April 1994 PRELIMINARY VERSION: DO NOT QUOTE Monetary Populism in Nineteenth-Century America: A New Interpretation Jeffry A. Frieden Department of Political Science University of California, Los Angeles Los Angeles, CA 90024-1472 The author acknowledges financial support from the UCLA Academic Senate Committee on Research, Center for American Politics and Public Policy, and International Studies and Overseas Programs. For the first century and a half of the American republic, monetary policy was at the center of national political debates. The high point of conflict over American monetary policy was reached between the 1870s and the 1890s, when the Greenback and Populist movements split groups, parties, and regions, defined the agenda of several national elections, and gave rise to some of the more vivid political rhetoric in the country's history. In the monetary arena, the defining principles of populism were hostility to the gold standard and desire for reflationary monetary policies. However, it is not easy to square the great divisiveness and fervor that characterized the movement with most of what we know about the United States after the Civil War, or about the politics of monetary policy. The standard story is that the mostly agrarian populists were exercised about their mortgages, and wanted inflation to lighten the burden of nominally denominated debt. However, neither logic nor history is particularly supportive of this interpretation, which has led many historians to pursue non- economic (cultural, religious, ideological) roots of American populism. In this paper, I present a new explanation for the sources of monetary populism. Rather than focusing on the anticipated impact of populist monetary policies on the nominal price level, I argue that these policies were primarily designed to engineer a dollar devaluation. The principal impact of the devaluation would be on the dollar prices of tradable goods, especially agricultural products. While there was concern about the overall nominal price level -- thus the real weight of nominal debt contracts -- this was secondary to concern about adverse relative price trends that a devaluation would reverse. This emphasis on the open-economy (real exchange rate) rather than closed-economy (nominal price level) effects of monetary populism allows me to resolve some thorny logical and empirical puzzles. It also ties the American populist experience into broader problems in the politics of exchange rates, and provides the basis for understanding both the features of American monetary policy in this era and why monetary politics has looked so different since the 1930s. The next section of the paper describes the broad outlines of American monetary populism. Section 2 summarizes the historiographic and economic debates about interpreting it, highlighting the theoretical and empirical problems with the prevailing focus on the impact of populist policies on the nominal price level. Section 3 presents my explanation of the issue, which involves recasting it as a conflict over the exchange rate and the exchange-rate regime with all of the distributional consequences these entail. In Section 4, I subject my argument to a series of statistical evaluations, examining how socio-economic characteristics of American states related to Congressional voting by their representatives on a variety of populist monetary initiatives. Section 5 discusses some remaining puzzles and controversies, and possible historical and contemporary implications of this study. 1. Monetary populism: Politics and policies American monetary populism began after the Civil War, and took many forms. Some populists formed and joined third parties, notably the Greenback and People's parties; others worked within the Democratic and Republican parties. Some populists were unremittingly radical, even anti-capitalist; others were staid pillars of the business community. For some, monetary populism was irretrievably linked to other policies on trade, regulation, and international affairs; for others, The Money Question was all that mattered. Here I focus entirely on monetary aspects of populism, which in itself is a daunting problem. During the Civil War the United States went off gold, and prices more than doubled under a paper currency ("greenback") standard. After 1865 debate began over whether and how the dollar should return to a fixed rate against gold ("resume"). Immediately after the Civil War ended, the Treasury shrank the money supply to appreciate the dollar and move toward resumption of gold convertibility at the pre-war rate. This real appreciation put pressure on tradables producers, especially manufacturers, as the data in Table 1 indicate. The Greanback movement developed as a response among the iron and steel manufacturers of Pennsylvania, eventually gaining support from railroadmen, as well as investment bankers and others tied to these industries.[1] In response to political pressure, the Grant administration reversed course. George Boutwell, a congressman from a manufacturing district who was Secretary of the Treasury and dominated economic policymaking during Grant's first term (1869-1873), purposely postponed resumption (of gold payments) and "contraction" (of the money supply to allow resumption); the country went through a sustained expansion. The Panic of 1873 and the ensuing slowdown in economic growth rekindled Greenback sentiment, and drew two important groups into the soft-money camp. Farmers were originally indifferent, for farm prices held up quite well in the first years after the Civil War. However, the Panic of 1873 initiated a secular decline in farm prices and led to agrarian enthusiasm for Greenback populism. The second important group brought into play after 1873 was silver miners. In 1873 silver was removed from circulation. At around that time, the price of silver began to decline relative to gold. Miners and Greenbackers devised a common program to meet both their needs. If silver were "re-monetized" at the old 16:1 rate against gold, the government would be forced to buy silver at well above the market rate. This would act as a subsidy to the miners; it would inject money into the economy as the government bought up silver; and it would force the country off gold and onto a de facto silver standard. A powerful alliance of Midwestern manufacturers, farmers, associated nontradables producers, and miners opposed the return to gold under the banners of greenback issue and free silver. Supporters of gold were concentrated in the Northeast, among the financial and commercial communities with strong ties to European trade and payments. New England textile and machinery manufacturers, heavily oriented toward world markets, also backed gold. This fundamental disagreement caused bitter political battles. In April 1874 Congress passed an Inflation Bill by a wide margin, thereby mandating expansion of the supply of paper money. Over 954 of Northeastern Congressmen opposed the bill, while over three quarters of Congressmen from the agrarian South and the agrarian and industrial West (including manufacturing Pennsylvania) supported it. Northeastarn hard-money interests immediately mounted a furious campaign to overturn the bill. President Ulysses Grant came through with a veto, but this cost the Republicans the congressional elections of 1874. After the Republican electoral debacle, Grant and Republican Party leaders attempted a display of party unity to salvage their chances for the 1876 presidential elections. In January 1875 they convinced lame duck Republicans to vote for the Resumption Act, mandating a return to gold on January 1, 1879. In 1876, Republican Rutherford B. Hayes, a supporter of hard money, was elected. Congress remained dominated by soft-money interests, and one of its first acts in February 1877 was to pass the moderately silverite Bland-Allison Act; President Hayes' veto was overridden. However, in late 1877 an attempt to repeal the Resumption Act was defeated: it passed the House and failed by one vote in the Senate. Hayes' Treasury Secretary, John Sherman, had in fact worked with Republicans and Democrats alike to find a compromise, and had determined that Bland-Allison was the price of defeating repeal of the Resumption Act. Even so, Hayes was forced to wield the blunt instrument of patronage in order to gather enough votes to save the gold standard.[2] In the meantime, some soft money supporters became disgusted with the two major parties. They founded the Greenback Party, which stood strongly for devaluation and a flexible exchange rate. The party ran Peter Cooper for president in 1876, with little success, but did better in 1878 congressional elections. However, by then Hayes and Sherman had traded for or bought enough votes in Congress to save resumption, and the country went back to gold at the beginning of 1879. Silver sentiment simmered in farm and mining regions through the 1880s, but erupted fiercely amid the agricultural depression that began in 1888.