Professional Documents
Culture Documents
in
Accounting
Vol. 2, Nos. 1–2 (2007) 1–174
c 2008 G. B. Waymire and S. Basu
DOI: 10.1561/1400000011
1
Department of Accounting, Goizueta Business School, Emory University,
Atlanta, GA 30322, USA, Gregory Waymire@bus.emory.edu
2
Department of Accounting, Fox School of Business and Mangement,
Temple University, Philadelphia, PA 19122, USA,
Sudipta.Basu@temple.edu
Abstract
We develop our paper by defining “accounting history research” and
posing six big picture questions about historical accounting evolution.
The paper selectively summarizes accounting history over the past ten
thousand years, organized around our six questions. This review pro-
vides useful examples that we draw upon for subsequent sections, but
can also be used as a US accounting history primer. We explain how
accounting history can inform scholars studying modern institutions
by analyzing several exemplary research papers. The paper discusses
numerous empirical studies using archival accounting data and suggests
further questions that can build upon and extend published research.
Finally, we discuss the implications of our evolutionary perspective
for accounting research, and identify numerous research opportunities
* This essay is dedicated to the memory of our colleague, George Benston, who passed away
earlier this year. George was a consummate scholar who believed strongly in the value
of economics-based policy research informed by intelligent data analysis. He also was a
person of remarkable energy and zest for life. We shall miss him.
under each of our six big picture questions that will together help us
build a coherent evolutionary theory of accounting.
3
4 Introduction: Why Accounting History Research is Valuable
1 To illustrate, there were 16 archival papers published in the January 2007 issue of The
Accounting Review, the March 2007 issue of the Journal of Accounting Research, and
the March 2007 issue of the Journal of Accounting and Economics. Of these, only three
studies include any sample data from before 1980 whereas 12 have samples drawn entirely
from the post-1990 period. None of the three studies using data from before 1980 are ones
where historical data is necessary to execute the study (two are asset pricing studies while
one is a methodological analysis of the properties of accounting conservatism measures).
Fleischman and Radcliffe (2005) discuss the lack of historical publications in major US
journals within the broader context of how accounting history research developed in the
1990’s.
2 Human recordkeeping is first evidenced by the tokens of Ancient Mesopotamia that date
back to approximately 8,000 BCE, and humans invented writing to keep better accounting
records (see Basu and Waymire (2006), pp. 213–217, and references therein).
5
7
8 Quantitative Accounting History Research
Libby et al., 2003, p. 846). For present purposes, we define accounting as a process that
involves recording historical data about economic transactions, which are then classified,
aggregated, and summarized in ways that allow re-construction of the past to facilitate
decision-making.
2 A nice review of the use of historical methods in accounting research is available in Previts
et al. (1990a,b).
3 See entry in Wikipedia at http://en.wikipedia.org/wiki/Cliometrics.
4 Examples of their early work include Fogel and Engerman (1974) and North and Thomas
(1973). McCloskey and Hersh (1990) provide a broad listing of cliometric research.
9
choices are not readily apparent for cases where tax and reg-
ulatory effects are absent?
(6) Why is accounting evolution subject to major discontinu-
ities that arise from changes made in response to economic
crises? Do these discontinuities, or punctuated equilibria,
arise because the fitness consequences of poor accounting
choices are revealed only by systemic failures affecting mul-
tiple firms simultaneously?
11
12 Overview of Questions that Arise in Accounting History Research
Table 3.1 Some important events in accounting history organized around six general themes.
A: The Importance of Recordkeeping (Section 3.1)
A1. 8,000 BCE. Earliest known use of clay tokens as symbolic representations of
transactions (Schmandt-Besserat, 1992).
A2. 3,200 BCE. Tokens begin to be sealed inside baked clay balls with seals impressed
on the outside to improve data integrity regarding complex exchange involving future
promises.
A3. 3,200 BCE. First known invention of writing, which was for the purpose of recording
transactions. Numerals invented around the same time (Nissen et al., 1993).
A4. 1,750 BCE. The legal code of Hammurabi requires transaction records as evidence in
cases of commercial disputes.
A5. 250 BCE. Athens has buildings serving as public archives for financial records,
accounts, lease contracts, and wills that could be used in legal disputes (Sickinger,
1999, pp. 122–134).
A6. 1200 AD. Audits required in English merchant guilds (Watts and Zimmerman, 1983,
p. 616).
A7. 1673. Ordinance of 1673 in France requires that journals be kept documenting
receivables and payables. This was retained in the French Commercial Code of 1807.
Well-kept books could be presented as evidence in legal proceedings over unpaid
obligations (Howard, 1932, pp. 91–102).
A8. 2002. Sarbanes–Oxley Act signed into law by President George W. Bush on July 30.
Section 1102 of the law establishes criminal penalties for tampering with the
transactional records of a company subject to US federal securities laws.
D: Accounting and its Relation to Law, Regulation, and Taxation (Section 3.4)
D1. 8,000 BCE. Earliest known use of clay tokens as symbolic representations to signify
the payment of communal tribute (Schmandt-Besserat, 1992).
D2. 1,750 BCE. The legal Code of Hammurabi requires transaction records as evidence
in cases of commercial disputes.
D3. 1433. Bank of Medici regularly produces balance sheets, which at that time were
routinely prepared for both purposes of control as well as property taxation (de
Roover, 1955, p. 414).
D4. 1673. Ordinance of 1673 in France requires that journals be kept documenting
receivables and payables. This was retained in the French Commercial Code of 1807.
Well-kept books could be presented as evidence in legal proceedings over unpaid
obligations (Howard, 1932, pp. 91–102).
D5. 1817. Accountants begin assisting British courts in bankruptcy cases (“Historical
Dates in Accounting,” The Accounting Review, July 1954).
D6. 1844. Joint Stock Companies Act of 1844, the first Companies Act, is written into
British law. Main provisions were to establish principles on differences between
partnerships and corporations, incorporation by registration rather than act of
(Continued)
15
2 Kohler (1952) defines recordkeeping as the collection and maintenance of objective and
verifiable (i.e., legal) evidence of transactions and bookkeeping as the systematic analysis,
classification, and aggregation of transactions.
18 Overview of Questions that Arise in Accounting History Research
Accounting history can be divided into two periods, the first being
that before double-entry bookkeeping appeared in Italy in the 1200s
and was summarized in Paciolo’s (1494) printed text (Chatfield, 1974,
p. 4). Throughout this era, accounting was primarily recordkeeping,
although in some places it had been elaborated into single-entry book-
keeping. Recordkeeping is of continuing economic significance since it
forms the core of modern accounting systems based on journalization
using double entry (Ijiri, 1975; Littleton, 1933; Potter, 1952). Thus,
understanding the economic value of recordkeeping, independent of
other steps in the accounting process, is necessary for understanding
the deeper functions of modern accounting. Eight events relevant to
the history of recordkeeping are listed in panel A of Table 3.1.
Archaeologists document that people recorded economic exchange
for at least 10,000 years (Schmandt-Besserat, 1992), although scratches
on stones and bones may have been used to record exchange as far back
as 75,000 BCE (Ifrah, 2001, Chap. 4). The earliest known records of
economic transactions are the tokens of ancient Mesopotamia, which
date back to 8,000 BCE at the time of early human settlements based
on agriculture (event #A1 in Table 3.1). This recordkeeping technology
was improved when multiple tokens for items in a bundled transaction
involving future performance obligations were baked in hollow balls
(“bullae”) after witnesses had affixed their seals (event #A2). This
technology improved the integrity of transactional data by rendering
the data “hard” in the sense that ex post it “will be difficult for people
to disagree” (Ijiri, 1975, p. 36). Further improvements in recordkeeping
resulted from written language and numerals, which were first invented
by the Sumerians around 3,200 BCE for purposes of recordkeeping
(event #A3) (Nissen et al., 1993).3
Transactional records have long served to resolve commercial dis-
putes. The Code of Hammurabi (c. 1750 BCE) specified how receipts
were to be used as evidence in legal proceedings (event #A4) (Saggs,
1989; Versteeg, 2000). By the middle of the third century B.C., the
Athenians maintained public archives of financial records, accounts,
3 Recent claims that writing emerging earlier in Egypt are disputed (Mattessich, 2002),
while similar claims for India (http://www.rediff.com/news/1999/jul/14harap.htm) seem
unfounded.
