Mechanization, Transportation, and the Location of Industry in Germany, 1846–1907

This dissertation investigates the question: why do some regions industrialize and others do not? The large variation in industrial development across countries and across regions within these countries has long been of interest to economists. Increasing regional disparities in the emerging economies of India and China suggest that this debate is still relevant for modern industrial development. My dissertation looks at an example of successful catch-up growth from the nineteenth century that provides an opportunity to compare very different regional experiences within a large economy. The German Empire of 1871 transformed from a mostly agricultural economy into an industrial leader over the second half of the nineteenth century, but industrial growth was concentrated in a few regions and several areas lost traditional manufacturing centers in the process.

The central thesis of this dissertation is that technological change in manufacturing and transportation contributed to the regional concentration of industrial development. In other words, the industrialization process itself caused regional disparities. Specifically, the adoption of steam-powered machinery created incentives for manufacturers to concentrate production in central areas and around coal fields. At the same time, the railroad boom lowered transport costs, making it feasible to serve distant markets from these central locations. As a result, the Ruhr Area gained industrial employment in large numbers while regions in Bavaria and East Elbia lost traditional manufacturing centers.

The first chapter focuses on the transition from water to steam and later electric power in manufacturing and its impact on the regional concentration of employment. Water-powered establishments had to locate on streams with sufficient power and away from neighbors to avoid conflict over water rights. A new establishment could diminish the power that was available to upstream neighbors if it created too much backwater. Similarly, downstream wheels could be hurt by water storage if dams reduced the water flow while they filled up reservoirs. Evidence that such conflicts existed in Germany comes from Hunsrueck Mountains, where at least sixteen neighboring ironworks fought repeatedly over water rights in the eighteenth century, and Bergneustadt, where the Prussian water authority resolved similar complaints over storage levels in the early twentieth century (Nicke 1999, pp. 110–11; and Schmitt 1961, p. 174).

The transition to steam power ended this potential for conflict among neighbors and introduced new incentives for manufacturers to form industry clusters. The introduction of steam engines created dependence on coal, which was expensive to transport, encouraging the location of manufacturing plants near coal mines. Installation and maintenance of steam engines and machinery also required the help of trained technicians, who were still scarce in Germany in the middle of the nineteenth century (Redlich 1944, p. 123). This suggests that manufacturers had an incentive to locate near mining and manufacturing centers with enough demand for machines and repair services to attract skilled engineers.

To empirically test the impact of mechanization on industry location, I compute indices of geographic concentration and co-location with coal mining for 124 manufacturing industries in 1875, 1895, and 1907 (Ellison and Glaeser 1994). These indices are regressed on measures of industrial power use with industry-fixed effects to control for unobserved characteristics. The results show that industries with larger water power establishments were geographically less concentrated, substantiating the claim that water power operations prevented the formation of industry clusters. In contrast to water wheels, the use of steam engines raised the geographic concentration of manufacturing industries. In some industries like machines, artificial dyes, and rubber, the growth of industry clusters away from coal mining regions suggests that this rise in concentration was not necessarily connected to cost savings for the transportation of fuel. Instead, it is likely that these industries became more concentrated because manufacturers took advantage of agglomeration economies to reduce the costs of introducing powered machinery. In industries like iron production, screws, and cotton spinning, new clusters were located in coal mining regions, which suggests that geographic concentration was the result of cost savings for the transportation of coal.

Lastly, the regression analysis indicates that the introduction of electric motors lowered the geographic concentration of industries. This suggests that the adoption of electric power did not contribute to the formation of industry clusters, but instead allowed manufacturers to spread out across more regions. One explanation for this finding is that the introduction of electricity led to a revival of manufacturing in southern regions that did not have coal deposits but possessed large water power sites which could be used for electric power generation.

The second dissertation chapter focuses on the importance of cheap coal for regional industrialization in the nineteenth century. The main goal of this chapter is to separately identify the impact of access to coal and access to the population centers that formed around coal mines. The introduction of steam engines and coke smelting created dependency on coal in manufacturing but it also contributed to the formation of new urban centers around coal mines. Especially in industries like textile production, where coal was a relatively minor input, these new population centers might have been the real reason for firms to locate production close to coal mines.

To separate these two factors, I estimate a location choice model in which firms decide to locate their employees in a particular district based on its profitability. This profitability is a function of the district's access to coal and its access to consumer markets. Digital maps of these districts and Germany's transportation network allow me to calculate the costs of shipping coal from each mining district to a potential manufacturing site. Similarly, I calculate the cost of shipping manufactured output to regional consumer markets in Germany and neighboring countries. Data on manufacturing employment and transportation costs from 1846, 1861, 1875, and 1882 allows me to exploit variation across districts and over time as transportation costs fell and manufacturing employment expanded.

