Journal of Accounting and Economics 55 (2013) 66–90

Contents lists available at SciVerse ScienceDirect

Journal of Accounting and Economics journal homepage: www.elsevier.com/locate/jae

Towards an understanding of the role of standard setters in standard setting$
Abigail Allen, Karthik Ramanna n
Harvard Business School, USA

a r t i c l e in f o

abstract

Article history:
Received 15 September 2010
Received in revised form
24 May 2012
Accepted 25 May 2012
Available online 7 June 2012

We investigate the effect of standard setters in standard setting. We examine how certain professional and political characteristics of FASB members and SEC commissioners predict the accounting ‘‘reliability’’ and ‘‘relevance’’ of proposed standards.
Notably, we find FASB members with backgrounds in financial services are more likely to propose standards that decrease ‘‘reliability’’ and increase ‘‘relevance,’’ partly due to their tendency to propose fair-value methods. We find opposite results for FASB members affiliated with the Democratic Party, although only when excluding financialservices background as an independent variable. Jackknife procedures show that results are robust to omitting any individual standard setter.
& 2012 Elsevier B.V. All rights reserved.

JEL classification:
D72
D78
G18
K22
L51
M41
Keywords:
Accounting
FASB
Politics
Relevance
Reliability
Standard setting

1. Introduction
As the Financial Accounting Standards Board (FASB) closes in on four decades, the role of its standards in shaping U.S. and international corporate reporting is widely acknowledged. An empirical literature on the political economy of FASB standard setting has emerged over that period to explore the origins of accounting standards largely through an analysis of constituent comment-letter lobbying (e.g., Watts and Zimmerman, 1978). But such comment-letter lobbying is only part of the political economy that determines accounting standards (e.g., Ramanna, 2008, studies the role of congressional intervention). At the core of the standard-setting process are the individuals that comprise the FASB and its sanctioning authority, the Securities and Exchange Commission (SEC). In this paper, we develop and test some exploratory hypotheses with a view towards building an understanding of the role of FASB and SEC regulators in U.S. GAAP.

$
We thank Mary Barth, Robert Bloomfield (the reviewer), Michelle Hanlon (the editor), Paul Healy, S.P. Kothari, Krishna Palepu, Sugata
Roychowdhury, Doug Skinner, Ross Watts, and seminar participants at Minnesota, MIT, the 2010 HBS IMO Conference, and the 2011 AAA FARS
Conference for helpful comments; Beiting Cheng, Michael Kregar, Scott Renner, and the HBS division of research for excellent research assistance; and
Charry Boris, Ron Guerrette, and the FAF for generous help with sourcing data on comment letters and exposure drafts. Any errors are our responsibility. n Corresponding author. Tel.: þ 1 585 317 6250.
E-mail address: kramanna@hbs.edu (K. Ramanna).

0165-4101/$ - see front matter & 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.jacceco.2012.05.003 A. Allen, K. Ramanna / Journal of Accounting and Economics 55 (2013) 66–90

67

Although the idea that FASB and SEC regulators can matter in standard setting is intuitively appealing, it has not been subject to empirical testing. This is due in part to limited data availability, but also in part to neoclassical economics, which is widely used in accounting research and tends to view ‘‘individuals’’ as ‘‘so empirically unimportant as to allow the use of
Occam’s razor in positive models’’ (e.g., Kalt and Zupan, 1984, p. 279). Recently, however, empiricists in finance and accounting have begun exploring the role of individuals on equilibrium outcomes, particularly in the context of individual managers and firm policies (e.g., Bertrand and Schoar, 2003; Bamber et al., 2010; Dyrenge et al., 2010). Moreover, in the regulation literature itself, there is some evidence of regulators’ preferences mattering in outcomes at both the congressional (e.g., Kau and Rubin, 1979) and the bureaucratic agency levels (e.g., Gormley, 1979). Thus, in the context of accounting standard setting, tests of influence of FASB and SEC regulators can help refine our understanding of the political economy of U.S. GAAP.
We conduct our study through an analysis of FASB exposure drafts proposed from 1973 (the FASB’s inception) through
2007. There are 149 such exposure drafts in our sample after data limitations. Our primary tests involve regressing assessments of the nature of an exposure draft on the average background characteristics of extant FASB and SEC regulators. We evaluate a proposed Statement of Financial Accounting Standards (SFAS) by focusing in particular on its impact on accounting ‘‘relevance’’ and ‘‘reliability’’—two characteristics usually cited as fundamental accounting properties in accounting textbooks (e.g., Stickney et al., 2010, pp. 23, 114). There are no obvious metrics to use in evaluating exposure drafts; our choice of ‘‘relevance’’ and ‘‘reliability’’ reflects our judgment on their importance to accounting. Since at least the publication of its conceptual statements in the late 1970s (e.g., FASB 1978, 1980), the FASB itself has viewed
‘‘relevance’’ and ‘‘reliability’’ as ‘‘the two primary qualities that make accounting information useful for decision making’’
(FASB, 1980, p. 5), adding that ‘‘serious disagreement’’ often arises ‘‘about whether the superior relevance of the results of one [accounting] method outweighs the superior reliability of the results of [another]’’ (FASB, 1980, p. 8). Moreover, the increased prominence, since the mid-1990s, of fair-value accounting in standard setting has generated additional interest in the ‘‘trade-off’’ between ‘‘relevance’’ and ‘‘reliability.’’ The FASB has often justified the increased use of fair values by arguing it will increase the ‘‘relevance’’ of accounting numbers (e.g., Johnson, 2005). In contrast, some academics have argued accounting estimates generated under fair-value accounting will decrease the ‘‘reliability’’ of financial reports (e.g.,
Watts, 2003).1
To obtain assessments of exposure drafts’ impact on ‘‘relevance’’ and ‘‘reliability’’ that are independent of researcher judgment, we develop a measure based on comment letters filed by the Big 8/6/5/4 auditors (hereafter, ‘‘Big N auditors’’).
There are 908 such comment letters in our sample after data limitations. The advantage to using Big N auditors’ comment letters is that they are available on most exposure drafts in our sample period and are contemporaneous (i.e., no hindsight bias). The letters are, however, likely to reflect the auditors’ private incentives, which can confound inferences if endogenous to our explanatory variables (i.e., the characteristics of FASB and SEC regulators). To mitigate this concern, in robustness tests we use an alternative assessment of the exposure drafts from two seasoned research assistants (with over
30 years of combined experience in accounting) blind to the objective of this study.
We build a biographical database of all 39 FASB members and all 41 SEC commissioners serving between 1973 and
2007. Drawing on empirical political-economy research that has examined the characteristics of regulators on regulation
(see Dal Bo, 2006, for a review), we focus on two sets of characteristics: professional and political. The professional characteristics are length of regulatory tenure, industry background in auditing, and industry background in investment banking/investment management (hereafter, ‘‘financial services’’)2; the political characteristics are affiliations, if any, with the Democratic and Republican parties. Prior research has consistently found high correlations between regulators’ professional and political characteristics and so has examined these characteristics both independently and jointly in multivariate regressions. We adopt this approach in our empirical design.
In examining professional characteristics independently, we find that longer average FASB and SEC tenures are associated with exposure drafts perceived by auditors as decreasing accounting ‘‘reliability;’’ but, we find no evidence of an association between the regulators’ tenures and exposure drafts’ ‘‘relevance.’’ If decreased ‘‘reliability’’ is an undesirable property of accounting (e.g., Watts, 2003), the result is consistent with longer regulatory tenures compromising accounting quality.3 Concerning industry backgrounds, we expect regulators with prior employment in auditing to be more sympathetic to accounting ‘‘reliability’’ (since ‘‘reliable’’ accounting lowers auditors’ litigation risk; e.g., Watts, 2003); in contrast, we expect members with prior employment in financial services to be sympathetic to valuation-relevant accounting (e.g., ICI, 2008), and thus more likely to promote ‘‘relevance’’ at the expense of ‘‘reliability’’ (e.g., Johnson, 2005).
We do not find results associating regulators’ careers in auditing with ‘‘reliability’’ and ‘‘relevance.’’ However, we find evidence that exposure drafts proposed by FASB members and SEC commissioners with prior experience in financial services are viewed by the Big N auditors as decreasing accounting ‘‘reliability.’’ Further, in the case of FASB members, experience in financial services is associated with exposure drafts viewed by the Big N auditors as increasing accounting

1
While ‘‘reliability’’ and ‘‘relevance’’ can be trade-offs in some circumstances, it is an empirical question as to whether these concepts are always at odds with each other.
2
When studying industry background, the regulatory literature has focused on industries most closely associated with the regulations being studied.
In our setting, we focus on auditing and financial services, viewing them as front-line intermediaries in the production and use of accounting information.
3
Stigler (1971) argues that longer regulatory tenures compromise regulation by promoting greater ‘‘coziness’’ between regulators and the regulated.

68

A. Allen, K. Ramanna / Journal of Accounting and Economics 55 (2013) 66–90

‘‘relevance.’’ Additional analysis suggests these associations are partly due to the tendency of FASB members with financial-services backgrounds (the proportion of which increases in our sample period) to propose standards that use fairvalue methods.
In studying regulators’ political characteristics, we are motivated by prior political-science research that has shown that political affiliations are salient predictors of regulator behavior: e.g., Cohen (1986) provides evidence that Democratic regulators are on average less sympathetic to corporate interests. In examining political characteristics independently, we find evidence that increased proportional membership of Democrats on the FASB is associated with exposure drafts that are perceived by the Big N auditors as both increasing accounting ‘‘reliability’’ and decreasing accounting ‘‘relevance.’’
However, when we examine the regulators’ professional and political characteristics jointly, we find that the results on backgrounds in financial services alone survive. Thus, in our population and time period, political affiliation does not appear to be a significant factor beyond financial services affiliation. We note that there is no ex-ante theory that suggests either professional or political characteristics are more important than the other in explaining regulatory decisions (e.g.,
Dal Bo, 2006), thus future research is needed to draw more definitive conclusions.
There are certain other caveats to our analysis. First, the small population of regulators in our study might mean that influential observations are driving reported statistical significance. We mitigate this concern through a jackknifing procedure where we re-estimate all regressions successively eliminating each regulator to determine if she/he is instrumental to inferences: this procedure does not alter inferences on variables discussed as statistically significant.
Second, the scope of our study is limited by our choice of dependent and independent variables: other dependent variables
(e.g., ‘‘comparability,’’ compliance costs, net-income effect) and independent variables (e.g., age, gender, education) can be considered. Thus, our findings should be interpreted as the result of a first look at the relationship between standard setters and GAAP.
These caveats notwithstanding, the results described above are robust to numerous substantive and econometric controls, including controls for cross-sectional dependence of observations, auditor-specific effects, and aggregate market conditions. Additionally, we conduct a number of sensitivity tests, including (i) using research assistants’ (instead of the
Big N auditors’) evaluations of exposure drafts; (ii) assigning greater weight to FASB and SEC chairmen when calculating the average background characteristics of extant regulators (to assess if chairmen are more important in standard setting); and (iii) restricting our analysis to periods of economic growth (to assess the sensitivity of our findings to broader macroeconomic conditions). These results are discussed in Section 5.
Broadly, the evidence in this paper suggests individual standard setters have equilibrium effects on standard setting.
Kothari et al. (2010) summarize two theories to explain accounting standard setting: ‘‘capture’’ and ‘‘ideology.’’ Under capture theory, constituent lobbying determines standard-setting outcomes since regulators are ‘‘captured’’ by their special-interest constituents; under ideology theory, constituent lobbying is only one input to standard setting, which is also influenced by regulators’ ideologies. If accounting standard setting is more aptly described by ideology theory, one would expect to see the systematic impact of regulators’ characteristics in accounting standards, as we find. However, empirically it is difficult to rule out ‘‘capture’’ because the selection of regulators is itself a political process, which may be beholden to special interests. For example, our findings associating the growing proportion of FASB members from financial services to fair-value standards can be explained by the growth of the financial-services sector over our sample period: changing political economies associated with the growth of finance may have resulted in the increased proportion of finance-industry veterans on the FASB, who in turn proposed fair-value standards. Going forward, a research program in this area that draws on our initial look at the question can provide additional insights into the role of individual regulators and special-interest politics on the nature of accounting regulation.
The rest of the paper is organized as follows. Section 2 lays out the motivation for our research-design choices and discusses associated limitations. Section 3 describes the construction of variables and develops associated hypotheses.
Section 4 discusses descriptive statistics and the multivariate regression strategy. Section 5 presents and interprets the multivariate results, including robustness tests. Section 6 concludes.

2. Motivating research-design choices
2.1. Which dependent variables?
To empirically assess the role of standard setters in standard setting, we require a reasonable and parsimonious metric to evaluate proposed standards. The analogous literature that explores the role of individual managers in firm policies generally employs explicit performance and governance metrics such as earnings, stock returns, disclosure standards, and accounting quality.4 Such obvious metrics are not applicable to our setting. In evaluating standards, we use ‘‘reliability’’ and ‘‘relevance’’ as discussed above. ‘‘Reliability’’ and ‘‘relevance’’ are widely viewed as being among accounting’s
‘‘fundamental qualitative characteristics’’ by both academics (e.g., Stickney et al., 2010, p. 765; Dyckman et al., 2011) and

4

See, for example, Bamber et al. (2010), Dyreng et al. (2010), and Ge et al. (2010).

