JSTOR

A torrent of numbers accompanies both the bureaucratic communication and the public discourse that are characteristic of modernity. And there is no area where their dominance is more pronounced than with regard to economic affairs. From as early as the seventeenth century, “political arithmetic” was a favorite tool of economic punditry, but it was in the late nineteenth and the early twentieth century that the production of economic data became quasi-industrial.1 States and interest groups constructed systems of economic statistics in the form that we know them today, as instruments of real-time observation with a new and reified conception of “the economy” as their object.2 Clearly the history of statistics is intertwined with the history of what Michel Foucault termed governmentality.3 However, since they circulated both inside and outside the structures of power, the role of the new data in enhancing governmental capacity, even in a broad Foucauldian sense, was never straightforward.4 At moments of crisis, rather than smoothly tracing the course of history, the flow of numbers could become a turbulent maelstrom. By the early twentieth century, the mechanisms of economic reporting were so powerfully developed that they had the capacity to swamp political actors with wave upon wave of bad economic news.5 Furthermore, as competing interest groups equipped themselves with teams of economists and statisticians, the “facts” themselves became the object of disorienting controversy.6

There was a time when the flood of data filling the basement of the modern archive inspired hopes of a more “social scientific” history. And the project of quantitative history is by no means dead.7 However, across most of the diverse discipline we call history, this enthusiasm has cooled and in some cases curdled into an attitude of hostility and incomprehension.8 The history of this divorce has varied across national academic communities and between institutions.9 Not everywhere has there been a complete rupture. For those inspired by the “linguistic turn” or Foucault's concept of power/knowledge, there were ways of continuing to engage with quantitative sources.10 A characteristically creative response came from Joan Wallach Scott, in her pathbreaking analysis of the categories of nineteenth-century French industrial statistics.11 But, as William H. Sewell, Jr., observes in Logics of History, although Scott was still writing about statistics, she had discarded the questions asked by quantitative historians. Once a statistical inquiry's “categories had been subjected to criticism and reinterpretation,” Scott left open the question of whether or not the actual data contained in the reports could be used “in a statistically critical fashion.” Did the numbers have anything left to tell us about the extra-textual world? For Sewell, “the fact that the question was not posed … speaks volumes about the epistemic assumptions of history's cultural turn.”12 Of course, in the 1980s, Scott was already embarked on a trajectory of unusual epistemological radicalism. But historians across the discipline have turned away from materialist notions of “the social” in favor of a textual-linguistic conception of culture. Quantitative sources and the question of what they might tell us about the world have become a minority interest.

Although Sewell is properly appreciative of what we have learned since the 1980s, in Logics of History he presents a powerful critique of the “textual turn.” Instead, he argues that we need to expand our understanding of language games to include a far wider range of meaningful expression. In addition, he calls for a reengagement with what he suggestively calls the “built environment,” the world that people make and communicate about. On this ontological foundation, Sewell calls for a return to questions of macro-explanation. And in this connection he insists that if we are not to deprive ourselves of an essential basis of historical judgment, we must reengage with quantitative source material.13

Sewell's critique is certainly welcome. But it is also provocative. In particular, one cannot help but wonder at his willingness to ground his bipartite description of humanity as homo faber and homo sapiens on what he refers to without hesitation as the “nature of things.”14 This, Sewell admits, is an “Enlightenment conceit.” And it would seem to stand in contradiction to his emphatic embrace of the nineteenth- and twentieth-century hermeneutical tradition, for which the question of such “natural” foundations is a fundamental problem.15 Furthermore, although one can only agree with Sewell that we should reject any fundamental epistemic divide between quantifying and non-quantifying approaches to history, the question is how to effect this reengagement without relapsing into naïve positivism. We have learned in the last generation that statistics have a history. Science studies has illuminated the profoundly perplexing, Janus-faced aspect of modern fact-making, helping us to understand how facts might appear at some times as self-conscious constructions and at other times as mere reflections of reality. But with regard to contending histories of the early Weimar Republic, even historians who accept Sewell's basic premise of epistemic continuity find themselves facing difficult questions of method and narrative style when handling contested historical statistics. And if we wish to write critical, political histories, our problem is even more severe. If we want to probe the options open to historical actors, then, as Sewell argues, we have little option but to actually employ the statistics we know to be historically constructed as tools for dissecting historical reality. Insofar as there are solutions to these many problems, they clearly cannot lie in strategies of disciplinary purity buttressed by dubious epistemological posturing. The solutions should surely be sought in a widening and deepening of our common strategies of interpretation.

There seems no better field in which to explore this complexity than the history of the Weimar Republic. The Weimar Republic was one of the protean moments of modernity.16 Sophisticated but often violent debate raged over every aspect of the ordering of politics, culture, and society. Not for nothing can the first German democracy claim to be one of the birthplaces of the modern critique of ideology and the cynical discipline of the sociology of knowledge.17 Not surprisingly, the economy bulked large in many of these arguments.18 The high-pressure fusion of political drama and theoretical controversy has made both the final crisis of the Weimar Republic between 1929 and 1933 and the earlier inflationary disaster of 1918–1924 into perennial topics of scholarly debate.19 And in these arguments, questions of fact are simply inescapable. Interweaving stories about the production of Weimar's trade statistics with the story of their use by both contemporaries and historians demonstrates the fruitfulness of an approach that, following Sewell, we might call hermeneutical quantification.20

