by Dean Karlan
Conley and Udry wanted to know: Was this happening? Were farmers actually sharing information? And if so, how did it flow?
Since it has neither size nor shape, nor color, odor, or taste, information is a very hard thing to track. Spoken words, the vehicles for most of the farmers’ information, disappear into thin air. But even if we cannot see the thing itself, we can follow its impact. This was essentially Conley and Udry’s strategy—to model the spread of information between pineapple growers by monitoring their fertilizer decisions for evidence of learning.
To begin, they had to map the pathways along which information could potentially travel. So they visited the homes of 180 pineapple growers in Akwapim South and asked which other growers they spoke to about farming. The result was a series of interconnected networks—“information neighborhoods”—linking each grower to the people he might learn from or teach.
For the next two years, Conley and Udry checked in on the farmers regularly to see the process at work. Sure enough, the information neighborhoods were crackling with activity. People were learning. When a farmer experimented with a new fertilizer regimen, for instance, his information neighbors took note. If the new treatment led to better-than-average profits, the neighbors were more likely to follow suit. On the other hand, when a farmer saw low profits from a given fertilizer regimen, his information neighbors were more likely to try something different at the next planting opportunity.
Beyond confirming the presence of learning in general, the fertilizer data also revealed a natural hierarchy. Whom do people listen to the most, and who does the most listening? Farmers new to pineapple growing were the most susceptible to influence from their information neighbors, while veteran pineapple growers had the most clout. The veterans’ wealth of experience translated into influence in two ways: First, their information neighborhoods tended to be larger; and second, the people in those neighborhoods paid close attention to—and were more likely to emulate—their actions.
The news that greenhorns take their cues from old hands may not be so surprising, but it is important if we want to design programs that work.
For instance, suppose you had found the perfect crop variety and fertilizer mix, but you only had enough money to fund a dozen demonstration plots. Where should they go? You could choose at random, or space them out evenly across the region, or place them near the biggest village; but there is a better way to get the message out. Just as you would put your ads on the Web sites that get the most traffic, you should put your demonstration plots in the most-watched fields. Given the findings of the viral pineapples study, those fields belong to seasoned farmers.
The Collapse of DrumNet
Because he lived some two hundred years before the concept came into being, the eighteenth-century Scottish poet Robert Burns couldn’t have known the first thing about development economics. So he certainly would not have suspected, when he penned them, that his famous lines would be so relevant to the world of antipoverty programs at the dawn of the new millennium. He wrote, The best laid schemes of mice and men
Go often awry,
And leave us naught but grief and pain,
For promised joy!
Or, to paraphrase: Despite our efforts to anticipate problems and plan ahead, things go haywire, reducing our grand visions to rubble.
Which brings us back to DrumNet, the Kenyan program that encouraged and helped farmers to grow French beans and baby corn for export to Europe. When we left them, the DrumNet farmers were skating right along. In 2004, participants had made a tidy profit selling the first season’s harvest to the exporters, who in turn sold it to European grocery stores. When they returned to their fields the following year, the DrumNet farmers faced a new challenge: EurepGAP, a set of food safety standards for agricultural products, adopted by most European grocery stores in early 2005.
If they wanted to continue exporting their produce, DrumNet participants would have to satisfy the criteria for EurepGAP certification. That meant farmers had to have a grading shed, a secure chemical storage facility with flush toilets and cement floors, an arsenal of mechanized sprayers, modern protective suits to wear when handling chemicals, detailed records of the seed varieties and fertilizer treatments for every plant they raised, and professional analysis of their water and soil done each year. Sadly, this extensive laundry list was simply out of reach for most DrumNet farmers, and it was a stretch even for the few who potentially could manage it. Not a single DrumNet participant could have earned EurepGAP certification without making changes to his farm.
So they had a choice: buy in or get out. Invest in sheds, storage, sprayers, and the rest of it, or quit the program. One independent study estimated the cost per farmer of complying with EurepGAP at $581, or about eighteen months’ income for a typical DrumNet member. That wasn’t much of a choice at all; the vast majority just couldn’t afford it.
Something had to give, and that something was DrumNet itself. In 2006, at the end of its second growing season, the program collapsed when the exporters declared they would not buy the produce, none of which had been EurepGAP certified. Without the exporters to take delivery of the crops, the chain broke down. Trucks never came to pick up the bushels of baby corn and great sacks of French beans, so farmers were stuck with them. Some of the harvest simply rotted; the portion they did manage to sell went to middlemen, usually at a significant loss.
By the time another year had passed, the Gichugu Constituency DrumNet program was nothing more than a bad memory. The farmers reverted to growing maize, potatoes, and kale, just as they had a few years before—just as their parents and grandparents had for ages.
Strong Foundations
Though behavioral insights like viral pineapples and the Savings and Fertilizer Initiative program have a great deal to offer, they cannot carry the day all by themselves. The foundation on which development programs are built—the bedrock of open markets, legal systems, infrastructure, and the like—have to be in order too.
