I went into Robert Bates’ recent book When Things Fell Apart: State Failure in Late-Century Africa with high expectations.  His 1982 work Markets and States in Tropical Africa: The Political Basis of Agricultural Policies remains one of the best pieces of African political analysis I’ve ever read, displaying an admirable willingness to take African domestic politics seriously, and thereby producing a insightful overview at the process of agricultural policy-making.  (I must confess here that I thought this sounded like a resoundingly dull topic until I read the book, at which point I realized that Bates had captured many of the fundamental dynamics of African politics in the post-colonial era.)  At any rate, I fully expected the same insight from When Things Fell Apart – and was disappointed to put it down feeling like I had learned little of value.  The book is a decent primer on African politics in the 1980s and 1990s for the reader who’s unfamiliar with the topic, but it doesn’t even define “state failure” clearly, let alone provide an adequate metric for determining why some African states failed and others didn’t.  One could be excused for coming away with the conclusion that the entire continent imploded in the 1990s.  Such an important topic deserves a more clearly-written book (as other reviewers seem to agree).

A bit of background: my primer on African politics in the 1980s and 1990s was Michael Bratton & Nicholas van der Walle’s 1997 book Democratic Experiments in Africa: Regime Transitions in Comparative Perspective (which I read for a course with the excellent Peter Lewis).  Democratic Experiments isn’t what I’d call compulsively readable, but it does draw a clear picture of the generalized path that many African countries followed from authoritarianism to democratization.  Generally speaking, the profound economic crisis of the 1980s unsettled many African autocrats’ grips on power, and by the end of the decade protests over economic stagnation and single-party rule were common.  Many leaders reluctantly switched to multiparty democracy under such pressure, whilst others were able to repress domestic dissent but were still pushed towards democracy by international donors.  Of course, some states avoided democratization entirely, as in Sudan with its civil war, Somalia with its total collapse, and Zaire with Mobutu’s preternatural skill at clinging to power.

So, in short: prolonged economic stagnation weakens leaders’ holds on power and provokes political unrest, which might lead to democracy (Ghana), civil war (Somalia), or both in succession (Rwanda & Burundi).  I was really hoping that Bates’ book would start from this point and investigate the determinants of why some nations collapsed and others weathered this period intact.  Instead, Bates takes the same stylized elements as Bratton & van der Walle (economic crisis + resultant instability in the political system), adds a structural assumption about politicians’ heightened discount rates in times of crisis, and comes up with a recipe for violent political infighting over access to resources and resultant state collapse.  This is the flip side of the processes of democratization explored in Democratic Experiments, and it’s a plausible analysis for many of the states that did in fact fall into civil conflict.

However, Bates never actually defines “state failure,” nor identifies the set of African states or even the specific time periods to which this analysis applies. His working definition appears to mean “civil war during the 1990s,” based on statements like the opening sentence of his conclusion, which discusses Liberia and Somalia.  It’s an odd oversight for a political scientist, given that the Liberian civil war is arguably a different type of “state failure” than the complete collapse of the central government in Somalia.  (Rwanda and the DRC are also great examples: both were sites of great violence in the 1990s, but today Rwanda has one of the strongest governments in Africa, whilst the Congolese government doesn’t even control all of its own territory.)  For that matter, evidence for the book’s arguments is culled from across the continent, including anecdotes from states such as Ghana and Zambia (which have never had civil wars) alongside the obvious choices of war-torn Burundi and DRC.  Bates’ reasonable analysis of the risk factors for state collapse is greatly weakened by this inability to apply his hypotheses to specific African states.

Only eight months after I finished the book, I thought I’d finally review More Than Good Intentions, by Dean Karlan and Jacob Appel.  Dean is my supervisor at Innovations for Poverty Action, as well as the organization’s founder, and what I particularly appreciated about the book was its clear explanation of how academic research in behavioral economics can lead to solutions for real problems of social policy in the developing world.  Given the level of popular discontent with neoclassical economics, fairly or unfairly, after the financial crisis, works like this go a long way towards demonstrating that the economist’s conceptual toolbox can contribute to making the world a better place.

More Than Good Intentions opens with a short review of the Easterly-vs.-Sachs saga, and essentially sidesteps the debate about whether aid ever works with a call for more evaluation of extant aid programs.  Their chosen tool is the randomized controlled trial.  Of course, there are any number of development problems that are not amenable to randomized evaluation.  Questions about the ethnicized distribution of government resources or the transnational funding networks of rebel groups really call out for other epistemological approaches.  What RCTs can do well is evaluate program-based aid in contexts where funding shortages mean that some potential beneficiaries can’t be included, and this is precisely the approach taken by the research projects summarized in the book.

