1 Introduction

Institutional ranking has attracted attention because it signals the quality of an educational program, draws research funding, and is influential on the recruitment of faculty and students. Various media such as the U.S. News & World Report, Business Week, Forbes, the Financial Times, Money, and Fiske Guide to Colleges have undertaken statistical and reputational rankings of colleges to provide information to the public. Increasingly, the importance and validity of college rankings has become a hotly debated issue.

While ranking agencies use different instruments to derive the ranking, faculty research productivity always, directly or indirectly, plays a significant role in the determination of overall academic reputation. Studies in research productivity are abundant. For example, Hasselback and Reinstein (1995), Brown (1996), and Stammerjohan and Hall (2002) examine rakings in accounting; Niemi (1987), Alexander and Mabry (1994), and Borokhovich et al. (1995) in finance; Conroy et al. (1995), Scott and Mitias (1996), and Collins et al. (2000) in economics. All of these studies, however, focus their rankings on the North America institutions. Few exceptions, such as Chan et al. (2005a, b) study accounting and finance research output in Asia-Pacific countries, respectively. To be sure, the ranking competition goes beyond the US, and has attracted international attention. Footnote 1 The interest in global ranking is not without precedence. For example, the Financial Times publishes its world rankings of the top 100 MBA programs annually. The globalization of the world’s economies definitely stimulates a stronger interest in a worldwide ranking for academic institutions.

The objective of this study is twofold. First, we study the global research output in finance during a 15-year period from 1990 to 2004. Unlike previous studies in the literature, our sample extends to a larger set of finance journals and a considerably longer period. We rank the production of finance literature by countries. Since productivity varies significantly across countries, we attempt to find the sources of such research productivity variations among different countries. Second, we rank the production of finance literature by academic institutions. This global ranking allows an institution to find its academic ranks in the world during the study period. We then examine the cross-institutional variations of research productivity among academic institutions.

The results of our research offer two major conclusions: First, based upon 21 finance journals, the top five most productive countries are in the following order: U.S, U.K., Canada, Hong Kong, and Australia. Footnote 2 We find that per capita GNP, English-speaking country factor, and rule of laws are significantly related to the production of finance literature. Second, New York University, the University of Pennsylvania, Harvard University, the University of Chicago, and UCLA take the top five spots among the 1,126 academic institutions with most JF-pages appearing in 21 finance journals during the fifteen-year period from1990 to 2004. The share of the U.S. institutions among the top-100 programs is overwhelming; 78 out of the top-100 institutions are U.S. universities. Cross-institutional variations in finance research can be explained by their financial resources, faculty size, and research catalyst effect.

The rest of the paper is organized as follows: Section 2 shows the data source and methodology. Section 3 presents descriptive statistics. Ranking by countries is reported in Sect. 4, and ranking by institutions in Sect. 5. Section 6 concludes.

2 Data and methodology

We manually collect all data from the hard copies of a set of 21 core finance journals for a period of 15 years from 1990 to 2004. The data include authors’ names, their affiliations, the country origin of the institutions, and the length of the article. The set of 21 finance journals are: Financial Analysts Journal, Financial Management, Financial Review, Journal of Banking and Finance, Journal of Business, Journal of Corporate Finance, Journal of Empirical Finance, Journal of Finance, Journal of Financial Economics, Journal of Financial and Quantitative Analysis, Journal of Financial Markets, Journal of Financial Research, Journal of Portfolio Management, Journal of Financial Intermediation, Journal of Financial Services Research, Journal of Futures Markets, Journal of International Money and Finance, Journal of Business Finance and Accounting, Review of Financial Studies, Pacific-Basin Finance Journal, and Review of Quantitative Finance and Accounting.

We select these 21 journals for several reasons. First, 16 of the 21 journals have been used in earlier research such as Chan et al. (2002) and Heck and Cooley (2005). The five new journals (Journal of Corporate Finance, Journal of Empirical Finance, Journal of Financial Markets, Pacific-Basin Finance Journal, and Review of Quantitative Finance and Accounting) have received attention from faculty members as quality research outlets but were not included in prior research. Second, these 21 journals include journals with a general scope (e.g., Journal of Finance) as well as specialized journals, such as Journal of Financial Markets, Journal of Futures Markets, and Journal of Corporate Finance. The inclusion of journals with general scope and journals with specialty takes into account the research productivity of faculty with different research interests and expertise. These journals are considered influential not only by the Americans, but also by the European and Asian authors as suggested in Oltheten et al. (2005).

