable equal to 1 if the ?rm engages an independent
auditor to conduct an audit of ?nancial statements pre-
pared in conformity with GAAP, and 0 otherwise;
COMM_RBC – (H3) Total commercial and industrial loans
(including real estate) divided by risk-based capital at
December 31, 2009; SOLD – (H3) Indicator variable equal
to 1 if the ?rm sold loans during the two years ending
December 31, 2009, and 0 otherwise; NPA_PCT – (Control)
Average nonperforming assets divided by the average bal-
ance in loans computed over the year ended December 31,
2009; CAPITAL – (Control) Tier 1 regulatory capital as of
December 31, 2009 divided by average assets over the year
ended December 31, 2009; ENFORCEMENT – (Control) Indi-
cator variable equal to 1 if the bank was under an active
enforcement order in the 5 years preceding the beginning
of the comment period, and 0 otherwise; SMALL – (Control)
Indicator variable set to 1 if the average balance of total as-
sets for the year ended December 31, 2009 is less than or
equal to $10 billion, and 0 otherwise; SIZE – (Control) Nat-
ural log of average assets computed over the year ended
December 31, 2009.
All continuous variables are winsorized at the top and
bottom 1%. TIMELINESS is included to test H1, and is a
proxy for banks’ historical use of slack available in current
GAAP for loan losses for which we predict a negative coef-
?cient. The Appendix provides detailed speci?cation of
TIMELINESS. High levels of TIMELINESS are associated with
lower historical use of accounting slack. TIMELINESS is
measured as the increase in explanatory power for banks’
loan loss provision of future non-performing assets as
compared to past non-performing assets.
15
Because we re-
quire at least 5 degrees of freedom in the ?rm-speci?c esti-
mation, we cannot compute TIMELINESS for all ?rms in the
sample.
14
In January 2014, the letters were available athttp://www.fasb.org/jsp/
FASB/CommentLetter_C/CommentLetterPage&cid=1218220137090&pro-
ject_id=1810-100&page_number=1.
15
Greater timeliness is not equivalent to conservatism because timeliness
is symmetric and captures the extent to which the current provision
captures both positive and negative future changes in non-performing
assets. For example, a provision would be timely if future expected
decreases in nonperforming loans were recognized as a reduction in the
current provision; however, the timely recognition of expected future
decreases in nonperforming assets would probably not be considered
conservative.
L.D. Hodder, P.E. Hopkins / Accounting, Organizations and Society 39 (2014) 117–133 123
Fig. 1. Examples of Fair Value Related Letters. Panels A and B of Fig. 1 provide examples of comment letters opposing and supporting fair value
measurement for banks, respectively. The letter in Panel A also provides an example of a bank manager representing himself as an investor and failing to
disclose that, at the time he submitted the letter, he was also the CEO and Chairman of the Board of Directors of a bank holding company. We were able to
con?rm the letter writer’s identity by matching his name and home address to current bank regulatory reports and SEC ?lings for a predecessor bank at
which he was an executive manager.
124 L.D. Hodder, P.E. Hopkins / Accounting, Organizations and Society 39 (2014) 117–133
For our tests of H2, we use three proxies for the demand
for transparency in banks: DEP_INS, CREDITOR, and
AUDITED, all obtained from SNL, Inc. DEP_INS is based on
banks’ decisions to purchase supplemental deposit
insurance through the temporary Transaction Account
Guarantee Program (TAGP) or Debt Guarantee Program
(DGP). The one-time supplemental insurance was offered
by the FDIC to forestall potential depositor runs at a time
Fig. 1 (continued)
L.D. Hodder, P.E. Hopkins / Accounting, Organizations and Society 39 (2014) 117–133 125
when con?dence in ?nancial institutions was at a nadir.
Because premiums were ?xed per unit coverage, supple-
mental insurance provided the most bene?t to ?rms with
?nancial conditions about which investors were most
uncertain. To the extent that greater uncertainty is associ-
ated with lower transparency, we expect DEP_INS to be
positively associated with banks’ decisions to lobby against
the proposed exposure draft.
Because banks with outside creditors have greater
incentives to reduce agency costs by providing transparent
?nancial information, we expect a negative coef?cient on
CREDITOR. For similar reasons, we include an indicator
variable for AUDITED, which is equal to 1 if the ?rm elects
to engage an independent auditor to attest that its general
purpose ?nancial statements are prepared in accordance
with GAAP. Firms distributing audited GAAP ?nancial
statements have made a commitment to more transparent
disclosure and, by engaging an auditor, have constrained
their own discretion. We expect a negative coef?cient on
AUDITED.
In H3, we investigate whether bank representatives’
decisions to submit comment letters are associated with
factors explicitly mentioned in the comment letters. We
include two variables designed to capture reasons most
commonly included in the comment letters. First, bank
representatives expressed concern over measurement of
hard-to-value loans. Speci?cally, they noted that a move
from current GAAP to the ED’s fair value measurement
for loans could cause banks with higher concentrations of
harder-to-value assets to incur signi?cant direct and indi-
rect costs in the preparation of periodic ?nancial state-
ments and regulatory reports. To control for variation in
the relative dif?culty of valuing loans across banks, we
include COMM_RBC, which is equal to total commercial
and industrial loans (including real estate) divided by
risk-based capital at December 31, 2009. These loan types
are most frequently mentioned as examples of hard-to-va-
lue loans. If negative comment letters are more likely from
banks with these types of loans, the coef?cient on
COMM_RBC will be positive.
A second and even more ubiquitous justi?cation in-
cluded in bank representatives’ negative comment letters
is that the respondents’ banks do not use fair values for
internal decision-making because fair values are irrelevant
to managers and investors. To control for this stated belief,
we identify a context in which fair values are clearly rele-
vant: the decision to sell certain loans in the portfolio. We
argue that banks making informed judgments to sell sub-
sets of loans for purposes that include liquidity, credit risk
management, and interest rate risk management, must
understand values, liquidity risks, and market risks inher-
ent in all eligible loan assets.