3.2 How Could Double-entry Bookkeeping Enable Large-scale Capitalism? 19
lease contracts, and wills (Sickinger, 1999, pp. 122–134), which could be
used in legal dispute resolution (event #A5). A continuing paramount
concern over accounting data integrity is evidenced by the use of
manorial audits in Britain around 1200 AD (event #A6) and subse-
quent laws requiring that records be kept in a specific form so that
they can serve as evidence in legal proceedings (event #A7). The
fundamental importance of transactional records characterized by high
integrity continues to be an important issue. Section 1102 of Sarbanes–
Oxley of 2002 establishes criminal penalties for tampering with the
transactional records of a company subject to US federal securities
laws (event #A8).
The economic function of recordkeeping is an issue we will return
to later in this survey, especially in the section on the evolutionary
function of accounting. Specifically, we will explore how hard records
leverage evolved human capabilities for beneficial exchange in building
the kinds of advanced societies seen around the globe today (Basu and
Waymire, 2006).
4 Werner Sombart is translated in English as asserting: “One cannot say whether capital-
ism created double-entry bookkeeping, as a tool in its expansion, or whether perhaps,
conversely, double-entry bookkeeping created capitalism. . . ” (Lane and Riemersma, 1953,
p. 38).
20 Overview of Questions that Arise in Accounting History Research
5 Accountingfor wasting assets was clearly recognized before the mid-19th century, but was
not economically significant until large corporations with permanent capital investment
came into being. Mason (1933) lists several examples of the accounting for wasting assets
dating back to the 17th century.
3.3 How Can Accounting “Norms” Form Spontaneously 21
6 Hopwood’s (1987) three case studies provide a richly textured analysis of the bi-directional
relation between accounting and organizational complexity at the individual firm level that
complements and deepens our description of the broad economy level.
22 Overview of Questions that Arise in Accounting History Research
7 The requirement that professional accountants must pass a test to attain professional
status is likely much older. For instance, professional scribes in ancient Mesopotamia were
subject to educational requirements (Pearce, 1995).
3.4 Accounting’s Relation with Law, Regulation, and Taxation 23
28
4.1 The Role of Thought Experiments in Research 29
2 Two highly successful scientists, Thomas Edison and Elihu Thomson, led scientific inno-
vation in these firms. Edison’s involvement in the merged entity ended a short time after
GE’s creation. Thomson continued on with GE for the rest of his career, having been
responsible for 696 patents, the third highest in American history as of 1990 (Carlson,
1991). He died in the 1930s.
32 Accounting History Research as a Stimulant to Thought Experiments
Table 4.1 General electric income statement and balance sheet for fiscal 1907.
little if any weight should be given to the figures at which intangibles assets appear on the
balance sheet. Such intangibles may have a very large value indeed, but it is the income
account and not the balance sheet that offers the clue to this value. In other words, it
is the earning power of these intangibles, rather than their balance sheet valuation, that
really counts.”
34 Accounting History Research as a Stimulant to Thought Experiments
GE’s policy was still praised by writers nearly 20 years later when
Sturgis (1925, p. 122) noted:
The obvious question about this case is why GE adopted this con-
servative reporting policy. It surely was not to conceal the existence
of economically valuable intangibles — GE’s 1907 annual report made
extensive disclosure of product innovations and sales by product type.
Also, the standard explanations for conservatism as arising from debt
contracting, taxation, and shareholder litigation are implausible. GE’s
outstanding debt as a percentage of total assets was trivial (about 2%
as of 1907); income taxation did not begin until 1909, and even then
only at a rate of 1%; and shareholder litigation over financial reporting
was minimal in this era.5 The threat of anti-trust regulation is more
5 However,in an influential decision in the case of Newton v. Birmingham Small Arms Co.
Ltd. (1906), an English court held, “Assets are often, by reason of prudence, estimated
4.2 The Role of Historical Case Studies 35
plausible, but GE had not been targeted like the major railroads and
visible industrials such as Standard Oil.
The most likely explanation for GE’s conservative reporting policy is
that it helped in raising equity capital, which was its primary vehicle of
external financing. Dividend policy was viewed as the primary driver of
value in the early 20th century US equity markets (Graham and Dodd,
1934, pp. 299–306). Because GE was a highly profitable entity and its
conservative reporting policy constrained excessive dividend payouts in
any given year, it could consistently pay annual cash dividends of $8
per common share for at least 25 years after 1900. This allowed the
company to raise considerable equity capital in the two decades after
1907.6 GE’s sales grew from $60.1 million in 1907 to $312.6 million
in fiscal 1927. This suggests that GE was an early growth stock that
secured substantial equity capital to finance its growth. GE’s reporting
and stable dividend policies likely sustained its ability to raise capital,
at least in part.
GE’s policy of maintaining steady high dividends was a challenge
because GE primarily sold equipment to firms and municipalities that
sought to exploit expanding US electrification. GE’s revenues, like those
of other durable goods producers, were highly sensitive to the busi-
ness cycle. GE was financially distressed and almost did not survive
the depression of 1893. Shortly thereafter, GE initiated its conserva-
tive reporting policy with large asset writedowns of patents as well as
securities investments and fixed assets. Lower asset values and immedi-
ate expensing of many investments meant that GE was booking large
“secret reserves.” As a result, the payment of excessive dividends was
and stated to be estimated, at less than their probable real value. The purpose of the
balance-sheet is primarily to shew that the financial position of the company is at least as
good as there stated, not to shew that it is not or may not be better” (Dickerson, 1966).
This precedent may have shaped GE’s expectations regarding any legal proceedings.
6 GE had about 7,000 shareholders when their reported common stock and surplus was $79.5
million in 1907 compared to 50,000 shareholders 20 years later when their total common
stock and surplus was reported to be $295.4 million. The number of shareholders gradually
trended up from 3,000 in 1900 to around 17,000 in 1920, and then surged dramatically
to over 25,000 in 1921. Details of the number of GE shareholders were taken from the
General Electric Annual Report for 1927.
36 Accounting History Research as a Stimulant to Thought Experiments
7 GE’s sales declined sharply during the downturn of 1908 when it reported sales of $44.5
million for the 1/31/09 fiscal year, down from $71.0 million in the prior year. In contrast
to GE’s conservative and transparent reporting policy, one of GE’s primary competitors,
Westinghouse, did not disclose any annual reports and held no shareholder meetings during
the years between 1897 and 1905 (Brief, 1987). Westinghouse was placed in receivership
after the depression of 1907, and George Westinghouse, the company’s founder, lost control
of the company.
4.2 The Role of Historical Case Studies 37
8 We are grateful to Eric Gottlieb for collecting these data for us.
38 Accounting History Research as a Stimulant to Thought Experiments
9 Investorsevidently were not misled by GE’s conservative reporting policy. GE was one of
only a handful of industrial common stocks quoted above the par value of shares (typically
$100 per share). GE’s market-to-book at 12/31/1908 was 1.27, which was greater than any
other firm in the Sivakumar and Waymire (1993) sample. Similarly, for firms that report
more conservatively, Craig et al. (1987) and Beaver et al. (1989) find higher price-earnings
ratios and higher market-to-book ratios, respectively.