In the autumn of 1922, the birth crisis of the Weimar Republic was reaching its climax.21 Eighteen months earlier, on May 5, 1921, Germany had faced the draconian London reparations ultimatum. Desperate to avoid further Franco-British sanctions, including the military occupation of large parts of western Germany, a center-left government headed by the populist Center Party politician Joseph Wirth had decided to continue the policy of “fulfillment” pursued since the acceptance of the Treaty of Versailles in 1919. Wirth, along with his liberal and social democratic coalition partners, hoped that a show of compliance would yield concessions from the reparations creditors.22 By 1922, this commitment had cost the lives of Finance Minster Matthias Erzberger and Foreign Minister Walther Rathenau, victims of right-wing gunmen egged on by vicious antisemitic invective.23 And no concessions had resulted from Britain and France. Wirth and his coalition partners thus faced a crucial choice: Should they make a new and serious effort at domestic consolidation, in the hope that this further show of goodwill would persuade the Entente to reduce reparations? Or had the gamble on democracy at home and cooperation abroad, the gamble on which the Weimar Republic had been founded, already failed? Was there really no alternative to confrontation? On both sides of the Rhine, nationalists were bracing for a new resort to violence.24 To avoid the catastrophe of a Ruhr invasion, it seemed that a fundamental decision on economic and financial policy could no longer be avoided, and this raised the question of what state the German economy was actually in.

At this pivotal moment in Germany's history, on October 24, 1922, the embattled chancellor took time to write the following letter to his minister of economic affairs, the Social Democrat Robert Schmidt:

The entire basis of the “fulfillment” strategy on which Wirth and his colleagues had staked their country's future was thrown into question by the apparent evidence that Germany's accelerating inflation had resulted not in a trade surplus, but in a deficit of more than 1 billion prewar Gold marks. As Wirth insisted,

For anyone interested in the role of statistical information in modern government, Wirth's letter offers a ready-made case study of the interaction between power, on the one hand, and the technicalities of statistics, on the other. Here we have the German chancellor insisting in an urgent letter addressed to a senior ministerial colleague on the need to open the black box of his nation's trade statistics.26 It is a source that confronts us in the most direct way possible with some of the basic and powerful questions asked by recent reflections on the history and sociology of statistics, notably the work of Alain Desrosières, Laurent Thévenot, and Theodore M. Porter. In Statistics and the German State, 1900–1945: The Making of Modern Economic Knowledge (Cambridge, 2001), I developed this approach in relation to the history of German economic statistics.27 The crisis of 1918–1923 appeared in that narrative as a moment of fundamental transition from the Wilhelmine era, with its culture of wide-ranging professorial engagement (Gelehrtenpolitik), to the new era of technocratic policy-science epitomized by institutions such as the Institute for Business-Cycle Research (Institut für Konjunkturforschung), founded in conjunction with the Reich's Statistical Office (Statistisches Reichsamt) in 1925. Although not cited explicitly in Statistics and the German State, the problems of Germany's trade statistics can clearly be located within this wider process of governmental transformation. It was problems such as those facing Wirth that caused the Weimar political class to invest in a new system of national economic statistics with which to underpin a new apparatus of national economic policymaking. In the process, statisticians and their clients both in and outside government reconceived the state's relationship to civil society in the medium of economic information.

Viewed in these terms, the history of statistics certainly yields important insights into the history of the modern state and its evolving use of expert knowledge. But, like Scott's essay, Statistics and the German State did not address Sewell's follow-on question. The construction of the statistical organization as a system of power and knowledge was at the center of attention, whereas the statistics as historical sources were left largely uninterrogated. To avoid confusion, it is worth stressing that this silence did not arise from an epistemological critique, from doubts about our access to a world beyond the text. Indeed, in analyzing modern economic knowledge in historical and pragmatic terms, Statistics and the German State, like the French work that inspired it, was in fact attempting to sidestep the entire epistemological problematic. Having located the subject of statistical knowledge in history, we were not about to reinstate a Kantian dichotomy and place the statistical object beyond human reach. This literature's failure to address itself more closely to the actual content of the historical data arose rather from the original, productive “shift” with which any such history of statistics must begin, namely the insistence that there is more to numbers than the question of “accuracy.” The polemical energy involved in tearing up the empirical foundations of economic and social history, to reveal them as rooted in institutional, political, and indeed economic history, overshot its mark. Paradoxically, by casting the question of “accuracy” aside, we disarmed ourselves. If our aim is to de-reify the data and to demonstrate the centrality to modern societal organization of the processes through which quantitative truth claims are made and put to use, then there is in fact no better way to do this than to focus as closely as possible on the disputes that arise in relation to particular truth claims at moments of crisis. What caused no lesser figure than the German chancellor to occupy himself with the organization of the Statistical Office was the fact that he desperately wanted to know the state of the balance of trade. And if we want to go further in understanding how knowledge interacted with power, we should address ourselves to the frantic efforts made by his subordinates to answer this question. We should overcome our inhibitions and move from a generalized history of statistics as a form of governmental knowledge to a history of the construction and use of particular facts.

The question of whether Germany was in surplus or deficit was of such vital importance to Wirth and his colleagues because it was a fundamental factor in the argument over Germany's ability to pay reparations. If Germany was to make real transfers of reparations—that is, if it was not to pay reparations simply by taking up new credits from other lenders—then it needed to achieve an excess of exports over imports; in other words, it needed a trade surplus. This accounting identity was beyond dispute. But what implications it carried for German policy depended on how the balance of trade was inserted into a wider worldview.

