- Email: [email protected]
One list to fit them all: What do we learn from journal ranking?
Finance Research Letters xxx (xxxx) xxx–xxx
Contents lists available at ScienceDirect
Finance Research Letters journal homepage: www.elsevier.com/locate/frl
One list to fit them all: What do we learn from journal ranking? Konstantinos Eleftherioua, , Michael Polemisa,b ⁎
a b
Department of Economics, University of Piraeus, 80 Karaoli & Dimitriou Street, 18534, Piraeus, Greece Hellenic Competition Commission, Athens, Greece
ARTICLE INFO
ABSTRACT
JEL classifications: C23 A2 A11 I22
In this paper, we apply the Phillips and Sul (2007; 2009) methodology to investigate the convergence pattern of finance journals across thirteen established academic journal lists. The results reveal that the majority of sample journals do converge across the academic lists. The estimated transition paths confirm the empirical analysis, revealing a “focal” point for the research institutions to minimize the discrepancies appeared by the journal lists proliferation.
Keywords: Convergence Finance journals Club clustering Transition paths
1. Introduction It is widely acknowledged by the scientific community that publishing a research paper in a highly ranked economic journal is of great importance to researchers in terms of hiring or being tenured and promoted by an academic department or research institute (Gibson et al., 2014; Kosteas, 2015). This happens since journal rankings constitute a focal point (“signal”) for attracting young economists and retaining older ones in prestigious institutions (Kalaitzidakis et al., 2011). Moreover, journal rankings contribute to the allocation of research funding across the economic departments (Oswald, 2007). Journal rankings are also appeared in other disciplines such as finance, business and econometrics (Chung and Cox 1990; Baltagi, 2007). As a result, there is a fast growing number of academic journal lists consisting of different types (i.e., citation based, expert based, hybrid) and rating categories (i.e., continuous or discrete score) as reported by Vogel et al. (2017). A key question then arises by academicians in the sense of selecting the most efficient journal list to appropriate assess the research output. We attempt to exemplify this research question by investigating the existence of convergence patterns among the sample lists included in our empirical analysis and identify their transitions paths over the examined period (1999–2018). As a consequence, our study departs from the existing literature (see among others Halkos and Tzeremes, 2011; Zheng and Kaiser, 2011; Gibson et al., 2014; Kosteas, 2015; Malesios, 2015) since we attempt to uncover common trends on eighteen field finance journals appeared in thirteen sample journal lists. In this way, we are able to draw some policy implications concerning the dissemination of research conducted by finance departments worldwide. The main contribution of this study is that we examine for the first time, to the best of our knowledge, the convergence patterns among thirteen major scientific journal lists. Specifically, we study the convergence of journal rankings, which are related to different
⁎
Corresponding author. E-mail addresses: [email protected] (K. Eleftheriou), [email protected] (M. Polemis).
https://doi.org/10.1016/j.frl.2019.08.026 Received 15 July 2019; Accepted 28 August 2019 1544-6123/ © 2019 Elsevier Inc. All rights reserved.
Please cite this article as: Konstantinos Eleftheriou and Michael Polemis, Finance Research Letters, https://doi.org/10.1016/j.frl.2019.08.026
Finance Research Letters xxx (xxxx) xxx–xxx
K. Eleftheriou and M. Polemis
Table 1 Summary statistics. Variables
Observations
Mean
Standard deviation
Min
Max
Journal of Risk and Uncertainty Journal of Finance Journal of Financial and Quantitative Analysis Financial Management Journal of Accounting, Auditing and Finance Journal of Financial Research Journal of Futures Markets Journal of Financial Economics Journal of Business Finance and Accounting Journal of Banking and Finance Review of Financial Studies Journal of Empirical Finance Journal of Corporate Finance Finance and Stochastics Mathematical Finance Journal of Financial Intermediation Journal of Derivatives European Financial Management
260 260 260 260 260 260 260 260 260 260 260 260 260 260 260 260 260 260
2,9 4,0 3,8 3,0 2,3 2,0 2,5 4,0 3,2 3,4 4,0 3,1 3,3 3,1 3,1 3,4 2,1 2,1
0,9 0,0 0,4 0,8 0,8 0,9 0,9 0,0 0,5 0,5 0,0 0,7 0,7 0,7 0,7 0,5 0,8 0,8
1 4 3 1 1 1 1 4 2 3 4 1 2 2 2 2 1 1
4 4 4 4 3 3 4 4 4 4 4 4 4 4 4 4 4 3
top field journals from the finance discipline by employing the Phillips and Sul (2007, 2009) methodology.1 The rest of the paper is structured as follows: Section 2 presents, in detail, the methodological framework of our analysis. Section 3 describes the dataset, while Section 4 presents and discusses our main findings. Finally, Section 5 concludes the paper. 2. Econometric methodology The methodological framework used in the present study is the one proposed by Phillips and Sul (2007, 2009). The Phillips and Sul (PS) methodology as applied in our case can be outlined as follows: If Xit is the solely factor of a panel data set (X denotes the rating for a given journal, i denotes the journal list and t the time), then