[3] In 1890 a Republican Congress and President agreed to the moderately reflationary Sherman Silver Purchase (or Treasury Note) Act. This doubled the amount of silver purchased by the Treasury under the 1878 Bland-Allison Act. The bill was too mild to satisfy anti-gold interests, but strong enough to cause fear in financial markets. These fears grew after the 1892 elections gave the Democrats, who had run on a silverite platform, control of the presidency and both houses of Congress for the first time since the 1850s. In the meantime the new People's (Populist) Party was scoring electoral successes throughout the West and South, and soft money ranked at the top of the Populists' demands. In the opposing corner, Northeastern commercial and financial interests remained at the core of the hard-money camp. The bankers' position had if anything hardened: many on Wall Street had come to hope that New York would soon be an international financial center, for which ironclad commitment to gold was a prerequisite. Manufacturers were more receptive to hard-money arguments than they had been in the 1870s, for at least two reasons. First, as the data in Table 2 indicate, declining prices of manufactured products were more than compensated by rapid productivity increases, so that few manufacturers felt substantially disadvantaged by the deflation. Second, manufacturers' interest in the money question had become secondary to their concern to defend tariff protection, which was under attack. They were willing to forgo support for silver if tariff protection were continued. In 1893 the country was hit by a panic which supporters of financial orthodoxy blamed on uncertainty about the country's commitment to the gold standard. Despite his party's platform, President Grover Cleveland was a firm supporter of gold, and during the Panic of 1893 he pushed Congress to repeal the silver purchase clause of the Sherman Act. After this repudiation of soft money, the Democrats lost the 1894 midterm elections, and in turn the gold conservatives lost the battle for control of the Democratic Party to free silver supporters. The 1896 election was fought largely over the gold standard. The Democrats and Populists jointly fielded William Jennings Bryan; the Republican candidate, moderate silver supporter William McKinley, came out for gold to appease hard-money interests. The narrow defeat of Bryan sealed the fate of silver, and in any event anti-gold sentiment dampened over the next few years as economic conditions improved. In 1900 Congress passed the Gold Standard Act, Bryan was defeated a second time, and the country's commitment to the gold standard was firm. Throughout this thirty-year period, the specific policies preferred by monetary populists varied, but always around common themes. The two constants were going off the gold standard, and establishing a government commitment to "price stability," that is a monetary policy to counteract deflation. For example, the 1892 Omaha Platform, on which the Populist presidential candidate won more than a million votes, began its demands with: First, Money. We demand a national currency, safe, sound, and flexible, issued by the general government only, a full legal tender for all debts, public and private.... (a) We demand free and unlimited coinage of silver and gold at the present legal ratio of sixteen to one. (b) We demand that the amount of circulating medium be speedily increased to not less than fifty dollars per capita.[4] The most radical position was that of the original Greenbackers, who wanted a fiat currency that would float freely against gold, as the original greenbacks did between 1862 and 1879. Silverites wanted free coinage of silver at 16:1 (16 ounces of silver per ounce of gold). Both policies would have de-linked the dollar from gold, whether directly in the Greenback case or indirectly with free silver, as government silver purchases and the issue of silver dollars (and silver certificates) would have forced a change in the gold parity, that is a depreciation of the dollar relative to gold.[5] The Populists also called for the prohibition of gold clauses, contractual agreements to index payments to amounts in gold rather than in nominal dollars as a hedge against potential devaluation. Although Populism is commonly associated with free silver, most of the Populists' monetary thinkers shared the Greenback view. They typically regarded free silver as inferior to greenbacks, but necessary to get the support of Western mining interests, which were politically crucial both because they were overrepresanted in the Senate and because they were a large faction within the Republican Party. In the words of Populist Senator William Allen of Nebraska, "We believe it possible so to regulate the issue of money as to make it of approximately the same value at all times." Another put it cogently: "It is the cardinal faith of Populism...that money can be created by the Government in any desired quantity, out of any substance, with no basis but itself."[6] For Ignatius Donnelly, the great Populist leader from Minnesota, "Scientific paper money, irredemable, based on the credit and wealth of the nation, issued and its volume controlled by the nation without the intervention of banks is the populist ideal money." With some resignation, however, Donnelly realized that while "there is no doubt that both gold and silver should be discarded, and an international legal tender paper money established...that is a vast reformation the world is not yet ready for."[7] In this context, the Populists settled for free silver. Associated with going off gold was a wide range of monetary proposals. Typically, they called for the Treasury to increase the supply of currency by a substantial amount (usually to reach $50 per person in the population), and to regulate it to avoid deflation. Closely tied to the monetary proposals were credit and banking schemes. Foremost among them was the "sub-treasury plan," a design to establish rural lending agencies ("sub- treasuries") that would make loans upon demand with land and crops in the ground as collateral, at very low interest rates (typically one or two percent annually). And there were various proposals to restrict the operations of national banks. It is probably worth emphasizing that few of these proposals are outlandish to modern eyes; certainly they do not appear to embody the threat to Western civilization that some of the Populists' opponents claimed to see. A dollar devaluation, whether the dollar were a paper currency or backed by a fixed amount of silver, was a reasonable response by a major commodity exporter to a global decline in commodity prices (indeed, most of the world's commodity exporting nations did undertake a depreciation in this period). The precise impact of free silver at 16:1 on prices is difficult to calculate, as it depends on the effect of increased American demand on the world price of silver, but Milton Friedman has made rough estimates. By his reckoning, while in reality the cumulative 1876-1879 deflation was of 13.9 percent and that of 1889-1896 was of 14.3 percent, a 16:1 silver standard would have led to a 1.7 percent cumulative inflation in the former instance and cumulative inflation of 16.3 percent in the latter. In light of the serious socio-economic effects of the deflation that instead took place, this leads Friedman to the judgement that the commitment to gold was "a mistake that had highly adverse consequences."[8] By the same token, the sub-treasury plan bears much similarity to farm credit arrangements developed in the inter-war period, especially during the New Deal. While there are many grounds for complaints against farm subsidies in principle (especially by non-farmers), the credit-based schemes are among the less inefficient ways of effecting transfers to agricultural producers. The explicit monetary proposals had more problems, especially if increasing the currency supply to $50 per person is taken to require an immediate Treasury response. This would have mandated a three-fold increase in the supply of currency. By one estimate, had this been undertaken in 1892 it would have increased the broad money stock (M-2) from $4.5 billion to $15.8 billion, and would have caused a one-time inflation of 247 percent as prices adjusted to the new money supply.[9] But the $50 per person demand was generally regarded as a target rather than a requirement, and it is almost certain that a soft-money Administration would simply have instructed the Treasury to reflate by moderate amounts. These, then, are the general outlines of the Populist movement and its monetary demands. While the movement was not ultimately successful, its pressures did affect the making of American monetary and other economic policies. Populism marked one of the more important watersheds in American political history, as it led up to the crucial realignment election of 1896, whose results set the tenor of American politics until the New Deal. And populist monetary ideas -- albeit not in their anachronistic silverite version -- continue to recur in the United States as elsewhere, especially in times of macroeconomic distress.[10] From the standpoint of scholars of macroeconomic policy and politics, the experience is of profound importance, as it is perhaps the most prominent and protracted set of monetary- policy debates in modern history. 