The review process in economics: is it too fast?
Azar, Ofer H.
1. Introduction
The academic publishing process is an extremely important topic
because it affects the productivity of scholars in producing and
disseminating new knowledge, and yet it has received relatively little
attention in the academic literature. Research about the academic review
process is an important tool to making more informed decisions about how
we should shape this process. Such research, for example, may allow us
to make better decisions in issues such as the publication delay,
submission fees, and single- versus double-blind review. Although some
studies on the process of academic research have been written and even
published in top journals, (1) the research in this area is scant compared to its importance.
The long time it takes an article from its first submission to a
journal to its publication is one of the main criticisms of the academic
review process in certain disciplines. Especially upset about this long
delay are untenured faculty, who need to publish several articles in a
few years in order to get tenure. The first-response time (the time from
submission of the manuscript to receipt of the initial editorial
decision about it; henceforth denoted FRT) is a particularly important
part of the delay; as opposed to the time it takes to revise the paper
or the time from acceptance to publication, the FRT delays all
manuscripts submitted, not only the few whose authors are asked to
revise and resubmit or the few that are accepted. Consequently, the
average paper is delayed by the FRT several times (about three to six
times according to Azar 2004).
The long FRT in economics journals (often three to six months)
seems unnecessary. After all, referees usually do not need more than a
few hours to read a paper and write a report on it; neither do editors
need much time to make a decision once they obtain the referees'
reports. The short FRTs in leading journals in finance and accounting
(often one to two months) suggest that shorter FRTs are possible.
Indeed, editors of many economics journals try to reduce the FRT in
their journals, their motivation often being either to benefit the
profession or to attract more submissions. Whatever the editors'
motivation is, most people believe that these efforts are welfare
increasing. This article suggests that this common belief is not
necessarily correct.
The article argues that the current FRT may be below optimal, so
that efforts to reduce it are counterproductive, even though I claim
that reducing the FRT will not harm the quality of the review process.
The reason that reducing the FRT may be harmful is that it will increase
the number of submissions of low-quality papers to top journals, thus
increasing the workload of referees and editors without any significant
benefit in terms of the quality of research published. Moreover, the
increased number of submissions will raise the rejection rate, and each
paper will be rejected more times on average before it is published, so
the total time from initial submission to publication may not decrease
at all.
2. Are the Efforts to Reduce the First-Response Time Beneficial?
The aspect of the review process that receives maybe the most
criticism in economics is the long FRT. (2) Authors, especially
untenured ones, are upset that it takes several months to receive a
decision about the submitted manuscript. After all, the refereeing task
takes only a few hours. Hamermesh (1994), for example, suggests that it
takes six hours to referee an average paper. The Canadian Journal of
Economics provides advice to referees in which it states "The
amount of time taken with a paper can vary enormously--anything from a
couple of hours to a couple of days of full-time effort. A typical
report should probably take 3 or 4 hours." (3)
If it takes only a few hours to referee a paper, why does it take
several months to get an editorial decision? The main reason is that it
takes the referees a long time to return their reports, usually not
because they need a lot of time to ponder about the paper but because
papers wait a long time to be read. This may be the result of the
referee having higher-priority tasks, of procrastination, and maybe of
fear that prompt response will result in additional refereeing
assignments too soon.
The delay caused by the refereeing process makes the dissemination of research slower, and this is particularly important because new
research builds on previous work, so any delay causes the entire chain
of research to be delayed. Moreover, when it takes a long time from
writing an article to its publication, this reduces the chances that a
policy-oriented article will be published in time to be relevant,
deterring economists from writing such papers (Borts 1981). These costs
of the delay brought several economists to suggest ways to reduce the
delay (Hamermesh 1994; Pressman 1994; Szenberg 1994). Editors often
express their desire to shorten the review time (Ellison 2002a). (4)
Their reason, however, is often to attract authors rather than the
profession's welfare (Stulz 2000).
