Much conventional wisdom is that, in the countries of the postindustrial core of the world economy, the role of the state in the economy reached its apogee in the 1970s, and that the coming of Reagan in the United States, Thatcher to Great Britain, and neoliberalism everywhere began a process of rollback: a return to more “market” and less “state” in the postindustrial “mixed” economies that dominate the world and are the most successful political systems of economic regulation that the world has ever seen.

To some degree this conventional wisdom is true. Governments in the 1970s tried to shape the industrial structure of their economies—engaged in “indicative planning” and other such exercises—to a degree rarely seen today. Governments both liberal and conservative in the 1970s put their thumbs on the scales in support of union movements in a way that they do not any more. The idea that activist fiscal policy could and should be used to guarantee full employment has proved to be yet another utopian delusion.

But in two very important areas the state has not retreated but has advanced. The first is that of central banking—an extraordinary island of central planning in our modern economies, paradoxically the domain of technocrats springing from a monetarist tradition that fears and hates the idea of central planning. The second is the domain of Peter Lindert's Growing Public: social spending. Public spending on health, education, welfare, and other forms of social insurance has not peaked and has not retreated. It has, rather, continued to grow as a share of national income, and promises to grow still further in the coming decades as aging populations, improving medical technologies, and rising educational expectations call forth additional demands.

The instinctive reaction of neoclassically trained economists—even those who put a very high weight on equality in their implicit and explicit social welfare functions—is that these high levels of social spending are at worst a disaster and at best a pity. It is not the case that most social spending is redistributive. Middle-class people go to subsidized museums where admission fees have been reduced because of their taxes. Taxes collected from middle-class people pay for doctors and nurses to treat middle-class people “for free.” The middle-class young pay taxes that are used for pay-as-you-go pensions for the middle-class old. Everywhere there are tax wedges pushing private rewards to work and industry away from social values, and for what? So that resources can be collected from some middle-class people and used to benefit other middle-class people.

Neoclassically trained economists will, when pushed, admit to a valid insurance motive as well as a valid redistributive motive. But, they will say, the logic of the social insurance and welfare states is not economic, it is political. Better if some way could be found to float every middle-class tub on its own bottom, and reserve taxing and spending for true redistribution and insurance.

Peter Lindert's Growing Public challenges this picture. Lindert believes and argues that the economic efficiency costs of growing public social spending are small. Lindert wants to argue that a large public spending total is, in today's developed world, more likely to be a sign of a nation that is successful as a political-economic enterprise than the reverse.

To this task Lindert brings, as Bob Margo wrote in his review of Growing Public for eh.net, a “command of a vast array of historical and contemporary evidence” that is “incredible.” He uses this command to make a complicated historical, comparative, and panel-data statistical argument that attacks two puzzles, which Lindert calls the “Robin Hood” and the “Free Lunch” puzzles. The “Robin Hood” puzzle is that among advanced countries social spending tends to be highest where people do not need it that much: it is highest where poverty and inequality are lowest. The “Free Lunch” puzzle is that Sweden and company are not punished for their high levels of social spending with substantial efficiency losses that leave them relatively impoverished. These puzzles are broad and exist at a more than cross-national level: even within the United States, social spending on the state level has been on a rise and has been positively associated with state-level wealth. Connecticut, New Jersey, and California are not punished for their generosity relative to Wyoming, Mississippi, and South Carolina.

Lindert lists what he sees as his nine major conclusions starting on page 20 of the first volume of Growing Public. I see fewer major conclusions: six. They are: (1) that the rise of social spending tracks the rise of political voice; (2) that decentralized school funding gives effective political voice to local interests that want high education spending; (3) that social spending is a superior good, so the spending share demanded rises with income; (4) that Roman Catholicism was a negative influence on taxes and transfers up until but not after the destruction of the antidemocratic right in World War II; (5) that social insurance states are significantly larger where the working class and the poor look like each other and trade places; and (6) that larger social insurance states with more redistribution are not associated with heavy drag on economic growth for two reasons—first, nations with larger social insurance states have less inefficient tax systems, and second, universal benefits avoid the surprisingly heavy incentive and administrative costs of strict means testing of benefits.