16
In contrast, representatives
from banks not engaging in loan sales might be motivated
to submit negative comment letters because they believe
fair value measurement is irrelevant given their business
models. To provide an additional text of H3, we include
SOLD as an indicator variable equal to 1 if the ?rm sold
any of its loans during the two years ending December 31,
2009, and 0 otherwise. If bankers’ assertions are accurate,
then we expect a negative coef?cient on SOLD.
We also include several control variables designed to
capture banks’ propensity to lobby. To control for the po-
tential effects of banks’ loan quality on the likelihood of
negative comment letter submission, we include NPA_PCT,
which is the average balance of nonperforming assets di-
vided by the average balance in loans computed over the
year ended December 31, 2009. The coef?cient on
NPA_PCT will be positive if banks with poor loan quality
were more likely to oppose the standard.
Dhaliwal (1982) proposes that capital structure will
in?uence the choice to lobby. Speci?cally, if a proposed
accounting standard presents a threat to capital levels or
is believed to increase capital volatility, then capital levels
should be negatively related to the choice to lobby against
the standard. This is particularly true in the banking indus-
try because regulators impose minimum capital standards.
For these reasons we expect a negative sign on the control
variable, CAPITAL.
An implicit assumption underlying banks’ resistance to
proposed fair value standards is that RAP-GAAP conformity
has caused the incurred loss model to become an impor-
tant source of regulatory slack that equates to forbearance
for troubled banks. If this is true, then those banks not eli-
gible for forbearance should be less likely to lobby for
accounting slack. Speci?cally, banks identi?ed by regula-
tors as troubled are given very little slack and are subjected
to enhanced monitoring, additional reporting, and the
imposition of higher minimum capital levels. For these
banks, regulatory constraints are already binding and
accounting slack is less valuable. To incorporate the effects
of active enforcement, we include an indicator variable,
ENFORCEMENT, which is equal to 1 if the ?rm has been
subject to a regulatory order in the ?ve years preceding
the beginning of the comment period, and 0 otherwise.
We predict a negative sign on the control variable,
ENFORCEMENT.
Two factors lead us to believe that an unusually high
proportion of small banks will respond to the ED. First,
RAP-GAAP conformity causes even the smallest banks to
be subject to the ED’s mandated fair value measurement
of loans in their regulatory ?lings, and loans are—by far—
the single largest class of assets on commercial banks’ bal-
ance sheets. Second, the American Bankers Association
(ABA) and state banking associations mounted an intense
lobbying campaign targeted speci?cally at smaller regional
and community banks (Ciesielski, 2010, p. 1). As part of
this campaign, these associations made available resources
that facilitated comment letter submission by less-sophis-
ticated banks. To control for this, we include SMALL, which
is equal to 1 if AVG_ASSETS (i.e., average daily balance of
total assets for the year ended December 31, 2009 reported
on the bank’s call report) is less than or equal to $10 billion,
and 0 otherwise. We predict a positive sign on the control
variable, SMALL.
By separately controlling for small banks, we are able to
also control for bank size within the small versus large par-
tition of banks. To accomplish this we also include SIZE,
which is the natural log of average assets computed over
16
Even if banks choose only to sell subsets of loan assets, all loan assets
are eligible for sale. Markets exist for most loan assets, even those with
unique terms (Ryan 2007).
126 L.D. Hodder, P.E. Hopkins / Accounting, Organizations and Society 39 (2014) 117–133
the year ended December 31, 2009. After controlling for
SMALL in the regression equation, we expect larger banks
within each of the small and large partitions to submit
comment letters. Therefore, we predict a positive sign on
the control variable, SIZE.
17
Results
Hypotheses tests
Panel A of Table 1 presents descriptive statistics for the
2971 comment letters obtained from the FASB in response
to the 2010 Exposure Draft. Consistent with prior litera-
ture, we ?nd that preparers are much more likely than
users to submit letters, with preparers accounting for
91.7% of all letters compared to 5.3% of letters attributable
to individual investors and investment advisors (Ryan
et al., 2000; Sutton, 1984; Tutticci et al., 1994; Weetman
et al., 1996).
18
As we previously noted, ?nancial statement
users (e.g., current and prospective investors) may be reluc-
tant to invest in accounting-standards-related lobbying be-
cause the cost of disclosure is borne by the disclosing
parties (i.e., it is only indirectly shared by current investors),
and the resulting bene?ts of enhanced disclosure is a public
good (Olson, 1965; Sutton, 1984).
Representatives from ?nancial ?rms, which include
banks, credit unions, specialty ?nance, and insurance com-
panies, prepared the vast majority of comment letters,
accounting for 2525 (i.e., 85%) out of 2971 separate pieces
of correspondence. The content of the correspondence ad-
dresses a variety of concerns, including fair value measure-
ment, hedging, convergence with IFRS, and the equity
method of accounting. However, the vast majority of re-
sponses was submitted by commercial-bank representa-
tives, and the subject of this correspondence is generally
limited to criticism of the proposed expansion of fair value
measurement for loans.
For purposes of analysis, we focus on comments
addressing fair value measurement and code the comments
Table 1
Descriptive statistics comment letters.
Af?liation subgroup Number Overall percent (%) Positive Negative Mixed Subgroup percent negative (%)
Panel A: Comment letters categorized by source and overall letter position
Preparers
Financial ?rms 2525 85.0 0 2521 4 99.8
Bank Trade Assoc. 50 1.7 0 50 0 100.0
Other ?rms 150 5.0 7 116 27 77.3
Total preparers 2725 91.7 7 2687 31 98.6
Users
Investors/advisors 131 4.4 12 114 5 87.0
Other/individuals 27 0.9 3 18 6 66.7
Total users 158 5.3 15 132 11 83.5
Other
Public accountants 35 1.2 0 33 2 94.3
Regulators 18 0.6 0 15 3 83.3
Other 35 1.2 3 24 8 68.6
Total other 88 3.0 3 72 13 81.8
Overall total 2971 100.0 25 2891 55
Overall total percent 0.8% 97.3% 1.9%
Number
Panel B: Reconciliation of total comment letters to sample of commercial banks submitting letters
Total letters 2971
Non-?nancial ?rms 446
Duplicate letters from individual ?rms 828
Potential ?nancial ?rm sample 1697
Missing required data ?elds 650
Final sample 1047
It reports descriptive statistics for all 2971 comment letters received by the Financial Accounting Standards Board (FASB) in response to its Exposure Draft
titled, Accounting for Financial Instruments and Revisions to the Accounting for Derivative Instruments and Hedging Activities (FASB, 2010). Panel A categorizes
the comment letters according to their source. Panel B provides a reconciliation of the total number of letters received by the FASB to the ?nal unique ?rm
sample used in our statistical analyses of commercial banks.