Are the efforts made by editors and others to shorten the FRT
beneficial from a social point of view? Most scholars think that the
answer is positive, as this enables faster dissemination of knowledge.
The few who think otherwise usually argue that shortening the delay will
reduce the quality of the review because referees will have less time to
prepare their reports. This argument, however, is hard to reconcile with
the fact that most of the delay is caused when the manuscript just waits
to be read. (5)
What I argue, however, is that even if shortening the FRT has no
effect on the review quality, it might not be optimal to shorten it
(obviously, if one believes that shortening the FRT reduces review
quality, this makes my claim even stronger). The reason is rooted in the
special structure of costs and benefits in the academic profession.
Basically, the idea is that the private monetary cost to submit an
existing manuscript to another journal is negligible compared to the
private benefits from a publication in a good journal. This cost is also
much smaller than the social cost of the review process. As a result, if
the FRT is very short, authors have an incentive to submit their
manuscript to many more journals than a social planner would like them
to. Authors do not internalize the costs that they impose on editors and
referees when they submit a paper. The FRT is an additional submission
cost from the author's perspective, and it therefore increases the
private costs of submission, reducing the number of submissions and
alleviating the workload on editors and referees. As a result, given the
current low submission fees in economics, shortening the editorial delay
without taking measures to prevent excessive submissions may in fact
reduce social welfare. (6) The following sections elaborate on these
ideas.
3. Why Does a Lower First-Response Time Lead to More Submissions?
To show why a lower FRT leads to more submissions, I present a
simple model about how the optimal submission strategy is determined.
The optimal submission strategy is a very complicated problem to solve
analytically, so to make the model traceable I use almost the simplest
framework possible and ignore interesting issues such as the differences
in FRTs between journals (for a discussion and empirical analysis of the
optimal submission strategy, see Oster 1980).
Assume that for a certain manuscript, there is a finite set of
journals that may publish it, and that they can be ranked according to
their quality, where quality is determined according to how much an
author gains from having a publication in the journal. Denote the number
of relevant journals by K, and let 1 be the highest-quality journal, 2
the second highest and so on. Let [G.sub.i] be the present value of the
gains from having a paper accepted by journal i (the i-th best journal),
for example, increased salary (the gains from publications are discussed
in detail in the following sections). By definition, [G.sub.1] [greater
than or equal to] [G.sub.2] [greater than or equal to] ... [greater than
or equal to] [G.sub.K].
The author can rank the quality of his paper, where quality J means
that the paper will surely be accepted by journals J, J+1, ..., K.
Clearly, the author will never submit the paper to the journals J+1,
J+2, ..., K, since he is better off submitting to journal J. There is
also a positive (but smaller than 1) probability that the paper will be
accepted in journals better than J; the probability of acceptance of a
quality-J paper in journal i is denoted by [q.sub.i](J). By definition,
[q.sub.i](J) = 1 for all i [greater than or equal to] J.
For simplicity I assume that [G.sub.1][q.sub.i](J) [greater than or
equal to] [G.sub.2][q.sub.2](J) [greater than or equal to] ... [greater
than or equal to] [G.sub.J-1][q.sub.J-1](J). It may be, however, that
[G.sub.J][q.sub.J](J) (which is equal to [G.sub.J]) is higher than
[G.sub.J-1][q.sub.J-1](J), and even higher than [G.sub.1][q.sub.1](J). I
also assume that each submission has a cost of c < [G.sub.K]. Let us
define [delta] = 1/[(1 + interest rate).sup.d], where d is the FRT.
Assuming that the author submits the manuscript to the next journal
immediately after receiving a rejection, the time between subsequent
submissions of the manuscript is equal to d. It follows that [delta] is
the discount factor according to which the author discounts the payoff
from the next submission.
Because both [G.sub.i] and [G.sub.i][q.sub.i] are nonincreasing in
i for all i < J, the author's optimal strategy is to submit the
paper first to the best m journals in a decreasing order (0 [less than
or equal to] m [less than or equal to] J - 1) and then to journal J.