Similar to Chan et al. (2002), we have four potential caveats in this database. First, not all authors in these 21 journals are finance faculty. While finance faculty members write a vast majority of the articles published in these journals, authors from other disciplines, such as accounting and economics, also contribute to the finance literature. Since often authors’ departmental affiliations are typically not specified in these journal articles, it is cost prohibitive to classify authors based upon their disciplines. Most importantly, there is no reliable source to identify the departmental affiliations of all authors during the entire 15-year period. Therefore, this study may overstate the ranks of certain finance departments. Nevertheless, as Chan et al. (2002) have argued, the impact of this bias, if any, is negligible because publications of finance articles by authors from other disciplines within the same institution also enhance the reputation of the finance program in that same institution.

Second, some journals such as Journal of Business, Review of Quantitative Finance and Accounting, and Journal of Business Finance and Accounting publish non-finance but finance-related papers (i.e., economics and accounting in these cases) as well. Many of these papers, however, can be related to finance research and any subjective exclusion of some articles may create another form of bias. Hence, we include all papers from all 21 journals.

Third, some elite economics journals (e.g., American Economic Review, and Journal of Political Economy) are not included in this study, although these journals occasionally publish influential research related to finance. They are not included because these economics journals publish mostly non-financial economics articles. The same arguments apply to top accounting journals such as Journal of Accounting Research.

Fourth, while all 21 journals are considered major finance journals, their quality is by no means identical. Aggregating all journals thus results in bias against elite journals. However, since a commonly used benchmark to account for journal quality, Social Science Citation Index (SSCI) is not available for some finance journals; adjusting journal quality becomes a difficult task. We employ two measures to mitigate this issue. First, using the methodology of Chan et al. (2002), we calculate the Journal of Finance equivalent pages (hereafter JF-pages) for each article. Since elite journals usually publish longer articles that are more thorough, using JF-pages as a weighting scheme helps to mitigate the differential qualities among journals. Second, we also rank institutions based upon top-5 finance journals only.

To measure research productivity, we made some adjustments to the raw data. First, similar to Chan et al. (2002), we first calculate the JF-pages for each article; then we adjust the JF-pages per author by dividing the article with the number of authors for multi-authored papers. Second, when an author has more than one affiliation, his/her contribution is divided equally among the stated institutions. For example, if an article has three co-authors (Professors A, B, and C) with the first author having two affiliations (X and Y) and the second and the third author each has one affiliation (W and Z), then institutions X and Y each receives 1/6 credit for the article and institutions W and Z each receives 1/3 credit for the article. Moreover, we proofread the manually collected data for possible errors. Additional verifications by studying university catalogs and websites are also conducted in case of doubt. We find that some authors or institutions use slightly different names over the 15-year sample period. For instance, we find several universities changed their names. We convert all the old names to the new names in such cases. An example is replacing Memphis State University by the University of Memphis because they represent the same institution with a name change occurred in the mid 1990s.

3 Descriptive statistics

For the period of 1990–2004, all 21 journals contain 11,501 weighted articles written by 8,554 authors from 1,126 universities and 1,035 non-academic institutions. Footnote 3 Among these 8,554 authors, 6,767 (79% of all authors) are affiliated with academic institutions writing a total of 9,633.7 weighted articles (83.8% of all articles). We plot a cumulative percentage of JF-pages authored by academic authors against the total number of universities in Fig. 1. The distribution is highly skewed. The top-5, top-10, top-25, and top-50 universities account for 10.3, 16.6, 29.5, and 43.7% of the total number of JF-pages, respectively. Therefore, 4.4% (50/1,126) of the universities account for more than 43% of the total production in finance literature. We also compute the Gini coefficient of finance publishing. The results are in the Appendix. Essentially, Gini coefficient measures the degree of concentration (inequality) in a distribution, with zero being no concentration (perfect equality) and one being total concentration (perfectly inequality). For all 1,126 institutions, the Gini coefficient is 0.7725, which indicates a high degree of concentration in finance research (i.e., a steeper Lorenz Curve). Among the North America, Asia-Pacific, and European regions, the Gini coefficients are 0.7587, 0.7306, and 0.6924, respectively. Therefore, concentration in finance research is observed across all regions of the world.

