Government effectiveness and support for democracy

I just published an article on EJPR with the same title as this post. For those of you that may be interested on the subject but not so interested as to read the whole thing, here's an attempted summary.

Looking at survey data, many scholars have found that whether individuals prefer Democracy to alternative types of regimes is something that is unrelated with individual ("subjective") perceptions of governmental or economic performance, or even with "objective" measures of economic performance. The conclusion that has been reached on this basis can be summarized thus: "popular belief in the superiority of democracy is not susceptible to the ups-and-downs of government performance or the short-term economic fluctuation.” Instead, it “is largely a principled affair”, "a stable cognitive value cultivated through the socialization process in the society."

This finding seems to fit the basic conjecture that David Easton made a long time ago about the difference between "specific" and "diffuse support". Specific support is directed to "the perceived decisions, policies, actions, and the general style of (…) authorities.” In turn, diffuse support,“representing as it does attachment to political objects for their own sake, will not be easily dislodged because of current dissatisfaction with what the government does.” To put this in a simpler way, people's reasons to be happy or unhappy with the way the regime they live under works may very well be shaped by short-term factors related to the performance of governments or of the economy. But whether they are "democrats", or better put, whether they believe democracy to be "the only game in town" and to see it as "legitimate", is shaped by deeper, more structural, longer-term forces.

My point in this article is, I hope, very simple. I question the previous ideas in three ways. First, upon careful rereading of Easton and others, we find that this is not exactly what they conjectured. There's a very nice quote from a 1975 piece by Easton on this subject:

"Diffuse support may also, however, derive from experience. If only because this is a source usually associated with specific support, its significance for diffuse support may easily be overlooked or underemphasized. Members do not come to identify with basic political objects only because they have learned to do so through inducements offered by others – a critical aspect of socialization processes. If they did, diffuse support would have entirely the appearance of a non-rational phenomenon. Rather, on the basis of their own experiences, members may also judge the worth of supporting these objects for their own sake. Such attachment may be a product of spill-over effects from evaluations of a series of outputs and of performance over a long period of time" (Easton 1975: 446) 

Others, like Lipset, Dahl or Linz, made similar conjectures. Rational people living under ineffective regimes will sooner or later question their legitimacy.

Second, following recent work on the subject, I argue that one of the reasons why this has not been borne out by the data may be the use of inappropriate measures of regime support. Asking people directly about whether they "prefer democracy to other regimes" or whether they think "democracy is a good thing" is just one way of measuring support. There are different and more indirect ways, which focus more on whether people reject autocratic alternatives or whether they perceive an inevitable trade-off between "democracy" and universally valued outcomes, such as prosperity, decisiveness, and order. If we employ those measures, arguably less prone to "democratic lip service" and demonstrably more valid in cross-national research, maybe we will find that support for democracy is less prevalent than what people think, and more vulnerable to regime performance that what has been suggested.

Third, and finally, I argue that measures of regime performance could focus more on outputs than outcomes. In other words, effectiveness should be measured in terms of the quality of policy formulation and implementation, rather than economic outcomes or the perception of those outcomes. In spite of its arguable failings, there is an available measure that fulfills these criteria and has been used extensively in cross-national research: the World Bank's "government effectiveness" indicator.

What follows is quite simple. I use the World Values Survey integrated data file, construct three measures of regime support - one more "explicit" (EDS), another capturing rejection of autocratic alternatives (DAP), and another capturing whether people reject an inevitable trade-off between democracy and universally valued goals (DPE) - and pose two hypotheses.

H1. In democracies, greater effectiveness is linked to stronger democratic support.

H2. In non-democracies, greater effectiveness is linked to weaker democratic support (this is not really what one would like to test. Instead, one would like to test whether effectiveness in non-democracies increases support for whatever type of non-democracy people live under, but we simply have no good measures of that).

I use a multilevel model applied to data from never less than 50 countries and 76 surveys (depending on the availability of data), taking into account that respondents are clustered in country-years and those in countries, add a series of contextual and individual-level controls (GDP per capita, Ethnic fractionalization, Years under democracy, Income inequality, Age, Gender, Education, Income, Social trust, etc), and estimate the impact of Government effectiveness on the three measures of democratic support in different regime-types ("democracies" vs. "non-democracies", measured in two alternative ways). The take home figure is this, showing the marginal effects of effectiveness on these measures of democratic support:

When we use explicit measures of democratic support, support for H1 is absent. However, when we use alternative measures of support, effectiveness has a relevant positive impact in democratic support in democratic regimes (one standard deviation increase in effectiveness increases democratic support - DAP or DPE - by 2/3 of a standard deviation). Support for H2 is slimmer. Still, several of the marginal effects are negative, and two are significant at conventional levels. In short, effectiveness in democracies is a correlate of democratic support. Note also that all other macro-level predictors fail to perform as some of the literature suggests they should, at least once effectiveness is taken into account.

 It's a very simple idea, but I hope it is reasonably well executed. The message is less optimistic than what a democrat would like. Democracies are not immune to the consequences of government ineffectiveness and bad policy-making. Ineffective democracies are likely to suffer in terms of their legitimacy near mass publics. And there are signs, to be confirmed with better data, that effective autocracies may be more stable than what we think, by diminishing demand for democracy and increasing their own legitimacy.

That's it. I plan to continue to work on this issue. Comments very welcome.