How to Write a Reference Letter for the Economics Job Market
Why I am writing this
Over the last few years I have served on many hiring committees. Last year alone, I went through close to 1,000 applications. This guide collects what I have learned from reading the letters of recommendation in those applications, and what would have made my job as a committee member easier.
The main lesson is that there is a code for writing letters of recommendation. Top departments in the US know this code. Letter writers at less highly ranked or non-US institutions often do not. This matters because a hiring committee cannot tell the two cases apart. When a letter deviates strongly from the code, I do not know whether the writer is simply not aware of it, or whether the writer wants to send a signal about the student. Since the second interpretation is always possible, the deviation hurts the candidate when we decide whom to move forward.
Below are a few practical steps for letter writers. The list is not exhaustive, and I am happy to be corrected.
Why letters matter. A few findings from the research on letters in our own job market:
- Hirtle and Kovner analyze 6,400 letters for more than 2,200 economics and finance PhD candidates. They find that letter length and a recommendation to “top” departments are positively associated with initial placements and with top journal publications (Hirtle and Kovner, NY Fed Staff Report 1129).
- Slonimczyk uses an LLM to measure the sentiment of confidential letters and finds that both letter quality and length predict job market success. Disagreement among the letter writers of one candidate hurts the candidate (Slonimczyk, arXiv).
- Blickle and Parlatore argue that experienced readers learn to interpret the style of well-known letter writers. This is one reason why letters from known advisors carry more weight, and why letters travel less well across fields (Liberty Street Economics, 2025).
Length: a short letter is a bad signal
A short letter is often read as a sign that the student is weak. The committee asks itself a simple question: why did the letter writer not want to write more about this person? One page is too short. In a survey of letter writers run by the health economics association ASHEcon, around 80% of respondents considered any letter below 1.5 pages a “short letter” (ASHEcon newsletter, 2020).
How long is a normal letter? The same survey reports an average of 2.6 pages for letters that respondents write, and a median of 3 pages for letters they receive. Hirtle and Kovner find an average of around 1,150 words per letter, with the middle half of letters between 780 and 1,430 words (NY Fed Staff Report 1129). Chris Blattman puts good letters at 2 to 3 pages (Blattman, 2022). In my experience, well-known letter writers from strong departments aim for the longer end, with letters of up to 6 pages. The survey respondents do see 4 pages or more as too long, so length alone is not the goal. Aim for around 2.5 to 4 pages of real content.
Content: what goes into the letter
A good letter follows a predictable structure, and the job market paper is its center. The order below is the one I found to be most common.
1. The candidate as a researcher (1 to 2 paragraphs)
Start with how you know the student. Are you their advisor, did you work with them on a project, or did you teach them in a PhD course? If you have a joint project, describe the student’s role in it.
Then describe the candidate as a researcher. Name their fields. Say what they are good at: whether they know the literature well, how mature they are as a researcher, and how they interact with their peers in seminars and reading groups. Keep this part research related.
Comparisons are the strongest statements you can make. “This is the best student I have ever supervised” is very powerful, provided you have supervised enough students for it to mean something. “This student is comparable to Donald Duck, who took an assistant professor position at Harvard” is even stronger, if you know such a student. If the candidate is not your top student, the letter should still be very positive. Comparisons to their cohort at the same university are also common. One rule from Hamermesh is worth keeping: do not compare the candidate to someone currently in the department the student is applying to (Hamermesh, 1993).
Placement statements are a two-edged sword. “I believe the student should be placed in a top 10 US department” puts your candidate at the very top. I have seen students with that statement end up at very good, not top-10 departments of course. But a department ranked lower might think the candidate is out of reach and might not invite them. “I recommend the student to all departments outside the top 20” is also a positive statement, and it tells lower-ranked departments that they have a real chance. Note that Hirtle and Kovner find such “top” recommendations matter for placement (NY Fed Staff Report 1129), so think about which departments your student actually targets.
2. The job market paper (at least 3 paragraphs)
The job market paper should take up most of the letter. Explain what the paper does, how it fits into the literature, and what its contribution is. PhD candidates have limited experience in selling their own work. Letter writers are often better at this, and the committee knows it.
Give your assessment of where the paper could be published. This also sends a signal. If the candidate is an economist, do not name a non-economics journal. Ideally, the paper can be published somewhere between the top 5 and a top field journal.
If the paper is coauthored, name the coauthors. Ideally, they are junior. If the coauthor is you, explain your role and the student’s role. Recruiters want to see that the candidate can work as an independent researcher.
3. Other papers (around 1 paragraph each)
Most candidates have around two papers besides the job market paper. Describe each briefly: the question, the approach, and how far along it is.
4. Presentations, teaching, and service
Mention whether the candidate has presented at conferences and how good they are at presenting. A paragraph on teaching and academic service is a plus, but not strictly necessary.
5. Closing
End with a short paragraph that summarizes your recommendation in one or two clear sentences.
Wording and formatting
US letters are extremely positive. A letter that reads as good in Europe might read as weak in the US. As one career advice blog puts it for international writers, American letters must be entirely, energetically, and overtly positive (The Professor Is In, 2016). This is not only a concern for candidates applying to the US. European departments also receive many applications from US programs, and from students whose letter writers were trained in the US. When some candidates come with enthusiastic US-style letters and others with sober European letters, the students with the sober letters are at a disadvantage.
Make the letter easy to skim. The best letters I have read put a few bold highlights in each paragraph. A committee member should see at a glance where you discuss the candidate as a researcher, where you discuss the job market paper, and where you discuss the other papers.
Open each paragraph with a topic sentence. The first sentence should say what the paragraph is about, for example “Her job market paper studies how tariffs on fertilizer affect food prices.” A reader who only reads the first sentence of each paragraph should still get the main message of that paragraph.
Sources
- Blattman, C. (2022). Everything you need to know about letters of recommendation. Blog post.
- Blickle, K. S. and Parlatore, C. (2025). Letters of Recommendation in the PhD Job Market: Lessons from Specialized Banks. Liberty Street Economics, Federal Reserve Bank of New York.
- Hamermesh, D. S. (1993). Professional Etiquette for the Mature Economist. American Economic Review Papers and Proceedings 83(2), 34–38.
- Hirtle, B. and Kovner, A. (2024). Demographic Differences in Letters of Recommendation for Economics Ph.D. Students. Federal Reserve Bank of New York Staff Report 1129. Blog summary on Liberty Street Economics.
- Kelsky, K. (2016). How to Write a Recommendation Letter. The Professor Is In.
- Slonimczyk, F. (2025). This Candidate is [MASK]: Prompt-based Sentiment Extraction and Reference Letters. arXiv 2410.16325.
- Tello-Trillo, S. (2020). Results from ASHEcon Survey on Letters of Recommendation: Part 1. ASHEcon Newsletter.