40 Accounting History Research as a Stimulant to Thought Experiments
income changes are positively associated with stock returns in the year
leading up to the annual earnings announcement date. The second is
that managers are less likely to disclose unfavorable news in the absence
of a disclosure mandate or other penalties (Dye, 1983; Patell, 1976;
Penman, 1980; Verrecchia, 1983).
Sivakumar and Waymire (1993, 1994) asked whether early 20th
century disclosures would be skewed toward good news and whether
the minimal accounting and auditing requirements of the era would
lessen or even eliminate the price-informativeness of income numbers.
Accounting income changes were compared to a benchmark of dividend
changes since dividend declarations were a primary information source
in the early 20th century (Graham and Dodd, 1934, pp. 325–338). Div-
idend declarations were credible because directors could be held legally
liable for paying excessive dividends that impaired the firm’s capital
(Berle and Means, 1932, pp. 135–136).
The Sivakumar and Waymire papers study 51 industrial corpora-
tions with common stock traded on the New York Stock Exchange
for at least two years between January 6, 1905 and June 3, 1910, the
latter date coinciding with the NYSE’s formal abolition of trading
in “unlisted” securities. Twenty-three of the 51 firms in the sample
had been traded on unlisted status for at least part of the 1905–
1910 period. These unlisted stocks are interesting because they had
not signed formal listing agreements requiring disclosure of an annual
report to the exchange. Thus, the Sivakumar and Waymire (1993,
1994) sample and time period was well suited to analyzing the impor-
tance of income disclosure in a largely voluntary income-reporting
environment.
There is substantial evidence of income and income component
disclosure although their extent varied considerably. Of the 217 sample
annual reports published in the Commercial and Financial Chronicle,
188 reports (87%) included at least a partial income statement. Of
these 188 reports, 97 (52%) included a sales number and 19 (of 51)
reporting firms always included sales information in their income
statement. Likewise, a majority of firms provided interim earnings
and marketing-production information at least once during the sample
period. Thus, while income and income-related disclosure was neither
4.3 The Role of Small Sample Studies Using Historical Data 43
10 About half (n = 26) of the 51 firms examined by Sivakumar and Waymire (1993) paid
dividends at some point in the 1905–1910 period.
11 The pattern of stock price changes coincident with Sivakumar and Waymire’s dividend
change sample is similar to the price behavior accompanying dividend changes in the
modern era when financial reporting policies were far better developed (Aharony and
Swary, 1980; Eades et al., 1985; Healy and Palepu, 1988).
44 Accounting History Research as a Stimulant to Thought Experiments
12 Thefirst two editions of The Accountant’s Handbook (Paton, 1932; Saliers, 1923) provide
extensive descriptions of the state of accounting practice before the Securities Acts.
4.4 What Price Do We Pay for Ignorance of Accounting’s Institutional History? 45
In reviewing the history of their creation, the SEC also notes on its
website:
13 FASB has also stated similar rationales on its website (www.fasb.org/facts): “The mission
of the FASB is to establish and improve standards of financial accounting and reporting
for the guidance and education of the public, including issuers, auditors, and users of
financial information. Accounting standards are essential to the efficient functioning of
the economy because decisions about the allocation of resources rely heavily on credible,
concise, transparent, and understandable financial information.”
46 Accounting History Research as a Stimulant to Thought Experiments
thirty years and eleven best-selling editions, the text has built a repu-
tation for accuracy, comprehensiveness, and student success.”14
Similar to the SEC’s statements, Kieso et al. (2006, pp. 6, 7) write:
when they note that the FASB is making “good progress” on their objectives (as stated
by Chairman Bob Herz) and they provide “several examples of where the board has
exerted leadership.” From this, they conclude that improved “financial reporting should
follow.”
16 The statements in Kieso et al. (2006) are not unique; many US accounting textbooks
routinely cite the investor protection rationale for federal regulation through the SEC
(Libby et al., 2003, p. 22; Pratt, 2003, p. 23; Reimers, 2003, pp. 9–11; Soffer and Soffer,
2003, pp. 7–13; Robinson et al. 2004, pp. 10, 11).
4.4 What Price Do We Pay for Ignorance of Accounting’s Institutional History? 47
book, Main Street and Wall Street (Ripley, 1927). Ripley’s article gen-
erated immediate and vigorous debate among financial news editors,
corporate executives, and professional accountants about the quality of
information available to the shareholders of publicly traded firms, and
Ripley was among a group of activists whose views influenced the pas-
sage of federal securities laws in the 1930’s (Hawkins, 1986, pp. 272–290
and 395, 396).
Ripley argued that, due to separation of ownership and control,
firms were operated more for managers’ self-interest rather than for the
interests of shareholders (Ripley, 1927, pp. 37, 38). Because managers
controlled financial reporting mechanisms, they could mask the effects
of this agency problem by manipulating accounting numbers, delaying
public disclosure, or withholding relevant information altogether. He
asserted that these practices harmed investors in part because equity
prices would not reflect available information, thereby creating an infor-
mational advantage for insiders (Ripley, 1927, pp. 204–207).17
Ripley testified in support of the Securities Acts, and helped shape
the thinking of those drafting these laws (Miranti, 1986, pp. 460, 461).
Foremost among them was lawyer Adolph Berle, who co-authored the
classic study on separation of ownership and control (Berle and Means,
1932). Berle had been Ripley’s student at Harvard, and was heavily
involved in drafting the Securities Exchange Act of 1934 (Parrish, 1970,
pp. 113, 114).18 Brief (1987, p. 147) concludes that Ripley’s impact
lasted beyond the 1930s by shaping perceptions that pre-SEC account-
ing data were routinely distorted and, hence, were largely useless for
equity pricing and investment decisions.
So, what evidence does Ripley provide to support his assertions
in his influential article and book? The 1926 article, which provided
the basis for chapter six of Main Street and Wall Street, only cites a
series of specific cases of alleged reporting weaknesses without providing
17 Ripley had articulated this view for years before his 1926 piece was published — see
Ripley (1911).
18 In the preface to the first edition of Berle and Means (1932), Berle writes: “All students
of these and allied problems, and we among them, owe a debt to Professor William Z.
Ripley of Harvard University, who must be recognized as having pioneered this area” (see
Berle and Means, 1932, liv).
4.4 What Price Do We Pay for Ignorance of Accounting’s Institutional History? 49
Table 4.2 Summary of criticisms of early 20th century financial reporting practices in 100
cases.
Number of Cases Cited 100
Number of Unique Firms Cited 69
Type of Company
Industrial Corporation 55 (80%)
Other (Interstate Railroad, Utility, Street Railway, etc.) 14 (20%)
Nature of citation
Unfavorable 72 (72%)
Favorable 28 (28%)
Period in which case occurred
After 1920 67 (67%)
Before 1920 33 (33%)
Cases with accounting or disclosure issue raised (n = 95)
Balance Sheet
Surplus/Impaired Capital from Excess Dividends 10
Assets: General 6
Assets: Large Intangible & Overcapitalization 6
Unconsolidated Subsidiaries: Liabilities, etc. 6
Assets: Low Intangibles 4
Disclosure: Balance Sheet &/or Components 4
Fixed Assets & Depreciation Reserve 1
Total Balance Sheet 37
Income Statement
Depreciation Expense 11
Disclosure: Income Stmt &/or Components 10
Interim earnings disclosure 8
Income manipulation & hidden reserves 6
Income manipulation: General 3
Treasury stock transactions 1
Total Income Statement 39
Other
Disclosure: General 19
Grand Total 95
Source: Chapter 6 of Main Street and Wall Street by W. Z. Ripley.