This strategy has the obvious stopping rule: once the paper is accepted
at a certain journal, the author does not submit it anymore. To find the
optimal value of m, the author first considers two options: (A) submit
the manuscript immediately to journal J; (B) submit the manuscript first
to journal 1 and if rejected to J. Notice that if the utility from (B)
exceeds that from (A), it is better to submit the manuscript first to 1,
but not necessarily to then submit to J if it is rejected. The utility
from (A) is [G.sub.J] - c, whereas the utility from (B) is -c +
[q.sub.1](J)[G.sub.1] + [1 - [q.sub.1](J)][delta]([G.sub.J] - c), so it
is optimal to submit immediately to J (to choose m = 0) if and only if
[G.sub.J] > [q.sub.1](J)[G.sub.1] + [1 -
[q.sub.1](J)][delta]([G.sub.J] - c).
Similarly, if the author submits to journal 1 and receives a
rejection, he compares the utility from submitting to J immediately and
submitting first to 2 and if rejected to J. Submitting to J at this
point (i.e., choosing m = 1) is optimal if and only if [G.sub.J]
[greater than or equal to] [q.sub.2](J)[G.sub.2] + [1 -
[q.sub.2](J)][delta]([G.sub.J] - c). We can analyze the optimal decision
at any point in a similar fashion. The result is that the author submits
to journal i rather than to J as long as
(1) [q.sub.i](J)[G.sub.i] + [1 - [q.sub.i](J)][delta]([G.sub.J] -
c) > [G.sub.J]
and once this inequality is violated for a certain journal i, he
submits the paper to J. (7)
Given the value of [G.sub.J], if the value of [q.sub.i](J)[G.sub.i]
+ [1 - [q.sub.i](J)][delta]([G.sub.J] - c) is increased for all i, the
number of journals that the author tries before submitting to J (which
we defined as m) is also (weakly) increased. One way to increase the
value of [q.sub.i](J)[G.sub.i] + [1 - [q.sub.i](J)][delta]([G.sub.J] -
c) for all i is to reduce c. This implies that if the submission cost is
reduced, the author chooses to submit his paper to more top journals
before submitting it to the journal where it is accepted for sure. The
same idea applies to the FRT, which can be thought of as the time cost
of submission. Because [q.sub.i](J)[G.sub.i] + [1 -
[q.sub.i](J)][delta]([G.sub.J] - c) is increasing in [delta], it is
decreasing in d. It follows that a shorter FRT (lower d) causes
[m.sup.*] (the optimal value of m) to be higher.
In addition, the average number of submissions is increasing in m.
To see this, notice that the expected number of submissions is equal to
n(m) = [q.sub.1] + 2(1 - [q.sub.1])[q.sub.2] + 3(1 - [q.sub.1])(1 -
[q.sub.2])[q.sub.3] + ... + (m + 1)(1 - [q.sub.1])(1 - [q.sub.2]) ... (1
- [q.sub.m]) [using [q.sub.i] rather than [q.sub.i](J) to simplify the
notation]. It is immediately apparent that n(m) is increasing in m, and
therefore n[[m.sup.*](d)] is decreasing in d, implying that lower FRTs
increase the number of submissions. The fact that the number of
submissions is decreasing in the FRT suggests that there is a cost to
shortening the FRT, namely, the opportunity cost of the time of referees
and editors.
4. Private Costs and Benefits of Submissions
Based on an average of several studies, a publication in economics
journals ranked 1-10 (level 1), 11-55 (level 2), or 56-100 (level 3)
increases salary by about 3%, 1.7%, or 0.7%, respectively. (8) The costs
of submitting an existing manuscript have three main parts: the value of
the time required for printing and mailing the manuscript, the
submission fee, and the monetary value of the delay in reaping the
monetary rewards from publication because of the refereeing process.
