Lee Crawfurd

The Iran War Oil Shock Will Hit the Hungry Hardest—Cash Should Be Part of the Response (new @CGDev blog with Eeshani Kandpal)

Can test results lead to real education reform? @Alenestour and @leecrawfurd show how PASEC results have shaped policy and improved learning in Francophone Africa. At a few cents per child, can countries afford to lose it? Also available in French: cgdev.org/blog/pisa-shoc…

An astonishing rise in the percentage of Chinese authors in the top 5% of scientific journals. nber.org/papers/w34694 nber.org/digest?page=1&…

@bswud I can see the appeal, we're mostly not at the frontier. But the price is rhetorical justification for foreign aid cuts that will cost lives because "we're poor too" when its just not on remotely the same scale.

I think it's a helpful framing device, but I can see why someone who mainly worked on developing countries would object to it. It would be like me in 1700 saying 'the Netherlands is a developing country', when in fact it's still going to be the richest country in Europe for another 150 years. It's just a sclerotic rentier country, declining relatively, but from a very high base.

'The War on Prices' is partly related to the fact that we've just had the biggest inflation wave we've seen in decades, and there is no indication that the fiscal/monetary authorities intend to have inflation go below target for a while to balance it out. But it's also a sign of reduced state capacity. Even relatively primitive states, like middle-income countries and the UK of the 1500s, could fix prices. It's a classic tool that countries reach for when they aren't able to pull off the sorts of projects that would cut the underlying actual costs. The magic wand.

@bswud I don't disagree with the specifics here but the "Britain is a developing country" line is misleading and minimises the actual large economic & political differences between Britain and many poorer countries

Britain needs to reckon with the fact that it is not just a developing country economically, but also in political culture. - There are widespread perceptions of corruption. - Governments are not capable of pulling off large projects without enormous consent generation schemes through established interest groups (NGOs) and massive side payments - The government is not trusted with expropriation tools - Even left-wing governments cannot raise broad-based taxes (and the far left opposition don’t make the case for broad-based taxes, but that they can extract loads of money from the rich and other scapegoats) - Even right-wing governments cannot take away state welfare entitlements - There is a dizzying array of inconsistent privileges - Parties are becoming less ideological and more tribal. There are explicit ethnoreligious parties standing, and bloc voting is becoming more and more common. - Everyone thinks that all politicians are liars - Almost no one is willing to take a hit in the interests of the country, and no one is expected to. I think all of these things are connected, and I also think it’s foolish and self defeating to pretend we are Britain of the 1950s, or Denmark, and that we can simply implement the most efficient policies by deciding to — we just need more political will! Instead, we need to be realistic about what a country in our situation can achieve. We need to come up with ways to steadily build state legitimacy and state capacity, by stigmatising dishonesty and using the tools of the past, which worked when we were last in this situation. The government can’t be trusted to spend money, so taxes need to be hypothecated to things voters want if we want to raise more. Large projects need to involve more specific deals with losers, overriding objectors with local support not (nonexistent) national fiat. Anything controversial needs to be approved in a party’s manifesto, or in a referendum. If it can’t be voted through, it cannot be implemented. If the voter doesn’t want it, they need to be convinced, or it can’t be done.

Child penalties in women's employment tend to be smaller in developing countries - but that doesn't mean equality! It means different drivers in the gap Interesting analysis from @s_anukriti

UK aid was at 0.5% vs. a government target of 0.7%. But instead it's being cut to 0.3%. UK aid was unusually focused on cost-effective interventions, so that cut of ~£6bn per year, which could cost hundreds of thousands of lives.

New in @AEA_Journals' American Economic Review: migration doesn't hollow out the home economy — it builds it up. More than 75% of the long-run income gains from migration are domestic. The home economy itself grows. Here's what we found: 🧵

I created brief summaries of every paper presented at the recent #OxCSAE2026 conference. Please check it out! And share widely. The width and depth of research is eally rincredible. lnkd.in/eKDtb23Q Blog is inspired by @DaveEvansPhD's blogs, and powered by #ClaudeCode

@Research_FRI @EzraKarger @Jabaluck @Afinetheorem @BasilHalperin @toddrjones @connacher_ @pawtrammell @mattsreynolds1 @danmayland @PTetlock Do you see more optimistic views for people who took the survey later (eg Feb 26 vs Oct 25)? I think we're already at the Slow Progress scenario now, so surprised how likely people think that is as the outcome for 2030.

Thank you to all of our coauthors: @EzraKarger, Otto Kuusela, @Jabaluck, @Afinetheorem, @BasilHalperin, @toddrjones, @connacher_, @pawtrammell, @mattsreynolds1, @danmayland, Ria Viswanathan, Ananaya Mittal, Rebecca Ceppas de Castro, Josh Rosenberg, and @PTetlock For full results, including the impact on sectors and occupations, read our full paper here: forecastingresearch.org/s/forecasting-… For a summary of the results, see our policy briefing here: forecastingresearch.org/s/forecasting-… And visit our Substack: forecastingresearch.substack.com/forecasting-th…

We completed the most comprehensive study of how economists and AI experts think AI will affect the U.S. economy. They predict major AI progress—but no dramatic break from economic trends: GDP growth rates similar to today's and a moderate decline in labor force participation. However, when asked to consider what would happen in a world with extremely rapid progress in AI capabilities by 2030, they predict significant economic impacts by 2050: • Annualized GDP growth of 3.5% (compared to 2.4% in 2025) • A labor force participation rate of 55% (roughly 10 million fewer jobs) • 80% of wealth held by the top 10% (highest since 1939) 🧵 Here's what we found:

@Jabaluck @EzraKarger Surprised how likely people rate the slow progress scenario, aren't we already there right now?

Incidentally, I did almost none of the work on this paper beyond sending a few emails with my thoughts, so kudos to @EzraKarger and team!

I was surprised by the results of this study: I thought AI researchers would forecast faster AI progress than economists, but there was little difference. Most differences are about forecasted economic impacts for given capabilities, and more within group than between group.

Last Week @Alenestour and I laid out the case for international learning assessments in Africa. This week @Alenestour & @barbarabruns lay out how PASEC and others can do even better cgdev.org/blog/nine-cent…

@Nerland87 Labor for reconstruction and population for geopolitical standing?

@leecrawfurd I'm mildly optimistic that this will increase demand for labor in medium term and increase desire to have a larger population.

"Flights from Nepal [to the Gulf] have virtually disappeared since the United States and Israel began their war on Iran on Feb. 28. ... In Nepal, remittances contributed roughly 26 percent of its GDP in 2024" archive.is/20260330170903…

@jpmvbastos Nice thread! To try and steelman Hickel for a sec, there are a few cites to papers that use UNSC membership as an instrument for IMF programs. I'd be curious to read some interrogation of that design as I have some doubts about the strength and excludability of that instrument.

I'm not going over all of them, but if this took me 30 minutes to find, you can figure out what a full audit would show. Btw, I'm happy to be proven wrong on the other citations -- if you find something that *is* causal, let me know.

I took the time to look at @jasonhickel's study. Despite the "caused" on the first page, they say they are not making causal claims themselves; only some references have causal arguments.

In-person training > app-based training