Xit =
(1)
it µt ,
where δit measures the deviation of journal list's i ranking from the common trend μt and can be represented as it
=
i
+
i it L (t )
(2)
1t a,
where δi is fixed, ξit is weakly dependent over t with ξit ∼ iid(0, 1) and L(t) is a slowly varying function with L(t) → ∞ when t → ∞. The null hypothesis of convergence for all i (H0) versus the alternative of non-convergence for some i (HA) can be expressed as
H0 :
i
=
and a
0; HA:
i
(3)
or a < 0
The null hypothesis in (3) can be tested through the following regression:
log
H1 Ht
2 log L (t ) = c^ + b^ log t + u^t ,
(4)
for t = [rT], [rT] + 1, ..., T with some r > 0.2 In (4), Ht = (1/ N )
N i=1
(hit
1)2 , hit =
Xit N 1
N i = 1 Xit
=
, N 1 iN= 1 it it
L(t) = log (t + 1) and
b^ = 2a^ where a^ is the least-squares estimate of a in H0 (null hypothesis). Based on the above analysis, PS argue that the hypothesis of convergence can be tested through a one-sided t-test. Specifically, the alternative hypothesis of non-convergence [see (3)] cannot be rejected at the 5% level if t b^ < 1.65. Given the absence of sample convergence, we proceed in identifying convergence clubs by implementing the following clustering procedure proposed by Phillips and Sul (2007): (i) we order the N journal lists for each journal title putting the list having the highest last period ranking first, the one with the next highest ranking in the last period second and so forth; (ii) we form all the possible core clubs by selecting the first k highest-ordered lists (for each journal) [see step (i)], with 2 ≤ k ≤ N and we calculate the convergence statistic tk for the logt regressions [eq. (4)]. Then, the size k* of the club is determined by the maximum tk, given tk > −1.65; (iii) from the remaining N − k* journal lists, we add one list at a time to the previous defined core clubs if the t-statistic is greater than zero; (iv) for the remaining lists (if any), we repeat steps (i)-(iii) until the formation of new clubs is not possible. Finally, we apply Phillips and Sul (2009) procedure to determine the existence of further convergence clubs and calculate their transition paths. 1 2
For an overview of the advantages of this methodology over other convergence approaches, see Apergis et al. (2013). Following Phillips and Sul (2007, 2009), r is set equal to 0.3. 2
Finance Research Letters xxx (xxxx) xxx–xxx
K. Eleftheriou and M. Polemis
Table 2 Club convergence results. Journal
Panel A: Phillips and Sul (2007) log t t-stat
Journal of Risk and Uncertainty Full sample −0.3316 (0.3189) Journal of Financial and Quantitative Analysis Full sample −1.0175 (0.1340) Club 1 [1,2,3,4,5,6,8,10,11,12,13] −0.7894 (0.0143) Club 2 [7,9] – Financial Management Full sample −0.6539 (0.2856) Club 1 [7,8,10,13] 0.6107 (0.3514) Club 2 [1,2,3,4,5,6,9,11,12] – Journal of Accounting, Auditing and Finance Full sample −0.8200 (0.0667) Club 1 [1,3,4,6,11,12,13] −0.7894 (0.0143) Club 2 [2,5,7,8] −0.7894 (0.0143) Club 3 [9,10] – Journal of Financial Research Full sample −0.8724 (0.0531) Club 1 [1,3,5,11,13] −0.7894 (0.0143) Club 2 [2,4,6,7,8,9,10,12] −0.2416 (0.2458) Journal of Futures Markets Full sample −0.9156 (0.0366) Club 1 [1,3,4,6,11,12,13] −0.7894 (0.0143) Club 2 [2,5,7,8,9,10] 0.3291 (0.2432) Journal of Business Finance and Accounting Full sample −0.8750 (0.1545) Club 1 [2,8,10,11] −0.7894 (0.0143) Club 2 [1,3,4,5,6,7,9,12,13] −0.7973 (0.2753) Journal of Banking and Finance Full sample −0.7894 (0.0143) Club 1 [1,3,4,5,8,10,11] −0.7894 (0.0143) Club 2 [2,6,7,9,12,13] −0.7894 (0.0143) Journal of Empirical Finance Full sample −1.1008 (0.2206) Club 1 [5,6,8,10] 0.3639 (0.4714) Club 2 [1,3,4,7,11,12,13] 0.0957 (0.9364) Club 3 [2,9] – Journal of Corporate Finance Full sample −1.2621 (0.0715) Club 1 [1,2,3,4,5,8,10,11] −0.0523 (0.2638) Club 2 [6,7,9,12,13] −1.5769 (0.1254) Finance and Stochastics Full sample −0.7696 (0.1377) Club 1 [4,5,8,10] 0.3639 (0.4714) Club 2 [1,3,6,9,11,12,13] −0.7894 (0.0143) Club 3 [2,7] – Mathematical Finance Full sample −1.2056 (0.1176) Club 1 [1,4,8,10,11] −0.7894 (0.0143) Club 2 [2,3,5,6,9,12,13] −1.0991 (0.2153) Non converging Group 3 [7] Journal of Financial Intermediation Full sample −0.9516 (0.3699) Club 1 [1,3,4,8,10,13] −0.1331 (0.2506) Club 2 [2,5,6,7,9,11,12] – Journal of Derivatives Full sample −1.3496 (0.0783) Club 1 [3,8,13] 2.1193 (0.8668) Club 2 [1,2,4,5,6,7,9,10,11,12] −1.4009 (0.2329) European Financial Management Full sample −1.0121 (0.1323) Club 1 [1,3,4,5,11,13] −0.7894 (0.0143) Club 2 [2,6,7,8,12] −0.7894 (0.0143) Club 3 [9,10] –
New club
Panel B: Phillips and Sul (2009) Final club log t