19 Nonetheless,his treatment of institutional details and data is far more cursory in Main
Street and Wall Street than in his earlier academic writings (Ripley, 1905, 1912, 1915).
However, it is likely that his audience was more familiar with the details of these recent
cases than readers almost a hundred years later.
50 Accounting History Research as a Stimulant to Thought Experiments
(untabulated) suggests that they are cursory — i.e., they occupy only
a sentence or two and omit overall materiality considerations in most
cases.
To be sure, Ripley cannot be faulted for providing no systematic
quantitative evidence that would directly address the effects of poor
accounting information on managerial and investor behavior in the
periods he considers. Still, it is quite troubling that accounting aca-
demics have been disproportionately influenced by Ripley’s and others’
beliefs about the value of regulation with little evidence to support such
beliefs (Benston, 1969; Sunder, 2005).20 This illustrates the dangers in
over reliance on the case method. Formal hypothesis tests using larger
samples are valuable in historical research because inferences from case
studies may not broadly generalize.
20 Benston (1969, pp. 517, 518) criticizes the weak evidence presented in the hearings that
led to passage of the Securities Exchange Act of 1934 and was later summarized in a
book by Ferdinand Pecora, the chief counsel of the US Senate Committee on Banking and
Currency (Pecora, 1939). Only recently have scholars skeptically examined the Securities
Acts from a public choice perspective, and Mahoney (2001 Abstract) concludes that the
net effect of the Securities Act of 1933 “was to reduce competition among investment
banks. In particular, the act protected separate wholesale and retail investment banks
from integrated investment banks.” Similarly, O’Connor (2004) argues that accountants
captured the SEC by extending the “best practice” of auditing to a broad mandate,
thus securing a professional monopoly and Partnoy (1999) describes the governmental
oligopolization of the credit rating industry at the same time.
4.5 How Can Researchers Become More Historically Informed? 51
Table 4.3 Alphabetical listing (by Author) of accounting and other historical sources.
A: Books on the General History of Accounting
Baxter, W. and S. Davidson (1962). Studies in Accounting Theory. This book contains
several essays that are classics in accounting. These include Hatfield’s “Historical
Defense of Bookkeeping,” Baxter’s “Recommendations on Accounting Theory,”
Yamey’s “The Case Law Relating to Company Dividends,” and many others.
Chatfield, M. (1974). A History of Accounting Thought. This excellent book provides a
summary of the history of accounting from its origins in ancient recordkeeping
practices up through modern standard-setting institutions.
Hausdorfer, W. (1986). Accounting Bibliography. Hausdorfer compiled several
bibliographies of historical accounting literature including lists of primary sources and
materials in several languages with their locations, lists of previous biographies and
bibliographies, and previous works on accounting history. It is about 50 years out of
date, but still a useful resource.
Littleton, A. C. (1933). Accounting Evolution to 1900. Littleton traces the development
of modern accounting in more depth than Chatfield. His chapters on the origins of
double entry, the appearance of financial statements in going concerns, and the
origins of British accounting regulation are especially informative.
Littleton, A. C. and B. S. Yamey (1956). Studies in the History of Accounting. This
beautiful book of essays includes marvelous pieces on the history of primitive
accounting devices like tally sticks, accounting rules of the British Companies Acts
before 1900, and early railroad accounting.
Mattessich, R. (2000). The Beginnings of Accounting and Accounting Thought.
Mattessich wrote several insightful papers on the development of accounting
technology in ancient Mesopotamia and accounting texts in ancient India that are
collected together in this volume.
Nissen, H. J., P. Damerow, and R. K. Englund (1993). Archaic Bookkeeping: Writing
and Techniques of Economic Administration in the Ancient Near East. This book
chronicles how basic accounting provided the impetus to the first human writing
systems observed in ancient Mesopotamia.
Parker, R. H. (1980). Bibliographies for Accounting Historians. This book reprints 8
bibliographies published during 1903–1977 and one unpublished bibliography.
Previts, G. J. and B. D. Merino (1998). A History of Accountancy in the United States.
This is the standard reference for understanding how accounting was used in colonial
America and then developed into its modern institutional form. The book does an
especially nice job of integrating the appearance of formal institutional structures
with the underlying social forces at play in broader American culture.
Schmandt-Besserat, D. (1992). Before Writing, Volume I: From Counting to Cuneiform.
Schmandt-Besserat’s work describes the archaeological basis for accounting in ancient
Mesopotamia. Her work proceeds from a careful description of the archaeological
evidence to present the hypothesis that ancient tokens were accounting devices that
served as the precursor to writing (also done for purposes of accounting).
ten Have, O. (1976). A History of Accountancy. An interpretive history of accounting
from ancient to modern times that is especially noteworthy in discussing the
transmission of accounting knowledge through time.
Zeff, S. A. (1972). Forging Accounting Principles in Five Countries: A History and
Analysis of Trends. This book is a thorough comparative study of the development of
accounting standards Canada, England, Mexico, Scotland, and the United States.
(Continued)
52 Accounting History Research as a Stimulant to Thought Experiments
56
5.1 Using Historical Data to Test Hypotheses: Examples from Economics 57
reference to the original tally stick and the record of remaining obliga-
tions (represented by the tally’s other half held by the Exchequer).
Watts and Zimmerman (1983) test agency cost explanations for
auditing by examining the original charters and by-laws of English
merchant guilds, merchant adventurers/regulated companies, and joint
stock companies from the 15th to the 17th century. Merchant guilds
appeared shortly after the Norman conquest of Britain in 1066 AD
and represent the earliest British incorporated enterprises. Merchant
adventurers and regulated companies that operated as foreign trade
monopolies appeared during the mid-13th century through the 17th
century. The first British joint stock company, the direct forerunner of
the modern corporation, appeared in 1555.
Watts and Zimmerman (1983) show that the charters and by-laws
of these organizations typically include clauses stipulating an audit of
the accounts. This suggests that auditing is an ex ante contracting
choice made when the firm is organized. In addition, teams of directors
and/or shareholders typically conducted the audit, which suggests that
audits were designed with foreknowledge that collusion between the
manager and auditor was possible. Fines and loss of reputation for
non-performance were used to promote independent audits and limit
manager–auditor collusion.
Watts and Zimmerman’s evidence is consistent with the hypothesis
that the origins of auditing result from inherent and powerful economic
forces rather than regulatory action. Their evidence is consistent with
others’ claims that auditing and internal control procedures are ubiq-
uitous to complex economic organizations (Chatfield, 1974).
single annual report until after its founder died in late 1907 (Chernow,
1998, p. 556; Porter et al. 1995). These firms’ common stocks were
actively traded and had large market capitalizations primarily based on
their dividend record. In contrast, many other firms published income
numbers voluntarily during the early 1900s, even though there was no
law requiring this.
Merino and Neimark (1982) argue that regulatory threats — e.g.,
investigations by the Congressional Industrial Commission of 1898–
1902 — provided a strong incentive for industrial firms to publicly
disclose more information about their activities. That is, politicians
clamored for greater publicity of corporate affairs to deter monopolies,
and in response firms increased disclosure to lessen political costs of reg-
ulatory intervention — e.g., prosecutions under the Sherman Antitrust
Act of 1890, one of which dissolved Standard Oil in 1911 (Chernow,
1998, pp. 537–559). Despite this argument’s intuitive appeal, it provides
an incomplete explanation since some firms were reporting extensive
information well before this time.
