Geek Heresy (15 page)

Critics have noted, however, that not everything can be tested with an RCT.
¹⁶
Some questions are inherently harder to explore with the technique. In practice, certain kinds of questions are systematically neglected, and some programs never receive the imprimatur of an RCT. The methodology excels at comparing packaged interventions against one another. If you want to know whether camera-based monitoring or a midday meal is more likely to help students do better on tests, an RCT is great. But comparing packaged interventions against non-packageable alternatives is much more difficult.

Here’s why. In order to run an effective RCT, researchers must carefully ensure a number of conditions. The experimental group must be selected so as to be without bias (hence, randomization). The control group must remain unaffected by the experiment. The intervention must be implemented exactly as defined. Data must be collected. And all of this must be done to exacting standards. To guarantee these requirements, researchers prefer to run RCTs in partnership with capable organizations such as Seva Mandir – that is, organizations capable of following onerous instructions. But by working with competent implementers, the experiments necessarily occur in special conditions that aren’t common elsewhere.

When the results of an experiment can’t be generalized beyond their immediate context, scientists call it a problem of “external validity.” Duflo and other randomistas say external validity can come by repeating studies in different contexts. Technically, that’s true, but it’s not easy to do for testing the value of capable implementation. The very thing you need for a good controlled trial would have to be suspended for a part of the experiment.
¹⁷
This is the RCT’s Achilles’ heel.

You can hear reverberations of Rossi’s Iron Law of Evaluation. There is, he wrote, a world of difference “between running a program on a small scale with highly skilled and very devoted personnel and running a program with the lesser skilled and less devoted personnel.”
¹⁸
The Law of Amplification is also in effect: Packaged interventions work in proportion to the capacity brought to bear. If you conduct experiments entirely in air where institutional capacity can be taken for granted, then you don’t know how the intervention would run under water, where capacity might not be present.

To be clear, I’m not against RCTs in and of themselves. I have applied RCT methodology in my own research. I believe that packaged interventions should be verified by RCTs more often. I’m on the board of JPAL’s sister organization, Innovations for Poverty Action, a nonprofit that also runs RCTs and whose work I deeply respect and support.

But if RCTs are an essential tool of decision-making for social policy, they still need to be placed within a larger conceptual framework. If
a world-renowned economist and careful experimentalist such as Duflo can’t avoid the pathologies of packaged interventions, it’s not clear that other researchers running RCTs can either. A single methodology cannot be the sole paradigm for determining what’s right for social change. The problem is not RCTs themselves as much as careless interpretation of their results.
¹⁹
An RCT is just one good tool in the toolbox of program evaluation. Good tools are important, but it’s even more important that architects and artisans use the right combination of tools in the right way for each decision-making task.

Eradicating Poverty Through Profits?

Another fashionable trend sees practitioners applying for-profit business approaches to social causes. The idea was put forth seductively by C. K. Prahalad, a professor of business at the University of Michigan. In his 2004 book
The Fortune at the Bottom of the Pyramid
, he wrote that the 4 billion people in the world who live on less than $2 a day could be enriched if they were viewed as a business opportunity.
²⁰
Prahalad’s motto is captured in the book’s subtitle:
Eradicating Poverty Through Profits
.

According to Prahalad, governments and nonprofits have been going about things all wrong, especially when it comes to poverty. They commit two sins hateful to any business: First, they don’t cover their own costs; second, they’re unable to reach large numbers of people. “Charity might feel good,” Prahalad observed, “but it rarely solves the problem in a scalable and sustainable fashion.”
²¹

Prahalad’s proposal was to sell low-cost products and services with the goal of fostering poor people’s “capacity to consume.” These people, he said, were “value-conscious consumers,” and affordable consumption opportunities served them by “offering them choices and encouraging self-esteem.”
²²
Prahalad’s rhetoric rehashed the themes of mainstream economics: individual choice, market freedom, and poverty alleviation without politics. He promised that there were “sustainable win-win scenarios where the poor are actively engaged and, at the same time, the companies providing products and services to them are profitable.”
²³

If this sounds too good to be true, that’s because, by and large, it is. Although Prahalad’s book is crammed with case studies, none of the examples actually shows a firm increasing its profit margins while improving the lives of poor customers. Aneel Karnani, another University of Michigan professor, took a critical look at Prahalad’s nine case studies and found that four were companies catering primarily to middle-income consumers, two were nonprofits sustained in part by public funding, and two were projects that hadn’t seen any profit. Only one was a for-profit corporation that made money selling to poor consumers.
²⁴

That last case is worth examining, because, as much as it seems to fit Prahalad’s thesis, it still comes up short. The company in question was the Unilever subsidiary Hindustan Lever Limited (HLL), and the product was soap, possibly Prahalad’s most widely cited example. Diarrhea is deadly for many children in India, but it can often be prevented by hand washing. So HLL sold Lifebuoy soap in small sachets that were affordably priced for “bottom of the pyramid” consumers.

It’s one thing to sell soap, though, and it’s another to instill hand-washing habits in people who’ve never learned the germ theory of disease. HLL tried two approaches. It ran a marketing campaign designed by the Madison Avenue giant Ogilvy & Mather, and it established a public-private partnership with the Indian government. Together, HLL and the state ran a multimillion-dollar, multi-platform campaign to convince people to wash their hands. Which worked better, the corporate marketing scheme or the public-private one? By Prahalad’s own admission, “although scalability seems to be greater with the [public-private partnership], benefits to corporate sales lie with” paid marketing efforts. In other words, while profits were best served by a corporate campaign, greater health impact came by way of taxpayer-funded social marketing.
²⁵

So none of the nine cases Prahalad held out as paragons of “eradicating poverty with profits” were any such thing. Either they didn’t serve the poorest people, or they didn’t make much of a profit. This shouldn’t come as a surprise. We already know how hard it is for the private sector to deliver services to poor populations on its own. For instance, despite many experiments by determined entrepreneurs, there
isn’t a system in the world that delivers universal health care or universal education without at least some support from the public sector. Private efforts either fail to generate enough revenue to keep going, or have to swim upstream to richer waters.