For nationalists such as the economist and politician Karl Helfferich, it was axiomatic that Germany, as an overcrowded country with insufficient “living space” and only coal by way of raw materials, was a nation fated to a “structural” trade deficit.28 The concept of economic “structure” became newly fashionable in the 1920s as an intellectualized expression of the nationalists' refusal to accommodate themselves to the postwar order. The term was used to describe the unalterable, entrenched difficulties facing any country attempting the kind of adjustment that was assumed by liberal economic theory, but which was denounced as impossible for Germany.29 And the nationalists had data to back them up. In the five years before World War I, despite its reputation as a champion exporter, Wilhelmine Germany's average annual trade deficit had been 1.46 billion Marks. To finance this “structural” imbalance, Germany had relied on “invisible earnings” from shipping, insurance services, and property income totaling as much as 1.75 billion Marks per annum.30 In a good year such as 1913, this enabled Germany, despite its trade deficit, to achieve a healthy current account surplus, adding further to its holdings of foreign assets. The war had destroyed this equilibrium. German exports had collapsed. Furthermore, the main sources of “invisible” income had been stripped away by the vengeful foe. To nationalist opponents of reparations, it was clear that it was impossible for Germany to achieve the underlying surplus of foreign exchange out of which it might pay reparations. Attempting to pay would involve throwing newly printed German money onto the exchanges to buy foreign currency. This could only engender inflation and domestic social instability. To avoid this disastrous scenario, Germany should defy the Entente and renege on the treaty.

The attraction of the nationalist position was that it proclaimed Germany's inability to pay as a quasi-natural fact. But if taken seriously, it carried with it the risk of catastrophic confrontation with Britain and France. Furthermore, it flew in the face of conventional liberal trade theory, which insisted that the balance of trade was not a structural given, but was determined by the balance of competitiveness. This, in turn, depended on prevailing international exchange rates and the price of goods in Germany relative to those abroad.31 Prices and exchange rates were at least to some degree within the power of the German government to control. Crucially, the nationalist position was vulnerable to empirical refutation. If the structural reading of Germany's economic situation was to have any credibility, there could be no trade surplus.

And it was precisely on the possibility of achieving a trade surplus that the pragmatic foreign policy of the Weimar Republic had rested since 1919.32 The so-called policy of “aggressive fulfillment” took it to be essential that Germany should make an effort to pay reparations, because otherwise it would face sanctions, threatening the very survival of the German state. If Germany started from a position of trade deficit, then the initial efforts to buy up foreign exchange would certainly depreciate the Mark, as the nationalists feared. But once this depreciation gathered momentum, it would unleash a wave of highly competitive German exports that would drive British, French, and American goods out of world markets. As John Maynard Keynes predicted in his global bestseller The Economic Consequences of the Peace, Germany could achieve the trade surplus necessary to pay reparations, but only at the expense of mass unemployment in the countries demanding reparations from it.33 Keynes feared Bolshevism. The Germans hoped that Britain and France would see reason and reduce their reparations demands.

Desperate as the situation clearly was, the advocates of aggressive fulfillment insisted that Germany still had some control over its own fate. But it was clear that theirs was a strategy fraught with risk, and not one that could be continued indefinitely. Managing an aggressive depreciation was a delicate balancing act. If the German currency depreciated too far and too fast, the entire country would be vulnerable to being bought up by rapacious foreign speculators. Alternatively, if German producers and workers got wise to the depreciation and began to adjust their wages and prices too rapidly in the opposite direction, then rising costs would offset any advantage to exporters. Germany might even see its competitiveness eroded, encouraging imports rather than exports. In any case, regardless of the effect on international competitiveness, any severe depreciation of the currency was bound to damage savers and undermine the stability of bourgeois society and long-term investment. Given these daunting risks, everything depended on skillful economic management and the speed with which Germany's diplomats could extract concessions from Britain and France.

We of course know that France refused to budge, and the punitive occupation of the Ruhr in January 1923 resulted in a hyperinflationary disaster that jeopardized the long-term prospects for political stability in Germany.34 Not surprisingly, this has focused historians' attention on a third option—the possibility, after a period of depreciation, of stabilizing the German currency and domestic prices at levels sufficiently competitive to enable the long-term repayment of reparations. Given the damage that aggressive fulfillment was doing to Germany itself, it was the choice between outright confrontation and stabilization that Weimar's politics actually revolved around between 1920 and 1924. And it was this choice that was at the heart of two major historical works on the German inflation that appeared virtually simultaneously in the mid-1990s, Gerald Feldman's magnum opus The Great Disorder and Niall Ferguson's ambitious first book, Paper and Iron. Many considerations entered into the stabilization debate. But in assessing the feasibility of stabilization, everything again depended on the balance of trade. Given the damage that “shock therapy” might do to Germany's fragile polity, economic stabilization could be attempted, it was agreed in Berlin, only if the trade balance indicated that the country was already reasonably close to a sustainable surplus.35 Indeed, so central is this question of the trade balance that Feldman and Ferguson both felt it necessary to make brief forays into the history of the Statistical Office, both citing Wirth's letter to Schmidt of October 24, 1922. Revealingly, however, they approached the question of Weimar's trade statistics in different ways.

As Feldman demonstrates in his extraordinary epic, the tragedy of Weimar politics originated in the fact that although stabilization had by 1922 come to be widely seen as the only real alternative to apocalyptic confrontation, its supporters could not agree on how it should be implemented. Big business predictably favored a “capitalist” deflation, based around a rollback of the concessions won by German labor in 1918–1919, a return to longer working hours, a curtailment of trade union power, cuts to welfare spending, and reduced taxes. In the summer of 1920, Britain and the United States had imposed deflations along these lines.36 Germany's political system was of course far weaker than theirs. So the more appropriate analogy is perhaps Italy, which also adopted deflation in 1920.37 The resulting financial crisis was to contribute to the erosion of the liberal center ground and the emergence by 1921–1922 of a fascist threat.38 Mussolini's March on Rome took place less than a week after Wirth penned his exasperated note to Schmidt.