Another hypothesis is that some firms could not raise enough financ-
ing solely based on a stable dividend history, and hence, chose to dis-
close more information. Large-scale US industrial corporations began
appearing en masse in the 1890s as part of the “Great Merger Wave”
(Navin and Sears, 1955). Hawkins (1963, p. 140) suggests that these
new organizations adopted more transparent reporting policies to help
raise the large amounts of equity financing their expanding operations
needed (e.g., General Electric — see Section 4.2).2 Consistent with this
argument, Hawkins (1963) notes that investment bankers in the pre-
SEC era advocated increasing corporate transparency and disclosure.
But why did some firms voluntarily report income to raise equity
capital whereas others apparently believed that they could raise cap-
ital and sustain investor trust solely through a stable dividend pol-
icy? The answer likely reflects differences in these firms’ operations
and investment opportunities. Trusts like American Sugar Refining
and Standard Oil arose from horizontal mergers aiming to monopolize
2 Hawkins (1963) does not attribute increased disclosure after 1900 solely to the demand for
financing. Indeed, he suggests that regulatory threats exerted a major influence on firms’
willingness to disclose more information about their performance and financial position.
62 The Use of Historical Data in Testing Economic Hypotheses
5 The Jiang et al. (2005) sample includes firms with a pool initiating in 1928 or 1929 that
was focused on trading in common stock, that was mentioned in subsequent Congressional
hearings, and where a copy of the pool agreement was available in Congressional investi-
gators’ records stored in the National Archives (Jiang et al., 2005, pp. 152–155; Mahoney
1999). Mahoney (1999) provides an in-depth discussion of the contractual and institutional
arrangements governing 1920’s investment pools.
66 The Use of Historical Data in Testing Economic Hypotheses
6 Banerjeeand Eckard (2001) try to link their findings to differences in firm-specific per-
formance disclosures such as earnings. These attempts are hard to interpret since they
rely on measures of firm visibility rather than direct measures of corporate disclosures or
dividend policies for firms in that era, which vary considerably between firms (Sivakumar
and Waymire, 1993, 1994).
5.2 Using Historical Data to Study the Emergence of Accounting Institutions 67
7 They exclude 170 of the 710 NYSE firms in the Center for Research on Security Prices
(CRSP) database at the start of October 1929 that have idiosyncratic industry-specific
reporting practices or missing data.
8 Economic and political forces vary in their influence on different aspects of reporting
policy. The explanatory power of the overall quality model and the components models
is modest — e.g., adjusted R2 of 0.11 for the overall model. This suggests that economic
factors generally explain less than one-fourth of the cross-sectional variation in empirical
measures of financial reporting and disclosure policies, similar to models using modern
data (e.g., Botosan (1997) and Clarkson et al. (1999)).
68 The Use of Historical Data in Testing Economic Hypotheses
9 The Pennsylvania Railroad had the most shareholders at 29,000. Three other railroads had
over 10,000 shareholders, nine had between 5,000 and 10,000 shareholders, and another
21 had between 1,000 and 5,000 shareholders. However, railroad ownership was usually
highly concentrated with a few controlling investors owning large numbers of shares. In
fact, some very large railroads were owned by just a few individuals or corporations —
e.g., the Erie Railroad had only 17 shareholders.
10 Loree (1922, pp. 188–200) provides a succinct history of each of the major sources of U.S.
railroad information.
70 The Use of Historical Data in Testing Economic Hypotheses
considered the first stock market service in this country.” John Moody
followed suit in 1900 with a similar manual that compiled information
applicable to industrial securities, but provided no investment analysis
and recommendations.
Today’s investment analysis still leans heavily on financial ratios
that arose in this era. Management uses of ratios tended to focus
on operating efficiency and profitability. The mid-19th century rail-
roads developed the “operating ratio,” total operating expenses divided
by revenues, to measure railroad operating efficiency (Mundy, 1912,
pp. 31–34). In 1903, the DuPont Corporation began using the return
on capital invested as its internal measure of financial performance, and
a year later was calculating this measure on a monthly basis for each
of its 13 products (Chandler, 1977, pp. 445, 446). By World War I, the
company had disaggregated this ratio into the product of asset turnover
and profit margin, and further disaggregated these components, initi-
ating the DuPont method (Chandler, 1977, pp. 446–448; Davis, 1950,
p. 7; Horrigan, 1968, p. 286). Credit analysts focused more on balance
sheet measures of liquidity such as the current ratio, which were in
widespread use by 1900 (Horrigan, 1968, p. 285).
Investment recommendations based on financial statement analysis
began on a large scale when John Moody published Moody’s Analyses
of Railroad Investments in 1909. The initial Moody’s manual totaled
nearly 600 pages with the first third of the manual explaining the basis
for the ratings system and the extensive data (both quantitative and
qualitative) used in preparing individual ratings. Unlike current edi-
tions, early Moody’s Manuals rated preferred and common stocks as well
as bonds. Overall ratings (e.g., Aaa, Aa, and so forth) and their compo-
nents (e.g., quantified measures of safety and qualitative descriptions of
salability) were provided. Moody started with railroads because of their
market prominence and because their accounting reports were available
in standardized Interstate Commerce Commission reports and financial
press articles. Such was not the case for major industrial corporations.11
11 Sivakumar and Waymire (1993, 2003) describe evidence that far more accounting infor-
mation was available for railroads than industrials in this era. Differences in the quality
of financial disclosure between railroads and industrials of this era can be seen in the
5.2 Using Historical Data to Study the Emergence of Accounting Institutions 71
Commercial and Financial Chronicle, which printed the full text of corporate annual
reports in some cases.
12 The industrials and utilities were split into separate Moody’s manuals in 1920 and a
13 Thislong-held view is a prime example of what Stigler (1988) refers to a widely accepted
“fact” that is incorrect (Harris, 1994, p. 610).
74 The Use of Historical Data in Testing Economic Hypotheses
Table 5.1 Historical summaries of 19th century UK bankruptcy statutes, shareholder pro-
tection laws, and the number of practicing accountants.
Year(s)
# New law
Year Bankruptcy # passed
(Crisis commissions Bankruptcy after
in Bold) sealed petitions crisis Description of law(s)
Panel A: Chronological presentation of major 19th century UK bankruptcy laws in rela-
tion to the number of bankruptcies two years before & two years after law enacted
the Company” (Hein, 1963, p. 509), which led to the rise of the major
accounting firms like Price Waterhouse (Watts and Zimmerman, 1983,
pp. 628, 629), and empowered auditors to “examine” directors and
officers. The Act also introduced minority investor protection by giv-
ing one-fifth of the shareholders, by number and value, the right to
request the government Board of Trade to appoint inspectors to exam-
ine the company’s affairs (O’Connor, 2004, p. 761). The Punishment
of Frauds Act of 1857 criminalized fraud through falsification of or
tampering with books of account. The 1862 Companies Act added an
annual audit to the default provisions and introduced the concept of
an “audit committee” nominated by shareholders but which excluded
directors or officers (O’Connor, 2004, pp. 763, 764). The Companies Act
underwent many minor revisions thereafter until the 1900 Companies
Act mandated many of the “optional” default regulations of prior acts
including reinstating the requirement that an annual audited balance
sheet be disclosed (O’Connor, 2004, p. 764). The 1907 Companies Acts
required companies to submit annual reports to the Registry Office,
which had to include audited balance sheets (O’Connor, 2004, p. 767).