t-stat
1+2
Club 1
−0.7894 (0.0143)
−55.2828⁎⁎
2+3
Club 2
−0.8262 (0.0813)
−10.1568⁎⁎
1+2
Club 1
−0.0519 (0.4707)
−0.1102
3
Club 2
–
–
1+2
Club 1
−0.4134 (0.1601)
−2.5825⁎⁎
2+3
Club 2
−1.0873 (0.1210)
−8.9826⁎⁎
1+2
Club 1
−0.9609 (0.1177)
−8.1658⁎⁎
2+3
Club 2
−1.5108 (0.2014)
−7.5017⁎⁎
1+2
Club 1
−0.7894 (0.0143)
−55.2828⁎⁎
2+3
Club 2
−1.0848 (0.1804)
−6.0129⁎⁎
−1.0396 −7.5938⁎⁎ −55.2828⁎⁎ – −2.2898⁎⁎ 1.7377 – −12.2937⁎⁎ −55.2828⁎⁎ −55.2828⁎⁎ – −16.4421⁎⁎ −55.2828⁎⁎ −0.9830 −24.9857⁎⁎ −55.2828⁎⁎ 1.3535 −5.6621⁎⁎ −55.2828⁎⁎ −2.8963⁎⁎ −55.2828⁎⁎ −55.2828⁎⁎ −55.2828⁎⁎ −4.9899⁎⁎ 0.7720 0.1022 – −17.6582⁎⁎ −0.1983 −12.5797⁎⁎ −5.5877⁎⁎ 0.7720 −55.2828⁎⁎ – −10.2477⁎⁎ −55.2828⁎⁎ −5.1059⁎⁎ −2.5724⁎⁎ −0.5311 – −17.2411⁎⁎ 2.4449 −6.0162⁎⁎ −7.6516⁎⁎ −55.2828⁎⁎ −55.2828⁎⁎ –
Notes: The numbers in parentheses denote the standard errors. The term log t denotes the convergence coefficient, while t-stat is the convergence test statistic. The latter is distributed as a simple one-sided t-test with a critical value of −1.65. ⁎⁎ denotes rejection of the null hypothesis (convergence) at 5% level of statistical significance. Convergence is implied when t-stat is equal to ‘-’.
3
Finance Research Letters xxx (xxxx) xxx–xxx
K. Eleftheriou and M. Polemis
Fig. 1. Transition paths.
3. Data and statistics We use a balanced panel data set of thirteen academic journal lists over a twenty year period yielding 260 observations. The variables used in the analysis refer to eighteen finance journal rankings appeared in all of the academic sample lists (see Appendix for details).3 To account for different rating categories among the journal lists, we appropriately convert the rating scale of each list to a four-level scale (i.e., from 1 for the lowest quality ranking to 4 for the highest) applicable to each sample journal. Moreover, since journal lists are not updated each year, we use the last available version of each list to complete missing observations. Table 1 presents the main descriptive statistics for the sample variables. We observe that three finance journals possess on average the highest journal ranking among the sample lists (i.e., Journal of Finance, Journal of Financial Economics, and Review of Financial Studies). This can be explained by the fact that these journals are ranked with the highest journal score awarded by the thirteen academic journal lists (Min = 4).4 On the contrary, three journals such as the Journal of Financial Research, followed by the Journal of Derivatives and the European Financial Management have the lowest mean (i.e., approximately equal to 2). 4. Results and discussion Table 2 reports the empirical findings of the club clustering algorithm based on Phillips and Sul (2007) methodology. The results indicate a rejection of the null hypothesis of full panel convergence at the 5% level of significance in all but one (Journal of Risk and Uncertainty) of the fifteen journal cases since t-stat < −1.65 (see Panel A). Therefore the investigation of potential club clusters among these lists is required. As it is evident from Table 2, in the majority of the cases, the club clustering algorithm identifies one convergence club (see Fig. 2) among the lists included in the analysis since the null hypothesis of convergence cannot be rejected. In three cases there are two convergence clubs reported (Financial Management, Finance and Stochastics, and Journal of Financial Intermediation). Finally, the Journal of Empirical Finance is the only journal with three convergence clubs. As a final step, we apply the club merging analysis along the lines of Phillips and Sul (2009). Based on our findings (see Panel B), we argue that the merging of clubs is not supported since the hypothesis of convergence is rejected in all but one (Journal of Empirical Finance) of the reported cases. Therefore, the initially formed clubs as described above are the appropriate ones. The transition paths of the estimated convergence clubs per journal (see Fig. 1) corroborate the above results. To sum up, the majority of the lists converge for the following five journals: Financial Management, Journal of Financial Research, Journal of Empirical Finance, Journal of Corporate Finance and Journal of Financial Intermediation. Only few lists converge for five journals (Journal of Financial and Quantitative Analysis, Finance and Stochastics, Journal of Future Markets, Journal of Derivatives and European Financial Management). On the other hand, there are no converging lists for the following three sample journals: 3 4
For our estimations, we used the STATA codes provided by Du (2017). Therefore, these three journals are excluded from the econometric analysis. 4
Finance Research Letters xxx (xxxx) xxx–xxx
K. Eleftheriou and M. Polemis
Fig. 2. Distribution of clubs. Notes: Each graph illustrates the corresponding journal list. The horizontal axis depicts the clubs for each journal, while the vertical axis the number of the club.