What’s most twisted about Prahalad’s logic is the suggestion that poor people can somehow consume their way out of poverty. As if their poverty were a function of what they can buy, not what they can earn. Karnani more sensibly argued that the best way to support poor people was to help them become higher-income producers, not consumers.
²⁶
Well-paid employment makes people richer. Godfather of microfinance Muhammad Yunus has come to a similar conclusion and urges social activists to start what he calls “social businesses.”
²⁷
They differ from Prahalad in seeking to employ low-income workers, not just sell to them.

Asocial Enterprises

Prahalad has had some influence in corporate circles, but it’s a variation of his idea called “social enterprise” that has really taken off. In business schools, engineering departments, and venture capital firms, social enterprises – start-up businesses that try to serve a social good through a viable business – are all the rage. Social entrepreneurs model themselves on the Steve Jobses and Mark Zuckerbergs of the world, not realizing that successful businesses are successful because they have carefully chosen their customers, not because they have a foolproof Midas touch. Apple is a profitable company not just because it designs superior products but also because it chooses the world’s wealthiest people as its market. It would hardly survive if it were constrained to selling $400 iPhones to individuals who earn less than that in a year. Even ad-based business models only work if advertisers are willing to pay for ads, which they won’t do if the eyeballs belong to people who lack disposable cash.

Toms Shoes has found a way around this problem. Its founder, Blake Mycoskie, is considered a social-enterprise pioneer. Toms is a for-profit company whose marketing strategy – touted in boldface at the top of its website – is “one for one”: For every pair of shoes sold, Toms donates a pair to a “person in need.”
²⁸
The firm is wildly successful.
Since its founding in 2006, it has handed out over 10 million shoes, which implies cumulative revenue of about half a billion dollars. It has recently expanded into eyeglasses with the same one-for-one promise. Bill Clinton once introduced Mycoskie – whose rakish good looks recall Jim Morrison – as “one of the most interesting entrepreneurs [I’ve] ever met.”
²⁹

Despite its performance, Toms has also come under scathing criticism. Some say that giving away shoes to poor communities stunts local economies and perpetuates a culture of dependence. Others note that Toms could redirect its giving to something more lasting than straw-and-canvas shoes. Even if their gifts cost just $5 each to make and transport, that’s still $50 million that could have been better spent. Still others find Toms to be little more than a sweatshop, exploiting cheap Chinese labor and photos of barefoot children for hefty profits.
³⁰

There is some truth to these points, but the real problem is deeper and subtler. I suspect Mycoskie’s critics wouldn’t be mollified that he has responded to them by opening factories in countries where he gives away shoes.
³¹
There is something more insidious at work.

Though Toms trumpets its commitment to addressing “hardships faced by children growing up without shoes,” it’s not very open about what customers are actually paying for. Toms doesn’t disclose its financial statements. What we do know is this: Thanks to a recent agreement to sell 50 percent of his stake to Bain Capital, Mycoskie stands to gain as much as $300 million on top of whatever he’s paid himself so far as sole owner and CEO.
³²

Imagine if you made a $50 donation to a nonprofit, and out of that, say, $10 went to the cause while $10 came back to you as a thank-you gift and $30 facilitated the executive director’s multimillion-dollar bonus. You wouldn’t stand for it. Yet this is effectively the Toms “business model.” And from it, the founder not only richly profits but then is weirdly hailed as a social activist’s hero. Something doesn’t seem quite right, and a comment made by a principal at Bain Capital – the same company that cast a dark shadow on Mitt Romney’s presidential campaign – provides a clue why: “We believe that the one-to-one promise is fundamental to the brand.”
³³
Or, to put it another way, Toms is
successful because of a cynical marketing pitch catering to slacktivist consumers who equate generosity with purchasing two cheaply made pairs of shoes and giving one away.

In the end, Toms is a shoe company with a social responsibility arm. Mycoskie deserves credit for being a shrewd entrepreneur and spending some of his growing fortune on charity. Otherwise, though, Toms is not that different from Nike: They both sell overpriced shoes to brand-conscious customers, exploit cheap developing-world labor to pay their executives well, and spend a portion of their revenue on charitable causes (in Nike’s case, on the nonprofit Nike Foundation).

But Toms does one more thing: By misleadingly presenting itself as primarily interested in charity, the company diverts the goodwill of people who might otherwise engage more deeply in a cause. In what psychologists call moral self-licensing, people use past good deeds – even minor ones – to excuse future apathy.
³⁴
So there’s a good chance that many Toms customers skimp on more worthwhile efforts, something they probably wouldn’t do if they bought their shoes from Nike, which runs its own social responsibility initiative but with less self-congratulatory fanfare. The greater jeopardy, though, is a broader
societal
self-licensing: By playing up efforts like Toms, we as a society fool ourselves into believing that the world’s problems can be solved by enlightened consumerism.

That’s the problem with social-enterprise hyperbole in general. Its noisy buzz draws attention away from effective government and nonprofit approaches to social causes.
³⁵
Large aid agencies such as United States Agency for International Development (USAID) and the Ford Foundation have begun spending their precious funds on for-profit entities. They hope that their one-time donations will turn into everlasting economic engines that also churn out public goods. As Jonathan Franzen observed in his novel
The Corrections
, “The more patently satirical the promises, the lustier the influx of American capital.”
³⁶