17
Size has been used in prior research as a proxy for political costs (Watts
& Zimmerman, 1978) because theory suggests that large ?rms are more
likely to lobby for income-decreasing accounting standards that reduce
political visibility. Further, scale economies are likely important both for
the probability of engaging in lobbying activity and for the potential gains
from in?uence (Hill, Kelly, Lockhart, & Van Ness, 2011). Given the relative
large number of small commercial banks that submitted comment letters,
these theoretical relations do not appear to apply in the present setting. The
results of our analyses are qualitatively similar (i.e., coef?cient direction,
magnitude and signi?cance levels for accounting-slack and agency-related
independent variables) if we omit SMALL, SIZE or both variables from our
speci?cation.
18
Our tabulation may understate the number of users, because some user
groups, including Financial Executives International surveyed their mem-
bers and conveyed the tabulated results in a single letter. Extrapolation of
survey results to letter counts would have resulted greater user participa-
tion and a higher ratio of favorable responses to the ED.
L.D. Hodder, P.E. Hopkins / Accounting, Organizations and Society 39 (2014) 117–133 127
as ‘‘Positive,’’ ‘‘Negative,’’ or ‘‘Mixed.’’
19
Panel A of Fig. 1 pro-
vides an example of a comment letter which we coded ‘‘Neg-
ative’’ with respect to fair value. Panel B provides an example
of a ‘‘Positive’’ response. Responses that were neither clearly
‘‘Positive’’ nor ‘‘Negative’’ were coded ‘‘Mixed.’’ Our coding
reveals that ?nancial ?rms are almost uniformly negative
about fair value (99.8%), while non-?nancial ?rms are also
negative, but signi?cantly less so (77.3%, X
2
p
= 0.000). Those
classi?ed as users are less negative about fair value than pre-
parers (X
2
p
= 0.000). Because no commercial banks take a
favorable position with respect to fair value of loans, we treat
the entire commercial bank sample as ‘‘Negative’’ for the
remaining analyses.
Panel B of Table 1 provides a reconciliation of the num-
ber of letters to the ?nal unique-?rm sample used for the
remaining statistical analyses. Non-?nancial ?rms
accounted for 446 letters. Many ?nancial ?rms were
represented multiple times, leading to deletion of 828
pieces of correspondence to arrive at our unique ?rm sam-
ple. The maximum number of letters sent by individuals
associated with a single ?rm was 46 (untabulated). Credit
unions, savings banks, specialty ?nance, and commercial
banks missing necessary quarterly data items account for
another 650 pieces of omitted correspondence. The ?nal
sample comprises 1047 unique commercial banks.
Table 2 provides descriptive statistics for our regression
variables. Except as otherwise noted, regression variables
are measured as of December 31, 2009. Statistical compar-
isons reported in Table 2 suggest that lobbying banks differ
from non-lobbying banks on several dimensions.
20
Related
to our proxy for banks’ propensity to use prior accounting
slack, TIMELINESS, measured as the change in R
2
between
Table 2
Descriptive statistics for lobbying and non-lobbying banks.
Lobbying banks (n = 1047) Non-lobbying banks (n = 4242)
Mean Q1 Median Q3 Mean Q1 Median Q3 t Test Nonpara-metric
H1: Proxy for prior use of accounting slack
TIMELINESS 0.103 0.030 0.069 0.138 0.113 0.033 0.078 0.156 ÃÃÃ ###
H2: Proxies for the demand for transparency
DEP_INS 0.854 0.817 + ###
CREDITOR 0.023 0.031 +
AUDITED 0.452 0.482 + #
H3: Proxies for reasons provided in comment letters
COMM_RBC 3.351 1.907 3.205 4.515 3.359 1.652 3.085 4.742
SOLD 0.105 0.085 + ##
Control and other variables
NPA_PCT 0.021 0.004 0.012 0.026 0.024 0.003 0.012 0.029 ÃÃÃ ##
CAPITAL 9.780 8.260 9.160 10.700 10.167 8.140 9.300 11.220 ÃÃÃ ###
ENFORCEMENT 0.325 0.376 + ###
AVG_ASSETS 0.344 0.080 0.163 0.333 1.477 0.064 0.131 0.283 ÃÃÃ
SMALL 0.939 0.929 +
SIZE 12.046 11.291 11.999 12.717 11.936 11.069 11.782 12.553 ÃÃ
R
2
Eq. (A2) 0.298 0.150 0.258 0.413 0.322 0.160 0.279 0.451 ÃÃÃ ###
ALLOW/LOANS 0.016 0.012 0.015 0.019 0.018 0.012 0.015 0.021 ÃÃÃ ##
Notes: The ?nal sample includes 5289 banks, which is comprised of 1047 unique lobbying banks from which the FASB received comment letters (i.e.,
described in Table 1) and 4242 non-lobbying banks from which the FASB did not receive comment letters and for which we are able to obtain necessary
?nancial data. LETTER is an indicator variable equal to 1 if the bank submitted a comment letter in response to the ED (i.e., left half of Panel A), and 0
otherwise (i.e., right half of Panel A); TIMELINESS is the ?rm-speci?c difference in adjusted R
2
from estimating Eq. (A2) compared to that derived from
estimating Eq. (A1) described in Appendix A; DEP_INS is an indicator variable equal to 1 if the ?rm elected to participate in the Transaction Account
Guarantee Program or the Debt Guarantee Program when these programs were offered by the Federal Deposit Insurance Corporation, and 0 otherwise;
CREDITOR is an indicator variable equal to 1 if the ?rm has outstanding non-deposit long term debt on December 31, 2009, and 0 otherwise; AUDITED is an
indicator variable equal to 1 if the ?rm engages an independent auditor, and 0 otherwise; COMM_RBC is total commercial and industrial loans (including
real estate) divided by risk-based capital at December 31, 2009; SOLD is an indicator variable equal to 1 if the ?rm sold any of its loans during the two years
ending December 31, 2009, and 0 otherwise; NPA_PCT is the average nonperforming assets divided by the average balance in loans computed over the year
ended December 31, 2009; CAPITAL is Tier 1 regulatory capital as of December 31, 2009 divided by average assets over the year ended December 31, 2009;
ENFORCEMENT is an indicator variable equal to 1 if the bank was under an active enforcement order in the 5 years preceding the beginning of the comment
period, and 0 otherwise; AVG_ASSETS is the average daily balance of total assets for the year ended December 31, 2009 reported on the bank’s call report;
SMALL is a an indicator variable set to 1 if AVG_ASSETS is less than or equal to $10 billion, and 0 otherwise; SIZE is the natural log of AVG_ASSETS; R
2
Eq.
