The new government wellbeing statistics suggest that London is the most miserable place in the country. Worse, from a personal perspective, in 4 years time I'll sit in the late 40s age group which will make me part of the unhappiest group in the the UK (men 45-49 living in London). Time to move, perhaps?
This begs the question what we learn from spatial differences in wellbeing? As usual, the situation is more complex than the simple headlines would suggest. First, it's possible that small regional differences occur by chance. This can occur even when surveys are nationally representative because the number of people interviewed in any one location will be considerably smaller. Second, the survey tells us that individual characteristics affect stated happiness. From SERC work on wages we know that areas differences in wages look much smaller once we allow for the fact that different types of people live in different places.
Indeed, a recent paper by Oswald and Wu for the US (published in the Review of Economics and Statistics 2011) show that the relative position of US states in terms of wellbeing shifts around a lot if you control for observed non-financial individual characteristics (i.e. not income). Again, from our work in the UK on wages we know that controlling for the unobserved characteristics of people would change these area effects even more. In short, when who you are is important in determining how happy you are, simple comparisons of areas are much less useful than they might seem.
A further complication is also highlight by Oswald and Wu's work. Once we recognise that people are mobile across space, economic theory suggests that people shouldn't be able to make themselves better off by moving. As Oswald and Wu explain: "If Vermont, for example, offers a more attractive level of well-being to representative individual A than does Ohio, then we would expect to see Ohio citizens like individual A try to move to Vermont. That kind of migratory flow will cease only when a receiving region has become less desirable as an area in which to live. The economic equilibrium ought to be one of strict equality of utility [...]. If the economist’s arbitrage theory across regions is correct and well-being data are a useful proxy for utility, then its prediction should be detectable in an empirical test for state-by-state equality of well-being
for a person of given characteristics".
In other words, once we control for individual characteristics then there shouldn't be any spatial differences in wellbeing measures providing that the wellbeing measure is actually capturing the factors that determine people's 'wellbeing' (in the sense of utility). In practice, for the US, Oswald and Wu find that some spatial differences in wellbeing do persist but they also find support for a weak form of spatial arbitrage - once we control for individual incomes wellbeing is lower in states with higher average per capita incomes. In other words higher incomes are compensation for living in less nice places in terms of the non-financial benefits they deliver.
So where does all of this leave this soon to be middle aged London based male? The evidence suggests that moving is unlikely to do much for my happiness (unless I am doing it because for personal reasons I know think me and my family would be happiest somewhere else). Instead, I should try to keep my wife happy (married people are happier), enjoy the kids (at least until they are 11) and start saving for that Porsche.