Not surprisingly, Wirth could not persuade his coalition partners to take the risks involved in this path. As an alternative, elements in the German labor movement and the Ministry for Economic Affairs developed a strategy of what might be termed social democratic stabilization, involving a state-managed effort to engineer a trade surplus by means of import controls, backed up by an egalitarian program of progressive taxation, rationing, and domestic redistribution.39 If pushed for historical analogies, one might point to the British Labour government after 1945, with its policy of “austerity.”40 It was a strategy along these lines that Robert Schmidt and his allies—notably State Secretary Julius Hirsch—continued to advocate into the autumn of 1922, up to the point at which Wirth's government collapsed, and Wilhelm Cuno and the advocates of nationalist confrontation took power.

The grandeur of Feldman's Great Disorder derives not only from the depth of its scholarship and its biblical extent, but also from its tragic sense. Feldman plunges us into the full complexity of the political and economic argument in Weimar Germany, and without adopting a safe point of distance. He provides a panorama of the actions and thoughts of dozens of protagonists from all sides, yet he is clearly not indifferent between the parties. Against the irrational, nationalist antisemites of the right, Feldman sets a compelling portrait of those who struggled for stabilization, and particularly those who advocated the policy of social democratic stabilization.41 For Feldman, the period between October and November 1922, at which point the Social Democrats withdrew from Wirth's coalition and his government collapsed, was “the last opportunity for stabilization that has genuine historicity.”42

But what, then, are the implications of Wirth's cry of confusion and dismay for Feldman's narrative? The advocates of a social democratic stabilization needed to show that there had been at least some improvement in Germany's trade position by the autumn of 1922. Wirth clearly wished that he could simply have relied on the relatively optimistic data coming from the Reich's Statistical Office. The prospect of having to intervene in the statistical laboratory itself filled him with a mixture of frustration and anxiety. The historian who perches on the shoulders of men such as Wirth cannot but share this mixture of emotions. For anyone who remains committed to the project of constructing encompassing macro-narratives, a source such as Wirth's letter is a trap door that opens onto entire realms of irreducible and threatening complexity. After all, to make the judgments that propel narratives such as Feldman's, we need to form some idea of the “real” constraints acting on men such as Wirth. But here is the rub: more often than not, it is on historical statistics that we must rely to define these constraints. In the autumn of 1922, it was not just Wirth's world that was thrown into confusion, but the world of his historians as well.

Of course, Feldman was far too scrupulous a scholar to have simply swept Chancellor Wirth's letter under the carpet. Indeed, he describes it as a “document of exceptional importance,” citing it at length to indicate Wirth's increasingly impatient mood by the autumn of 1922 and his willingness to doubt the conventional wisdom of fulfillment.43 But Feldman never addressed head-on the substantial question of the accuracy of the trade data. Instead, a few pages after his discussion of Wirth's outrage, Feldman cites other numbers compiled by the Reparations Commission, which purport to show that despite Wirth's anxieties, Germany's trade balance was actually close to equilibrium. (See Table 1.)44 This is clearly helpful to Feldman's position. But how these figures could be reconciled with Wirth's anxiety that Germany might in fact be running a large trade deficit is a question that Feldman leaves unexplored.

TABLE 1 Germany's Contested Trade Balance 1913–1924 (Million Gold Marks)

Open New Window

Like the exclusively political treatment of economic statistics offered in Statistics and the German State, but for different reasons, Feldman fails to reach closure. He acknowledged that knowing the state of the trade balance was crucial to Wirth. He acknowledged that there were fundamental problems with the data. Furthermore, he could not bracket Wirth's doubts as the passing anxieties of a harassed politician, because his entire massive oeuvre hinges on the claim that stabilization in 1922 was a real possibility, and this depended on there being a favorable balance between exports and imports. And yet Feldman provides us with no truly convincing grounds for suspending Wirth's doubts about the optimistic picture presented by the Reparations Commission.

In Feldman's defense, one could reply that he was simply staying true to the conventions of the genre within which he wrote. To confront his readers with a complex technical discussion of the data would have been out of character with the grand sweep of his narrative. Furthermore, it might have required Feldman to adopt the point of view of a truly external narrator commanding a position of omniscience above that available to his protagonists—a position that he generally refused. It is this refusal, in turn, that grounds the sympathy that he maintained with even his more disagreeable characters. As Feldman never lets us forget, most of what we know as historians, we owe to the efforts by historical actors to make sense of their surroundings.45

But to a self-confident economic historian firmly camped on the quantitative side of the methodological divide, much of the preceding discussion is likely to seem merely otiose. The empirical question that both Statistics in the German State and Feldman's Great Disorder skirted around is the kind of question that is meat and drink to quantitative historians. Cognizant of relevant technical problems, we must simply apply ourselves to the relevant published and unpublished sources and come up with our best guess as to the trade balance.46

There are, of course, numerous histories of the German inflation written in this objectifying mode.47 However, for a near-perfect contrast to Feldman, we need look no further than Niall Ferguson's early work Paper and Iron. Superficially, the main difference between Ferguson and Feldman is political. Whereas Feldman's sympathies are with the advocates of what we have dubbed a social democratic stabilization, and whereas Feldman is still willing to hold open the possibility of a stabilization in the autumn of 1922, Ferguson argues that the debate on fiscal reform in Germany between May 1921 and November 1922 was “to some extent phoney … as the Chancellor himself was not in earnest about trying to balance the budget.”48 Instead, Ferguson argues for a stabilization along capitalist lines to have been implemented in the first half of 1920. With the United States and Britain leading the way, 1920 witnessed an unprecedented worldwide deflation. Even Germany's inflation came to a temporary halt, ushering in a period of so-called “relative stabilization.” Ferguson acknowledges that the retrenchment necessary to make this stabilization permanent might have occasioned a bout of protest from the left. But on the basis of his Hamburg case study, he judges that fears of revolution were exaggerated. Furthermore, whatever repression might have been needed to deal with the Communists was more than justified in light of the damage done to the fabric of German society by the subsequent hyperinflation.