Journal of Business, Finance and Accounting, Journal of Banking and Finance and Mathematical Finance. 5. Conclusion In this paper, we implement the Phillips and Sul (2007, 2009) methodology to examine the convergence pattern of eighteen major finance journals across thirteen established academic journal lists. We conclude that with the exception of the three elite finance journals (i.e., Journal of Finance, Journal of Financial Economics and Review of Financial Studies) and the Journal of Risk and Uncertainty, journal lists’ rankings do not converge for the rest of the sample journals. However, journal lists tend to cluster into groups, indicating that at least some of them exhibit some uniformity in ranking finance journals. Given that employment in academia is heavily based on the quality of publications as measured by journal lists, our findings have important implications for the mobility of finance scholars across academic institutions in different countries. Declaration of Competing Interest None. Acknowledgments We would like to thank Patroklos Patsoulis for excellent research assistance in collecting the data and Thanasis Stengos for useful comments and suggestions. Appendix A. Finance journals [abbreviation] Journal of Risk and Uncertainty [JRU] Journal of Finance [JF] 5
Finance Research Letters xxx (xxxx) xxx–xxx
K. Eleftheriou and M. Polemis
Journal of Financial and Quantitative Analysis [JFQA] Financial Management [FM] Journal of Accounting, Auditing and Finance [JAAF] Journal of Financial Research [JFR] Journal of Futures Markets [JFM] Journal of Financial Economics [JFE] Journal of Business Finance and Accounting [JBFA] Journal of Banking and Finance [JBF] Review of Financial Studies [RFS] Journal of Empirical Finance [JEF] Journal of Corporate Finance [JCF] Finance and Stochastics [FS] Mathematical Finance [MF] Journal of Financial Intermediation [JFI] Journal of Derivatives [JD] European Financial Management [EFM] B. Journal lists used in the analysis [list #] Association of Business Schools Academic Journal Quality Guide (CABS) [1] Center National de la Recherche Scientifique (CNRS) [2] Australian Business Deans Council Journal Rankings List (ABDC) [3] ESSEC Business School Paris (ESSEC) [4] Foundation National pour l'Enseignement de la Gestion des Entreprises (FNEG) [5] Erasmus Research Institute of Management Journals Listing (EJL) [6] Association of Professors of Business in German Speaking Countries (VHB) [7] Danish Ministry Journal list (DEN) [8] SCImago (SJR) [9] High Council for Evaluation of Research and Higher Education (HCERES) [10] University of Queensland Adjusted ERA Rankings List (UQ) [11] European Journal of Information Systems 2007 Mingers & Harzing (EJIS) [12] WU Wien Journal Rating May 2008 (WIE) [13] References Apergis, N., Christou, C., Hassapis, C., 2013. Convergence in public expenditures across EU countries: evidence from club convergence. Econ. Financ. Res. 1 (1), 45–59. Baltagi, B.H., 2007. Worldwide econometrics rankings: 1989–2005. Econ. Theory 23 (05), 952–1012. Chung, K., Cox, R., 1990. Patterns of productivity in the finance literature: a study of the bibliometric distributions. J. Financ. 45 (1), 301–309. Du, K., 2017. Econometric convergence test and club clustering using Stata. Stata J. 17 (4), 882–900. Gibson, J., Anderson, D.L., Tressler, J., 2014. Which journal rankings best explain academic salaries? Evidence from the University of California. Econ. Inq. 52 (4), 1322–1340. Halkos, E.G., Tzeremes, G.N., 2011. Measuring Economic Journals’ citation efficiency: a data envelopment analysis approach. Scientometrics 88, 979–1001. Kalaitzidakis, P., Mamuneas, T.P., Stengos, T., 2011. An updated ranking of academic journals in economics. Can. J. Econ. 44 (4), 1525–1538. Kosteas, V.D., 2015. Journal impact factors and month of publication. Econ. Lett. 135, 77–79. Malesios, C., 2015. Some variations on the standard theoretical models for the H-index: a comparative analysis. J. Assoc. Inf. Sci. Technol. 66 (1), 2384–2388. Oswald, A.J., 2007. An examination of the reliability of prestigious scholarly journals: evidence and implications for decision-makers. Economica 74 (293), 21–31. Phillips, P.C.B., Sul, D., 2007. Transition modeling and econometric convergence tests. Econometrica 75 (6), 1771–1855. Phillips, P.C.B., Sul, D., 2009. Economic transition and growth. J. Appl. Econom. 24 (7), 1153–1185. Vogel, R., Hattke, F., Petersen, J., 2017. Journal rankings in management and business studies: what rules do we play by? Res. Policy 46 (10), 1707–1722. Zheng, Y., Kaiser, H.M., 2011. Price premiums for journal quality and journal governance: evidence from economics journals. Econ. Lett. 112 (1), 125–127.
