Is it getting harder to publish?

Another week, another rejection letter. Am I writing bad papers, or is it getting harder to publish?

I can think of several reasons why it might be getting harder to publish.

1. Many new faculty members were hired in the last five to ten years. All of those people are trying to get tenure. In order to get tenure, they need to publish - so there are more and more people chasing the available journal slots.

2. The technology of research has changed. Many data sets are readily available on-line. Stata makes erstwhile complicated procedures, for example, calculating marginal effects on a probit model, as easy as typing mfx and hitting enter. So it's easier to generate research findings.

3. The technology of submitting articles to journals has changed. Back in the day, submitting an article to a journal took hours - to go to the library and find out the submission information from the hard copy of the journal, print out four copies of the article, plus cover letter, and mail it all out. Now it's just a matter of pressing a few buttons: "attach file" "submit."

But on the other hand:

4. The technology of publishing has changed. If typesetting is outsourced to low wage destinations and everything is distributed electronically, per page costs may have decreased in real terms (I don't know, I'm just speculating here), which would allow journals to accept more articles.

What's the evidence? Each year, the Canadian Journal of Economics provides data on acceptance rates and the number of articles submitted to the journal (it's at the end of the November issue under "News of the Association").

There is an upwards trend in the number of submissions, providing some support for the hypotheses one through three above.

There is perhaps some suggestion of a downwards trend in acceptance rates. The News of the Association in the early 1990s reported acceptance rates of 25 percent and noted that these were "roughly the historical norm." Acceptance rates now are about 18 percent. But as the diagram shows, acceptance rates jump up and down a bit.  The current low rates could just represent random variation.

 

A journal's acceptance rate is determined in part by the editor of the journal and in part by whoever decides the page budget (that is, the number of pages in each volume of the journal). There is some danger, therefore, in attempting to explain acceptance rates in terms of big demographic/technological trends - they might simply reflect the whims of the editor.

And now...find another journal, press "attach" and "submit" one more time...

 

Comments

[Nick Rowe]

There's also an upward trend in the number of journals, right? That gives rise to 2 thoughts:

1. If the number of journals has grown more slowly/quickly than the number of economists, that would make it harder/easier to publish over time.

2. Look at all the journals that existed (say) 50 years ago. Are those journals now "top" journals? If so, it is harder to publish in a "top" journal than it was 50 years ago. (What we now think of as "top" journals weren't top then.)

Sometimes I'm thankful I'm old.

[Frances Woolley]

Nick: "If the number of journals has grown more slowly/quickly than the number of economists, that would make it harder/easier to publish over time."

If you were to extend the acceptance rate graph back for more years, there would be other periods when the CJE acceptance rates were under 20 percent, reflecting, I suspect, changes in the journal's reputation over time, demographics and, as you say, the number of journals.

"Look at all the journals that existed (say) 50 years ago. Are those journals now "top" journals?" some are, e.g. JPE, AER, EJ. Some not so much, e.g. Scottish journal of political economy (57 years old). But you're right, it is hard to think of examples of old low-ranked journals.

So if we go far enough back in time an AER was a fairly ordinary publication?

[Jack]

Couple thoughts:
(1) More economists, more *international* economists submitting to Anglo journals, and also increased publication expectations coming from lower-ranked universities, where it used to be enough to just present in conferences;

(2) The lower acceptance rates are if anything biased upward because there are more lower-tier journals today, so the average quality of submissions at top journals (say, top general, and top field) has increased a lot, in addition to the higher number of submissions

So yes, it is getting harder and harder to publish in good journals, and increasingly departments are discouraging submission of papers to lower ranked journals (where quality control may be an issue)

[Account Deleted]

Frances:

I'm certainly finding it harder to publish. All of your explanations sound plausible. As another, how about the uber-specialization even within academic disciplines? It certainly makes a putative match between a given article and the editor handling it more difficult.

In certain fields, such as development, it's now de rigeur to do randomized field trials. These are heavy on infrastructure and money and eliminates most economists (including me) who don't have the resources.

That's why I am following Nick and you into the econo-blogosphere!

Cheers

Vivek

[Frances Woolley]

Jack: The News of the Association does give a breakdown of submissions by place of origin, but I don't recall the numbers off hand.

I do know that the acceptance rate for Canadian-based economists at the CJE is much higher than the acceptance rate for non-Canadian-based economists, which I suspect reflects the quality of their submissions (Canadians know it's not easy to publish in the CJE so don't bother to submit poor quality papers). This has been true for years. So overall acceptance rates underestimate the acceptance rates for Canadian-based economists.

Your post raises some questions about why we're all in this rat race anyways - what is the social value of research that editors and referees consider unpublishable?

And an unrelated note: what is remarkable is not that the CJE acceptance rate appears to have declined slightly over time, but that it hasn't shot up far more than it has - I suspect this reflects a deliberate policy decision to manage the journal in a manner consistent with the best interests of the Canadian economics profession.

[Mike Moffatt]

I've purposely been staying out of these threads RE: journal publishing. I think by now everyone knows my position on the system - the incentives are so mismatched that we need to rip it up and start again. But as a self-selected outsider, that's a really easy position for me to take.

I just wanted to send a quick note to let you know how impressed I am at both the quality of the posts and comments, as well as the willingness of *insiders* to ask tough questions about the system. It's not at all an easy thing to do. Great job Frances et. al!

[Jim Sentance]

It's not just new faculty looking for tenure, but graduate students looking for a leg up in the employment market. Having been there (demand side) for the past two years it's pretty clear that the objective of most job market candidates is to have a publication or two, or at least and R&R, before heading out. Multiply that by the hundreds on the market every year...

[Frances Woolley]

Jim - pity the poor journal editors!

[Larry Willmore]

"The technology of research has changed. Many data sets are readily available on-line. Stata makes erstwhile complicated procedures, for example, calculating marginal effects on a probit model, as easy as typing mfx and hitting enter. So it's easier to generate research findings."

It's also easier to generate and publish rubbish. I just finished some research on temporary migration in China, estimating a multinomial logit (MNL) model. In a review of the literature, every author - with one exception - who entered an independent variable and the square of the variable into the model and reported "marginal effects" reported effects for variable x AND effects for the square of x. This is absurd, as it is impossible to change x without changing its square. It also led them to claim that a variable had a positive effect on migration when, in fact, the relationship was positive to a point, and then became negative. They simply highlighted the positive coefficient on x, ignoring the negative coefficient on x squared. I am not making this up. These papers passed peer review and were published in reputable journals. This is the danger when it is possible to generate results by mindlessly "typing mfx and hitting enter".

[James in Ottawa]

Here is a link to an argument that it has gotten easier to publish but that effort/quality has gone down. Some food for thought in this discussion.

[Frances Woolley]

David Green, editor of the Canadian Journal of Economics, adds these comments:
This does interest me. As one of the entries said, I
did run some numbers a year or so back to see about
country of origin of the submissions (I wrote it up in my
annual report so it should be posted as part of the CEA
annual report - I think its the May 2009 report). What I
found was that the acceptance rate for submissions from
Canadian institution based people remains in the mid to
high 20 percent range and has been around that for a long
time. Our overall acceptance rate has declined because we
have had a burst of submissions from outside Canada that
tend to have a much lower acceptance rate.
The timing of the burst in submissions to the CJE
seems to coincide pretty closely with when we went to all
on-line submissions. Now I think its easy for submissions
to come from various places in the world. I saw a write-up
about the JHR that showed a similar boost in submissions
when they went on-line.