The political stakes in this argument could hardly be higher. The history of the Weimar Republic is written under the shadow of Nazism. Feldman's ultimate goal was to show how German social democracy might have retained its determining influence in the Weimar Republic. Ferguson, by contrast, is trying to map a consistent capitalist route around the disaster of National Socialism. But they are also separated by subtle yet profound differences in methodology and narrative emplotment. By contrast with Feldman's equivocation on the matter of the data, Ferguson follows Wirth's letter to its logical conclusion. The official statistics that showed a reassuring trade surplus were bogus. Casting himself in the role of the pioneering economic historian, Ferguson sets out to “rework” the data, producing a new time series that confirms Wirth's worst fears. The strategy of aggressive depreciation had backfired.49 Rather than driving German exports into foreign markets, the surge in domestic inflation that had accompanied the external depreciation of the Mark had reduced competitiveness and generated a gaping trade deficit. The only moment at which the costs of stabilization might have been bearable was 1920, in which year Ferguson, too, believes that Germany achieved a small trade surplus.

By contrast with the painstaking subtlety of Feldman's magnum opus, Ferguson cannot be accused of taking his historical protagonists too seriously. The reference to Wirth's efforts at stabilization as “phoney” is characteristic of the caustic treatment he metes out. At one point, Ferguson even indulges in sensationalist speculation about the way in which a supposed homosexual attraction between Keynes and the German negotiator Carl Melchior might have influenced the campaign against reparations.50 But for all the journalistic color, Ferguson is absolutely consistent in his methodology. First, resolve the quantitative issue. Produce new data, which expose the real failure of the strategy of fulfillment (Table 1). Then, with Excel-generated graphs to hand, hold up the weak-kneed political strategy of the Weimar government for indictment. We are presented with a consistent narrative of ironic, critical distance. What must concern us are the costs of this distance. In particular, Ferguson fails to fully explain how his new data for 1922 could differ by such a wide margin from previously accepted accounts. Ferguson, like Feldman, claimed that his numbers ultimately derived from official sources, in his case the monthly trade statistics of the Reich. But were these not contaminated by the same problems that had rendered previous judgments uncertain?

Clearly, statistics are troublesome material for historians. The three histories we have discussed all attempt to make use of the data, but each does so in a way that exhibits characteristic blind spots. By insisting that statistics were made and made for a purpose, Statistics and the German State ended up deflecting attention away from the descriptive content of the data. In constructing his epic of political struggle, Feldman needs the data to provide a factual frame, but cannot make them freeze into a solid background. Ferguson, qua economic historian, uses his command of the quantitative sources to assert a vantage point of distance, but can do so only by ignoring the fact that his data are in fact fragile historical artifacts. In the final analysis, economic historians, like the rest of us, feed off the scraps left by contemporaries, which sucks them into a dependence that their narrative posture denies.

Is there not a hidden logic to these recurring frustrations? Are we not dealing here with the consequences of what Bruno Latour famously referred to as the Janus-faced quality of modern knowledge?51 On the one hand, statistics of society and the economy are self-consciously and very overtly the products of a labor of construction. On the other hand, once established, they become objective facts, mere passive reflections of an external economic or social reality. Our three histories are trapped by this duality. Statistics and the German State writes the history only of statistical construction. Ferguson, by contrast, treats the data as the objects the statisticians wished for them to become. Both positions are consistent, but at the price of seeing only one “face” of the process. Feldman seeks to straddle the divide, but ends up with an argument marred by inconclusiveness. We can escape this impasse by constructing a history of the politics of the German trade balance that takes in both sides of this Janus-faced reality, seeking to account for both the constructedness and the givenness of historical statistics. To paraphrase Latour, what we need is a three-dimensional, stereoscopic rendering of the recalcitrantly bifurcated history of fact-making.52

If we pursue the question of the German trade balance back through the archive, it becomes clear that the problems for the statisticians did not start in 1922. Their first brush with nationalist indignation came two years earlier. In 1919, there had been little doubt that Germany had run a substantial trade deficit. Imports had rushed in to satisfy pent-up demand, and exports were far below prewar levels. But in the autumn the exchange rate collapsed, making German goods more competitive, and in December 1919 the Ministry of Economic Affairs imposed strict trade controls. Then, in the early summer and autumn of 1920, the German currency strengthened on the exchanges. There was little doubt, therefore, that the trade balance had improved. But by how much? For much of 1920, the diplomats were locked in a series of difficult negotiations over reparations, at Spa and then Brussels. Despite the signs of stabilization, Germany's negotiators insisted that the trade balance remained vulnerable and that any large reparations demands were likely to result in an immediate currency crisis and renewed inflation. The Allies, by contrast, insisted that the signs of improvement were real and that the healthy trade balance could become permanent, if only the German government was willing to raise taxes and cut spending.

When the trade figures for 1919 and early 1920 were first released, they offered little support for the German position. From early 1920, they showed a steady improvement in the balance of trade, culminating in April and May 1920 with an export surplus of more than 1.6 billion Marks.53 True to its constructive attitude toward fulfillment, the statisticians' chief political patron, the Ministry of Economic Affairs, did not flinch. The numbers were included in an official German memorandum to the Brussels Conference, with no more than an obscure technical appendix by way of explanation. But the nationalist press in Berlin was outraged. The incompetent statisticians had robbed Germany's delegation of its best defense.54 Meanwhile, the French and Belgians seized on Germany's healthy trade data to argue for increased reparations.55 Under pressure from the representatives of the member states and the Reichsbank, the embarrassed Reich's Ministry of Economic Affairs was forced to issue an exculpatory press release.56

