1
Economics Departmental Rankings in Korea: A Decade Later
(forthcoming, Journal of Asian Economics)
Jang C. Jin*
Department of Decision Sciences and Managerial Economics
Faculty of Business Administration
Chinese University of Hong Kong
Shatin, N.T.
Hong Kong
Phone: (852) 2609-7902
Fax: (852) 2603-5104
Email:
Economics Departmental Rankings in Korea: A Decade Later
I. Introduction
About a decade ago when a ranking of economics departments in Korea was first published (Jin, 1996), it surprised many Korean economists. The rankings were seen as a threat to the economics profession there. As a result, the rankings were seldom mentioned in public and academics refused to discuss them, even in private. However, as students were given a right to evaluate professors and had increasing opportunities to go abroad, university administrators were compelled to change their attitude toward professors’ performance, particularly their publication in internationally-recognized journals. Some schools even started providing cash incentives for international journal publications although this type of academic research, in addition to teaching and service, is assumed to be a duty in most developed countries.
Particularly, since the mid-1990s, economics professors in Korea have been motivated to publish in international journals more than ever; their publication in international journals has begun to be appreciated. In addition, many Korean scholars returned from overseas universities in recent years, and labor mobility even within the domestic market has become a fashion. However, there have been no studies evaluating the recent changes in research productivity of the economics profession in Korea, and little has been done for analysis of the causes and effects of increased productivity.1
This article documents and qualifies the recent changes in research productivity in Korea. Section II updates departmental rankings reflecting the past ten years. Section III investigates the sources of rising productivity in research. The impact of the improved productivity on school reputation is also estimated. Section IV compares domestic performance with other East Asian and U.S universities. Finally, implications and conclusions are presented in section V.
II. Economics Departmental Rankings
Since ranking research performance of economics departments is a sensitive issue, several strict rules consistent with previous research (Jin, 1996) are applied. First, departmental rankings are based upon page counts of articles published in international refereed journals over the period 1970–2003. The EconLit CD-ROM (December 2003) is used to track page counts as it includes all international journals (currently about 400 major journals) that have been appeared in the Journal of Economic Literature (JEL). Total pages are counted only for the articles clearly subjected to a refereeing process. Excluded are conference proceedings, monographs, books, book reviews, collective volume articles, PhD dissertations, and working papers. NBER working papers, IMF staff papers, and Carnegie-Rochester conference series on public policy are the only non-refereed publications included for the rankings as they are frequently cited in the economics literature. It would be desirable to include domestic Korean language journals, but this was not done since inclusion of the many school-published journals in Korean is a formidable task and beyond the scope of this paper.2
Second, the ranking methodology counts the ‘stock’ of past and current publications that have been published by current faculty members in each department (e.g., Scott and Mitias, 1996). Faculty names were taken from the websites of economics departments of the universities in the study (as of December 2003).3 Third, only economics departments are rated. Although there are research-active economists working in departments such as international trade, agricultural economics, finance, and so forth, they are not counted. Several research institutes are not included either, because their job obligations differ from academia. Fourth, economics departments are excluded from the ranking if only one faculty member is active in research. Fifth, if a journal article is coauthored by 'n' people, 1/n of the pages are credited to each coauthor. Finally, branch campuses were not independently counted.4
Figure 1 shows page counts of international refereed journals for top 40 universities in Korea over the sample period 1970-2003. There are more than 200 tertiary institutions in Korea, but only forty universities were included. The remaining 160 plus schools do not appear in the ranking, either because their faculty had zero publications during the sample period 1970-2003 or because they did not have a department of economics. Korea University and Seoul National University are ranked as the top two schools with 1837.8 and 1783.4 published pages respectively. The improvement of Hanyang University to number three from its previous rank number seven is impressive although its performance is about one half of the top two schools. After that, Sogang follows in 4th, while some may be disappointed by Yonsei’s 5th place finish. Relatively new and small schools also entered the top tier. For example, Sungkyunkwan, Kookmin, Chonnam, Dongguk, and Sookmyung are now ranked in top-10 schools. Other than that, weak research performance in lower ranks appears to be similar to each other, and the ranks will be changed dramatically if one star performance is included.5
Table 1 presents the research performance further for two sub-periods and compares the changes in the ranks. For the first sub-period 1970-1993, the order of top four schools (Seoul, Korea, Yonsei, and Sogang) are identical to the finding in Jin (1996) although actual number of pages are slightly different. For the lower ranked schools, more variations are found. The differences found here from the past ranking are mainly due to changes in current faculty members of the departments. More specifically, during the past ten years, a number of fine scholars were increasingly appreciated by the Korean job market and moved to better universities. For example, when a fine research scholar entered a job market, economics departments were buying his or her future flow of research as well as the stock of past and current research the scholar would bring to a department.
