So far, the effective altruist strategy for global poverty has followed a high-certainty, low-reward approach. GiveWell mostly looks at charities with a strong evidence base, such as bednets and cash transfers. But there's also a low-certainty, high-reward approach: promote catch-up economic growth. Poverty is strongly correlated with economic development (urbanization, industrialization, etc), so encouraging development would have large effects on poverty. Whereas cash transfers have a large probability of a small effect, promoting growth has a small probability of a large effect. (In general, we should diversify across high- and low-risk strategies.) In short, can we do “hits-based development”?
How can we affect growth? Tractability is the main problem for hits-based development, since GDP growth rates are notoriously difficult to change. However, there are a few promising options. One specific mechanism is to train developing-country economists, who then work in government and influence policy in a pro-growth direction, ultimately increasing the probability of a growth episode. The key is that local experts have local knowledge of culture, politics, and law, which allows them to understand the impediments to growth in a way that foreign World Bank consultants cannot. The goal is not to find the specific interventions that will boost growth in a specific country, but rather to find the experts who will be most capable of finding those interventions.
For example, Lant Pritchett mentions a think tank in India that influenced its liberalizing reforms, which preceded a large growth episode. One way to think of the mechanism is having an economist "in the room" at a key moment to prevent the president from enacting a policy that would lead to hyperinflation (say). Since such an action is highly consequential, but also unlikely, it plausibly has high expected value. A different angle is that local experts will improve the quality of solutions to important problems (urbanization, industrialization, health care, etc). This translates into a concrete goal: increase the number of local experts in developing countries, to increase the chance of a growth episode.
In terms of crowdedness, there are, for example, very few Africans doing economics PhDs in the US. Some developing countries (like China) already have many economists, and do not need to be targeted. One feature of this proposal is that, in principle, it has a clear stopping point: train developing-country economists until you reach N per country.
There are many ways that extra funding could relax budget constraints: GRE fees; PhD application fees; scholarships for undergrads, masters, and PhD students; research projects; thesis prizes; subsidizing RAships; TA buyouts; conference fees and travel; think tanks; translating textbooks; bootcamps/workshops/conferences; sabbaticals/research visits; and so on.
From initial conversations, it seems that the key constraint is the number of well-trained students who can get accepted into top PhD programs. Hence, funding should be targeted to training undergrads and Master's students, as the African School of Economics is doing. Hence, one option is to fund ASE and its satellite campuses. Donations need not be restricted to scholarships, for two reasons. First, because of the fungibility problem, such restrictions are difficult to guarantee. Second, ASE's values are consistent with EA values. This matches GiveWell's advice to "find an organization whose existing priorities you are comfortable with – and give unrestricted."
 See recent discussions on the forum here and here.
 Wantchekon says that his knowledge of culture and history led him to the research question of the effect of slavery on trust: "It was his data-mining skills that helped him find the answer. But it was his Beninese background that raised the question."
 See also: "I think you invest in people in those countries who are researching, acting, and investigating, and talking about precisely those questions [about property rights, corruption, democracy, transparency]. I think investing $15 million in a group in Nigeria who was asking in an evidence- and experience-based way how to reduce corruption in Nigeria—I think that seems like a pretty plausible way to get those questions answered."
 On this view, it's worth training a generation of economists, with only one becoming a presidential advisor, in order to prevent a single hyperinflation.
 Note that ASE is partly funded by Princeton and the World Bank.