IPA’s weekly links

Guest post by Jeff Mosenkis of Innovations for Poverty Action.


  • The summer ape blockbuster you’re been waiting for is here. In Science economists Seema Jayachandran, and Joost de Laat team up with  satellite researchers Eric Lambin, Charlotte Stanton, Robin Audy, and Nancy Thomas (with some help from IPA and Uganda’s Chimpanzee Sanctuary and Wildlife Conservation Trust). They ran the first RCT showing that just paying farmers in Uganda a little bit not to cut down forest on their land where endangered chimps live cut deforestation in half.
    • But the real policy importance comes from their cost-effectiveness analysis. Each village kept about 3000 tons of CO2 out of the atmosphere during the 2-year study, and each of those tons cost only about 46 cents. This is something like 10-50 times cheaper than achieving similar results through energy technology subsidies in the U.S.
    • More in the New York Times or Atlantic, and this thoughtful thread on the topic from an expert (h/t David Evans). You can call in and ask Seema about it yourself on Public Radio’s Science Friday today sometime between 2-2:30 Eastern (stream from wnyc.org)


  • Some of the experts quoted in the stories above note that what really makes the study powerful is that they didn’t stop at the cutting deforestation in half part, but also ran the numbers to find out how cheap it was. Turns out this is pretty rare, but really important to people who need to actually implement policies. In education, for example, a cognitive behavioral therapy intervention had about the same effect as remedial games, but one cost $15 per student, the other cost $4,400 per student (albeit in different countries). Most research papers leave the price tag out, or as Caitlin Tulloch puts it on the Life of The Mind podcast:

It is perpetually amusing to me that economists are uniquely bad at talking about how much things cost.


  • In that podcast she talks eloquently and compellingly about leading the group that is systematically doing cost-effective analysis across all of the International Rescue Committee’s programs. Particularly interesting about doing it for such a large org is with lots of work in different locations, she begins to see what looks like a mystery for individual RCTs – that programs work differently and cost different amounts in different places. She suggests that when you begin to aggregate across individual evaluations, you start to see patterns that you don’t see in any one individual study and realize why seemingly conflicting individual results are really expressions of a larger systematic pattern.
  • Kids in Cameroon do muuuuch better than German kids on the marshmallow test. Researchers speculate b/c parenting styles of the Nso ethnic group there emphasize emotional self-control and deference to parents.
  • Interesting thread from a labor economist on what job creation policies might look like in the U.S. if they focused on created jobs rather than training workers (h/t Sally Hudson)


The best tweet so far on the deforestation study:


IPA’s weekly links

Guest post by Jeff Mosenkis of Innovations for Poverty Action.

  • North Korea’s surprising, lucrative relationship with Africa (via Kim Yi Dionne)
  • In an inexplicable lapse some congressional staffer has surely been punished for, the House Foreign Affairs Committee invited three eminently qualified women to testify about women’s empowerment in the developing world. Even more encouraging was that the hearing was titled “Beyond Microfinance.” Mary Ellen Iskenderian, head of the financial inclusion org Women’s World Banking, Georgetown’s Melanne Verveer, the first U.S. Ambassador for Global Women’s Issues, and MIT economist Tavneet Suri all converged on a similar message about moving away from ineffective programs towards ones that have been shown to work. Watch the video and read their full written testimonies here.
  • What it’s like to run for office as a woman in Kenya.
  • Owen Ozier on three tricks for not letting preanalysis plans mess up your plans.
  • Don’t listen to the EconTalk interview with law prof Robin Feldman about the stunningly clever ways drug companies maintain brand name monopolies, even in the face of laws designed to limit them. Because if you do, you might need blood pressure medication, which probably is off patent but still not available in generic. But you will learn a lot about how monopolies can flourish even in systems designed to encourage competition.
  • NYTimes asks its Wirecutter site for its favorite luggage for frequent travelers (and a DIY version of “smart luggage” with a built-in finder). Also, beware of people checking live alligators as baggage (but maybe the only legit alligator skin luggage on the carousel).
  • Long read on Philip Morris’ secret campaign to undermine anti-smoking rules in poor countries. But congrats to global health expert Amanda Glassman, whose research turned up in their power point presentations. (We can only hope one day the International Criminal Court brings their power point designers to justice).
  • This was mind-blowing:


IPA’s weekly links

Guest post by Jeff Mosenkis of Innovations for Poverty Action.

