How’s Your Government? International Evidence Linking Good Government and Well-Being

Download 102 Kb.
Size102 Kb.
  1   2   3   4   5   6   7   8   9

How’s Your Government?

International Evidence Linking Good Government and Well-Being

John F. Helliwell and Haifang Huang*

1. Introduction

Defining and evaluating ‘good government’ requires some heroic assumptions. Which features of the operation of a government are valuable and to whom? If a government’s goodness is to be defined by the nature and consequences of its policies, which policies and which results should be used as litmus tests? While most analysts would agree that the quality of a government depends on the extent to which it improves the welfare of its citizens (with due regard for the interests of and relations with others), the lack of readily available measures of welfare has led to a wide range of techniques being used for evaluation. Political scientists have tended to give pride of place to governmental systems that embody democratic principles and practices supported by the rule of law. Economists have tended to evaluate systems of government, and institutions in general, in terms of their perceived ability to foster economic growth (sometimes with regard also for the distribution of the resulting goods and services).

In this paper we adopt a much broader perspective. We employ survey measures of life satisfaction as though they were direct measures of utility, and use them to evaluate alternative features and forms of government. Since measures of life satisfaction are now available for many countries, they offer a potential new tool for comparative political economy. If these data are to provide valuable new insights, they must be seen to tell at least a slightly different story from other evaluation methods, and to do so in a convincing way. This is an ambitious research agenda, in which the results we report here are neither the first nor the last words. In this paper we first show that several different measures of governmental quality are highly correlated with cross-national differences in life satisfaction. Although many of the same measures of governmental quality have been shown to affect prospects for economic growth, our evidence suggests that the income channel is a relatively minor part of the story, since the partial effects of government quality on life satisfaction are largely unaffected by the inclusion of average per capita incomes. This demonstration will occupy the next section of the paper. Then there are two sections of more detailed results, the first testing the robustness and likely channels of the links between governmental quality and life satisfaction, and the second using our data and methods to cast a slightly different light on some recently studied issues in comparative political economy. For example, it has been found that proportional and majoritarian democratic systems appear to have different implications for the size of government spending. But is this higher spending a good thing, as would be suggested by the argument that it reflects the design and implementation of broadly-based welfare systems, or a bad thing, as surmised by those1 who have reported that high ratios of government spending are associated with lower economic growth? Our direct measures of life satisfaction provide a possible method for evaluating the net welfare effects of alternative political institutions. We conclude with a summary of our results and proposals for further research.
Before launching into a presentation of results using subjective assessments of life satisfaction as a measure of utility, it is perhaps necessary to answer, at least in a provisional way, some of the sceptical questions that might be raised about our proposed research agenda. Many social scientists have viewed subjective data with suspicion, regarding such evidence as much inferior to more explicitly behavioural data. Second, many users of life satisfaction or happiness data have encountered difficulties posed by excluded variables and reverse causation. For example, cross-sectional studies based on individual data have shown an apparently large effect linking marriage to higher life satisfaction. Thinking in excluded variable terms, sceptics have argued that since inherent genetic personality differences have been shown to be strong determinants of life satisfaction, these excluded traits are likely to affect both happiness and marriage status because those with optimistic and extroverted personalities are likely to be both happier and married. Seen as an issue in reverse causation, any random factor that made someone happier would also make them more attractive to others. Such correlation with unmeasured personality differences is likely to be more important at the individual than at the national aggregate level, since personality types are likely to average out in large samples. But it has also been argued that even at the national level there may be differences in personality or mood that would affect responses to life satisfaction questions in ways not connected with ‘real’ differences in the quality of life2.
Perhaps the strongest defence against contamination by personality differences is to take such traits directly into account. Entering measured personality differences at the individual level3 and at the national level4, does indeed show them to have substantial importance in explaining life satisfaction, especially at the individual level. But in neither case does explicit allowance for these differences change the main results that were previously apparent.
The strongest support for the ‘reality’ of the subjective assessments of life satisfaction is that they tell a remarkably consistent story to that provided by suicide data5 The ability of the same variables to explain both life satisfaction and suicide in a consistent manner is especially remarkable because suicides tend to be drawn from those at the extreme lower tail of the distribution of life satisfaction6, and there is no guarantee that the determinants should be so similar. The fact that these two such different types of data tell the same story provides strong support for subjective assessments of life satisfaction. To counter the objection against the use of subjective data (an objection that is in any event seldom supported by research findings), the suicide data are based on ultimate and irrevocable decisions. The suicide data are also free of suspicions of reverse causation. Thus the fact that the suicide and life satisfaction data are well explained by the same model suggests that the life satisfaction results cannot be dismissed out of hand, whether by rejection of subjective data, by suggestions that the questions might be interpreted so differently in different countries as to make comparisons useless, or by the possible existence of reverse causation. Thus we are convinced that the life satisfaction data have strong claims to be used as direct measures of utility. Since they have been widely collected in the past, and are very easy and cheap to collect in the context of almost any survey, we think that they provide a promising tool for comparative assessments of the quality of governance.

Download 102 Kb.

Share with your friends:
  1   2   3   4   5   6   7   8   9

The database is protected by copyright © 2022
send message

    Main page