Were you happy yesterday? Policymakers want to know
Are you happy? Can you put a number on your happiness level between one and 10?
Those looking to assess social well-being, including a number of economists, are increasingly asking these questions. Yet do attempts at quantifying our happiness in this manner serve a useful function in an analysis of social conditions?
While questions over the value of economic growth as the sole measure of the success or failure of a society grow ever more insistent, the search for other ways to gauge progress has been moving apace. In 2007 the European Commission, European Parliament, Club of Rome, OECD and WWF held a conference on looking “Beyond GDP”. One of the things to emerge was a consensus that GDP was an inadequate tool for assessing well-being and that new, more timely indicators needed to be developed to map social trends.
Out of this was born the Office for National Statistics’ Measuring National Well-Being Programme designed to demonstrate a “commitment to developing wider measures of well-being so that government policies can be more tailored to the things that matter”. The first Annual Report was published last year and included the “National Well-being wheel of measures”. It revealed, among other things, that 75.9% of people surveyed reported a medium to high rating of satisfaction with their lives overall while a full 71.1% rated their happiness yesterday as medium or high.
If that all seems a little fluffy, that’s because it is – and intentionally so. Measuring so-called “soft” factors in order to complement existing hard data is precisely the point of Happiness Indices and their ilk. Of course in doing so happiness in inevitably shifted from a subjective feeling of elation (of varying intensities) to a statistical snapshot of national comfort levels – a sort of scientific experiment in which there is no control group.
From a policy perspective, however, I think this poses some deeper problems. As Arik Levinson pointed out in his recent NBER paper “Happiness, Behavioral Economics, and Public Policy”:
“When it comes to using happiness for public policy, there is conflict between these two central issues: habituation and the nature of the happiness question. It seems intuitive that life satisfaction should be the goal of policy, not momentary happiness. That suggests we want to identify equation based on average difference between locations over many years, not daily fluctuations…that would affect experience utility. But habituation suggests people's reported happiness may be a poor indicator of those long-term differences in life satisfaction. Moreover, if the goal of public policy is long-term assessments of welfare, we should identify equation using life-satisfaction rather than momentary happiness. But daily fluctuations in a public good like air pollution should not, in theory, affect people's overall assessment of their entire lives. Why would a person's life satisfaction be worse just because they are asked about it on a smoggy day?”
This helps explain why the ONS distinguishes between short-term happiness experienced yesterday and the general sense of overall well-being. But although distinguishing between the two makes methodological sense, in practice it is exceedingly difficult for individuals to assess their general sense of well-being as distinct to their emotional state in the here-and-now.
Behavioural economists refer to this problem as “projection bias”. That is, as Levinson defines it, people tend to misestimate their future desires based on current circumstances (and indeed are likely to skew their past desires in a similar manner). It is for this reason that economists have traditionally relied on revealed preferences – what people actually choose to spend their money on for example – to assess utility.
Of course, whether revealed preferences can really tell us about utility is a subject not without its contentious issues – see this post from Unlearning Economics for example. Yet I’m not convinced that happiness measurements really get us much further in understanding the dynamic preferences of individuals or groups within a society and how policy can help shape and be shaped by them.
To me, such measures appear vulnerable to sending the wrong signals to policymakers. Take David Spencer’s example of labour markets. During an economic downturn those who manage to keep their jobs while others around them lose theirs might report an increase in satisfaction even as their individual working conditions deteriorate. Policymakers could assume that this apparent uptick in happiness is a signal of positive developments in labour markets.
Even more fundamentally, the targeting of happiness could potentially slant policy away from reducing the misery of those who are suffering most in society. As a recent paper by Orsolya Lelkes argued, “whilst happiness is in large part an idiosyncratic thing, unhappiness is more closely correlated with social conditions”. Targeting minimising misery rather than maximising happiness could therefore be the more socially beneficial strategy.
So is measuring happiness the best way to achieve social harmony? I’m afraid I tend to side with Norbert Schwarz and Fritz Strack in their observation:
“What is being assessed, and how, seems too context dependent to provide reliable information about a population’s well-being, let alone information that can guide public policy.”