2. Monetary populism: Problems of interpretation In spite of the voluminous historical record on Populist monetary demands and debates, and its great political prominence, attempts to interpret the phenomenon have left scholars somewhat perplexed. Prevailing efforts to understand monetary populism as simply a struggle between debtors and creditors over the nominal price level have serious theoretical and empirical problems, which has led some scholars to embark on imaginative culturalist constructions of the true nature of Populism. Early on, the story of Populism was told in terms of beleaguered farm debtors whose infatuation with reflation flowed from their belief that a rise in the overall price level would reduce the real burden of their nominally denominated debts. This was certainly part of the background to Populism, but reliance on it as a full explanation of the phenomenon faces major difficulties. The first set of puzzles associated with the standard story is theoretical. It is hard to imagine why farmers should have been concerned about the overall nominal price level in and of itself. Inflation would have been of unmitigated value to them only under very special circumstances. Certainly any nominal contracts -- such as farm mortgages -- would have been less burdensome. But the price of other inputs, such as fertilizer, transportation, and other goods and services, would have risen; and the price of their output (whether wheat, corn, or other crops) could not have been guaranteed to have risen as much. In other words, farmers cared about both input and output prices; debt was simply one input among many. A full explanation of support for Populist monetary policies needs to take into account their potential effect on both sides of the scissors allegedly affecting farmers, rising debt and declining farm prices. A focus solely on debt can only be half (or less) of the story. Furthermore, the inflation in question would hardly have been unanticipated. The average farm mortgage in the West was for five years, and many were at much shorter maturities. Given the great monetary uncertainties of the day, lenders would have been foolish not to protect themselves against potential inflation, especially when this inflation would have had predictable political origins. If political trends led to expectations of inflation, lenders would either have charged an appropriate risk premium or might have insisted on writing gold clauses into debt contracts. Even without this, the lag between political action, policy implementation, and price effects; the short maturities of farm mortgages; and the inevitable readjustment of new loans to new monetary conditions, all mean that the average farm Populist could have expected barely more than a year's debt service relief had soft money prevailed. This fact, observed at the time by Irving Fisher, would have substantially reduced the ability of free silver to provide relief to indebted farmers.[11] These general analytical concerns are compounded by daunting empirical issues. Some commonly expressed empirical concerns are in fact not so puzzling as they might appear. During the 1950s and 1960s, new economic historians found that, despite farmers' complaints, there was little systematic evidence of a secular deterioration in farmers' terms of trade over the late nineteenth century, nor did interest rates appear particularly outlandish (given appropriate risk premia), nor was there much evidence of overpricing of railroad transportation.[12] Now, such findings may lead some economists to wonder why farmers may have wanted government measures to make them better off, but few political scientists would be troubled: there is no necessary reason why economic actors should demand favorable policies only when there are economically justifiable reasons for such demands. But other empirical findings are a more direct challenge to the standard story. As already mentioned, most mortgages were at very short term. One survey of representative counties found that mortgages in the West North Central and South Central states, where Populist support was strongest, averaged 3.8 and 2.7 years in maturity, respectively.[13] While inflation might provide some modest relief, farmers could hardly have expected it to last very long, as renewed mortgages would have reflected new price trends. Also puzzling is the fact that, in the face of alleged concern about overall price trends affecting nominal debt contracts, gold clauses were very rare in farm mortgages while they were relatively more common among other forms of debt. From a political standpoint, the correlation between levels of debt and Populism is not so strong as might be expected. Some studies have found such a relationship among farm communities, while others have been less confident that indebtedness was a good predictor of support for Populism.[14] The most careful study of voting for Populist candidates in the literature, on county patterns in Kansas in the 1890s, found that controlling for other economic characteristics, value of mortgages was actually associated with lower levels of Populist support. On the other hand, farm occupation and the variability of farm yields were positively correlated with Populist votes, and wealth (whether in banks or in land) and farm income were negatively correlated with Populist support. Once these factors (and several others) are included in a multivariate analysis, debt levels enter negatively, if at all.[15] Just as striking is that other very heavily indebted groups of the national population were not drawn to the Populist cause. Urban residential mortgages were the fastest growing segment of the loan market during the Populist period, but middle-class homeowners in the big cities are widely believed to have been quite hostile to Populist monetary schemes. So too were many manufacturing and commercial groups large net debtors, but their attitude toward easy money tended to be one of lukewarm support at best, great hostility at worst. It is also the case that while Populist rhetoric did mention farm debt, it tended to focus more on other issues. First and foremost was the apparent tendency for farm prices to decline more than those of other goods. Associated with this was the concern about the prices of particular inputs, such as railroad transportation (alleged to be artificially expensive due to monopolistic practices) and farm implements (alleged to be artificially expensive due to protective tariffs). For example, one survey of Kansas farmers in 1893 found that 65 percent attributed farm distress to low prices, 15 percent to drought, 11 percent to "money scarcity" and "interest rates" combined, 5 percent to railroads, and the rest to scattered other factors.[16] None of this is meant to imply that farm debt was not of concern to American farmers in the late nineteenth century, or that raising the nominal price level would not have provided them with some debt relief. It is to argue that the depth and breadth of Populist monetary sentiment is hard to square with a story based entirely on farm mortgages. Farmers' debts, to reiterate, were at very short maturities, would surely have adjusted rapidly to the new price level, and were not very well correlated with support for Populism; nor were other debtors in the population sympathetic to Populist monetary proposals. It might be objected that it was the inter-regional character of farm debt that led to the inter-regional divisions observed during the Populist era. But even in those parts of the country with the largest proportions of mortgages held in other regions, the vast majority of farm mortgages were held locally. By one estimate, in the early 1890s in the West North Central and West South Central regions, those with the smallest local mortgage share, some 85 percent of all mortgages were held locally.[17] One attempt to address these concerns has been to re- evaluate farmers' complaints. Less aggregate analyses -- typically differentiating shorter- and longer-term trends, and among states and regions -- have confirmed the early view that many farmers were subject to great volatility in income, and that in many areas there was a significant deterioration in farmers' terms of trade for several years at a time.[18] Newer studies have also confirmed that these agrarian difficulties were roughly correlated (both over time and across place) with patterns of support for Populism (and pro-populist Democrats).[19] And, although the argument here is more controversial, there is some recent support for the view that Midwestern, Western, and Southern farmers faced quasi-monopolistic banking conditions that may have raised mortgage rates to them above what would have been justified purely by financial considerations such as risk premia.[20] These studies have largely superseded the widespread cliometric view, common in the 1960s and 1970s, that there were few economic bases for farm discontent in the Populist era. However, they leave unexplained why agrarian demands focused so heavily as they did, and in the ways that they did, on monetary issues. While there were clear -- if regionally and temporally differentiated -- reasons for farmers to pursue political action to better their lot, it is still a puzzle why this action took the form of a peculiar insistence on reflation. This has contributed to attempts by some historians to argue that Populism had important, perhaps predominant, non-economic components. A forerunner of more extreme views was the argument made by Anne Mayhew, to the effect that farm protest was aimed more at the process of commercialization of agriculture than at any particular economic conditions: mainly subsistence farmers whose sales were discretionary found themselves forced to specialize and rely entirely on the marketplace.[21] Probably most influential among American historians has been the work of Lawrence Goodwyn. While Goodwyn recognizes the importance of economic issues to the Populists, he denies the centrality of economic roots or meaning to the movement. For him, Populism was first and foremost "a new way of looking at things -- a new culture," and its centerpiece was a "cooperative commonwealth," an alternative to primitive capitalism based on the belief that "the mature corporate state would, unless restructured, erode the democratic promise of America."[22] Goodwyn's culturalist focus on farmers' attempts to fashion a new social order leads him to directly deny the importance of monetary issues: "the groundswell for silver derived less from broad public understanding of currency than from moral outrage at the apparently surreptitious means by which the bill [to de- monetize silver in 1873] had achieved congressional approval." Hard money "in a cultural sense...rested on moral values rooted in both religion and primitive capitalism;" soft money was "a political ideology grounded in a desire of non-bankers to cope with changing commercial power relationships within an industrializing society."