Or to be more precise, a new study by John Helliwell and Haifang Huang has found cross-country evidence linking the quality of government with measures of life satisfaction. This follows last year's paper by the two authors on well-being in the workplace, which found many workers want trust, not money.
Their new paper (NBER Working Paper No. 11988) is entitled How's Your Government? International Evidence Linking Good Government and Well-Being. Helliwell and Huang used the World Values Survey measures of life satisfaction "as though they were direct measures of utility", to evaluate the effect of alternative features and forms of government:
We find that life satisfaction is more closely linked to several World Bank measures of the quality of government than to real per capita incomes, in simple correlations and more fully specified models explaining international differences in life satisfaction. We test for differences in the relative importance of different aspects of good government, and find a hierarchy of preferences that depends on the level of development.
The ability of governments to provide a trustworthy environment, and to deliver services honestly and efficiently, appears to be of paramount importance for countries with worse governance and lower incomes. The balance changes once acceptable levels of efficiency, trust and incomes are achieved, when more value is attached to building and maintaining the institutions of electoral democracy.
Looking at the scatter plots, the correlation between life satisfaction and quality of government is remarkably linear. One interesting and seemingly robust result is that "countries with some element of proportional voting do have higher levels of life satisfaction."
The authors conclude that "these data provide the broadest and least assumption-driven way to evaluate the quality of government". They call for more data:
If our case is accepted, at least in a provisional way, then life satisfaction data should be collected much more broadly and routinely. Since the relevant questions can be added at little or no cost to surveys being used for other purposes, there is ample scope for rapid increases in the relevant pool of data.
For those without NBER paper access, a November 2005 version of the papers is available here (PDF).