John Loeser

1/N 🚨🚨🚨 New WP!!! 🚨🚨🚨 (ungated documents1.worldbank.org/curated/en/404…) w/ @fkondylis we ask: do we need a “big push” to lift households out of poverty? we find: leaner cash based programs reliably reduce poverty more cost effectively than more expensive (but also more impactful) programs

🆕 How Rwanda used industrial policy to slow the spread of COVID-19 Today on VoxDev, @byrne_kecon (@qmuleconomics), Florence Kondylis (@wb_research), @loeserjohn & Denis Mukama (@rrainfo) discuss licensing textile manufacturers to produce masks: voxdev.org/topic/macroeco…

1/ Happy to release StatsClaw — an open-source multi-agent workflow for building statistical software with AI. w/ @Maple_Optboy Site: statsclaw.ai Paper: bit.ly/statsclaw

Are you interested in Tax research? This is your chance to submit your paper and win pleasant cash prizes. For more information regarding this call for papers, click on this link rra.gov.rw/fileadmin/Fold…

🚨New in AER🚨: "Negative Control Falsification Tests for IV Designs" (@DanieliOren, @DanielNevo, Itai Walk, Bar Weinstein, @danzeltzer). doi.org/10.1257/aer.20… Common placebo test implementations may reject valid IVs. We're coming to save your IVs! 🧵

ARE research - Irrigation is central to sub-Saharan food security and climate resilience are.berkeley.edu/news/irrigatio…

@jt_kerwin @jmwooldridge after - before

@jmwooldridge Are there any causal inference methods that “allow” for selection into the treatment? It seems like they all find ways of ruling it out (or give biased estimates)

I regularly get a common/question about how diff-in-diffs isn't that useful because it doesn't allow "selection" into treatment. Of course it does, but the nature of selection is restricted. One way to interpret the usual parallel trends assumption is as "stable selection bias."

Call for papers: UC Berkeley conference honoring Elisabeth Sadoulet — October 17, 2026. Abstract submission deadline: March 15. Conference papers will be invited to submit to a special issue of the Journal of Development Economics in her honor. are.berkeley.edu/news/announcem…

Here's the prompt: # Starter Prompt for Claude Code Copy and paste everything below into Claude Code to begin the project: --- ## CRITICAL WORKFLOW REQUIREMENTS **YOU MUST FOLLOW THESE RULES. THEY ARE NON-NEGOTIABLE.** 1. **NEVER claim something works without running a test to prove it.** After writing any code, immediately write and run a test. If you cannot test it, say so explicitly. 2. **Work modularly.** Complete one module at a time. After each module, report what you built, show test results, and wait for confirmation before proceeding. 3. **Iterate and fix errors yourself.** Do not rely on the user to report errors back to you. Run the code, observe the output, and fix problems before presenting results. 4. **Be explicit about unknowns.** If you're uncertain about something, say so. Don't guess. 5. **Use python3 and pip3.** Always use `python3` (not `python`) and `pip3` (not `pip`) for all commands. --- ## PROJECT INSTRUCTIONS Read `INSTRUCTIONS.md` in this directory. It contains detailed guidance for replicating and extending a published political science paper on vote-by-mail. **Your task**: Follow the instructions phase by phase, starting with Phase 0 (Project Setup). **Critical rule**: At each checkpoint marked with 🛑, you must: 1. Summarize what you have completed 2. Present key outputs for review 3. List any issues or concerns 4. **STOP and wait for my explicit approval before proceeding to the next phase** Do not skip checkpoints. Do not proceed past a 🛑 without my approval. **Begin now with Phase 0.** # end of prompt And here's the instructions file: github.com/andybhall/vbm-…

My fellow farmers and I are excited to release this preview of our full-length feature to be released next fall

We have a new working paper: Using national slum polygons, satellite imagery, and SDiD, we compare in-situ upgrading vs relocation in Chile. Upgrading improves housing and nearby areas—at lower cost nber.org/papers/w34560 w/ @paul_gertler , Raimundo Undurraga and Joaquin Urrego

Who wins when public transit challenges private transit? Last week on VoxDev, @danbjork @Columbia_CDEP, @Alice_Duh @WorldBank, @nagpal_geetika @wb_research & @NTsivanidis @BerkeleyHaas @berkeleyecon explored the impacts of new public buses in Lagos: voxdev.org/topic/infrastr…

Remembering Betty Sadoulet: 1945-2025 CEGA is saddened by the passing of faculty affiliate Betty Sadoulet. Betty leaves behind a remarkable legacy in the field of international economic development and will be deeply missed. Learn more about her work: go.cega.org/bettysadoulet

Glad this paper with @guido_imbens is out in the JEP. The LaLonde paper has had a big impact on my academic journey and continues to teach us about the challenges and possibilities of conducting credible inference using nonexperimental data. aeaweb.org/articles?id=10… Many thanks to Tim for the tireless editing and suggestions. @TimothyTTaylor

@KhoaVuUmn documents.worldbank.org/en/publication…

New IV alert 🚨🚨

Energy-efficient biomass cookstoves cut fuel use and, in Rwanda, do not trigger significant local rebound effects – consumption, fuel collection time, and prices remain largely unchanged for those not adopting cleaner cookstoves. voxdev.org/topic/energy-e…

Very happy that this new paper "Irrigation infrastructure and satellite-measured land cultivation impacts: Evidence from the Senegal river valley" is now forthcoming at Journal of Development Economics! sciencedirect.com/science/articl…

Only a few days left before the deadline. Don't miss the chance to interact with researchers and policymakers doing impactful research!

Partnering with business for development economics research 📢 Last week on #VoxDevTalks, @fannan2316 (@AreBerkeley) discussed how academics can partner with the private sector to conduct research and have impact: voxdev.org/topic/finance/…

@borusyak @instrumenthull i think same proof would work if also (eps1, Y) perp Z given W. not sure what that means but also not sure what completeness really means…

@loeserjohn @instrumenthull Thanks. Given the context, we are happy with Y(W,eps2) but we'd like to allow X(Z,W,Y,eps1)

Help us out, math nerds! math.stackexchange.com/questions/5086…