The rest of the book is thematically structured around financial activities (borrowing, saving, consumption) and non-financial activities (agriculture, healthcare, education).  The financial sections of the book are the best non-technical introduction to the topic that I’ve seen. Dean’s interests tend towards microfinance and decision-making, and these chapters give a thorough overview of contemporary Western narratives around microfinance and the many reasons why the financial needs of the poor are more varied than simply “getting a loan.”  For instance, whilst an RCT conducted in South Africa showed that randomly extending microfinance loans to people who otherwise wouldn’t have gotten one did raise those clients’ incomes, qualitative data from the Philippines also showed that the rigid structure of microfinance products drives many people back to the neighborhood moneylender.  (The takeaway here isn’t that moneylenders are evil, but that microfinance banks might take a lesson in customizable loan repayments from them.)  Another RCT in Peru used list randomization [PDF] to show that nearly a third of microfinance clients use their loans for household consumption instead of business needs – technically a violation of their loan agreements, but a more accurate reflection of their current financial needs.

The non-financial chapters are also consistently interesting, although they tend towards summarizing notable research and policy innovations rather than placing the results within a global context.  They’re like the greatest hits of development research.  For instance, the agriculture chapter doesn’t provide an overview of agricultural modernization attempts in Africa, but it does shed light on why Kenyan farmers don’t purchase fertilizer when they need it (it’s hard for them to save money after the harvest), and how Ghanaian pineapple farmers spread information about new agricultural technologies.  Similarly, the education chapter doesn’t go into great depth about the history of universal primary education, but it does demonstrate that programs as simple as providing uniforms or cash grants to poor students can dramatically improve attendance.  One of the most remarkable studies of recent years showed that treating Kenyan students for intestinal worms with a twenty-cent pill reduced absence rates by up to 25%.  This result was so spectacular that the researchers started an NGO, Deworm the World, dedicated to reproducing this success.

All in all, More Than Good Intentions makes a strong case for the relevance of behavioral economics to development policy.  It’s also an excellent popular introduction to some of the fundamental questions of foreign aid and development economics.  I gave a copy to my parents to answer their perennial question of “so what exactly do you do in development work again?”  So far it seems to be working.

Tiller

Maize

Playing in the shea nuts

N&V also included a small methodological note which I found especially telling in re: the social importance of land in Burundi.  The sample for the data underlying this paper was drawn from households who completed the 1998 Burundi Priority Survey, which was a joint project of the Burundi Institute of Statistics & Economic Studies and the World Bank.  Interestingly, despite the fact that N&V collected a second round of data in 2007, after multiple years of war, they noted only 13% attrition from their original sample.  As they write, “In Burundi the pressure [on] land is extraordinary high…  As a result people may have only have fled [from the conflict] at the very last minute, if there was no other option, and return[ed] immediately after the violence…to ensure their claim to land. Most often, our survey team would find the households in the same location as in 1998″ (p. 6).  Pretty remarkable.

Cover.  As you’ll see, the intended audience is one of male farmers, who grow maize as an accepted male crop – but their wives still play a significant if unacknowledged role.

Skipping to the middle of the book for a second, this illustration of the farmer in his field captures a lot about the intended audience.  Note the straw hat, simple sandals, manual distribution of fertilizer, and use of a repurposed condensed milk can to hold it.

Contrast this to the implicitly better-off fertilizer salesman, who shows up to deliver fertilizer in a baseball cap and closed-toe shoes.  (Jan Chipchase has a good post mentioning footwear as status markers.)  There’s not much romanticization of the economic prospects of smallholder farming here.

There’s some great detail as well in the section on how to avoid storing fertilizer improperly.  The protagonist obviously isn’t doing too badly, because he has a metal roof, but leaks are a persistent problem.  His young daughter, here seen splashing in the spilled fertilizer, is wearing appropriate earrings and waist beads.  (And yes, it’s quite normal that she’s wearing nothing else.)

What does it mean for farmers to recognize themselves in this officially sanctioned view of how to farm?

