A conversation with Lant Pritchett on June 18, 2012


Linear vs. Transformative Philanthropy

Prof. Pritchett made a distinction between "linear philanthropy" and "transformative philanthropy" as follows:

"Linear philanthropy” is scaling up an intervention with demonstrated impact, which has identifiable beneficiaries, and which does good on average. It’s attractive because

Funding deworming and increasing school enrollment are examples of linear philanthropy.

“Transformative philanthropy” involves trying to be a part of a transformative movement, which eventually mobilizes others, so that early contributions to the movement are highly leveraged. This involves taking actions that have an incremental impact today, but which potentially accumulate in a non-linear way.

The civil rights movement is an example of transformative philanthropy. In 1959 it looked as though changing people’s views on civil rights would be hopeless, but looking back we can see that if a philanthropist in 1959 played even a small role in creating the civil rights movement, the benefits would be incalculable.

Another example of transformative philanthropy is related to India’s recovery from its economic crisis of 1991. Other countries had previously had similar crises and failed to implement good policies that would have allowed them to recover from their crises. By way of contrast, India implemented good policies and recovered in a short time frame. Most of the key actors who ensured that India implemented the policies that it did were influenced by a think tank established by the Ford Foundation ten years before the crisis. The think tank exposed Indians to relevant ideas from the developed world about liberalization. The difference between (a) India’s upward economic trajectory and (b) what its upward economic trajectory would have been if it had been unsuccessful in recovering from the 1991 crisis is in the trillions of dollars. As such, the Ford Foundation’s investment in the think tank had a huge impact. For the ten years preceding the crisis, it looked like the think tank was having no impact, but it turned out to have a huge impact.

Prof. Pritchett worked on the Copenhagen Consensus, but was ambivalent about the project because the project had a high focus on linear philanthropy to the exclusion of transformative philanthropy. He believes that current fashions in development economics are overly focused on immediate, tangible results to the exclusion of transformative philanthropy projects that could have a much greater impact. He believes that linear philanthropy does have merits, and that major philanthropists should invest in both linear and transformative philanthropy.

Prof. Pritchett believes that philanthropists engaging in transformative philanthropy should work on causes which they’re passionate about, committed to, and willing to work on for many years. This is because successful transformative philanthropy requires active participation and because having an effect with transformative philanthropy requires a long sustained effort.

Prof. Pritchett recommended having one core transformative philanthropic focus. He raised doubt about the possibility of an organization having multiple transformative agendas. Holden said that pursuing multiple transformative agendas takes more staff capacity and money than pursuing one does, but that he doesn’t think that doing so is infeasible for a philanthropist of the right scale. He mentioned that most large foundations (and in particular the Hewlett Foundation) have multiple transformative agendas. Prof. Pritchett said that it’s trickier to have multipurpose organizations than single purpose organizations, that there’s often little overlap between transformative foci, and that pushing forward even a single transformative agenda requires very large scale and a long timeframe.

Prof. Pritchett gave three examples of areas that he believes are good targets for transformative philanthropy: increasing labor mobility, opportunity capitalism, and improving learning in schools in the developing world.

Increasing labor mobility

Prof. Prichett made the following case:

The easiest way to increase a poor person’s wealth is to let him or her move to a rich country. Most poor people have low productivity because of the environment that they inhabit rather than because they have intrinsically low productivity. Because of this, they can make much more money if they move to a better economic environment.

A woman from Bangladesh could make more money working for three months at a ski resort than all of the microcredit that she would have access to over her lifetime in Bangladesh. Local interventions designed to help poor people in a country like Bangladesh tend to raise their incomes by amounts in the neighborhood of 15%, whereas allowing them to work in a developed country tends to raise their income by amounts in the neighborhood of 1000%. There is evidence for this from randomized controlled trials.

Because the difference in earning power across countries is so large, the potential gains from liberalizing immigration are in the trillions of dollars. However, the political obstacles to liberalizing immigration, e.g. in the US, are enormous because voters in the US are opposed to immigration. Liberalizing immigration would require a long-term transformative effort to change a lot of parties’ minds, analogous to the civil rights movement.

