About

Dr. David Reinstein · PhD in Economics, University of California, Berkeley

I’m an economist with 25 years of experience at UC Berkeley, the University of Essex, the University of Exeter, Rethink Priorities, and in other consulting and teaching roles. I founded and co-direct The Unjournal, a nonprofit (over $650k in grants) that commissions public expert evaluation and rating of impactful research.

Previous positions: Senior Economist, Rethink Priorities (2021–2023); Visiting Assistant Professor, University of Maastricht (2021–2024); Senior Lecturer, University of Exeter (2016–2021); Lecturer, University of Essex (2006–2015).

What I work on

Research evaluation and open science: how research gets evaluated, rated and used, mainly through The Unjournal, and more recently what AI models can and can’t do here.
Effective charitable giving and other-regarding behavior: what motivates it (see this synthesis), particularly the impact of effectiveness information.
Applied microeconomics and social science: social, psychological and institutional influences on life choices and consumer behavior.
Tools and policies aiming to improve social outcomes, and measuring whether they do.

I combine microeconomic modeling, statistical analysis of observational data, and field experiments and trials. I’m interested in causal inference, meta-analysis, predictive machine learning, Bayesian approaches and Monte Carlo policy simulations.

I’ve published in journals such as the Journal of Public Economics, the Journal of Economics and Management Strategy, and Social Psychological and Personality Science. Recent work and selected publications are on the Research page; the full list is in my CV.

The new way

Peer review is great (and peer evaluation is better), but in the 2020s we can do better than paying Elsevier & co to “publish” static PDFs for us. I’m working to research, write and get critical feedback in dynamic, collaborative, open and continuously updated formats.

That idea became The Unjournal, which I founded and co-direct. We’ve now published 55+ public evaluation packages, and our Pivotal Questions work connects research to specific decisions funders face. More on The Unjournal.

Earlier work in this mode

Mostly from before The Unjournal. For current work, see Research and Projects.

A broad project of synthesis and meta-analysis: Increasing effective charitable giving: The puzzle, what we know, what we need to know next (see the project page)
Reporting on an experimental research project: Impact of impact treatments on giving: field experiments and synthesis (listed under research and giving on my projects page)
Teaching and research ‘web books’ (also under teaching and methods notes): Researching and writing for Economics students, Microeconomics (MSc teaching), and Statistics, econometrics, experiment and survey methods, data science: notes

Tools, apps and side projects

I build small web tools and experiments, many with AI coding agents: Blues Flow for studying the 12-bar blues, the Word Bocce word game, the Math Estimator games for kids, a Library of Things finder, an AI jazz composition trial, and research tools like the Impact Products Directory. All of them, with screenshots. The Unjournal’s own tools are on the Unjournal tools page.

Other projects and affiliations

Ironically, I’m “humbled” by the honor of “winning” grants (so humbled that I feel compelled to announce it on my CV here).

Effective Altruism Market Testing team (EAMT): public ‘gitbook’ action and reporting space. I directed this team. It is no longer running trials, but I keep the resources up and am still involved in research and discussion in this area.
Earlier impact project: Innovations in Fundraising (archived copy)
Older blog posts (2010–2022)

Colophon

This website is built with Hugo and the blogdown R package. I adapted it by hand in 2021–22 from a custom theme by my colleague Willem Sleegers (itself inspired by Three.js Journey). Earlier versions are on the Internet Archive.

How this site uses AI

Since late 2025 I’ve updated this site mainly with an AI coding agent, Claude Code (Anthropic). Each page ends with a short “AI use” note saying what applies to it. In general:

Old blog posts (2010–2022) and the original pages were written by me, before AI writing tools. The agent has only fixed dead links in them.
In October 2026 I asked the agent to redesign the site and bring it up to date. It drafted the new home and Research pages, rewrote The Unjournal page from my earlier text, tightened this page, and updated the CV. AI agents searched the web and my own files for recent papers, talks and posts, and checked each link. I set the direction, supplied facts and figures, and revised the result over several rounds of feedback before anything went live.
The research, evaluations and opinions described here are my own and my co-authors'. Many of the tools and side projects on my projects page were built with AI coding agents; those produced largely by the AI are labelled as such.
The design and code changes (layout, styling, fixes) since late 2025 were made by the agent at my direction.

If you spot something wrong, please let me know (contact details are on my CV).

AI use: Originally written by me (2022). In October 2026 I had Claude Code shorten and reorganize it; most of the wording is still mine, and I reviewed the changes. How this site uses AI