This week’s FPD Chief Economist Talk featured Sendhil Mullainathan, Professor of Economics at Harvard University and co-founder of Ideas42, a non-profit that uses insights from behavioral economics to inform the design of products, innovations and policies. Sendhil is well known as one of the leading thinkers in behavioral economics and much of his research has focused on topics at the intersection of psychology and development economics, ranging from corruption in the allocation of drivers’ licenses to the role of psychology in the take-up of micro-finance and consumer loans — all very important issues that matter for what we do at the Bank and so Sendhil’s talk provided great food for thought!
So what exactly is behavioral economics? In simple terms it is the realization that very often people don’t act as rational cost benefit optimizers as traditional economics would have us believe. In fact, they rarely do. Most of our decisions are subject to a wide variety of behavioral biases and this has real consequences for economic outcomes: we are impatient, so we have trouble committing to save. We tend to be overly confident about our skills and ability to predict the future, and this can lead to some terrible investment decisions. We respond to how choices are framed and presented to us, so we go for the credit card that has the flashiest advertising blitz, rather than the card with the most reasonable terms.
Behavioral economics is by now a well established field: many studies have documented the importance of psychology in all kinds of economic decisions and the assertion that people are not always and everywhere rational in their economic behavior is unlikely to raise many economist eyebrows today. Much of Sendhil’s talk was therefore devoted to thinking about what’s next, and emphasizing the point that –given how much we know about the importance of psychology in economic decisions — it is about time that we do more to translate insights from behavioral economics into the design of better products and policies — “behavioral design” to use Sendhil’s term.
There is certainly no shortage of applications for this agenda, and some of the most interesting applications are to be found in the financial sector. A particularly good example is the design of commitment savings products: by and large people understand that having precautionary savings is a good thing and would tell you that “yes” they definitely intend to save if you asked them. In reality, there are many reasons why that intention might not turn into a reality –just like your plan to hit the gym every day of the week. The challenge of behavioral design is to figure out what causes this divergence between intention and outcome and design a product or policy that addresses the behavioral pattern. In the case of savings this could be as simple as an SMS reminder or as involved as a commitment savings product that ties your hands and deducts an automatic amount from your paycheck which you might be tempted to spend rather than save if you had the choice.
Another highly relevant application for behavioral design that has gotten much attention since the financial crisis is the area of consumer finance and consumer protection. It is worth noting that there is also a “dark side” of behavioral design: many financial products, such as payday loans and certain types of credit cards for example exploit behavioral biases of consumers —providers make money by betting on consumers’ impatience, inattention to fees and interest rates and tendencies to procrastinate. Smart behavioral design is probably as much about understanding which behavioral biases are likely to be exploited in the marketplace as figuring out how to design products and policies that are more consistent with how people behave and can remedy some of these issues.
What’s true for the design of products is also true for the design of development policies more broadly, where taking lessons from behavioral economics on board can significantly improve the odds that good intentions actually translate into good outcomes: anti-corruption policies are more effective if they not only punish but also reduce opportunities for corruption. Participation in retirement savings plans is higher if people are defaulted into enrolling into 401k plans rather than having to sign up on their own (a famous example from the early days of behavioral economics).
Finally, an interesting aside that got some airtime in the last part of Sendhil’s talk is that these ideas also should also make us think about how we approach impact evaluation. In many cases we measure the outcome of a policy intervention without much regard to the underlying mechanism that generates –or prevents— certain outcomes and patterns of behavior. An important point to take away from Sendhil’s pitch for behavioral design is that we would be well advised to pay more attention to diagnostics. Rather than simply measuring what comes out of the ‘black box’ of a development program, we would be well advised to test and understand models of human behavior that can inform the design of products and policies beyond the setting of one particular program or intervention.
Here’s a video link to Sendhil’s talk if you want to spend more time thinking outside the box.