A. Allen, K. Ramanna / Journal of Accounting and Economics 55 (2013) 66–90

69

the FASB (e.g., FASB, 1978, 1980).5,6 In addition to ‘‘reliability’’ and ‘‘relevance,’’ there are likely other possible metrics to evaluate accounting standards, including ‘‘comparability,’’ ‘‘consistency,’’ and whether the standards are income increasing. In this sense, there is considerable scope for additional research along the lines we have pursued.
2.2. Which independent variables?
In selecting the characteristics of FASB members and SEC commissioners to study, we are motivated by prior politicaleconomy research in this area. Dal Bo (2006), in a recent review, notes that empirical research on the role of regulators in regulation, while (p. 215) ‘‘well short of abundant,’’ has largely focused on regulators’ professional characteristics – particularly, industry backgrounds – and regulators’ political party affiliation. In addition, he points to empirical work on the role of regulators’ terms-in-office on regulatory outcomes. Given the exploratory nature of our study, we focus on these independent variables.
On industry backgrounds, ex ante, we have a broad choice of industry classifications to organize the data (e.g., SIC codes).
However, given the limited number of FASB members (n¼39) and SEC commissioners (n¼41) in our sample period, we are unable to use such broad-based industry classifications. Prior empirical research on regulators’ industry backgrounds has focused on industries most closely associated with those regulations (e.g., Cohen, 1986, studies whether Federal Communications Commission, FCC, regulators with broadcasting industry experience are more supportive of that industry). In our case of studying accounting standard setting, we identify auditing and financial services as the most closely associated industries. We focus on backgrounds in auditing because accounting and auditing are joint products in financial reporting and because of the historical evidence on the close input of the audit industry in standard setting (e.g., Watts and Zimmerman, 1982, 1983). We focus on backgrounds in financial services because the financial services industry is a front-line intermediary in using accounting information. This includes investment management, which uses accounting information on the buy side, and investment banking, which uses accounting information on the sell-side. Thus, we expect an investigation of standard setters with backgrounds in auditing and financial services to provide a useful lens into standard setting.7
2.3. Limitations of the research design
We attempt to provide some empirical evidence on the role of standard setters in standard setting. Such evidence can complement existing findings on the role of constituent comment-letter lobbying and congressional intervention in standard setting (see Kothari et al., 2010, for a recent review). Empirically, we focus on the association between standard setters and the exposure drafts they propose. Exposure drafts appear prior to direct comment-letter lobbying and thus provide a relatively clean setting (relative to final standards) to examine the role of standard setters. Of course, constituent lobbying can influence the exposure draft process as well, but such ex-ante lobbying is difficult to observe, and our research design does not address its possible effects on standard setting. Further, it is possible that the selection of regulators to the FASB and SEC is itself a function of constituent lobbying. Such lobbying, in turn, is likely driven by extant economic and political circumstances, for example, macroeconomic conditions, globalization (e.g., growth of IFRS), the rise of the financial services sector, or the rise of information technologies.8 While our research design allows us to infer a role for standard setters in standard setting, it does not allow us to conclusively establish whether this role derives from some intrinsic ideology of regulators or from prevailing political economies.
3. Variable measurement and hypotheses
3.1. Dependent variables: decreased ‘‘reliability’’ and increased ‘‘relevance’’
To evaluate the FASB exposure drafts in our sample period independently of researcher judgment, we rely on two separate methods. First, we examine relevant comment letters filed by the Big N auditors. Second, we use two research
5
The classification of standards as along ‘‘reliability’’ and ‘‘relevance’’ can also be related to research on the demands of debtholders versus equityholders on financial reporting practices. Debtholders are usually seen as demanding ‘‘reliability’’ (e.g., Watts, 2003), while equityholders are seen as demanding either ‘‘relevance’’ (e.g., Barth, 2006) or both ‘‘reliability’’ and ‘‘relevance’’ (e.g., LaFond and Watts, 2008).
6
Recently, the FASB modified its conceptual framework to move away from ‘‘reliability’’ towards ‘‘representational faithfulness.’’ This change was likely made (at least in part) due to criticisms that the FASB was undermining the ‘‘reliability’’ of accounting standards (e.g., Watts, 2003). In response to these criticisms, some FASB members argued that ‘‘reliability’’ had been ‘‘misunderstood’’ to mean ‘‘verifiability;’’ the concept of ‘‘representational faithfulness’’ was advanced to replace ‘‘reliability’’ (e.g., Schipper, 2005). The change was proposed at a joint FASB-IASB board meeting on May 25, 2005, and the change was introduced into the conceptual framework in 2010. Since the change was initiated towards the end of our sample period, and went into effect after our sample period, we use ‘‘reliability,’’ not ‘‘representational faithfulness,’’ in our analyses.
7
In addition to tenure lengths, industry backgrounds, and party affiliation, it is possible that other characteristics of FASB and SEC regulators also matter. For example, in the context of studying the idiosyncratic styles of CEOs, Bertrand and Schoar (2003) examine whether managers’ age, gender, and education matter. In unreported tests, we examine whether such characteristics of FASB and SEC regulators are systematically associated with the standards they propose. We also test whether the regulators’ backgrounds in academia and government systematically vary with their proposals. The results are inconclusive.
8
On the role of macroeconomics on regulation, Bertomeu and Magee (2011) propose a model where accounting regulators are subject to different political pressures during different stages of the economic cycle.

70

A. Allen, K. Ramanna / Journal of Accounting and Economics 55 (2013) 66–90

Table 1
The sample is the 157 proposed SFAS issued between 1973 and 2007 on which the Big N auditors filed comment letters.
Era

Big 8

Big 6

Big 5

Big 4

Period
Audit firms

1973–1989
Arthur Anderson

1989–1998
Arthur Anderson

1998–2002
Arthur Anderson

2002–2007

Ernst & Young

Ernst & Young

Ernst & Young

Deloitte Touche

Deloitte Touche

Deloitte Touche

KPMG
PWC

KPMG
PWC

10
0

19
0

10

19

Arthur Young
Ernst & Whinney/Ernst & Ernst
Touche Ross
Deloitte, Haskin & Sells

)
)

Peat Marwick
Coopers Lybrand
Price Waterhouse
# of SFAS issued
# SFAS w/zero CLs
Remaining sample

KPMG
Coopers Lybrand
Price Waterhouse

104
4
SFAS 89, 98, 101, 103
100

30
2
SFAS 109, 124
28

)

The four proposed SFAS from 1973–1989 with no comment letters from the Big 8 are: SFAS 89 (which made supplementary information of price-level information voluntary); SFAS 98 (accounting for sale-leasebacks); SFAS 101 (disclosure issues in certain regulated entities); and SFAS 103 (resetting the effective date of another standard). The two proposed SFAS from 1989–1998 with no comment letters from the Big 6 are: SFAS 109 (accounting for income taxes) and SFAS 124 (accounting for certain investments held by non-profits).

assistants who are blind to the objectives of the study to manually assess the exposure drafts (this process is described later in the sub-section).
The key advantage to using Big N auditors’ comment letters is that they provide a consistent and contemporaneous source of exposure-draft evaluations. The evaluations are consistent in that the Big N auditors comment on a large majority of exposure drafts in our sample period, so we do not have to rely on evaluations from disparate sources. The evaluations are contemporaneous in that the letters do not suffer from hindsight bias. Moreover, Big N auditors are sophisticated consumers of accounting standards, so we expect their evaluations to have information content.
The changing industrial organization of the U.S. auditing oligopoly means that our set of ‘‘Big N auditors’’ begins with the
‘‘Big 8’’ in 1973 and ends with the ‘‘Big 4’’ in 2007. Table 1 provides a timeline of the changing dynamics of the U.S. audit industry. There are 170 distinct FASB exposure drafts that became 163 distinct SFAS in our sample period, 1973–2007. The absence of Big N auditor comments letters on six SFAS over that period decreases our sample size to 157 SFAS (Table 1 provides details on the SFAS without comment letters). These 157 SFAS can be traced back to 149 distinct exposure drafts (several exposure drafts resulted in multiple SFAS). There are collectively 908 unique comment letters by the Big N auditors on the 149 exposure drafts. We obtain paper copies of these comment letters from the FASB archives in Norwalk, Connecticut, and then digitize the comment letters using a combination of optical character recognition software and manual transcription. The digitized letters are then analyzed for contextually relevant occurrences of word stems ‘‘relevan’’ and ‘‘reliab’’ to create our auditor-based measures of the exposure drafts’ impact on decreased ‘‘reliability’’ and increased ‘‘relevance’’ using a process described in Appendix A. Based on that process, we define two variables, inc_relv and dec_relb, intended to capture the intensity of auditors’ concerns that a proposed standard will increase ‘‘relevance’’ and decrease ‘‘reliability,’’ respectively. The variables inc_relv and dec_relb are defined as follows. For each Big N auditor comment letter ‘‘i’’ on a proposed SFAS ‘‘j’’: inc_relvij ¼ 1À

WC_inc_relvij
WC ij

ð1Þ

dec_relbij ¼ 1À

WC_dec_relbij
WC ij

ð2Þ

In the above equations, WC_inc_relvij is the word count of the first instance of the word stem ‘‘relevan’’ used in the context of increased ‘‘relevance’’ in comment letter ‘‘i’’ on proposed SFAS ‘‘j;’’ WC_dec_relbij is the word count of the first instance of the word stem ‘‘reliab’’ used in the context of decreased ‘‘reliability’’ in comment letter ‘‘i’’ on proposed SFAS
‘‘j;’’ WC ij is the total word count of comment letter ‘‘i’’ on proposed SFAS ‘‘j.’’ In measuring inc_relv and dec_relb, we focus on the relative positions of the word stems ‘‘relevan’’ and ‘‘reliab’’ within a comment letter in order to get a measure of the relative importance of the auditors’ sentiments on ‘‘relevance’’ and ‘‘reliability.’’ The implicit assumption is that the stronger an auditor feels on ‘‘relevance’’ or ‘‘reliability,’’ the earlier the concept will be discussed in the comment letter.9
9
This assumption is consistent with the usual format of comment letters, which generally begin with an introductory paragraph highlighting key issues before tackling technical details in the body of the letter. Thus, if ‘‘relevance’’ and ‘‘reliability’’ are sufficiently important concerns for a letter writer, we expect the terms to be mentioned in the introductory paragraph, resulting in higher scores on inc_relv and dec_relb. In untabulated analysis, we tested the robustness of our results to this assumption by defining alternative binary dependent variables that are not sensitive to the relative location of substantive references to ‘‘relevance’’ and ‘‘reliability.’’ Results of this analysis are inconsistent with the concern that location-based construction may be driving our primary results.

A. Allen, K. Ramanna / Journal of Accounting and Economics 55 (2013) 66–90

71

By construction, inc_relv and dec_relb are confined to the range [0, 1] and are expected to increase in the strength of an auditor’s opinion of an exposure draft’s impact on increased ‘‘relevance’’ and decreased ‘‘reliability,’’ respectively.
The Big N auditors’ evaluations of exposure drafts are likely to be influenced by their private incentives: for example, if auditors are biased towards identifying decreased ‘‘reliability’’ over increased ‘‘relevance’’ because of litigation concerns, or if auditors are biased by the extant composition of their client base.10 In our tests, we do not expect these incentives to be correlated with our independent variables (i.e., the characteristics of FASB and SEC regulators), thus we expect these biases to add a scalar or a random variable to the regressand. Nevertheless, to mitigate the concern that auditor incentives can affect inferences in our tests, we supplement our auditor-based evaluations of the FASB exposure drafts with manual assessments by two research assistants who are blind to the objectives of the study but have extensive experience and practical familiarity with accounting. We use the standard dual-coder model in having the research assistants evaluate the exposure drafts. That is, the research assistants first independently evaluate each exposure draft based on a rubric discussed in Appendix B; then, the research assistants meet to resolve, if possible, instances of disagreement in their assessments. Of the 170 exposure drafts that became the 163 SFAS in our sample period, we are able to obtain, from the
FASB archives, copies of 145 exposure drafts representing 137 distinct SFAS. Copies of the remaining exposure drafts, all dating from the 1980s and before, are not readily available in the FASB archive.11 The 145 exposure drafts are manually assessed and then merged with the 149 exposure drafts for which we have auditor-based evaluations, yielding a common sample of 126 exposure drafts. In the subsequent section, we explore the correlation in our dependent variables across the auditor and research-assistant evaluations. The research assistants’ evaluations of decreased ‘‘reliability’’ and increased
‘‘relevance’’ are denoted Manual_dec_relb and Manual_inc_relv, respectively.
3.2. Independent variables: characteristics of FASB members and SEC commissioners
As noted earlier, our primary tests focus on the professional characteristics (i.e., tenure lengths and industry backgrounds) and political characteristics (i.e., party affiliations) of FASB members and SEC commissioners, because prior literature has studied these variables in the context of other regulators (e.g., Gormley, 1979; Cohen, 1986; Dal Bo, 2006,
Leaver, 2009). The first FASB members took office in 1973 (shortly after the FASB’s founding), and there have been 39 members on the board through December 2007. For each of these 39 members, we collect data on their length of tenure on the FASB, their backgrounds, if any, in auditing and financial services, and their political affiliations. In the same period, there have been 41 SEC commissioners, and we collect similar data on the commissioners.
Data on the duration of service on the board and the most recent employer prior to appointment to the board for FASB members are obtained primarily from two sources: (1) press notices issued by the FASB at a member’s initial appointment; and (2) the FASB’s annual informational bulletin, ‘‘Facts about FASB.’’ We create two non-exhaustive indicator variables to classify the members’ pre-FASB employers for further analysis: the first variable identifies whether a member worked for an audit firm prior to joining the board; the second whether the member worked for an investment bank or investment management firm. The equivalent data on the SEC commissioners’ tenure and professional background are obtained from the SEC’s historical archives, as well as from newspaper biographies of the commissioners (usually published upon the commissioners’ initial appointment).
In addition, we also build a database of the 39 FASB members’ political affiliations. Conceptually, we are interested in whether the members identify as Democrats or Republicans. Since members of the FASB are not explicit political appointees (they are appointed by the non-governmental Financial Accounting Foundation), the members’ party affiliations are not readily known. Thus, we infer members’ political identities by studying the history of their campaign contributions (if any). The Federal Election Commission (FEC) archives data on campaign contributions over $200 by U.S. individuals. Members contributing to the Democratic Party are coded as Democrats; those contributing to the Republican
Party are coded Republicans; while members not contributing to either party are not assigned a political identity.12 In the case of SEC commissioners, party affiliations are declared at or prior to appointment, so political identities need not be inferred from campaign contributions.
Our empirical tests are concerned with evaluating the influence of FASB and SEC regulators on exposure drafts.
Accordingly, for each exposure draft in our sample, we average the personal characteristics of all FASB members and SEC commissioners in office at the time. For example, for the exposure draft that became SFAS 106, we average across the seven FASB members and five SEC commissioners in office as of February 1989 (the date the exposure draft was issued) their lengths of service on the board (hereafter, Tenure FASB and Tenure SEC, respectively). Similarly, we compute across the members and commissioners, the proportion with prior employment in auditing (hereafter, % Auditor FASB and % Auditor
SEC, respectively), the proportion with prior employment in investment banking/investment management (hereafter,
% Financial FASB and % Financial SEC, respectively), the proportion contributing to the Democratic Party (hereafter,
10

Nelson et al. (2002) provide some survey-based evidence on auditors’ incentives.
In the case of several exposure drafts from the 1980s and before, only one paper copy exists at the FASB archive. The FASB publications department is in the process of digitizing all historic records, but the exposure drafts missing from our study were not available at the time we conducted the analysis. 12
To the extent that the FEC database is not comprehensive, our measure of political contributions is measured with error. However, we are not aware of any reason for the FEC excluding contributors over $200.
11

72

A. Allen, K. Ramanna / Journal of Accounting and Economics 55 (2013) 66–90

% Dem Donor FASB and % Democrat SEC, respectively). We also compute the proportion of FASB members contributing to the
Republican Party (hereafter, % Rep Donor FASB). An equivalent variable for SEC commissioners is obviated by the fact that the proportion of Republicans and Democrats in the SEC sample is collectively exhaustive. In subsequent empirical tests, we do not include % Auditor SEC because only one of the 41 SEC commissioners that served during our sample period worked for an audit firm prior to appointment to the commission.
The assumption implicit in averaging FASB members’ and SEC commissioners’ characteristics by exposure draft is that these documents represent the average position of the members and commissioners, respectively, in office at the time.13 In sensitivity tests described later, we examine the robustness of our results to assigning greater weight to FASB and SEC chairmen when calculating the average background characteristics.

