Tips on Writing a Referee’s Report

As research academics we spend a substantial amount of time r eviewing papers for scholarly journals. While not as important as publishin g our own research, the quality of our work as referees is important, both for our pro fession and for our success as scholars. This note presents some suggestions fo r writing good referee reports.


Introduction
As research academics we spend a substantial amount of time reviewing papers for scholarly journals.While not as important as publishing our own research, the quality of our work as referees is important, both for our profession and for our success as scholars.The public benefit for the profession is that feedback from referees helps us to do research that better Invited article.Accepted 18 April 2013.Published on-line 30 May 2013.Supervising editor: Ricardo P. C. Leal.This article was originally prepared as a teaching note for our doctoral students.We owe thanks to Phillip Dybvig, Rocky Higgins, Jon Karpoff and Paul Malatesta for helpful comments on earlier versions.
*University of Southern California, California, United States.E-mail: ferson@ marshall.usc.edu**University of Southern California, California, United States.E-mail: matsusak@ marshall.usc.educontributes to knowledge.There is also a significant private career benefit to writing conscientious referee reports.The quality of a scholar's work as a referee signals his or her competence as a researcher and enhances the scholar's reputation with journal editors.Good refereeing may lead to appointments to conference program committees.Activities such as refereeing and program committee work are considered in faculty review and tenure decisions.Some people believe that a reputation for being a good referee is correlated with getting good referees on their own papers.Journals tend to draw their editorial boards from the pool of good referees.At the same time, most would agree that too much time spent reviewing a bad paper is not a good allocation of resources.A junior (well, any) faculty member's time is limited, and refereeing activities should not crowd out a scholar's own research.This note presents some suggestions for writing good referee reports.
A typical referee report consists of two parts.The first is a cover letter to the editor, not usually seen by the author(s) of the paper.The second part is the main body of the report, which typically includes a very brief summary of the paper, followed by criticisms and constructive comments for the authors' benefit.The main body of the report is sent to the author(s), but the name of the referee typically is not disclosed.As a referee, the paper's authors' names may or may not be revealed to you, depending on the journal.In recent years most journals have moved to web based processing systems.As a result, the previous practice of offering margin comments for the authors' benefit has become uncommon.
Occasionally you will be asked to review a paper, where it is obvious that the paper is not appropriate for the journal that has asked you to review it.In such a case, the question is whether to review the paper anyway, or to simply respond with a brief statement as to why the paper is not appropriate for the journal.Another, related situation is when a journal sends you a paper that you are not really qualified to review.The journal business is increasingly competitive, and turn-around time is one of the important dimensions on which journals compete.Editors would rather know right away if the paper is not appropriate for the journal or the referee.We suggest an immediate response to this effect, either by email to the editor or through the web-based system.The editorial office would like to quickly ask the next referee.
What characterizes the very best reports?If the recommendation is to reject, then one might say that the gold standard would be met if the 10 Rev. Bras.Financ ¸as (Online), Rio de Janeiro, Vol. 11, No. 1, March 2013 authors "liked" the rejection.Not that anyone likes being rejected by a journal, but it would be great if the author would find the report so valuable for the research going forward, that it turned the pain of rejection into a positive experience.This is typically not achieved.If the recommendation is to revise and resubmit, a gold standard report provides explicit, feasible advice on how the author should improve the ultimate impact of the paper on the profession and the paper's contribution to knowledge.