The data in panels A and B of Table 5.1 establishes a relation
between macroeconomic performance and accounting and auditing reg-
ulation that had consequences on the market for audits and financial
information. It also had a major effect on the supply of accounting
services. Panel C of Table 5.1 reproduces data on the number of accoun-
tants listed in London from 1799 and 1870 (Littleton, 1933). Of partic-
ular note is the more than tenfold increase in the number of practicing
accountants in London between 1820 and 1870.
The limited data from Littleton (1933) shown in Table 5.1 cannot
speak to the broader economic consequences of creditor and investor
legislation, but it does suggest that pure “public interest” explana-
tions are likely incomplete. Public interest theories of regulation have
been empirically tested on the first US state laws regulating security
Company’s Affairs.” The 1856 Act also strengthened the auditor independence provision
of the 1845 Act by requiring that (emphasis added) “No Person is eligible as an Auditor
who is interested otherwise than as a Shareholder in any Transaction of the Company;
and no Director or other Officer of the Company is eligible during his Continuance in
Office” (Hein, 1963, p. 509; O’Connor, 2004, p. 761).
78 The Use of Historical Data in Testing Economic Hypotheses
sales — i.e., the “Blue Sky” laws.15 The first Blue Sky law was adopted
by Kansas in 1911, and 47 of the 48 states had adopted such laws by
1931. These laws fall into three categories (Mahoney, 2003): (1) merit
review statutes where an appointed official was granted broad discre-
tion in approving all securities sold in the state, (2) ex ante fraud laws
where security sales required state pre-approval but the official had far
less discretion than under a merit review statute, and (3) ex post fraud
statutes that required no government pre-clearance but included anti-
fraud provisions. Kansas enacted a merit review law that was imitated
by eleven states in either 1912 or 1913 (no states enacted such laws after
1913). Twenty-five states enacted ex ante fraud laws between 1913 and
1923, while ten states passed ex post fraud statutes between 1912 and
1931 (Mahoney, 2003, Table 3.1).
Mahoney (2003) treats the Blue Sky laws as a naturally occurring
experiment that allows him to investigate two research questions. First,
why did some firms adopt Blue Sky laws early while others waited?
Second, why did different states adopt different types of Blue Sky laws?
Three hypotheses have been advanced to explain interstate differ-
ences in the timing and type of Blue Sky law adoption. The public
interest hypothesis says that these laws were a response to investor
fraud (Seligman, 1983, pp. 18–33). A second public choice hypothesis
implies that self-interested constituent groups lobbied state regulators
to adopt such laws because their private interests would be favored
(Macey and Miller, 1991). Under the public choice hypothesis, small
local banks actively sought to curtail competition from major corpo-
rations with publicly traded securities and the large investment and
commercial banks that sold these securities. A third political ideology
hypothesis suggests that states with Populist or Progressive movements
would be more likely to adopt certain types of legislation.16 In other
words, Blue Sky laws would be just one component of a broader leg-
islative agenda within such states.
15 Such a law was “popularly known as a ‘blue sky’ law, since it was intended to check stock
swindlers so barefaced they ‘would sell building lots in the blue sky’” (Seligman, 2002,
p. 44).
16 For historical research on the Populist and Progressive movements, see Hofstadter (1955),
Mahoney’s first set of tests examines why some states adopted Blue
Sky laws earlier than others. These tests find no positive association
between fraud incidences and when Blue Sky laws were adopted; indeed,
states with more frauds adopted Blue Sky laws later, which is incon-
sistent with the public interest hypothesis. The best predictor of when
a state passed a Blue Sky law is the extent of its Progressive political
influence. States with a greater securities industry presence tended to
write such laws later. Thus, the results from timing tests are more con-
sistent with the political ideology and public choice hypotheses than
the public interest hypothesis.
Mahoney (2003) examines why states adopted different types of
Blue Sky laws in a second set of tests. These tests indicate that the
public choice hypothesis best explains between-state differences in the
nature of Blue Sky law adopted; states with a large number of small
banks preferred merit review statutes whereas states with greater secu-
rities industry presence did not prefer such laws. Again, there is no evi-
dence that the incidence of fraud is positively associated with the type
of statute enacted, suggesting that either the public choice hypothesis
is invalid or that the fraud incidence measure is noisy.
17 The regulatory scheme in the United States was no doubt strongly influenced by British
securities legislation enacted in the 19th century (see Section 5.1). There are differences in
the objectives of the British and American systems and accounting practice in Britain is
frequently described as principles-based while US practice is labeled as more rules-based.
Differences between the systems are discussed and debated in Bush (2005) and Benston
(2008) (see also Benston (1976)).
84 The Use of Historical Data in Testing Economic Hypotheses
non-NYSE issues, it was present for all issues regardless of where the
issue subsequently traded.
Other studies have examined the effects of recurring disclosure
under the 1934 Act. Benston (1973) compared changes in stock return
volatility around 1934 for two sets of NYSE firms: (1) “Disclosers” (i.e.,
those disclosing annual sales prior to 1934), and (2) “Non-disclosers”
(those not disclosing annual sales prior to 1934). Benston (1973) found
no significant differences in return volatility changes between these
firms. This evidence, combined with his review of other academic stud-
ies, court cases, and pre-SEC disclosures, led him to conclude that “the
disclosure requirements of the Securities Exchange Act of 1934 had
no measurable positive effect on the securities traded on the NYSE”
(Benston, 1973, p. 153).
As with Stigler’s (1964) paper, proponents of SEC disclosure
mandates have criticized Benston’s study (e.g., Seligman (1983)).
Not surprisingly, subsequent studies suggest that the SEC’s disclo-
sure mandates may have had more complex effects. Greenstone et al.
(2006) document that market values of over-the-counter (OTC) firms
increased when the SEC mandated new disclosures for them in 1964.
Distinctively, the paper identifies firms that had not previously dis-
closed information meeting the new SEC requirements and shows
that they experienced large stock price increases when they complied.
OTC firms also had a dramatic reduction in stock return volatility
around the SEC’s mandatory disclosure requirements in 1964 (Ferrell,
2007).
Bushee and Leuz (2005) examine the 1999 extension of SEC dis-
closure rules to firms traded on the Over-the-Counter Bulletin Board
(OTCBB). This SEC regulation led over 2,000 noncompliant firms to
leave the OTCBB, reducing their market values. Already compliant
firms experienced increased returns and liquidity when the new SEC
rules were announced. Interestingly, previously noncompliant firms that
improved disclosure to remain on OTCBB experienced increased liq-
uidity but negative stock returns. The Bushee and Leuz results are
provocative because they suggest that while mandating disclosure
improves a market’s liquidity, it also reduces shareholder wealth for
many firms, especially those driven out of the market.
5.3 Using Historical Data to Study the Effects of Accounting Regulation 85
18 In1917, the AIA worked with the Federal Trade Commission to develop a set of account-
ing and auditing guidelines that was published by the Federal Reserve Board (1917), but
firms were not required to follow these guidelines (see Moonitz (1970)). A revision was
published in 1929.
5.3 Using Historical Data to Study the Effects of Accounting Regulation 87
prices firms’ securities are more useful in this regard (Ball and Brown,
1968).
Ely and Waymire (1999a) estimate the following cross-sectional
model for each of their 67 annual samples:
ri = γ0 + γ1 ∆Ei + γ2 Ei + εi , (5.1)
Table 5.2 Summary of estimation results for yearly models (ri = γ0 + γ1 ∆Ei + γ2 Ei + ε∗i )
of the earnings-return relation across different accounting standard setting regimes for yearly
NYSE samples, 1927–1993 from Ely and Waymire (1999a).