The thumbnail sketch takeaway is that large-scale government spending on social welfare and social insurance is here to stay: it has, Lindert argues, few if any economic efficiency costs, offers powerful redistributive and insurance benefits to society as a whole, and offers powerful political benefits as well, for the modern social insurance and social welfare statements are very popular among the voters. In Lindert's view, we should not expect the social insurance state to be eroded over time. It is not under pressure to shrink as a result of its inefficient handing of too-great tasks to the thumb-fingered government. Rather, the social insurance state is here to stay as the most successful set of human governance institutions of all time.

To my mind the historical and comparative analyses are the strongest. I, at least, find the regression analyses offered in volume 2 of Growing Public as less convincing than the other components of Lindert's tripartite historical-comparative-statistical story that make up the bulk of volume 1. Serial correlation is rampant, identification disputable, the number of truly independent cases low. Hypothesis testing is, as a result, unconvincing—or would be unconvincing if Lindert rested his case on panel-data regressions alone. These problems are inevitable in large-scale historical- comparative work, and will be until we can run natural or unnatural experiments on whole planets set up with similar initial conditions. The regressions work best, I think, as ways of summarizing the data: ways to convince readers that the stories of historical causation and the comparative arguments made in Lindert's case studies are in fact representative of the universe of advanced postindustrial nations, and not atypical. At this task Lindert's regressions are completely convincing, especially because of the large amount of heavy lifting that has gone into the quantitative-comparative foundation of Growing Public. To get a sense of how much heavy lifting was involved, the curious and researchers can find Lindert's data for Growing Public at <http://www.econ.ucdavis.edu/faculty/fzlinder/Lindert%20data%20CUP%20book/CUP%20data%20page.html>.

One reason for Lindert's success, I believe, is that his argument has its roots not so much in James Buchanan-Gordon Tullock–style public choice as in a more sensitive and catholic tradition that I think of in terms of Albert Hirschman's Exit, Voice, and Loyalty. Individuals are loyal to others in that they have a utility function that can depend not only on their own consumption but on that of other groups, and so pressure the government for redistributive programs and transfers. How much individuals can pressure the government depends on their voice—which in Lindert's analysis is taken as a given historical trend rather than as a dynamic process in its own right.

Lindert finds confirmation of his general voice-and-loyalty model in the pre-twentieth-century history of poor relief. The early history of poor relief is simple: there was not very much of it. In England starting about 1750, however, and only in England, poor relief becomes much more generous. Lindert follows George Boyer in interpreting English poor relief as a program for redistributing wealth upwards rather than downwards. According to Boyer and Lindert, increasing commercialization of agriculture created a high, seasonal demand for labor on the part of the truly rich landed elite who controlled the levers of political power in eighteenth-century England. By taxing the general rate-payers of the parishes to provide seasonal poor relief, the truly rich elite could reduce the wage costs they would otherwise have to pay if their workers had to subsist for 12 months on what they were paid for working eight. The rural poor benefited, but the rural really rich benefited as well. With the spread of the franchise at the coming of Reform, the political “voice” that supported this program lost out in the more-general clamor. Poor relief expenditures fell, not to rise again until the political incorporation of the working and other poor into the electorate in the twentieth century.

The other major pre-twentieth-century focus of Lindert's book is education. Why is it that the leaders in education were America, Germany, and—it turns out, somewhat surprisingly—France? Lindert attributes their lead in education to a combination of local control and widespread voting rights. Where education policy and spending is decided on at the national level, things are too dispersed to support a political pressure coalition to create high education spending. Your national-level taxes will be used to pay for the schooling of their children, and you do not know them. By contrast, when school funding is decided on the local level the degree of social cohesion and loyalty is strong enough to support a large and well-funded public school system. Both widespread voting rights and local control are needed if a country was going to be a leader in public education before the twentieth century.

In the twentieth century, Peter Lindert takes on the puzzle of why high education spending and high education duration in the United States are associated not with higher but with lower scores on standardized tests in America than in other countries that spend less and have earlier ages at which students leave school. The right-wing view is that American education has been captured by the teachers' lobby, and that as a result America gets relatively little bang for the bucks it spends on education, at least on primary and secondary education. Lindert does not really want to accept this interpretation, but if I read the book correctly he cannot find an alternative interpretation that fits the facts. He prefers to leave this situation as an unsolved puzzle.