[23] A similar view is that of Scott McNall, who explicitly objects to seeing "a direct link between economic problems and political behavior," emphasizing instead the farmers' alleged attempts "to create a distinct movement- culture," in which they "tried to create themselves as a class, one which would serve as a Jeffersonian counterweight to the power of entrepreneurs."[24] While non-economic factors certainly played an important part in American politics in the late nineteenth century, it is hard to believe that economic issues were purely a smokescreen for a broader cultural conflict. The Populists worked out elaborate policy proposals, and flooded the country with millions of books and pamphlets intended to educate farmers and others about the implications of Populist economic policies. Some prior economic interpretations had problems; this is not a reason to jettison attempts to explain at least some of the Populist phenomenon on the basis of the economic interests of socio- political agents. But this requires rethinking the economic impact of soft money. 3. Reinterpreting monetary populism in an open economy Prevailing analyses of Populist monetary proposals focus on their impact on the nominal price level. To the contrary, I believe their principal goal, and the essential cause of differential patterns of support and opposition, was to affect relative prices. First, soft money was intended to force a devaluation of the dollar. Second, soft money implied going off the gold standard altogether and making the dollar a silver- backed currency that would float against gold. Empasizing the effects of reflation on the aggregate price level is appropriate in a closed economy. But the American economy in the late nineteenth century was not closed. American financial markets were tightly integrated with those abroad. After all, the United States was the world's largest borrower: net capital inflows to the United States averaged over 7.3 percent of Net Capital Formation between 1884 and 1893.[25] Although there were significant barriers to the import of many manufactured goods, trade was over 12 percent of GDP (more than double the post-World War Two figure); a fifth of the country's total farm output was exported, with much higher proportions for such crops as wheat and cotton. In an open economy with a floating exchange rate, the principal effect of reflationary policies is to depreciate the currency, raising thereby the price of tradable goods relative to nontradable goods and services. Risk-adjusted interest rates are bound by integration with financial markets abroad; the local- currency price of tradables is bound by world prices and the exchange rate.[26] Depreciation thus helps producers of tradables at the expense of producers of non-tradables. For a country on (or considering) a fixed exchange rate system, such as the gold standard, reflation cum depreciation in fact requires either a temporary or a permanent abandonment of the commitment to a fixed rate. During the gold standard era, it was widely believed that a devaluation would destroy the credibility of the monetary authorities to maintain a gold parity in general. While some involved in the monetary debates did want a managed devaluation with the dollar re-fixed at a lower (depreciated) rate, most regarded this as infeasible without an international agreement.[27] In this context, the implementation of soft-money proposals would have had two predictable effects: to depreciate the dollar and to float it. The distributional impact of these policies is substantially different from simply raising the nominal price level across the board, which only affects nominal contracts. Inasmuch as the distributional effects were reflected in politics, this alternative emphasis gives rise to different expectations of the politics of Populism as well. The first set of distributional effects has to do with the depreciation of the currency.[28] A depreciation raises the relative price of traded goods, and lowers the relative price of nontradable goods and services. This is especially beneficial to import-competers and export-competers. In the American context, as the country was a major agricultural and mineral exporter, a depreciation would certainly have helped those farmers and miners who produced goods that entered into international trade (i.e. almost all of them). For manufacturers, this would also have been the case, as they were largely competing with imports from Europe. However, this support would have been tempered in the 1890s by the imposition of America's high tariffs on manufactures. Inasmuch as the tariffs were prohibitive (which many were), this would essentially have made the protected goods nontraded. Even where they were not, a specific tariff could accomplish essentially the same effect on output prices as a devaluation, and did not involve raising imported input prices. This alternative, needless to say, was not readily available to farmers and miners producing for export.[29] For them, a devaluation was an excellent, and eminently feasible, way to use government policy to raise their output prices. The second dimension, the preferred currrency regime (fixed or floating against gold, in the event), was of course related to the level of the exchange rate but is not the same. A floating rate was particularly noxious to those economic agents heavily involved in international trade and payments, especially the big Northeastern commercial and financial houses and some exporters of heavy custom-made machinery, who regarded the commitment to gold as a fundamental necessity for their ability to keep their international business.[30] However, for non-tradables producers without international ties -- for whom the exchange rate was essentially irrelevant -- the sacrifices associated with a fixed rate were unjustifiable. So too would those tradables producers operating solely in the home market be indifferent or hostile to a fixed rate, which if anything might encourage import competition. Certainly this characterization is stylized and incomplete. It does not adequately take into account the nuances in economic actors' preferences, and the detailed differences that depended on demand and supply elasticities, the depth of forward markets, and a host of other things. Nor does it indicate the degree to which groups might prefer particular outcomes. After all, exporters might have a weak preference for a fixed rate but a strong preference for a depreciated one; nontradables producers might have a strong preference for a floating rate but a weak preference for an appreciated one, and so on.[31] In practice, as mentioned above, the two dimensions tended to collapse into one in the monetary debates -- as they often do even today. One side wanted a strong fixed rate, the other a weak floating rate. At the center of the first, hard-money, camp were international financial and commercial groups; at the center of the second, soft-money, camp were export-competing farmers and miners. Domestic nontradables producers, by most accounts, were divided but generally not heavily involved in the debates. Manufacturers tended toward a soft-money stance, but this was tempered both by the relief provided by trade protection and by the fact that the core of the soft money coalition in the 1890s was hostile to industrial tariffs. This open-economy perspective on the monetary debates of the 1890s, then, leads to predictions about anticipated political cleavages that differ substantially from the traditional closed- economy view. The standard story, relying on preferences over the overall nominal price level, expects solely a debtor-creditor divide, with debtors inflationists and creditors deflationists. An open-economy account, relying on preferences over the exchange rate, expects most crucially that soft-money advocates will be exposed tradables producers while hard-money advocates will be globally-oriented economic actors. The position of domestic nontradables producers is somewhat ambiguous -- they would prefer an appreciated floating rate, but that option was not on the policy agenda. Again, all of this is relatively crude and abstracts from many economic and political details of the American political economy in the 1890s. Economically, probably the most important, to reiterate, was the protected nature of American manufacturing. Politically, again, it was the real or potential linkage between debates on money and trade: if soft money meant free trade (as it did for many agrarian populists), manufacturers were against it. Indeed, in the 1896 Presidential election the Republicans made an explicit connection of this sort, promising industrialists high tariffs in return for their support for gold. It can also be noted that the implications of debtor status are somewhat confounded by the undoubted existence of credit rationing. If, as is almost certainly the case, only the more creditworthy individuals could borrow, then measured levels of farm indebtedness, for example, would be biased toward more prosperous farmers. At any rate, the difference in predictions as to political position between the two perspectives is relatively clear. I now turn to an empirical evaluation of my argument. 