Pre-CAP CAP APB FASB
(1927–1938) (1939–1959) (1960–1973) (1974–1993)
Median of annual 15.7% 23.1% 13.3% 13.9%
adjusted R2
Median of annual sum 0.81 1.25 2.37 0.86
of coefficients
(γ1 + γ2 )
Median of annual % of 22.3% 8.0% 7.0% 14.6%
losses
∗ r equals the 16-month market adjusted return on security i, ∆E equals the change in
i i
annual earnings scaled by beginning of period price for security i, and Ei equals the level
of annual earnings scaled by beginning of period price for security i. The regression model
is estimated using cross-sectional data for yearly samples of NYSE common stocks for each
of the 67 years between 1927 and 1993, inclusive.
90 The Use of Historical Data in Testing Economic Hypotheses
Ei = α0 + α1 Di + β0 Ri + β1 Ri Di + ei . (5.2)
19 TheGreat Depression causes the higher loss frequency. For instance, the percentage of
firms with negative net income in the 1931, 1932, and 1933 samples equals 39%, 63.6%,
and 47.5% respectively.
5.3 Using Historical Data to Study the Effects of Accounting Regulation 91
Table 5.3 Summary of estimation results for yearly models (Ei = α0 + α1 ∆Di + β0 Ri +
β1 Ri Di + e∗i ) of Basu (1997) asymmetric timeliness regression model reported in
Holthausen and Watts (2001, Table 2).
Mean Basu Mean of annual
Reporting regime as described timeliness median market
Period by Holthausen and Watts measure (β1 ) to book ratio
1927–1941 Prestandard-setting, low 0.09∗∗ 0.77
litigation
1942–1946 Price controls, standard-setting, −0.32 1.03
low litigation
1947–1950 Standard-setting, low litigation 0.01 0.90
1951–1953 Price controls, standard-setting, 0.07 0.93
low litigation
1954–1966 Standard-setting, low litigation 0.06∗∗ 1.31
1967–1975 Standard-setting, high litigation 0.05∗∗ 1.30
1976–1982 Standard-setting, litigation 0.16∗∗ 1.01
1983–1993 Standard-setting, high litigation 0.43∗∗ 1.67
∗ E equals the 16-month market adjusted return on security i, ∆D equals the level of
i i
annual earnings scaled by beginning of period price for security i, Di equals one if Ri is
negative, and ei is a residual. The regression model is estimated using cross-sectional data
for yearly samples of 100 NYSE common stocks for each of the 67 years between 1927 and
1993, inclusive. Ely and Waymire (1999a) collected the regression data originally.
∗∗ Significantly different from zero at .05 level or better with one-tailed test.
have computed annual median market-to-book ratios from the Ely and
Waymire (1999a) samples in constructing this column.
Two aspects of these findings are relevant for present purposes. The
first is that significant conservatism is present for the earliest period
they examine (1927–1941). This is consistent with the hypothesis that
accounting practice reflected longstanding principles even before the
advent of modern standard setting. The second is the presence of con-
servatism in all periods after 1954. This is also evident in the levels of
market-to-book ratios, which reach their maximum value in the 1983–
1993 sub-period.20 This is consistent with conservatism being a deeply
ingrained feature of practice that is difficult for standard setters to
eliminate (Watts, 2003).
20 The significantly higher conservatism in the most recent period also helps explain the
lower R2 s that Ely and Waymire (1999a) find for the FASB era because their linear
regression model is less well specified for this period. Ryan and Zarowin (2003) show
that the apparent secular decline in R2 after 1950 is completely explained by increasing
conservatism and greater market efficiency (increase in prices leading earnings).
92 The Use of Historical Data in Testing Economic Hypotheses
94
6.1 An Evolutionary Story of Accounting 95
1 Human cultures that develop better capabilities to adapt their behavior through routines
and rituals will more likely survive and procreate (Donald, 1991). Thus, human ability to
perform complex tasks is enhanced if somebody learns how to store information external
to the brain so as not to be forgotten. For example, knowledge stored in a book (e.g.,
The Idiot’s Guide to Finding and Decorating a Bigger and Better Cave) can be passed
on to the future generations or other cultures, even if that was not the author’s original
intent.
96 Future Research: Toward an Evolutionary Theory of Accounting
2 In the United States, these processes became far more important around 1900 when an
accountancy profession gained increasing social acceptance (Miranti, 1986; Previts and
Merino, 1998).
3 Accounting has no analogue to marketing research that allows for the ex ante evaluation
of the consequences of a specific policy. This is likely one reason that empirical models of
accounting choice have such low explanatory power; see Fields et al. (2001) for a review of
accounting choice research. Similarly, the lack of direct ex post feedback allows standard-
setters to use inaccurate subjective models as the basis for “intelligent design” attempts for
long periods of time, and only large crises cause regulators to attempt to learn and improve
their models (see Sargent (2008) for an insightful historical and theoretical analysis of
learning by monetary authorities).
6.1 An Evolutionary Story of Accounting 101
4A parallel can be drawn to the costs of excessive traffic signals and uniform speed limits,
whose often arbitrary enforcement can cause drivers to watch out for signs rather than
monitor traffic conditions when they drive (Staddon, 2008). Several European cities have
removed all traffic signs from their roads and report fewer fatal traffic accidents and better
traffic flow as a result (e.g., Schulz (2006) and Neate (2008)).
6.1 An Evolutionary Story of Accounting 103
We presently know little about the forces that have shaped account-
ing’s evolution. Thus, our focus in the next sub-section will be to iden-
tify areas for future research using the six “big picture” questions that
we posed earlier in the survey.
5 The ability of humans to form contracts is unique relative to other species (Wilson, 1998,
pp. 186, 187).
6.2 Future Research on Accounting Evolution 105
100%
D: Credit 32.5%
20%
0%
Community Size
Fig. 6.1 Cumulative percentage of SCCS Societies where a particular economic institution
is present plotted against community size (from Basu et al. (2008b), Figure 4a).
110 Future Research: Toward an Evolutionary Theory of Accounting
Basu et al. (2008a,b) provide fairly strong support for Basu and
Waymire’s (2006) prediction that the basic accounting institution of
recordkeeping fundamentally alters the nature of an economy and pro-
motes its further expansion. There are of course a number of other
empirical analyses that could be conducted (see Basu and Waymire
(2006, pp. 218–223)). We regard the agent-based simulations and neu-
roscientific experiments as being especially promising because they rep-
resent new research techniques and data that enable investigation of
other aspects of the same research question. Because these methods
can also be used to analyze the effects of double-entry bookkeeping and
evolved accounting principles, we defer discussion of these approaches
to subsequent sub-sections.
and (3) there is little a priori reason that accounts would be useful
for deciding whether to start a new organization, enter a new market,
or expand within an existing market (see Yamey (1964, pp. 119–133)).
Most (1972) agrees with many of the specifics of Yamey’s argument, but
reaches an overall different conclusion after reformulating Sombart’s
hypothesis in light of assertions by Schumpeter that double entry was
a device to deal with cognitive constraints inherent to fundamentally
uncertain economic environments (Most, 1972, pp. 731–733).6 Most
(1972) takes the position that Sombart’s hypothesis asserts that double-
entry bookkeeping fulfills a planning function that is more fundamental
than other functions it came to serve (e.g., stewardship and valuation).
Most’s (1972, p. 734) restatement of Sombart’s hypothesis is
(emphasis in original):
6 In
constructing this argument, Most (1972) relies on arguments by Frank Knight (1921)
and John Maynard Keynes (1936) about the economics of fundamental uncertainty.
6.2 Future Research on Accounting Evolution 115
7 Yamey (1964, p. 127) also recognized the flexibility of double entry but significantly down-
played its importance. Yamey (2005) critiques the arguments and evidence in Aho (1985)
and Carruthers and Espeland (1991), and remains steadfastly immune to any “Sombartian
intoxication.”