In addition, the ranks in recent ten years 1994-2003 are found to be similar to the overall ranking of the entire sample period 1970-2003 in the first column except a few changes. Thus, it may not be too much to say that recent performance dominates and it nearly determines the overall ranking. The last column in Table 1 then reports the changes in the rank over time, which evaluates how much each school was improved especially for the past ten years. Korea University is found to outperform all others, followed by Seoul National. Hanyang University has also improved significantly. The improvements in Chonnam, Hallym, Pusan, and Yeongnam are also impressive. However, many schools, including Seoul National and Yonsei Universities, did not follow a current trend in research performance in Korea.6
Table 2 presents the per capita productivity in which a wide range in the size of a department is controlled for the number of faculty members in that department. The first column shows average productivity that one faculty member has published in international refereed journals over the entire sample period 1970-2003. Korea University is again found to be at the top, and Seoul National follows. In other words, a faculty member in Korea University is, on average, more productive than in Seoul National.
To check on the robustness of the results, the sample was split into two sub-periods 1970-1993 and 1994-2003, based upon the degree of openness in Korean universities. The university openness is defined here as how many publications the university “buys” from overseas. Since the mid-1990s, many universities have “purchased” overseas international publications by hiring well-published academics from foreign countries. The invited overseas scholars might also have a spillover effect on other faculty members. This is analogous to the role of foreign direct investment (FDI) in the process of economic development. A domestic economy will not only be improved directly by foreign capital inflows, but domestic technology will also be improved indirectly through the spillover effect. Thus, the sample period beginning from 1994 may be thought of as representing an ‘FDI-type’ research, whereas the sample period prior to 1994 may be thought of as representing an ‘import-substitution’ research.
The second and third columns in Table 2 present annual averages of the per capita productivity for two sub-periods. For the first sub-period, Seoul National University is ranked first and Korea University second; however, the order is reversed for the most recent ten-year period. For the rest of universities in the sample, the latter sub-period nearly determines the overall ranking as in Table 1. In addition, an annual average productivity across universities has been increased sharply for the years characterizing an FDI-type research. In particular, the average productivity appears to be trivial prior to 1994, but it jumps 200-1000 percents and even 3000 percents for the period 1994-2003 (see the last column in Table 2). In contrast, low productivity during the first sub-period might be related to the lack of incentives and motivations in the 1970s and 1980s.
III. Regression Results
Figure 2 shows total research productivity of all economics departments over the period 1970-2002. The most recent year 2003 was not included here because the complete publication of the year did not appear in the December 2003 issue of EconLit CD-ROM because of publication lag. It was found that the economics profession was nearly inactive in academic research until the mid-1980s. In the late 1980s and the early 1990s, total publication in Korea slightly increased but hovered around 250 pages per year. Since the mid-1990s, the publication has been increased substantially and reached a peak in 1997 and another peak is also found in 2001. The sharp rise in the late 1990s and early 2000s is very impressive. Greater disagreement, however, arises over the probable causes for the recent increase in productivity. Some economists argue that a purchase of overseas publication is the primary cause for the increased productivity in recent years, whereas most others claim it as domestic; domestic faculty members are publishing more than before. One may also wonder if the increased research performance influences school reputation.
Table 3 shows correlation coefficients (r) for several relevant variables. Research productivity that is measured in total page counts appears to be highly correlated with the stock of overseas publications purchased (r=0.79), which were originally affiliated to a foreign institution at the time of publication but are counted here for his or her current affiliation in Korea. The research productivity of current faculty members is also highly correlated with the students’ college entrance exam (or college aptitude test) scores (r=0.69), which are reported as a minimum to get in the economics department of each university. The exam scores are thus used as a proxy for student quality.