4. The politics of monetary populism: Evidence from Congressional voting I propose to assess the economic roots of support for monetary populism by looking in some detail at the relationship between Congressional voting on monetary issues in the 1890s, on the one hand, and economic characteristics of the constituencies represented by members of Congress, on the other. For the economic background, we can refer back to Table 1. This indicates the general pattern of relative price movements over the course of the 1880s and 1890s. Although specifics vary from series to series, it appears clear that there was a consistent tendency for tradables prices to decline and nontradables prices (especially construction) to decline less or even rise. For example, between 1879 and 1894, according to the Gallon series, while farm prices declined 19 percent, construction prices rose 16; according to the Warren and Pearson series, analogous figures show a 12 percent decline against a 3 percent decline. This would certainly be expected to give rise to discontent among tradables producers. Manufacturers indeed responded with strong demands for trade protection, which they received. Farmers and miners, who sold heavily into world markets, pushed for a devaluation -- or at least that is my argument. To evaluate this argument, I looked at the most contentious Congressional votes from the 1892-1896 height of Populist sentiment (the votes are described in Appendix B). The six House and four Senate votes analyzed are widely recognized as central to the struggle between gold and silver. In the statistical analysis, I pool the House and Senate votes; this helps counteract some of the effects of bill-specific strategic voting and other contingent considerations. A detailed statistical analysis of each of the ten votes is in Appendix A. I then evaluated the impact on Congressional voting behavior of some economic characteristics of constituencies. All data, except for the party affiliation of the senator or member of Congress, are for states -- data at the level of the Congressional district are not available. The 1890 census, as usual, reported employment figures; those used here are simply measured as a share of the total labor force for the state. Fortunately, due to Populist agitation, the 1890 census included a great deal of information about landed indebtedness. The figure used here is simply the total value of real estate debt for the state as a propertion of the total value of all land in the state.[32] The relevant point to recall is that my perspective and that of the standard story imply different sources of support for soft-money policies. The traditional interpretation of Populism leads to the expectation that support for monetary populism would come primarily from debtors, for whom reflation would reduce the real burden of nominal debts. My view is that support for soft money would be concentrated in those groups expected to benefit from a real depreciation, especially tradables producers. In the 1890s American context, one large group of such producers, manufacturers, had obtained a substitute, trade protection, but farmers and miners had no available alternative policies other than depreciation. I expect little or no relationship between levels of indebtedness and preferred monetary policies, as the impact of such policies on output prices might easily swamp their impact on debt service payments. Table 3 presents relevant information for the four Senate votes analyzed. The first column displays the means of all variables, divided according to whether the Senate votes in question were for the soft-money or hard-money position. For example, while 53.9 percent of all the Senate votes cast were cast by Democrats, 67.8 percent of the votes for soft money and 40.3 percent of those for hard money were Democrats. Other variables are for the senators' states. So real estate debt averaged 11.5 percent of the total value of land in states whose senators voted soft money, while it was 15.2 percent in states whose senators voted hard money. In can readily he seen that a hard-money stance was associated with Republicans, higher levels of debt, and more employment in professional and personal services, trade and transport, and mechanical and manufacturing. A soft-money stance was associated with Democrats and more employment in agricultural and mining activities. Simply looking at means does not control for relationships among the variables, so the second column shows the results of a logit analysis. The votes are coded so that a positive coefficient implies a hard-money position, a negative coefficient soft money (personal service employment, which varied little, was excluded to avoid the employment variables summing to one). Again it can be seen that hard money was associated with higher levels of debt, and more employment in professional service and manufacturing (although the coefficient on professional service is not statistically significant). Soft money was associated with the Democrats, and with more employment in agriculture and mining, and trade and transport (with agriculture and mining not significant). Studies of voting in the Senate are notoriously unreliable. States can be very heterogeneous, and it is common for the two senators to stake out a different constituencies, functionally or geographically (upstate New York and New York City, upstate and downstate Illinois, northern and southern Califoria, for example). There are also relatively few senators. And in many of the votes in question here, the Senate vote was relatively meaningless posturing; often the bill's fata depended on developments in the House which the senators anticipated. Analyzing votes in the House of Representatives, however, presents the problem that the economic data are only available for states. I employed three different techniques to address this problem. First, I used state-level data for all members of Congress from a state in a logit analysis. That is, while each individual member of Congress's vote was a separate observation of the dependent variable, all members from any one state were assigned the same economic data. This imposes the clearly incorrect presumption that all congressional districts in a state are economically identical. Second, instead of looking at votes by individual members of Congress, I calculated the percentage of each state's delegation voting for the hard-money position on the bill and used this as the dependent variable. That is, each observation of the dependent variable here was the proportion of the state's members of Congress voting for hard money. This has the advantage of allowing states to vary along a continuum; however, it presumes that a dichotomous choice was really continuous and has other misleading properties. Third, I used the same dependent variable as in the second method, but recognizing that this is constrained to vary between zero and one, I transformed it into the log of the odds ratio. In the second and third specifications, the economic variables were as in the first, but the partisan variable was re-calculated as the share of the state's delegation that was Democrats. Table 4 presents the means of all variables. As in the Senate, a hard-money stance was associated with higher debts, and more employment in professional and personal service, trade and transport, and manufacturing. Soft money was associated with the Democrats, and with more employment in agriculture and mining. Table 5 displays the three statistical techniques described above. It can be seen that the signs of the coefficients for all economic variables are in the direction I expect; the great majority also reach statistical significance. It is especially gratifying that, despite the very different character of the three techniques, the results are quite stable across them. The r-squared, pseudo r-squared, and percent correctly predicted statistics all indicate that the explanatory variables "explain" a relatively large share of the variance in the voting. The OLS results in column 2 of Table 5 can be read to imply the following relationships, looking only at those reaching statistical significance. A one percentage point increase in real estate debt as a share of land value in a state was associated with a 2.256 percentage point increase in the share of the state's delegation voting for hard money. A one percentage point increase in employment in agriculture and mining as a share of total employment in a state was associated with a 2.044 percentage point decline (increase) in the share of the state's delegation voting for hard money (soft money); of trade and transport, with a 8.096 percentage point decline. The vary large coefficients on trade and transport and professional service are a function of the relatively low levels of employment in these sectors. Two things emerge quite clearly from these statistical analyses. First, support for Populism actually declined with levels of indebtedness. This is strong disconfirmation of the traditional approach. It is in line with my framework, as I expected there not to be strong support for Populism from debtors, other things equal; certainly nothing in my view required debtors to oppose Populism, though. It is likely that the relationship observed is due to the wealth effect associated with credit rationing, discussed above. Second, agricultural and mining states tended strongly to support Populist monetary policies, while states with lots of manufacturing and professional services tended strongly to oppose them. The latter finding is very much in line with my argument; voting on these monetary policies was closely linked to characteristics of the state's output, thus presumably to relative prices. It -- along with the finding on debt -- is very much at odds with the standard nominal price level story. One interesting outcome is the quite extraordinary degree to which trade and transport employment is associated with support for soft money. This relationship is the strongest, and most frequently statistically significant, in all the analyses of all the Congressional votes, across all specifications. It also is not evident from the means (i.e. not controlling for other explanatory variables). Although it would be foolish to attempt a full interpretation of this, my sense is that this picks up the support of many nontradables producers for a floating exchange rate. But I have no way of evaluating this statistically any further, and the historical accounts essentially ignore the role of these economic groups in the Populist debates. It is important to emphasize that my purpose here was not to explain the voting outcomes themselves, that is, why a bill passed or failed in the House and Senate. To understand this requires a much more developed analysis of the institutional environment, partisan effects, and strategic considerations. My goal is much more modest, to see if the voting patterns of the 1890s were consistent with my proposed reinterpretation of the politics of monetary populism. I believe that the results are a convincing refutation of the standard story, for which the principal support for soft money should have come from debtors. I also believe that the results represent strong evidence for my approach, for which support for soft money should have depended on the expected relative price effects of a depreciation, and thus come from tradables producers. Of course, my interpretation also relies on some contingencies, such as the complicated preferences of manufacturers protected by very high tariffs. But in general I find substantial support for my perspective. 