3.3. Hypotheses development
Tenure FASB and Tenure SEC can be used to assess the impact of the average length of standard-setters’ terms on regulatory capture. In the classical economic theory of regulation (Stigler, 1971), longer terms (i.e., higher values of Tenure
FASB and Tenure SEC) signify greater ‘‘coziness’’ between regulators and the regulated, compromising regulatory outcomes.
However, Leaver (2009) develops and tests a model of regulation where longer terms insulate regulators from political pressure, thus improving regulatory outcomes. If decreased ‘‘reliability’’ is an undesirable accounting property, a positive association between Tenure FASB/Tenure SEC and our proxies for decreased ‘‘reliability’’ (i.e., dec_relb and Manual_dec_relb) is consistent with longer term-lengths compromising regulatory outcomes per Stigler’s theory. Similarly, if increased
‘‘relevance’’ augments accounting, a negative association between Tenure FASB/Tenure SEC and our proxies for increased
‘‘relevance’’ (i.e., inc_relv and Manual_inc_relv) is consistent with Stigler’s theory.
% Auditor FASB, % Financial FASB, and % Financial SEC can be used to assess the impact of FASB members’ and SEC commissioners’ industry backgrounds on accounting standard setting. Prior research in political science has shown that regulators tend to be more supportive of the industries they hail from (perhaps because they seek employment or consulting opportunities in those industries upon completion of their regulatory terms, e.g., Cohen, 1986). Given their role in assuring financial reports, and the substantial legal liability associated with this role (e.g., Kellogg, 1984; Watts, 2003), we expect auditors, ex ante, to be more sympathetic to standards promoting ‘‘reliability’’ at the expense of ‘‘relevance.’’
Moreover, if accounting regulators’ industry backgrounds matter in standard setting, FASB members and SEC commissioners with backgrounds in auditing will, ceteris paribus, be more likely associated with standards promoting ‘‘reliability’’
(potentially over ‘‘relevance’’). Thus, we predict negative coefficients between % Auditor FASB and our proxies for both decreased ‘‘reliability’’ and increased ‘‘relevance.’’ In contrast, ceteris paribus, we expect FASB members and SEC commissioners with backgrounds in financial services (defined as investment banking and investment management) to be more supportive of standards expected to improve accounting’s relevance through the use of fair values.14 Moreover, if the FASB is correct about its arguments linking fair values to increased ‘‘relevance’’ and, sometimes, decreased ‘‘reliability’’
(Johnson, 2005), regulatory backgrounds in financial services are likely to result in standards with such properties. Thus, we predict positive coefficients between % Financial FASB/% Financial SEC and our proxies for both increased ‘‘relevance’’ and decreased ‘‘reliability.’’
The empirical literature in political science has also considered the implications of regulators’ political affiliations on regulations, finding that Democratic regulators are on average less sympathetic to regulations benefiting corporate interests (Dal Bo, 2006). Extending this finding to accounting regulations, we can expect Democratic FASB and SEC regulators to be more sympathetic to standards that mitigate corporations’ information advantage over outsiders.
Evidence that corporations’ information advantage benefits managers, e.g., Healy and Whalen (1999), is germane to this prediction. Such benefits can engender anti-corporate sentiment (e.g., a perception that managers exploit information advantages to receive ‘‘excess compensation’’) that is more likely to resonate with Democrats. Linking Democrats’ relative focus on mitigating corporations’ information advantage to promoting ‘‘reliability’’ over ‘‘relevance’’ is trickier. On the one hand, increased ‘‘reliability’’ over ‘‘relevance’’ can mitigate corporations’ information asymmetry over outsiders because:
(1) ceteris paribus, managers are inherently more likely to emphasize good news over bad news (e.g., Kothari et al., 2009);
(2) regulatory solutions that are focused on mitigating corporations’ information advantage emphasize, on average, timely discussion of bad news (e.g., Watts, 2003); and (3) such solutions – conservatism and verifiability – result in greater
‘‘reliability’’ over ‘‘relevance’’ (e.g., Kothari et al., 2010, p. 256). On the other hand, firms themselves have incentives to prefer ‘‘reliability’’ to ‘‘relevance,’’ for example, corporations can benefit from accounting conservatism (e.g., through lower capital costs; LaFond and Watts, 2008; Zhang, 2008). Thus, the equilibrium relation between % Dem Donor FASB/% Democrat
SEC and our proxies for increased ‘‘relevance’’/decreased ‘‘reliability’’ is an empirical question. Ex ante, we have no
13
The maximum number of FASB members (SEC commissioners) at any given time during our sample period is seven (five). However, because new members do not immediately take office upon the resignation of another member, the size of the board can on occasion be less than seven (five).
14
Anecdotal evidence is consistent with this conjecture: e.g., the Investment Company Institute, the U.S. industry association for investment management firms, has strongly supported the use of fair-value accounting (ICI, 2008). Further, the Big 3 investment banks – Goldman Sachs, Morgan
Stanley, and Merrill Lynch – were all enthusiastic supporters of fair-value-based rules for mergers and acquisitions, including in subsequent goodwill impairment testing, during the standard-setting process for SFAS 141 and 142 (e.g., Ramanna, 2006).

A. Allen, K. Ramanna / Journal of Accounting and Economics 55 (2013) 66–90

73

Table 2
Summary statistics of and correlations between measures of decreased ‘‘reliability’’ and increased ‘‘relevance’’
The sample is based on 157 proposed SFAS issued between 1973 and 2007. inc_relv is an assessment that a proposed SFAS will increase accounting
‘‘relevance’’ as expressed by the Big 8/6/5/4 auditors (hereafter ‘‘Big N auditors’’) in their comment letters. dec_relb is an assessment that a proposed SFAS will decrease accounting ‘‘reliability’’ as expressed by the Big N auditors in their comment letters. Manual_inc_relv is an assessment that a proposed SFAS will increase ‘‘relevance’’ as determined by two independent reviewers. Manual_dec_relb is an assessment that a proposed SFAS will decrease ‘‘reliability’’ as determined by two independent reviewers. See Section 3 for details.
Panel A: Summary statistics
Variable

Mean

S.D.

Maximum

Minimum

0.07
0.04
0.31
0.65

dec_relb inc_relv Manual_dec_relb
Manual_inc_relv

Median
0.00
0.00
0.00
0.00

0.22
0.17
0.46
1.20

0.99
0.98
1.00
5.00

0.00
0.00
0.00
0.00

Panel B: Pearson correlations (Spearman above the diagonal)
Variable
(1)
(2)
(3)
(4)

dec_relb inc_relv Manual_dec_relb
Manual_inc_relv

(1)

(2)

(3)

(4)

1.000
0.225***
0.341***
0.502***

0.205**
1.000
0.147*
0.203**

0.347***
0.151*
1.000
0.609***

0.502***
0.238***
0.596***
1.000

Significance levels: (*) 10% level, (**) 5% level, and (***) 1% level using a 2 tailed test with S.E. clustered by SFAS.

prediction on % Rep Donor FASB. Nevertheless, we include this variable in our analysis because % Dem Donor FASB and % Rep
Donor FASB are not collectively exhaustive, and an analysis with % Rep Donor FASB can provide additional insights.15

4. Descriptive statistics and multivariate research design
4.1. Descriptive statistics
Appendix C provides a summary definition of all variables in the study. Table 2, Panel A, reports summary statistics for our measures of decreased ‘‘reliability’’ (dec_relb and Manual_dec_relb) and increased ‘‘relevance’’ (inc_relv and Manual_ inc_relv). The comment-letter-based statistics are for the 908 Big N auditor comment letters, and the manually assessed statistics are for the 145 exposure drafts examined by our research assistants. The mean value of dec_relb (Manual_ dec_relb) is 0.07 (0.31) and the median value is zero (zero). There is considerable variation in dec_relb (standard deviation is 0.22), and much of the variation is across (and not within) proposed standards. The maximum average value of dec_relb is observed on the exposure draft for SFAS 141R, Business Combinations. A major provision in this exposure draft was to allow an acquirer to recognize acquired net assets at their fair values, without regard to the cost of the acquisition.
Eliminating acquisition cost as the upper bound for net-asset-value recognition can introduce considerable subjectivity in financial reporting; thus it seems reasonable that SFAS 141R’s exposure draft received a high dec_relb score.
The mean value of inc_relv (Manual_inc_relv) is 0.04 (0.65). The median values of inc_relv and Manual_inc_relv are zero.
The standard deviation of inc_relv is 0.17 (over four times the mean), suggesting, as with dec_relb, that there is considerable variance among comment letters in their assessments on increased ‘‘relevance.’’ In unreported tests, we find that over two-thirds of this variation is across (and not within) proposed standards. The maximum average value of inc_relv for any given proposed SFAS is observed on the exposure draft for SFAS 159, The Fair Value Option for Financial
Assets and Financial Liabilities. SFAS 159 is a standard intended to ‘‘improve relevance of financial statements’’ (FAS, 2007), so the high inc_relv score on the exposure draft is consistent with inc_relv measuring increased ‘‘relevance.’’
In untabulated tests, we further examine the validity of inc_relv and dec_relb as measures of increased ‘‘relevance’’ and decreased ‘‘reliability,’’ respectively. Specifically, we randomly sampled 54 of the 908 big auditor comment letters (6%) to manually assess whether the letters expressed sentiments on increased ‘‘relevance’’ and decreased ‘‘reliability.’’ In all but five of the 54 sampled letters (9%), our evaluation agreed with inc_relv and dec_relb. In all five exceptions, inc_relv and dec_relb were coded zero because the actual word stems ‘‘relevan’’ and ‘‘reliab’’ were never used, while our manual assessment was that the letters did in fact express sentiments on increased relevance and/or decreased reliability (i.e., there are no false positives in the coding of inc_relv and dec_relb). The 9% misclassification refers exclusively to false negatives, which essentially result in a low power issue, biasing against finding results.
15
Two additional factors can confound predictions on political affiliation. First, the variables % Dem Donor FASB and % Rep Donor FASB are not collectively exhaustive because we cannot identify the political affiliation, if any, for FASB members in our sample who have never made campaign contributions in excess of $200. Second, the political distance between Democrats and Republicans on the FASB is unlikely to be as wide as that in the general population, because
FASB members are usually drawn from the relatively homogenous business community (including investors’ representatives).

74

A. Allen, K. Ramanna / Journal of Accounting and Economics 55 (2013) 66–90

Table 2, Panel B, reports Pearson (Spearman above the diagonal) correlation coefficients between the dependent variables discussed above. The p-values on the correlation coefficients are based on clustering at the SFAS level.16 The variables dec_relb and Manual_dec_relb (inc_relv and Manual_inc_relv) are significantly correlated with each other, r ¼ 0.341 (r ¼0.203), at the 95% confidence level or higher, suggesting that our comment-letter-based proxies and our manually assessed proxies capture similar concepts. The various proxies for decreased ‘‘reliability’’ and increased
‘‘relevance’’ are also significantly correlated with each other. This result is consistent with the FASB’s conception of
‘‘relevance’’ and ‘‘reliability’’ as trade-offs.
Table 3, Panel A, reports summary statistics for the FASB members’ and SEC commissioners’ personal characteristics. These measures constitute the set of explanatory variables in subsequent regression-based tests. The mean value of Tenure FASB is 4.2 and the median is 4.3, suggesting that, on average, an exposure draft is issued by a board with just over four years of individual service experience. In contrast, the mean and median values of Tenure SEC are 3.1 and 3.0, respectively, suggesting SEC commissioners are on average less experienced in their extant jobs. Fig. 1 plots the time series of Tenure FASB and Tenure SEC over the sample period. There does not appear to be any discernible time trend in average service experience on the two bodies.
On average, about 40% of FASB members were most recently employed in auditing (% Auditor FASB), while about 4% of
FASB members were most recently employed in investment banking/investment management (% Financial FASB). Fig. 2 plots the time series of these two variables over the 1973–2007 period: % Auditor FASB appears to have held steady over time, while % Financial FASB, which was zero through about the mid-1990s, appears to have increased to just under 30% in
2007. The average proportion of SEC commissioners most recently employed in financial services (% Financial SEC), at 15%, is higher than the corresponding FASB statistic. Fig. 2 also plots the trend in % Financial SEC, which appears to show considerable time series variation.
The average (median) proportion of FASB members contributing to the Democratic Party, % Dem Donor FASB, is 16.73%
(14%). The statistics are similar for % Rep Donor FASB at 18% (14%). Fig. 3 plots the time series of these two variables: % Dem
Donor FASB is higher than % Rep Donor FASB in the first few years of the FASB’s existence, while % Rep Donor FASB is higher in the period between 1995 and 2002. The average proportion of Democratic SEC commissioners (% Democrat SEC) is 45%, which indicates the average statistic for Republican SEC commissioners in about 55%. Overall, the partisan proportions for
SEC commissioners are higher than those for FASB members because the former are known with certainty and are collectively exhaustive in the sample. The time series variation in % Democrat SEC (Fig. 3) is predictable, given that commissioners are appointed by the U.S. president.17
Pearson correlations (Spearman above the diagonal) between the explanatory variables in Panel A, Table 3 are shown in Panel
B, Table 3. Statistical inferences are based on clustering by year. There are strong correlations between the background variables
(i.e., tenure and prior employment) and the personal politics variables among FASB members and SEC commissioners. For example, Tenure FASB is positively associated with % Rep Donor FASB (0.428) and negatively associated with % Dem Donor FASB
(À 0.521); % Auditor FASB is positively associated with both % Rep Donor FASB (0.216) and % Dem Donor FASB (0.519). Also, %
Financial SEC is negatively associated with % Democrat SEC. These correlations are consistent with findings in prior research involving the backgrounds and personal politics of FCC commissioners (e.g., Gormley, 1979; Cohen, 1986).

4.2. Multivariate research design
We are interested in assessing how our measures of FASB proposals’ impact on ‘‘reliability’’ and ‘‘relevance’’ vary with characteristics of standard setters. Accordingly, the dependent variables in our regressions are variously, dec_relb, inc_relv,
Manual_dec_relb, and Manual_inc_relv. In specifying the explanatory variables in these regressions, we follow prior research on regulators by examining the effect of professional and political characteristics both independently and jointly.
In the first set of regressions, we only include as explanatory variables the measures of FASB and SEC regulators’ professional characteristics: Tenure FASB, % Auditor FASB, % Financial FASB, Tenure SEC, and % Financial SEC. We do not include measures of the regulators’ political characteristics because of the high observed correlations between political and personal characteristics.18 Appropriately, results from such regressions must be interpreted as exploratory, not definitive. The formal specification for our first set of regressions is given by
DepVar ij ¼ f ðTenure FASBj ,% Auditor FASBj ,% Financial FASBj ,Tenure SEC j ,% Financial SEC j Þ

ð3Þ

pffiffiffiffiffiffiffiffiffiffi pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
That is, significance of correlation coefficients is computed using the t distribution as Prððrn nÀ2= ð1Àr2 ÞÞ4 tðnÀ2ÞÞ, where r is the Pearson correlation coefficient and n is the number of clusters (i.e., SFAS).
17
No more than three of the five SEC commissioners at any given time can belong to the same party; so, for example, a Democratic U.S. president cannot name five Democrats to the commission. Nevertheless, the proportion of SEC commissioners from the same party does sometimes exceed threefifths because of vacancies and time lags between appointments.
18
In his review of the literature on regulators’ impact on regulation, Dal Bo, 2006 notes that (p. 217) ‘‘although industry background seems to matter, it is not clear that it has a very strong effect once one considers the role of political affiliations.’’ He attributes this result to the high correlations, noting, for example, that in the case of the FCC, ‘‘no Democratic administration appointed a commissioner with [broadcasting] industry background’’ during the
1955–1974 period. In essence, there is no ex-ante theory that suggests either professional or political characteristics are more important than the other in explaining regulatory decisions, and, given the given the high correlations and small sample sizes in these regressions, there is some value to examining professional and political characteristics independently.
16

Table 3
Panel A
Summary statistics on explanatory variables
The sample is based on the 157 proposed SFAS issued between 1973 and 2007 on which the Big N auditors filed comment letters. Tenure FASB is an ED-level measure of the average tenure in years of all extant
FASB members. % Auditor FASB is an ED-level measure of the proportion of extant FASB members with most recent former employ in auditing. % Financial FASB is an ED-level measure of the proportion of extant
FASB members with most recent former employ in investment banking/investment management. %Rep Donor FASB is an ED-level measure of the proportion of extant FASB members making campaign contributions to the Republican Party or candidates. % Dem Donor FASB is an ED-level measure of the proportion of extant FASB members making campaign contributions to the Democratic Party or candidates.
Tenure SEC is an ED-level measure of the average tenure in years of all extant SEC commissioners. % Financial SEC is an ED-level measure of the proportion of extant SEC commissioners with most recent former employ in financial services. % Democrat SEC is an ED-level measure of the proportion of extant Democratic SEC commissioners.
Variable

Median

S.D.

