In the course of his interesting discussion of “The Politics of Political Economists,” Professor Stigler challenges the alleged view of Professor Mises that “economic statistics, or more generally quantitative economics — generates a radical political viewpoint.”1 Stigler asserts that the empirical student acquires a “real feeling” for the functioning of an economic system, and “has had the complexities of the economy burned into his soul.”
Without going into the question of Mises’s precise viewpoint on this issue, I think it important to note that Stigler has overlooked several fundamental considerations.
In the first place, statistics are desperately needed for any sort of government planning of the economic system. In a free-market economy, the individual business firm has little or no need of statistics. It need only know its prices and costs. Costs are largely discovered internally within the firm and are not the general data of the economy which we usually refer to as “statistics.”
The “automatic” market, then, requires virtually no gathering of statistics; government intervention, on the other hand, whether piecemeal or fully socialist, could do literally nothing without extensive ingathering of masses of statistics. Statistics are the bureaucrat’s only form of economic knowledge, replacing the intuitive, “qualitative” knowledge of the entrepreneur, guided only by the quantitative profit-and-loss test.2 Accordingly, the drive for government intervention, and the drive for more statistics, have gone hand-in-hand.3
The enormous expansion of governmental activity in the gathering and disseminating of statistics in the last 25 years is surely more than coincidentally related to the similar expansion of the role of government in regulating and manipulating the economy. One of the leading authorities on the growth of government expenditures has put it this way:
Advance in economic science and statistics improved our knowledge of interstate and intrastate differences in needs and capacities and may have helped stimulate the system of state and federal grants-in-aid. It strengthened belief in the possibilities of dealing with social problems by collective action. It made for increase in the statistical and other fact-finding activities of government.4
We need not detail here the extensive use that has been made of national-income and gross-national-product statistics, as well as other statistical measures, in the attempts of the federal government at combating business cycles or unemployment.
Nor is this just a contemporary story. An authoritative work on British government puts the case thus:
the minor role of government during the nineteenth century reflects more than the absence of violent economic disruption; it also reflects the infancy of the economic and social sciences. Compared with recent decades, the volume of systematic information about social conditions was very small, which meant that the existence of problems was hard to establish persuasively.… If the volume of unemployment is unknown, the gravity of the problem is in doubt.
The accumulation of factual information about social conditions and the development of economics and the social sciences increased the pressure for government intervention.… Surveys like Charles Booth’s Life and Labor of the People in London revealed conditions which shocked public opinion in the late eighties and nineties. As statistics improved and students of social conditions multiplied, the continued existence of such conditions was kept before the public. Increasing knowledge of them aroused influential circles and furnished working class movements with factual weapons.5
Surely the role of the Fabian Society’s industrious empirical studies in furthering the cause of socialism in Great Britain is too well-known to need stressing here.
On the continent and in America in the late 19th century, it is well-known that the rebels against laissez-faire and the classical political economy stressed their replacement with induction from economic history and statistics. That was the goal of the German Historical School and its Verein für Sozialpolitik, and of the young, German-trained exponents of the “new political economy” of government intervention in the 1870s and 1880s.6 One of their leaders, Richard T. Ely, who called the new approach the “look and see” method, made it clear that the aim of fact gathering was to “mold the forces at work in society and to improve existing conditions”; they believed that as economists they had a responsibility for “shaping the character of the national economy.”7
And let us not overlook the eminent interventionist sociologist Lester Frank Ward, whose proposed “scientific,” “positive,” planned economy, would consist of “social engineering” based on statistical information fed from all parts of the country into a central bureau of statistics.8
Nor was it only abstract speculators who expressed such views. Statisticians themselves participated in this movement. As early as 1863, Samuel B. Ruggles, American delegate to the International Statistical Congress in Berlin, declared that “statistics are the very eyes of the statesman, enabling him to survey and scan with clear and comprehensive vision the whole structure and economy of the body politic.” One of the founders of the Verein für Sozialpolitik was the famous statistician Ernst Engel, head of the Royal Statistical Bureau of Prussia.9 US Federal Reserve Board Chairman Jerome Powell holds a news conference after a Federal Open Market Committee meeting in Washington, DC, December 19, 2018.
And Carrol D. Wright, one of the early commissioners of labor in the United States and a man greatly influenced by Engel, urged the collection of statistics of unemployment because he wanted to find a remedy (presumably via government action). Wright hailed the new German school as including men of all lands “who seek by legitimate means, and without revolution, the amelioration of unfortunate industrial and social relations.” Henry Carter Adams, a student of Engel’s, who established the Statistical Bureau of the Interstate Commerce Commission, believed that “ever-increasing statistical activity by the government was essential not only for the sake of controlling naturally monopolistic industries, but also for the efficient functioning of competition wherever possible.”10And certainly one of the great spurs toward constructing index numbers of wholesale and other prices was the desire to have government stabilize the price level.11
Unquestionably one of the prime founders of modern statistical inquiry in economics was Wesley C. Mitchell. There is no doubt that Mitchell aspired to lay the basis for “scientific” government planning. Thus:
[Quoting from Mitchell] “clearly the type of social invention most needed today is one that offers definite techniques through which the social system can be controlled and operated to the optimum advantage of its members.” To this end he [Mitchell] constantly sought to extend, improve, and refine the gathering and compilation of data.… Mitchell believed that business-cycle analysis … might indicate the means to the achievement of orderly social control of business activity.