6
Contents lists available at ScienceDirect
Finance Research Letters journal homepage: www.elsevier.com/locate/frl
One list to fit them all: What do we learn from journal ranking? Konstantinos Eleftherioua, , Michael Polemisa,b ⁎
a b
Department of Economics, University of Piraeus, 80 Karaoli & Dimitriou Street, 18534, Piraeus, Greece Hellenic Competition Commission, Athens, Greece
ARTICLE INFO
ABSTRACT
JEL classifications: C23 A2 A11 I22
In this paper, we apply the Phillips and Sul (2007; 2009) methodology to investigate the convergence pattern of finance journals across thirteen established academic journal lists. The results reveal that the majority of sample journals do converge across the academic lists. The estimated transition paths confirm the empirical analysis, revealing a “focal” point for the research institutions to minimize the discrepancies appeared by the journal lists proliferation.
Keywords: Convergence Finance journals Club clustering Transition paths
1. Introduction It is widely acknowledged by the scientific community that publishing a research paper in a highly ranked economic journal is of great importance to researchers in terms of hiring or being tenured and promoted by an academic department or research institute (Gibson et al., 2014; Kosteas, 2015). This happens since journal rankings constitute a focal point (“signal”) for attracting young economists and retaining older ones in prestigious institutions (Kalaitzidakis et al., 2011). Moreover, journal rankings contribute to the allocation of research funding across the economic departments (Oswald, 2007). Journal rankings are also appeared in other disciplines such as finance, business and econometrics (Chung and Cox 1990; Baltagi, 2007). As a result, there is a fast growing number of academic journal lists consisting of different types (i.e., citation based, expert based, hybrid) and rating categories (i.e., continuous or discrete score) as reported by Vogel et al. (2017). A key question then arises by academicians in the sense of selecting the most efficient journal list to appropriate assess the research output. We attempt to exemplify this research question by investigating the existence of convergence patterns among the sample lists included in our empirical analysis and identify their transitions paths over the examined period (1999–2018). As a consequence, our study departs from the existing literature (see among others Halkos and Tzeremes, 2011; Zheng and Kaiser, 2011; Gibson et al., 2014; Kosteas, 2015; Malesios, 2015) since we attempt to uncover common trends on eighteen field finance journals appeared in thirteen sample journal lists. In this way, we are able to draw some policy implications concerning the dissemination of research conducted by finance departments worldwide. The main contribution of this study is that we examine for the first time, to the best of our knowledge, the convergence patterns among thirteen major scientific journal lists. Specifically, we study the convergence of journal rankings, which are related to different
⁎
Corresponding author. E-mail addresses: [email protected] (K. Eleftheriou), [email protected] (M. Polemis).
https://doi.org/10.1016/j.frl.2019.08.026 Received 15 July 2019; Accepted 28 August 2019 1544-6123/ © 2019 Elsevier Inc. All rights reserved.
Please cite this article as: Konstantinos Eleftheriou and Michael Polemis, Finance Research Letters, https://doi.org/10.1016/j.frl.2019.08.026
Finance Research Letters xxx (xxxx) xxx–xxx
K. Eleftheriou and M. Polemis
Table 1 Summary statistics. Variables
Observations
Mean
Standard deviation
Min
Max
Journal of Risk and Uncertainty Journal of Finance Journal of Financial and Quantitative Analysis Financial Management Journal of Accounting, Auditing and Finance Journal of Financial Research Journal of Futures Markets Journal of Financial Economics Journal of Business Finance and Accounting Journal of Banking and Finance Review of Financial Studies Journal of Empirical Finance Journal of Corporate Finance Finance and Stochastics Mathematical Finance Journal of Financial Intermediation Journal of Derivatives European Financial Management
260 260 260 260 260 260 260 260 260 260 260 260 260 260 260 260 260 260
2,9 4,0 3,8 3,0 2,3 2,0 2,5 4,0 3,2 3,4 4,0 3,1 3,3 3,1 3,1 3,4 2,1 2,1
0,9 0,0 0,4 0,8 0,8 0,9 0,9 0,0 0,5 0,5 0,0 0,7 0,7 0,7 0,7 0,5 0,8 0,8
1 4 3 1 1 1 1 4 2 3 4 1 2 2 2 2 1 1
4 4 4 4 3 3 4 4 4 4 4 4 4 4 4 4 4 3
top field journals from the finance discipline by employing the Phillips and Sul (2007, 2009) methodology.1 The rest of the paper is structured as follows: Section 2 presents, in detail, the methodological framework of our analysis. Section 3 describes the dataset, while Section 4 presents and discusses our main findings. Finally, Section 5 concludes the paper. 2. Econometric methodology The methodological framework used in the present study is the one proposed by Phillips and Sul (2007, 2009). The Phillips and Sul (PS) methodology as applied in our case can be outlined as follows: If Xit is the solely factor of a panel data set (X denotes the rating for a given journal, i denotes the journal list and t the time), then