Maintaining attention in an interview can be hard

Above: Maintaining attention for a long interview is always a challenge


  • Fake news is already disrupting Kenya’s election.
  • Qualtrics, the research software company, did a randomized experiment testing the kinds of extra questions researchers embed in surveys to make sure respondents are paying attention and answering thoughtfully. They found including those questions earlier in a survey actually led to respondents performing worse on later questions. They now recommend against including those kinds of questions. (h/t Sanjay Srivastava)
  • It’s worth listening to the pair of EconTalk podcasts that arose out of Chris and Lant Pritchett’s back and forth on Chris’ open letter to Bill Gates suggesting he bet on cash rather than handing out chickens as aid.
    • Lant responded that the type of work that microeconomists do (particularly the crowd who does randomized controlled trials) pales in potential compared to the massive effects on every aspect of poverty that comes with the kind of large-scale economic growth that China and India have seen. He argues that everybody in development should be on “team economic growth.”
    • Chris was on this week with a great discussion about why he still believes these questions are important to answer. He argues that these kinds of studies do more than just answer a question at hand, they’ve really advanced our understanding of what keeps people in poverty. (It’s also worth listening to the brief discussion on what kind of research young researchers should be doing.)
  • Along those lines, it’s also worth checking out this tidbit from the very good interview of Tyler Cowen by Stripe CEO Patrick Collison (iTunes, SoundCloud, transcript). It followed a discussion questioning how much concrete understanding of the economy macroeconomics has brought us:

COLLISON: Yeah. What are the, say, top two most underinvested areas of economics today?

COWEN: Culture and economics, for me, is by far the most underinvested. I still think randomized control trials, they’re expensive, but you do actually learn things from them, which are probably true. That’s remarkable.


COWEN: They contain actual knowledge. Now, it’s true the questions you can ask are narrower, but it seems odd to turn down the reward of actual knowledge, right? [laughs]

COLLISON: You recently linked to the new book whose title is escaping me — you’ll probably remember it — on the series of interviews on random control trials in economics. And there’s all these questions about to what degree they have external validity and so on. Do you think the critics are overstating the case?

COWEN: One of the main criticisms is, if you do randomized control trials, you’re studying something like, “Well, does paying mothers to bring their children in for vaccines work in getting the mothers to bring the children in?” You’re not asking big-picture questions of political economy. But big-picture questions of political economy — they can be very hard to control. There’s no one who can steer, say, what will happen with India or Kenya, but you can change some policy regarding, “Do you reward mothers for bringing their children in for vaccinations?”

You know the subtitle of our blog, “Small Steps Toward a Much Better World”: there’s something to that. We can make a lot of these small steps. It’s also related to the correct attitude about management. A lot of good management is doing very small things and not always some grand philosophy. So I think this is actually still underrated.

But with Lant’s always insightful critiques including how we’re thinking about measurement completely wrong in education, we created this helpful predictor to know what he’ll be writing about next:


Everything You and Your Colleagues Have Ever Done is Useless in The Field Of:

IPA’s weekly links

Guest post by Jeff Mosenkis of Innovations for Poverty Action.


  • We don’t do enough thinking in the U.S., much less in developing countries, about end of life “palliative” care, helping people with difficult terminal illnesses suffer less. But a lot of suffering happens at the end of life; if your goal is to alleviate suffering, pain management is doable. The BBC has a very interesting story of the woman who singlehandedly brought palliative care to Mongolia.
  • David Evans summarizes a post and paper from Ben Piper at RTI doing cost-effectiveness analysis of technology programs to improve education. Often even when new technologies work, they’re more expensive and the same outcomes could have been achieved through traditional means.
  • Dropping laptops into Peruvian classrooms through the One Laptop Per Child program didn’t do much (surprise).
  • Don Green is teaching a 3-day advanced course in field methods at J-PAL in August.
  • Economist Jess Hoel offers advice for women academics on careers, including publishing, accountability groups, and networking. Full notes on her website, http://www.jessicabhoel.com/ (top right).
  • Reuters reports the President is about to sign an executive order requiring 100% of U.S. food aid to be transported on U.S. ships (up from the current 50%) to boost U.S. jobs. This might be a problem for anybody counting on that food because of low U.S.-based shipping capacity. The requirement would increase costs while introducing significant delays.
  • We already have a policy experiment along similar lines in the U.S.: the Jones Act requires cargo being shipped between ports in the U.S. to be carried on U.S. ships, and Planet Money documents how nearly everybody hates it. When New Jersey was ready to buy 40,000 tons of road salt from Maine during a harsh winter the only ship available to take it that qualified under the Jones Act was a small barge. So they did it by making lots of trips to transport it in small batches, passing by the big ship that was already in port. Joe Stiglitz explained that the Jones Act costs the U.S. a quarter million dollars per job saved.
    On the plus side, maybe jacking up the cost of food aid would move more aid to cash?

I’ll let this one speak for itself

I don’t know why economists have a bad rep – this from a neuroscientist who studies the trolley dilemma (people are generally willing to kill one person to save five by flipping a switch, but not by pushing the one person in front of the trolley to stop it).Trolley

The article goes on to describe a student project that found Buddhist monks also in that group. So, interesting intellectual bedfellows.

(image credit above)

IPA’s weekly links

Guest post by Jeff Mosenkis of Innovations for Poverty Action.