Table 1 presents data about submission fees in different journals. The
value of the delay in publication caused by the refereeing process (from
the author's perspective) depends on the FRT; Table 2 presents the
FRT in various journals, showing that it is on average a little more
than 4 months. One can compute the cost of the editorial delay for his
salary, years until retirement, discount rate, and assumptions about
where the article will eventually be published. (9) Except for faculty
very close to retirement or faculty in countries where the contribution
of publications to salary is insignificant, it is the case that the
editorial delay is the major cost of submitting an existing manuscript
to a journal.
5. Optimal Submission Strategy
The optimal submission strategy given the hundreds of journals in
economics is a very complicated problem (for discussion of this problem
and numerical analysis for eight journals, see Oster 1980).
Consequently, I take a simpler approach, which is to find out the
optimal cutoff probability between submitting to level-1 journals and to
lower-quality journals: a cutoff probability of 4%, for example, means
that an author would find it optimal to submit his paper to level-1
journals only if he estimates that his acceptance chances (in each
journal separately) are higher than 4%. Otherwise, he is better off
submitting the paper to a lower-ranked journal in which he has higher
acceptance chances. The exact algorithm according to which the cutoff
probabilities were computed is omitted for the sake of brevity but is
available from the author upon request. It is based on comparing the
cost of submitting to a level-1 journal (which consists of the three
costs mentioned earlier) to the benefit (which is influenced by the
probability that the paper will be accepted in a level-1 journal and the
additional monetary benefit of publication in level 1 compared to level
2). Table 3 reports the cutoff probabilities for different values of the
FRT and the submission fee. (10) Today, the FRT of level-1 journals is
about 4 months, and submission fees are around $50, so the optimal
cutoff probability is about 4.5%. (11)
6. First-Response Times and the Number of Submissions
If submission fee on average is $50, how do different FRTs affect
the behavior of authors? Suppose that we could reduce the FRT to only
two months. We see from Table 3 that the cutoff probability will change
from 4.5% to 2.3%. What does it mean in terms of the number of
submissions? Because acceptance rates in the top five journals are
around 9%, and in the next five around 16%, it probably means many more
submissions. (12) The reason is that the distribution of the quality of
papers is very skewed. Many of the papers accepted at top journals had
an a priori acceptance probability much higher than the average
acceptance rate of 9%. Because the average a priori acceptance
probability is equal to 9%, this implies that hundreds of papers
submitted to each top journal have a priori acceptance probability lower
than 9%. It follows that reducing the cutoff probability from 4.5% to
2.3%, for example, can result in hundreds of additional submissions to
each of the top journals. Many authors of low-quality papers who today
are deterred by the long editorial delay and do not submit to top
journals will give luck a chance if delays are shorter. The same thing
will happen to lower-quality papers that today are submitted to level-3
journals but will be submitted to level-2 journals if the delay becomes
significantly shorter.
7. Additional Effects of Shorter First-Response Times
The discussion so far suggests that shortening the FRT will cause
many submissions of low-quality papers to good journals and thus create
a lot of additional workload for referees and editors. There are
additional important points in favor of a high FRT, however. Because a
reduced FRT will increase the number of submissions to top journals,
acceptance rates will drop, and papers will suffer more rejections
before they are published. The time they spend being rejected from
journals increases the total publication delay and may offset and even
exceed the time saved by shortening the FRT. As a result, we may not
only increase the workload of referees and editors but also increase the
total time that a paper spends from its initial submission to its
publication. Moreover, the increased number of submissions is likely to
lead journals to use less qualified referees, and referees to spend less
time reviewing each submission, both reducing the quality of the
refereeing process.
On the other hand, if referees provide helpful comments to rejected
papers also, and if authors revise their papers accordingly, papers
submitted to top journals and rejected become better, offsetting some of
the additional time costs of referees and editors. However, shortening
the delay might also induce authors to submit their papers in an earlier
stage and with a lower quality than they do today.
Another issue is the matching between journal quality and article
quality. Creating a good match is the main reason for the refereeing
process: it allows readers to focus their reading on top-quality
articles, and it facilitates the job of promotion and tenure committees.
