by Dean Karlan
Since prices didn’t change the kinds of people who got nets or the way they were used, the difference between cost-sharing and free distribution could be easily summarized: A lot fewer people ended up protected, and the providers of the nets saved some money. Unfortunately, they weren’t saving much. Each net cost about six dollars to produce, so when Population Services International sold nets to Kenyans for seventy-five cents, according to their prevailing policy, they were already bearing the vast majority of the cost. Covering the last seventy-five cents would have increased their cost per net by about 13 percent, but then they could have served four times as many people!
In fact, given the indirect social benefits of protection (i.e., breaking the chain of transmission), spending a little bit more to boost demand for the nets probably made economic sense too. Cohen and Dupas crunched the numbers from the providers’ perspective and determined that, on average, it was likely cheaper to save a life by giving the nets away than by selling them.
The Most Important Drops in the Bucket
Malaria isn’t the only global health scourge that has the attention of the development community. Diarrheal diseases claim two million lives (mostly children’s) each year worldwide—a loss that is doubly tragic because it is so needless. There are cheap, highly effective ways to treat and prevent diarrhea, but they are woefully underused.
Back in chapter 3, I mentioned Sendhil Mullainathan’s discussion of oral rehydration salts as an example of the “Last-Mile Problem”: We have a perfectly viable solution, but we’ve failed to get it into the hands of the people who need it most. Chlorine—dilute chlorine solution for drinking water, to be precise—is a very similar case.
When human and animal waste collects near a water source like a spring, a well, a borehole, or a stream, water drawn from that source is susceptible to contamination by E. coli and other diarrhea-causing bacteria. Even water drawn from a clean source can become contaminated if it is stored in a dirty container. But a few drops of chlorine will reliably eradicate diarrheal bacteria from ten to twenty liters of water, even in a dirty container. It’s powerful stuff.
Most of the people of Busia, Kenya (the same village with the school uniform giveaway program we saw in the last chapter), know a thing or two about chlorine. If you asked, 70 percent of people would tell you that dirty drinking water can cause diarrhea. Better yet, closer to 90 percent would tell you that they had heard about WaterGuard, the brand of dilute chlorine solution sold at more than a dozen shops in town. This broad awareness of both the problem and a solution was largely due to the efforts of Population Services International, which introduced WaterGuard in 2003 and has advertised it widely ever since. Just as it does with bed nets, Population Services International sells WaterGuard at a nominal price rather than giving it away. A month’s supply for a household in Busia is sold for about thirty cents, or roughly a quarter of a day’s wage for a typical agricultural worker.
The only problem is that this cheap, well-known solution to a pervasive problem has failed to catch on.
People still fall ill frequently with diarrhea; some die. Michael Kremer, the Harvard economist whom we last saw searching for the best way to get Kenyan kids into school in chapter 9, decided to try to tackle the diarrhea problem as well. There were a lot of ways he could dream up to get people to use more chlorine—from simply giving it away, to community education programs, to cajoling individuals—but it wasn’t obvious which would work best. So he tested them all.
Kremer, along with Sendhil Mullainathan, Edward Miguel, Clair Null of Emory University, and Alix Zwane from the Bill & Melinda Gates Foundation (and a former board member of IPA), designed a series of RCTs to find out how different inducements to use chlorine stacked up against one another.
First, they tried lowering the price by distributing free bottles of WaterGuard to some houses and half-off coupons to others. Cutting the price in half did increase the portion of houses that used chlorine from about 5 to 10 percent. But giving it away was an obvious choice from the public health perspective. Houses receiving free chlorine saw a huge jump in usage, to roughly 70 percent.
The researchers suspected that prices might not tell the whole story. They had ideas from behavioral economics too—ideas that had to do with social learning, attention, and trust. They tested the effectiveness of one-on-one and village-wide encouragement of chlorine use by NGO workers, the importance of social networks in getting people to adopt chlorine, and the impact of paying local promoters to pitch the product within villages. There were some things to say about the first two—village-wide marketing worked slightly better than one-to-one, and community leaders’ chlorine usage appeared to influence other people’s decisions somewhat—but they only led to small and short-term increases in overall usage. Having a local promoter drawn from the village itself, on the other hand, drove up usage both immediately and persistently.
In some sense, the most powerful local promotion device is the example of others in the community. Maybe the way to drive usage, the researchers thought, is to make chlorine use public and visible. So they also designed and tested a chlorine dispenser: a stand with a (free) bottle of WaterGuard locked inside and a special spigot that delivers exactly enough chlorine to disinfect a standard twenty-liter jerry can of water. Instead of adding chlorine at home, as one would with retail WaterGuard, the dispensers were located right at water sources. You add the chlorine when you fetch the water, and it does its work as you walk home with your jerry can balanced on your head. This setup also has the benefit of a natural attention mechanism: The chlorine is right there when you are preparing your water, begging you to put a dollop inside the jerry can. This is akin to selling fertilizer coupons at harvesttime, right when cash is in hand. Capturing people’s attention at the critical moment is one of the many reasons that timing matters.