Xit =
(1)
it µt ,
where δit measures the deviation of journal list's i ranking from the common trend μt and can be represented as it
=
i
+
i it L (t )
(2)
1t a,
where δi is fixed, ξit is weakly dependent over t with ξit ∼ iid(0, 1) and L(t) is a slowly varying function with L(t) → ∞ when t → ∞. The null hypothesis of convergence for all i (H0) versus the alternative of non-convergence for some i (HA) can be expressed as
H0 :
i
=
and a
0; HA:
i
(3)
or a < 0
The null hypothesis in (3) can be tested through the following regression:
log
H1 Ht
2 log L (t ) = c^ + b^ log t + u^t ,
(4)
for t = [rT], [rT] + 1, ..., T with some r > 0.2 In (4), Ht = (1/ N )
N i=1
(hit
1)2 , hit =
Xit N 1
N i = 1 Xit
=
, N 1 iN= 1 it it
L(t) = log (t + 1) and
b^ = 2a^ where a^ is the least-squares estimate of a in H0 (null hypothesis). Based on the above analysis, PS argue that the hypothesis of convergence can be tested through a one-sided t-test. Specifically, the alternative hypothesis of non-convergence [see (3)] cannot be rejected at the 5% level if t b^ < 1.65. Given the absence of sample convergence, we proceed in identifying convergence clubs by implementing the following clustering procedure proposed by Phillips and Sul (2007): (i) we order the N journal lists for each journal title putting the list having the highest last period ranking first, the one with the next highest ranking in the last period second and so forth; (ii) we form all the possible core clubs by selecting the first k highest-ordered lists (for each journal) [see step (i)], with 2 ≤ k ≤ N and we calculate the convergence statistic tk for the logt regressions [eq. (4)]. Then, the size k* of the club is determined by the maximum tk, given tk > −1.65; (iii) from the remaining N − k* journal lists, we add one list at a time to the previous defined core clubs if the t-statistic is greater than zero; (iv) for the remaining lists (if any), we repeat steps (i)-(iii) until the formation of new clubs is not possible. Finally, we apply Phillips and Sul (2009) procedure to determine the existence of further convergence clubs and calculate their transition paths. 1 2
For an overview of the advantages of this methodology over other convergence approaches, see Apergis et al. (2013). Following Phillips and Sul (2007, 2009), r is set equal to 0.3. 2
Finance Research Letters xxx (xxxx) xxx–xxx
K. Eleftheriou and M. Polemis
Table 2 Club convergence results. Journal
Panel A: Phillips and Sul (2007) log t t-stat
Journal of Risk and Uncertainty Full sample −0.3316 (0.3189) Journal of Financial and Quantitative Analysis Full sample −1.0175 (0.1340) Club 1 [1,2,3,4,5,6,8,10,11,12,13] −0.7894 (0.0143) Club 2 [7,9] – Financial Management Full sample −0.6539 (0.2856) Club 1 [7,8,10,13] 0.6107 (0.3514) Club 2 [1,2,3,4,5,6,9,11,12] – Journal of Accounting, Auditing and Finance Full sample −0.8200 (0.0667) Club 1 [1,3,4,6,11,12,13] −0.7894 (0.0143) Club 2 [2,5,7,8] −0.7894 (0.0143) Club 3 [9,10] – Journal of Financial Research Full sample −0.8724 (0.0531) Club 1 [1,3,5,11,13] −0.7894 (0.0143) Club 2 [2,4,6,7,8,9,10,12] −0.2416 (0.2458) Journal of Futures Markets Full sample −0.9156 (0.0366) Club 1 [1,3,4,6,11,12,13] −0.7894 (0.0143) Club 2 [2,5,7,8,9,10] 0.3291 (0.2432) Journal of Business Finance and Accounting Full sample −0.8750 (0.1545) Club 1 [2,8,10,11] −0.7894 (0.0143) Club 2 [1,3,4,5,6,7,9,12,13] −0.7973 (0.2753) Journal of Banking and Finance Full sample −0.7894 (0.0143) Club 1 [1,3,4,5,8,10,11] −0.7894 (0.0143) Club 2 [2,6,7,9,12,13] −0.7894 (0.0143) Journal of Empirical Finance Full sample −1.1008 (0.2206) Club 1 [5,6,8,10] 0.3639 (0.4714) Club 2 [1,3,4,7,11,12,13] 0.0957 (0.9364) Club 3 [2,9] – Journal of Corporate Finance Full sample −1.2621 (0.0715) Club 1 [1,2,3,4,5,8,10,11] −0.0523 (0.2638) Club 2 [6,7,9,12,13] −1.5769 (0.1254) Finance and Stochastics Full sample −0.7696 (0.1377) Club 1 [4,5,8,10] 0.3639 (0.4714) Club 2 [1,3,6,9,11,12,13] −0.7894 (0.0143) Club 3 [2,7] – Mathematical Finance Full sample −1.2056 (0.1176) Club 1 [1,4,8,10,11] −0.7894 (0.0143) Club 2 [2,3,5,6,9,12,13] −1.0991 (0.2153) Non converging Group 3 [7] Journal of Financial Intermediation Full sample −0.9516 (0.3699) Club 1 [1,3,4,8,10,13] −0.1331 (0.2506) Club 2 [2,5,6,7,9,11,12] – Journal of Derivatives Full sample −1.3496 (0.0783) Club 1 [3,8,13] 2.1193 (0.8668) Club 2 [1,2,4,5,6,7,9,10,11,12] −1.4009 (0.2329) European Financial Management Full sample −1.0121 (0.1323) Club 1 [1,3,4,5,11,13] −0.7894 (0.0143) Club 2 [2,6,7,8,12] −0.7894 (0.0143) Club 3 [9,10] –
New club
Panel B: Phillips and Sul (2009) Final club log t
t-stat
1+2
Club 1
−0.7894 (0.0143)
−55.2828⁎⁎