  • Three relatively recent additions to development Twitter worth noticing if you haven’t already, Tavneet Suri, Nava Ashraf, Seema Jayachandran. Here’s proof from just this week:
  • There’s a new website devoted to making development research easily accessible, VoxDev.org. Editor Tavneet Suri says:

Here is our vision: We want to bring cutting-edge research to the forefront of decision making – for policymakers, the private sector, NGOs, the civil society, and everyone in between. We believe that the role of research should be elevated to inform economic decisions at every level. VoxDev will be the place where rigorous research on international development moves into practice; we will translate the state-of-the-art into pragmatic policy advice.

  • From Nava Ashraf, video of Al Roth talking market design (version with slides here)
  • From Seema Jayachandran – parents’ preference for having boys over girls in surveys is tricky to measure because of a literally “odd” phenomenon. Most parents, even in places like India, want the same number of boys and girls. The preference appears when they’re considering odd numbers of children and have to make a choice. This makes standard questions about wanting boys vs. girls a bit misleading.
  • Outgoing WHO director Margaret Chan looks back on her tenure. She says the dirtiest fights in global health were with tobacco companies. She also worries about moves to make poor countries more self-sufficient in health because of risks beyond policymakers’ control:

In large parts of the developing world, especially in Africa, small-holder farmers in the informal sector remain the backbone of the economy, severely limiting domestic resources derived from taxes. Much of this agriculture depends on rain and is highly vulnerable to extreme weather events, which are becoming more common as the climate changes.

  • David Evans has a roundup of 2 days of great education papers at the RISE conference on improving education systems, including links to individual videos.
  • Speaking of education, my colleagues in Kenya are holding a conference and looking for presentations on research directly applicable to curriculum reform for Kenya (deadline July 7). The research doesn’t have to be from Kenya, but should be from somewhere similar enough in education to be applicable.

And incentives in everything – killer whales have started shaking down fishermen for their fish (via Chris). But there was an example in 1920’s Austrailia where orcas and humans cooperated in hunting for whales. For decades “The Law of the Tongue” meant that the orcas would help herd the whales for the humans, who would then harpoon the whales and share the carcass with the orcas (the orcas took the tongue). But like every good multiple round econ game, it worked great for both sides till one side got greedy (h/t my dad).

(Photo credit above)

IPA’s weekly links

Guest post by Jeff Mosenkis of Innovations for Poverty Action.


  • It’s a puzzle why more people don’t use long-acting reversible contraceptives like IUDs or hormone shots. Berk Ozler reports on qualitative findings from Cameroon about why adolescent girls don’t seem interested in them.
  • A new working paper suggests that how refugees fare economically in the U.S. is heavily dependent on at what age they arrive, perhaps because learning English is easier if they come earlier. The authors estimate over their first 20 years in the U.S. refugees pay about $21,000 more in taxes than they receive in benefits. (h/t Dylan Matthews)
  • J-PAL has a fellowship for quantitative-oriented Ph.D. students to work on data transparency from their home universities, the June 30 deadline is approaching:

Our fellowship program offers financial support (tuition assistance of up to $12,000 and a stipend of $13,000) for one semester (approximately 4.5 months) for current PhD students. While preference will be given to students from economics programs, graduate students from other disciplines with strong quantitative and programming skills (STATA, R etc.) are very welcome to apply. During the fellowship, students will work with our affiliated professors from dozens of universities around the world, re-analyzing RCTs from scratch (starting with the creation of the data set from “raw survey” data, up to production of the final econometric analysis included in the working papers).

  • The AEA’s chart of the week above, “The Resource Curse In Action,” overlays geocoded data of mines and areas of conflict, comes from this paper showing how conflicts in Africa rose around mines as commodity prices increased.  Also on mines:
    • Another paper finds sanctions in the Dodd-Frank act against conflict minerals in the DRC:

…increased the probability of infant deaths in villages near the policy-targeted mines by at least 143 percent. We find suggestive evidence that the legislation-induced boycott did so by reducing mothers’ consumption of infant health care goods and services.

  • I  recall (but can’t find now) a story suggesting that creating certified mineral providers whom companies could source from led to worse conditions for the workers because it created monopolies who could further exploit them.
  • A former illegal miner in Sierra Leone describes how he paid for his start in college with a portion of the proceeds from one single diamond he found, and is now a BBC reporter.
  • Angrist, Lavy, Leder-Luis, & Shany, re-examine “Maimonides Rule” (named for the Medieval Jewish philosopher), about the relationship between class size and student achievement. While it applied in Israeli class data from 1991, it does not appear in more recent data. (h/t Dina Pomeranz)

Though it turns out, Maimonides himself was railing against absent teachers (a current problem in many countries) back in 12th Century Egypt (h/t Josh Yuter)

Absent teachers Maimonides

IPA’s weekly links

Guest post by Jeff Mosenkis of Innovations for Poverty Action.