(A2) is the adjusted R
2
from the ?rm-speci?c estimation of Eq. (A2) described in Appendix A; ALLOW/LOANS is the ratio of the allowance for loan losses
divided by loans at December 31, 2009.
The t-test column reports two-sample pooled t-test signi?cance levels with Satterthwaite approximations of degrees of freedom in cases of unequal sample
variance. ÃÃÃ, ÃÃ, and à denote parametric statistical signi?cance at the 1%, 5%, and 10% (two-tail) levels, respectively. For variables measured on a ratio
scale, we report Savage two-sample nonparametric signi?cance levels. Indicator variables are denoted ‘‘+’’ in the t-test signi?cance column, and are only
tested using nonparametric X
2
test statistics (i.e., we do not report parametric signi?cance levels for indicator variables). ###, ##, and # denote non-
parametric statistical signi?cance at the 1%, 5%, and 10% (two-tail) levels, respectively.
19
Responses were coded twice by independent research assistants. There
were no cases of disagreement.
20
Table 2 reports two-sample pooled t-test signi?cance levels with
Satterthwaite approximations of degrees of freedom in cases of unequal
sample variance. For variables measured on a ratio scale, we report Savage
two-sample nonparametric signi?cance levels. Indicator variables are only
tested using nonparametric X
2
test statistics.
128 L.D. Hodder, P.E. Hopkins / Accounting, Organizations and Society 39 (2014) 117–133
Eqs. (A2) and (A1), is larger for non-lobbying ?rms (0.113
compared to 0.103; p < 0.01). This is consistent with non-
lobbying ?rms’ provisions incorporating more timely for-
ward-looking information in the years preceding the com-
ment period.
Turning to our H2 proxies for the demand for transpar-
ency, banks receiving additional deposit insurance are
more likely to lobby against fair value (85.4% compared
to 81.7%; p < 0.01). In addition, banks choosing to provide
audited ?nancial statements prepared in accordance with
GAAP appear to be less likely to submit comment letters,
with 45.2% of lobbying banks purchasing ?nancial state-
ment audits, compared to 48.2% of non-lobbying banks
(p < .10). In contrast, the simple univariate comparisons
of CREDITOR reveal that lobbying and non-lobbying banks
are not statistically different in their propensity to have
non-deposit, long-term debt. However, overall, univariate
results are consistent with our hypotheses.
Table 2 reveals an interesting pattern in our H3 proxies
for reasons provided by responding banks in their com-
ment letters. First, there appears to be no statistical differ-
ence in the relative proportions of hard to value
commercial and industrial loans across lobbying and
non-lobbying banks. Second, banks opposing the ED actu-
ally had higher relative amounts of loans sold in the period
leading to the issuance of the FASB’s ED (0.105 versus
0.085, p < 0.05). Thus, there is no univariate evidence to
suggest that the reasons for opposition to the ED stated
in the comment letters drive the propensity to oppose
the ED.
We now turn to our control and other variables. Coun-
ter to the argument that banks with higher levels of non-
performing assets will resist a move to fair value
measurement, the relative proportion of NPA_PCT is higher
in non-lobbying banks than in banks opposing the ED
(0.024 compared to 0.021, p < 0.05). The average value of
CAPITAL of non-lobbying banks is signi?cantly higher than
that of lobbying banks (10.167% compared to 9.780%;
p < 0.01). A smaller percentage of banks with open
ENFORCEMENT actions in the 5 years preceding the com-
ment period choose to lobby (32.5% compared to 37.6%;
p < 0.01), consistent with these banks (1) reaping lower
bene?ts from accounting slack and (2) having decreased
desire for visibility given the extreme level of regulatory
scrutiny accompanying ENFORCEMENT.
The size distribution AVG_ASSETS is highly skewed for
both groups, as suggested by the fact that medians are
much smaller than the means. The mean value of AVG_AS-
SETS for non-lobbying ?rms is larger than lobbying ?rms
(average assets of $1.477 billion compared to $0.344 bil-
lion); however, the median values do not statistically dif-
fer. Given the skewed distribution, we use the natural log
of average assets in our regressions to capture SIZE. We
?nd that the mean value for SIZE is signi?cantly larger
for lobbying ?rms (12.046) than that of non-lobbying ?rms
(11.936) (p < 0.05), but the medians are, again, not statisti-
cally different. We include the variable SMALL because,
compared to companies’ responses to other FASB propos-
als, an unusually large number of small, private banks sub-
mitted comment letters. Interestingly, the proportion of
small banks submitting comment letters appears to
represent the overall proportion of small banks in the en-
tire sample because the difference in SMALL between lob-
bying (93.9%) and non-lobbying banks (92.9%) is not
statistically different.
