I knew Milton Friedman, but not well; and I am not competent to express an informed opinion on his major academic work, which was in macroeconomics. The economists of his generation with whom I principally associated were George Stigler, Ronald Coase, and Aaron Director (Friedman's brother-in-law)--microeconomists who had a major impact on the law and economics movement.
I did, however, read a few of Friedman's essays. Two in particular struck me around the time I came to Chicago. One was his essay on the methodology of positive economics, in which he argued that the way to test a theory was not by assessing the realism of its assumptions, but by assessing the accuracy of its predictions. Economics makes heavy use of unrealistic assumptions, primarily concerning rationality, and yet the predictions generated by models based on those assumptions are often accurate. Where they are inaccurate, this is a spur to reexamining the assumptions and perhaps modifying them, as is occurring in such fields as finance, where assuming a more complex human psychology than finance theorists traditionally assumed has helped to explain anomalies (from a rational-choice perspective) in the behavior of financial markets.
The emphasis on predictions connects Friedman's essay to Karl Popper's philosophy of science, in which the scientific method is viewed as a matter of making bold hypotheses, confronting them with data, and ascribing tentative (always tentative) validity to the hypotheses that survive the confrontation. Popper's methodology of fallibilism has strong affinities with Friedman's methodology. Both are strongly empiricist. Stigler in conversation merged these two closely related approaches, and I was very struck by the melded approach.
The other essay of Friedman's that struck me was an essay on taxation in which he argued, contrary to the conventional view at the time (though I gather the argument was not original with him), that there was no theoretical reason for supposing income taxes superior in point of efficient resource allocation to excise taxes. An excise tax--say, a 10 percent tax on yachts--drives a wedge between cost and price and so deflects buyers to substitutes that may cost more to produce but look cheaper because they are not taxed at so high a rate. (The effect is the same as monopoly pricing.) But Friedman argued that income taxes have the same effect, by driving a wedge between the cost of work and the wage (price) received by the worker, thus deflecting him to untaxed substitutes, such as leisure, or to jobs that generate untaxed benefits, including leisure in the case of teaching (for example), but also prestige, amenities, tax-favored fringe benefits, and job security. This idea of the parity of excise and income taxes has wide-ranging implications for public policy, since the tendency (still) is to neglect the misallocative effects of income taxation--a neglect of which I think even Friedman was sometimes guilty, as I am about to argue.
Perhaps his most important general contribution to economic policy was the simple, but when he first propounded it largely ignored or rejected, point that people have a better sense of their interests than third parties, including government officials, do. Friedman argued this point with reference to a host of issues, including the choice between a volunteer and a conscript army. With conscription, government officials determine the most productive use of an individual: should he be a soldier, or a worker in an essential industry, or a student, and if a soldier should he be an infantryman, a medic, etc.? In a volunteer army, in contrast, the determination is made by the individual--he chooses whether to be a soldier or not, and (within limits) if he decides to be a soldier what branch, specialty, etc., to work in. A volunteer army should provide a better matching of person to job than conscription, and in addition should create a more efficient balance between labor and capital inputs into military activity by pricing labor at its civilian opportunity costs.
But this is in general rather than in every case. The smaller the armed forces and the less risk of death or serious injury in military service, the more efficient a volunteer army is relative to a conscript one. These conditions are not satisfied in a general war in which a significant fraction of the young adult population is needed for the proper conduct of the war and the risk of death or serious injury is substantial--the situation in World War II. For then the government's heavy demand for military labor, coupled with the high cost of military service to soldiers at significant risk, would drive the market wage rate for such service through the roof. Very heavy taxes would be required to defray the expense of a volunteer army in these circumstances and those taxes would have misallocative effects that might well exceed the misallocative effects of conscription.
I mention this example because I find slightly off-putting what I sensed to be a dogmatic streak in Milton Friedman. I think his belief in the superior efficiency of free markets to government as a means of resource allocation, though fruitful and largely correct, was embraced by him as an article of faith and not merely as a hypothesis. I think he considered it almost a personal affront that the Scandinavian nations, particularly Sweden, could achieve and maintain very high levels of economic output despite very high rates of taxation, an enormous public sector, and extensive wealth redistribution resulting in much greater economic equality than in the United States. I don't think his analytic apparatus could explain such an anomaly.
I also think that Friedman, again more as a matter of faith than of science, exaggerated the correlation between economic and political freedom. A country can be highly productive though it has an authoritarian political system, as in China, or democratic and impoverished, as was true for the first half century or so of India's democracy and remains true to a considerable extent, since India remains extremely poor though it has a large and thriving middle class--an expanding island in the sea of misery. What is true is that commercial values are in tension with aristocratic and militaristic values that support authoritarian government, and also that as people become economically independent they are less subservient, and so less willing to submit to control by politicians; and also that they become more concerned with the protection of property rights, which authoritarian government threatens. But Friedman seemed to share Friedrich Hayek's extreme and inaccurate view that socialism of the sort that Britain embraced under the old Labour Party was incompatible with democracy, and I don't think that there is a good theoretical or empirical basis for that view. The Road to Serfdom flunks the test of accuracy of prediction!
I imagine that without the element of faith that I have been stressing, Friedman might have lacked the moral courage to propound his libertarian views in the chilly intellectual and political climate in which he first advanced them. So it should probably be reckoned on balance a good thing, though not to my personal taste. His advocacy of school vouchers, the volunteer army (in the era in which he advocated it--which we are still in), and the negative income tax demonstrates the fruitfulness of his master micreconomic insight that, in general, people know better than government how to manage their lives. But perhaps not always.