Agriculture in Kenya


I know this week for many of us it’s been hard to pay attention to what else has been going on in the world, what with the release of Stata 15 and all, but I’ll try to help with some stuff you may have missed:

  • The World Bank’s Hedy Sladovich &Emanuela Galasso put together several very accessibly written one-paragraph summaries of recent findings on what works in early childhood development.
    • It pairs nicely with this (mostly Western-based) review from 10 researchers on what works in pre-K. If you’re busy, skip to the last page for the consensus statement.  (Both via David Evans)
  • But if you’re more into land and agriculture, you might prefer Markus Goldstein, Niklas Buehren and Muthoni Ngatia’s one-sentence summaries of 45 recent conference presentations on African agriculture.
  • A bunch of big name authors find returns to nudges for governments are generally pretty high, though they vary (helpfully, the authors compare different types of nudges in the same category so you can see the variance). It’s important to know that while the effects aren’t usually huge, they’re worthwhile because they’re so cheap (such as switching a default option in an existing program or sending a differently worded message), as they note:

Because traditional interventions are intended to change behavior by altering the cost-benefit calculation that individuals undertake when focusing on a particular decision, these interventions face the challenge that individuals’ ability (and desire) to engage high-level cognitive capacities is often limited (Shah, Mullainathan, & Shafir, 2012).  Nudges, by contrast, can succeed because they account for individuals’ intuitions, emotions, and automatic decision-making processes. These processes can be triggered or enlisted with simple cues and subtle changes to the choice environment, so nudges can be effective yet cheap, generating high impact per dollar spent.

And AreMenTalkingTooMuch.com is a handy timer you can use in meetings:







IPA’s weekly links

Guest post by Jeff Mosenkis of Innovations for Poverty Action.


  • Why Poverty is Like a Disease” looks at epigenetics and how poverty interacts with one’s own biological development. It’s well written, by someone who describes what it was like growing up poor and how hard it was to escape. (h/t David Batcheck)
  • Our World in Data asks “How Much Will it Cost to Mitigate Climate Change?”
  • Some nice news via Dina Pomeranz: globally deaths from diarrheal disease fell by a third between 2005 and 2015. Researchers attribute much of the drop to the development of a vaccine, but also say good sanitation is key.
    • An interesting process story behind a current sanitation RCT; A design firm talks about the process for coming up with a handwashing station appropriate for use in rural Kenya. They describe how they couldn’t just drop in what they thought was a clever design because it kept failing, and how they kept going back to the drawing board. (8,000-person RCT results coming soon)
  • Poachers are targeting newly discovered rare species apparently by reading the scientific papers that describe where to find them.
  • The Freakonomics podcast asks “Are The Rich Really Less Generous Than the Poor?” (as posited by previous lab experiments). They discuss a field experiment in the Netherlands in which the researchers left people fake misaddressed letters with cash (or the equivalent of a check, which could only be redeemed by the person named). Economist Jan Stoop dressed in a fake postal uniform and dropped them in mailboxes in rich and poor households to see if the recipients would re-mail them to the intended destination. Going against conventional wisdom, the rich returned the letters more often, but this gap was entirely explained by a mental scarcity model, which is based on the body of research showing how the pressures of poverty can sap mental bandwidth. The farther away in the month from pay period the poor were (when presumably money was tighter and pressures of life greater), the less often they took the time to re-mail the envelopes.
  • The authors mention in the Freakonomics episode that they decided to add time from paycheck as a covariate based on a chance comment from an audience member at a talk. A new meta-analysis of studies in the American Journal of Political Science finds covariates can make all the difference:


    In almost 40% of observational studies, we find that researchers achieve conventional levels of statistical significance through covariate adjustments. Although that discretion may be justified, researchers almost never disclose or justify it.


    That they were able to discover this seems to have come down to one editor’s decision (h/t Ryan Briggs):

Our analysis is possible because the American Journal of Political Science (AJPS), a leading political science journal, began enforcing the posting of replication data and code. In practice, then Editor Rick Wilson would not proceed with the copyediting of accepted papers until the authors posted their data to Dataverse. We therefore analyze AJPS volumes 56-59 (2012-2015).


And a reminder about those “most in each state” maps from XKCD:


IPA’s weekly links



  • World Bank Chief Economist Paul Romer is giving up some of his management duties, in part due to frictions over trying to get researchers to communicate better, pressuring them to create shorter presentations and use fewer run-on sentences (like this one). Bloomberg and FT’s reports seem gossipy and read like only part of the story. Romer has responded with a post on his personal blog, and offered a public mirror of the internal blog he uses to communicate with World Bank staff.
    • Taking on global poverty and trying to get economists to communicate clearly seems ambitious. But from my experience in both areas I have faith there’s a solution out there to the first.
  • If you’re doing any traveling this weekend, some podcasts:
    • The two Planet Money episodes on how the biggest shock in modern currency, surprise demonetization, happened in India. Turns out it was the brainchild of an engineer who just sold his idea relentlessly for years with a powerpoint of cartoons and got to PM Modi early on (Part 1, Part 2, or iTunes episodes 770 & 771).
    • If you’re in the humanitarian relief world – Tiny Spark talks with a veteran aid worker and a psychologist who specializes in aid workers on how to avoid burnout. One of the most interesting things was how much bad management contributes to aid worker demoralization. (iTunes)
    • Tyler Cowen talks with Raj Chetty (web or iTunes)
  • On that last one (getting back to World Bank) – it’s worth noting Cowen and Chetty both attribute some of the success of Chetty’s papers to the attention his group pays to the writing. He explains that it takes many revisions that go into writing it in a way that makes circuitous methods accessible:

    CHETTY: That is our strategy. I think the other thing that’s extremely important is, we spend a lot of time on trying to achieve clarity. There are ways to write papers in economics that are more accessible to the public and thereby have greater  impact, and there are ways to write papers that are more technically oriented and narrow the set of readers.