5. Puzzles. implications, and conclusions This analysis is only the first step in understanding the political economy of American Populism. However, it is a necessary first step. For a long time, scholars have worked with an inadequate characterization of the expected economic effects of Populist monetary policies, and particularly of the expected impact of these policies on groups in society. This is especially limiting for those wanting to understand the politics of Populism, for a prerequisite to analyzing the politics of economic policies is to understand how these policies were expected to affect politically relevant groups. This does not by any means require that the political actors in question had a full understanding of the economics of these issues, or that they organized themselves precisely along the lines these economic effects might suggest. Certainly American farmers wanted relief from debt, and certainly soft money would have provided some relief. But this cannot be the whole story, and I argue it was not even the major part of the story. The statistical evidence reported above, I believe, bear out my assertion. Indeed, Americans were not so oblivious to the sorts of considerations I discuss as much of the literature tends to imply. On the purely anecdotal level, many Populist complaints focused on how price trends seemed to be affecting agricultural products particularly adversely. A quaint example is found in the most famous propaganda tract of monetary populism of the 1890s, a small hook called Coin's Financial School. The fictitious protagonist of the pamphlet, after discussing the impact of deflation on debtors, went on to treat of other prices. He imagined a farmer going forth to "see if he could buy as much of this world's goods with 50 cents as ha formerly could with a dollar." We will suppose, before starting, he goes to pay his taxes. He will find that his 50-cent wheat will not pay as much as his $1.40 wheat did in 1873. He will find his taxes just as much, and it will take all of twice as much wheat to pay them as in 1873. While passing out of the Court House suppose he meets a county official and should ask him what salary was paid to his office in 1873 and now. The answer would be the same number of dollars now as in 1873. The same is true of city, state, and national officers, also with the army, navy, and official abroad....He starts for the depot and to get there he takes a street-car. He finds the fare the same as in 1873. He gets on a Pullman car to find the cost the same as in 1873. He registers at a first-class hotel. He finds the cost about the same as in 1873. He sands a telegram, and finds it costs the same as in 1873. He gets a shave with the same result. He buys tea and coffee, with the same result....He finds interest, except in cities on first-class loans, about as high as in 1873. Should he now meet the man who told him that his 50-cent wheat would buy as much of the world s goods as it ever did, it might result seriously for the other fellow."[33] Every price comparison mentioned here is with a non-traded good or service: government, transportation, lodging, communications, a shave, restaurant food, and financial services. This is probably not coincidental, for the simple reason that a comparison with manufactured goods prices would almost certainly have shown (as in Table 1) a relative trend in favor of farm prices. This is not to say that the Populists, or anyone else, had a clear sense of the differential price effects of a real appreciation, only that they had a general understanding of contemporary price trends. After all, Americans had had fifteen years' experience with a floating, depreciated currency in the Greenback era (which most Populist monetary thinkers had analyzed closely, and anyone older than about 35 in 1890 remembered). They had also, by 1896, had nearly ten years' experience with what we would consider a real appreciation of the gold dollar. It is not surprising that there was general awareness of the broad contours of the relative price trends the country had been experiencing, and those likely to result from re-floating the dollar against gold. But gaining a better understanding of this aspect of the political economy of Populism is only a beginning. There are many characteristics of the era well worth further political economy investigation. To mention but a few, they include the bargains that brought manufacturing over to the hard-money camp, the use of partisan and patronage weapons to affect Congressional voting, the relationship between the monetary and non-monetary (especially financial and fiscal) components of Populist programs, and the international political aspects of the bimetallism controversies of the period. By the same token, we still do not fully understand many of the more purely economic issues of the day, such as why gold clauses (indexing) were not more common, why export subsidies were never considered, and what the economic impact of the monetary measures taken were.[34] This analysis may have implications beyond the understanding of a somewhat arcane episode in American political and economic history. Staying for the moment in the realm of monetary-policy history, massive agitation for soft money was not a phenomenon confined to the United States. Support for devaluation, depreciation, and silver -- and opposition to the gold standard -- -- was something of an international movement from the 1870s until the 1930s, ebbing and flowing as world relative prices changed. It won the backing of German Junkers, Argentine farmers and ranchers, Italian industrialists, and many others for decades, and in many guises. So a clearer perspective on the political economy of the American variant of this trend may help shed some light on other national experiences.[35] The age of monetary populism in America may help illuminate other issues as well. One is the peculiar tendency of monetary policy to vary greatly in political prominence. In the late nineteenth and early twentieth centuries, monetary policy was very contentious in almost all countries; from the 1930s until the 1970s, it was relatively less conflictual. It has, over the past twenty years, migrated back toward center stage. If my argument is correct, and the very heated monetary policy debates of the past are best understood as a function of the open economy, it follows that the more open the economy, the more monetary policy will tend to implicate the sorts of distributionally distinct interest groups discussed in the American context. This would lead to an expectation that conflict over monetary policy should he a function of the openness of the national economy, both over time and across countries.[36] Another issue that this episode, and its interpretation, highlight is the potential substitutability between monetary policy, on the one hand, and other redistributive policies. If monetary policy is aimed at the aggregate nominal price level -- at changing relations between debtors and creditors -- only the broadest-gauged of alternative policies can be imagined. If, however, it is intended to redress more specific relative price trends, such as a decline in the price of commodity exports, then alternate policy reponses are easier to implement. In the American case, the relevant observation is that the real appreciation of the dollar led to two kinds of demands from tradables producers: for a nominal devaluation from farmers and miners, and for trade protection from manufacturers. The broad question of policy alternatives is an important one, especially where some policies are preferred on efficiency or other grounds. Another point that can be made is that the expected distributional effects of monetary policy are important to the burgeoning analyses of macroeconomic political economy. Most of this literature is based on closed-economy assumptions. For example, in some analyses, Left parties and the poor are expected to be more inflation-acceptant than Right parties and the rich; the reasons adduced range from the existence of an exploitable Phillips curve, through different net debtor positions, to rank empirical observation. An open-economy context more appropriate to today's developed countries would lead to a different set of political expectations, mora or less along the lines described in the 1890s American setting.