For the twentieth century as a whole, Lindert provides an explanation of the “Robin Hood” puzzle that I find satisfactory. Cross-class loyalty between the middle class that dominates the electorate and the poor who need redistributive transfers, increasing wealth, aging populations, and the extent of the franchise together provide a sufficient explanation for a rising demand for social insurance, health, and education spending that political parties seeking to win elections have competed to supply. Countries where inequality is rampant are, by and large, also countries where there is little cross- class loyalty between middle and poor (and also are countries with younger populations and, often, restricted franchises). Before World War II, these also tend to be places where the religious Catholic hierarchy has aligned itself with the antidemocratic right. Thus the countries where it is easiest to assemble the political coalition for redistributive social spending are also the countries where the objective need for it is least.

I find Lindert's explanation of the “Free Lunch” puzzle to be less satisfactory. In a case study of Sweden that examines “universalism” in benefit programs and analyzes the care with which governments design their tax policies, Lindert argues against the “elegant fiction” of the Laffer curve in its believable form: the claim that revenues rise less than one-for-one with tax rates because a developed country with high tax rates will suffer from considerable tax deadweight-loss-caused drag on its economic growth. Lindert's case hinges on a negative: he cannot find any pattern in which larger social spending is associated with lower levels or growth rates of GDP.

I was unconvinced by the case study of Sweden. Sweden is just too weird a country for what goes on there to be taken as representative or illustrative of processes that go on in other advanced postindustrial economies. The peculiar nature of Sweden's Labor Organization, the role of the Wallenberg family, the extremely high export orientation of the Swedish economy, its extraordinary homogeneity, the fact that it had a House of Peasants in the Ståndsriksdagen since the 1436 meeting at Uppsala in the aftermath of Engelbrekt Engelbrektsson's rebellion, and so on make me extremely leery of drawing conclusions from and attempting to apply the Swedish or more broadly the Scandinavian model elsewhere.

The rest of Lindert's explanation of the “Free Lunch” puzzle is his belief that the relatively high marginal tax rates and notches created by means-testing social welfare programs do have significant efficiency costs through their incentive effect, and his belief that the fiscal stakes matter in that a country such as Germany that seeks to raise a huge amount of money—nearly 50 percent of GDP, say—will pay great attention to making sure that its taxes are as close to nondistortionary as possible, whereas a country such as the United States that seeks to raise less money—30 percent of GDP, say—will pay much less attention to the details of tax code and wind up with as large efficiency losses from its tax system as the higher-tax country.

It is certainly possible. Adam Smith did write in the Wealth of Nations that the French found their tax system three times as burdensome as the British found theirs, even though the French system raised only one-third the per capita revenue of the British. Most public finance economists teach that the deadweight losses from the U.S. system of capital taxation are very high. But I cannot convince myself that these are laws of nature rather than accidents of history. In my mind's eye I do see the possibility of a United States with equally redistributionist taxation as a whole but also with much lower capital taxation and stronger economic growth. In my mind's eye I do see continental European countries that might have relied less on sales and more on wealth taxes to fund the expansion of social spending that the left wanted for its own sake and that the center-right felt it had to support in order to cement the national consensus for the Cold War. That it is hard to argue convincingly that continental Europe suffers in economic growth from the high average tax rates needed to maintain its social spending does not mean that the standard blackboard explanations of deadweight loss are wrong. But here I sound like some Marxist arguing that the failure of the rate of profit to fall is still consistent with the claim that the rate of profit tends to fall. Lindert has history on his side. All I have is economic theory.

In the middle of volume 1 Peter Lindert leaves economic history behind and gazes into his crystal ball. Like many others, he sees current levels of social insurance benefits as unsustainable. Governments cannot keep their pension and health promises, he believes and argues in chapter 8—especially as faster economic growth raises expectations about the scale of publicly provided geriatric health care. The tax base cannot, Lindert believes, be willing to support the further expansion of tax rates that would be required. Lindert's argument here is not inconsistent with his explanation of the “Free Lunch” puzzle, but the fit is awkward. It is, after all, the parents and relatives of the current taxpayers who are receiving publicly provided health and pensions. As long as they remain middle-class societies, they may view cutbacks in the pension that Uncle Fred in the back room is receiving with as much alarm as they would view increases in the tax rates on their own income.

In chapter 9 Peter Lindert asks a different question: does the history of the most developed countries present to the less developed an image of their own future? Lindert says, “yes.” The political and economic processes that led to the rise of social spending in Europe and America are at work in the rest of the world as well. In time, developing countries will face the same opportunities, problems, and challenges that have led to growing public spending in richer parts of the world, and for some of these countries—China, for instance—the future will soon be here.