How do more submissions affect this matching? If authors know very
little about the quality of their papers, and referees are very accurate
in their evaluation, inducing more people to submit to top journals will
increase the quality of top journals (some cases in which good papers
are not submitted to top journals will be eliminated), improving the
matching between article and journal qualities. If authors have a good
idea about the quality of their papers, and referees make some mistakes,
however, more submissions of low-quality papers (induced by a shorter
FRT) can actually reduce the average quality of top journals and hurt
the sorting function of journals.
Many economists feel that untenured faculty suffer the most from
the long FRTs because they have limited time to obtain sufficient
publications for tenure. This is incorrect, however, because untenured
professors compete among themselves. If shorter FRTs, for example, will
allow assistant professors to have more publications in their first few
years, the number of publications tenure committees require for tenure
will increase as well (Pressman 1994).
8. Conclusion
In light of recent efforts by editors to reduce the FRT, I examine
whether doing so is socially beneficial. I argue that the editorial
delay constitutes the major cost of submitting an existing manuscript to
a journal. Because the rewards of publication in top journals are very
high, a reduction in the editorial delay and therefore in the submission
cost will induce many more submissions of low-quality papers to top
journals. This has large costs in terms of the additional time that
editors and referees will have to waste to handle these papers.
Moreover, an increase in submissions will increase the rejection rate
and the average number of times that a paper is rejected before being
published. As a result, the total time from the first submission to
publication (potentially in a different journal) may not decrease much
and may even increase. If this total time decreases, it is hard to
compare the cost of referees' and editors' time with the
benefit of faster dissemination of new research. The discussion
suggests, however, that it is certainly possible that a shorter FRT is
not beneficial. In fact, a longer FRT might even be better than the
current FRTs for the same reasons that shortening the FRT might not be
beneficial. It follows that the efforts of editors to reduce the FRT,
although promoting the interest of the journal to attract authors, may
be socially undesirable.
References
Azar, Ofer H. 2004. Rejections and the importance of first response
times. International Journal of Social Economics 31:259-74.
Azar, Ofer H. 2005. The academic review process: How can we make it
more efficient? American Economist. In press.
Blank, Rebecca. 1991. The effects of double-blind versus
single-blind reviewing: Experimental evidence from the American Economic
Review. American Economic Review 81:1041-67.
Borts, George H. 1981. Report of the Managing Editor: American
Economic Review. American Economic Review 71:452-64.
Ellison, Glenn. 2002a. The slowdown of the economics publishing
process. Journal of Political Economy 110:947-93.
Ellison, Glenn. 2002b. Evolving standards for academic publishing:
A q-r theory. Journal of Political Economy 110:994-1034.
Engers, Maxim, and Joshua S. Gans. 1998. Why referees are not paid
(enough). American Economic Review 88:1341-9.
Hamermesh, Daniel S. 1994. Facts and myths about refereeing.
Journal of Economic Perspectives 8:153-63.
Hamermesh, Daniel S., and Sharon M. Oster. 2002. Tools or toys? The
impact of high technology on scholarly productivity. Economic Inquiry
40:539-55.
Laband, David N. 1990. Is there value-added from the review process
in economics? Preliminary evidence from authors. Quarterly Journal of
Economics 105:341-52.
Laband, David N., and Michael J. Pierre. 1994. Favoritism versus
search for good papers: Empirical evidence regarding the behavior of
journal editors. Journal of Political Economy 102:194-203.
Moore, William J., Robert J. Newmam and Geoffrey K. Turnbull. 2001.
Reputational capital and academic pay. Economic Inquiry 39:663-71.
Oster, Sharon. 1980. The optimal order for submitting manuscripts.
American Economic Review 70:444-8.
Pressman, Steven. 1994. Simultaneous journal submissions: the case
against. American Journal of Economics and Sociology 53:316-33.
Price, Gregory N., and Laura Razzolini. 2002. The returns to
seniority in the labor market for academic economists. Unpublished
paper.
Sauer, Raymond D. 1988. Estimates of the returns to quality and
coauthorship in economic academia. Journal of Political Economy
96:855-66.