he [Mitchell] envisaged the great contribution that government could make to the understanding of economic and social problems if the statistical data gathered independently by various Federal agencies were systematized and planned so that the interrelationships among them could be studied. The idea of developing social statistics, not merely as a record but as a basis for planning, emerged early in his own work.13
The federal government’s own account of the growth of its statistical agencies differs little from the above examples. The Bureau of the Budget, during President Eisenhower’s not rabidly socialistic administration, explained the continued growth of federal statistics as follows:
National growth and prosperity demanded an enlightened conduct of public affairs with the aid of factual information. The ultimate responsibility of the Federal Government for underwriting the health of the national economy has always been implicit in the American system.14
Then, speaking of the New Deal era after 1933, the bureau added:
A realization grew in the Congress and in high administration circles that sound and positive proposals to combat the depression required analysis based upon reliable information. As a result … statistical expansion was resumed at an accelerated pace.15
Suffice it then to say that a leading cause of the proliferation of governmental statistics is the need for statistical data in government economic planning. But the relationship works also in reverse: the growth of statistics, often developed originally for its own sake, ends by multiplying the avenues of government intervention and planning. In short, statistics do not have to be developed originally for politicoeconomic ends; their own autonomous development, directly or indirectly, opens up new fields for interventionists to exploit.
Each new statistical technique, whether it be flow of funds, interindustry economics, or activity analysis, soon acquires its own subdivision and application in government. A particular example is input–output analysis, which began as a purely theoretical attempt to lend empirical content to the Walrasian system of general equilibrium. It has now advanced to the point where its champions hail it as providing
an integrated picture of the industrial mechanism. They believe it can measure with fair accuracy the changes in inter-industry relations that would follow assumed changes in the “final bill of goods…” In practice, the most important change in the bill of goods is that called for by way of large-scale rearmament. It is hardly astonishing, therefore, that most of the development and application of input–output studies have been connected with industrial mobilization.16
There are other reasons why the statistically oriented will tend to become interventionists. For one thing, the economic statistician will tend to be impatient of all theory as “armchair speculation,” and hence will tend to advocate piecemeal, pragmatic, decide-every-case-on-its-“merits” type of government planning. It is perhaps true, as Stigler declares, that few empirical economists have become outright socialists or communists; such a course would be much too theoretical for them. But neither do they become adherents of laissez-faire; instead, a case-by-case, ad hoc approach drives them down the path of a muddled government interventionism.
I do not know whether, as Stigler asserts, “the most radical wing of the new dealers was not distinguished for its empirical knowledge of the American economy.” But certainly the Tugwells and the Stuart Chases and the Veblenians proclaimed their empiricism often enough. And historians of the New Deal generally praise it highly for its flexible, pragmatic approach.
Another reason why statistics and political pragmatism are mutually congenial is that the very hallmark of the pragmatic approach is to begin by looking for problems or “problem areas” in the society. The pragmatist looks for areas where the economy and society fall short of the Garden of Eden, and these, of course, abound. Poverty, unemployment, old people with scurvy, young people with cavities — the list is indeed endless. And as each problem multiplies under the care of his eager research, the pragmatist calls ever more stridently for government to do something — quickly — to solve the problem. Only hard-headed, deductive, a prioristic economic theory can teach him about ends and means, allocation of resources, opportunity cost, and the other rigors of the economic discipline.
Considering the above discussion, it is no wonder that conservative members of Congress, in the days before they were indoctrinated in the modern economic niceties by the Joint Committee on the Economic Report, were very suspicious of the seemingly harmless expansion of federal statistical activities. Thus, in 1945, Representative Frank Keefe, conservative Republican congressman from Wisconsin, was in the process of questioning Dr. A. Ford Hinrichs, head of the Bureau of Labor Statistics, on the latter’s request for increased appropriations. In the course of the questioning, Keefe’s misgivings about government statistics emerged as a cry from the heart—unsophisticated perhaps, but at least of sound conservative instinct:
There is no doubt but what it would be nice to have a whole lot of statistics.… I am just wondering whether we are not embarking on a program that is dangerous when we keep adding and adding and adding to this thing.
We have been planning and getting statistics ever since 1932 to try to meet a situation that was domestic in character, but were never able to even meet that question.… Now we are involved in an international question…. It looks to me as though we spend a tremendous amount of time with graphs and charts and statistics and planning. What my people are interested in is, what is it all about? Where are we going, and where are you going?17
I think we can conclude that the nub of the difference between Stigler and myself is this: to him a radical or nonconservative is essentially a socialist or a communist. To me, a nonconservative is someone who advocates intervention rather than laissez-faire. The difference is one of frame of reference. If we define conservatism as Stigler does, then it is true that most economists are conservatives; if we define it as believing in laissez-faire, then the conclusion must be very different. For the key then becomes not so much economics and 18noneconomics as theory versus empiricism. Empiricists will tend less to be full-scale socialists, but will also drift generally toward intervention.
Still, when all is said and done, it is probably true that even the proportion of believers in laissez-faire is much greater among economists than in other academic disciplines, and that the “average” point on the ideological spectrum in economics is considerably “to the right” of the average in other fields of study. It appears that the economic discipline, per se, imposes a rightward shift in ideological belief. And this, after all, is the main point of Stigler’s article.