Threshold citation analysis of influential articles, journals, institutions and researchers in accounting

2.1. Publication-based approach

The publication-based approach, as used in the ranking of accounting departments and authors, relies on the counting of articles that have appeared in a set of accounting journals. Zivney et al. (1995) provide an early comprehensive study of institutional ranking. Chan et al. (2007) examine a global institutional ranking and find that authors whose degrees are from elite PhD-granting institutions tend to publish disproportionately more articles in leading accounting journals relative to other authors. Other studies offer different perspectives within the counting approach. For instance, Rama et al. (1997) and Kirchmeyer et al. (2003) consider the gender difference in accounting research by counting the number of articles by male vis-à-vis female accounting researchers. Chung et al. (1992) examine the empirical distribution of the accounting research output with respect to institutions and authors.

To differentiate between the quality among different journals, the literature also provides alternative rankings by limiting the scope to a few selected premier accounting journals. For instance, Chan et al. (2007) also examine an alternative institutional ranking using the 5 premier accounting journals (Accounting, Organizations and Society, Accounting Review, Contemporary Accounting Research, the Journal of Accounting Research and the Journal of Accounting and Economics). There are two major assumptions in this publication-based approach. First, the quantity of research is assumed to also represent the quality of research, which might or might not be the case. In other words, this approach does not distinguish influential research works that have been frequently cited from those that are less frequently cited even though the two research works might have been published in the same publication outlet. Second, as suggested by Reinstein and Calderon (2006), researchers need to impose a subjective decision to include only a pre determined set of accounting journals in their database. Therefore, the publication-based approach ignores influential research works in other publication outlets, such as book chapters, conference proceedings, monographs or journals in other disciplines.

2.2. Citation-based approach

The citation-based approach explicitly considers the quality of research as the ranking criterion. The approach uses citation as a quality marker. That is, a high-quality accounting article is the one that received more citations. The approach has been widely used in ranking journals, institutions and authors in accounting. Sriram and Gopalakrishnan (1994) use citations to rank accounting doctoral programmes and their graduates. Seetharaman and Islam (1995) examine journal rankings based on citations received by respective journals. Brown (1996) identifies influential accounting articles, authors, PhD-granting institutions and programmes by studying the citations of the ‘top 100’ articles in the Accounting Research Directory.

Citation studies in other disciplines often rely on the citation data from the Social Science Citation Index (SSCI). However, the accounting journals coverage in the SSCI is rather limited. It is not surprising to learn that authors in accounting citation studies use their own developed database and, therefore, the databases are usually limited in scope of citation coverage. As discussed in Reinstein and Calderon (2006), the first citation study in accounting, McRae (1974), examines only the citations of 17 accounting articles. Brown (1996) studies the citations of ‘top 100’ accounting articles. In addition, these studies do not distinguish self-citations by the authors and ‘negative’ citations. These limitations undermine the accuracy of the citation-based approach in the ranking of accounting departments, authors and journals.

2.3. Survey approach

In the survey approach, opinions regarding journal and institution quality are obtained. Examples of the survey approach include Schroeder et al. (1988), Brown and Huefner (1994), Ballas and Theoharakis (2003) and Herron and Hall (2004). The logic behind the survey approach is that expert opinions are valuable and that perceptions matter. Opinion surveys suffer from possible sample bias and possible questionnaire design flaws. Most of the time, the ‘experts’ have different perceptions about quality (Lowe and Locke, 2005).

2.4. Derived-product approach

The derived-product approach has several strands. The ranking criterion includes using the placement record of PhD graduates (see Hasselback and Reinstein, 1995; Stammerjohan and Hall, 2002), representation in editorial board memberships (Mittermaier, 1991), and the number of times a research work is downloaded (see Brown, 2003; Brown and Laksmana, 2004). Although these ranking methods are logical, they also have limitations. For schools without PhD programmes in accounting, the publication record of PhD graduates does not apply to non-doctoral-granting institutions and, hence, there is no way of gauging the performance of non-doctoral-granting schools. Editorial membership representation, although also widely used in other disciplines, penalizes accounting departments whose faculty members might be too busy with their own teaching, research, services or other consulting activities to serve on the editorial board of journals (Chan and Fok, 2003). The electronic download method is novel but is confined to research in journal form. All these drawbacks limit the use of the derived-product approach to rank accounting departments.

In summary, although no approach is perfect, the threshold citation approach, which incorporates the quantity of research and the quality of research, offers a viable alternative to mitigate the limitations of the quantity-orientated publication-based approach and the quality-orientated citation-based approach.

4.1. General

Tables 1–3 provide the descriptive information on these influential research works. Table 1 classifies these frequently cited works by disciplines of the publication outlets. The top 3 disciplines influencing accounting research are accounting, finance and economics. As expected, 7220 citations (more than 77 per cent) of these frequently cited works are in accounting. The second most cited field in accounting research is finance, followed by economics. In other words, more than 93 per cent of the frequently cited research works in accounting is from these 3 disciplines.