Related to banks’ past use of the accounting slack avail-
able in the incurred loss model for loans, we ?nd that the
average R
2
corresponding to the ?rm-speci?c estimation
of Eq. (A2) is signi?cantly lower for lobbying ?rms
(0.298) than for non-lobbying ?rms (0.322), suggesting
that lobbying banks’ provisions are more likely to deviate
from amounts predicted by past and forward-looking
explanatory variables (p < 0.01). We also ?nd that, in abso-
lute terms, lobbying ?rms have signi?cantly lower allow-
ances for loan losses as a percentage of loans (0.016
compared to 0.018, p < .01).
Table 3 presents statistics from our estimations of Eq.
(1) with TIMELINESS serving as our H1 proxy for banks’
historical use of the slack available in the incurred-loss
model for loans; DEP_INS, CREDITOR and AUDITED serving
as H2 proxies for banks’ demand for decreased transpar-
ency, and COMM_RBC and SOLD serving as our H3 proxies
for reasons commonly included in banks’ comment letters.
Overall, the model explanatory power is signi?cant for Eq.
(1) (log-likelihood of 76.78, p < 0.000), with the overall
Table 3
Logistic regression of banks’ probability of lobbying on hypothesized
determinants.
Pred. Eq. (1)
Sign Coeff. X
2
Intercept ? À2.251 42.64
ÃÃÃ
H1: Prior use of accounting slack
TIMELINESS
t
À À0.489 6.96
ÃÃÃ
H2: Demand for transparency
DEP_INS
t
+ 0.141 6.17
ÃÃÃ
CREDITOR
t
À À0.268 3.99
ÃÃ
AUDITED
t
À À0.094 4.97
ÃÃ
H3: Reasons in comment letters
COMM_RBC
t
+ À0.001 2.28
SOLD
t
À 0.204 5.37
ÃÃ
Control variables
NPA_PCT
t
+ À1.605 6.00
ÃÃ
CAPITAL
t
À À0.016 6.69
ÃÃÃ
ENFORCEMENT
t
À À0.126 7.97
ÃÃÃ
SMALL
t
+ 0.433 13.89
ÃÃÃ
SIZE
t
+ 0.106 19.72
ÃÃÃ
Model Log-likelihood 76.78
DF 11
p-Value 0.10), while repre-
sentatives from banks that sell loans were more likely to
submit negative comment letters (b
6
= 0.204, X
2
= 5.37,
p < 0.05).
Turning to our control variables, we ?nd that represen-
tatives from banks with higher levels of nonperforming
assets were actually less likely to submit negative com-
ment letters (b
7
= À1.605, X
2
= 6.00, p < 0.05). We also ?nd
that, as predicted, our control variables CAPITAL and
ENFORCEMENT are signi?cantly negatively associated with
lobbying (b
8
= À0.016, X
2
= 6.69, p < 0.01 and b
9
= À0.126,
X
2
= 7.97, p < 0.01 respectively). The results for these vari-
ables are also consistent with the signi?cance of regulatory
factors in banks’ decisions to lobby against fair value
accounting for loans. Low CAPITAL levels immediately
preceding the comment period in?uenced ?rms to lobby
for accounting slack that would be bene?cial in maintain-
ing compliance with minimum regulatory capital
standards, while the presence of ENFORCEMENT damp-
ened banks’ propensity to lobby, consistent with bankers
perceiving lower bene?ts of accounting slack when regula-
tory scrutiny is high.
The signi?cantly positive coef?cient on SMALL
(b
10
= 0.433, X
2
= 13.89, p < 0.01) con?rms our observation
that a disproportionally large number of small banks re-
sponded to the FASB’s ED, while the positive and signi?-
cant coef?cient on SIZE (b
11
= 0.106, X
2
= 19.72, p < 0.01)
is consistent with the notion that after controlling for
SMALL (and other factors), the larger banks within the
small/large partitions have a greater propensity to lobby.
21
Taken together, the results reported in Table 3 suggest that
our proxies for historical use of accounting slack (H1) and
lower demand for transparency (H2) are robust determi-
nants of bank representatives’ choices to submit comment
letters critical of the FASB’s ED. In contrast, our proxies for
the reasons for opposition most commonly included in com-
ment letters were either not signi?cant or opposite of the
predicted relation.
Robustness
As described in our discussion of model speci?cation,
we use include TIMELINESS (Beatty & Liao, 2011) as our
proxy for banks’ historical use of slack available in current
GAAP for loan losses. To mitigate concerns that our results
are contingent on the speci?c proxy we use for the histor-
ical use of accounting slack, we also conduct our analyses
using three other measures of historical slack usage by
banks: LLCONSERV, LLACCRUAL and STD_ROA. LLCONSERV
is based on the relative size of the allowance for loan losses
controlling for non-performing assets. This is consistent
with Watts’ (2003) notion of conservatism, which he
describes as an intentional downward bias in reporting
the value of net assets. We compute LLCONSERV as the
residual from a regression of ALLOWANCE/LOANS on
non-performing assets (NPA) as of December 31, 2009.
Higher (lower) values of LLCONSERV indicate that banks
are using accounting slack to report relatively higher (low-
er) loan-loss allowances, which means they are reporting
relatively lower (higher) net asset balances for loans after
controlling for the amount of problem assets.
We calculate LLACCRUAL as the (signed) ?rm-speci?c
residuals from a cross-sectional estimation of Eq. (A2)
(speci?ed in the Appendix) over the period 2010Q1
through 2011Q3. We structure this ?rm-speci?c time ser-
ies measure to re?ect short-term earnings management,
which is distinct from our measure of TIMELINESS. LLAC-
CRUAL is estimated in cross-section, and captures the
residual short-term discretion in the provision after con-
trolling for the variables that in?uence TIMELINESS. Firms
with more negative (positive) realizations of LLACCRUAL
relatively underprovided (overprovided) for expected fu-
ture losses.
Our last alternative proxy, STD_ROA, is consistent with
Bushman and Williams’ (2012) ?nding that banks’ use of
loan loss discretion to report smooth earnings reduces
transparency and hinders the ability of outsiders to
21
Consistent with Beatty et al. (2002) we also consider whether a ?rm is
publicly traded or private and ?nd that publicly traded ?rms are less likely
to lobby against the ED. Inferences are unchanged by inclusion of PUBLIC.