Dispensers were the best approach of all. They were at least as effective in driving usage as providing free WaterGuard directly to households, and they didn’t require costly door-to-door distribution. Better yet, the dispensers appeared to engender real, lasting changes in behavior. People in dispenser communities appeared to be disinfecting more, and more consistently, as time went on. In communities where free WaterGuard was provided to households, chlorine usage peaked just weeks after distribution and slid down thereafter; in dispenser communities, usage continued to increase for months after installation, and remained high even a year and a half later.
Beatrice and Agnes
It looks like chlorine dispensers might be the answer to the Last-Mile Problem for chlorine—or at least part of the answer. But there are still big unanswered questions. First and foremost, who’s going to pay? We saw that charging even a nominal price (fifteen cents per family per month) for retail WaterGuard drove away most people; but with a shared community dispenser, maybe the cost could be split among many families, or maybe people would be more willing to pay because of social pressure.
One morning in the field was enough to show that there’s probably not a tidy answer. In the summer of 2009, Jake went with staff from the IPA office in Busia to talk with some of the “refillers,” who had volunteered to maintain the chlorine dispensers by checking regularly that they had enough WaterGuard and reporting problems to IPA staff.
The first dispenser they visited was in a rural area. They drove out of town, turning after some time into what looked like a wall of maize but was actually a narrow path between two fields. Soon they came to a cluster of five small mud buildings, where they parked. Then they were led out of the compound, down a narrow dirt path, and onto a gently sloping hillside where young maize and sorghum grew shoulder-high on either side. They came upon a spring, next to which a chlorine dispenser had been set in the ground on a pole.
After a few minutes, a woman in a bright blue polyester blouse emblazoned with a wild silver design came down the path and introduced herself as Beatrice, the refiller. She explained that she had been nominated for the position because sh
e lives close to the spring, is literate, and has a mobile phone. She said that, since the advent of the dispenser, the community had put together some collective initiatives of its own, namely chickenraising groups. Everyone could agree the dispenser had done a lot of good, both health- and economics-wise, the latter because of the money saved on treating cases of diarrhea and typhoid.
When Jake asked where she saw the dispenser program going in the future, Beatrice said that IPA should continue to provide free WaterGuard for a long time so that they could enjoy their health. But, he asked, what if the subsidy stopped? Would the community pull together and pay full price to keep the dispenser stocked? She seemed doubtful. That would mean hitting up people for contributions, a job nobody was keen on doing. No, Beatrice sighed, an end to the subsidy would probably mean an end to the dispenser.
About a half hour later, and just a mile or two back on the main road toward town, Agnes strode through the metal gate of her compound, confident and self-possessed, and introduced herself. She was another refiller. Agnes smiled a lot, was not timid, and had funky teeth. She told Jake that she and her husband, the owner of the compound and landlord to a number of tenants, had constructed a well in the compound’s courtyard some years ago. They offered the water for free to their tenants, and sold it to other households in the neighborhood. As it turned out, initially the water was no good. People fell sick—and complained—regularly. Agnes had tried disinfecting the well by sprinkling chlorine crystals into it, and she had tried advising her clients to boil the water, but neither tactic worked.
When IPA came along with the chlorine dispenser, she saw an immediate improvement. There was less sickness, both in and outside her immediate family. Her tenants and clients were happier, and they became more numerous. When Jake met her she was serving twenty-three neighborhood compounds besides her own.
Asked what she would do if they began charging for the WaterGuard, Agnes didn’t hesitate. She knew the chlorine was a worthwhile expense, and she would be prepared to go on buying it. Never mind her tenants’ and clients’ satisfaction; the money she saved on treatment for her family members alone, she said, would justify paying full price.
No Size Fits All
Beatrice and Agnes, living just a couple miles apart, were in different worlds as far as the chlorine dispensers were concerned. Agnes would have been ready to press on without subsidy on the spot; Beatrice would have needed to do some serious organizing and political maneuvering. In the same vein, the problems posed by malaria are as different as Davis Charway and the girls of Lunyofu Primary School.
The solutions are likely to be just as varied.
Rigorous testing showed that pregnant Kenyans would likely be better served by free distribution of bed nets than by cost-sharing, but that doesn’t mean we should abandon market-based solutions altogether. They have their place as sure as Agnes does; what we need to do is figure out when and where the different solutions work, so we can apply them just when conditions are right. We can’t write global prescriptions about specific program designs until we do that figuring. And that’s exactly why IPA continues to test different ways of running the chlorine dispenser program, both in and beyond Busia.
As far as general directives go, I agree with Sachs that nobody should be priced out of protection, and with Easterly that resources—and the will it takes to mobilize them—are both too valuable and too scarce to waste. What’s essential is that we pursue effectiveness in our campaigns with the same vigor and tenacity as we pursue the ultimate goal of eradication. If we do any less we will fail on both accounts.
