Today’s post is by Angus Deaton, Dwight D. Eisenhower Professor of Economics and International Affairs at the Woodrow Wilson School of Public and International Affairs and the Economics Department at Princeton University.
The happiness revolution is in full swing. Country statistical offices are beginning to collect data on wellbeing, the OECD has prepared a manual on how best to do it, and politicians are promising to focus, not just on economic growth and better material living standards, but on the much wider range of things that are important to people. This broadening of policy is much to be welcomed, as is the wider understanding of the inadequacies of GDP as an exclusive national target. And even if the data on wellbeing have their problems and their critics, a flawed but broader measure, such as wellbeing, can lead to better policy than a flawed but narrow one, such as income.
What are some of the things that we can do with wellbeing measures that cannot be done with standard, income based, policymaking? The concept of utility has been visible everywhere in economic texts and in economic thinking, but invisible in our databases. For goods and services that are bought and sold in markets, the invisibility of utility is not a problem because we can use prices as a guide to how much things matter to people. But that leaves uncovered the vast range of public goods and services—schools, hospitals, parks, airports, even the legal and political systems—that do not have prices. Conventional policy analysis sometimes works with “shadow” prices, the prices that would exist if these things were bought and sold. But these are hard to get right, and the numbers are often controversial. But if we can make utility visible, and measure it, then we can directly calculate how wellbeing responds to a park, or to a faster train line, and we can replace guesswork with measurement. Indeed, the “happiness” literature has estimated thousands of “happiness regressions” that show how self-reported wellbeing is affected by a wide range of circumstances, such as how income stacks up against time spent with friends, what value people place on lower pollution, or on living in a green and pleasant place.
There are many areas where people should make choices for themselves, and governments should either not interfere or should tread very lightly. Providing information is one example of light treading. If happiness regressions tell us that people are more satisfied with their lives when they choose to be clergymen, and not bar owners, or when they live in the mountains rather than in the plains, there is a role for making those facts known, perhaps endorsed by governments who might vet the quality of the research. This is what David Halpern calls “de-shrouding,” telling people how their choices might affect their wellbeing, as judged by the wellbeing of others who have already made those choices. Deshrouding would have no effect if people were fully informed, but people are often not very good at anticipating the long-term consequences of their choices. Happiness research has the potential to allow people to do better.
Even so, the results of happiness regressions must be interpreted with care. That a circumstance is correlated with life satisfaction does not automatically mean that it would be a good policy to promote more of it. Take the example of parents and children, where much of the literature finds that people who live with children are less satisfied with their lives than those who do not. Some of this is likely real, and there is good evidence that children bring a wide range of emotional experiences, including anger, stress, and worry, but also happiness and joy. Yet children do not come to people at random (blind storks?), but (usually) come to people who want them, just as those who are not parents are (usually) those who do not want children. Those who do and do not have children are different in many ways, including their tastes, not least their tastes for being parents. Comparing the wellbeing of parents and non-parents is therefore not a useful guide to young couples who are trying to decide whether or not to have children.
Consider also spatial differences in wellbeing. For example, we may discover that those who live near an airport have lower life satisfaction than those who do not. On a naïve view, people who bought houses under the flight path had no idea how much the noise would affect their lives. On a more sophisticated view, people understood that the noise was bad, but chose to live there because house prices were low, so that the noise actually provided them with an opportunity to live in a larger and more convenient house. In such a case, their low wellbeing reflects, not the noise, but that their incomes are too low to permit them to have a nice house that is also quiet. A policy of limiting noise or restricting landing hours may make such people worse off, because house prices rise, and a choice is no longer available to them. Of course, we can over-rationalize behavior, and people often do make mistakes, so that, in most cases, we have to deal with some mixture of the naïve and sophisticated.
An important question—still unanswered—is whether self-reported wellbeing is all that matters, which would simplify policymaking—or whether it is simply one thing that matters, so that people will trade it off for other good things. To illustrate the why this matters, consider that, in the U.S., once we make adjustment for educational and income differences, African Americans report substantially higher life evaluation than Whites. But African Americans suffer many deprivations beyond income and education; they are more likely to be in jail, to be banned from voting, to live in areas with low-quality schools and hospitals, to be unemployed, and to die young. Does their higher life evaluation mean that these other deprivations can be safely ignored? If not, and we refocus policy entirely towards wellbeing, we are making a mistake, just as when policy focused entirely on income.