In a study of soccer ball manufacturers in Pakistan, Eric Verhoogen of SIPA and Amit Khandelwal of Columbia Business School found that how workers are paid, and whether or not incentives are offered, can promote or stifle the adoption of a new technology.
Columbia News wrote about the study here.
Verhoogen, who is an associate professor of international and public affairs and economics, followed a somewhat unorthodox path to SIPA. “After college at Harvard, I was a high-school teacher in Los Angeles, and then I was a labor organizer in Minnesota and Ohio,” he told SIPA News. “Then tried my hand at journalism in Berkeley and at the Nation [magazine]. And then I started grad school at UMass Amherst and later transferred to Berkeley.” After earning his PhD in 2004, he came directly to SIPA, where he received tenure in 2010.
In a recent conversation with SIPA News, Verhoogen shared additional information about the soccer-ball study and discussed some other things he’s working on.
You found clearly that misaligned incentives can compromise the adoption of otherwise beneficial technology. Why is this significant, and what are the challenges inherent in a study like this?
It’s an interesting, important question for economic development and growth more generally. It’s also a hard question to study because it’s hard to observe technology use by manufacturing firms and rare to have information about the actual cost and benefits of technology.
This is partly because technologies vary a lot across firms, and particularly across sectors. And unlike other types of data, it’s hard to collect via survey—sometimes firms don’t want to share information. Economists have other methods to estimate productivity, but they’re all pretty indirect.
Why soccer balls? I realize it’s just a coincidence that we’re in the middle of the World Cup tournament.
With soccer-ball producers [in Sialkot, Pakistan], you have a pretty large number of firms, 135, producing a standardized product using similar technology. So the same basic production process is used by large and small firms alike. We thought we could introduce a new technology that would be useful to these guys, to producers, and focus on the diffusion process.
As the Columbia News story explains, Verhoogen and his team developed a fabric-cutting die that would enable producers to use fabric more efficiently, creating an opportunity to cut costs and increase profits.
Is it unusual that your team of researchers gave the manufacturers a technological advancement? Does it impact the study somehow?
In development economics, there’s been a broad trend over 15 or 20 years, of having more of these experimental interventions. There’s a large literature on technology adoption in agriculture where researchers share information about improved production processes. What’s more unusual about our study is that we are focusing on larger manufacturing firms and especially that we invented the technology we gave out.
So what happened when you introduced the new technology?
We gave the dies out in May 2012, and to be honest we were expecting very fast adoption. We were planning to focus on the diffusion process, seeing how the technology spread to firms we didn’t give it to.
We had evidence to indicate that the technology was working, that it was more efficient, but after 15 months only six firms had adopted the new technology. This was a puzzlingly low adoption rate, so we decided to write a paper about that.
The number-one reason firms didn’t adopt the new technology was that the employees were unwilling to use it. What became clear was that the cutters actually cutting the material are paid a piece rate per pentagon or hexagon. They want to go as fast as possible and don’t care about waste.
Our new technology slowed them down initially, certainly for the first month or two, and given their wage contract they have no incentive to adopt new technology. So we formulated this hypothesis that the misalignment of incentives was a key constraint to adoption.
We did a second experiment to probe this—we explained the misalignment and said we would pay a lump-sum bonus of one month’s salary, about $150, to the cutter if in one month he could demonstrate competence in the new technology.
The incentive program led to a 26 percent increase in probability of adoption of the new treatment. That such a small incentive targeted at workers could have a significant effect indicated to us that the misalignment of incentives is why the technology wasn’t being adopted.
Can you elaborate on the significance of your findings, and the study?
One piece of the big picture is that you have to have employee buy-in. Workers will only cooperate in the adoption of new technologies if they expect to gain—and if they don’t cooperate, they can effectively block it.
Also, by introducing the innovation we were able to actually observe the process and statistically distinguish between different hypotheses, as opposed to in case studies. This was a particularly clean setting, and we have a strong argument that the new technology is beneficial for essentially all firms.
I think this sort of thing happens all the time in many different settings. We happened to be able to observe it in one setting, but we think there are many incremental changes that could be made in different settings, and make a big difference.
Traditional economists sometimes say there can’t be a $100 bill on the sidewalk because if there were, someone would pick it up. We think this is a $100 bill on the sidewalk, but firms aren’t picking it up.
You’ve said you have a second paper in the works about diffusion—which is what you originally thought you’d be writing about. What can you tell us about it?
That paper is in progress, and the main question is what are the channels through which technology diffuses. Who do people talk to, who do they subcontract for, who are they related to, where are workers from, where have they worked in past, what religion are they, et cetera. We’re keeping track of information flow, keeping track of technology, which channels are most important for technology diffusion.
Whether ideas spill over is an important question for policy purposes. If the originating firms are not capturing the full benefit of an idea, they may invest less than the socially optimal amount to gain or acquire the idea. It’s classic market failure in the market for ideas. There’s less innovation because firms can’t reap the full benefit.
What’s next for Sialkot’s manufacturers?
One direction that we’re going, related to the idea of spillover, is to work with the government of the state of Punjab in Pakistan to develop an innovation fund. Individual firms may not have incentive to invest in innovation as much as government or other social planners, so there may be a natural role for governments to subsidize innovation. Coming out of this project, we want to convince officials to provide a fund for matching grants to firms for new machines, new processes, new products. Officials have allocated $5 million—but currently it’s stalled. We’re working on trying to get it going again.
You’re also the director of SIPA’s Center for Development Economics and Policy. How has CDEP been received since it formally launched in November 2013?
There’s a lot of enthusiasm about development economics at SIPA. There’s been a great response from students and faculty members, and also from people outside SIPA.
We have a couple of initiatives that are gaining momentum. One is a human capital initiative for human education and health issues—what leads someone to acquire education, what factors shape education and health, and what are the consequences of that for a labor market. Another is our firms and innovations initiative, which examines issues around industrial upgrading in developing countries— the question of why some countries can grow and thrive in world economy and some less so.
Another coincidence with the World Cup… you’re also pursuing research in Brazil.
In Brazil, with support from the President’s Global Innovation Fund, I have a project on the interaction between labor market regulation and innovation at the firm level.
The question is, how do firms respond to labor market regulation? Economists tend to think of the effect of labor regulation as uniformly negative, but we’re investigating whether there are less familiar but important positive effects on firm behavior.
For example, the minimum wage in Brazil has risen a lot. The minimum wage affects the relative cost of hiring different types of workers, more low-skill than high-skill. If you give firms incentive to upgrade the composition of their workforce, that may in turn induce them to use higher quality inputs, to produce higher quality outputs for sale to richer people in Brazil or richer export markets.
You’ve lived and worked in many different and interesting places. After almost 10 years here, how does SIPA measure up?
I very much like being at SIPA and teaching SIPA students because it keeps me grounded in the world. Our students have experience in the world and they’re planning to go back and be involved in things on the ground—I think it’s healthy and stimulating for me to be exposed to them and to be at a place that respects policy-oriented work. I got into this job to make the world a better place and I haven’t given up hope that that’s possible.