However, AUDIT, becomes insigni?cant. AUDIT is correlated with PUBLIC
because all public ?rms must be audited. When the sample is restricted to
private ?rms, AUDIT is signi?cantly negative, and all inferences remain
unchanged.
130 L.D. Hodder, P.E. Hopkins / Accounting, Organizations and Society 39 (2014) 117–133
monitor. We calculate STD_ROA as the standard deviation
of quarterly net income divided by average assets for the
quarter, computed over the period 2001–2011.
In untabulated regression analyses, we ?nd that each of
these alternative proxies loads signi?cantly (and in the
predicted direction) when they are substituted for TIMELI-
NESS in equation (1) (lowest t statistic = 7.24, p < .001).
In addition, when all four proxies are simultaneously
included in our regression analysis, only LLACCRUAL
becomes statistically insigni?cant, with all three agency-
related proxies retaining statistical signi?cance. Taken
together, these supplemental analyses suggest that TIME-
LINESS is a robust proxy for banks’ historical use of
accounting slack in current GAAP for loan losses, and that
the inferences from our hypotheses are robust to
alternative measures of accounting slack.
Summary and conclusions
In 2010, the ABA and state banking associations
engaged in an intensive campaign to convince representa-
tives from member banks to lobby the FASB against the
ED’s proposal to measure and report loans at fair value
(Ciesielski, 2010, p. 1). Despite the fact that the FAF
(1977) states that ‘‘it would not be appropriate to establish
a standard based solely on a canvass of the constituents’’
(p. 19), it appears that the FASB decided to scuttle fair
value recognition for loans because of the ‘‘overwhelm-
ingly negative reaction to its proposal from companies
and investors’’ (Rapoport, 2011). Given the importance of
due process and the in?uence of comment letters, we seek
to better understand why representatives from 1047
unique public and private commercial banks submitted
comment letters opposing the FASB’s proposal, while rep-
resentatives from 4242 commercial banks (for which we
have suf?cient data to conduct our analyses) chose to
remain silent.
We provide evidence suggesting that bank representa-
tives’ objections to the ED are motivated less by stated con-
ceptual concerns and more by self-interest. The incurred
loss model for accounting for loan losses makes available
accounting slack that can be used to manage capital and
earnings. Our investigation reveals that representatives
lobbying the FASB against fair value measurement for
loans are more likely to be af?liated with a bank that his-
torically has used accounting slack inherent in the incurred
loss model to less-timely record loan losses. One objection
to fair values is that their use may increase measurement
discretion. Our ?nding that banks historically using
accounting slack are more likely to oppose the ED suggests
that the FASB’s fair value proposal is perceived to decrease
available accounting slack related to loans.
The fact that the ED retains amortized cost reporting on
the face of the ?nancial statements, and also requires the
credit-loss component of fair value changes to be esti-
mated and charged against net income, suggests that
?nancial statement users and regulators will have more
information under the proposal. Given that the demand
for transparent information is increasing in agency costs,
bank representatives obtaining more (unobservable)
private bene?ts relative to agency costs are more likely
to resist a movement to more transparent accounting stan-
dards. In our analysis of transparency-related motives, we
?nd that banks submitting comment letters opposing the
loan-reporting provisions of the ED are (1) more likely to
elect to participate in the new optional, supplemental,
?xed-price deposit insurance and debt guarantee programs
offered by the FDIC, (2) less likely to have nonguaranteed
outside creditors and (3) less likely to obtain ?nancial
statement audits of their non-regulatory GAAP ?nancial
statements. These results are consistent with bank manag-
ers and shareholders af?liated with lobbying banks bene-
?tting from the lower agency costs that accompany
government guarantees of indebtedness and the lower lev-
els of external monitoring that otherwise would be pro-
vided by unguaranteed debt holders and by auditors.
Our results are inconsistent with bankers’ arguments
that their opposition to fair value measurement is driven
primarily by concerns about costs, decreased transparency,
measurement dif?culty, or a mismatch with their existing
business model. Similarly sized banks using accounting
slack to delay the recognition of loan losses are unlikely
to face systematically higher implementation costs than
those not using such slack. Further, banks using accounting
slack (1) to manage earnings, (2) to less-timely recognize
anticipated non-performing assets, and (3) to present rela-
tively higher net asset balances than banks with similar as-
set quality, are also unlikely to be concerned with
potentially diminished transparency or lower reliability
associated with fair value estimation. Moreover, we ?nd
no evidence that representatives from banks with more-
dif?cult-to-value loans are more likely to lobby against fair
value measurement, and our results suggest that represen-
tatives from banks that use loan fair values for internal
decision-making are actually more likely to lobby against
fair value measurement.
Conceptually, fair value recognition should increase
transparency (Linsmeier, 2011). For example, Blankespoor,
Linsmeier, Petroni and Shakespeare (2013) ?nd that fair
value measurement of ?nancial instruments better de-
scribes banks’ credit risk. For public banks disclosing fair
values annually, Barth, Beaver and Landsman (1996) ?nd
that loan fair values are value-relevant. Further, Hodder,
Hopkins, and Wahlen (2006) suggest that income volatility
measured under full fair value accounting is signi?cantly
more informative about banks’ risks—and better re?ects
the risk priced in banks share prices and expected re-
turns—than volatility measured under the current US
reporting system. As compared to disclosure, ?nancial
statement recognition may increase the precision of fair
value estimates, leading to higher ?nancial statement
quality (Libby, Nelson, & Hunton, 2006).
Our ?ndings contribute to the continuing debates re-
lated to fair value measurement of ?nancial instruments,
the incurred-loss model for loan losses, RAP-GAAP confor-
mity and the impact of regulators on the promulgation of
general-purpose ?nancial accounting standards. Our re-
sults should also be considered in the context of due pro-
cess for setting accounting standards and for legislating
?nancial regulation. Our analysis reveals (1) that a nontriv-
ial number of individuals identi?ed as investors were, in
L.D. Hodder, P.E. Hopkins / Accounting, Organizations and Society 39 (2014) 117–133 131
reality, bank-af?liated individuals who concealed their
af?liation, (2) that the negative comment letters were con-
centrated among bank representatives, and (3) that com-
ment letters were submitted by representatives from
banks that had a lower demand for transparency and sys-
tematically utilized the accounting slack available in cur-
rent GAAP for loan losses.