Holden raised the point that there are already powerful entities promoting immigration for self-interested reasons, whereas in the historical examples of transformative philanthropy that Prof. Pritchett mentioned, there weren’t pre-existing powerful entities promoting the relevant change. Holden said that this is evidence that liberalizing immigration may be less tractable than the historical examples that Prof. Pritchett mentioned. Prof. Pritchett said that this could be the case, but that he thinks that the efforts of self-interested groups to promote immigration are unlikely to work because they are transparently self-interested. Holden raised the counterargument that the self-interested parties are able to fund academics to write about the humanitarian case for the views that they’re trying to promote. Prof. Pritchett said that though the cause may be intractable, he thinks that it’s not, raising the point that until the most recent century, labor mobility was the norm. Prof. Pritchett mentioned that he wrote a book called “Let Their People Come,” in which he tried to make a humanitarian case for immigration.

The missing middle

According to Prof. Prichett:

The United States has many medium sized firms whereas poor countries tend to have many small firms and many large firms but few medium sized firms.In developing countries, children don’t have the foundational skills of reading and basic math that they need to build job skills. Knowing the Pythagorean theorem, linear relationships, and ratios might be relevant to being a carpenter, and children don’t emerge from schools learning these things.

The situation seems to arise from poor countries’ corporate regulatory systems. Poor countries are generally not able to enforce their regulatory standards. Prof. Pritchett has coauthored a paper titled “Deals vs. Rules” providing empirical data that what firms report to doing doesn’t match up with the laws where they work. Regulatory standards are enforced asymmetrically: large firms can afford to buy their way out of them and small firms are too small for the government to keep track of. As a result, medium firms end up subject to more regulations than small firms or large firms, and thus are at a disadvantage relative to them. Furthermore, large firms are able to use their influence to block their smaller competitors from growing.

Prof. Pritchett sees an opportunity for transformative philanthropy in trying to change the regulatory environments in poor countries, so that successful small firms have the opportunity to expand to medium firms, and so that medium firms can prosper. This would be a hard battle to fight because both large firms (which benefit from the current policies) and the left wing are in favor of the current regulations. The payoff of success would be in the trillions of dollars. Prof. Pritchett hasn’t thought about the details of how one might try to change this situation himself on a practical level yet, in part because it is hard and in part because he hasn’t perceived much interest in the area.

A focus on medium firms would contrast with the currently fashionable focus on microfinance, which is oriented toward single person businesses. Pritchett believes that there should be more focus not on businesses with one or two people, but on helping high-potential small businesses scale from e.g. having two employees to having 200 employees. If small firms scaled like this more often, there would be more higher paying jobs, and poverty reduction is mostly about people getting higher paying jobs.

Holden asked whether Prof. Pritchett has made a pitch for this cause to the Gates Foundation or the Hewlett Foundation, remarking that the Hewlett Foundation seems more interested in transformative philanthropy than linear philanthropy. Prof. Pritchett said that he has spoken to Hewlett about some issues but never really been asked to engage by Gates.  His general (if unfounded) impression is that they have some aversion to getting involved with capitalism — they feel safer focusing on delivering goods and services.  It seems to him The Hewlett Foundation and Gates Foundation are definitely on board with transformative ways of delivering services, but transformative efforts to get economies operating more efficiently are much harder for philanthropy to engage in, even when they want to based on his own experience with Google.org.

Improving learning in schools in the developing world

According to Prof. Prichett:

One of the Millennium Development Goals is getting all children to complete primary school. Lots of groups, including the Hewlett and Gates Foundations have been making progress on this; it’s easy linear philanthropy. Things like building schools and ensuring that each school has a girls’ bathroom are relatively straightforward.

A big problem is that though most kids have some schooling, the quality of learning in the schools is very poor. The Hewlett Foundation (among others) supported a basic skills test by Pratham/ASER, which found that 50% of fifth grade children can’t read. On PISA, a global standardized test, the mean score for US students was 500 and the standard deviation for US students was 80. By contrast, the mean score for Indian students (in two states) was around 300. This means that the average Indian student’s performance is in the second percentile of US students’. Thirty percent of Pakistani 6th graders can’t divide a three-digit number by a one-digit number (data from LEAPS).