Maximum

Minimum

FASB & SEC professional characteristics
Tenure FASB
% Auditor FASB
% Financial FASB
Tenure SEC
% Financial SEC

4.2
39.52%
4.35%
3.1
15.15%

4.3
42.86%
0.00%
3.0
20.00%

1.5
7.80%
8.05%
1.2
16.62%

6.7
57.14%
28.57%
6.2
66.67%

0.6
16.67%
0.00%
0.2
0.00%

FASB & SEC political characteristics
% Rep Donor FASB
% Dem Donor FASB
% Democrat SEC

18.01%
16.73%
44.99%

14.29%
14.29%
40.00%

12.37%
17.20%
20.22%

42.86%
66.67%
100.00%

0.00%
0.00%
0.00%

Panel B
Pearson correlations between explanatory variables (Spearman above the diagonal)
The sample is based on the 157 proposed SFAS issued between 1973 and 2007 on which the Big N auditors filed comment letters. Tenure FASB is an ED-level measure of the average tenure in years of all extant FASB members. % Auditor FASB is an ED-level measure of the proportion of extant FASB members with most recent former employ in auditing. % Financial FASB is an ED-level measure of the proportion of extant FASB members with most recent former employ in investment banking/investment management. %Rep Donor FASB is an ED-level measure of the proportion of extant FASB members making campaign contributions to the Republican Party or candidates. % Dem Donor FASB is an ED-level measure of the proportion of extant FASB members making campaign contributions to the Democratic Party or candidates. Tenure SEC is an ED-level measure of the average tenure in years of all extant SEC commissioners. % Financial SEC is an ED-level measure of the proportion of extant SEC commissioners with most recent former employ in financial services. % Democrat SEC is an ED-level measure of the proportion of extant Democratic SEC commissioners.
FASB/SEC professional characteristics
(1)
FASB/SEC Professional Characteristics
(1) Tenure FASB
1.000
(2) % Auditor FASB
À 0.276nnn
(3) % Financial FASB
À 0.035
(4) Tenure SEC
0.053
(5)% Financial SEC
À 0.122
FASB/SEC Political Characteristics
(6) % Rep Donor FASB
(7) % Dem Donor FASB
(8) % Democrat SEC

0.428nnn
À 0.521nnn
0.068

(2)

(3)

FASB/SEC political characteristics
(4)

(5)

(6)

(7)

À 0.255nnn
1.000
À 0.311nnn
0.282nnn
À 0.215nnn

0.022
À 0.405nnn
1.000
À 0.365nnn
0.403nnn

0.002
0.295nnn
À 0.356nnn
1.000
À 0.320nnn

À 0.155n
À 0.175nn
0.351nnn
À 0.306nnn
1.000

0.441nnn
0.229nnn
À 0.066
0.531nnn
À 0.222nnn

À 0.447nnn
0.787nnn
À 0.448nnn
0.093
À 0.245nnn

0.216nnn
0.519nnn
0.149n

À 0.067
À 0.320nnn
0.336nnn

0.529nnn
À 0.030
0.135n

À 0.275nnn
À 0.249nnn
À 0.341nnn

1.000
À 0.284nnn
0.284nnn

À 0.108
1.000
0.022

(8)

0.020
0.175nn
0.348nnn
0.193nn
À 0.146n

A. Allen, K. Ramanna / Journal of Accounting and Economics 55 (2013) 66–90

Mean

0.346nnn
0.052
1.000

Significance levels (n) 10% level, (nn) 5% level, and (nnn) 1% level using a 2 tailed test w/S.E. clustered by year.

75

76

A. Allen, K. Ramanna / Journal of Accounting and Economics 55 (2013) 66–90

Fig. 1. Average tenure of FASB members and SEC commissioners by proposed SFAS, 1973–2007. The sample is the 157 proposed SFAS issued between
1973 and 2007 on which the Big N auditors filed comment letters. Tenure FASB is an exposure draft (ED)-level measure of the average tenure in years of all extant FASB members. Tenure SEC is an ED-level measure of the average tenure in years of all extant SEC commissioners.

Fig. 2. Proportion of FASB members and SEC commissioners with prior employment in auditing and financial services. The sample is the 157 proposed
SFAS issued between 1973 and 2007 on which the Big N auditors filed comment letters. % Auditor FASB is an ED-level measure of the proportion of extant
FASB members with most recent former employ in auditing. % Financial FASB is an ED-level measure of the proportion of extant FASB members with most recent former employ in investment banking/investment management. % Financial SEC is an ED-level measure of the proportion of extant SEC commissioners with most recent former employ in financial services.

Fig. 3. Proportion of FASB members and SEC commissioners by political identity. The sample is the 157 proposed SFAS issued between 1973 and 2007 on which the Big N auditors filed comment letters. %Rep Donor FASB is an ED-level measure of the proportion of extant FASB members making campaign contributions to the Republican Party or candidates. % Dem Donor FASB is an ED-level measure of the proportion of extant FASB members making campaign contributions to the Democratic Party or candidates. % Democrat SEC is an ED-level measure of the proportion of extant Democratic SEC commissioners. In Eq. (3), ‘‘i’’ is a big auditor comment letter and ‘‘j’’ is an exposure draft. Standard errors in estimating Eq. (3) are clustered two-ways, by proposed SFAS and big auditor (using the method described in Petersen, 2009). We estimate two specifications of Eq. (3) (and all subsequent regressions), one with Big N auditor fixed effects and one without.

A. Allen, K. Ramanna / Journal of Accounting and Economics 55 (2013) 66–90

77

The Big N auditor fixed effects specifically identify the ‘‘Big 5’’ auditors; thus for example, a comment letter by Touche
Ross from the period preceding the establishment of Deloitte & Touche will be identified by a Deloitte & Touche fixed effect. We test for the association between our dependent variables and the FASB and SEC regulators’ political characteristics
(i.e., % Dem Donor FASB, % Rep Donor FASB, % Democrat SEC) in a second set of regressions. The formal specification for our second set of regressions is given by
DepVar ij ¼ f ð% Rep Donor FASBj ,% Dem Donor FASBj ,% Democrat SEC j Þ

ð4Þ

In Eq. (4), DepVar and the subscripts ‘‘i’’ and ‘‘j’’ are as defined in Eq. (3). Standard error clusters are also as described earlier. In a final set of regressions, we include all independent variables described in Eqs. (3) and (4). Coefficients in all regressions are estimated using ordinary least squares (OLS). We report results both with and without controls for two market-based variables: the annual value-weighted market return (VWRETD) and the standard deviation of the daily value-weighted market return (sd_VWRETD) in the twelve months preceding the issuance of a proposed SFAS.
5. Multivariate results
5.1. Results using Big N auditors’ comment letters
Table 4 reports OLS estimation results where the measure of decreased ‘‘reliability’’ from auditor comment letters
(dec_relb) is the dependent variable. There are seven columns to Table 4. In the first three columns, FASB members’ and SEC commissioners’ professional characteristics are the explanatory variables (as in Eq. (3)); in columns four to six, FASB members’ and SEC commissioners’ political affiliations are the explanatory variables (as in Eq. (4)); the seventh column

Table 4
OLS regression of dec_relb on the characteristics of FASB members and SEC commissioners.
Sample is 908 big auditor comment letters written on 149 exposure drafts that became 157 SFAS issued between 1973 and 2007. dec_relb is an assessment that a proposed SFAS will decrease accounting ‘‘reliability’’ as expressed by the Big N auditors in their comment letters. See Section 3 for details. Tenure FASB is an ED-level measure of the average tenure in years of all extant FASB members. % Auditor FASB is an ED-level measure of the proportion of extant FASB members with most recent former employ in auditing. % Financial FASB is an ED-level measure of the proportion of extant FASB members with most recent former employ in investment banking/investment management. %Rep Donor FASB is an ED-level measure of the proportion of extant FASB members making campaign contributions to the Republican Party or candidates. % Dem Donor FASB is an ED-level measure of the proportion of extant FASB members making campaign contributions to the Democratic Party or candidates. Tenure SEC is an ED-level measure of the average tenure in years of all extant SEC commissioners. % Financial SEC is an ED-level measure of the proportion of extant SEC commissioners with most recent former employ in financial services. % Democrat SEC is an ED-level measure of the proportion of extant Democratic SEC commissioners. The market variables are
VWRETD and sd_VWRETD. VWRETD is the annual value-weighted market return (from CRSP) for the 12 months directly preceding the month in which a proposed SFAS was issued. sd_VWRETD is the standard deviation of daily VWRETD for the 12 months directly preceding the month in which a proposed
SFAS was issued. Figures in italics and parentheses are standard errors.
Professional
(1)
Tenure FASB
% Auditor FASB
% Financial FASB
Tenure SEC
% Financial SEC

0.0086
(0.0065)
À 0.1824
(0.2119)
0.6668nnn
(0.2047)
0.0196n
(0.0106)
0.1804
(0.1108)

(2)

Political
(3)

n

0.0137
(0.0077)
À 0.1597
(0.2099)
0.6672nnn
(0.1994)
0.0216n
(0.0110)
0.2101n
(0.1254)

(1)

(1)

0.0140
(0.0077)
À 0.1679
(0.2069)
0.6438nnn
(0.2024)
0.0220nn
(0.0110)
0.2075n
(0.1258)

No
No
SFAS, Auditor

Yes
No
SFAS, Auditor

Yes
Auditor
SFAS, Auditor

À 0.0516
(0.1270)
À 0.2540nnn
(0.0764)
0.0703
(0.0793)
No
No
SFAS, Auditor

908
0.1013

908
0.1067

908
0.1233

908
0.0383

% Dem Donor FASB
% Democrat SEC

No. Obs.
R2

(3)

n

% Rep Donor FASB

Market vars
Fixed effects
S.E. cluster

(2)

Joint

Significance levels (n) 10% level, (nn) 5% level, (nnn) 1% level using a 2 tailed test.

À 0.07579
(0.1144)
À 0.26618nnn
(0.0809)
0.080215
(0.0845)
Yes
No
SFAS, Auditor

À 0.0686
(0.1143)
À 0.2615nnn
(0.0787)
0.0794
(0.0849)
Yes
Auditor
SFAS, Auditor

0.0129
(0.0148)
À 0.1943
(0.2142)
0.5506nn
(0.2211)
0.0202
(0.0128)
0.2393n
(0.1381)
0.0099
(0.1889)
À 0.0102
(0.0664)
0.0513
(0.0939)
Yes
Auditor
SFAS, Auditor

908
0.0412

908
0.0616

908
0.1245

78

A. Allen, K. Ramanna / Journal of Accounting and Economics 55 (2013) 66–90

reports the regression combining all explanatory variables. In the first and fourth columns, we do not include the marketbased variables, VWRETD and sd_VWRETD, as temporal economic controls; in all other columns, these variables are included. In addition, columns two and five do not include auditor fixed effects, whereas columns three, six, and seven do.
In the following discussion, we focus on the results from columns three, six, and seven, since these columns have the most exhaustive specifications, only discussing the other columns when inferences differ. All regressions in Table 4 use the sample of 908 comment letters. Standard errors in all regressions are clustered by Big N auditor and SFAS, and are robust to heteroskedasticity.
When professional characteristics are examined independently, we find both Tenure FASB and Tenure SEC are positively associated with decreased ‘‘reliability,’’ suggesting that longer terms of service on the FASB and SEC are associated with a perception of decreased accounting ‘‘reliability’’ (the coefficient on Tenure FASB is insignificant when market-based controls are excluded). If decreased ‘‘reliability’’ is an undesirable accounting property, this result is consistent with longer term-lengths compromising regulatory outcomes, per Stigler’s theory of regulation. To put the coefficients’ magnitudes in perspective, the implication from column (3) is that a one standard deviation increase in FASB tenure (SEC tenure) is associated with a decrease in
‘‘reliability’’ that is about 30% (38%) of the mean dec_relb value. We also find evidence that % Financial FASB and % Financial SEC are positive and significant predictors of FASB proposals perceived as decreasing accounting ‘‘reliability.’’ A one standard deviation increase in % Financial FASB (% Financial SEC) is associated with a decrease in ‘‘reliability’’ that is about 74% (49%) of the mean dec_relb value. This evidence is consistent with the proposition that a prior career in investment banking/investment management predisposes standard setters to produce standards that deemphasize accounting ‘‘reliability.’’ Contrary to our expectations, we find no evidence in Table 4 linking % Auditor FASB and decreased ‘‘reliability.’’
When political characteristics are examined independently, the coefficient on % Dem Donor FASB is significant and negative in explaining dec_relb. The implication from column (6) is that a one standard deviation increase in % Dem Donor
FASB is associated with an increase in ‘‘reliability’’ that is about 64% of the mean dec_relb value. The evidence suggests that increased proportional representation of Democrats on the FASB is associated with the production of standards that are viewed as increasing accounting ‘‘reliability.’’ We do not find a similar result with the proportion of Democrats on the SEC.
In combining all explanatory variables in column (7), only the results on % Financial FASB and % Financial SEC are statistically significant. This result is consistent with prior studies that combine regulators’ professional and political characteristics, where high correlations between these variables and the small population size are seen to confound statistical inferences (Dal Bo, 2006).
However, in unreported tests we find variance inflation factors from this regression are inconsistent with severe multicollinearity suggesting that, for our sample, financial services affiliation is the overriding explanatory variable.
Table 5 reports OLS estimation results where the measure of increased ‘‘relevance’’ from auditor comment letters
(inc_relv) is the dependent variable. Table 5 is otherwise identical to Table 4 in all respects. As in Table 4, we focus on discussing results from columns three, six, and seven of Table 5. When professional characteristics alone are the explanatory variables, we find only % Financial FASB is a positive and significant predictor of FASB proposals perceived as increasing accounting ‘‘relevance.’’ In column (3), one standard deviation increase in % Financial FASB is associated with an increase in ‘‘relevance’’ that is about 73% of its mean value. This evidence is consistent with the proposition that a prior career in investment banking/investment management predisposes standard setters to produce standards that increase accounting ‘‘relevance.’’ When political characteristics alone are the explanatory variables, we find % Dem Donor FASB is significant and negative in explaining inc_relv. In column (6), one standard deviation increase in % Dem Donor FASB is associated with a decrease in ‘‘relevance’’ that is about 65% of the mean inc_relv value. Column (6) also reveals a statistically negative association between % Rep Donor FASB and proposals perceived as increasing ‘‘relevance.’’ We are not aware of a theory to interpret this result. In combining all explanatory variables in column (7) of Table 5, only the coefficient on % Financial FASB is statistically significant.
To summarize the key findings from Tables 4 and 5: across tests using auditor comment letters, the data are consistent with the proposition that a prior career in financial services predisposes FASB standard setters to favor accounting
‘‘relevance’’ over ‘‘reliability.’’
5.2. Results using manual assessments of exposure drafts
Our primary comment-letter-based measures of decreased ‘‘reliability’’ and increased ‘‘relevance’’ are sensitive to auditors’ distinct incentives, which may be endogenous to our explanatory variables. Accordingly, we use manual assessments by two research assistants, as discussed in Section 3 and Appendix B, as alternative dependent variables
(Manual_dec_relb and Manual_inc_relv) to address this concern.
Table 6 Panel A presents the descriptive statistics for: (A) the exposure drafts common to both our manual and comment-letter sample (n ¼126); (B) the sub-sample of exposure drafts for which we only have manual assessments
(n ¼19); and (C) the sub-sample of exposure drafts for which we only have auditor comment letters (n ¼23). Using a twosample differences-in-means t-test we compare the average values of explanatory variables across the three groups. Of particular note, Tenure FASB and Tenure SEC are significantly lower in sub-sample (C), while % Dem Donor FASB is significantly higher. These differences are largely caused by data availability for the manually assessed sub-sample. That sub-sample (columns (A) and (B)) excludes several exposure drafts from the early years of the FASB (1980s and before), a period characterized by lower values for Tenure FASB and Tenure SEC and higher values for % Dem Donor FASB, as shown in Figs. 1–3.