Fig. 1
figure 1

Cumulative percentage of JF-pages appeared in 21 finance journals for 1,126 universities (1990–2004). This figure plots a cumulative percentage of JF-pages authored by academics against the total number of universities. The distribution is skewed. The top-5, top-10, top-25, and top-50 universities account for 10.3%, 16.6%, 29.5%, and 43.7% of total number of JF-pages respectively

Full size image

Table 1 reports the summary statistics of research productivity by academic institutions and by academic authors. We report the JF-pages as well as the weighted number and the unweighted number of articles published. In Panel (A), the mean values of the JF-pages, weighted number and unweighted number of articles per academic institution are 162.84, 8.56, and 17.51, respectively. Since the median values of these variables are very small relative to their respective means, the distribution is highly skewed as shown in Fig. 1. The skewness and kurtosis statistics are positive and large for all research productivity measures. The maximum JF-pages per institution is an impressive 4,971.91, while the minimum is a miniscule 1.04.

Table 1 Summary statistics of the research productivity in a set of 21 finance journals from 1991 to 2004
Full size table

Panel (B) of Table 1 summarizes research output by authors affiliated with academic institutions. A small number of these authors, however, may have both academic and non-academic affiliations. A total of 6,767 academic authors contributed to writings in these 21 journals. An average author would produce 27.22 JF-pages, 1.43 weighted articles, or 2.83 unweighted articles during the period of 1990–2004. The median value of JF-pages is 13.59, the median weighted article is 0.83, and the median unweighted article is 1.0. Since all median values are smaller than the means, the distribution is again skewed although the skewness is smaller than that is reported in Panel (A) based upon institutions. Similar to Panel (A), both research productivity measures also show large skewness and kurtosis. The most productive author produces 562.12 JF-pages; 24 weighted articles; or 43 unweighted articles during this period. The least productive one comes up with 0.65 JF-pages. Footnote 4

In Panel (C) we report the frequency of publications for individual authors affiliated with academics. Among the 6,767 academic authors, 3,583 (53% of total) have published only one unweighted article in the 21 finance journals during the entire 15-year period. Cumulatively, 69% of all authors have published two articles or less during the 15-year period. Publishing ten or more articles in 15 years, therefore, places a researcher in the top five percentile of the productivity distribution. Less than two percent of the authors publish more than 14 unweighted articles, i.e., approximately one article (single or coauthored) per year.

Figure 2 shows the total JF-pages published each year over the 15-year sampling period. For all 21 journals, the total JF-pages increased from 8,050 in 1990 to 18,019 in 2004, representing a 123.8% increase, or an annual increasing rate of 5.5%. For the top-5 finance journals, Footnote 5 the total JF-pages also increased from 3,899 in 1990 to 7,458 in 2004, representing a 91.2% increase, or an annual growth rate of 4.4%. The growth in JF-pages over time reflects both the increases in the number of articles published and the average length of manuscripts. Footnote 6

Fig. 2
figure 2

Total number of JF-pages published in the Top-5 and all 21 journals. This diagram plots the total number of JF-pages published in the top-5 and all 21 finance journals from 1990 to 2004

Full size image

In addition to these aggregate data, we examine in Figs. 3 and 4 the annual research output from 1990 to 2004 partitioned by regions of the world, i.e., North America (including US and Canada), Europe, Asia-Pacific, and others. Figure 3 shows the share of research output by world regions using all 21 journals. North America produced 89.45% of the total JF-pages in 1990, and it steadily declined to 66.24% in 2004. On the other hand, the share of both Europe and Asia-Pacific regions gained ground during this 15-year period. Specifically, Europe’s share of JF-pages increased from 6.35% in 1990 to 20% in 2004, and Asia-Pacific from 2.79% to 12.05%. All other countries contributed 1.39% of the total JF-pages in 1990, and 1.64% in 2004.

Fig. 3
figure 3

Share of total JF-pages in various regions. This diagram plots the share of total JF-pages in 21 finance journals contributed by institutions in North America (US and Canada), Europe, Asia-Pacific, and others

Full size image
Fig. 4
figure 4

Share of total Top-5 JF-pages in various regions. This diagram plots the share of total JF-pages in the top-5 finance journals contributed by institutions in North America (US and Canada), Europe, Asia-Pacific, and others

Full size image

Figure 4 shows the share of research output by regions using only the top-5 journals. Although North America also lost share during the 15-year period from 96.38% in 1990 to 86.27% in 2004, the decline is not as steep as using all 21 journals. Europe gained share from 1.59% in 1990 to 6.46% in 2004, while Asia-Pacific from 0.49% to 5.62%. North American, mainly US, institutions, therefore, still have a quasi-monopoly position in the top-5 journals.