Changes to the bedrock can hurt (as when the adoption of EurepGAP export standards demolished DrumNet) but they can also help. They have the power to open up vast new economic spaces, to push back the boundaries encircling the poor. In the state of Kerala, on India’s southwest coast, a broader economic horizon meant more people got to eat fish.
Since at least the year zero, the fishermen of Kerala have been hauling in netsful of the sea’s glinting silver bounty and dragging them ashore to sell in the markets that dot the state’s hundred-mile coastline. For most of the past two thousand years, they steered their boats into shore each evening by the market closest to their homes, where their day’s catch would be bought by retail vendors. This system had the advantage of simplicity, but it was far from efficient. The number of fish hauled into the market on a given day often exceeded the number of hungry mouths to eat them, or vice versa. Ironically, towns just a few miles apart often had opposite problems. Looking across the entire state, supply and demand might have been in balance, but buyers and sellers were being mismatched—to their mutual disadvantage—by the fishermen’s inflexible custom of selling exclusively at their local markets.
In 1997, Robert Jensen, a UCLA economist (and filmmaker!) organized a weekly survey of fifteen coastal markets to track the sales of fish over time. What he saw were a lot of missed opportunities. On a typical day, eight of the fifteen markets had an excess of either buyers or sellers, which meant hopeful customers went home disappointed in some towns while surplus product was left to spoil in others. Prices, working to align supply and demand, fluctuated wildly from day to day, up and down the coast. Where there were too many fish, vendors literally had to give them away; where there weren’t enough, buyers paid ten rupees for a kilo of sardines. The upshot was that the market was an uncertain place for both fish eaters and fishermen.
That would all change with the introduction of cellular phones in Kerala between 1997 and 2000. The installation of towers along the coast wasn’t part of any aid
program, but an infrastructure investment by a for-profit communications firm. It also happened to be a perfect solution to the fish markets’ problem. Now, instead of blindly choosing to sell at the home market each afternoon, a fishing boat could call around to check the prices at nearby markets and set a course for the most profitable one.
In estimating the overall impact of cell phone access on the market for fish, Jensen took advantage of the fact that the cellular network expanded tower by tower, gradually opening new areas of the coast to communications. It created what economists call a “natural experiment”—something that looks and acts a lot like an RCT, except that people are assigned to treatment or control by a natural process (in this case, the construction of cell phone towers) rather than by explicit and purposeful randomization.
When each new region of Kerala came into the cellular coverage area, prices and availability of fish stabilized virtually overnight. The chart of weekly sardine prices across that region’s markets, which had looked like the seismograph of a catastrophic earthquake, converged to a steady plateau around six rupees per kilo. Now customers could count on finding reasonably priced fish whenever they went to market. Fishermen could count on selling their catch every day and earning a steady income. The cell phone company must have been pleased with all the ship-to-shore communications, and even economists could take some satisfaction in the elimination of a deadweight loss: Unsold fish were no longer being left out to spoil. It truly was a win for everybody.
Why Fighting Poverty IS Like Rocket Science After All
Harvard economist Michael Kremer is an inveterate optimist and a firm believer that we can make great strides against poverty if we use tools and methods proven to work. That’s why, in lieu of vacation, he spends his free time interviewing schoolchildren in Kenya about their exposure to hookworm, and why he thinks constantly about how to get governments and organizations to scale up proven-effective ideas. He is also the first in the triumvirate of MacArthur Foundation “genius” grant winners who are part of this movement (Esther Duflo and Sendhil Mullainathan being the other two). If he really thought economic development was like a space shuttle disaster, he wouldn’t invest so much of himself in it. Still, he has been known to lead off a discussion about development with the story of the Challenger.
Watching as the shuttle, breaching gently like a great metal whale over the Atlantic Ocean, bursts suddenly into flame and disperses in spidery fingers of white smoke, one’s immediate thought is that only a total mechanical failure could be to blame. The voice of mission control crackles in an eerie monotone while falling debris traces the gnarled shape of a bare tree branch across a cloudless sky: “Obviously a major malfunction . . .”
But weeks later, when the wreckage had been recovered from the ocean floor and painstakingly inspected, NASA determined that the culprit was a simple rubber O-ring. It had become brittle in the cold morning air on launch day and failed to hold its seal.
Kremer’s point is that the Challenger, a mechanical wonder comprising thousands upon thousands of moving parts, millions of horsepower, and three fifteen-story-high fuel tanks, relied completely on a flimsy piece of rubber much like the one found in your bathroom faucet. In fact, it relied completely on innumerable such commonplace and forgettable things. The failure of a single rivet, valve, or circuit could just as easily have caused a catastrophe.
Like space shuttles, development programs are complex systems with many potential blowout points—prices, credit, infrastructure, technology, law, trust, and even weather. The failure of DrumNet shows just how vulnerable they are. Who would have thought that the buying policies of Italian supermarkets could derail Kenyan farmers?
The good news is that every potential blowout point is also an opportunity. O-rings are not hard to fix; you just have to keep testing. After all, who would have thought that a simple fertilizer coupon could change generations of farming practices?