2+3
Club 2
−0.8262 (0.0813)
−10.1568⁎⁎
1+2
Club 1
−0.0519 (0.4707)
−0.1102
3
Club 2
–
–
1+2
Club 1
−0.4134 (0.1601)
−2.5825⁎⁎
2+3
Club 2
−1.0873 (0.1210)
−8.9826⁎⁎
1+2
Club 1
−0.9609 (0.1177)
−8.1658⁎⁎
2+3
Club 2
−1.5108 (0.2014)
−7.5017⁎⁎
1+2
Club 1
−0.7894 (0.0143)
−55.2828⁎⁎
2+3
Club 2
−1.0848 (0.1804)
−6.0129⁎⁎
−1.0396 −7.5938⁎⁎ −55.2828⁎⁎ – −2.2898⁎⁎ 1.7377 – −12.2937⁎⁎ −55.2828⁎⁎ −55.2828⁎⁎ – −16.4421⁎⁎ −55.2828⁎⁎ −0.9830 −24.9857⁎⁎ −55.2828⁎⁎ 1.3535 −5.6621⁎⁎ −55.2828⁎⁎ −2.8963⁎⁎ −55.2828⁎⁎ −55.2828⁎⁎ −55.2828⁎⁎ −4.9899⁎⁎ 0.7720 0.1022 – −17.6582⁎⁎ −0.1983 −12.5797⁎⁎ −5.5877⁎⁎ 0.7720 −55.2828⁎⁎ – −10.2477⁎⁎ −55.2828⁎⁎ −5.1059⁎⁎ −2.5724⁎⁎ −0.5311 – −17.2411⁎⁎ 2.4449 −6.0162⁎⁎ −7.6516⁎⁎ −55.2828⁎⁎ −55.2828⁎⁎ –
Notes: The numbers in parentheses denote the standard errors. The term log t denotes the convergence coefficient, while t-stat is the convergence test statistic. The latter is distributed as a simple one-sided t-test with a critical value of −1.65. ⁎⁎ denotes rejection of the null hypothesis (convergence) at 5% level of statistical significance. Convergence is implied when t-stat is equal to ‘-’.
3
Finance Research Letters xxx (xxxx) xxx–xxx
K. Eleftheriou and M. Polemis
Fig. 1. Transition paths.
3. Data and statistics We use a balanced panel data set of thirteen academic journal lists over a twenty year period yielding 260 observations. The variables used in the analysis refer to eighteen finance journal rankings appeared in all of the academic sample lists (see Appendix for details).3 To account for different rating categories among the journal lists, we appropriately convert the rating scale of each list to a four-level scale (i.e., from 1 for the lowest quality ranking to 4 for the highest) applicable to each sample journal. Moreover, since journal lists are not updated each year, we use the last available version of each list to complete missing observations. Table 1 presents the main descriptive statistics for the sample variables. We observe that three finance journals possess on average the highest journal ranking among the sample lists (i.e., Journal of Finance, Journal of Financial Economics, and Review of Financial Studies). This can be explained by the fact that these journals are ranked with the highest journal score awarded by the thirteen academic journal lists (Min = 4).4 On the contrary, three journals such as the Journal of Financial Research, followed by the Journal of Derivatives and the European Financial Management have the lowest mean (i.e., approximately equal to 2). 4. Results and discussion Table 2 reports the empirical findings of the club clustering algorithm based on Phillips and Sul (2007) methodology. The results indicate a rejection of the null hypothesis of full panel convergence at the 5% level of significance in all but one (Journal of Risk and Uncertainty) of the fifteen journal cases since t-stat < −1.65 (see Panel A). Therefore the investigation of potential club clusters among these lists is required. As it is evident from Table 2, in the majority of the cases, the club clustering algorithm identifies one convergence club (see Fig. 2) among the lists included in the analysis since the null hypothesis of convergence cannot be rejected. In three cases there are two convergence clubs reported (Financial Management, Finance and Stochastics, and Journal of Financial Intermediation). Finally, the Journal of Empirical Finance is the only journal with three convergence clubs. As a final step, we apply the club merging analysis along the lines of Phillips and Sul (2009). Based on our findings (see Panel B), we argue that the merging of clubs is not supported since the hypothesis of convergence is rejected in all but one (Journal of Empirical Finance) of the reported cases. Therefore, the initially formed clubs as described above are the appropriate ones. The transition paths of the estimated convergence clubs per journal (see Fig. 1) corroborate the above results. To sum up, the majority of the lists converge for the following five journals: Financial Management, Journal of Financial Research, Journal of Empirical Finance, Journal of Corporate Finance and Journal of Financial Intermediation. Only few lists converge for five journals (Journal of Financial and Quantitative Analysis, Finance and Stochastics, Journal of Future Markets, Journal of Derivatives and European Financial Management). On the other hand, there are no converging lists for the following three sample journals: 3 4
For our estimations, we used the STATA codes provided by Du (2017). Therefore, these three journals are excluded from the econometric analysis. 4
Finance Research Letters xxx (xxxx) xxx–xxx
K. Eleftheriou and M. Polemis
Fig. 2. Distribution of clubs. Notes: Each graph illustrates the corresponding journal list. The horizontal axis depicts the clubs for each journal, while the vertical axis the number of the club.