To the extent comment letter submissions continue to
be part of regulatory and accounting-standard-setting
due process, regulators and standard setters may wish to
consider ways to effectively identify all letter writers and
their af?liations, and to explicitly consider the incentives
of all participants in the rule making process. Dispropor-
tionate preparer response can result in disproportionate
weighting of preparer concerns relative to potential public
bene?ts. Indeed, standard setters appear to be in?uenced
by comment letters, and research ?nds that standard set-
ters, when explaining standard setting results, are more
likely to mention preparer groups as exerting in?uence
through the comment letter process and other forms of
lobbying (Hope & Briggs 1982; Hope & Gray, 1982; Jupe,
2000; Ang, Gallery, & Sidhu, 2000). In the context of the
ED, some preparers represented themselves as users,
rather than preparers, perhaps to appear less self-serving
in their motivations. Our results suggest that dispropor-
tionate weighting of preparer feedback on proposed
accounting standards may have unintended consequences
with respect to public markets: successful lobbying that
results in less transparency may facilitate managers’ reten-
tion of private bene?ts while increasing public costs.
News accounts suggest that the FASB decided to drop
the ED because of the negative reaction to fair value mea-
surement for loans (Rapoport 2011).
22
This episode raises
important questions about the extent to which feedback
from vested interests should in?uence the standard setting
process. The Financial Accounting Foundation (FAF) notes
that standard setters must exercise judgment and consider
constituents’ comments in context: ‘‘(b)ecause standard set-
ting requires some perspective, it would not be appropriate
to establish a standard based solely on a canvass of the con-
stituents’’ (FAF, 1977, p. 19).
Some argue that accounting should be protected from
politically motivated or special interest lobbying (e.g. Ger-
both 1973; Horngren 1973). Because conceptually sound
accounting standards are likely to affect different compa-
nies in different ways, or favor one constituent group over
another, standard setting inherently involves decisions
that affect relative social welfare. As Solomons (1978)
notes, ‘‘[n]eutrality, in the sense in which the term is used
here does not imply that no one gets hurt’’ (p. 70). He sug-
gests that reactions to proposed accounting standards
must be considered in context with ultimate decisions
based on conceptual and social merits. The results of our
inquiry support this view.
Appendix A. Supplementary material
Supplementary data associated with this article can be
found, in the online version, athttp://dx.doi.org/10.1016/
j.aos.2013.10.002.
References
American Bankers Association (2009). Comment letter addressed to
members of the house ?nancial services committee concerning support
the Perlmutter-Lucas revised accounting amendment to H.R. 3996.
(November 18).
American Institute of Certi?ed Public Accountants (2008). Comment letter
Ms. Tamara Oyre concerning the review of the International Accounting
Standards Committee Foundation’s, Constitution (September 15).
Ang, N., Gallery, N., & Sidhu, B. K. (2000). The incentives of Australian
public companies lobbying against proposed superannuation
accounting standards. Abacus, 36(1), 40–70.
Barth, M., Beaver, W., & Landsman, W. (1996). Value relevance of banks’
fair value disclosures under SFAS No. 107. The Accounting Review, 71,
513–537.
Beatty, A., Chamberlain, S. L., & Magliolo, J. (1995). Managing ?nancial
reports of commercial banks: The in?uence of taxes, regulatory
capital, and earnings. Journal of Accounting Research, 33, 231–261.
Beatty, A. L., Ke, B., & Petroni, K. R. (2002). Earnings management to avoid
earnings declines across publicly and privately held banks. The
Accounting Review, 77, 547–570.
Beatty, A., & Liao, S. (2011). Do delays in expected loss recognition affect
banks’ willingness to lend? Journal of Accounting and Economics, 52,
1–20.
Blankespoor, E., Linsmeier, T., Petroni, K., & Shakespeare, C. (2013). Fair
value accounting for ?nancial instruments: Does it improve the
association between bank leverage and credit risk?. The Accounting
Review.http://dx.doi.org/10.2308/accr-50419 (forthcoming).
Bushman, R. M., & Williams, C. D. (2013). Delayed expected loss recognition
and the risk pro?le of banks (February 28). .
Bushman, R. M., & Williams, C. D. (2012). Accounting discretion, loan loss
provisioning, and discipline of Banks’ risk-taking. Journal of
Accounting and Economics, 54(1), 1–18.
Ciesielski, J. T. (2010). It’s déjà vu all over again: The fair value accounting
wars. The Analyst’s Accounting Observer, 19 (August 16).
Collins, J. T., Shackelford, D. A., & Wahlen, J. M. (1995). Bank differences in
the coordination of regulatory capital, earnings, and taxes. Journal of
Accounting Research, 33, 263–291.
Dechow, P. M., Hutton, A. P., & Sloan, R. G. (1996). Economic consequences
of accounting for stock-based compensation. Journal of Accounting
Research, 34(Suppl.), 1–20.
Dhaliwal, D. (1982). Some economic determinants of management
lobbying for alternative methods of accounting: Evidence from the
accounting for interest costs issue. Journal of Business Finance and
Accounting, 9, 255–265.
Dugan, J. C. (2009). The need to preserve uniform standards for national
banks. Speech delivered on September 24, 2009. .
Financial Accounting Foundation (1977). The structure of establishing
?nancial accounting standards. Report of the Structure Committee of
the FAF: Norwalk, CT.
Financial Accounting Standards Board (2009). Minutes of the March 18,
2009 Board meeting—Loan loss disclosures (April 28).
Financial Accounting Standards Board (2010). Accounting for ?nancial
instruments and revisions to the accounting for derivative instruments
and hedging activities. Exposure draft of proposed accounting
standards update of topic 825 and topic 815 (May 26).
Financial Crisis Advisory Group (2009). Report of the Financial Crisis
Advisory Group (July 28). .