  • Two job opportunities:
  • Rachel Glennerster and Mary Ann Bates have a piece in the Stanford Social Innovation Review about how to decide if a successful program will work in another context. They propose four steps:
    • Step 1: What is the disaggregated theory behind the program?
    • Step 2: Do the local conditions hold for that theory to apply?
    • Step 3: How strong is the evidence for the required general behavioral change?
    • Step 4: What is the evidence that the implementation process can be carried out well?

On the Development Impact Blog, David Evans walks through applying those steps to decide whether a technology-aided teaching program that worked in urban India would also work in rural Tanzania.

  • Also in the same issue of SSIR, Blattman, Faye, Karlan, Niehaus, and Udry have a straightforward piece (written for the general public) on what we know and still need to find out about just giving the poor cash.

And the award for best RAs of the year goes to the ones on this Argentina study that found having someone else in a public bathroom increased urinal flushing and hand washing:

Researchers first identified three male restrooms that allowed observers to reliably hear whether a user flushed the urinal and washed his hands. For each user who entered the restroom, researchers alternated assignment between the treatment and comparison group. When a treatment group-assigned individual entered the restroom, an observer entered the restroom eight seconds afterwards and placed himself at another urinal. The observer acted as another user of the restroom without flushing the urinal or washing his hands and stayed in the restroom until the participant left. For individuals assigned to the comparison group, the observer waited outside the restroom and listened to determine whether the participant flushed the urinal and/or washed his hands. To avoid any possible influence by a third party, observers excluded instances when other people were also using the restroom.

Wonder if the RAs were trained in physics.

Final lecture: Order & Violence

Screenshot 2017-05-22 14.21.24

There were supposed to be 10 weeks to the quarter but turns out UChicago doesn’t do a makeup day for Memorial Day! So some major surgery later, my last two lectures fit into one. This is just as well because I don’t know that much about democratization.

Thanks to everyone for the comments and suggestions over the quarter. One of these days, maybe soon, I would like to write this up into a book. So critical comments are valuable.

My syllabus is here and the slides are here:

  • Week 9 – Democracy promotion & wrap-up
  • Week 8 – State building without war making
  • Week 7 – Armed interventions
  • Week 6 – Legacies of Imperialism and post-colonial politics
  • Week 5 – The State and Society
  • Week 4 – Institutions
  • Week 3 – State development
  • Week 2 – Conflict
  • Week 1 – The demand for order

IPA’s weekly links

Guest post by Jeff Mosenkis of Innovations for Poverty Action.

Injectable birth control

  • An attempt to measure and rank the most effective ways to reduce CO2 in the atmosphere ranked educating girls and family planning (globally) above green energy. Fewer unplanned births means fewer carbon footprints. (h/t Osman Siddiqi)
    • On Monday, President Trump expanded the “global gag rule” to mean no funding can go to any NGO that also discusses abortion with beneficiaries anywhere in their operations. This expansion grows its impact from about $600 million to $8.8 billion in funding.
  • There’s a birth control that women can inject themselves (photo above) with minimal training. Each one lasts 3 months, and the price has just come down from $1 to 85 cents a dose.
  • Does foreign aid work? Featuring Charles Kenny, Rachel Glennerster, and Amanda Glassman.
  • A profile of the federal judge leading Brazil’s anti-corruption campaign, “Operation Car Wash,” which is modeled after Italy’s fairly successful anti-corruption campaign in the 1990s.
  • This week’s The Weeds podcast from Vox discusses the Universal Basic Income experiment IPA is conducting with GiveDirectly in Kenya, along with the accompanying general equilibrium one examining how the large influx of cash affects the local economy. As a bonus they also talk about the controversial private school instruction program which we’re also evaluating in Liberia.  The prior “Weeds in The Wild” episode went to visit a Kenyan village in the UBI experiment. (h/t Chris Boyer)
    • One fear of UBIs is that they’ll reduce motivation for work. A review of previous UBI programs in the U.S. and Canada finds no evidence for that, but evidence of positive effects in consumption, education, and health.
    • An analysis of a UBI in Iran also found little reduction in adult recipients working, with some increases for service sector workers. (h/t David McKenzie & Dina Pomeranz)
  • Melissa Dell & Ben Olken find lasting positive effects of Dutch colonialism in Java. Areas close to where colonial sugar was grown in the 19th Century have better infrastructure, more schools, and higher education levels today.

