The article started as a long paper on distributional inequality in the advanced capitalist societies that I presented at the 1976 annual meeting of the APSA. Using OECD data on the post-tax distribution of income in 12 countries, I found that the extent of inequality varied inversely with the extent to which Social Democratic and other leftist parties had governed in recent years. Control of government by left-of-center parties was closely associated with the extent of and growth in the extractive capacity of the state, defined in terms of the ratio of all public revenues to GDP, and the extent of and growth in extractive capacity were closely associated with the degree of distributional equality. But there were perplexing anomalies. For example, the Netherlands, where the Left had seldom governed, experienced the largest increase in the ratio of revenues to GDP since 1960 and was the most egalitarian in terms of the post-tax distribution of income. Belgium and Ireland also experienced unusually large increases in the ratio of revenues to GDP despite infrequent control of government by left-of-center parties.

Mindful of those countries' high dependence on trade and drawing upon Robert Dahl and Edward Tufte's (1973) discussion of the relationship between small size and trade dependence, the work of Odd Aukrust (1977) and Assar Lindbeck (1975) on “open” economies, that of Gerhard Lehmbruch (1977) on incomes policies, and that of Robert Gilpin (1975) on neomercantilism, I incorporated a measure of openness—the ratio of exports and imports to GDP—in the analysis. The measure was very closely related to the extent of the increase between 1960 and 1975 in 18 countries in the ratios of public revenues to GDP and social spending to GDP and, indeed, trumped all other explanatory variables, including the partisanship of government.

It is no doubt the article's conclusion that trade openness was the most important source of the expansion of the public economy between 1960 and 1975 that has caused it to appear on this list. That conclusion suggested that any account of change over time in taxes or spending would have to take the effects of openness into consideration.

In recent decades the strong positive cross-national correlation that I reported between openness and the increase in the relative size of the public economy has become slightly negative. The largest increases in the relative size of the public economy after the mid-1970s did not occur in the countries that were most dependent on trade but, rather, in countries that were, if anything, relatively closed. In some of the countries that were most dependent on trade—most notably, Belgium, the Netherlands, and Ireland, the earlier anomalies—the relative size of the public economy increased only slightly or decreased (see Cameron and Kim 2006).

Notwithstanding the change from a strong positive to a moderately negative cross-national relationship, the article's suggestion that a high degree of openness may, depending upon the partisanship of the government and the structure of labor market institutions, cause governments to offer compensation through social policy in exchange for wage moderation remains, I think, an intriguing hypothesis in an era marked by increasing global interdependence and economic “openness.”

A large literature has accumulated that uses pooled cross-sectional time-series analysis to determine, among other things, whether openness affects the magnitude and direction of annual changes in social spending. The results vary but generally suggest that neither the extent of openness nor the magnitude of increase in openness is associated with increases in social spending. However, using that mode of analysis, we have found that, although increases in the extent of openness had a contractionary effect on total revenues and expenditures in the period after 1970, large trade deficits and deteriorations in the balance of trade had, as the compensation hypothesis would suggest, strong expansionary effects on expenditures and budget deficits (see Cameron and Kim 2006).