11
TO MATE
The Naked Truth
While I was studying for my Ph.D., Paul Gertler, a professor at UC Berkeley (who we saw in the last chapter working with the Mexican government to evaluate Progresa), visited MIT for a day to give a seminar. This is a common process at many universities, for an outside professor to come and give a ninety-minute talk based on one of his or her current research papers. It’s a way for economists to keep current and get feedback on projects or papers before they are published. Aside from the talk, the visiting professor usually schedules appointments throughout the day to meet with faculty and, occasionally, graduate students.
I had never had such a meeting before, but Esther, my adviser, e-mailed me and said, “Get on Paul’s schedule.” I did. As I walked to our appointment I was a bit nervous, having no idea how these conversations really proceeded. Paul did, of course. He asked me what I was doing, and I told him I was about to leave for South Africa to set up an experiment to measure the impact of microcredit. (The experiment totally failed, and I learned an early and now-obvious lesson: The staff of the partner organization really should want to be evaluated, otherwise they will find any one of a hundred ways to ruin an RCT.)
He said—and I’m fairly certain this is an exact quote—“Great! While you are there, how about finding out the price of whores with and without condoms?”
I laughed but never did follow up on this. I didn’t fully absorb that hookers’ prices could be a topic for serious economic research.
I was wrong. Turns out Paul wasn’t joking. He was talking about a new way of looking at an issue that affects almost every person on Earth—man and woman; black, white, and brown; rich and poor alike. It’s sex.
Sex is a great equalizer—first because practically everybody does it, but more important because it strips us down to our naked selves. I’m not talking about clothes. Sex is a primal activity. It’s in our biology. In some sense, we are most definitively Humans—and most definitively not Econs—when we’re doing the deed. In that space of urge, impulse, and heavy breathing, a lot fades into the background. There are really not so many differences between rich and poor in between the sheets.
Which helps to explain why we all make mistakes about protection, no matter where we live or how well-off we are. In the heat of the moment, the probability of contracting a disease or having an unwanted pregnancy—the downside of unprotected sex—is not salient, if it is known at all; the upside, meanwhile, is beating down the door. The throes of passion simply are not the best setting to run through a cost-benefit analysis.
But someone who faces the decision regularly, and who doesn’t do it for the passion, might learn to do better. After all, practice makes perfect, right? If anyone should know, it’s the professionals. Prostitutes, that is.
So that’s what Paul was really talking about when he asked me to talk to hookers in South Africa. Fortunately he marched ahead with his idea even though I missed my chance to participate. He got together with Manisha Shah, also of UC Berkeley, and Stefano Bertozzi of Mexico’s National Institute of Public Health, and went to Mexico City. They leveraged the expertise of local pimps, police officers, taxi drivers, medical workers, and bar owners to find and interview about a thousand sex workers near Mexico City in 2001. They asked the women about the details of their last few “transactions.” It turned out the women knew all about safe and unsafe sex—their sliding prices proved it.
Condom use was the standard, but it wasn’t a hard-and-fast rule (no pun intended). According to the interviews, prostitutes protected themselves nine times out of ten. And the tenth time, when they agreed to go without, they charged extra—about 23 percent more, on average. Here was the evidence that they knew about the relevant hazards: They were demanding more money for taking additional risks.
Professional as they were, these women didn’t have anything as simple or direct as a menu with options and prices. Agreements were born out of a bargaining process.
Most of the time, the prostitute suggested he use a condom and the john agreed. In the other cases, the john usually asked to have unprotected sex. Knowing that she had something he wanted, the prostitute now had a chance to extract something in return. It reflects the bargaining nature of the transaction that women identified as “very attractive” made the most from this opportunity, capturing a 47 percent premium for unprotected sex—more t
han double the average.
So the prostitutes knew about the risks they were taking. They also evidently knew about supply and demand. All other things equal, they would have accepted a lower price in return for protected sex; but when they discovered that the customer actually preferred that, too, they were more than happy to make him pay extra for it. In cases where the john was the one to suggest condom use, prostitutes managed to extract an 8 percent premium for agreeing. That’s just plain savvy.
Bad Information, Bad Choices
So that’s the view from the trenches, where these battles are fought night in and night out. But, of course, most sex on Earth is not had by sex workers; so to address issues of reproductive health more generally, we need to see what the amateurs are doing. Do they know the risks of unsafe sex? If so, how are they being compensated for those risks when they take them? And if not, could learning the relevant information lead them to behave differently?
On the whole, public knowledge about sexual health in developing countries is limited. It suffers both from a lack of information, and also from a frightening amount of misinformation . I’m not just talking about the fine points here. Dr. Manto Tshabalala-Msimang, South Africa’s minister of health from 1999 to 2008, was famously off the mark on AIDS. During her tenure she warned against the use of antiretroviral drugs (she contended they were toxic) and instead advocated a strictly nutritional approach to prevention and treatment, earning the nickname “Dr. Beetroot.”
She spoke with conviction, but her program was straight out of medieval times: “Shall I repeat garlic, shall I talk about beetroot, shall I talk about lemon? These delay the development of HIV to AIDS-defining conditions, and that’s the truth.”