And if you thought it was hard to make sure your survey questions preserve their meaning when translated into multiple languages, be thankful you’re not the sign language interpreter for Snoop Dogg and the Wu-Tang Clan.

Her prep work also includes researching dialectal signs to ensure accuracy and authenticity. An Atlanta rapper will use different slang than a Queens one, and ASL speakers from different regions also use different signs, so knowing how a word like guns and brother are signed in a given region is crucial for authenticity.

Sorry, there’s pretty much no video with Snoop Dogg or Wu-Tang lyrics I can post (but you can find an extended interview with interpreter and taekwondo black belt Holly Maniatty here).

Week 8 of my Order & Violence course: State building without war making

Screenshot 2017-05-17 10.10.44

This week’s lecture draws heavily on one of the most important books on development I’ve ever read: Building State Capability by Harvard’s Matt Andrews, Lant Pritchett, and Michael Woodcock. It is open access and you can download it for free. If you hate trees but love the authors, buy a print copy.

My syllabus is here and the slides are here:

  • Week 8 – State building without war making
  • Week 7 – Armed interventions
  • Week 6 – Legacies of Imperialism and post-colonial politics
  • Week 5 – The State and Society
  • Week 4 – Institutions
  • Week 3 – State development
  • Week 2 – Conflict
  • Week 1 – The demand for order

The conformity of economics papers?


Noah Smith has a fantastic essay on the problem with economics papers. Some excerpts:

…in practice, I think what often happens in econ is more like the following:

1. Some papers make structural models, observe that these models can fit (or sort-of fit) a couple of stylized facts, and call it a day. Economists who like these theories (based on intuition, plausibility, or the fact that their dissertation adviser made the model) then use them for policy predictions forever after, without ever checking them rigorously against empirical evidence.

2. Other papers do purely empirical work, using simple linear models. Economists then use these linear models to make policy predictions (“Minimum wages don’t have significant disemployment effects”).

3. A third group of papers do empirical work, observe the results, and then make one structural model per paper to “explain” the empirical result they just found. These models are generally never used or seen again.

A lot of young, smart economists trying to make it in the academic world these days seem to write papers that fall into Group 3.

…It’s easy to see this pro-forma model-making as a sort of conformity signaling – young, empirically-minded economists going the extra mile to prove that they don’t think the work of the older “theory generation” (who are now their advisers, reviewers, editors and senior colleagues) was for naught.

…But what is the result of all this pro-forma model-making? To some degree it’s just a waste of time and effort, generating models that will never actually be used for anything.

…Paul Romer worries about this in his “mathiness” essay:

[T]he new equilibrium: empirical work is science; theory is entertainment. Presenting a model is like doing a card trick. Everybody knows that there will be some sleight of hand. There is no intent to deceive because no one takes it seriously.

Pure theory papers seem to rarely get checked against data, leading to an accumulation on the shelves of models that support any and every conclusion. Meanwhile, pure empirical papers don’t often get used as guides to finding good structural models, but are simply linearly extrapolated.

Noah flags a bigger problem, however:

In other words, econ seems too focused on “theory vs. evidence” instead of using the two in conjunction. And when they do get used in conjunction, it’s often in a tacked-on, pro-forma sort of way, without a real meaningful interplay between the two.

…I see very few economists explicitly calling for the kind of “combined approach” to modeling that exists in other sciences – i.e., using evidence to continuously restrict the set of usable models.

Another view, however, is that for many problems it’s too soon to build a model. We need more data points and stylized facts before our models will be any good. For a given question, it’s rare to have more than a couple good empirical studies, especially outside the US.

I don’t completely buy this view. Theories and models will be useful along the way to guide where we seek data or how we design experiments. And I personally have found simple general models help me in interpreting the results of my own experiments. So I like seeing them. But I think we are a long way from having enough evidence to restrict the set of useable models.

IPA’s weekly links

Guest post by Jeff Mosenkis of Innovations for Poverty Action.BhubHomepage

  • Chris Blattman had a nice interview with Stephen Ladek of the Terms of Reference podcast (iTunes). Chris talked about his approach to what RCTs are good for – less about that specific program but more about testing our assumptions behind the mechanism through which something might work. They also get into how he got started studying conflict and resolution (because someone offered to let him borrow some Land Rovers), but why he’s now putting his money on that as the enduring question of development.
  • Yesterday IPA and ideas42, with Penn’s Center for Health Incentives & Behavioral Economics and Harvard’s Behavioral Insights Group, launched a new website to make nudge-style behavioral interventions easy for non-researchers to access and understand. It’s called the Behavioral Evidence Hub, www.bhub.org, and has easy to browse and skim explanations of programs that have worked in the past and the principles they use. It’s also got checklists for designing new programs drawing on evidence from past studies.
    • We’re still collecting examples of rigorously tested behavioral interventions. If you know of good candidates, please submit them on the site.
  • People seem optimistic about the new USAID administrator nominee, Mark Green.
  • It might get lost amidst a higher-profile departure this week, but the U.S. Census director resigned, and that’s really important.
  • For the geeks, a really nice discussion on the Development Impact blog on list randomization, and how to get respondents to tell the truth about very personal things like HIV status.
  • From Uganda, a new comedy film about two guys who try to make money starting an NGO.
  • If you want an update on the “China’s great/terrible for Africa” debate, China doubled its investment in Africa in 2016 as the U.S. and the U.K. were pulling back, and is the biggest foreign direct investor in Africa. A long read from the NYTimes Magazine profiles an engineering project in Namibia and questions the China good/bad dichotomy. It’s complicated, and the relationships and markets are maturing and changing the dynamic.