116 Future Research: Toward an Evolutionary Theory of Accounting
8 Ofcourse, Sombart’s hypothesis implies more generally that accounting provides informa-
tion useful to any economic actor besides management in “making sense” of an organi-
zation’s performance within its local environment. An implication of this “sense-making”
view would be that accounting data could be useful to a security analyst drawing infer-
ences based on qualitative analysis, which would provide the basis for forecasting earnings
necessary for equity valuation tasks (Palepu and Healy, 2007).
6.2 Future Research on Accounting Evolution 119
where tax and regulatory factors are absent? In this case, it is harder to
evaluate how a specific accounting method affects the firm’s prospects
for survival and resource acquisition. Clearly the portfolio of account-
ing methods and judgments must matter, but how much a particular
method is preferred is less obvious. Absent tax or regulatory forces, it
seems that the choice of straight-line or accelerated depreciation just
does not matter as much as say how the firm prices its products or
invests in research and development or whether or not it closes an
existing production facility. Stated differently, the signal-to-noise ratio
regarding the effect of a specific accounting method is very low in the
absence of regulatory or tax effects.
The obvious retort to this line of thinking is that it must matter to
some degree, and if we could only figure it out, we could understand
better why firms choose specific accounting methods. Stated differently,
if the net benefits of an accounting method were more obvious to deci-
sion makers, then researchers would be able to empirically identify such
effects. Or, even if decision makers were not very well informed, mar-
ket forces would render organizations that make inferior choices extinct.
Fields et al. (2001) report that our empirical models of specific account-
ing method choices perform poorly. One possible explanation is that
the typical rational choice theories are not descriptive because deci-
sion makers cannot identify which methods will produce better results.
Instead decision makers adopt mental models that help rationalize prior
successes and failures but that will only be revealed to be inaccurate
in novel states of the world. Shared subjective models produce self-
confirming equilibria that generate “rational” actions in normal states
but generate dysfunctional actions in unusual situations.9
The question thus becomes: how do “new” accounting methods orig-
inate and then subsequently spread in use? The most likely answer lies
in new forms of transactions. For example, consolidated reporting in
the United States grew out of the experience with the creation of new
large entities created by mergers in the latter part of the 19th century.
9 deFigueireda et al. (2006) use inaccurate mental models and self-confirming equilibria to
explain several historical puzzles regarding the American Revolution. Sargent (2008) uses
a similar framework to understand recurrent failures in monetary policy over hundreds of
years.
120 Future Research: Toward an Evolutionary Theory of Accounting
10 This search could be internal rather than external. Early railroad depreciation methods
such as retirement and replacement accounting are similar to FIFO and LIFO inventory
accounting, and were only gradually replaced by systematic depreciation methods.
11 Within the context of genetic transmission and natural selection, some have argued that
docile and altruistic behaviors essentially “hitchhike” off other genes that enhance per-
sonal fitness (Gintis, 2003; Simon, 1990).
12 In conversations at the 2007 AAA annual meeting, archaeologist Denise Schmandt-
Nelson et al. (2002) report survey evidence from auditors that earn-
ings management through transaction structuring is more prevalent for
bright-line accounting standards. This research can be extended using
experiments that examine how transaction structuring to avoid rigid
standards and regulation lead to economic inefficiencies.14 For instance,
trust game experiments that allow for interaction between actors but
require standardized financial reporting might be designed to evalu-
ate the effects of historical cost accounting (representation based on
the firm’s local conditions) versus fair value accounting (representation
based on more global conditions presumed to be applicable to all firms).
14 Lys and Vincent (1995) estimate that AT&T spent between $50 million and $500 million
to satisfy pooling accounting requirements when it acquired NCR in 1991, and conclude
that this value-destroying merger was at least partially related to the 1984 consent decree
with the Department of Justice that led to the breakup of AT&T. Case studies can also
shed light considerable light on these questions.
15 Easterly (2008) labels these approaches as “bottom-up” and “top-down” in distinguishing
16 The problem here is one of a low signal-to-noise ratio independent of tax and regulatory
effects. This is generally consistent with the weak pattern of association that emerges in
the empirical literature on accounting method choices (Fields et al., 2001).
17 Radner (1995) demonstrates that action based on a survival motive can have very differ-
ent implications for aggregate economic outcomes in a world of uncertainty than profit
maximizing actions.
126 Future Research: Toward an Evolutionary Theory of Accounting
18 Others note much earlier links between accounting practices and subsequent principles
such as the entity theory (Littleton, 1933, pp. 183–204).
6.2 Future Research on Accounting Evolution 127
et al., 2006).
6.2 Future Research on Accounting Evolution 129
(Barton and Waymire, 2004; Johnson et al., 2000; Mitton, 2002). John-
son et al. (2000) and Mitton (2002) investigate the East Asian crisis of
1997–1998 and Barton and Waymire (2004) examine the stock market
crash of 1929.
The 1997–1998 East Asian crisis affected equity securities in Asian
emerging markets. The primary issue associated with this crisis was
a loss of investor confidence that triggered a large net capital outflow
from these economies. This, in turn, created strong incentives for man-
agers of firms in these economies to expropriate shareholder wealth by
diverting corporate assets for their personal use. Johnson et al. (2000,
pp. 145–151) use a static model to hypothesize that a loss of investor
confidence is associated with a lower expected return from investment
in a particular firm, which in turn increases incentives for expropri-
ation by the manager. Thus, their hypothesis is that the effect of a
crisis will be greater when shareholder protection from expropriation
through governance institutions is weaker. They use several variables as
empirical proxies for governance, one of which is the presence of more
stringent accounting standards in the economy.
Johnson et al. (2000) estimate a series of regression models using
the country level measures employed previously by La Porta et al.
(1998). They find evidence that several of the governance variables
help explain cross-country differences in the 1997–1998 exchange rate
depreciation and stock market decline in 25 emerging market coun-
tries. Their results, however, provide no support for the hypothesis that
stronger accounting standards are associated with these cross-country
differences.
Mitton (2002) extends the Johnson et al. (2000) study to more
closely examine the relation between accounting policies and stock price
performance during the East Asian crisis at the individual firm level.
Mitton’s sample includes 398 firms in five specific countries affected by
the East Asian crisis: Indonesia, Korea, Malaysia, the Philippines, and
Thailand. Mitton’s empirical proxies for disclosure quality are based
on whether the firm (1) has an ADR listed in the United States at
the start of the crisis (10.3% of firms; N = 41), or (2) is audited by a
Big Six auditor (29.6% of firms; N = 118). The motivation for these
proxies is that firms with an ADR comply with the more stringent US
130 Future Research: Toward an Evolutionary Theory of Accounting
21 Theaverage market decline of the firms in Mitton’s sample over the entire crisis was
−68.7% (Mitton, 2002, p. 220).
6.2 Future Research on Accounting Evolution 131
22 Barton and Waymire (2004, Table 12) find a statistically significant relation between
October 1929 returns and proxies for noisy fundamentals and the availability of divi-
dends as an anchor for valuation, providing additional evidence consistent with their
predictions.
132 Future Research: Toward an Evolutionary Theory of Accounting
25 The propensity of humans to believe they can control complex systems has been termed a
“fatal conceit” by Frederich Hayek (1988) that often produces “unintended consequences”
(Merton, 1936). This suggests that accounting research on regulating a complex system
will likely be interdisciplinary and involve research on how policymakers write rules, the
effects of such rules, and how policymakers alter the very language of discourse that affects
the academic research produced to evaluate the consequences of regulation (White, 2005).
7
Concluding Remarks
136
137
139
References
140
References 141