[37] This is relevant both to analyses of potential political business cycles, and of the potential impact of such institutions as independent central hanks on macroeconomic policy outcomes.[38] Indeed, inasmuch as monetary politics in an open economy is irrevocably linked to the exchange rate, and the exchange rate is the responsibility of the Finance Ministry rather than the central bank (as it is in most countries), the actual implications of central bank independence are much murkier. In any case, all these extensions and more are possible. The more limited point worth making is that this paper presents evidence that at least in one historically salient instance, prevailing approaches to the politics of monetary policy have been very misleading. Traditional analyses of monetary populism in nineteenth century America -- and of monetary politics in most countries most of the time -- have focused on the expected effects of inflation or deflation on the aggregate nominal price level. In an open economy, this focus is unjustified. In an open economy, support for an expansionary monetary policy typically comes from those who would benefit from a depreciation of the exchange rate -- not simply, only, or even at all from those who would be served by an inflating away of their nominally denominated debts. In the United States in the 1890s, indeed, support for monetary populism -- opposition to the gold standard -- did not come from the country's indebted classes. Indeed, those with higher levels of real estate debt typically opposed Populism. Soft money -- silver, leaving the gold standard, floating the dollar, depreciating -- was supported first and foremost by farmers and miners the (domestic-currency) relative prices of whose products would have been raised by a dollar devaluation. It was also supported by those in such nontradables activities as trade and transportation, for whom a stable exchange rate was unimportant. Monetary populism was opposed by manufacturers who could achieve analogous, and mors targeted, relative price results by way of trade barriers. It was also almost certainly opposed by the country's international financial and commercial groups, although these are not identifiable in the data I have used. I have argued that American monetary populism was motivated by relative price concerns rather than preferences about the overall nominal price level. This view is, I believe, both theoretically more sound than the traditional view and analytically more in accord with what we know about the period. It also appears fully borne out by the evidence brought to bear here about the impact of economic characteristics of the American states on House and Senate voting on monetary policy issues. Table 1 Representative relative price indices, 1860-1899 (1870=100) Product 1869 1874 1879 1889 1894 1899 GALLMAM SERIES[a] Traded goods Agriculture 147 128 100 85 81 87 Manufacturing 151 123 100 90 66 80 Mining 178 144 100 82 75 85 Nontraded goods and services Construction 133 124 100 119 116 126 WARREN AND PEARSON SERIES[b] Traded goods Farm products 178 142 100 93 88 89 Textile prod. 170 132 100 87 72 75 Metal prod. 169 145 100 87 57 87 Nontraded goods and services Building mat. 149 136 100 109 97 107 Table 1, continued Representative relative price indices, 1860-1894 (1879=100) Product 1869 1874 1879 1889 1894 1899 SHAW SERIES[c] Traded goods Cons. dur. 144 n.a. 100 98 87 84 Prod. dur. 172 n.a. 100 93 82 92 Nontraded goods and services Perishables 163 n.a. 100 102 88 87 Const. mat. 123 n.a. 100 104 88 99 BRADY SERIES[d] Traded goods Salt 289 n.a. 100 92 n.a. 74 Cotton goods 192 n.a. 100 92 n.a. 74 Machine prod. 159 n.a. 100 45 n.a. 39 Agr. mach. 151 n.a. 100 84 n.a. 81 Nontraded goods and services Bakery prod. 168 n.a. 100 119 n.a. 118 Fresh meat 142 n.a. 100 107 n.a. 111 Fresh fish 185 n.a. 100 101 n.a. 102 Const. costs 110 n.a. 100 108 n.a. n.a. Notes to Table 1 n.a.=not available a. Robert Gallman, "Commodity Output, 1839-1899," in Trends in the American Economy in the Nineteenth Century Studies in Income and Wealth Volume 24 (Princeton: Princeton University Press, 1960), pp. 13-43; his Variant A is used for construction. b. Farm products, textile products, metals and metal products, and building materials calculated from George Warren and Frank Pearson, Gold and Prices (New York: John Wiley and Sons, 1935), pp. 30-32. c. Consumer durables, producer durables, perishable products, and construction materials calculated from William Shaw, Value of Commodity Output since 1869 (New York: National Bureau of Economic Research 1947), pp. 290-295. d. Salt, cotton woven goods, machine-shop products, agricultural machines, bakery products, fresh meat, fresh fish, and construction costs (for houses, churches and schools) calculated from Dorothy Brady, "Price Deflators for Final Product Estimates, in Output, Employment, and Productivity in the United States After 1800 Studies in Income and Wealth Volume 30 (New York: Columbia University Press, 1966), pp. 91-111. Table 2 Value Added per Gainful Worker, 1869-1899 (1870=100) Sector 1869 1879 1889 1899 Agriculture 84 100 105 118 Manufacturing 84 100 146 159 and Mining Construction[a] 82 100 95 90 a. This is Gallon's Variant A. Source: Gallman (see Table 1, note a). Table 3 Pooled results for four Senate votes: Means of all explanatory variables and statistical evaluation (logit only) Variables Means Logit Constant ---- 1.467 (0.32) Democratic Party .539 -0.798** Soft money .678 (1.98) Hard money .403 Real estate debt as share of land value .133 9.707** Soft money .115 (2.89) Hard money .152 Share of labor force in: Agriculture and mining .448 -4.256 Soft money .542 (0.80) Hard money .350 Professional service .041 14.716 Soft money .039 (0.61) Hard money .043 Personal service .185 --- Soft money .173 Hard money .196 Trade and transport .133 -33.272** Soft money .117 (2.40) Hard money .150 Mechanical/manufacturing .194 20.640** Soft money .129 (2.92) Hard money .261 Number of observations 295 295 Percent correctly --- 83.05% predicted pseudo r-squared --- .474 Table 4 Pooled results for six House of Representatives votes: Means of all explanatory variables Variables Means Democratic Party .685 Soft money .824 Hard money .577 Real estate debt as share of land value .149 Soft money .119 Hard money .175 Share of labor force in: Agriculture and mining .413 Soft money .533 Hard money .313 Professional service .041 Soft money .039 Hard money .043 Personal service .188 Soft money .169 Hard money .204 Trade and transport .142 Soft money .117 Hard money .163 Mechanical/manufacturing .215 Soft money .142 Hard money .276 Table 5 Pooled results for six House of Representatives votes: Statistical evaluation (1) (2) (3) Logit OLS Odds ratio Constant 3.764* 1.537** 8.020** (1.66) (6.60) (4.49) Democratic Party -0.558** 0.116 1.400 (3.58) (0.89) (1.39) Real estate debt as 9.863** 2.256** 16.137** share of land value (6.36) (3.17) (2.95) Share of labor force in: Agriculture and mining -7.071** -2.044* -15.099 (2.67) (1.68) (1.60) Professional service 10.591 8.612 65.673 (1.05) (1.36) (1.35) Trade and transport -30.963** -8.096** -66.849** (5.11) (2.53) (2.72) Mechanical/manufacturing 12.543** 0.864 8.935 (3.65) (0.74) (0.99) Number of observations 1738 44 44 Percent correctly 78.83% --- --- predicted pseudo r-squared .470 --- --- Adjusted r-squared --- .612 .620 Note: A negative coefficient indicates a soft-money stance. Notes to Tables 3-5 All means are statistically significant at the 95 percent level of confidence. * = statistically significant at the 90 percent level of confidence ** = statistically significant at the 95 percent level of confidence t-statistics in parentheses. The pseudo r-squared statistic reported is C/(N+C) where C is the -2 log likelihood ratio and N is the number of observations. This is the variant suggested by John Aldrich and Forrest Nelson, Linear Probability. Logit. and Probit Models (Newbury Park: Sage, 1984), p. 57. Dependent variables: For logit, votes of individual members of Congress or the Senate (0 = soft money, 1 = hard money.) For OLS, share of each state's congressional delegation voting for the hard-money position on each bill. For odds ratio, the log of the odds that the representative member of each state's congressional delegation voted for the hard-money position on each bill. Thus a positive coefficient is hard-money, a negative coefficient is soft-money. Explanatory variables: For logit, Democratic Party is specific to each individual member of Congress; for OLS and odds ratio, Democratic Party is the Democrats' percentage of the state's congressional delegation. All other explanatory variables are for states. For a more detailed statistical analysis of each of the ten votes, see Appendix A. For a description of the votes, see Appendix B. Source for all tables: Appendix A Detailed analysis of votes included in Tables 3-5 The tables that follow present means and statistical analyses of the six votes on bills in the House of Representatives, and the four votes on bills in the Senate, that were pooled to construct Tables 3, 4, and 5. They also present an analysis of Congressional voting that looks only at Democratic members of Congress. A short description of the ten votes is provided in Appendix B. Table A-1 State economic conditions and House of Representatives support for monetary populism in the 1890s: Means of all explanatory variables for some important votes Variables Table Free Free Coinage Bland Coinage Bill Bill Amendment Democratic Party .730 .753 .629 Soft money .911 .929 .866 Hard money .558 .609 .511 Real estate debt as share of land value .149 .149 .149 Soft money .118 .113 .119 Hard money .180 .181 .166 Share of labor force in: Agriculture and mining .407 .409 .411 Soft money .521 .539 .552 Hard money .293 .296 .333 Professional service .041 .041 .041 Soft money .039 .038 .038 Hard money .044 .043 .043 Personal service .189 .189 .188 Soft money .172 .169 .167 Hard money .207 .205 .200 Trade and transport .143 .143 .143 Soft money .120 .115 .114 Hard money .166 .167 .159 Mechanical/manufacturing .219 .218 .216 Soft money .149 .139 .130 Hard money .290 .288 .264 Table A-l, continued Variables Repeal of Free Silver Gold bonds Sherman Act Override Authorization Democratic Party .670 .687 .648 Soft money .773 .850 .630 Hard money .623 .497 .669 Real estate debt as .151 .150 .148 share of land value Soft money .126 .113 .125 Hard money .164 .196 .175 Share of labor force in: Agriculture and mining .408 .430 .417 Soft money .553 .541 .505 Hard money .336 .289 .311 Professional service .041 .041 .041 Soft money .038 .038 .040 Hard money .042 .045 .042 Personal service .189 .185 .188 Soft money .162 .169 .172 Hard money .202 .205 .207 Trade and transport .143 .140 .141 Soft money .116 .115 .123 Hard money .156 .172 .162 Mechanical/manufacturing .219 .204 .213 Soft money .130 .137 .159 Hard money .263 .289 .277 Table A-2 State economic conditions and Democratic House of Representatives support for monetary populism in the 1890s: Means of all explanatory variables for some important votes Variables Table Free Free Coinage Bland Coinage Bill Bill Amendment Real estate debt as .138 .138 .132 share of land value Soft money .109 .104 .102 Hard money .184 .180 .158 Share of labor force in: Agriculture and mining .434 .433 .456 Soft money .530 .547 .579 Hard money .286 .294 .352 Professional service .039 .039 .038 Soft money .037 .037 .034 Hard money .042 .042 .041 Personal service .185 .185 .184 Soft money .171 .169 .163 Hard money .208 .205 .201 Trade and transport .135 .135 .131 Soft money .115 .110 .103 Hard money .167 .167 .154 Mechanical/manufacturing .206 .206 .192 Soft money .147 .137 .120 Hard money .297 .292 .252 Table A-2, continued Variables Repeal of Free Silver Gold bonds Sherman Act Override Authorization Real estate debt as .137 .132 .132 share of land value Soft money .107 .098 .103 Hard money .153 .201 .163 Share of labor force in: Agriculture and mining .450 .473 .460 Soft money .604 .570 .571 Hard money .364 .279 .342 Professional service .039 .038 .038 Soft money .033 .035 .034 Hard money .041 .043 .041 Personal service .184 .180 .183 Soft money .154 .164 .164 Hard money .201 .213 .203 Trade and transport .132 .127 .129 Soft money .099 .110 .104 Hard money .151 .178 .156 Mechanical/manufacturing .196 .182 .189 Soft money .111 .127 .127 Hard money .244 .293 .257 Table A-3 State economic conditions and House of Representatives voting on some important monetary bills: Statistical evaluation a. To table a Free Coinage Bill (H.R. 4426), March 24, 1992. Failed, 148-149. Democrats 81-130, Republicans 67-11, Populists 0-8. A negative vote (coefficient) is soft-money. (2) Logit OLS Odds ratio Constant 7.860 2.475** 13.484** (1.30) (9.07) (5.69) Democratic Party -1.976** -0.055 -0.439 (4.53) (0.43) (0.39) Real estate debt as 10.939** 2.240** 14.512* share of land value (2.90) (2.66) (1.98) Share of labor force in: Agriculture and mining -10.821 -2.937** -20.350 (1.52) (2.02) (1.61) Professional service 21.165 6.962 68.515 (0.83) (1.04) (1.17) Trade and transport -44.554** -10.243** -85.341** (2.67) (2.72) (2.61) Mechanical/manufacturing 9.829 0.362 9.142 (1.13) (0.26) (0.75) Number of observations 300 44 44 Percent correctly 82.334 --- --- predicted pseudo r-squared .438 --- --- Adjusted r-squared --- .578 .553 54 Table A-3, continued b. Resolution to consider a Free Coinage Bill (S 51), July 13, 1892. Failed, 136-154. Democrats 121-94, Republicans 10-60, Populists 7-0. A positive vote (coefficient) is soft-money. (1) (2) (3) Logit OLS Odds ratio Constant 6.255 -0.850** -8.758** (0.89) (3.11) (3.61) Democratic Party 2.185** 0.081 0.798 (4.20) (0.62) (0.69) Real estate debt as -14.469** -2.045** -17.959** share of land value (2.81) (2.37) (2.33) Share of labor force in: Agriculture and mining -5.690 2.135 14.078 (0.69) (1.47) (1.09) Professional service 29.889 -4.557 -28.311 (1.05) (0.67) (0.47) Trade and transport 8.965 7.904** 63.766 (0.52) (2.10) (1.90) Mechanical/manufacturing -32.166** -1.017 -13.658 (2.81) (0.72) (1.09) Number of observations 283 43 43 Percent correctly 83.04% --- --- predicted pseudo r-squared .414 --- --- Adjusted r-squared --- .577 .558 Table A-3, continued c. The Bland Amendment, to set the mint prices of silver and gold at a ratio of sixteen to one (HR1 Amendment), August 28, 1893. Failed, 125-226. Democrats 97-115' Republicans 15-110; Populists 8-0. A positive vote (coefficient) is soft-money. (1) (2) (3) Logit OLS Odds ratio Constant -0.151 0.900** -10.417** (0.03) (2.89) (3.96) Democratic Party 1.286** -0.121 -1.162 (3.07) (0.70) (0.80) Real estate debt as -5.358 -1.490 -13.218 share of land value (1.30) (1.51) (1.58) Share of labor force in: Agriculture and mining 1.828 2.417 17.237 (0.28) (1.46) (1.23) Professional service -35.431 -17.511** -125.461* (1.35) (2.01) (1.70) Trade and transport 37.634** 10.497** 86.759** (2.38) (2.40) (2.34) Mechanical/manufacturing -29.039** -0.740 -9.762 (3.16) (0.47) (0.73) N 337 43 43 pseudo r-squared .559 --- --- Percent correctly 82.794 --- --- predicted adjusted r-squared --- .416 .400 Table A-4, continued d. To repeal the Purchase Clause of the Sherman Act (HR 1), November 1, 1893. Passed, 194-94. Democrats 123-72; Republicans 71-17; Populists 0-5. A negative vote (coefficient) is soft-money. (1) (2) (3) Logit OLS Odds ratio Constant 8.223 2.883** 12.277** (1.37) (10.11) (5.22) Democratic Party 0.351 0.340** 3.204** (0.76) (2.18) (2.48) Real estate debt as 4.747 2.052** 18.088** share of land value (1.15) (2.22) (2.38) Share of labor force in: Agriculture and mining -12.134* -3.734** -21.705* (1.73) (2.38) (1.68) Professional service 58.688* 17.475** 130.361* (1.76) (2.20) (1.99) Trade and transport -69.050** -12.877** -92.630** (3.90) (3.25) (2.84) Mechanical/manufacturing 23.278** -0.610 4.387 (2.32) (0.40) (0.35) Number of observations 282 41 41 Percent correctly 82.624 --- --- predicted pseudo r-squared .430 --- --- Adjusted r-squared --- .509 .519 Table A-5, continued e. To override the President's veto of the Silver Seignorage Bill (HR 4956), April 4, 1894. Passed (but insufficient to override veto), 144-114. Democrats 119-59; Republicans 18-55; Populists 7-0. A positive vote (coefficient) is soft-money. (1) (2) (3) Logit OLS Odds ratio Constant 2.504 -0.195 -2.090 (0.26) (0.63) (0.78) Democratic Party 1.320** 0.155 0.878 (2.57) (0.93) (0.60) Real estate debt as -23.786** -3.198** -27.368** share of land value (3.97) (3.24) (3.14) Share of labor force in: Agriculture and mining 1.796 1.284 7.416 (0.20) (0.76) (0.50) Professional service 15.146 2.447 14.594 (0.47) (0.29) (0.19) Trade and transport 29.274** 5.737 43.930 (2.23) (1.33) (1.15) Mechanical/manufacturing -28.539** -1.588 -18.880 (2.23) (0.94) (1.27) Number of observations 246 42 42 Percent correctly 86.184 --- --- predicted pseudo r-squared .379 --- --- Adjusted r-squared --- .528 .532 Table A-6, continued f. To authorize gold bonds (HR 8705), February 7, 1895. Failed, 135-162. Democrats 89-95; Republicans 46-58; Populists 0-9. A negative vote (coefficient) is soft-money. (1) (2) (3) Logit OLS Odds ratio Constant 9.978** 1.769** 10.547** (1.96) (6.55) (4.41) Democratic Party 1.843** 0.358** 3.796** (4.51) (2.40) (2.87) Real estate debt as 15.192** 2.185** 17.658** share of land value (3.84) (2.53) (2.33) Share of labor force in: Agriculture and mining -17.603** -2.607* -23.043* (2.89) (1.82) (1.81) Professional service 4.942 6.083 81.407 (0.20) (0.81) (1.22) Trade and transport -40.219** -7.434* -68.328** (2.69) (1.96) (2.03) Mechanical/manufacturing -2.817 -0.021 -1.018 (0.41) (0.01) (0.08) Number of observations 290 43 43 Percent correctly 78.974 --- --- predicted pseudo r-squared .478 --- --- Adjusted r-squared --- .505 .474 Table A-7 State economic conditions and Democratic House of Representatives support for monetary populism in the 1890s: Statistical evaluation (logit only) Table Free Free Coinage Bland Coinage Bill Bill Amendment positive=hard positive=soft money money Constant 11.5494 4.468 -4.647 (1.74) (0.58) (0.75) Real estate debt as 15.616** -11.886* -7.822 share of land value (3.12) (1.93) (1.38) Share of labor force in: Agriculture and mining -17.745** -1.705 8.624 (2.26) (0.19) (1.17) Professional service 25.599 64.375* -34.131 (0.85) (1.80) (0.97) Trade and transport -55.043** 3.110 45.152** (2.79) (0.17) (2.33) Mechanical/manufacturing -0.399 -25.598** -18.448* (0.04) (1.99) (1.77) Number of observations 219 213 212 Percent correctly 83.56% 84.51% 78.77% predicted pseudo r-squared .450 .436 .474 Table A-7, continued Repeal of Free Silver Gold bonds Sherman Act Override Authorization positive=hard positive=soft positive=hard money money money Constant 17.465** -11.304 7.433 (2.23) (1.16) (1.17) Real estate debt as 0.408 -39.544** 11.305** share of land value (0.06) (3.87) (2.06) Share of labor force in: Agriculture and mining -22.643** 11.450* -12.287* (2.42) (1.80) (1.64) Professional service 69.032** -11.196 6.089 (2.42) (0.23) (0.18) Trade and transport -82.265** 77.150** -35.104* (3.18) (2.69) (1.87) Mechanical/manufacturing 11.938 -11.152 6.613 (0.89) (0.72) (0.65) Number of observations 189 169 188 Percent correctly 82.544 86.984 78.724 predicted pseudo r-squared .519 .532 .474 Note: Sample includes only Democratic members of Congress. Table A-8 State economic conditions and Senate support for monetary populism in the 1890s: Means of all explanatory variables for some important votes Variables Free Coinage Peffer Repeal of Bill Amendment Sherman Act Democratic Party .500 .547 .544 Soft money