Stulz, Renee M. 2000. Report of the Editor for his tenure and 1999.
Journal of Finance 55:1861-92.
Szenberg, Michael. 1994. Disseminating scholarly output: the case
for eliminating the exclusivity of journal submissions. American Journal
of Economics and Sociology 53:303-15.
(1) See, for example, Laband (1990), Blank (1991), Hamermesh
(1994), Laband and Piette (1994), Engers and Gans (1998), Moore, Newman,
and Tumbull (2001), Ellison (2002a, b), and Hamermesh and Oster (2002).
(2) In what follows, I sometimes use "editorial delay" or
just "delay" rather than "FRT," but they all mean
the same thing.
(3) See on-line at http://economics.ca/cje/en/referees.php.
(4) See also the editors' message of the Review of Economic
Studies at http://www.restud.com/report.htm.
(5) Another argument why shorter delay might reduce the review
quality is that to reduce the delay the editors would have to use less
qualified referees. Indeed, Hamermesh (1994) finds evidence that heavily
cited economists take a few more weeks to referee papers than others.
Whether those economists provide better referee reports is an
interesting question for future research, as they might be very busy and
therefore dedicate less time to their report.
(6) For a discussion of several potential measures to reduce the
FRT while preventing frivolous submissions, see Azar (2005).
(7) If the inequality (1) is satisfied for journal z and is
violated for journal z + 1, it is also satisfied for journals 1, 2, ...,
z - 1, and is also violated for z+ 1, z + 2, ..., J - 1. This follows
from the fact that [q.sub.i](J)[G.sub.i] + [1 -
[q.sub.i](J)][delta]([G.sub.J] - c) is decreasing in i for all i < J.
To see this, consider two journals x and y, where x < y < J. We
want to show that [q.sub.x](J)[G.sub.x] + [1 -
[q.sub.x](J)][delta]([G.sub.J] - c) [greater than or equal to]
[q.sub.y](J)[G.sub.y] + [1 - [q.sub.y](J)][delta]([G.sub.J] - c). If
[q.sub.x](J) [less than or equal to] [q.sub.y](J), this follows
immediately [recall that c < [G.sub.K] < [G.sub.J] and [q.s [G.sub.x] [greater than or equal to] [q.sub.y](J)[G.sub.y] If
[q.sub.x](J) [greater than or equal to] [q.sub.y](J), notice that
[q.sub.x](J)[G.sub.x] + [1 - [q.sub.x](J)][delta]([G.sub.J] c) [greater
than or equal to] [q.sub.x](J)[G.sub.y] + [1 - [q.sub.x]
(J)][delta]([G.sub.J] - c) [greater than or equal to]
[q.sub.y](J)[G.sub.y] + [1 - [q.sub.y](J)][delta]([G.sub.J] - c), where
the first inequality follows from [G.sub.x] [greater than or equal to]
[G.sub.y] and the second inequality follows from [q.sub.x](J) [greater
than or equal to] [q.sub.y](J) and [G.sub.y] [greater than or equal to]
[G.sub.J] > [delta]([G.sub.J] - c).
(8) Moore, Newman, and Turnbull (2001) found that a publication in
economics journals ranked 1-10, 11-55, and 56 and below, increases
salary by 2.9%, 1.7%, and 0.1%, respectively. The true contribution to
salary is slightly higher, however, because of the additional effect of
citations on salary. Sauer (1988) finds that including the effect of
citations on salary, publication in the top journal is worth an increase
of 3.8% in salary, and publications in the journals ranked as 10th,
20th, 40th, and 80th are worth 61.6%, 53.1%, 34.1%, and 18.9% of the
value of publication in the top journal. Price and Razzolini (2002)
estimate wage equations from censored salary data generated by grant
applications submitted to the National Science Foundation Economics
Program. A publication in the top six economics journals increases
salary by 0.5-3.6% (in the various specifications), and a publication in
any economics journal increases salary by 0.2-0.5%.
