“Hey, honey, what happened? How is it that we eat meat so rarely? Have prices risen lately?” These questions, when generalized to every American household, have extremely important political implications. Thomas A. Stapleford’s very detailed and precise book explains the various ways in which the US federal administration, mainly through its Bureau of Labor Statistics (BLS), has answered since the end of the nineteenth century, and analyzes how the answer is tied to political concerns. His book, which pertains to the now-long tradition of socio-historical studies of statistics, describes the subtle and intricate relations between what might appear nowadays as a purely technical and neutral statistical measurement—the Consumer Price Index—and politics.
To tell his story, Stapleford sits precisely where the statisticians deliver their data and are confronted by those who use them: unions, other agencies of the administration, and, finally, the White House. The archives he studies in the BLS are mainly those of the commissioner or his deputy commissioners (very little hierarchically below); the literature he analyzes is mainly written for outsiders of the BLS. The evidently political character of these interlocutors explains why the book justly deserves the subtitle of “political history.” Let us consider three such examples
First, Stapleford explains how, during WWI, the first stable index—called the Cost of Living Index—was established by the BLS to facilitate federal arbitration of hours and wages between unions and employers. He shows that the very existence of such an index had important political consequences in labor relations because it was there to be mobilized by actors during negotiations. This fundamental political nature was, of course, amplified when the actual figure it gave favored one or the other camp.
Second, it is not the single uses, but the technicalities of producing this index that is also shown to bear on politics. For example, in the 1920s, an important theoretical decision had to be taken concerning the index. In effect, a cost of living index measured the variations of the price of a basket of goods. But this basket itself evolved. A worker’s family of 1900 did not buy the same things as an equivalent family in 1930. The problem, then, was that the price variation of the basket that varied had no clear meaning: no one could infer whether the cost of living was cheaper or dearer because the living itself was different. Therefore, it was necessary to stabilize this basket, but how? One could either use a “constant utility” basket (that is, a basket that contains changing products but provides the same utility through time for a family), or one could use a “constant goods” basket, containing always the same goods through a long period, even though their utility changed. The first choice had the advantage of being easy to conceptualize, but proved extremely difficult to put into practice (how to measure the utility of goods?). The second choice had a less clear interpretation, but was much easier to construct. The BLS, under the influence of institutionalists and especially Wesley Mitchell—themselves dedicated to “pragmatic” measurement (but it is unclear whether Stapleford uses this adjective in the philosophical or in the day-to-day meaning)—chose the “constant goods” approach. But the immediate consequence was that the index did not account for the change of quality of the basket, which would become object of fierce debates with unions during WWII.
Finally, the book shows how these statistics also had important consequences for the role played by the federal government in the economy. For example, after the chapter on the New Deal explaining how the personnel changed in the BLS, which became much more academic than it was before, there are two chapters on WWII. The first one is about the use of the index by the government to fight war inflation and the 1943 “little steel” decision, fixing the maximum wage raised to 15% of its January 1941 level; and the second is about the dissatisfaction of unions towards the BLS and their efforts to replace federal figures and experts with their own surveys during the war.
However, despite the very important qualities of the book, it deserves some criticism. First, some of the influences on the actors building the index do not receive the full treatment they deserve. Second, the very high level of academic capacity of the author leads him to a sort of theoretical conservatism, sometimes self-contradictory.
I have three remarks about the actors’ resources. First, at the turn of twentieth century, labor statistics was a highly internationalized topic; European countries and international institutes promoted many exchanges. Carol C. Wright, for example, who created the BLS, spent much time in the sessions of the International Statistical Institute, and had very close ties with British statisticians. Although Stapleford mentions some of them, it is arguable that they were much more influential than they appear here.
Second, a surprise of the book is the very late influence of both unions and the Bureau of Agricultural Economics (BAE) on the BLS index. In effect, within the US, from the mid-nineteenth century on, one of the most influential government statistical agencies, maybe as much as the Bureau of the Census, was the BAE and its antecedent, in particular the Division of Crop and Livestock Estimates (forgotten today because of the fall of the agricultural economical role, but then evident). Since 1863 it produced, among other data, a monthly series called the “Prices of Farm Products,” which Stapleford barely mentions, and only with respect to WWII. On their part, unions also had an early influence, and this time directly on the young BLS. As early as the 1890s, New York State, soon followed by Massachusetts BLS, asked the former, through an institutionalized system of correspondence, to estimate the number of their members unemployed. Every trimester, major unions sent such reports to the bureaus, which were the basis for the first unemployment series. Hence, it is surprising that these two sets of institutions (BEA and unions) appear mainly during discussions of WWII in the book (see Topalov Reference Topalov1994). It is true that unions are cited during the discussion of WWI, but not as a resource for producing figures.
Finally, Stapleford argues very convincingly that the main theoretical resource for the Cost of Living Index was institutional economics. But Stapleford’s treatment of this group as homogenous poses a problem, especially when we keep in mind that the BAE, whose price index during the New Deal appears to stem from exactly the opposite hypotheses from those of the BLS, was also a nest for institutionalists, especially M. Ezekiel (see Taylor and Taylor Reference Taylor and Taylor1953). It would be interesting to understand how this unique resource had two contradictory effects.
Now, two points about method. First, Stapleford opens his argument by placing himself under the authority of Max Weber, saying that he will describe a rationalization process: “A history of the ties between cost of living statistics and American Governance is thus a history of such [Weberian] rationalization process” (p. 7). This goal is very classic—sort of a pre-Science and Technology Studies approach. But, in fact, Stapleford shows the contrary throughout his book (especially concerning the “neutrality” of statistics, which, as said above, appear much more political than neutral), and concludes that “the strict separation of the political and the technical that is typically used to justify rationalized governance does not exist.” Therefore, holding to his first insight, he proposes in the epilogue to “construct a new vision of ‘rationalized’ governance” (p. 385) all by himself. But the fact that the actors he studied did not help him with this goal might lead one to think that the very Weberian question was not the best possible.
Second, the nine chapters are more or less period driven. The author opens each one of them with a presentation of the historical context and then enters the details of his own statistical story. Of course, this is how history is usually written because it helps the reader understand the argument. But the problem is that it erases one of the crucial and specific tasks with which federal statisticians are confronted: to properly define what they call the “universe”; that is, to define America itself, the territory to which their cost of living applied, and the time period that is of interest for them in a way compatible with their practices, graspable by their techniques. Statisticians are context makers, not takers (Didier Reference Didier2009). But the author’s writing method leads him to ignore the point. (Exceptions include the very interesting discussion about the kind of families taken into account [p. 85] and the kinds of cities [p. 163], but why and how these families and cities are chosen is not documented.) Stapleford justifies the word “politics” he uses in his title, but not “America,” which is nonetheless a huge problem—and no less political—for federal statisticians.
Stapleford’s historiographic choices may have caused his story to lack inventiveness, but the whole book remains extremely informative, very well written, and of a high professional standard. This, along with the centrality of his subject, makes it clearly a ‘must-have’ resource for students of anything from labor to statistics and/or the federal administration.