A. Allen, K. Ramanna / Journal of Accounting and Economics 55 (2013) 66–90

79

Table 5
OLS regression of inc_relv on the characteristics of FASB members and SEC commissioners.
Sample is 908 big auditor comment letters written on 149 exposure drafts that became 157 SFAS issued between 1973 and 2007. inc_relv is an assessment that a proposed SFAS will increase accounting ‘‘relevance’’ as expressed by the Big 8/6/5/4 auditors (hereafter ‘‘Big N auditors’’) in their comment letters. See Section 3 for details. Tenure FASB is an ED-level measure of the average tenure in years of all extant FASB members. % Auditor FASB is an ED-level measure of the proportion of extant FASB members with most recent former employ in auditing. % Financial FASB is an ED-level measure of the proportion of extant FASB members with most recent former employ in investment banking/investment management. %Rep Donor FASB is an ED-level measure of the proportion of extant FASB members making campaign contributions to the Republican Party or candidates. % Dem Donor FASB is an EDlevel measure of the proportion of extant FASB members making campaign contributions to the Democratic Party or candidates. Tenure SEC is an ED-level measure of the average tenure in years of all extant SEC commissioners. % Financial SEC is an ED-level measure of the proportion of extant SEC commissioners with most recent former employ in financial services. % Democrat SEC is an ED-level measure of the proportion of extant Democratic SEC commissioners. The market variables are VWRETD and sd_VWRETD. VWRETD is the annual value-weighted market return (from CRSP) for the 12 months directly preceding the month in which a proposed SFAS was issued. sd_VWRETD is the standard deviation of daily VWRETD for the 12 months directly preceding the month in which a proposed SFAS was issued. Figures in italics and parentheses are standard errors.
Professional
(1)
Tenure FASB
% Auditor FASB
% Financial FASB
Tenure SEC
% Financial SEC

0.0015
(0.0031)
À 0.1852
(0.1362)
0.3388n
(0.1973)
0.0025
(0.0037)
0.0927
(0.0717)

(2)
À 0.0013
(0.0031)
À 0.2096
(0.1410)
0.3516n
(0.1935)
0.0021
(0.0038)
0.0708
(0.0777)

Political
(3)

No
No
SFAS, Auditor

Yes
No
SFAS, Auditor

Yes
Auditor
SFAS, Auditor

À 0.1722n
(0.1017)
À 0.1608nnn
(0.0609)
À 0.0043
(0.0884)
No
No
SFAS, Auditor

908
0.0594

908
0.0634

908
0.0681

908
0.03

% Dem Donor FASB
% Democrat SEC

No. Obs.
R2

(2)

(3)

(1)

À 0.0014
(0.0031)
À 0.2105
(0.1412)
0.3634n
(0.1937)
0.0019
(0.0039)
0.0720
(0.0782)

% Rep Donor FASB

Market vars
Fixed effects
S.E. cluster

(1)

Joint

À 0.158942n
(0.0921)
À 0.1497nn
(0.0598)
À 0.01291
(0.0874)
Yes
No
SFAS, Auditor

À 0.1624n
(0.0942)
À 0.1520nn
(0.0611)
À 0.0122
(0.0881)
Yes
Auditor
SFAS, Auditor

À 0.0001
(0.0051)
À 0.1150
(0.1042)
0.5514nnn
(0.1933)
0.0056
(0.0053)
À 0.0052
(0.0676)
À 0.0441
(0.1275)
À 0.0432
(0.0627)
À 0.1090
(0.0963)
Yes
Auditor
SFAS, Auditor

908
0.032

908
0.0349

908
0.0775

Significance levels (n) 10% level, (nn) 5% level, (nnn) 1% level using a 2 tailed test.

Table 6 Panel B presents OLS estimation results where Manual_dec_relb and Manual_inc_relv are the dependent variables. There are six columns to Table 6 Panel B: Manual_dec_relb is the dependent variable for the first three columns,
Manual_inc_relv for the next three. The first column for each dependent variable includes only regulators’ professional characteristics as independent variables; the second column for each dependent variable includes only regulators’ political characteristics as independent variables; the final column for each dependent variable includes both professional and political characteristics. In all columns, we include auditor fixed effects and the market-based controls. Each regression is based on 126 observations, one for each exposure draft where both auditor comment letters and manual assessments are available. Standard errors are heteroskedasticity robust.
The results in Table 6 Panel B show that % Financial FASB is a significant determinant of both Manual_dec_relb and
Manual_inc_relv, which is consistent with regression results using auditor comment letters (Tables 4 and 5). As in Table 4, we find a significant negative coefficient on % Dem Donor FASB in regressions on Manual_dec_relb that include only political variables; we do not find a similar result on Manual_inc_relv. In contrast to Table 4, in Table 6 Panel B we do not find significant coefficients on Tenure FASB, Tenure SEC, and % Financial SEC in regressions on Manual_dec_relb. The non-results on the tenure variables are likely explained by the exclusion of several exposure drafts in Table 6 Panel B regressions due to data limitations, as discussed above.
Overall, to summarize the key findings from Tables 4–6: across tests using auditor comment letters and manual assessments of exposure drafts, the data are consistent with the proposition that a prior career in financial services predisposes FASB standard setters to favor accounting ‘‘relevance’’ over ‘‘reliability.’’ Our coding rubric for the manual assessment of exposure drafts’ focus on ‘‘relevance’’ over ‘‘reliability’’ relies on the use of fair-value methods in these proposals (see Appendix B for details). Thus, the key finding on financial services affiliation can be explained, in part, as the tendency of regulators with a financial services background to propose standards that use fair-value methods in recognition and disclosure. When combined with the descriptive evidence from Fig. 2 and Table 3, which shows an increase in the proportion of FASB members from financial services from the mid-1990s through 2007, this result can provide a partial explanation for the growth of fair-value accounting.

80

A. Allen, K. Ramanna / Journal of Accounting and Economics 55 (2013) 66–90

Table 6
Panel A
Differences in means of explanatory variables across the comment-letter and manually assessed sub-samples.
Two-sample differences-in-means t-tests are performed on pairs of three distinct sub-samples. Sub-sample A is the 126 exposure drafts for which we have both manual assessments and auditor comment letters. Sub-sample B is the 19 exposure drafts for which we have manual assessments but no auditor comment letters. Sub-sample C is the 23 exposure drafts for which we have auditor comment letters but no manual assessments. Tenure FASB is an ED-level measure of the average tenure in years of all extant FASB members. % Auditor FASB is an ED-level measure of the proportion of extant
FASB members with most recent former employ in auditing. % Financial FASB is an ED-level measure of the proportion of extant FASB members with most recent former employ in investment banking/investment management. %Rep Donor FASB is an ED-level measure of the proportion of extant FASB members making campaign contributions to the Republican Party or candidates. % Dem Donor FASB is an ED-level measure of the proportion of extant
FASB members making campaign contributions to the Democratic Party or candidates. Tenure SEC is an ED-level measure of the average tenure in years of all extant SEC commissioners. % Financial SEC is an ED-level measure of the proportion of extant SEC commissioners with most recent former employ in financial services. % Democrat SEC is an ED-level measure of the proportion of extant Democratic SEC commissioners.
A

# of exposure drafts
Big N auditor comment letters available
Manual assessments available

B

B&C

C

C&A

4.335
0.389
0.064
3.018
0.188
0.191
0.126
0.453

Tenure FASB
% Auditor FASB
% Financial FASB
Tenure SEC
% Financial SEC
% Rep Donor FASB
% Dem Donor FASB
% Democrat SEC

A&B

4.277
0.396
0.038
3.028
0.137
0.173
0.140
0.464

nnn

2.907
0.419
0.000
2.427
0.076
0.087
0.405
0.389

nnn

126
Yes
Yes

19
No
Yes

n

nn nnn nnn nnn nn

nnn nnn nn

23
Yes
No

Significance levels (n) 10% level, (nn) 5% level, and (nnn) 1% level using a 2 tailed two-tailed t-test.
Panel B
OLS regression of Manual_dec_relb and Manual_inc_relv on the characteristics of FASB members and SEC commissioners
Sample is the 126 exposure drafts for which we have both auditor comment letters and manual assessments (See Table 6A). Manual_inc_relv is an assessment that a proposed SFAS will increase ‘‘relevance’’ as determined by two independent reviewers. Manual_dec_relb is an assessment that a proposed SFAS will decrease ‘‘reliability’’ as determined by two independent reviewers. See Section 3 for details. Tenure FASB is an ED-level measure of the average tenure in years of all extant FASB members. % Auditor FASB is an ED-level measure of the proportion of extant FASB members with most recent former employ in auditing. %
Financial FASB is an ED-level measure of the proportion of extant FASB members with most recent former employ in investment banking/investment management. %Rep Donor FASB is an ED-level measure of the proportion of extant FASB members making campaign contributions to the Republican Party or candidates. % Dem Donor FASB is an ED-level measure of the proportion of extant FASB members making campaign contributions to the Democratic Party or candidates. Tenure SEC is an ED-level measure of the average tenure in years of all extant SEC commissioners. % Financial SEC is an ED-level measure of the proportion of extant SEC commissioners with most recent former employ in financial services. % Democrat SEC is an ED-level measure of the proportion of extant Democratic SEC commissioners. The market variables are VWRETD and sd_VWRETD. VWRETD is the annual value-weighted market return (from CRSP) for the 12 months directly preceding the month in which a proposed SFAS was issued. sd_VWRETD is the standard deviation of daily VWRETD for the 12 months directly preceding the month in which a proposed SFAS was issued. Figures in italics and parentheses are standard errors.
Manual_dec_relb
(1)
Tenure FASB
% Auditor FASB
% Financial FASB
Tenure SEC
% Financial SEC

(2)

(3)

(1)

0.0551
(0.0448)
À 0.2959
(0.7059)
1.9236nnn
(0.6928)
0.0588
(0.0495)
0.0261
(0.3493)
À 0.7157
(0.5874)
À 0.1197
(0.3951)
À 0.2491
(0.2971)
Yes
No
Robust

0.0015
(0.0943)
À 2.1387
(1.4911)
5.8551nnn
(1.6954)
0.0467
(0.1008)
0.8158
(0.8384)

126
0.1482

0.0313
(0.0352)
À 0.8485
(0.5551)
1.3714nn
(0.5541)
0.0250
(0.0419)
0.2801
(0.3030)

Market Vars
Fixed effects
S.E.

Yes
No
Robust

À 0.7617n
(0.4181)
À 0.6918nn
(0.3165)
0.0833
(0.2231)
Yes
No
Robust

No. Obs.
R2

126
0.1253

126
0.0717

% Rep Donor FASB
% Dem Donor FASB
% Democrat SEC

Manual_inc_relv

Significance levels (n) 10% level, (nn) 5% level, and (nnn) 1% level using a 2 tailed test.

(2)

(3)

Yes
No
Robust

À 2.3263nnn
(0.8732)
À 1.2018
(1.0331)
À 0.0319
(0.5728)
Yes
No
Robust

0.0302
(0.1241)
À 1.6739
(1.8935)
8.7955nnn
(1.9482)
0.1058
(0.1292)
À 0.3722
(0.9230)
À 0.2012
(1.3590)
0.8250
(1.3185)
À 1.8942nnn
(0.6855)
Yes
No
Robust

