A guide to graduate study in economics: ranking economics departments by fields of expertise.

1. Introduction

Each year, thousands of undergraduates apply for admission to graduate schools in economics intending to obtain a Ph.D. Many of these students have little idea on how to choose a graduate program, and many go to an undergraduate adviser looking for advice. Prospective graduate students and their advisers have little published research to help them in the process of choosing what schools best match the undergraduate's skills and interests.

This study highlights many of the characteristics of departments that offer doctoral degrees in economics and provides information on both overall productivity and productivity by subject field. This research is significant for those looking to obtain a Ph.D. in economics because the choice of where to attend graduate school has been shown to be important in both academic and nonacademic job markets. Research into the careers of Ph.D. economists (Barbezat 1992; McMillen and Singell 1994; Stock and Alston 2000; Siegfried and Stock 2004) consistently indicates that graduates from top-rated schools fare better in academic and nonacademic job markets than their peers from lower-ranked programs.

Based on the finding that the quality of the school influences outcomes in the job market, the best advice for those applying to graduate school in economics may simply be to apply to the best schools to which you will likely be admitted. Yet this advice is of little value for those who are unlikely to be admitted into a top program yet have a strong interest in one of the many subject fields of economics and a strong desire to pursue a particular field. This group of students is left getting advice from an undergraduate adviser who cannot be expected to know the strengths of economics departments across the country or to search the Web pages of all the programs that offer a Ph.D. looking for clues as to what school is the best match.

In this article, we provide information to undergraduate students and their advisers on the research strengths of 129 economics departments that offer Ph.D. degrees in the United States and to identify schools that are ranked highly in the many different subject fields of economics. This article should also provide guidance to departments hiring new Ph.D. candidates within a specific field and to job candidates looking for information on potential academic employers.

This article differs from the many papers ranking the quality of economics departments by identifying the relative strength of all Ph.D. programs and by specifically providing information on all the major subject fields in economics. Although Tschirhart (1989) ranks departments in fields of expertise, only a limited set of fields is identified, and departments are ranked using data that are now over 20 years old. U.S. News and Worm Report (1) also provides a ranking of economics departments by field. Their ranking is based on survey responses of department chairs who were asked to rank all departments on a five-point scale. Department rankings by field can also be found on the EconPhd.net website (http://www.econphd.net). This site ranks departments by field, using publications in 63 highly ranked economics journals during the 1993-2003 period. The data we used as the basis for this article are more comprehensive and cover a larger time frame. We used all journals in which economists at the Ph.D.-granting institutions in the United States had published during a 20-year period. Our data set consists of publications in 254 journals over the 20-year period 1985-2004. This analysis provides by far the most detailed, complete ranking of departments by field in the literature.

In addition to simply identifying the top 20 schools in each field, other information, not found elsewhere, is provided on the relative importance of the field at the school and how the scholarly output is distributed across the department's faculty. To measure the concentration of faculty in a field, we calculate a Herfindahl-Hirschman Index (HHI). The HHI is particularly important for an undergraduate to consider. Planning to obtain a Ph.D. from a school in hopes of studying with a single person is a risky undertaking not only because the faculty member may move but also because any single faculty member can mentor only a limited number of students.

We recognize that ranking departments is fraught with danger. Thursby (2000) has pointed out that using single measures of department productivity suggests differences between many departments that are meaningless, a finding we reiterate when solely aggregate measures of performance are used. However, by providing detailed information on departments by field and by identifying the publication patterns of the faculty within the field, we are able to highlight some differences that aggregate measures gloss over.

2. Methods

Similar to Tschirhart (1989), the data-gathering stage consists of four basic steps: (i) identifying all Ph.D.-granting institutions in economics as of the 2004 spring semester, (2) (ii) identifying all tenure-track or tenured faculty as of the 2004 spring semester, (iii) acquiring a list of faculty publications, and (iv) determining the quality of each publication.

To identify the universities offering doctoral degrees in economics, we used the website maintained by the University of Albany. (3) This site contained a list of all economics departments with Ph.D. programs at American and Canadian universities and was verified with Peterson's Guide to Graduate Schools. (4) Based on this, we identified 129 programs located in the United States that offered doctoral degrees in economics as of the spring of 2004.

The second step, identifying all tenure-track or tenured faculty for each university, was accomplished by accessing economics department Web sites. A slight shortcoming of this approach is that faculty lists are highly dependent on whether a department maintains and updates their faculty lists. Removing faculty members without any publications resulted in over 2600 faculty names. In the few cases where faculty appeared on multiple department websites, we included the faculty member in the department where he or she had a permanent and current affiliation. We recognize that there are some faculty who are members of a department other than economics (e.g., the Department of Managerial Economics and Decision Sciences at Northwestern University) yet contribute to the education of graduate students and are productive in the field of economics. Determining who these faculty are and the extent to which they are involved in the economics department made it impractical to include them in the analysis.

The third step focused on acquiring journal publications for each faculty member listed in the Journal of Economic Literature database Econlit. The database was queried for the publications of tenure-track faculty identified by the 129 departments. Faculty were dropped from the analysis if Econlit indicated that they had no published articles. This study focused on articles published between 1985 and 2004. Over this time period, Econlit cataloged over 38,000 publications of faculty who were employed in Ph.D. economics programs as of the spring of 2004. (5) Further, Econlit provided four essential pieces of information that would be needed for analysis: (i) article source, (ii) page numbers, (iii) number of authors, and (iv) Journal of Economic Literature subject codes. The article source would be needed in order to assess the quality of the article. The credit each author received for a publication was weighted by the number of authors and page length. The greater the number of coauthors, the less credit assigned to each coauthor, and the greater the length of the article, the greater the credit assigned to each coauthor. (6) The subject codes would be needed to sort articles by a field of expertise.

The final step was assigning a quality index, [Q.sub.j], to each journal. We used both the impact factors published in the 2004 Social Science Citation Index (SSCI scores) and rankings based on "citations per character in 1990" for articles published between 1985 and 1989 (JEL scores) proposed by Laband and Piette (1994). (7) Many publications contained at least one or both an SSCI and a JEL score. There were 107 journals containing both an SSCI and a JEL score. There were an additional 131 with only an SSCI score and an additional 16 with only a JEL score. Thus, the total number of journals indexed in the SSCI that we used in our analysis was 238, and the total number of journals indexed in the JEL that we used in our analysis was 123. Publications that had neither an SSCI nor a JEL score were dropped from the analysis. It should be noted that although the SSCI indexes 172 journals in the economics discipline, we use all publications identified by Econlit and indexed in the SSCI, even if outside the economics discipline, in calculating productivity.

Following Tschirhart (1989), articles were adjusted by number of authors and page length. The first step consisted of dividing the number of pages of article i, [pages.sub.i], by the number of authors (n), thus ensuring that each author received 1/n credit times the number of pages. The second step consisted of taking the value from the first step ([pages.sub.i] divided by n) and dividing it by the average length of all articles from the same journal j ([[bar.p].sub.j]). The weighting that each coauthor of article i in publication j, [W.sub.ij], receives is given by

[W.sub.ij] = [pages.sub.i]/[n.sub.i]/[[bar.p].sub.j]