This was first and foremost a political necessity. However, the nationalist critique had in fact exposed a genuine technical problem, a problem rooted in the nineteenth-century origins of official statistics. The system of trade statistics that the Weimar Republic inherited from the Kaisserreich was a weird hybrid of modern census technology and the “statistique des notables” of the early nineteenth century.57 Since the creation of the German nation-state in the 1870s, the basic recording of international trade had been done by the customs posts that compiled summaries of cross-border shipments for the Imperial Statistical Bureau.58 These, however, were records of physical weight, not value. To get from the tonnage of trade to values, the statisticians needed estimates of average prices for each class of goods. The necessary advice, documentation, and guesswork were provided by an annual meeting of the so-called Consultative Committee for Trade Statistics. This was staffed by an array of businessmen, managers, and “experts” delegated from trade associations and chambers of commerce. By the early twentieth century, this rough-and-ready procedure for estimating Germany's balance of trade had become the object of increasingly vocal criticism and was modified so that on the export side, the system now operated more formally. Exporters were required to declare both the tonnage and the value of their shipments, and the questionnaires were no longer compiled locally, but were shipped en masse to the Statistical Office for central processing with Hollerith punch-card machines. Just prior to 1914, a similar system of direct reporting was also planned for imports, but its implementation was held up by the inability of the Statistical Office to cope with millions of additional returns.

The inflationary conditions of the postwar period cruelly exposed the imbalances in this system of trade statistics with one leg (exports) in the twentieth century and the other (imports) mired deep in the early nineteenth century. In the autumn of 1920, with the nationalists railing against the trade surplus, the Statistical Office realized that different modes of reporting might seriously bias the results. Whereas exports were reported directly to Berlin on a weekly basis, and their values therefore kept up with inflation, imports were still valued annually by the Consultative Committee at out-of-date prices. Despite appeals from the statisticians, the interest groups had refused to provide staff for more regular meetings. The resulting undervaluation of imports biased the overall result toward a trade surplus. The organizational solution was obvious: to implement the long-overdue trade statistical reform, to require value declarations for both imports and exports, and to implement centralized processing for both sets of returns. It was only from the spring of 1921 that this fully symmetrical system of reporting began to operate. And it generated little controversy, because in 1921, with the rest of the world lapsing into depression, there could be no question that Germany did run a substantial trade deficit.59 But as soon as the surplus returned in the early autumn of 1922, so did the protests. And again it was the Reichsbank, with its intensely nationalistic leadership armed with a large staff of economic experts, that was the first to react. On October 10, Vice-President Otto Georg von Glasenapp of the Reichsbank gave a carefully edited interview to the United Press of the United States, insisting that “although on paper Germany showed a favorable trade balance in August, the exact reverse was really the case.”60 Behind closed doors, the language was less measured. On October 13, the hapless representatives of the Reich's Statistical Office were subjected to a withering attack by von Glasenapp, who accused them of nothing less than sabotage.61 A day later, it was the Reichsbank that alerted Chancellor Wirth to the “truly shocking fact that the German import statistics are completely wrong and virtually useless.”62 A week after Schmidt received his humiliating missive from Wirth, the Reichsbank followed it up with another memorandum stressing that figures showing Germany's true trade deficit were “a question of life and death for the German people and the German economy.”63

The mood in Berlin was clearly overwrought. But at first, the Reich's Ministry of Economic Affairs did not yield to the nationalist bombast. State Secretary Julius Hirsch and his colleagues conceded that there were technical difficulties with the data, and that Germany might well have been suffering a modest deficit in 1922. But relative to 1921, the balance had clearly improved. Furthermore, in treating the structural deficit as the unalterable source of Germany's economic difficulties, the Reichsbank and its nationalist supporters were confusing cause and effect. A surplus not only was possible, but could be made permanent if Berlin was willing to implement a determined domestic stabilization.64

But in the first days of November 1922, the advocates of stabilization lost the political battle. Wirth's government collapsed, to be replaced by the right-wing Cuno administration. It rejected domestic stabilization and steered Germany toward a showdown in the Ruhr. Hirsch resigned from the Ministry of Economic Affairs under threat from the same menacing antisemitic libel that had legitimized the killing of Erzberger and Rathenau. In less dramatic circumstances, the head of the Department for Trade Statistics in the Reich's Statistical Office was sacked.65 And the Statistical Office again changed its story.

In a detailed memorandum, for which the new minister of economic affairs took full responsibility, the estimates of the trade balance for 1922 were completely revised.66 And again, a good technical explanation was found.67 This time the problem was located not in the statistical machinery itself, but in the asynchronicity in the payment systems for foreign trade induced by the inflation. To avoid the impact of currency depreciation, German merchants sought to pay as early as possible for imports. Exporters, by contrast, took payment in foreign currency at the last moment possible. With the German currency in free fall, this slight time lag was enough to seriously understate the value of imports relative to exports. The solution, ironically, was a return to the old system of consultation. From November 1922, the national data emerged from a monthly meeting of experts at which the import returns were “adjusted” to bring them in line with current prices and exchange rates. The result, satisfyingly enough, was to show that between January and September 1922, rather than a steady improvement culminating in a surplus, Germany had in fact piled up a cumulative deficit of 1.5 billion Gold marks. The orthodox view was confirmed. Germany was fatally destabilized by a structural external imbalance.

This defeat for the advocates of fulfillment has echoed down the generations. The new data series that Niall Ferguson constructed to make his case against Gerald Feldman, in fact, derives directly from this “corrected” trade data published after October 1922 in the Monatliche Nachweise of the Statistisches Reichsamt.68 Unaware of the precise circumstances of their production, it was on numbers produced to satisfy the nationalists that Ferguson rested his case. Of course, Ferguson's purpose is quite different from that of Helfferich et al. He argues not for confrontation in 1923, but for stabilization in 1920. As we have seen, the nationalists insisted that there had been a trade deficit even in 1920. But Ferguson and the nationalists do agree on at least two points: their rejection of the entire strategy of aggressive fulfillment and their insistence that the “facts” refute the possibility of a stabilization in 1922, precisely the possibility that State Secretary Julius Hirsch and latterly Gerald Feldman wanted to hold open.