Railroad construction likely responded to the regional demand for transportation services, which raises concerns about the endogeneity of shipping costs. I address these concerns by using instrumental variables that are based on the straight line distances between districts and the growth of the national transportation network as a whole. An additional concern arises from the possibility of spatial spillovers between neighboring districts. I extend the methodology of Steven T. Berry, James Levinsohn, and Ariel Pakes to estimate the combinatorial logit model by Chaushie Chu, which offers a convenient way to allow for spatial autocorrelation in this setting (Berry, Levinsohn, and Pakes 1995; Chu 1989).

The results show that access to coal was more important than access to consumer markets for the location of employment in metal production and textiles. It is not surprising that coal mattered more than consumer markets for metal production, where fuel was the most important input measured in volume. It is surprising that textile production was concentrated in regions with cheap access to coal because fuel is generally not considered a major input for this sector. The explanation might be found by looking at machine tool making.

For machine tool makers, access to coal mattered more than access to consumer markets during industrial take-off between 1846 and 1861, but consumer markets became more important between 1875 and 1882. This finding matches the historical narrative about the German machine tool industry, which was tightly connected to mining during its early development but branched out to serve other industries later on (Redlich 1944). The result also suggests an explanation for the location of textile makers in coal mining regions. The mechanization of textile production required access to trained technicians. The fact that many of these technicians worked in coal mining might have been sufficient motivation for textile makers to locate in the same areas as well.

The third chapter takes a closer look at the effect of the railroad boom on industrial growth, paying attention to the different experiences of existing manufacturing centers and peripheral regions with low levels of industrial employment. On the one hand, transportation improvements could have promoted the growth of industry in peripheral regions where land and labor were presumably cheaper than in industrial areas. A reduction in transport costs could have attracted manufacturers to the periphery by making it cheaper to ship output back to markets in industrial regions.

On the other hand, transportation improvements could have increased the value of existing manufacturing centers, leading to more concentration in industry. It is possible that high transportation costs before the railroad boom acted as a protective barrier around regional markets. If location specific endowments or economies of scale and agglomeration in existing manufacturing centers were more valuable than cheap factor inputs in the periphery, removing this barrier could have led to a decline of manufacturing in regions with relatively low levels of industrial employment.

I test these predictions using a growth regression with per capita employment in manufacturing on the left-hand side and access to consumer markets on the right-hand side. The level of employment in manufacturing across all industries serves as a measure of industrial development in this setting with relatively large regions that each had a mix of employment in industry, agriculture, and services. The index of market access provides a convenient measure of transportation costs that considers all regional connections within the economy. An interaction term of growth in market access with the level of per capita manufacturing allows the impact of transportation improvements to vary across regions with different levels of industrial employment.

The results show that improvements in market access had a negative impact on manufacturing growth in regions with below median levels of per capita manufacturing. For regions above this mark, the impact of improved market access was positive. The finding indicates that transportation improvements did not support industrialization in regions without a strong manufacturing base but instead contributed to the geographic concentration of manufacturing.

Together the three dissertation chapters show that the regional concentration of early industrialization in Germany can be explained in part by technological change itself. One factor in this process, the dependence of manufacturers on coal, was specific to the era of steam engines. Most modern industries use electric motors, so they no longer need to have direct access to coal. The consequences of the railroad boom are more directly relevant for today's developing economies. Transportation investment is often used as a public policy tool to promote development in economically lagging regions by making it easier for them to participate in national markets. The findings in this dissertation caution against this use of transportation investment as a tool for regional convergence and suggest instead that regional imbalances may be an integral part of industrial development.

THERESA GUTBERLET, Rensselaer Polytechnic Institute

REFERENCES

• Berry, Steven T., Levinsohn, James, and Pakes, Ariel. “Automobile Prices in Market Equilibrium.” Econometrica 63 (1995): 841–90.

• Chu, Chaushie. “A Paired Combinatorial Logit Model for Travel Demand Analysis.” In Transport Policy, Management, & Technology Towards 2001: Selected Proceedings of The Fifth World Conference On Transport Research, Vol. 4. 1989.

• Ellison, Glenn, and Edward L. Glaeser. “Geographic Concentration in U.S. Manufacturing Industries: A Dartboard Approach.” NBER Working Paper No. 4840., Cambridge, MA, August 1994.

• Nicke, Herbert. Fleiss und Wasserkraft: Oberbergische Gewerbe- und Industriegeschichte am Beispiel der Familie Ising in Bergneustadt. Nümbrecht, Germany: Martina Galunder Verlag, 1999.

• Redlich, Fritz. “The Leaders of the German Steam-Engine Industry During the First Hundred Years.” The Journal of Economic History 4, no. 2 (1944): 121–48.