4 Ranking by Countries

Table 2 reports the ranking in aggregate research output by countries. We also report the number of institutions in each country contributed to the literature, the mean productivity of contributing institutions, and the standard deviation of JF-pages. We do not rank countries by their respective mean productivity per institution because such a measure could be misleading. Consider a hypothetical country that has 20 universities. Among these 20 universities, only one contributes five weighted articles to the finance literature, while the contribution from the other 19 institutions is nil, hence they are not ranked at all. If we use the mean productivity to rank countries, this hypothetical country would have been ranked high, but this is misleading because the five weighted articles is the mean productivity of a single institution, not the average of all 20 institutions.

Table 2 Summary statistics of JF-pages appeared in 21 finance journals by countries
Full size table

In total, 59 countries contributed to the production of finance literature in these 21 journals from 1990 to 2004. Column 3 reports the total JF-pages, and column 1 presents the ordinal rank of each country’s research output based upon JF-pages. The U.S. dominates the finance literature production with a lion’s share of 73% (133,667.6/183,360.3 JF-pages) of the total finance research published in these 21 journals, followed by U.K. (12,687.49 JF-pages or 6.9% of the total JF-pages), Canada (6,809.9 JF-pages or 3.7% of the total), Hong Kong (4,436.13 JF-pages, or 2.4% of the total), and Australia (3,720.93 JF-pages, or 2.0% of the total). The top-5 countries, therefore, account for 88% of the total JF-pages.

What might have contributed to the variations in finance research across countries? Although we use aggregate productivity to rank countries, the size of the population probably is not relevant. For example, the two most populous countries such as China and India produce only 155.32 and 104.46 JF-pages, respectively in these 21 journals. Indonesia, another country with large populations, produces a minuscule 13.11 JF-pages. On the contrary, Singapore, a country of only 3 million is ranked 10th. A country’s wealth, measured by her per capita GNP, on the other hand, might offer some explanations. After all, wealthy nations have more established institutions of higher education, which, ceteris paribus, should produce more finance research. The wealth of a nation, however, may not tell the whole story. Examining the statistics reported in Table 2 find that wealthy nations such as Japan are outranked by less wealthy nations such as Hong Kong, Singapore, Netherlands, Korea, and Taiwan, to name a few.

Another factor that may be related to finance literature production is language. Since finance literature is dominated by English-language literature, English-speaking countries naturally would have an edge in producing clearer English text. Hence, we conjecture that English-speaking countries, other factors the same, produce more finance literature.

Finally, seemingly unrelated, finance literature output may also be associated with the extent in which a country offers legal protection to her investors. When a country has a legal system that offers little protection to her investors, accounting standards are lax, the rule of laws are weak, and little incentive for the intellectual exchange in the arena of finance research would exist. Of course, it is also equally arguable that the lack of intellectual exchange in finance research leads to little regard for the investors’ protection laws. Therefore, our interest in not in the direction of causality; rather, we are interested in the relationship between finance literature output and the legal protection a country offer to her investors.

To examine this hypothesis, we include in the model per capital GNP and a binary variable which classifies countries on the basis of English-speaking. We also include several interesting variables in La Porta et al. (1998) including rule of laws, judicial system efficiency, index of accounting standard, and concentration of share ownership in the largest public companies. La Porta et al. find ownership concentration negatively related to investors’ protection. We construct our model of the finance literature production by countries as the following:

$$ \begin{aligned} Ln (JF\hbox{-}pages)_{i}=\alpha +\beta_{1} (English)+\beta_{2} Ln(per\ capita\ GNP)_{i}+\beta_{3} (Judicial)_{i}+\beta_{4}(Law)_{i}\\+\beta_{5} (ACCstd)_{i}+\beta_{6} (OWN)_{i}+\beta_{7} (Origin)_{i}+u_{i} \end{aligned} $$
(1)

The dependent variable in the model is the natural logarithm of the aggregate JF-pages produced by ith country (Ln (JF-pages)). We use the log transformation for the JF-pages because this variable is highly skewed. Certain exogenous variables in the model are based on La Porta et al. (1998) and are defined as:

  • Ln(per capital GNP) = natural logarithm of a country’s per capita GNP.

  • English = A binary variable such that English = 1 if a country is English-speaking; otherwise, English = 0.

  • Judicial = Efficiency of judicial system. Scale from zero to 10; zero being the least efficient.