In developing countries, children don’t have the foundational skills of reading and basic math that they need to build job skills. Knowing the Pythagorean theorem, linear relationships, and ratios might be relevant to being a carpenter, and children don’t emerge from schools learning these things.

Prof. Pritchett sees a need to assess how much children in the developing world are learning, whether they’re acquiring the skills that they need, if they’re not then why they’re not, and what we can do to improve things. Schooling is one of the most successful movements of our time, and improving its quality in the developing world is important. This is a hard problem and Prof. Pritchett believes that solving it will require a transformative movement.

Cari said that her impression is that there’s a movement around this subject that’s picking up steam. Prof. Pritchett that it is picking up steam, but from a very low base. He said that the number of people who are talking about shifting focus from school enrollment to learning in the developing world (as opposed to “quality” which is much more broadly defined) could fit into a single room. Prof. Pritchett has talked about the topic with the Hewlett Foundation. Cari said that her impression is that Hewlett’s global-education team is increasingly focused on learning rather than enrollment.

Prof. Pritchett continued by saying:

The availability of schooling doesn’t seem to be the key constraint on economic prosperity. Today, the adult population of Haiti has more schooling than the adult population of France did in 1975, but France was very prosperous in 1975 and Haiti is very poor today. Part of the problem here is that Haitian children are not learning much in school and part of the problem is that the Haitian economy is not set up to absorb the skills that schools teach.

Though correlational studies show a relationship between education and income, it seems that school enrollment hasn’t caused economic growth. Schooling has been expanded almost everywhere, but there hasn’t been broad-based economic growth almost everywhere. In some places, planned economies gave rise to a spurious correlation between education and income. For example, the 1970’s getting a government job in Africa and in Egypt required a schooling degree, and the government jobs paid well. Eighty percent of people with a junior high school degree or above were working for the government. This gave rise to a correlation between amount of schooling and income.


Holden described GiveWell’s recent interest in meta-research and asked Prof. Pritchett about ways in which academic research is failing to maximize its value to society and what can be done about them.

Prof. Pritchett is writing a paper on this topic, which he will send to GiveWell. He summarized his view as follows:

The real world operates at a finer level of granularity than the research world. The devil is in the details. The research world claims that conditional cash transfers (CCTs) are a proven intervention, but CCTs are highly variable; one might work and another might not. Some relevant variables are: (i) the amount of money delivered per child (ii) whether the cash transfers should go to the child, the child’s father or the child’s mother (iii) the frequency with which cash transferred are delivered and (iv) what the precise criteria are for delivering cash transfers. Academics haven’t engaged in discussion about design details of cash transfer programs.

Parts of the “randomized control trial” approach fits well with the “planner” approach to development (in Bill Easterly’s terminology). It doesn’t fit as well with searching for solutions to problems when we’re not close to finding a solution and don’t have agencies that are capable of implementing the interventions that we find. Prof. Pritchett has coauthored a paper called “Capability Traps.” India just isn’t capable of implementing a lot of interventions so knowing which ones that would work if they could be implemented but cannot be implemented just doesn’t seem like huge progress. It’s important to think not only about what would have a big impact if implemented, but also how to create agencies in India capable of implementing a lot of things.

Prof. Pritchett has a negative view of the randomista movement in that he believes that the potential gains have been exaggerated. According to Prof. Pritchett:

There was a wave of randomized evaluations about social policy in the United States in the 1970’s and that after ten years it collapsed because the practitioners didn’t think they were helping.  Studies of policy are embedded in a political context and hence claims about the impact of studies should be based on a positive political economy of policy formulation and implementation—and we are far from that.  It is not at all obvious that studies will have more impact on policy because they are “randomized” versus being more embedded in organizations and movements.

Randomization looks scientific, but more scientific methods only lead to more scientific outcomes if the entities being studied are the kinds of entities for which the empirical appoach of experimentation is appropriate (e.g. homogeneous or with the right kind of invariance laws). Learning about electrons doesn’t work the same way as learning about human beings and organizations. Electrons in Nepal are very similar to electrons in Alameda County but teachers in Nepal are not very similar to teachers in Alameda County.