Beware of Happiness Statistics

A recurrent complaint leveled at economists is that income per capita is not a good measure of welfare. In its most refined forms, the complaint is that gross domestic product (GDP) per capita is a poor indicator of well-being. This is why there have been frequent calls to shift to other measures of well-being where incomes would be delegated to a supporting role. More often than not, the proposed replacement measure is “happiness” – a measure of how much individuals are satisfied. 

Iceland’s government is the latest to invoke the need to use a happiness statistic in a prominent role in its policy objectives. Is that a wise choice? On most grounds, the answer is in the negative. On the few grounds where happiness is indicative, it seems to cut against certain priors held by those who advocate their use instead of GDP per capita. 

The first thing that must be noted is that “happiness” measures have certain limitations. A question is asked to participants in a survey who must, on a scale that generally goes from 0 to 10 (where 10 is the happiest), indicate how happy they are. By definition, this means that there is an upper limit to the statistic (it cannot exceed the top value). However, in spite of this limitation, it seems that incomes do match relatively well with happiness measures: richer individuals tend to be happier than poorer individuals. 

The second thing is that “happiness” is not really what is being measured in most surveys. Instead, the attention goes to “life satisfaction,” which asks participants to grade the satisfaction with the lives they live. This measure is broader than the narrower emotional happiness. It implicitly includes elements about one’s projected life path and past life events. It is a measure that is very rich in terms of capturing what subjectively matters to individuals in terms of well-being (see here, here and here). That measure is quite strongly linked with income as Nobel laureates Angus Deaton and Daniel Kahneman noted. 

To be sure, that broader measure of “happiness” is not solely influenced by incomes. Health, education and cultural expectations also factor heavily in determining how individuals assess their own well-being. Yet, the relation between income and happiness is constant across countries: richer means greater life satisfaction. More importantly, when one plots changes in income against changes in reported life satisfaction or happiness, it is quite clear that the two statistics move up together. 

Thus, it seems that there is little to gain in shifting to happiness statistics as it seems to go up with income. Moreover, it is worth pointing out the obvious; that improvements in health and education (which also affect personal perceptions of life satisfaction) will also generate increases in income. 

Nevertheless, there might be elements to gain from giving attention to happiness statistics. However, it is not in the way that most expect. With the exception of the conservative political scientist Arthur Brooks, most of those who argue for a greater emphasis on happiness statistics tend to the left of the political spectrum. This is surprising because of the changes in happiness in inequality and income inequality, the latter of which these left-leaning individuals care considerably about. 

While there is a relationship between happiness and income whereby increases in income lead to increases in happiness, there is no particular reason for believing that the gains are constant. More precisely, an extra dollar of income may have larger effects for those lower on the happiness scale. Not only that, but the gains in satisfaction from an extra dollar are not necessarily constant over time.

Consider the following thought experiment to see the relevance of this point. You are a poor individual in 1900. For the sake of illustration, let us imagine you are in the bottom 10% of the population. You gain a 1% increase in your real income. What can you do with that income? What goods can you consume? Most likely more food, more clothing or better shelter. Basically, you expend more on necessities. Now imagine that you are in the same relative position (still in the bottom 10%) and you gain the same proportional gain in income. What can you do with it? You can spend on an iPhone that brings you in contact with the wider world. You can spend on video games that provide incredible entertainment. You can buy a subscription to a streaming service that gives you access to a wide array of entertainment options. In essence, the range of what can bring us joy thanks to extra income has increased over time. 

How does this speak to inequality? Well, consider that numerous studies have found that there is falling inequality in life satisfaction even though there is rising income inequality. True, the rich have been getting richer faster to a certain (debated) degree. However, the poor and the middling classes have not stagnated. They have benefitted from increases in income. For them, the increases in income mattered much more than for the richest. If we were to give a greater weight to happiness statistics, this would entail that we would need to care a bit less about income inequality. Yet, few make this point. Probably because it cuts against the priors held by many. 

Should there be more emphasis on happiness statistics? As an economist, it is hard to argue against having more data. However, each measure has its uses that come with limitations and flaws. If we are to advocate their use, we ought to be careful in considering what they say and do not say. 

Related Articles – Economic Growth