(9) For example, with a salary of $90,000, a discount rate of 6%,
and 30 years until retirement, assuming that the paper will eventually
be published in a level 1 journal, the publication increases annual
salary by $2,700 (3% of $90,000), so an editorial delay of 4 months
costs the author about $900. Similarly, for papers that will be
published eventually in level 2 or 3 journals, the cost of the editorial
delay is about $510 or $210, respectively.
(10) I thank an anonymous referee for making the point that
"time costs" and "money costs" are potentially
substitutes. This can also be seen in Table 3: we can shorten the FRT
and keep the cutoff probability (and therefore the number of
submissions) unchanged if we increase the submission fees. To keep the
cutoff probability at 4.5%, for example, if we reduce the FRT from 4 to
2 months, we have to increase submission fees from $50 to $375.
Increasing submission fees, however, has various other effects, such as
discriminating against authors from developing countries and in favor of
authors whose submission fees are funded by their institution or a grant
(or those who have a generous and flexible research budget from which
they pay submission fees). Examining in more detail the issue of optimal
submission fees is a worthwhile project but is beyond the scope of the
current article.
(11) Table 3 is based on the benefits associated with a publication
in level-1 and level-2 journals, so it does not apply directly to the
comparison between level-2 and level-3 journals. The main point,
however, that a small reduction in FRT will have to be compensated by a
very significant increase in submission fees to keep the cutoff
probability unchanged (and thus to prevent the number of submissions
from increasing), is similar when we compare level-2 to level-3 journals
as well.
(12) Details about how the acceptance chances of the top journals
were computed are available from the author on request.
Offer H. Azar, Department of Business Administration, School of
Management, Ben-Gurion University of the Negev, P.O. Box 653, Beer Sheva 84105, Israel; E-mail: azat@som.bgu.ac.il.
The author thanks Gadi Barlevy, Jacques Cremer, James Dana, Eddie
Dekel, Glenn Ellison, Ricky Lain, Nisan Langberg, Nadav Levy, Robert
Porter, William Rogerson, Michael Whinston, Asher Wolinsky, and
especially the Editor, Laura Razzolini, and two anonymous referees for
helpful discussions and comments. Financial support from The Center for
the Study of Industrial Organization at Northwestern University is
gratefully acknowledged.
Received September 2003; accepted March 2005.
Table 1. Submission Fees in Various Journals
Submission Fee
(Members or Submission
Journal Subscribers) Fee (Others)
Economics journals
American Economic Review $100 $200
Canadian Journal of Economics $25 $65a
Econometrica $0 $30a
Economica $0 $49a
Economic Inquiry $100 $160
International Economic Review $65 $125
Journal of Economic Theory $0 $0
Journal of Labor Economics $0 $0
Journal of Mathematical Economics $0 $0
Journal of Monetary Economics $100 $175
Journal of Political Economy $75 $125
Quarterly Journal of Economics $0 $0
RAND Journal of Economics $50 $85
Review of Economics & Statistics $0 $58
Review of Economic Studies $0 $0
Southern Economic Journal $50 $110 (a)
Accounting journals
The Accounting Review $125 $200
Journal of Accounting & Economics $250 $300
Journal of Accounting Research $250 $250
Finance journals
Journal of Finance $70 $140
Journal of Financial Economics $500 $550
Review of Financial Studies $125 $175
The data were taken from the journals' websites in March 2005.
Where submission fees differ according to the author's geographic
location, they refer to U.S. submissions.
(a) Submission fee for nonmembers and nonsubscribers includes an
annual membership or subscription.
