Happiness and economists' obsession with income

This piece explores why economists are obsessed with income, why they are so challenged by the notion that income and happiness don't have a linear relationship, and why they should chill out because happiness and the Easterlin paradox really aren't a challenge to the economic way of thinking. 

Fundamental models in economics suggest that as someone’s income rises their utility will also increase. Empirical verification suggests that this is mostly correct, as are certain derived assumptions such as ‘people will choose more over less income’. A lot of these checks have used ‘revealed preferences’ to verify the claim (for very good reasons that I won’t go into here). That is, they note that people act in such a way as to maximise their income, ergo: they must prefer more income to less because income makes them happy. As a consequence of this kind of thinking, economics built a monolithic body of work that assumes that higher incomes make people happier and thereby enshrines higher incomes as the key goal of development and economic policy.

Then along came Easterlin. Starting in the late 70s Easterlin started to eschew the revealed preferences approach to instead ask people directly whether they were happy (which he took to be synonymous with utility). He found something striking: within countries, wealth seemed to predict happiness quite strongly, but across countries the effect was weak. Both across and within countries there appeared to be diminishing returns to happiness from income. Rather than a linear relationship, the more money you had the more you needed to bring about a further increase in your happiness. This collection of observations came to be known as the Easterlin Paradox.

Easterlin proposed and then verified a few ideas that might explain the paradox. The first was relative income effects: absolute income is important, but we like to be richer than our peers. If we are wealthy by historic standards (a modern day plumber has wealth far exceeding that of most Ancient Kings) but poor by the standards of our contemporary society, we will feel unhappy. The fact that capitalism’s dynamics means that fast growth and rising inequality are linked exacerbates this issue. The second was adaptation—we get used to money. The most famous study of this involved lottery winners and found that after two years they were back to the level of happiness they were at before they won big. A third factor was the opportunity cost of income, specifically leisure. To earn a high income often took a lot of time, which didn’t leave much left over to enjoy your money. A final factor was instability: people don’t like rapid change, but rapid change is exactly what you get during periods of fast income growth.   

The economics profession at first largely dismissed this work as bunkum, but once its veracity was no longer disputed, economics had a bit of a flip out. To this day economists continue to investigate, in painstaking detail, the relationship between income and happiness.  Just recently, an amazingly elite collection of economists got together in Paris to discuss research into happiness and income arising out of new data on China’s dramatic growth experience (book review coming in a few months once I get it published). They confirmed everything we already knew and added extremely little to our existing knowledge. What a waste of resources.

It seems that the paradox undermines the entire identity of some economists. A couple of weeks back, a rather prominent Australian economist told me, quoting a 2008 paper by Samuelson and Wolfers, that ‘the Easterlin Paradox is dead’. I was struck by the vehemence with which he said it. It seemed that it was of the utmost importance to him that this repugnant idea be buried. He made this statement with vehemence despite the fact that I had made no indication that I thought happiness research undermined economics. An added irony is that the S&W paper does not at all undermine the Easterlin paradox. It shows that income and happiness have a linear relationship on the log scale, which means precisely that there are diminishing marginal returns to happiness from income (see graph at top of page). The S&W paper undermines the idea that there is a satiation point, but only people with an anti-capitalist agenda ever argued that seriously.

Economics needs to do two things. First, it needs to chill out about the paradox. The paradox has almost no implications for the foundation theories and models of economics, as I will explain below. Second, it needs to stop investigating income because the marginal returns to those investigations is approaching nill. Instead, it needs to get together with psychologists and philosophers and build a richer concept of utility that can be employed where appropriate (there aren’t many places) to produce better theories (for example, of the objectives of public policy in rich countries).

Let’s consider the implications of happiness research for the fundamental ideas in economics. As I already said, the S&W paper shows that income and happiness have a linear relationship on the log scale. This holds at the country, state and individual level. Therefore, the basic idea that people like money and will act to increase their incomes in the absence of any opportunity costs because they will thereby increase their utility holds. Microeconomic theory robust is therefore robust to the Easterlin paradox. It can only break down under one condition: where the opportunity cost of pursuing income in terms of utility from other sources outweighs the returns from pursuing further income. The most obvious case is the marginal cost to leisure of increasing income by working more (another is environmental degradation). Both income and leisure are sources of utility (as modelled in the most basic individual labour supply model), both suffer from diminishing returns, and people will assess this trade-off at the margin. This insight then, which we can arrive at from the paradox, is nothing new to economics, even old-school neoclassical economics. It is only new and especially only a challenge if you literally see utility and income as synonymous. If that is the case then you are an idiot. Income is measured in dollars, utility is measured in utils. They aren’t the same thing. Economics and economists know this. Sometimes development is measured only in terms of dollars, but since Amartya Sen dropped his ‘development as freedom’ idea in the 80s (and even earlier when we started developing the human development index) we haven’t taken that approach. Indeed, I think anyone with a brain was only ever using income merely as a convenient proxy for measuring development progress, which is entirely justifed by the graph of the paradox above, where income initially has massive impacts on utility. The models underlying this approach to development economics sought to increase utility, not income. 

This has implications for people trying to replace GDP growth with other metrics, and for people trying to undermine the basis of economic thought. Let’s start with the latter. Income and utility dynamics will still hold when people buy butter at the supermarket, when firms try to maximise profits by engaging in strategic behaviour, when people bid for public goods etc. Almost all of the models economics uses will hold in the face of the paradox. They may need to be modified in practice when confronted by a complex scenario (especially those involving large sums of money vs. quantities of some other utility enhancing thing like environmental goods), but they do not need to be modified fundamentally. Certainly there are exceptions to the rule, but these are rare and economics has an entire branch (behavioural economics) dedicated to their study, so chill. 

With regards to other metrics, GDP growth will always be critical for employment growth, and employment is one of the strongest predictors of happiness, so there isn't much point bringing in happiness and crowding the dashboard, especially when GDP growth is clear whereas fluctuations in happiness are anything but. GDP growth still has strong effects on happiness at the lower end of the income scale, so it is still a critical metric of development. At the higher end we are reaching the flat part of the curve, so trade-offs are more live, but these don’t come up very often and are very difficult to quantify because every individual has a different happiness scale (i.e. 7/10 might mean something completely different to me than it does for you). Sometimes we can do it. For example, we can back out a dollar valuation by asking people how happy certain changes, some with dollar values and others without, would make them, and then compute the dollar value of non-dollar denominated changes via the value of their happiness change. This could be useful for public and environmental goods, for example. But at the macro level there is rarely a conflict between GDP growth and other sensible policy outcomes that isn’t currently already decided in favour of the other policy issues e.g. emissions trading schemes and taxation policy. We don’t need happiness to make these calls.