Search

A large and growing body of research demonstrates a correspondence between public opinion and policy behavior. Almost all of this research has focused on the US, however. Do similar patterns obtain in other countries? Or is the US unique? This paper represents one step towards answering these questions. We extend previous research on the dynamics of spending preferences and budgetary policy in the US, using comparable longitudinal measures of public preferences and government spending in the US, UK, and Canada. The nature and magnitude of opinion-policy connections vary across both policy functions and countries. We explore this variance here, and suggest several possible explanations. [ABSTRACT FROM AUTHOR]

This paper examines homogeneity and heterogeneity in preferences for public policy, across income and education levels and party identification, and across policy domains in the US, UK and Canada. Do preferences differ across segments of the public at particular points in time? What about over time - do we observe a uniform swing across different groups (parallel publics), or does the flow of opinion differ across party, income, and education? Our data on public preferences for government spending show cross-sectional heterogeneity across sub-aggregates, but very high levels of over-time homogeneity in trends across subgroups. The structure of preference change across subgroups is fairly similar; most notably, thermostatic responsiveness to policy holds generally across the subgroups. This is especially true for broad domestic aggregates of opinion and policy. There still are some differences in feedback, particularly across education levels, as the highly educated are significantly more responsive to spending change over time. People with low and middling levels of education still do change their preferences for spending in response to what policymakers do, just to a lesser extent, at least in certain policy domains. This fits with the large and growing literature that demonstrates real differences owing to information levels. There also are significant differences across partisan subgroupings in various domains. The implications of these findings are considered in the concluding section. ..PAT.-Unpublished Manuscript [ABSTRACT FROM AUTHOR]