• Schmitt, Robert. “Die Geschichte der Rheinboeller Huette.” Tradition: Zeitschrift fuer Firmengeschichte und Unternehmenbiographie 6 (1961): 155–88.

Bullion, Bills, and Arbitrage: Exchange Markets in Fourteenth- to Seventeenth-Century Europe

Markets are critical to economic growth. As claimed by Smithian growth theory, markets encourage the division of labor and specialization. Compared with other factors, such as investment and technology, markets played an even more important role in economic development in the early modern era, when the economy was dominated by agriculture and manufacture operated on a small scale often under the control of guilds. In markets, surpluses could be sold, demand for various commodities satisfied, and people could obtain signals for their production and consumption decisions. The ability of markets to encourage economic development greatly depended on the degree of their integration and efficiency.

Compared with the grain market, in which degree of integration is susceptible to nonmarket factors (Aloisio 2007; Studer 2009; Howell 2010; Pfister, Uebele, and Albers 2011), financial markets come close to satisfying the conditions of arbitrage: trading in homogeneous goods and much freer from constraints imposed by regulations and transportation. Hence, it provides a better ground from which to study market integration and efficiency than does the grain market. Perhaps owing to the availability of data, most studies examine financial market development only from the eighteenth century onwards. Recently, by analyzing the difference between the nominal exchange rates of currencies and their mint parity, scholars have extended the quantitative study of financial market integration back to the late medieval and early modern period. They reveal that though progress in integration was noticed in the early sixteenth century, financial markets in fourteenth-to-sixteenth century Europe took quite a long time (approximately one year) to adjust (Volckart and Wolf 2006; Bernholz and Kugler 2008; Kugler 2009; Chilosi and Volckart 2011). However, these findings are based on low-frequency, i.e. annual, data, and do not provide a benchmark for judging whether the degree of integration, in terms of the speed of adjustment, was high or low by the standards of the time. Therefore, it is difficult to identify whether the great length of adjustment is due to the use of annual data, which probably gives upwardly biased results; (Federico 2012, p. 487) or reflects technological determinants of integration in the contemporary conditions, e.g., the speed of transportation and communication; or indicates genuinely poor market integration.

Because of these two problems, research has not yet offered a satisfactory assessment of the integration and efficiency of financial markets in the late Middle Ages and early modern era, nor of what factors determined the degree of integration and efficiency and whether later financial innovations had any impact on them. In order to answer these questions and to show long-term development, this thesis employs foreign exchange rates, which enable us to construct a high-frequency data set, e.g., monthly and weekly entries, to estimate the degree of financial market integration in late medieval and early modern Europe. Moreover, the thesis takes the contemporary speed of communication as a benchmark by which to judge market performances, since the flow of information was crucial to financial arbitrage.

This thesis is composed of three self-contained research papers. A contextual chapter (chapter 1) is included to address some fundamental issues shared by the three papers and illustrates how they together deal with the main theme of this thesis. Chapters 2, 3, and 4 investigate foreign exchange markets in terms of integration and efficiency in three subperiods of the late medieval and early modern eras. Chapter 2, using Datini's data, considers market integration against the backdrop of the great increase in the cost and risk of conducting business in such a hostile environment as fourteenth-century Europe, when warfare and plague were rampant. Taking advantage of the Great Debasement, 1544–1551, chapter 3 examines the impact of information asymmetry on the effectiveness of exchange arbitrage. Chapter 4 investigates northern European exchange market integration during the Franco-Dutch war, and shows that because of the financial innovations introduced at the turn of the seventeenth century, the costs associated with financial transactions fell substantially.

Originally, bills of exchange were devised to transfer money between different places with different currencies. Due to slow communication at the time, they began to function as credit instruments. Consequently, from late mediaeval times sale credit was granted through exchange dealings, the problem of temporary liquidity was solved, and capital could be more easily allocated (De Roover 1963). The dual function of bills of exchange—as means of payment and instruments of credit—rendered them indispensable in long-distance trade. Hence, the effectiveness of arbitrage on exchange markets can be used to measure how efficiently capital would have been allocated internationally in the late medieval and early modern eras.

The choice of period is, to a great extent, data-driven, but nevertheless each period reflects different stages in the development of foreign exchange markets and the political and economic circumstances of the time. Except for those that appear in the fragmentary seventeenth-century price currents, mercantile documents such as correspondence and merchants’ account books and ledgers are the richest source of information about exchange rates before the eighteenth century. The data set used to study foreign exchange markets contains some published series and some new materials collected from the archives. For the late fourteenth century, this thesis employs the published exchange rates extracted from Datini's commercial documents.Footnote ¹ In contrast with the exchange rates from around 1400, based entirely on a single source, the sixteenth-century series is a new collection of exchange rates compiled from various mercantile documents. The series of exchange rates for the seventeenth century is composed of those recorded in the Marescoe-David letters, with the addition of two other commercial documents.