  • Law = Rule of laws. Scale from zero to 10; zero being the least.

  • ACCstd = Index of the quality of accounting standards. Higher index value means higher quality.

  • OWN = Average percentage of common shares owned by the three largest shareholders in the 10 largest non-financial, privately owned domestic firms in a given country.

  • Origin ∈ (French, Germany, Scan) = A binary variable so that French = 1 if a country’s origin of commerce law is French civil law; Germany = 1 if Germany civil law; and Scan = 1 if Scandinavian law.

Our sample size for this analysis ranges from 39 to 51 depending on the variables used in the model. Although, La Porta et al., has relevant data for 49 countries and we have data for 59 countries, the sample size is reduced in our study after matching their data with our data. Since there are six countries in La Porta et al., that published no finance articles, the matched sample is left censored at 0. Therefore, the appropriate method to estimate our model is the TOBIT analysis.

Table 3 reports the results for this empirical model. Several interesting results emerge. First, model (a) through model (f) reports regression results when each exogenous variable enters the analysis individually. Chi-square statistics, reported in the parentheses, indicate that Judicial, Law, ACCstd, OWN, and GNP are all statistically significant at the one percent level and carry the expected signs. English is also significant albeit at lower level, suggesting English-speaking countries do have advantage over non-English speaking countries.

Table 3 TOBIT results of finance literature productivity by countries
Full size table

Second, in model (g) where all investors’ protection variables, equity ownership concentration, and GNP are included in the same equation, only two variables are significant—Law and English. The rule of laws (Law) carries the expected positive sign and is significant at the one percent level, while English is significant at the ten percent level. Since the only investors’ protection variable that is significant in the multivariate setting is Law, in model (h) we include Law, English, and GNP in the same equation. In this model, GNP is significant at the one percent level, while both English and Law are significant at the five percent level.

Finally, in model (i) we include Law, GNP, English and three binary variables measuring the origin of the commerce law. According to La Porta et al, English-Common-Law countries offer investors better protection. Law, GNP, and Scan are all significant at the one percent level. All three measures of law origin carry negative sign although only Scan is statistically significant. Surprisingly, English is reduced to insignificant although it still carries the expected positive sign. Footnote 7 Overall, our TOBIT models suggest that a nation’s wealth, language, and rule of laws (and legal protection) are associated with a country’s finance literature output. Footnote 8

5 Ranking by academic institutions

In this section, we first rank institutional research productivity using JF-pages published in 21 finance journals. We also include the weighted and unweighted number of articles for reference. Table 4 presents the 100 institutions with the highest JF-pages appeared in 21 finance journals. Footnote 9 New York University, the University of Pennsylvania, Harvard University, the University of Chicago, and UCLA take the top 5 spots. The top-100 universities are overwhelmingly represented by the U.S. institutions and the top-18 are exclusively U.S. institutions. Out of these 100 institutions, U.S. universities account for 78 places, followed by U.K. (6), Canada (5), Hong Kong (4), Netherlands (2), Singapore (2), France (2), and Australia (1). Footnote 10 Table 4 also reveals the skewness of the JF-pages distribution. For a school to move from 100th to 75th, it needs to advance 175.62 JF-pages (from 504.63 to 680.25 JF-pages). For a similar ranking advance from 50th to 25th, a school would need 408.15 JF-pages (from 866.63 to 1,274.78 JF-pages).

Table 4 The 100 academic institutions with most JF-pages appeared in 21 finance journals (1990–2004)
Full size table

Although we include 21 journals that are ranked high in the finance literature, variations in journal quality still exist and this journal quality bias might penalize certain institutions that stress quality, while favor others that take a broader view of journal quality. Ideally, one would make explicit adjustment in journal quality to minimize this bias. However, a commonly acceptable measurement of journal quality, Social Science Citation Index (SSCI) is not available for most of the finance journals in all years. Furthermore, some would argue that SSCI impact factor does not necessary measure quality per se. Hence, the use of impact factor introduces a different type of bias. To provide an alternative view of institutional ranking, therefore, we report rankings based upon the top-5 finance journals in Table 5. The qualities of these top-5 finance journals are the least controversial.

Table 5 The 100 institutions with most JF-pages appeared in top-5 finance journals
Full size table

Using the top-5 finance journals only, the top five institutions are New York University, the University of Pennsylvania, the University of Chicago, Harvard University, and the University of Michigan. Comparing with the results reported in Table 4 where all 21 journals are employed, the top-8 schools are the same with minor changes in relative ranking. Again, U.S. institutions overwhelmingly dominate the top-100 list in Table 5. Non-U.S. institutions only claim 19 places.

