Are you sure that more money would make you happy?
- Author: Dr Carlos Cortinhas
- Date: 21 May 2015
- Copyright: Image appears courtesy of Getty Images
Conventional wisdom has always maintained that money can’t buy you happiness (1). But science, as it turns out, would seem to say otherwise.
The data for happiness studies invariably come from large scale surveys such as the World Values Survey or the Gallup World Poll, which include questions such as “All things considered, how satisfied are you with your life these days?” or “Taking all things together, how would you say things are these days—would you say you’re very happy, quite happy, not very happy or not at all happy?” The answers are then transformed to a scale that allows for international comparisons and for the study of changing trends over time.
According to the widely cited work of Richard Easterlin, money does make you happy, but only up to a certain point. In his now famous 1974 article (Easterlin (1974), Easterlin postulated that although high income correlates with happiness in within-country comparisons at any given time, that relationship was not clear when comparing happiness differences across countries at a given time. Furthermore, he also asserted that in the long-run, an increasing average income did not raise the average well-being within a country. These contradictory results have become known as the “Easterlin Paradox” and have generated much interest in the topic, creating a whole strand of ‘happiness literature’ within not just economics but also in psychology.
The Easterlin paradox is so called for two reasons: First, because the (aggregate) long-run trend of average happiness seems to be flat even when the average income increases over time. Figure 1 illustrates this point by using data for the United States between 1972 and 2014. Although real per capita GDP increases steadily over that time period, mean happiness is virtually unchanged (2). Second, this apparent lack of an association between happiness and income over time is in contradiction to the robust and positive association between income and happiness at a given time both at individual level but also across countries (using cross-sectional data) (3).
Figure 1: Happiness and GDP per capita, United States, 1972-2014
Source: Data on mean happiness was computed from the General Social Survey (GSS) in the same fashion as Easterly (2005) and data on per capita GDP was obtained from the World Bank’s World Development Indicators (2014).
A popular explanation for this result is that wealth beyond a certain amount might not make us happier: once we’ve achieved a reasonable degree of financial security (this is usually allocated at an annual income of roughly $15,000 per year) our basic needs are met and our sense of well-being does not improve as income rises. The argument for the existence of a ‘satiation point’ is claimed widely and seems to be based on the fact that when analysing the relationship between income per capita and well-being across countries the relationship is positive and concave as depicted in Figure 2 (4).
Figure 2: Well Being and GDP per capita across OECD countries
Source: Data on well-being comes from the OECD’s Better Life index and data on per capita GDP was obtained from the World Bank’s World Development Indicators (2014).
The quadratic regression fit (in blue in Figure 2) clearly shows a concave and positive relationship between well-being and average income and raises the hypothesis that “the elasticity of happiness with regards to income could eventually become indistinguishable from zero” (Senik (2014), p. 100), suggesting that the positive relationship between happiness and average income could break down at some high level of income.
A recent paper by Betsey Stevenson and Justin Wolfers in the prestigious American Economic Review (Stevenson and Wolfers (2013)) suggests that Easterlin and his followers got it wrong, or at least in part. After pouring over data from 140 countries, they concluded that rich people are happier than poor people, people in rich countries are happier than people in poor countries and when countries get richer their people tend to get happier too (5). They also tested for a cut-off point at $8,000, $15,000 and $25,000 but found no evidence of a significant break in “either the happiness-income relationship, or the life satisfaction-income relationship, even at annual incomes up to half a million dollars” (Stevenson and Wolfers (2013), p. 602). This result is in line with the results of other studies (e.g. Deaton (2008)), and suggests that the relationship between happiness and income is log-linear, positive, with diminishing returns and does not appear to confirm the existence of a satiation point. Perhaps then, the old adage that ‘the more you get, the more you want’, is true of us all.
The two main explanations for the Easterlin paradox in the existing literature (See Senik (2014) for a recent survey) according to Clark et al (2008) are that (i) income may be evaluated relative to others (social comparison) or (ii) to oneself in the past (habituation). The first is the old idea that people are concerned only with the wealth of their next door neighbours. So, if you are keeping up with – or better yet, surpassing – the Jones’s, you are happy. By that logic, people in impoverished countries could be happy if they were just a little bit better off than those around them. By contrast, people in rich countries will never be happy as long as their neighbours are doing as well as they are. The second explanation is that people’s aspirations rise along with their actual living conditions and as their well-being depends on the gap between their actual situation and their aspirations, growth in income will never allow them to fill the gap.