126
0.2771

126
0.0696

126
0.3294

A. Allen, K. Ramanna / Journal of Accounting and Economics 55 (2013) 66–90

81

5.3. Robustness and sensitivity tests
With the small population of FASB and SEC regulators in our sample, there is a concern that one individual with an extremely strong personality can be driving the results described thus far. The analogous literature on managers and firm policies employs technologies around job-switching to address this concern (e.g., Bertrand and Schoar, 2003). Due to the unique nature of the task we study, i.e., standard setting not corporate management, we cannot employ these technologies. An alternative robustness test is a jackknifing procedure where we re-estimate all our regressions successively eliminating each regulator (and reconstructing all independent variables accordingly) to determine if any member was instrumental to our statistical inferences. Unreported results obtained from this procedure are inconsistent with the proposition that any one FASB member is instrumental to the factors previously identified as statistically significant: Across 39 jackknifed subsamples – each eliminating one FASB member – we find no cases where elimination of an individual from our sample changes the sign or significance of our primary results.19
The implicit assumption underlying construction of our independent variables is that an exposure draft represents the average position of all extant FASB members and SEC commissioners; however, it is possible that the chairmen of these groups have greater influence than other members. In unreported robustness tests, we examine the effects of assigning greater weight to FASB and SEC chairmen when calculating the average background characteristics of an extant board. In particular, we assign the background characteristics of FASB and SEC chairmen twice the weight of non-chair members.
While the choice of doubling the weight on chairmen is admittedly arbitrary, the objective of this test is simply to assess whether the relative importance of FASB chairmen subsumes the results shown earlier. All substantive results discussed in
Tables 4 and 5 are robust to the procedure described above.
Finally, as discussed earlier, it is possible that the selection of a set of regulators on the FASB and SEC in a given time period depends, at least in part, on more fundamental macroeconomic conditions. Accordingly, we study the sensitivity of our results to these conditions. In restricting the sample to periods of expansion in the U.S. economy (as identified by the NBER), financial services background and Democratic Party affiliations remain significant predictors of increased ‘‘relevance’’ (inc_relv) and decreased ‘‘reliability’’ (dec_relb), consistent with results reported in Tables 4 and 5. The only result from those tables not carrying through is the negative coefficient on % Rep Donor FASB on inc_relv for which we have no ex-ante prediction and which is not consistent across all specifications.
6. Conclusions
Motivated by an interest in broadening the understanding of accounting standard setting beyond the role of constituent comment-letter lobbying and congressional intervention, we examine the role of FASB and SEC regulators in the process.
Specifically, we examine how the professional and political characteristics of these regulators vary in the nature of exposure drafts proposed from 1973 to 2007. Because there is no obvious metric to evaluate the proposals, we rely principally on Big N auditors’ contemporaneous evaluations of the exposure drafts along dimensions of ‘‘reliability’’ and ‘‘relevance.’’ Our focus on
‘‘reliability’’ and ‘‘relevance’’ reflects our judgment on their importance to accounting, also evidenced in several leading accounting textbooks and in the FASB’s conceptual framework. The regulators’ professional characteristics we study are tenure, background in auditing, and background in financial services; the political characteristics are affiliation, if any, with the
Democratic and Republican parties. Our key finding is that FASB members with a prior professional affiliation with the financial services industry are more likely to propose standards that decrease ‘‘reliability’’ and increase ‘‘relevance,’’ partly due to their tendency to propose fair-value methods of measurement. Given that the proportion of FASB members from the financial services industry has increased from the mid-1990s to 2007, this finding can provide a partial explanation for the growth of fair-value accounting. We also find that FASB members affiliated with the Democratic Party are more likely to propose standards that increase ‘‘reliability’’ and decrease ‘‘relevance,’’ although only when excluding financial-services affiliation as an independent variable. Since our statistical inferences are based on a small population of FASB and SEC regulators, we conduct jackknifed sensitivity analyses: we find no evidence that any one regulator is driving inferences.
Broadly, the paper provides a first empirical look at an important feature in the political economy of U.S. GAAP: the role of regulators at the FASB and SEC. While our research design does not allow us to distinguish whether the documented role of regulators derives from some intrinsic ideology of these individuals or from more primitive selection effects that place these regulators in office, our study takes the first important step of examining the impact of individual standard setters on standard setting (in the spirit of Bertrand and Schoar’s analogous study of managers on firm policies). Our study highlights opportunities for work on the question of how accounting regulators are chosen, including issues such as whether there is a ‘‘revolving door’’ between standard setters and special-interest groups. Moreover, as accounting institutions worldwide reorganize in response to globalization, such research can have important practical implications in the area of regulatory design.20
19
The successive elimination of two FASB members in the jackknife procedure does turn, in some cases, the previously insignificant coefficient on % Auditor
FASB significantly negative (as predicted): one of these members has a financial services background, the other an auditing background. One implication is that our failure to find evidence on % Auditor FASB in the regression that includes all independent variables is driven by the influential effects of these members.
20
For example, in the past five years, both Canada and China have undertaken some revamping of their standard-setting institutions (e.g., Ramanna and Cheng, 2009; Ramanna et al., 2010). Further, in the U.S., between 2008 and 2010, the FASB has pared down and increased its membership from seven to five and back to seven, in order to ‘‘protect and maintain its efficiency’’ (FAF, 2008, 2010). Given the paucity of evidence to guide such structural changes, most, if not all of the institutional transformations have been ad hoc.

82

A. Allen, K. Ramanna / Journal of Accounting and Economics 55 (2013) 66–90

Appendix A. Details of the process for creating auditor-based measures of decreased ‘‘reliability’’ and increased ‘‘relevance’’
We use a custom-designed Perl script to analyze the Big N auditors’ comments letters. For each comment letter, the Perl program first identifies all instances of the word stems ‘‘relevan’’ and ‘‘reliab.’’ The program then outputs: (1) the exact position within the comment letter where a word stem of interest occurs (the position of a word stem is reported as its word count from the beginning of the document); (2) the entire sentence containing the identified word stem; and (3) the total word count for the letter.
Next, a research assistant (RA) trained in accounting principles, but blind to the intent of our study, manually examines both the first sentence referencing ‘‘relevan’’ and the first sentence referencing ‘‘reliab.’’ On each sentence, the RA determines whether the word stem in question is being used in: (1) a positive context, i.e., whether the letter is indicating that the proposed standard will increase ‘‘relevance’’/‘‘reliability;’’ (2) a negative context, i.e., whether the letter is indicating that the proposed standard will decrease ‘‘relevance’’/‘‘reliability;’’ or (3) a context that is irrelevant to the use of ‘‘relevance’’ and ‘‘reliability’’ as accounting principles. Examples of the RA’s assessments from actual sentences in the comment letters are below.

 Positive context: ‘‘We support the approach followed in the Exposure Draft and believe that application of those



standards will provide relevant and understandable information as well as an appropriate balance between comparability and flexibility.’’ Source: Arthur Andersen’s comment letter on proposed SFAS 117.
Negative context: ‘‘We also believe the Proposed Standard exacerbates the complexities of Statement 125 and permits recognition of revenue that cannot be reliably measured.’’ Source: Deloitte’s comment letter on proposed SFAS 140.
Irrelevant usage: ‘‘The auditor should familiarize himself with the relevant provisions of the partnership agreement.’’
Source: Arthur Andersen’s comment letter on proposed SFAS 102.

In instances where the research assistant identifies the comment letter’s first use of ‘‘relevance’’/‘‘reliability’’ as irrelevant to accounting principles, the RA proceeds to the second sentence containing the word stem in question. This process continues until the RA encounters either a positive or negative use of ‘‘relevance’’/‘‘reliability’’ or the RA determines that all uses of ‘‘relevance’’/‘‘reliability’’ in the comment letter are irrelevant to accounting principles.

Appendix B. Coding rubric for research-assistant-based measures of decreased ‘‘reliability’’ and increased ‘‘relevance’’
The research assistants were instructed to evaluate the exposure drafts recording their perspective on whether the underlying proposal would decrease ‘‘reliability,’’ where ‘‘reliability’’ is defined as per the FASB as, ‘‘The quality of information that assures that information is reasonably free from error and bias and faithfully represents what it purports to represent.’’ The resulting variable is a binary indicator denoted Manual_dec_relb. To obtain research assistants’ assessments of exposure drafts’ increased ‘‘relevance,’’ we rely on the following procedure: we asked the research assistants to score each exposure draft on the nature of its use of fair-value accounting. Our focus on ‘‘fair values’’ in measuring ‘‘increased relevance’’ is motivated by the
FASB viewing the former as resulting in the latter (e.g., Johnson, 2005). In particular, research assistants scored each exposure draft on a score of 0–5, with unit scores for each of the following: (1) the introduction of fair-value accounting for asset writedowns; (2) the introduction of fair-value accounting for asset recognition and remeasurement; (3) the introduction of fair-value accounting for liability recognition and remeasurement; (4) the recognition of fair-value changes in the income statement; and
(5) the required disclosure of fair-value amounts. The resulting count variable is denoted, Manual_inc_relv.
Assessing Manual_dec_relb and the components of Manual_inc_relv requires the exercise of professional judgment.
Accordingly, both research assistants employed for this task are seasoned professionals, with MBA degrees from topranked U.S. business schools (as per U.S. News rankings) and with combined industrial work experience in finance and accounting exceeding thirty years. We recruited both research assistants specifically to evaluate the FASB exposure drafts, and both were selected for their practical familiarity with accounting.
Of the 145 exposure drafts coded by the two research assistants, 105 received identical evaluations on Manual_dec_relb, while 114 received identical evaluations on Manual_inc_relv. On the exposure drafts with differing evaluations, the research assistants were able to resolve all differences in subsequent discussions. At no point in this process were the research assistants apprised of the study’s hypotheses or its independent variables. Research assistants were compensated on a flat hourly wage (i.e., no performance-based pay).

Appendix C
For a summary definition of all variables, see Table C1.

Appendix D. Dependent variable scores by exposure draft
For dependent variable scores by exposure draft, see Table D1.

A. Allen, K. Ramanna / Journal of Accounting and Economics 55 (2013) 66–90

83

Table C1
Variable definitions.
Variable

Description

Dependent variables inc_relv dec_relb
Manual_inc_relv
Manual_dec_relb

Assessment that a proposed SFAS will increase accounting ‘‘relevance’’ as expressed by the Big 8/6/5/4 auditors (hereafter
‘‘Big N auditors’’) in their comment letters. See Section 3.
Assessment that a proposed SFAS will decrease accounting ‘‘reliability’’ as expressed by the Big N auditors in their comment letters. See Section 3.
Assessment that a proposed SFAS will increase ‘‘relevance’’ as determined by 2 independent reviewers. See Section 3.
Assessment that a proposed SFAS will decrease ‘‘reliability’’ as determined by 2 independent reviewers. See Section 3.

FASB & SEC professional characteristics
Tenure FASB
Exposure draft (ED)-level measure of the average tenure in years of all extant FASB members
% Auditor FASB
ED-level measure of the proportion of extant FASB members with most recent former employ in auditing.
% Financial FASB
ED-level measure of the proportion of extant FASB members with most recent former employ in investment banking/ investment management
Tenure SEC
ED-level measure of the average tenure in years of all extant SEC commissioners
% Financial SEC
ED-level measure of the proportion of extant SEC commissioners with most recent former employ in financial services
FASB & SEC political characteristics
%Rep Donor FASB
ED-level measure of the proportion of extant FASB members making campaign contributions to the Republican party or candidates. % Dem Donor FASB
ED-level measure of the proportion of extant FASB members making campaign contributions to the Democratic party or candidates % Democrat SEC
ED-level measure of the proportion of extant Democratic SEC commissioners.
Other variables
VWRETD
sd_VWRETD

Annual value-weighted market return (from CRSP) for the 12 months directly preceding the month in which ED was issued.
Standard deviation of daily VWRETD (CRSP) for the 12 months directly proceeding the month in which an ED was issued

Table D1
Dependent variable scores by exposure draft.
SFAS

ED title

ED date

dec_relb

inc_ relv

Manual_ dec_relb

Manual_ inc_relv

SFAS001

Disclosure of Foreign Currency Translation
Information
Accounting for Research and Development Costs
Reporting Accounting Changes in Interim Financial
Statements: an amendment of APB Opinion No. 28
Reporting Gains and Losses and Extinguishments of
Debt: an amendment of APB Opinion No. 30
Accounting for Contingencies
Classification of Short-term Obligations Expected to Be
Refinanced: an amendment of ARB No. 43, Chapter 3,
Section A
Accounting and Reporting by Development Stage
Companies, Subsidiaries, Divisions and Other
Components
Accounting for the Translation of Foreign Currency
Transactions and Foreign Currency Financial
Statements
Accounting for Income Taxes–Oil and Gas Producing
Companies: an amendment of APB Opinions No. 11 and 23
Extension of ‘‘Grandfather’’ Provisions for Business
Combinations: An Amendment of APB Opinion No. 16
Accounting for Contingencies–Transition Method: An
Amendment of FASB Statement No.5
Accounting for Certain Marketable Securities
Accounting for Leases
Accounting for Leases: Revision of Exposure Draft
Issued August 26, 1975
Financial Reporting for Segments of a Business
Enterprise
Restructuring of Debt in a Troubled Loan Situation
Accounting by Debtors and Creditors for Troubled
Debt Restructurings
Prior Period Adjustments
Accounting for Leases–Initial Direct Costs: An
Amendment of FASB Statement No. 13

10/19/73

0.000

0.000

NA

NA

06/05/74
11/11/74

0.000
0.000

0.000
0.000

NA
0.000

NA
0.000

01/31/75

0.000

0.000

NA

NA

10/21/74
11/11/74

0.000
0.000

0.000
0.000

NA
NA

NA
NA

07/19/74

0.000

0.000

NA

NA

12/31/74

0.000

0.000

NA

NA

04/25/75

0.000

0.000

NA

NA

09/08/75

0.000

0.000

NA

NA

10/31/75

0.000

0.000

NA

NA

11/06/75
08/26/75
07/22/76

0.000
0.000
0.000

0.000
0.000
0.000

NA
NA
NA

NA
NA
NA

09/30/75

0.094

0.000

NA

NA

11/07/75
12/30/76

0.000
NA

0.000
NA

NA
1.000

NA
0.000

07/29/76
08/08/77

0.000
0.000

0.000
0.000

NA
0.000

NA
0.000

SFAS002
SFAS003
SFAS004
SFAS005
SFAS006

SFAS007

SFAS008

SFAS009

SFAS010
SFAS011
SFAS012
SFAS013
SFAS013
SFAS014
SFAS015
SFAS015
SFAS016
SFAS017

84

A. Allen, K. Ramanna / Journal of Accounting and Economics 55 (2013) 66–90

Table D1 (continued )
SFAS
SFAS018

ED title

Financial Reporting for Segments of a Business
Enterprise–Interim Financial Statements: An
Amendment of FASB Statement No. 14
SFAS019
Financial Accounting and Reporting by Oil and Gas
Producing Companies
SFAS020
Accounting for Forward Exchange Contracts/an amendment of FASB Statement No. 8
SFAS021
Suspension of the Reporting of Earnings per Share and
Segment Information by Nonpublic Enterprises: an amendment of APB Opinion No. 15
SFAS022/
Accounting for Leases: I Inception of the Lease: An
SFAS023 Amendment of FASB Statement No. 13 II Changes in the Provisions of Lease Agreements Resulting from
Refundings of Tax-Exempt Debt: an amendment of
FASB Statement No. 13
SFAS024
Reporting Segment Information in Financial
Statements That Are Presented With Another
Enterprise’s Financial Report: an amendment of FASB
Statement No. 14
SFAS025
Suspension of Certain Accounting Requirements for Oil and Gas Producing Companies: an amendment of FASB
Statement No. 19
SFAS026
Profit Recognition on Sales-Type Leases of Real Estate: an amendment of FASB Statement No. 13
SFAS027
Classification of Renewals or Extensions of Existing
Sales-Type or Direct Financing Leases: an amendment of FASB Statement No. 13
SFAS028
Accounting for Sales with Leasebacks: an amendment of FASB Statement No. 13
SFAS029
Determining Contingent Rentals
SFAS030
Disclosure of Information about Major Customers: an amendment of FASB Statement No. 14
SFAS031
Accounting for Income Taxes Related to U.K. Tax
Legislation Concerning Stock Relief
SFAS032
Specialized Accounting and Reporting Principles and
Practices in AICPA Industry Accounting Guides,
Industry Audit Guides, and Statements of Position: an amendment of APB Opinion No. 20
SFAS033
Financial Reporting in Units of General Purchasing
Power
SFAS033
Financial Reporting and Changing Prices
SFAS033
Constant Dollar Accounting: supplement to an exposure draft of a proposed Statement of Financial
Accounting Standards, Financial Reporting in Units of
General Purchasing Power
SFAS034
Capitalization of Interest Cost
SFAS035
Accounting and Reporting by Defined Benefit Pension
Plans
SFAS035
Accounting and Reporting by Defined Benefit Pension
Plans: revision of exposure draft issued April 14, 1977
SFAS036
Disclosure of Pension and Other Post-Retirement
Benefit Information: an amendment of APB Opinion
No. 8
SFAS037
Balance Sheet Classification of Deferred Income Taxes: an amendment of APB Opinion No. 11
Financial Reporting and Changing Prices: Specialized
SFAS039/
SFAS040/ Assets—a supplement to FASB Statement No. 33
SFAS041
SFAS042
Determining Materiality for Capitalization of Interest
Cost: an amendment of FASB Statement No. 34
SFAS043
Accounting for Compensated Absences
SFAS044
Accounting for Intangible Assets of Motor Carriers: an amendment of Chapter 5 of ARB 43 and an interpretation of APB Opinions 17 and 30
SFAS045
Accounting for Franchise Fee Revenue
SFAS046
Financial Reporting and Changing Prices: Motion
Picture Films; a supplement to FASB Statement No. 33
SFAS047
Disclosure of Guarantees, Project Financing
Arrangements, and Other Similar Obligations: an amendment of FASB Statement No. 5