Journal of Business, Finance and Accounting, Journal of Banking and Finance and Mathematical Finance. 5. Conclusion In this paper, we implement the Phillips and Sul (2007, 2009) methodology to examine the convergence pattern of eighteen major finance journals across thirteen established academic journal lists. We conclude that with the exception of the three elite finance journals (i.e., Journal of Finance, Journal of Financial Economics and Review of Financial Studies) and the Journal of Risk and Uncertainty, journal lists’ rankings do not converge for the rest of the sample journals. However, journal lists tend to cluster into groups, indicating that at least some of them exhibit some uniformity in ranking finance journals. Given that employment in academia is heavily based on the quality of publications as measured by journal lists, our findings have important implications for the mobility of finance scholars across academic institutions in different countries. Declaration of Competing Interest None. Acknowledgments We would like to thank Patroklos Patsoulis for excellent research assistance in collecting the data and Thanasis Stengos for useful comments and suggestions. Appendix A. Finance journals [abbreviation] Journal of Risk and Uncertainty [JRU] Journal of Finance [JF] 5
Finance Research Letters xxx (xxxx) xxx–xxx
K. Eleftheriou and M. Polemis
Journal of Financial and Quantitative Analysis [JFQA] Financial Management [FM] Journal of Accounting, Auditing and Finance [JAAF] Journal of Financial Research [JFR] Journal of Futures Markets [JFM] Journal of Financial Economics [JFE] Journal of Business Finance and Accounting [JBFA] Journal of Banking and Finance [JBF] Review of Financial Studies [RFS] Journal of Empirical Finance [JEF] Journal of Corporate Finance [JCF] Finance and Stochastics [FS] Mathematical Finance [MF] Journal of Financial Intermediation [JFI] Journal of Derivatives [JD] European Financial Management [EFM] B. Journal lists used in the analysis [list #] Association of Business Schools Academic Journal Quality Guide (CABS) [1] Center National de la Recherche Scientifique (CNRS) [2] Australian Business Deans Council Journal Rankings List (ABDC) [3] ESSEC Business School Paris (ESSEC) [4] Foundation National pour l'Enseignement de la Gestion des Entreprises (FNEG) [5] Erasmus Research Institute of Management Journals Listing (EJL) [6] Association of Professors of Business in German Speaking Countries (VHB) [7] Danish Ministry Journal list (DEN) [8] SCImago (SJR) [9] High Council for Evaluation of Research and Higher Education (HCERES) [10] University of Queensland Adjusted ERA Rankings List (UQ) [11] European Journal of Information Systems 2007 Mingers & Harzing (EJIS) [12] WU Wien Journal Rating May 2008 (WIE) [13] References Apergis, N., Christou, C., Hassapis, C., 2013. Convergence in public expenditures across EU countries: evidence from club convergence. Econ. Financ. Res. 1 (1), 45–59. Baltagi, B.H., 2007. Worldwide econometrics rankings: 1989–2005. Econ. Theory 23 (05), 952–1012. Chung, K., Cox, R., 1990. Patterns of productivity in the finance literature: a study of the bibliometric distributions. J. Financ. 45 (1), 301–309. Du, K., 2017. Econometric convergence test and club clustering using Stata. Stata J. 17 (4), 882–900. Gibson, J., Anderson, D.L., Tressler, J., 2014. Which journal rankings best explain academic salaries? Evidence from the University of California. Econ. Inq. 52 (4), 1322–1340. Halkos, E.G., Tzeremes, G.N., 2011. Measuring Economic Journals’ citation efficiency: a data envelopment analysis approach. Scientometrics 88, 979–1001. Kalaitzidakis, P., Mamuneas, T.P., Stengos, T., 2011. An updated ranking of academic journals in economics. Can. J. Econ. 44 (4), 1525–1538. Kosteas, V.D., 2015. Journal impact factors and month of publication. Econ. Lett. 135, 77–79. Malesios, C., 2015. Some variations on the standard theoretical models for the H-index: a comparative analysis. J. Assoc. Inf. Sci. Technol. 66 (1), 2384–2388. Oswald, A.J., 2007. An examination of the reliability of prestigious scholarly journals: evidence and implications for decision-makers. Economica 74 (293), 21–31. Phillips, P.C.B., Sul, D., 2007. Transition modeling and econometric convergence tests. Econometrica 75 (6), 1771–1855. Phillips, P.C.B., Sul, D., 2009. Economic transition and growth. J. Appl. Econom. 24 (7), 1153–1185. Vogel, R., Hattke, F., Petersen, J., 2017. Journal rankings in management and business studies: what rules do we play by? Res. Policy 46 (10), 1707–1722. Zheng, Y., Kaiser, H.M., 2011. Price premiums for journal quality and journal governance: evidence from economics journals. Econ. Lett. 112 (1), 125–127.
6