Table 2. First-Response Times (FRT) in Various Journals (in Days)
Median Mean Source/
FRT FRT Period Journal Issue
Economics journals
Quarterly Journal NA 47 1997 Ellison (2002a)
of Economics
114
82
Canadian Journal 91 1/1/02- The journal's
of Economics 12/1/02 website
Journal of 103 108 2000/2001 September 2001
Economic History
Southern Economic 107 122 2001 October 2002
Journal
American Economic 109 122 7/1/00- May 2002
Review 6/30/01
Econometrica 110 122 2000 January 2002
98 92
108 122
Economic Journal 137 137 2000 RES Newsletter
(Jan 2003)
137 125
168 188
European 143 165 2000 May 2002
Economic Review
RAND Journal 153 131 2000 Summer 2002
of Economics
Economic Inquiry NA 159 1/1/02- October 2002
4/15/02
Journal of NA 167 2000 Ellison (2002a)
Political Economy
Review of 175 171 9/2000- The journal's
Economic Studies 2/2001 website
194 198
159 138
Accounting journals
Accounting Review 51 52 3/1/01- July 2002
2/28/02
Journal of Accounting 52 53 12 months August 2002
and Economics ending
4/2002
Finance journals
Journal of Financial 37 43 10/2001- The journal's
Economics 9/2002 website
Journal of Finance 39 44 3/l/00- The journal's
7/31/02 website
Median Mean
FRT FRT Comments
Economics journals
Quarterly Journal NA 47 All papers
of Economics
114 Accepted papers
only
82 Papers sent to
referees
Canadian Journal 91
of Economics
Journal of 103 108 Including
Economic History resubmissions
Southern Economic 107 122 New submissions
Journal only
American Economic 109 122 Rejected papers
Review only
Econometrica 110 122 New submissions
only
98 92 Revisions only
108 122 All papers
Economic Journal 137 137 All papers
137 125 Letters advising
rejection
168 188 Letters inviting
revision
European 143 165
Economic Review
RAND Journal 153 131 Simple average of
of Economics the four quarters
of the year
Economic Inquiry NA 159
Journal of NA 167
Political Economy
Review of 175 171 New submissions
Economic Studies only
194 198 First revision
159 138 Second revision
Accounting journals
Accounting Review 51 52 Including
resubmissions
Journal of Accounting 52 53
and Economics
Finance journals
Journal of Financial 37 43
Economics
Journal of Finance 39 44 Including
resubmissions
Additional details about the computations performed (in those
cases that the journals publish the distribution rather than
the mean or median) can be obtained from the author on request.
Table 3. Optimal Submission Strategy: Cutoff Probabilities between
Level-1 and Level-2 Journals
Fee
Delay
(Months) $0 $50 $100 $150 $200
0.04 0.2% 0.5% 0.8% 1.1% 1.4%
1 1.0% 1.3% 1.7% 2.0% 2.3%
2 2.0% 2.3% 2.6% 3.0% 3.3%
3 3.0% 3.4% 3.7% 4.1% 4.4%
4 4.2% 4.5% 4.9% 5.3% 5.6%
5 5.4% 5.8% 6.2% 6.5% 6.9%
6 6.7% 7.1% 7.5% 7.9% 8.2%
8 9.6% 9.9% 10.3% 10.7% 11.1%
10 12.5% 12.9% 13.3% 13.7% 14.0%
12 15.6% 16.0% 16.3% 16.7% 17.0%
Fee
Delay
(Months) $250 $300 $350 $400
0.04 1.7% 2.1% 2.4% 2.7%
1 2.6% 3.0% 3.3% 3.6%
2 3.7% 4.0% 4.3% 4.7%
3 4.8% 5.1% 5.5% 5.8%
4 6.0% 6.3% 6.7% 7.1%
5 7.3% 7.6% 8.0% 8.4%
6 8.6% 9.0% 9.3% 9.7%
8 11.4% 11.8% 12.2% 12.5%
10 14.4% 14.8% 15.1% 15.5%
12 17.4% 17.8% 18.1% 18.5%
The numbers in the table represent the cutoff probability when an
author has to choose whether to submit his article to a level-1
(top 10 journals) or level-2 (journals ranked 11-55) journal. If the
probability of acceptance in level-t journals is higher than the
cutoff probability, the author should submit to a level-1 journal,
otherwise he should submit to a level-2 journal.