ED date

dec_relb

inc_ relv

Manual_ dec_relb

Manual_ inc_relv

09/20/77

0.000

0.000

0.000

0.000

07/15/77

0.000

0.000

0.000

0.000

11/07/77

0.000

0.000

0.000

0.000

02/27/78

0.000

0.000

0.000

0.000

12/19/77

0.000

0.000

0.000

0.000

07/19/78

0.000

0.000

0.000

0.000

11/07/78

0.000

0.000

0.000

0.000

12/22/78

0.000

0.000

0.000

0.000

02/13/79

0.000

0.000

0.000

0.000

12/21/78

0.000

0.000

1.000

0.000

12/21/78
03/29/79

0.000
0.000

0.000
0.000

0.000
0.000

0.000
0.000

07/30/79

0.000

0.000

1.000

0.000

06/01/79

0.000

0.000

0.000

0.000

12/31/74

NA

NA

NA

NA

12/28/78
03/02/79

NA
0.086

NA
0.000

0.000
0.000

0.000
0.000

12/15/78
04/14/77

0.000
0.000

0.000
0.116

0.000
1.000

0.000
4.000

07/09/79

NA

NA

1.000

3.000

07/12/79

0.000

0.000

0.000

0.000

03/14/80

0.000

0.000

0.000

0.000

04/21/80

0.448

0.130

0.000

1.000

04/22/80

0.000

0.000

0.000

0.000

12/17/79
10/24/80

0.050
0.000

0.000
0.000

0.000
0.000

0.000
0.000

12/01/80
02/09/81

0.000
0.000

0.000
0.057

NA
0.000

NA
0.000

03/31/80

0.000

0.000

0.000

0.000

A. Allen, K. Ramanna / Journal of Accounting and Economics 55 (2013) 66–90

85

Table D1 (continued )
SFAS

ED title

ED date

11/14/80
Disclosure of Obligations: I Disclosure of Indirect
Guarantees of Indebtedness of Others: an interpretation of FASB Statement No. 5
02/09/81
SFAS048/
Accounting for Certain Product Sales I Revenue
SFAS049 Recognition When Right of Return Exists II Accounting for Product Financing Arrangements
06/12/81
Accounting by the Entertainment Industry I Motion
SFAS050/
SFAS051/ Picture Films II Broadcasting III Cable Television IV
SFAS053/ Records & Music
SFAS063
SFAS052
Foreign Currency Translation
08/28/80
SFAS052
Foreign Currency Translation; revision of exposure
06/30/81
draft issued August 28,1980
SFAS054
Financial Reporting and Changing Prices: Investment 11/16/81
Companies: an amendment of FASB Statement No.33
11/06/81
SFAS055
Determining whether a Convertible Security is a
Common Stock Equivalent: an amendment of APB
Opinion No. 15
11/06/81
SFAS056
Applicability of FASB Statement No. 32 to AICPA
Statements of Position and Guides on Accounting and
Auditing Matters: an amendment of FASB Statement
No. 32
SFAS057
Related Party Disclosures
11/06/81
SFAS058
Capitalization of Interest Cost in Financial Statements 09/30/81
That Include Investments Accounted for by The Equity
Method; an amendment of FASB
02/22/82
SFAS059
Deferral of the Effective Date of Certain Accounting
Requirements for Pension Plans of State and Local
Governmental Units: an amendment of FASB
Statement No. 35
SFAS060/
Accounting by the Insurance Industry I Accounting and 11/18/81
SFAS061 Reporting by Insurance Enterprises II Accounting for
Title Plant
SFAS062
Capitalization of Interest Cost in Situations Involving 12/22/81
Tax-Exempt Borrowings and Certain Gifts and Grants: an amendment of FASB Statement No. 34
SFAS064
Extinguishment of Debt Made to Satisfy Sinking-Fund
02/23/82
Requirements: an amendment of FASB Statement No. 4
SFAS065
Accounting for Certain Mortgage Banking Activities
02/03/82
12/15/81
SFAS066/
Accounting for Certain Real Estate Transactions I
SFAS067 Accounting for Costs and Initial Rental Operations of
Real Estate Projects II Accounting for Sales of Real
Estate
SFAS068
Research and Development Arrangements
04/27/82
SFAS069
Disclosures about Oil and Gas Producing Activities: an 04/15/82 amendment of FASB Statements 19 and 25
12/22/81
SFAS070
Financial Reporting and Changing Prices: Foreign
Currency Translation: an amendment of FASB
Statement No. 33
08/19/82
SFAS070
Financial Reporting and Changing Prices: Foreign
Currency Translation: an amendment of FASB
Statement No. 33 (Revision of 12/22/81 ED)
SFAS071
Accounting for the Effects of Regulation of an
03/04/82
Enterprise’s Prices Based on Its Costs
10/07/82
SFAS072
Accounting for Certain Acquisitions of Banking or
Thrift Institutions: an amendment of APB Opinion No.
17 and an interpretation of APB Opinion No. 16
SFAS073
Reporting a Change in Accounting for Railroad Track 04/12/83
Structures: an amendment of APB Opinion No. 20
SFAS074
Accounting for Special Termination Benefits Paid to
12/28/82
Employees
06/07/83
SFAS075
Deferral of the Effective Date of Certain Accounting
Requirements for Pension Plans of State and Local
Governmental Units: an amendment of FASB
Statement No. 35
10/13/82
SFAS076
Extinguishment of Debt and the Offsetting of
Restricted Assets against Related Debt: an amendment of APB Opinion No. 26 and FASB Statement No. 34
SFAS076
Extinguishment of Debt: an amendment of APB
07/14/83
Opinion No. 26 (Revision of 10/31/82 ED)

SFAS047

dec_relb

inc_ relv

Manual_ dec_relb

Manual_ inc_relv

NA

NA

0.000

0.000

0.000

0.000

NA

NA

0.000

0.000

NA

NA

NA
0.104

NA
0.000

0.000
NA

0.000
NA

0.000

0.000

0.000

0.000

0.093

0.000

0.000

0.000

0.000

0.000

1.000

0.000

0.125
0.000

0.000
0.000

0.000
0.000

0.000
0.000

0.000

0.000

0.000

0.000

0.000

0.000

0.000

0.000

0.000

0.000

1.000

0.000

0.000

0.000

0.000

0.000

0.000
0.000

0.000
0.000

0.000
0.000

0.000
0.000

0.000
0.000

0.000
0.000

1.000
0.000

0.000
1.000

NA

NA

0.000

0.000

0.000

0.000

0.000

0.000

0.061

0.000

1.000

0.000

0.000

0.000

0.000

0.000

0.000

0.000

0.000

0.000

0.000

0.000

1.000

0.000

0.000

0.000

0.000

0.000

NA

NA

0.000

0.000

0.000

0.000

0.000

0.000

86

A. Allen, K. Ramanna / Journal of Accounting and Economics 55 (2013) 66–90

Table D1 (continued )
SFAS
SFAS077

ED title

Accounting and Reporting by Transferors for Transfers of Receivables with Recourse
SFAS077
Reporting by Transferors for Transfers of Receivables with Recourse (Revision of 11/18/81 ED)
SFAS078
Classification of Obligations That Are Callable by the
Creditor: an amendment of Chapter 3A of ARB No. 43
SFAS079
Elimination of Certain Disclosures for Business
Combinations by Nonpublic Enterprises: an amendment of APB Opinion No. 16
SFAS080
Accounting for Futures Contracts
SFAS081
Disclosure of Postretirement Health Care and Life
Insurance Benefits Information
SFAS082
Financial Reporting and Changing Prices: Elimination of Certain Disclosures: an amendment of FASB
Statement No. 33
SFAS083
Designation of AICPA Guides and Statement of Position on Accounting by Brokers and Dealers in Securities, by
Employee Benefit Plans, and by Banks as Preferable for
Purposes of Applying APB Opinion 20: an amendment of FASB Statement No. 32 and a rescission of FASB
Interpretation No. 10
SFAS084
Induced Conversions of Convertible Debt: an amendment of APB Opinion No. 26
SFAS085
Yield Test for Determining whether a Convertible
Security is a Common Stock Equivalent: an amendment of APB Opinion No. 15
SFAS086
Accounting for the Costs of Computer Software to Be
Sold, Leased, or Otherwise Marketed
SFAS087
Employers’ Accounting for Pensions
SFAS088
Employers’ Accounting for Settlements and
Curtailments of Defined Benefit Pension Plans and for
Termination Benefits
SFAS089
Financial Reporting and Changing Prices: Current Cost
Information
SFAS089
Financial Reporting and Changing Prices
SFAS090/
Regulated Enterprises – Accounting for Phase-in Plans,
SFAS092 Abandonments, and Disallowances of Plant Costs: an amendment of FASB Statement No. 71
SFAS091
Accounting for Nonrefundable Fees and Costs
Associated with Originating and Acquiring Loans: an amendment of FASB Statements 13, 60, and 65 and a rescission of FASB Statement No. 17
SFAS093
Recognition of Depreciation by Not-for-Profit
Organizations
SFAS094
Consolidation of All Majority-Owned Subsidiaries—an amendment of ARB No. 51, with related amendments of APB Opinion No. 18 and ARB No. 43, Chapter 12
SFAS095
Reporting Income, Cash Flows, and Financial Position of Business Enterprises
SFAS095
Statement of Cash Flows
SFAS096
Accounting for Income Taxes
SFAS097
Accounting and Reporting by Insurance Enterprises for
Certain Long-Duration Insurance Contracts and for
Realized Gains and Losses from the Sale of Investments
SFAS098
Sale and Leaseback Transactions Involving Real Estate,
Sales-Type Leases of Real Estate, Definition of the
Lease Term, and Initial Direct Costs of Direct Financing
Leases
SFAS099
Deferral of the Effective Date of Recognition of
Depreciation by Not-for-Profit Organizations
SFAS100
Accounting for Income Taxes—Deferral of the Effective
Date of FASB Statement No. 96: an amendment of FASB
Statement No. 96
SFAS101
Regulated Enterprises—Accounting for the
Discontinuation of Application of FASB Statement No. 71
SFAS102
Statement of Cash Flows—Exemption of Certain
Enterprises and Classification of Cash Flows from
Certain Securities Held for Resale

ED date

dec_relb

inc_ relv

Manual_ dec_relb

Manual_ inc_relv

11/18/81

NA

NA

0.000

0.000

08/31/82

0.000

0.000

0.000

0.000

07/30/82

0.000

0.000

0.000

0.000

10/04/83

0.000

0.000

0.000

0.000

07/14/83
07/03/84

0.239
0.000

0.000
0.000

0.000
0.000

2.000
0.000

10/10/84

0.000

0.000

0.000

0.000

12/06/84

0.000

0.000

NA

NA

12/06/84

0.000

0.000

0.000

0.000

12/06/84

0.000

0.000

0.000

0.000

08/31/84

0.471

0.000

1.000

2.000

03/22/85
06/14/85

0.096
0.000

0.000
0.000

NA
0.000

NA
0.000

12/14/84

NA

NA

0.000

0.000

09/30/86
12/19/85

NA
0.000

NA
0.000

0.000
NA

0.000
NA

12/31/85

0.000

0.000

0.000

0.000

12/23/86

0.000

0.000

0.000

0.000

12/16/86

0.000

0.013

0.000

0.000

11/16/81

NA

NA

0.000

0.000

07/31/86
09/02/86
12/23/86

0.000
0.023
0.000

0.122
0.157
0.000

0.000
0.000
1.000

0.000
0.000
0.000

08/31/87

NA

NA

0.000

0.000

06/06/88

0.000

0.000

0.000

0.000

10/13/88

0.000

0.000

0.000

0.000

07/08/88

NA

NA

0.000

0.000

11/30/88

0.000

0.000

0.000

0.000

A. Allen, K. Ramanna / Journal of Accounting and Economics 55 (2013) 66–90

87

Table D1 (continued )
SFAS

ED title

ED date

dec_relb

inc_ relv

Manual_ dec_relb

Manual_ inc_relv

SFAS103

Accounting for Income Taxes—Deferral of the Effective
Date of FASB Statement No. 96: an amendment of FASB
Statement No. 96
Statement of Cash Flows—Net Reporting of Certain
Cash Receipts and Cash Payments and Classification of
Cash Flows from Hedging Transactions
Disclosure about Financial Instruments
Disclosure of Information about Financial Instruments with Off-Balance-Sheet Risk and Financial instruments with Concentrations of Credit Risk
Employers’ Accounting for Postretirement Benefits
Other Than Pensions
Disclosures about Market Value of Financial
Instruments
Accounting for Income Taxes—Deferral of the Effective
Date of Statement No. 96, an amendment of FASB
Statement No. 96
Accounting for Income Taxes
Reporting by Defined Benefit Pension Plans of
Investment Contracts: an amendment of FASB
Statement No. 35
Rescission of FASB Statement No. 32 and Technical
Corrections
Employers’ Accounting for Postretirement Benefits: an amendment of FASB Statements No. 5 and 43
Accounting and Reporting for Reinsurance of ShortDuration and Long-Duration Contracts
Accounting by Creditors for Impairment of a Loan: an amendment of FASB Statements No. 5 and 15
Accounting for Certain Investments in Debt and Equity
Securities
Accounting for Contributions Received and
Contributions Made and Capitalization of Works of Art,
Historical Treasurers, and Similar Assets
Accounting for Contributions Received and
Contributions Made (Revision of 10/31/90 ED)
Financial Statements of Not-for-Profit Organizations
Accounting by Creditors for Impairment of a
Loan—Income Recognition: an amendment of FASB
Statement No. 114
Disclosure about Derivative Financial Instruments and
Fair Value of Financial Instruments
Accounting and Reporting by Mutual Life Insurance
Enterprises and by Insurance Enterprises for Certain
Long-Duration Participating Contracts: an amendment of FASB Statements No. 60, 97, and 113 (Includes
Proposed AICPA Statement of Position, Accounting for
Certain Insurance Activities of Mutual Life Insurance
Enterprises)
Accounting for the Impairment of Long-Lived Assets
Accounting for Mortgage Servicing Rights and Excess
Servicing Receivables and for Securitization of Mortgage
Loans an amendment of FASB Statement No. 65
Accounting for Stock-Based Compensation
Share-Based Payment: an amendment of FASB
Statements No. 123 and 95
Accounting for Certain Investments Held by Not-forProfit Organizations