There is a very strong correlation between my returning to Africa and my completely neglecting this blog – which says less about African internet than about how busy I always find myself when I’m here!  I came into my current position with IPA at the beginning of a two-month household survey examining underinvestment in agriculture in northern Ghana, and since then our whole team has been working non-stop.  Our surveyors leave for the field between 7 and 8 am every day, so I’m usually at the office by 6.30 to make sure that everything’s prepared.  Then it’s a long day of tracking survey documents, sorting soil samples, assigning survey teams to new communities, preparing per diem payments, troubleshooting the netbooks & survey software used in the field, selecting respondents for field audits, taking calls from surveyors, and making frequent three-hour round trips up to our satellite office in Walewale, among any number of other things.  An early day might end at 7 pm, and a late one at 10 pm.  The sheer amount of work has forced me to grow more as a manager than I have in any other position I’ve yet had, which has been fantastic.  It simply doesn’t leave much space at the edges of my days for anything else.

Thinking some more about the geographic scale of different agricultural markets has led me to consider how information availability and pricing might differ between them.  Mobile phones are still a fairly rare commodity in rural Rwanda (outside of the towns), and if some set of extremely poor farmers were only selling extra produce hyperlocally, it seems that perhaps prices would either be set in total isolation from regional or national prices, or might be determined exclusively by a few people with mobiles.  (I wonder if outside information would be convincing in this scenario – if the phone guy says to the farmer, “this cassava is 5 francs cheaper in Nyamata,” would the farmer accept this as a negotiating tactic, or assume that he’s lying?  Perhaps it’s connected to how easy it is for the phone guy to actually access the cheaper cassava in Nyamata – the credibility of his threat.)

Then again, even with mobile technology to help information access along, its helpfulness still seems fundamentally predicated on A) physical mobility and B) social networks.  Information about great prices in a town up the road will be less useful to a farmer if he still can’t reach it easily, and receiving the information in the first place is still connected to one’s actual social network (and ability to pay for airtime, of course).  I wonder if the differing availability of mobiles to rural residents of different socioeconomic statuses may actually increase the vulnerability and exclusion of the poorest of the poor, rather like differing levels of access to microinsurance might actually push healthcare farther out of reach of the poorest.

This is probably analogous to any form of large-scale, cost-efficient production – the potential for socioeconomic displacement depends on the product you’re displacing.  And I suppose that’s also a commentary on the target market.  I’d guess that the vast majority of Rwandese agricultural production is by smallholder farmers for either hyperlocal (within a few miles) or local (say half a day’s walking distance) markets, and that only certain commodities are ever going to be linked to the Kigali low-income or high-income markets.  (There’s not really a hell of a lot of local stuff being sold to high-income people in Kigali anyway, except maybe the specialty vegetables at Simba. The wealthy mzungus & Rwandese have their own totally separate selection of imports, whose prices correspondingly reflect good information about international pricing and transport costs.  Which is to say, they cost an arm and a leg.)

So I’m wondering if perhaps there are many smallholder farmers who are producing the same types of commodities as might be produced commercially, but if they’re really not linked into the same markets that large-scale production would target.  It seems like this might be more of a threat to medium-scale enterprises that already serve perhaps a regional or national (Rwandese) market, but then that’s a difficult ethical issue from threatening the livelihoods of smallholders who have almost nothing to start with.  It’s a persistant challenge of doing business in poor countries – weighing the social benefits of increased food production & lower prices (marginal improvements to the well-being of many relatively poor consumers) vs. the possibility of undercutting the incomes of medium-scale farmers (major detriments to well-being of a few relatively poor producers).  As an aside, I think there’s also an interesting lesson about innovation in there – that its effects will go in different directions depending on whether it’s creating an improved version of an existing product or service, or whether it’s creating something so novel that new markets develop around it.  “Innovation” always seems to be referred to as a uniform category of activities, without regard for its differing impacts, at least in popular parlance.

I was struck recently by the New Times‘ “good news” that the rate of non-agricultural employment in Rwanda has doubled in the last 10 years!  That is, it’s apparently gone from 5% of the population to 10%.  Kigali is such a bubble; it’s easy to forget that there are many Rwandese people who will live out their days without ever seeing a multiple-story building or a paved road.

That said, this official stat about 10% non-ag employment/90% subsistence ag seems like at best a crude measure of people’s actual livelihood strategies.  I haven’t seen a specification of how much non-ag employment is formal or informal (or even what you’d consider to be “informal” around here), and in a similar sense, I’m sure that many rural residents participate in at least some non-ag activities as part of their income-smoothing plans.  I’d love to see a better measure of this data – to get a sense of how many people are exclusively dependent on agriculture (though perhaps with different types of income-generating activities within the sector), and how many are principally dependent on ag, with a significant amount of income still coming from other activities (carpentry, say, or teaching, or day labor on other farms…).  I also know a small subset of co-op members in Nyamata who grow most of their own food, even though they make a majority of their income from basket-weaving – I wonder how they’d be classified.