Underappreciated Aspects of the Ratchet Effect

Over the years, as I have defended my hypothesis about the ratchet effect of national emergencies in U.S. history since the Progressive Era (when the ideological conditions for the full operation of this effect were established), I have encountered many doubters and critics. My fellow economists have been especially disposed to reject my hypothesis.

The economists sometimes dismissed my idea on the grounds that national emergencies are simply transient or abnormal events—outliers, as the econometricians call them—that should either be purged from a long time series or be treated as stochastic deviations from a long-run trend caused by more “fundamental” factors, such as demographic shifts or changes in the franchise. At times, the economists dismissed my hypothesis because, they maintained, it fails to fit every case in every country in every period of history; as puerile positivists, they concluded that the hypothesis is wrong because a correct hypothesis must be, in effect, a flawless law of history. Most often, perhaps, the economists pooh-poohed my hypothesis because they misconstrued it: they took it to be, like the typical economist’s model of the growth of government, a hypothesis about government spending and nothing more. So, if a certain crisis did not appear to have had a persistent, long-run effect by increasing the estimated trend level of government spending (often after standardization for increases in population, the gross domestic product, and the price level), then it did not explain, in their eyes, “the growth of government.”

I have always insisted that modern government has many facets and that, at minimum, a study of its growth must consider not only government spending (or taxing or employing), but also the government’s scope and power. Changes in these latter aspects of government do not leave the same kind of easily retrieved record, or numerical data set, that economists typically work with—and without which they are more or less at sea, or in denial. Over the many years that I have pursued my research into the growth of government, I have repeatedly met with evidence of essential elements of the ratchet effect that lie completely beyond the purview of conventional economic research on this subject.

The most recent such evidence I’ve encountered appears near the end of Godfrey Hodgson’s book The Colonel: The Life and Wars of Henry Stimson, 1867-1950 (New York: Knopf, 1990). Stimson was one of the most important public (or private) figures of the twentieth century. Among other things, he was the human bridge between the original American imperialists of the late nineteenth and early twentieth centuries (e.g., Elihu Root, Theodore Roosevelt, Leonard Wood) and the Cold War directors and managers of the national security state (e.g., Dean Acheson, John J. McCloy, McGeorge Bundy, Robert Lovett). Many of these postwar movers and shakers had worked for or with Stimson at some point, especially at the War Department when he was secretary of war from 1940 to 1945.

In considering how the postwar defense and foreign-policy Establishment, with its characteristic policy proclivities, had been fostered by such men’s engagement in the Big One, Hodgson writes:

Government service in World War II—in the War Department or other civilian departments for the slightly older men, in the Office of Strategic Services or elite military units for the younger ones—gave a whole generation of ambitious and educated Americans a taste for power, as opposed to business success, and an orientation toward government service which they never lost. When they went back to their law offices or their classrooms, they took with them contacts, attitudes and beliefs they had learned in war service. (pp. 384-85, emphasis added)

Ponder this passage, especially the part I have italicized. Here is a legacy of national emergency that no econometrician can tease out of a statistical time series, yet it is the kind of legacy that gives the ratchet effect its greatest substance.

I have emphasized again and again that the legacies of national emergency are not merely fiscal, but also, more critically, institutional and ideological. From Hodgson’s observation, we can see both the institutional and the ideological legacies embodied in a generation of highly placed, closely connected individuals who exerted tremendous influence over the apparatus and conduct of U.S. foreign and defense policies for decades after World War II and whose influence may be seen in the government’s conduct of foreign and defense affairs even today, though in somewhat muted and altered form.

The ratchet effect is a more complex phenomenon than most of my fellow economists seem capable of understanding. To grasp it, one must go beyond economic analysis, and even beyond public choice analysis. One must plunge into the domain of historical analysis, where social scientific theories are necessary but not sufficient for our understanding and where one continually meets path dependencies and contingencies whose resolution is anything but determinate, where real human beings, relying on their attitudes, beliefs, and commitments (shaped by their particular experiences, acquaintances, and loyalties) make genuine—that is, not fully predetermined—choices and thereby set in motion new streams of cause and effect that ripple through time and space.

A century before I formulated my ideas about the ratchet effect, the great William Graham Sumner understood its essence clearly when he said:

It is not possible to experiment with a society and just drop the experiment whenever we choose. The experiment enters into the life of the society and never can be got out again.

For historians and social scientists, the challenge remains to explicate precisely how this “entering into the life of the society” has occurred as a result of the “experiments” the government has performed during each of the great national emergencies of the past century. Such understanding appears even more imperative at present, as the government engages in yet another great experiment.