9
TO LEARN
The Importance of Showing Up
In September 2000, the UN launched an ambitious campaign against poverty by announcing the Millennium Development Goals (MDGs), a set of eight development benchmarks that, if achieved, would constitute a drastic reduction in suffering and privation around the world. It was a big plan on a big stage, and a lot of people were enthusiastic about it. The declaration containing the MDGs was adopted by 189 nations and signed by 147 heads of state and governments.
Such a consensus was a watershed moment in the global conversation about development and poverty. At last, the world’s movers and shakers were in agreement (not perfect agreement, of course—many griped about their favorite initiatives being left out), if not about the principles that ought to guide development or the broad issues in most dire need of attention, then at least about specific targets and clearly defined indicators to measure them. Now, some argue that support for goals and targets can be too easily won—and means little—when the supporters themselves face no immediate consequences for failure. To quote Esther Duflo, “Nobody is going to come from Mars and say, ‘You didn’t reach the goals, so we will invade’—there is no onus [on the UN to reach them].”
In any event, one thing everybody could get behind was schooling, an easy rallying point. (After all, have you ever seen a political advertisement that announced a stand against education?) Right near the top of the list of MDGs—second only to “Eradicate extreme poverty and hunger”—is “Achieve universal primary education.” There are good reasons for the high ranking, not least of which is the fact that education trickles down into many of the development outcomes we care about. Educated people have better jobs, better health, and greater gender equality. Many also argue that education is an end in itself, that literacy and numeracy are essential to an active mental life that is ultimately its own reward. So, politically, education was an easy win for the UN.
Things were not so clear for Anthony, who lived in a small village in central Ghana. Like the UN, Anthony was convinced that education was valuable, both for its own sake and for the doors it would open in the future, and so he wanted more. Unfortunately, dignitaries’ signatures alone couldn’t help him.
Had he been asked to, though, he certainly could have appended his name to the UN’s landmark declaration; when Jake met him he had just finished practicing his penmanship.
Jake’s mom had come to visit him in Ghana during his time as a research assistant there. They had traveled along the pitted highway up to Lake Bosumtwi, a vast silvery disc bounded by the steep crater walls of an extinct volcano. They drove along the water’s edge on a dusty track that came over a rise and curved to a stop in a flat dirt basin shaded by tall, wide trees. The village kids must have heard the car coming, for they were on it before it pulled to a stop. They were a motley crew, different shapes and sizes: tall, skinny teenage boys projecting authority, rambunctious younger boys terrorizing one another, prepubescent girls with little brothers and sisters on their hips. You could tell from the commotion that they didn’t get too many visitors.
Jake and his mom were swept up in the little crowd and guided by some of the older boys through the trees and a small reed marsh beyond to the shore of the lake, where they were invited to stand and marvel. They did. One teenage boy, who looked to be about seventeen years old, walked up while Jake was looking out over the water and tapped him on the shoulder. He was smiling hugely. He said, “Good afternoon!”
“Good afternoon. Thank you for bringing us to the lake.”
“Oh, this lake.” He pointed and looked past his outstretched finger over the water and kept smiling. “That is Lake Bosumtwi.”
“Yes, I’ve heard.”
“And I am Anthony.”
“Hello, Anthony. It’s nice to meet you. I’m Jake.”
“Tjchek?”
“Or Jacob. Jacob is fine too.”
“Oh, Mister Jacob! We are very glad to meet you.” Anthony was still smiling. He had big, smooth, almond-shaped eyes that darted around and flas
hed excitedly.
“I’m glad too. You can just call me Jacob, though. I don’t need to be Mister.”
But the die had been cast. To Anthony, Jake was Mister Jacob that day and all the days that followed. When they were leaving the village a half hour later, he said Mister Jacob when he asked for Jake’s mobile phone number, and when he took it down in his foolscap notebook he wrote “Mr. Jacob” next to it in neat, deliberate print with a blue ballpoint pen. That was when he showed off the pages of penmanship practice. It was mostly his name written over and over, spaced out evenly across the lines so they were arranged in tidy columns. “I did them this morning,” he said.
Anthony called Jake a couple weeks later. He said he wanted to discuss something very important, and that he would come to Accra for a meeting. Because coming to the capital would cost him about six dollars (not to mention eight hours in a rickety van), Jake suggested that they should talk it out over the phone, but he wouldn’t budge. He would be there on Friday, he said.
Friday came and rain poured down so that the streets rushed like muddy streams in the spring melt. Somehow, Anthony fought his way out of the morass of the capital’s central trotro station, with its scores of lurching, skidding vans, and up the hill to the General Post Office, where Jake found him sitting under the awning at the appointed hour. His clothes were wet, but not muddy, and he wore the same wide smile under his darting eyes.
They stood beneath the awning, and sheets of water fell on three sides. They almost had to yell to be heard over the rain. Anthony described his predicament. He was just finishing the SSS (Senior Secondary School) exams, which would determine his eligibility for tertiary education—university, polytechnic, or vocational training college—and he was worried. His parents, he explained, had put all their educational eggs in one basket, and the basket was almost empty.