At this point, one might conclude that the history of German official statistics in fact reduces to a political power play, and that our journey through the interstices of the statistical apparatus has been a pointless detour. The numbers cannot be treated as independent evidence for either side of the argument, because they were nothing more than a political pinball. In the limit, we would be left with no basis on which to form a quantitative judgment about the state of the German economy. But this is reductive. The technical arguments advanced for each correction, although clearly convenient, deserve to be taken seriously. Embattled and fragile as it was, the Reich's statistical system retained at least a measure of autonomy, an autonomy best indicated by its continued ability to generate unwelcome surprises for the impatient politicians and civil servants, not only in 1920 and 1922 but again in 1923.

The evidence of an unalterable trade deficit suited the new government of Chancellor Cuno and his cabinet of nationalists, who replaced Wirth, Schmidt, et al. But it was anathema to the Entente. And because Germany continued to report the tonnage of trade as well as the hotly contested value figures, it was possible for further rounds of debate to ensue. In early 1923, British, French, and Italian experts on the Reparations Commission found a new series of problems, this time not in the import but in the export data. Comparisons of the papermark values declared by German exporters for their consignments at the customs post with the domestic prices charged for similar categories of goods indicated that German exporters were declaring artificially low prices. And German experts had been making similar discoveries of their own.69 The Reich's Statistical Office noted large discrepancies between the value of identical export consignments declared in papermarks, as opposed to those declared in inflation-proof Gold marks. It appeared that those exports valued in papermark terms were being recorded at the official minimum price rather than their actual market value. The consequence of these discoveries for the assessment of Germany's trading position was dramatic. After the scandal of 1922, the Statistical Office had rejiggered its procedures to arrive at an estimated deficit of 2.23 billion Gold marks for the entire year. By contrast, the Reparations Commission, allowing for the undervaluation of exports, estimated a deficit of little more than 250 million. This was very close to the figure that Hirsch and the Ministry of Economic Affairs had sought to defend in the autumn of 1922. Seventy years later, these were the numbers that Feldman cited to buttress his case for the possibility of stabilization. By July 1923, the archive reveals that the Reich's Statistical Office had convinced itself that the real figure for 1922 was a surplus in excess of 150 million Gold marks.70

This time, however, the Statistical Office did not make the mistake of casually announcing the inflammatory news to the general public. Instead, a series of confidential meetings were summoned under the auspices of the Ministry of Economic Affairs. The ensuing debate was full of bitter irony. The Ministry of Economic Affairs, the Statistical Office, and the Finance Ministry all insisted that since “no one abroad” regarded the current German trade statistics as anything other than self-serving eyewash, something had to be done to restore their credibility. The Reichsbank admitted that the numbers published following its dramatic intervention in the autumn of 1922 were probably useless. But it managed to secure agreement that they should be spared a further humiliating round of public revisions.71 It was left to the liberal economist Anton Feiler and former minister of economic affairs Schmidt, both of whom had been among the most important advocates of stabilization in the autumn of 1922, to spell out the devastating political implications of the new data: “The false statistics produced hitherto [since November 1922] were the basis for inactivity in fiscal and monetary policy and the basis for current economic policy.” At the time of the meeting in the summer of 1923, inflation had reached astronomic rates, and Germany was threatening to descend into anarchy. As Feiler and Schmidt insisted, “Doing away with this [statistical] foundation would put those in the right who had all along demanded a different policy”—in other words, a policy of stabilization.72

Once more, driven by the logic of its own internal system of relations, the statistical reporting system had overturned the nationalist shibboleth of a structural trade deficit. And it would be convenient, of course, if the argument had ended with this final set of revisions. In the face of continuing opposition from the Reichsbank, however, the news of the trade surplus was held back.73 Instead, a third set of figures was published, based on the notional exercise of applying 1913 unit prices to the volumes of the early 1920s.74 These showed the modest trade surplus that would have resulted if Germany had been able to import and export at prices prevailing in the halcyon Wilhelmine era. The statisticians felt able to defend these numbers because their method of calculation was at least transparent and formally consistent. The nationalists accepted them through gritted teeth, but only because it was obvious that they had no practical bearing on the situation as of 1922–1923.75 Furthermore, the Reichsbank still had room to insist that given the change in the terms of trade since 1913, at current prices a favorable trade balance was not plausible.76 It was not until 1924, in the comparative calm following the end of inflation and the reparations settlement brokered by the Americans, that the Reich's statisticians were finally permitted to publish a definitive official version showing a relatively healthy balance for both 1922 and 1923.77

Opening the black box of complexity that lurks within the statistical system has not, therefore, yielded consensus. Supporters of Feldman's interpretation may wish to make use of the paper trail we have unearthed to bolster their case. The view that the trade balance was in surplus clearly had strong support both inside and outside Germany. By contrast, the data on which Ferguson sought to base his case—the revised figures of the autumn of 1922—were in the end abandoned even by the Reichsbank. Nevertheless, the argument continued. Certainly, given the general unreliability of all the numbers and the convictions of well-informed contemporaries such as the Reichsbank, Ferguson's position cannot be conclusively rejected. If one contrasts the deflation in Britain and the United States with Germany's inflationary conditions, a trade deficit in 1922 can hardly be ruled out.

