This fascinating paper by Di Tella and MacCulloch shows that the simple fact that one's favored political party is in power has a big effect on happiness:
A surprising finding of the paper concerns the relative importance of politics. We include in our partisan happiness equations a variable that measures the ideological position of the government in power. It indicates that when the government leans more to the right ideologically, right-wing individuals tick up their happiness scores. In the same periods, left-wing individuals declare themselves to be more dissatisfied with their lives. The size of the coefficient is large and highly significant. A right-wing individual living under Mitterrand would be willing to put up with an increase of 11 percentage points in the inflation rate in order to see Margaret Thatcher take charge of the government. One possible explanation for this result is that there are other policies, not linked to macroeconomics in nature, along which governments differ and that our analysis ignores. These could include agricultural policy, the approach to fighting crime, the policy on abortion and other social issues, etc. But another possibility is that politics enters directly into the utility function (or that people simply care about winning). Furthermore, the variable capturing the ideological position of the government (Right Wing Government) is strongly correlated with inflation (negatively) and unemployment (positively). Thus, there seem to be two channels through which governments affect the well-being of their constituencies: a direct channel and an indirect effect through unemployment and inflation. Our results indicate that the color of the government matters for a large part of the population. [emphasis added]
My favorite hypothesis is that coalitional success enters directly into the welfare function. Now, this is fascinating for all sorts of reasons. For instance, it would seem, then, that the need to maintain a distinct and coherent coalitional identity will limit median-voter convergence. It also implies weird things for utilitarians who insist on maximizing relative to current preferences. If the utility hit to rightwingers out of power, for example, is greater than the utility to hit to leftwingers out of power, then, other things being equal utility-wise, it could turn out that a rightwinger minority should be put in power over a leftwinger majority. The general application of this kind of thinking is that partisans will try to convince their side that being out of power is really depressing, with the result that no matter who is in power, half the population is really depressed.
But let me instead point out a picayune possible implication for libertarians. People apparently like unemployment insurance because we tend to prefer that our income changes in a relatively smooth way, rather than suddenly and drastically. Could libertarianism be a utility smoothing individual political strategy. You're never in power, but then you're never out of power either. No ups, no down, no anxiety about the next raise always around the corner. In a world of partisan volatility, libertarianism is a kind of insurance against hedonic swings from politics. Whether this leaves us better off on net is an open question.