The jury still seems to be out on whether these two hypotheses explain the Easterlin Paradox. A number of great challenges remain that may prolong the controversy for years to come. First, the data on happiness tends to be measured in a bounded scale, which makes it “difficult to expect a continuous rise of average happiness over the long run” (Senik (2014), p. 103), and makes it likely that people’s interpretation of the different categories of happiness will not be constant over time. Second, the way aggregate measures of happiness are computed are subjective, hard to measure and appear to be “extremely insensitive indicators” (Johns and Ormerod (2007), p. 9). In fact, aggregate happiness measures are known to be uncorrelated with a number of the important factors that one would expect to strongly affect happiness, such as life expectancy, government spending or unemployment (6). Finally, researchers need better quality data on both individual and aggregate happiness/well-being, “particularly panel data, which allows for the correction of unobserved personality traits and correlated measurement errors, as well as for better determining the direction of causality” (7).
So, it may be that ‘money can’t buy you love’, but it seems that it can certainly buy you happiness.
Clark, A.E., Frijters, P. and Shields, M. (2008), ‘Relative Income, Happiness, and Utility: An Explanation for the Easterlin Paradox and Other Puzzles’, Journal of Economic Literature, 46(1): 95-144.
Deaton, A. (2008), ‘Income, Health and Well-being around the World: Evidence from the Gallup World Poll’, Journal of Economic Perspectives, 22: 53–72.
Easterlin, R. A. (1974), ‘Does Economic Growth Improve the Human Lot? Some Empirical Evidence.’ In Nations and Households in Economic Growth: Essays in Honor of Moses Abramovitz, edited by Paul A. David and Melvin W. Reder. New York: Academic Press.
-------- (2005), ‘Feeding the Illusion of Growth and Happiness: A Reply to Hagerty and Veenhoven’, Social Indicators Research, 74: 429–443.
Graham, C. (2008), ‘Economics of Happiness,’ The New Palgrave Dictionary of Economics (online), 2nd Edition. Eds. Durlauf, S. N. and Blume, L. E., Palgrave Macmillan.
Inglehart, R., Foa, R., Peterson, C., and Welzel, C. (2008), ‘Development, Freedom, and Rising Happiness: A Global Perspective (1981–2007)’, Perspectives on Psychological Science, 3(4): 264–85.
Johns, H. and Ormerod, P. (2007) ‘Happiness, Economics and Public Policy’, Institute of Economic Affairs, Research Monograph 62.
Senik, C. (2014), ‘Wealth and Happiness’, Oxford Review of Economic Policy, 30(1): 92-108.
Sacks, W.D., Stevenson, B. and Wolfers, J., (2010), “Subjective Well-Being, Income, Economic Development and Growth”, NBER Working Paper 16441.
Stevenson, B. and Wolfers, J., (2013), “Subjective Well-Being and Income: Is There Evidence of Satiation?”, American Economic Review: Papers & Proceedings, 103(3): 598-604.
1. In this article the (subjective) terms happiness, well-being and life-satisfaction will be used interchangeably. Although not the same, the results of previous studies do not seem to vary significantly whichever term is used.
2. In fact, the correlation between the two variables appears to be negative and statistically significant (r =-0.444 with a p-value of 0.025). Easterlin (2005) presented the results of a OLS regression of mean happiness (H) on real GDP per capita (Y) for the period 1972-2002 and the results showed the coefficient of Y to be negative and equal to -1.05e-6 (standard error = -0.81). A regression for the longer period 1972-2014 presented in figure 1 shows the coefficient to be -2.09e-06 but this time it is found to be statistically significant at the usual 5% level (p-value = 0.0158).
3. Although Easterlin (1974) stated that “whether any such positive association [between income and happiness] exists among countries at a given time remains uncertain” (p. 30), this association is now well established in the happiness literature.
4. This is a well-known result and similar graphs have been reported in a number of past papers, including Deaton (2008) and Inglehart et al (2008).
5. The latter point is actually explored in Sacks, Stevenson and Wolfers (2010).
6. Johns and Ormerod (2007) also report that “relationships between happiness and crime appear, tentatively, to throw up a positive correlation!”, (p. 9).
7. Graham (2008). The typical happiness equations have the standard form:W_it=α+βx_it+ϵ_it with W being the reported well-being of individual i at time t, and x is a vector of known variables, which typically include socio-demographic and socioeconomic characteristics. These types of equations rely on the assumption that X causes Y and do not attempt to capture the possibility of Y also causing X. A Granger causality test on the data presented in figure 1 above, however, does not support the hypothesis of dual causality as we can reject the null that ‘Income does not Granger cause Happiness’ at the 5% level of significance but not the reverse case.
Bio: Dr Carlos Cortinhas is Associate Professor of Economics and Director of Education (Economics) at the University of Exeter Business School.