If the census bureau can’t get their digitization act together in time, they may want to hire this guy. The best data entry person in the world may be working in a gas station convenience store, where he’s been taking inventory for 30 years.


IPA’s weekly links

Guest post by Jeff Mosenkis of Innovations for Poverty Action.


  • The datasets above all have the same means, SD, and correlations explain Justin Matejka & George Fitzmaurice of Autodesk in a modern update of Anscombe’s Quartet. But the lesson is the same – always plot your stuff. (h/t everybody)
  • Jobs postings: Summer jobs with Busara Center, full-year with Michael Clemens, and many at the IPA/J-PAL (and several other orgs) jobs portal.
  • Podcasts: Tyler Cowen talks with Cardiff Garcia on FT Alphachat about his new book on his philosophy of economics that he put online free, Stubborn Attachments. My takeaways:
    • A world of full-on Peter Singer effective altruists would be a world with no ice cream because every extra dollar above what one absolutely needs should go to alleviating someone else’s suffering. He doesn’t think that’s realistic to push for but does believe in a slightly scaled down version.
    • He also thinks people undervalue the usefulness of economic growth for alleviating suffering at large scale. So it may be more advantageous to invest a dollar in economic growth if it can alleviate 3 dollars’ worth of future suffering than in alleviating 1 dollar of suffering now.
    • Mainly, it was nice to listen to a conversation that looks at the big picture of how we should be thinking about using the tools of economics. It’s nice to imagine a world where economics draws more on philosophy than math.
  • Last week’s The Weeds discussed Dube & Harish’s paper looking at 500 years of wars, showing Queens were more likely to go to war than Kings (article about it from Pacific Standard). (h/t Dina Pomeranz)
  • How many people have to die for a disaster to make the news? One if it’s from a volcano, 2 from an earthquake, and 38,920 if from hunger (ungated here). Also, if it happens in an Olympic year, the chances of it making the news and getting disaster aid go down. (via Emily Troutman)
  • Using White House visitor logs, Brown & Huang find companies whose executives visited Obama’s White House were more likely to have stock prices go up in the following months (by about 1%), and to experience better regulatory treatment after meeting with federal officials. But those firms also took a bigger hit immediately after the Trump election. (Gated, article about in Bloomberg).
  • PubMed is adding in conflict of interest information on studies after abstracts, though research shows sometimes conflict disclosures can backfire for a number of reasons, including the perception of the conflict having been addressed, while the person receiving the information doesn’t sufficiently adjust for it.
  • Apparently (at least according to Wikipedia) in the years 1937-1942 there was another organization called IPA. A collaboration of social scientists, historians, educators, and journalists, it was devoted to flagging fake news. (h/t Colin Rust)

I don’t know why people think conflicts of interest are a problem. Ocean Spray was happy to be of assistance co-authoring a paper on the healthy effects of Cranberry Juice (which have been questioned elsewhere), and even helped out with the clinical trials registry (via that Vox article):






IPA’s weekly links

Guest post by Jeff Mosenkis of Innovations for Poverty Action.