Why do some institutions have higher research productivity than others? In this section, we try to find factors that are associated with the cross-institutional research output variations. These factors include financial resource of the institution, faculty size, and research catalyst effect among the faculty members. It is expected that these factors are positively correlated with our research output measures. The research catalyst effect measures how well the faculty members within each institution work together among themselves. Other things being the same, we expect an institution with faculty members working together more likely to produce more research. The catalyst effect enhances research productivity through two factors: the incentive it provides and the avoidance of dilution through article weighting scheme. Obviously, the weighting scheme will split the credit among coauthors if they are not from the same institution.

Faculty size is expected to be positively correlated with JF-pages since the aggregate JF-pages, not the mean JF-pages per institution is the dependent variable. Finally, institutions with richer finance resources are able to better support their faculty research with, for example, appropriate database, reduced teaching load, and research grants; hence more research output.

The empirical model is specified as:

$$ JF-pages_{i}=\alpha +\beta_{1} (Size)_{i} +\beta_{2} (Peer)_{i}+\beta_{3} (Per\ Capita\ Budget)_{i} + u_{i} $$

where

  • JF-pages = research productivity measured by JF-pages;

  • Size = the number of finance faculty members in the ith institution Footnote 11;

  • Peer = the number of peer-coauthored articles from the ith institution. This is a proxy of research catalyst effect;

  • Per Capita Budget = per capita business school budget as of January 2005; Footnote 12 this is a proxy of institutional financial resources.

In Table 6, we report the cross-institutional variations of research productivity. Footnote 13 We have the complete data for only 341 US AACSB-accredited schools. Catalyst effect proxy, financial resources, and faculty size are all statistically significant at the one-percent level with the expected signs. Larger faculty size, more generous budget, and more peer-coauthored articles all contribute to research output, hence higher institutional ranking. The adjusted R2 value of 0.867 is quite high for a cross-sectional regression analysis. Footnote 14 Comparing the parameter magnitudes, some interesting results emerge. The parameter of the faculty size coefficient is 12.9862 meaning, other things being held constant, each additional faculty contributes to approximately 13 more JF-pages for the institution. This number is statistically significant, but less so economically. For example, for a school to move from 75 percentile to 50 percentile, 33.6 additional faculty is needed if all other factors are held constant. Footnote 15 On the other hand, the magnitude of the parameter for the catalyst effect is both statistically significant and economically large. One additional peer-coauthored article contributes to 40 additional JF-pages. Therefore, for a school to move from 75 percentile to 50 percentile, 10.9 ((887–450)/40 = 10.9) additional peer-coauthored articles are needed, other things held constant. Finally, the parameter size of 0.9688 for the per capital budget means a one thousand dollar increase in per capita budget results in one additional JF-pages. Obviously, it is not cheap to enhance finance research.

Table 6 Cross-institutional variations of research output
Full size table

6 Conclusions

We study the ranking of finance programs globally using a set of 21 finance journals from 1990 to 2004. A total of 6,767 academic authors from 1,126 academic institutions published at least one unweighted article in this set of journals. An average institution published only 162.84 JF-pages over the 15-year period. As the distribution is highly skewed, the median JF-pages is only 28.00. Similar skewed distributions can be found for the numbers of publications by author. More than three-quarter of the 6,767 academic authors published three or fewer articles during the entire 15 years period. Therefore, publishing five or more unweighted articles would put an individual in the top fifteen percentile of the research output distribution. Publishing 15 articles, or one per year, would rank this individual in the top two percentile.

When ranked by countries, the U.S. dominates the finance research arena with a share of 73%, followed by U.K., Canada, Hong Kong, and Australia. We study factors that are associated with the finance literature production and find that countries with English-language speaking, better rule of laws (or, investors’ protection), and higher national wealth produce more finance literature.

We also rank finance programs based upon a full sample of 21 journals and a subset of top-5 journals. In both cases, the U.S. institutions dominate the top-100 list. When all 21 journals are employed, 78 U.S. universities are ranked in the top-100. When only the top-5 journals are employed as the base of ranking, 81 U.S. universities are ranked in the top-100, followed by Canada, and U.K.

Finally, we also provide explanations to the cross-institutional variations of research productivity among a subset of schools. We find that faculty size, financial resource, and research catalyst effect explain the variations.