How to Evaluate a Paper
In evaluating a paper, consider two main things: (1) the importance of the question or findngs in the context of the literature; and (2), the execution.The first criterion is more important.Even a perfectly executed paper is not suitable for a top journal if the question or findings are not important enough.Similarly, a paper with a potentially important finding might be attractive to a top journal even if the execution in the current version needs improvement.Junior scholars sometimes focus more on the mechanics of (2), perhaps because they are more comfortable with that than with their own judgment about ( 1).An editor will sometimes pick a junior scholar as the referee because they have concerns about technical aspects of the paper and want someone to ferret this out.However, you were probably asked to be the referee because the editor wants to hear your judgment and opinion about the overall contribution of the paper.
Of course, the importance of a paper is a matter of opinion, and this judgment requires having some context in the literature.What are the two or three main papers that are related to the submission, what is the state of knowledge about the question, and how does this paper potentially advance this state of knowledge?Placing the paper in the context of the literature should help you to think about its importance.A very brief summary can help the editor, or even the author, to understand the paper's place in the literature.A summary may help the author see connections to the literature that he or she did not recognize, or at least to consider how one specialist reader sees the connections.However, a referee report is not a literature survey, so don't write too many sentences about this in your report: one or two sentences, or at most a short paragraph, is all you should need to summarize your views on how the paper relates to the literature.
Execution includes the quality of the writing and exposition, the structure and elegance of the model, the appropriateness of the empirical methods, data, and so on.When you identify a problem with execution you should try to suggest a solution.If there is not a clear solution to a serious problem in the execution, that is probably grounds to recommend that the paper be rejected.
There is some difference of opinion on the right extent of referee involvement in a paper.Some have noted that the referee process in finance seems to have evolved to a situation where referees spend a lot of energy through several rounds of revisions, asking authors for many small changes that polish the paper to a fine sheen, as seen from their perspective, as opposed to focusing on whether the paper suitable for publication or not.This high level of involvement slows the publication process, the rate at which new results are disseminated, and the amount of time authors and referees have to do more impactful research work.Ultimately it is the author's paper, not the referee's paper.Sometimes, a referee will appropriately ask for robustness checks in order to help determine whether a result has some external validity.However, when deciding whether to ask authors to produce additional evidence, robustness tests, generality in the proofs and so on, you should consider the costs and the benefits.Will this change lead to a material improvement in the paper?If the change only leads to a minor improvement, it may not be worth the costs.We recommend that you not ask authors to make every change that would improve the paper, only those changes where the improvement is worth the costs in authors' and referee's time, and the delay in publication.

The Body of the Report
Most referee reports are between two and ten single-spaced pages in length.However, a simple recommendation to reject a low quality paper can be shorter.Some are only a paragraph or two, giving a summary judgment and its basis.While this type of review is less satisfactory from the author(s) perspective, some editors find the shorter review perfectly acceptable when the paper is straightforward and the reasons for a rejection are clear.Some of the most senior and established scholars tend to use the shorter form.However, personally -and other editors that we have spoken with have agreed -it is annoying when a junior scholar tries to mimic a senior scholar by adopting this form.An editor can tell the difference.
Since your identity is not to be revealed to the authors, do not put your name on the report, and if you send the report as a *.pdf through the web system, be sure to disable any indentifying code.Start the report with the title and manuscript number of the paper and the journal for which the report is written.The report should never be sent directly to the authors.Send it to the editor who asked you to do the review, or use the web-based management system.
The first paragraph of the report is typically a short summary of the paper.Make it clear and succinct, focusing on the main important features of the paper, how the paper is motivated in relation to the literature, and what the contributions of the paper are to the literature.You want this paragraph to convince the editor and the author that you understand what is going on in the paper, and it helps to jog the editor's mind in reading your report.You don't want too much detail here, as the editor should read the paper, too.(However, in many cases an editor will only skim a paper when it is in an early stage of the review process.)In this introductory paragraph, don't second guess the authors' interpretation of what the paper contributes; rather, take the authors' representation as given.You don't want to give a rejected author too easy a basis for disputing the validity of your report.(Example: "Look at the first paragraph.The stupid referee doesn't even understand what my paper is about!")If you disagree with the authors' interpretation, explain the reasons for your disagreement later in the report.
A second paragraph can be used to summarize your overall opinion of the strengths and weaknesses of the paper, and your view of its potential contribution to the literature.If you are negative on a paper due to a lack of importance, then make this clear to both the author and the editor.We have seen cases where a referee reacts negatively to a paper because it is not important enough, but instead of being explicit about that, the report focuses on a long list of minor secondary and technical issues with the execution.Sometimes, this reflects insecurity on the referee's part about their judgment on the importance.Comments on minor and technical points can be helpful to even a rejected author in revising the paper, but if the core issue is that the contribution is small, it is better to be explicit about that.It serves the authors and the editor best if you accurately represent the nature of your concerns.
If the report is lengthy, you might list the main issues that the report will address in more detail, and then describe how the rest of the report is organized around those issues.Normally, do not put your recommendation to reject or allow a revision in the report.That goes in the letter to the editor.
We have seen instances where referees explicitly state their opinion on the editorial decision (accept, revise, reject) in the initial paragraphs of a report.But this is the editor's, not the referee's decision to make.Sometimes an editor will make a decision other than the one recommended by a referee.For example, there may be a second referee with a different opinion.Your report should allow the editor this prerogative.However, it is important that the message to the authors, from the body of the report, be consistent with the recommendation that you make to the editor in your cover letter.If the cover letter tells the editor that the paper is horrible, but the body of the report seems glowing to the author, the author might be confused when the editor rejects the paper.
Subsequent paragraphs of the report should provide and explain your criticisms and suggestions for improving the paper.This point is particularly important.If you find some aspect of the paper unacceptable, do not simply note that fact, but try to offer the authors a concrete, constructive suggestion on how to address your concern.If you find a major, fatal flaw in the paper, discuss this early in the report.(For example, Theorem 1 is wrong and the whole paper relies on it.Present your counterexample to the theorem right away!)It is often useful to organize and group your comments into categories.For example, comments on the theory, the empirical methods, the authors' interpretation of the evidence, etc. may be grouped together.Comments directed at improving the exposition of the paper are also an appropriate category.It is useful to number your main points or otherwise set them out in some organized way for easy reference in the future.This is helpful in the event that a revised paper is later reviewed again, and makes it easy to see if the revision addresses the main points raised in the previous report.
As a general point, give some thought to the tone of your report.Your goal is to help the authors improve their paper by pointing out potential weaknesses in the argument and execution, and by suggesting constructive paths they might follow to address those weaknesses.The goal is not to simply accumulate a list of grievances to justify a rejection.Even the roughest papers may contain the kernel of an interesting idea that you can highlight and help the authors pursue.In short, you will write the most constructive reports if you approach them with the spirit of helping your fellow scholars improve their research, and so advance knowledge in the field.