Recently, school fame was ranked for thirty top universities in Korea based upon a survey data conducted by Joong-ang Daily Newspaper (December 17, 2003). Since the way of ranking the fame (e.g. 1=best, 30=worst) is to have an awkward negative relationship with other performance-based variables in the model, an inverse of the rank is used here to assign the highest score to the best school and the lowest to the worst. Among top-thirty universities surveyed, six schools were not found in our research-based ranking. Only twenty four schools are used for analysis. The lower values in Table 4 represent the correlation coefficients between variables when 24 schools are used, while the upper values represent correlation when all 40 schools are used. The results provide evidence that the school fame has a strong positive correlation not only with research performance of current faculty members, but with college entrance exam scores. In other words, the school fame is highly related to faculty performance as well as student quality. However, it is surprising to find that a tradition in higher education, which is measured as the number of years of school history, is no longer strongly related to school fame. The school history is also weakly related to student quality as well as faculty’s research performance.
Table 4 further investigates the sources of increased research productivity in recent years. For regression analysis, total pages published are used as a dependent variable. All variables except dummies are taken as logarithms as a heteroscedasticity problem may typically arise in this type of cross-school data, i.e. residuals were observed to be relatively large in higher-rank universities and became smaller in lower ranks. The log-linear model mitigates the measurement scale of the raw data, and thus changes in residuals appear to be smaller. In this case, the heteroscedasticity problem may not be serious. Table 4 provides evidence that changes in overseas publication has a positive impact on domestic productivity, and the impact appears to be significant at the 5% significance level. The size of the impact also appears to be similar in all different specifications of the model. The results are in general consistent with a recent trend that many fine scholars overseas returned to Korea although their performance tended to be less active after being returned. It is also found that the demand for overseas publication depends on the degree of openness of the universities. More open universities in higher ranks are generally found to have a high proportion of overseas publication (30-53% of total pages), while the proportions are relatively low (0-20%) for lower ranks.
However, the effect of a public school dummy variable that assigns 1’s if public and 0’s if private appears insignificant at the conventional significance levels. The public school includes several national universities, as well as city universities located in Seoul and Incheon. Unlike state universities in the U.S., provincial universities in Korea are fairly new and similar to vocational schools in the U.S. None of them are included in the ranking. The result in Table 4 shows that the public school dummy does not have a significant impact on research performance, and the insignificant effect thus suggests that national universities, on average, perform as much as private universities. This finding is, however, at odds with a general belief in which a decline in research activity until the mid-1990s was largely due to a rigidity of national universities in Korea.
In addition, smaller teaching loads were found to help improve research performance. The data for teaching load was recently surveyed by phone calls to all forty economics departments in the sample. Most schools were found to teach 2-4 classes per semester (equivalently, 6-12 hours per week). Although real teaching load may vary depending on undergraduate or graduate courses as well as the number of preparations, 3 classes per semester (9 hours per week) appeared to be an average among these top forty schools in Korea. To find the productivity effects of teaching load, a dummy variable of one was assigned if the current teaching load was less than nine hours per week, and zero otherwise. The results are in general consistent with the general belief that people may produce more if their teaching load is reduced; the impacts are statistically significant at the 10% significance level.7
Table 5 shows another estimation results in which school fame was regressed on college entrance exam, school tradition, and research performance. The variable for overseas publications was not included in this model due to a potential statistical problem of collinearity with total research performance (r = 0.79 in Table 3). The dependent variable in this case was an inverse of the rank of school fame surveyed, and only twenty four schools were estimated due to availability of the data for school fame. Although the degrees of freedom problem may arise, the effect of college entrance exam that is used as a proxy for student quality appears to be positive and statistically significant at the 5% significance level. Surprisingly, however, the number of years of school history has no significant effect on school fame; the effects are found to be positive but insignificant even at the 10% significance level. The fame, however, significantly depends upon research performance of current faculty members. The effect of research productivity on school fame appears to be positive and statistically significant. Therefore, research matters. Notice, however, that when research productivity is added in equations (3) and (4), the size of the effects of college entrance exam drops, but statistical significance remains almost intact. The size of the effects of research productivity also remains intact even in different model specifications (3) and (4). Thus, multicollinearity problems may not be serious in this model specification.
IV. International Comparison
Although the data show that most economists in top-tier Korean universities are now active in research, how do their schools compare with other East Asian and U.S. universities? Jin and Yao (1999) directly compared East Asian universities using top-36 journals over the period 1990-1996, and they found that Seoul National University at that time ranked number eight in East Asia though its performance was about 12% of the top school in East Asia.8 If the performance were further compared with U.S. universities, the two most productive schools in Korea would rank perhaps around 100th to 150th at that time.9