But if we cannot break the deadlock from internal archival sources, why not address Ferguson in his own terms? Why not ask the economic historians' question? Is there not some means of gaining “external” purchase on this question? Of course, for evidence about the state of the German economy, one is generally confined to German statistical sources, which in the period between 1918 and 1923 all suffer from the same profound problems. But in the case of foreign trade, this does not apply. German exports and imports, although widely dispersed, were concentrated on a manageable number of large economies. And in the trade of those countries, the German share was significant. In principle it should therefore be possible to construct at least an approximation of the German trade balance by compiling the figures for trade with Germany from the accounts of its major trading partners.

Efforts in this direction have been made in the past. Carl-Ludwig Holtfrerich examined Germany's trade balance with Britain, France, and the United States and came to the conclusion that Germany provided those economies with welcome export opportunities during the deflation of 1920–1921.78 With these countries, at least, Germany's trade balance certainly did deteriorate after 1920, giving support to Ferguson. However, Holtfrerich says little about 1922. Furthermore, given the truly global reach of Germany's trading connections, a three-country sample is too small to give an accurate picture of the overall balance. We need a more comprehensive overview.

Apart from its laboriousness, there are difficulties facing any such attempt. The system of international trade accounts is not fully consistent. One should not expect to find exports recorded in the German data as going to France showing up one for one in the French import account. To make matters more complicated, Germany was not the only country suffering from monetary disorder. After the suspension of the gold standard system in 1914, all the economies of the world suffered more or less severe inflation, and their exchange rates fluctuated in some cases quite violently. We must therefore make do with translating the trade accounts of Germany's largest trading partners into dollars, the most stable world currency. Also, if we are not to lose sight of our earlier argument, we must of course constantly remind ourselves that in compiling these data from standard handbook sources, we are relying on precisely the same kind of bureaucratic-technical apparatus that we have just deconstructed in Berlin. The assumption on which we proceed is simply that these other statistical agencies were in better shape than the Statistisches Reichsamt. (See Table 2.)

TABLE 2 German Trade with Major Commercial Partners 1913–1935 in m $

Open New Window

As our new trade data show, the depreciation of the German currency clearly did not create a huge surge of German exports into the markets that the advocates of aggressive fulfillment were particularly targeting, namely France and Britain, Germany's major reparations creditors. As Table 2 confirms, Germany's balances with those countries in fact deteriorated. However, these data also suggest that this was not typical of all of Germany's trading relations. With the smaller European economies, most notably the Netherlands, Germany achieved a trade surplus. And this is confirmed by unpublished data compiled by the Reich's Statistical Office, which showed exports to the Netherlands in the latter half of 1921 exceeding those to Britain, France, and the United States combined.79 If the focus, therefore, was less on applying political pressure than on achieving the overall trade surplus necessary to finance reparations, then the evidence suggests that the strategy was more successful than Ferguson is willing to admit. The balance of both internal and external evidence supports Feldman's view that the German trade account in 1922 was in a sufficiently good state to make stabilization at least a hypothetical possibility.

It might seem, therefore, that we have given the last word to the economic historians. The international data appear as a deus ex machina to pronounce a satisfyingly ironic conclusion to our protracted cogitations. Ferguson may have won the methodological battle, but he lost the historical war. Indeed, it seems that we might have spared ourselves the entire elaborate archival detour. Standard statistical sources were all that was necessary to decide on the possibilities open to Weimar's decision-makers. Economic history's victory in the methodological civil war hands victory in the argument over stabilization to Feldman. But in coming to this abrupt conclusion, we would be in danger of losing sight of what the discussion was actually about. What was fundamentally at stake in this argument was the possibility of rational politics in the early Weimar Republic. Compared to the dogmatism of the nationalists, the attraction of Feldman's heroes was precisely their resilient rationalism. They continued to bet on their capacity to read events and to manipulate the German economy. But this was a gamble at long odds, and by 1922 it had failed. It was not, as Ferguson suggests, a lack of will that made the debate over stabilization increasingly phony, but the chaotic course of inflation, which made well-informed debate impossible. By the autumn of 1922, the advocates of stabilization could no longer convince even Chancellor Wirth of the factual evidence on which they based their claim that a disaster could still be avoided. In this sense, if the overriding aim was to preserve the capacity of the German political system to act coherently, then by 1922 it was too late for rational politics.

But does this not push us back toward Ferguson and his counterfactual of a conservative stabilization in 1920? The balance of statistical evidence certainly points to a trade surplus in 1920. However, this evidence was contested by the nationalists even then. More importantly, in the wake of the Kapp Putsch of March 1920 and the bloody suppression of the general strike that followed it, neither left nor right had the stomach to impose a stabilization on their terms. It was not until the Entente announced the full extent of their reparations demands in the spring of 1921 and Germany was faced with the prospect of a descent into hyperinflation that the real argument over stabilization began. In this sense, Ferguson's counterfactual would also seem to fail the test of political plausibility.

But if this is our conclusion, is the methodological boot now not on the other foot? Perhaps it was the economic historian's insistence on determining what the trade balance actually was that was the irrelevance. If in judging the plausibility of a counterfactual all that matters are contemporary perceptions and politics, then perhaps we could have spared ourselves the distasteful detour into the world of statistical handbooks and anachronistic spreadsheets. All we needed to do was to insert a fleshed-out account of the chaos within the Statistical Office into Feldman's grand narrative of Weimar politics. This, however, would again be to risk collapsing the complex, stereoscopic rendition of reality that we have been struggling to develop. And preserving this stereoscopic view is crucial because only with both aspects of the data in view can we fully grasp the tragedy of the Weimar Republic. It is one thing to say that Weimar's politicians could not agree on the terms of stabilization. It is quite another to say that they could not agree despite the fact that stabilization was a real possibility.80 And we twist the knife even further when we argue that they were not in a position to seize the opportunity when it came in the autumn of 1922 precisely because their previous policy of aggressive fulfillment had undermined the conditions necessary for rational political decision-making.