Forssbaeck, J. (2011). Ownership structure, market discipline, and banks
risk-taking incentives under deposit insurance. Journal of Banking &
Finance, 35(10), 2666.
Gerboth, D. (1973). Research, intuition and politics in accounting theory.
Accounting Review, 48(3), 475–482.
22
The comment period for the ED ended on September 30, 2010. The
FASB announced, in January 2011, that it would not consider measuring
and reporting loans at fair value as part of a ?nancial instruments standard.
In interviews, FASB Chair Leslie Seidman characterized these actions as a
response to feedback on the proposal (Rapoport, 2011). In April 2013, the
FASB issued a new Exposure Draft, Financial Instruments—Overall (Subtopic
825-10): Recognition and Measurement of Financial Assets and Financial
Liabilities—Proposed Amendments to the FASB Accounting Standards Codi?ca-
tion
Ò
, that largely excludes loans from fair value measurement.
132 L.D. Hodder, P.E. Hopkins / Accounting, Organizations and Society 39 (2014) 117–133
Hill, M. D., Kelly, G. W., Lockhart, G. B., & Van Ness, R. A. (2011).
Determinants and effects of corporate lobbying (January 19). or .
Hochberg, Y. V., Sapienza, P., & Vissing-Jørgensen, A. (2009). A lobbying
approach to evaluating the Sarbanes-Oxley Act of 2002. Journal of
Accounting Research, 47, 519–583.
Hodder, L. D., Hopkins, P. E., & Wahlen, J. M. (2006). Risk-relevance of fair-
value income measures for commercial banks. The Accounting Review,
81, 337–375.
Hodder, L., Kohlbeck, M., & McAnally, M. L. (2002). Accounting choices
and risk management: SFAS No. 115 and U.S. bank holding
companies. Contemporary Accounting Research, 19, 225–270.
Hope, T., & Briggs, J. (1982). Accounting policy making – Some lessons
from deferred taxation debate. Accounting and Business Research,
12(46), 83–96.
Hope, T., & Gray, R. (1982). Power and policy making: the development of
an R&D standard. Journal of Business Finance and Accounting, 9(4),
531–558.
Horngren, C. (1973). The marketing of accounting standards. Journal of
Accountancy, 137(4), 61–66.
Jensen, M. C., & Meckling, W. H. (1976). Theory of the ?rm: Managerial
behavior, agency costs and ownership structure. Journal of Financial
Economics, 3, 305–360.
Johnson, S. (2009). Keep bank regulators way from FASB: SEC Commissioner.
CFO.com (November 17). .
Jupe, R. (2000). Self-referential lobbing of accounting standards board:
The case of Financial Reporting Standard No. 1. Critical Perspectives on
Accounting, 11(3), 337–359.
Kane, E. J. (1989). How incentive-incompatible deposit-insurance funds fail.
National Bureau of Economic Research Working Paper No. 2836.
.
Kelly, L. (1985). Corporate management lobbying on FAS No. 8: Some
further evidence. Journal of Accounting Research, 23, 619–632.
Libby, R., Nelson, M., & Hunton, J. (2006). Recognition v. disclosure:
Auditor tolerance for misstatement, and the reliability of stock-
compensation and lease information. Journal of Accounting Research,
44, 533–560.
Linsmeier, T. J. (2011). Financial reporting and ?nancial crises: The case
for measuring ?nancial instruments at fair value in the ?nancial
statements. Accounting Horizons, 25, 409–417.
Liu, C., & Ryan, S. G. (2006). Income smoothing over the business cycle:
Changes in banks’ coordinated management of provisions for loan
losses and loan charge-offs from the pre-1990 bust to the 1990s
boom. The Accounting Review, 81, 421–441.
Moyer, S. (1990). Capital adequacy ratio regulations and accounting
choices in commercial banks. Journal of Accounting and Economics, 13,
123–154.
Nichols, D. C., Wahlen, J. M., & Wieland, M. M. (2009). Publicly-traded vs.
privately-held: implications for bank pro?tability, growth risk, and
accounting conservatism. Review of Accounting Studies, 14, 88–122.
Olson, M. (1965). The logic of collective action. Cambridge, MA: Harvard
University Press.
Rapoport, M. (2011). Banks force retreat on fair-value plan. The Wall Street
Journal (January 26), C2.
Ryan, S. G. (2007). Financial instruments and institutions: Accounting and
disclosure rules (2nd ed.). New York: John Wiley & Sons.
Ryan, C., Dunstan, K., & Stanley, T. (2000). Local government accounting
standard-setting in Australia: Did constituents participate? Financial
Accountability and Management, 16, 373–396.
Schulz, S. M., & Hollister, J. (2003). Lobbying FASB on accounting for
investments. The Journal of Applied Business Research, 19, 85–93.
Solomons, D. (1978). The politicization of accounting. Journal of
Accountancy, 146(5), 65–72.
Sutton, T. G. (1984). Lobbying of accounting standard-setting bodies in
the UK and the USA: A Downsian analysis. Accounting, Organizations
and Society, 9, 81–95.
Tschirhart, J., O’Brien, J., Moise, M., & Yang, E. (2007). Bank commercial loan
fair value practices. Washington, DC: Federal Reserve Board. .
Tutticci, I., Dunstan, K., & Holmes, S. (1994). Respondent lobbying in the
Australian Accounting standard-setting process: ED49. Accounting,
Auditing & Accountability Journal, 7, 86–104.
Watts, R. L. (2003). Conservatism in accounting Part I: Explanations and
implications. Accounting Horizons, 17, 207–221.
Watts, R. L., & Zimmerman, J. L. (1978). Towards a positive theory of the
determination of accounting standards. The Accounting Review, 53,
112–134.
Weetman, P., Davie, E. S., & Collins, W. (1996). Lobbying on accounting
issues: Preparer/user imbalance in the case of the operating and
?nancial review? Accounting, Auditing & Accountability Journal, 9,
59–76.
L.D. Hodder, P.E. Hopkins / Accounting, Organizations and Society 39 (2014) 117–133 133
doc_213708268.pdf