  • Chris Blattman & Stefan Dercon have an op-ed in the New York Times, reporting on their study with IPA and the Ethiopian Development Research Institute. They randomly offered poor workers in Ethiopia who were applying for factory jobs the jobs they wanted or an alternative entrepreneurship opportunity. Turns out the jobs were pretty bad – most quit and those that stayed weren’t any better off than those who never god a job offer (the entrepreneurs did pretty well though.) Similarly to what Chris discussed with Russ Roberts on EconTalk, they conclude that middle management and professionalization is really what’s lacking in industrialization, along with a more robust social safety net to support those looking for better jobs.
  • Jonathan Morduch presented new findings from Bangladesh. He, with Lee, Ravindran, Shonchoy, and Zaman, encouraged rural migrants who were going to cities to find factory work to use mobile money, and found it bumped up remittances to their families back home. There were all sorts of positive savings and poverty outcomes (but not consumption). Similar to what Blattman and Dercon found in Ethiopia though, this came at the expense of worse health for the factory workers.
  • Some podcast recommendations:
    • On FT Alphachat, Princeton economist Anne Case discusses her triad of papers that have made a big splash about the middle-age white mortality and “deaths from despair” (overdoses, alcohol, and suicide). (iTunes)
      • In addition to the actual findings, what was really interesting was to hear was how she and her co-author have searched for explanations by engaging with colleagues in other fields like sociology and medicine (many of the findings were out there separately, but hadn’t been put together systematically), as well as how they engaged productively with methodological critiques.
    • Freakonomics has a couple of episodes with a collection of big name economists (I believe the technical term is a Davos) to ask how they’d design a new economy from the ground up. But better is a few episodes back “Why is My Life So Hard” where they talk to a couple of experimental psychologists who asked – if we know how good gratitude is for well-being, why are people always complaining about what’s going wrong in their life? They dub it  “headwinds/tailwinds asymmetry” – we take the good for granted but always notice the bad – and go into some fun studies (in pairs of siblings, each thinks the parents treated the other better) then how to avoid the trap. (Part of their inspiration is Louis C.K.’s observation, that Everything’s Amazing and Nobody’s Happy, similar observations to what Max Roser and Hans Rosling have made about the general state of humanity.) iTunes
    • Ezra Klein’s interview with computer scientist Cal Newport, about taking your life back from technology in the age of distraction, was very good (iTunes)
    • The first and most popular episode of Finding Impact speaks with Liz Jarman of Living Goods about how their entrepreneurial model of healthcare delivery in Uganda is reducing childhood mortality significantly, and how the RCT data helped change the way they expanded into Kenya. (iTunes)
  • Automakers met this week with the heads of the EPA and Department of Transportation about rolling back Obama-era fuel efficiency standards, but a new paper finds that higher efficiency standards led to lighter cars, and fewer deaths in traffic accidents. Using the statistical value of a life, they argue the lives saved alone economically justify the higher standards:

    This result is critical for policy: our illustrative calculations suggest that CAFE standards are more likely to have positive net benefits once we account for the saved lives. Further, CAFE may have positive net benefits even if the value of future fuel savings is not included as a benefit.

    (h/t Alexander Berger)

When studies have suggested a gender gap in tenure, I’m not sure they were talking about The Gap actually having a women’s clothing line called “The Tenure Track Professor” (J Crew’s Assistant Professor line seems less than practical too). Still waiting for a men’s relaxed fit Tenure Track Suit. (h/t Laura Seay)


Do sweatshops lift workers out of poverty? This experiment might surprise you.


Expecting to prove the experts right, we went to Ethiopia and — working with the Innovations for Poverty Action and the Ethiopian Development Research Institute — performed the first randomized trial of industrial employment on workers. Little did we anticipate that everything we believed would turn out to be wrong.

Stefan Dercon and I have an op-ed in today’s New York Times.

The best tweet so far:

IPA’s weekly links

Guest post by Jeff Mosenkis of Innovations for Poverty Action.


  • Microsoft’s Steve Ballmer started an initiative to make government data from all levels very easy to access. So if you’re in the middle of a Facebook debate and want to know how much the government spends on healthcare versus defense, or how medicare is doing, or how many firearms are found at TSA checkpoints every year since 2005, you can find an easy chart at https://usafacts.org/
  • Owen Barder argues that it would be much more effective for poor countries to prepare for disasters in advance, just like everybody else does, by taking out insurance. In a new short briefreport, and CGD event video, he argues that setting up insurance would make responding to a crisis like an Ebola outbreak orders of magnitude cheaper. An automatic payout would mean a much faster response than the current system of starting fundraising after the disaster occurs, and could be designed in a way that works with the private sector and governments to incentivize better development (freeing up governments to build effective water management in conjunction with guaranteed protections in case of flood.)
    • This article was written by someone from the African Risk Capacity (ARC) insurance org, which does just that. It argues nobody heard about the drought in Senegal last year because Senegal had bought into the insurance pool, which helped it cope much more effectively than Somalia is coping now.
    • But Barder adds a cautionary tale about ARC from last year. Malawi had agreed to terms requiring a payout in the case of a drought affecting more  than 1.39 million people. While an assessment found over 6 million people affected, ARC’s computer models (based on satellite cloud imagery and previous years’ data) only computed 21,000 people at risk. (Though ARC did later make an $8 million payout after a household survey and concluding farmers were growing different crops than their model had assumed.)
    • I’m sympathetic to the insurers though. Economists Dean Karlan and Chris Udry told Planet Money about some of their problems setting up a rainfall insurance program to test in Ghana. Insurance requires very specific and objective (non-debatable) criteria to trigger a payout, which often doesn’t jibe with how individuals are suffering, and also good data from the past to estimate future risk, both of which can be hard where rain is fickle.
  • Jonathan Morduch was interviewed about his new book with Rachel Schneider The Financial Diaries, the result of them tracking every dollar spent by 235 low- and moderate-income households for a year.
  • An investigative piece in BMJ finds Coca-Cola tried to influence how journalists covered scientific research. According to the piece (gated), the company funded conferences for journalists to attend, through grants to the University of Colorado so the journalists didn’t know the true sponsor. At the conferences, they heard research emphasizing that obesity was exercise-related rather than about diet, in line with Coca-Cola’s marketing strategy.

This is my colleague David Batcheck’s last day at IPA, and he left me with a warning about “Big Data”