The Cover Letter
This should be written as a formal letter to the editor.You can use an electronic version of your university letterhead for style points.Here is where you tell the editor your frank opinion of what you think that he or 14 Rev. Bras.Financ ¸as (Online), Rio de Janeiro, Vol. 11, No. 1, March 2013 Tips on Writing a Referee's Report she should do about the paper.This letter will not be shown to the authors in most cases, although editors may excerpt or paraphrase from the letter in justifying their decision to the authors.The bottom line is whether the paper should be: (1) rejected, using language that does not encourage a resubmission; (2) rejected, but a resubmission should be allowed which addresses the concerns in your report; (3) rejected, but a resubmission which responds to the suggestions and comments in your report should be encouraged; or (4) accepted for publication.It may be useful to begin a cover letter with a very abbreviated version of the first paragraph of the report, reminding the editor which of the many papers under review is being discussed.
Be explicit in the letter about your recommendation.Don't make the editor guess about what you would do if the choice was yours.(It is not, but an editor will often place a lot of weight on the advice of a good referee.)If this is the first submission of a paper, choice (4) acceptance, is rare.If you choose (1), rejection, then explain to the editor the reasoning that justifies the rejection.If you recommend that a resubmission be (2) allowed, or (3) encouraged, then be as explicit as possible about what you would recommend that the editor say to the authors in his or her letter to them.For example, do you view some of your suggestions as crucial and others as less important?The explanation to the editor need not be lengthy.Keep it to a sentence or two, unless you have things to say that you did not want to put in the report for the authors to see.
It is appropriate to alert the editor to any aspects of the paper where you do not feel competent to render useful judgments.Never attempt to bluff when you don't understand something.If you bluff and make a mistake it gives the rejected author a good reason to contest the rejection decision, and the editor will not by happy.It is better to admit that you are confused about or don't follow some argument.Then, the onus is on the author to explain things more clearly.
In some cases, you might have seen the paper as a referee for a different journal (presumably, it was rejected).It is appropriate to inform the editor about this.In most instances, you should let the editor know about this right away, and offer some first impression of how much or little the paper seems to have been revised compared to the version that you saw before.The editor might want to get a fresh read from a different referee, or might want you to review it again.An author would be foolish to send a rejected paper unchanged to a new journal, but we have seen this happen, and the editor should know if this is the case.If the paper was revised in response to your earlier report, the situation becomes similar to a second-round review of the paper.Part of the decision process that an editor undertakes about a paper involves a judgment over how responsive the authors are likely to be to the comments they receive during the review process.If you have information about this based on your previous review, the editor may find this useful.
It can be very rewarding to craft an excellent referee report, not just for the personal satisfaction in a job well done, but there is a sense of satisfaction in knowing that you have contributed positively to the quality of a research colleague's work and to the level of research in the profession.