10/19/89

NA

NA

NA

NA

07/25/89

0.000

0.151

0.000

0.000

11/30/87
07/21/89

NA
0.000

NA
0.000

1.000
0.000

1.000
0.000

02/14/89

0.533

0.000

1.000

4.000

12/31/90

0.244

0.593

1.000

1.000

06/17/91

0.000

0.000

0.000

0.000

06/05/91
03/20/92

NA
0.000

NA
0.326

0.000
1.000

0.000
1.000

06/30/92

0.000

0.000

0.000

0.000

05/12/92

0.000

0.000

0.000

0.000

03/20/92

0.049

0.121

0.000

0.000

06/30/92

0.210

0.124

1.000

2.000

09/09/92

0.507

0.000

1.000

4.000

10/31/90

0.000

0.000

1.000

2.000

11/17/92

0.379

0.000

1.000

0.000

10/23/92
03/31/94

0.000
0.000

0.589
0.000

0.000
0.000

0.000
0.000

04/14/94

0.000

0.137

0.000

1.000

03/24/94

0.000

0.000

0.000

0.000

11/29/93
06/28/94

0.000
0.148

0.000
0.030

1.000
1.000

3.000
4.000

06/30/93
03/31/04

0.372
0.318

0.000
0.466

1.000
0.000

2.000
2.000

03/31/95

NA

NA

0.000

3.000

10/24/95

0.510

0.120

1.000

2.000

09/20/96

0.000

0.000

0.000

0.000

11/11/96

0.000

0.000

0.000

0.000

01/19/96

0.000

0.135

0.000

0.000

06/20/96

0.000

0.000

0.000

0.000

SFAS104

SFAS105
SFAS105

SFAS106
SFAS107
SFAS108

SFAS109
SFAS110

SFAS111
SFAS112
SFAS113
SFAS114
SFAS115
SFAS116

SFAS116
SFAS117
SFAS118

SFAS119
SFAS120

SFAS121
SFAS122

SFAS123
SFAS123R
SFAS124

SFAS125

Accounting for Transfers and Servicing of Financial
Assets and Extinguishments of Liabilities
SFAS126
Elimination of Certain Disclosures about Financial
Instruments by Small Nonpublic Entities: an amendment of FASB Statement No. 107
SFAS127
Deferral of the Effective Date of Certain Provisions of
FASB Statement No. 125: an amendment of FASB
Statement No. 125
SFAS128/
Earnings per Share and Disclosure of Information
SFAS129 about Capital Structure
SFAS130
Reporting Comprehensive Income

88

A. Allen, K. Ramanna / Journal of Accounting and Economics 55 (2013) 66–90

Table D1 (continued )
SFAS

ED title

ED date

dec_relb

inc_ relv

Manual_ dec_relb

Manual_ inc_relv

SFAS131

Reporting Disaggregated Information about a Business
Enterprise
Employers’ Disclosures about Pensions and Other
Postretirement Benefits: an amendment of FASB
Statements No. 87, 88, and 106
Employers’ Disclosures about Pensions and Other
Postretirement Benefits: an amendment of FASB
Statements No. 87, 88, and 106 and a replacement of
FASB Statement No. 132
Accounting for Derivative and Similar Financial
Instruments and for Hedging Activities
Accounting for Mortgage-Backed Securities and
Certain Other Interests Retained after the
Securitization of Mortgage Loans Held for Sale by a
Mortgage Banking Enterprise: an amendment of FASB
Statement No. 65
Amendments to FASB Statement No. 66, Rescission of
FASB Statement No. 75, and Technical Corrections
Transfers of Assets in Which a Not-for-Profit
Organization Acts as an Agent, Trustee, or
Intermediary: an Interpretation of FASB Statement No.
116
Transfers of Assets involving a Not-for-Profit
Organization That Raises or Holds Contributions for
Others
Accounting for Derivative Instruments and Hedging
Activities—Deferral of the Elective Date of FASB
Statement No. 133: an amendment of FASB Statement
No. 133
Accounting for Certain Derivative instruments and
Certain Hedging Activities: an amendment of FASB
Statement No. 133
Rescission of FASB Statement No. 53
Accounting for Transfers of Financial Assets: an amendment of FASB Statement No. 125
Business Combinations and Intangible Assets

01/19/96

0.018

0.000

1.000

0.000

06/30/97

0.000

0.000

0.000

1.000

09/12/03

0.163

0.131

0.000

1.000

06/20/96

0.101

0.046

1.000

3.000

4/10/98

0.000

0.000

1.000

2.000

10/13/98

0.000

0.000

0.000

0.000

12/29/95

NA

NA

0.000

0.000

07/17/98

0.000

0.000

0.000

3.000

05/20/99

0.000

0.000

0.000

0.000

03/03/00

0.000

0.000

0.000

0.000

10/16/98
06/28/99

0.000
0.378

0.000
0.000

0.000
1.000

0.000
2.000

09/07/99

0.461

0.152

1.000

3.000

06/30/05

0.909

0.477

1.000

5.000

02/14/01

0.647

0.041

1.000

3.000

02/07/96

0.000

0.000

1.000

1.000

02/17/00

0.452

0.278

1.000

0.000

11/15/01

0.158

0.376

1.000

2.000

06/30/00

0.000

0.000

0.000

0.000

02/14/02

0.000

0.000

0.000

0.000

05/10/02

0.000

0.000

1.000

3.000

10/04/02

0.000

0.000

1.000

1.000

05/01/02

0.132

0.000

0.000

1.000

10/27/00

0.297

0.000

0.000

0.000

12/15/03

0.000

0.000

0.000

0.000

02/20/03

0.000

0.000

0.000

0.000

12/15/03

0.161

0.000

1.000

1.000

SFAS132

SFAS132R

SFAS133
SFAS134

SFAS135
SFAS136

SFAS136

SFAS137

SFAS138

SFAS139
SFAS140

SFAS141/
SFAS142
SFAS141R Business Combinations: a replacement of FASB
Statement No. 141
SFAS142
Business Combinations and Intangible
Assets—Accounting for Goodwill (Revision of 9/7/99 ED)
SFAS143
Accounting for Certain Liabilities Related to Closure or
Removal of Long-Lived Assets
SFAS143
Accounting for Obligations Associated with the
Retirement of Long-Lived Assets (Revision of 2/7/96 ED)
SFAS144/
Rescission of FASB Statements No. 4, 44, and 64 and
SFAS146 Technical Corrections
SFAS145
Accounting for the Impairment or Disposal of LongLived Assets and for Obligations Associated with
Disposal Activities
SFAS145
Rescission of FASB Statements No. 4, 44, and 64 and
Technical Corrections—Amendment of FASB
Statement No. 13 (Revision of 11/15/01 ED)
SFAS147
Acquisitions of Certain Financial Institutions: an amendment of FASB Statements No. 72 and No. 144 and FASB Interpretation No. 9
SFAS148
Accounting for Stock-Based Compensation–Transition and Disclosure: and amendment of FASB Statement
No. 123
SFAS149
Amendment of Statement 133 on Derivative
Instruments and Hedging Activities
SFAS150
Accounting for Financial Instruments with
Characteristics of Liabilities: Equity, or Both
SFAS151
Inventory Costs: an amendment of ARB No. 43,
Chapter 4
SFAS152
Accounting for Real Estate Time-Sharing Transactions: an amendment of FASB Statements No. 66 and 67
SFAS153
Exchanges of Productive Assets: an amendment of ABP
Opinion No. 29

A. Allen, K. Ramanna / Journal of Accounting and Economics 55 (2013) 66–90

89

Table D1 (continued )
SFAS

ED title

ED date

dec_relb

inc_ relv

Manual_ dec_relb

Manual_ inc_relv

SFAS154

Accounting Changes and Error Corrections: a replacement of ABP Opinion No. 20 and FASB
Statement No. 3
Accounting for Certain Hybrid Financial Instruments: an amendment of FASB Statements No. 133 and 140
Qualifying Special-Purpose Entities and Isolation of
Transferred Assets: an amendment of FASB Statement
No. 140
Accounting for Servicing of Financial Assets: an amendment of FASB Statement No. 140 (Revision of 6/
10/03 ED)
Accounting for Transfers of Financial Assets: an amendment of FASB Statement No. 140 (Revision of 6/
10/03 ED)
Fair Value Measurements
Employers’ Accounting for Defined Benefit Pension and Other Postretirement Plans: an amendment of
FASB Statements No. 87, 88, 106, and 132(R)
The Fair Value Option for Financial Assets and
Financial Liabilities: Including an amendment of FASB
Statement No. 115
Consolidated Financial Statements, Including
Accounting and Reporting of Non-controlling Interests in Subsidiaries: a replacement of ARB No. 51

12/15/03

0.000

0.000

1.000

0.000

08/11/05

0.190

0.397

1.000

2.000

06/10/03

NA

NA

1.000

0.000

08/11/05

0.003

0.113

1.000

4.000

08/11/05

NA

NA

1.000

0.000

06/23/04
03/31/06

0.599
0.000

0.245
0.000

0.000
0.000

2.000
1.000

01/25/06

0.451

0.669

1.000

4.000

06/30/05

0.586

0.000

1.000

4.000

SFAS155
SFAS156

SFAS156

SFAS156

SFAS157
SFAS158

SFAS159

SFAS160

References
Bamber, L., Jiang, J., Wang, I., 2010. What’s my style? The influence of top managers on voluntary corporate financial disclosure. The Accounting Review
85, 1131–1162.
Barth, M., 2006. Research, standard setting, and global financial reporting. Foundation and Trends in Accounting 1, 71–165.
Bertomeu, J., Magee, R., 2011. From low-quality reporting to financial crises: politics of disclosure regulation along the economic cycle. Journal of
Accounting and Economics 52, 209–227.
Bertrand, M., Schoar, A., 2003. Managing with style: the effect of managers on firm policies. Quarterly Journal of Economics 118, 1169–1208.
Cohen, J., 1986. The dynamics of the ‘‘revolving door’’ on the FCC. American Journal of Political Science 30, 689–708.
Dal Bo, E., 2006. Regulatory capture: a review. Oxford Review of Economic Policy 22, 203–225.
Dyckman, T., Magee, R., Pfeiffer, G., 2011. Financial Accounting (3e). Cambridge Business Publishers, Westmont, IL.
Dyreng, S., Hanlon, M., Maydew, E., 2010. The effects of executives on corporate tax avoidance. The Accounting Review 85, 1163–1189.
Financial Accounting Foundation (FAF), 2008. The Financial Accounting Foundation board of trustees approves changes to oversight, structure and operations of FAF, FASB and GASB. FAF News Release February 26, Norwalk, CT.
Financial Accounting Foundation (FAF), 2010. Financial Accounting Foundation to Increase Size of FASB. FAF News Release August 24, Norwalk, CT.
Financial Accounting Standards Board (FASB), 1978. Statement of Financial Accounting Concepts No. 1, Objectives of Financial Reporting by Business
Enterprises. FASB, Norwalk, CT.
Financial Accounting Standards Board (FASB), 1980. Statement of Financial Accounting Concepts No. 2, Qualitative Characteristics of Accounting
Information. FASB, Norwalk, CT.
Financial Accounting Standards Board (FAS), 2007. FASB Standard Establishes Fair Value Option for Financial Assets and Liabilities. FASB News Release
February 15, Norwalk, CT.
Ge, W., Matsumoto, D., Zhang, J., 2010. Do CFOs have style? An empirical investigation of the effect of individual CFOs on accounting practices.
Contemporary Accounting Research 28, 1141–1179.
Gormley, W., 1979. A test of the revolving door hypothesis at the FCC. American Journal of Political Science 23, 665–683.
Healy, P., Whalen, J., 1999. A review of the earnings management literature and its implications for standard setting. Accounting Horizons 13, 365–383.
Investment Company Institute (ICI), 2008. Comment Letter by CEO Paul S. Stevens on the SEC’s Request for Comment on Mark-to-market Accounting.
Available at: /http://www.ici.org/pdf/23068.pdfS (accessed April, 2011).
Johnson, L., 2005. Relevance and Reliability: The FASB Report. Available at: /www.fasb.orgS.
Kalt, J., Zupan, M., 1984. Capture and ideology in the economic theory of politics. American Economic Review 74, 279–300.
Kau, J., Rubin, P., 1979. Self-interest, ideology, and logrolling in congressional voting. Journal of Law and Economics 22, 365–384.
Kellogg, R., 1984. Accounting activities, security prices, and class action lawsuits. Journal of Accounting & Economics 6, 185–204.
Kothari, S., Ramanna, K., Skinner, D., 2010. Implications for GAAP from an analysis of positive research in accounting. Journal of Accounting & Economics
50, 246–286.
Kothari, S., Shu, S., Wysocki, P., 2009. Do managers withhold bad news? Journal of Accounting Research 47, 241–276.
LaFond, R., Watts, R., 2008. The information role of conservative financial statements. The Accounting Review 83, 447–478.
Leaver, C., 2009. Bureaucratic minimal squawk behavior: theory and evidence from regulatory agencies. American Economic Review 99, 572–607.
Nelson, M., Elliott, J., Tarpley, R., 2002. Evidence from auditors about managers’ and auditors’ earnings management decisions. The Accounting Review 77,
175–202.
Petersen, M., 2009. Estimating standard errors in finance panel data sets: comparing approaches. Review of Financial Studies 22, 435–480.
Ramanna, K., 2006. The implications of unverifiable fair-value accounting: evidence from the political economy of goodwill accounting. Presentation at the Journal of Accounting & Economics Conference, Northwestern University.
Ramanna, K., 2008. The implications of unverifiable fair-value accounting: evidence from the political economy of goodwill accounting. Journal of
Accounting & Economics 45, 253–281.
Ramanna, K., Cheng, B., 2009. Choosing a GAAP for Canada. Harvard Business School Case No. 110-023.

90

A. Allen, K. Ramanna / Journal of Accounting and Economics 55 (2013) 66–90

Ramanna, K., Donovan, G., Dai, N., 2010. IFRS in China. Harvard Business School Case No. 110-037.
Stickney, C., Weil, R., Schipper, K., Francis, J., 2010. Financial Accounting (13e). South-Western Cengage Learning, Mason, OH.
Schipper, K., 2005. Fair Values in Financial Reporting. Presentation at American Accounting Association Annual Meetings, August 2005. Available at:
/http://fars.org/2005AAAFairValueKSchipper.pdfS (last accessed September 15, 2009).
Stigler, G., 1971. The theory of economic regulation. Bell Journal of Economics 2, 3–21.
Watts, R., 2003. Conservatism in accounting Part I: explanations and implications. Accounting Horizons 17, 207–221.
Watts, R., Zimmerman, J., 1978. Towards a positive theory of the determination of accounting standards. The Accounting Review 53, 112–134.
Watts, R., Zimmerman, J., 1982. Auditors and the Determination of Accounting Standards. Working Paper, University of Rochester.
Watts, R., Zimmerman, J., 1983. Auditor Independence and Scope of Services. Working Paper, University of Rochester.
Zhang, J., 2008. The contracting benefits of accounting conservatism to lenders and borrowers. Journal of Accounting & Economics 45, 27–54.