Table 2 provides the specific publication outlets of these high-impact research works. The 957 frequently cited research works are published in 70 different publication outlets. To conserve space, we only list the outlets that are cited 10 or more times.² The top 5 publication outlets to look for influential research works in accounting research are the Journal of Accounting Research, Journal of Accounting and Economics, Accounting Review, the Journal of Finance and the Journal of Financial Economics. Specifically, 204 of the 957 frequently cited works are published in the Journal of Accounting Research, and this journal is cited 2126 times (22.79 per cent of all citations). The Journal of Accounting and Economics, Accounting Review, the Journal of Finance and the Journal of Financial Economics are cited 2055 times (22.03 per cent), 1552 times (16.64 per cent), 474 times (5.08 per cent) and 472 times (5.06 per cent), respectively. Accounting, Organizations and Society and Contemporary Accounting Research follow with 437 times (4.68 per cent) and 396 times (4.25 per cent), respectively. Although most of the publication outlets are academic journals, there are some interesting exceptions. The Handbook of Labor Economics and Fortune are cited 21 and 16 times, respectively. Table 2 also illustrates the limitation of counting publication output in a predetermined set of accounting journals within a given time period. For instance, by only concentrating on articles in leading accounting journals, important papers published in Econometrica and the Journal of Political Economy, among others, are neglected. Hence, the results in Table 2 suggest that there are indeed other outlets that published frequently cited research works in accounting research.

Table 3 identifies the 25 most frequently cited publications in the 5 accounting journals from 2000 to 2004.³ The 4 most frequently cited articles are: ‘A heteroskedasticity-consistent covariance matrix estimator and a direct test for heteroskedasticity’ by Halbert White (1980) in Econometrica; ‘Earnings, book values, and dividends in equity valuation’ by James A. Ohlson (1995) in Contemporary Accounting Research; ‘Earnings management to avoid earnings decreases and losses’ by David C. Burgstahler and Ilia D. Dichev (1997) in the Journal of Accounting and Economics; and ‘The effect of bonus schemes on accounting decisions’ by Paul M. Healy (1985) in the Journal of Accounting and Economics. These 4 articles are credited with 76, 60, 49 and 49 citations, respectively. These papers have a theoretical focus serving as foundations to accounting research. Of the top 25 frequently cited research works, 20 are published in accounting journals. The remaining influential research works are published in finance and economic journals. This highlights the fact that accounting has a strong foundation.

There are several additional noteworthy findings in Table 3. First, although some of the top 25 frequently cited research works are published in the 5 premier accounting journals (8 articles are published in the Journal of Accounting and Economics, 5 articles in the Journal of Accounting Research, 4 articles in Accounting Review and 2 articles in Contemporary Accounting Research), there are exceptions. An article in Econometrica ranked 1st, a Journal of Financial Economics article ranked 5th, an article in the Journal of Political Economy and 1 article in the Bell Journal of Economics tied for 17th, and a Journal of Finance article ranked 20th. Finally, the publication year ranges widely, from 1973 for an article by Eugene F. Fama and James D. MacBeth in the Journal of Political Economy to 2000 for an article by Ray Ball, S. P. Kothari and Ashok Robin in the Journal of Accounting and Economics.

4.2. Ranking of institutions and authors

Table 4 provides the ranking of the top 50 institutions in accounting research based on the impact of research works produced by their staff or faculty members. The threshold citation approach weights both the number of authors and the number of citations in generating the ranking and has a value ranging from a high of 785.25 for the University of Chicago to a low of 36.83 for the University of Missouri. The top 3 accounting programmes associated with frequently cited research works are the University of Chicago, the University of Pennsylvania and the University of Michigan. In comparison, Chan et al. (2007) show that the top 3 accounting departments using the publication-based approach of the same 5 premier accounting journals are the University of Pennsylvania, the University of Chicago and the University of Michigan.

Faculty members at the University of Rochester are credited with 73 high-impact research works and a threshold citation index of 491.33. Although scholars at Harvard University published 5 more frequently cited research works (78 distinct research works) than scholars at the University of Rochester, the threshold citation index for Harvard University is lower than the index for the University of Rochester. The main reason is that research works produced by faculty members at the University of Rochester are cited more frequently in the 5 accounting journals over a 5 year period (2000–2004) than those produced by faculty members at Harvard University.

There are several additional interesting results from Table 4. First, although US institutions dominate the top 20 institutions, 1 non-US university is credited with producing influential research works: the University of British Columbia, with a ranking of 17th. Further examination of the top 50 list reveals that scholars at 6 institutions outside the USA (the University of British Columbia, the University of New South Wales, the University of Alberta, Hong Kong University of Science and Technology, the University of Lancaster and Macquarie University) have engaged in high-impact research works, a respectable presence by non-US institutions. Second, unlike other disciplines, such as economics, finance, insurance or real estate, the top 50 influential institutions in accounting are all academic institutions. Finally, the threshold citation index in Table 4 is skewed towards top-ranked institutions. The implication is that it takes a lot of effort for an institution to move up in the ranking. To move 10 spots from the 50th position to the 40th position, an additional 14.17 citations would be required. However, to move from the 30th position to the 20th position, an additional 35.67 citations would be required. This indicates the substantial commitment of financial and human resources to produce the quantity of frequently cited research works needed to be highly ranked.

Table 5 presents the top 50 frequently cited authors using the threshold citation index, which simultaneously weights the number of authors and the number of citations in generating the ranking. The threshold citation index for the top 50 frequently cited authors has a range from 141.08 for Richard G. Sloan to 39.00 for Ravi Bhushan. The top 3 influential authors in accounting that are frequently cited in the published literature are Richard G. Sloan, Robert E. Verrecchia and Paul M. Healy.