Owing to transaction costs, the law of one price is held as long as the price differentials between two places remain in the band of transaction costs for arbitrage. The more integrated markets are, the quicker the adjustment will be and thus the shorter the time that the law of one price is violated. The threshold cointegration model is applied to measure the speed of adjustment and transaction cost of arbitrage. Due to slow communications, the estimated speeds of adjustment are deemed to be slow and will appear inefficient by modern standards. In order to properly interpret the results of the estimations, since arbitrage and price adjustment critically depended on the flow of information, the speed of communication in each period is used as a benchmark showing whether the foreign exchange markets in late medieval and early modern Europe were integrated and efficient.

This thesis finds that late medieval financial markets, as far as foreign exchange is concerned, were better integrated than existing studies have shown and arbitrage operated rapidly to exploit the profitable opportunities created by the government's interventions in exchange dealings. Looking at long-term developments, the degree of market integration was in close accordance with the contemporary speed of communication from the late fourteenth century to the late seventeenth century. However, the transaction costs associated with arbitrage on foreign exchange declined greatly, possibly due to the following financial innovations developed in the sixteenth and seventeenth centuries: discountability, negotiability and the endorsement of foreign bills, regular publication of exchange rates, and the establishment and engagement of public banks in handling bills of exchange (Neal 1990, pp. 1–43; Quinn and Roberds 2006). Together the results suggest that exchange markets in the late Middle Ages and early modern era were capable of allocating capital efficiently and bridging the demand and supply of funds effectively within the limits of contemporary conditions, particularly communication speeds. Consequently, though faced with the prevalent anti-usury doctrine, commercial ventures were financed in an economical way.

However, the thesis also shows that exchange market integration was temporarily hampered by warfare and monetary alterations. In the short term, market integration was disturbed by warfare which led to high transaction costs and interrupted trade. The effect of monetary alterations on market integration largely depended on the degree of information asymmetry. When information about monetary alteration was implicit and required some time to process, exchange markets could not adjust efficiently and a long interval elapsed before the restoration of equilibrium. The flow and quality of information critically determined the degree of financial market integration and efficiency in the period under survey.

In sum, this thesis advances two major contributions to the literature on financial markets. First, it uses monthly and weekly data in a medieval context and takes to a new level the quantitative analysis of late medieval financial markets. The high frequency data produces a more precise measurement of the degree of integration and, in contrast to the existing studies, the results do indicate that a high level of integration was achieved in the late Middle Ages. Second, this thesis takes the speed of communication into account and uses it to judge the estimated speed of integration. As financial integration was already known to be associated with the fall in information cost, (Chilosi and Volckart 2011) this thesis further connects the flow and quality of information to the speed of price adjustment. This is a new and useful idea in judging how well financial markets performed, since the flow of information plays an important role in financial arbitrage.

LING-FAN LI, Centre for Financial History, University of Cambridge

REFERENCES

• Aloisio, M. A. “A Test Case for Regional Market Integration? The Grain Trade Between Malta and Sicily in the Late Middle Ages.” In Money, Markets, and Trade in Late Medieval Europe: Essays in Honour of John H. A. Munro, edited by L. Armstrong, I. Elbl, and M. M. Elbl, 297–309. Leiden: Brill, 2007.

• Bernholz, P., and P. Kugler. “Financial Market Integration in The Early Modern Period in Spain: Results from a Threshold Error Correction Model.” Working Paper, 2008.

• Chilosi, D., and O. Volckart. “Money, States, and Empire: Financial Integration and Institutional Change in Central Europe, 1400–1520.” The Journal of Economic History 71, no. 3 (2011): 762–91.

• _____. “The Knowledge Economy of the Renaissance: The Advent of Printing and Financial Integration in Central Europe.” Czech Historical Review 109 (2011): 262–82.

• De Roover, R. “The Organization of Trade.” In Cambridge Economic History of Europe: Volume 3: Economic Organization and Policies in the Middle Ages, edited by M. M. Postan and E. E. Rich, 42–118. Cambridge: Cambridge University Press, 1963.

• _____. The Bruges Money Market Around 1400. Brussels: Paleis der Academien, 1968.

• Federico, G. “How Much Do We Know About Market Integration in Europe?” Economic History Review 65 (2012): 476–78.

• Howell, M. C. Commerce Before Capitalism in Europe, 1300–1600. Cambridge: Cambridge University Press, 2010.

• Kugler, P. “Financial Market Integration in Late Medieval Europe: Results from a Threshold Error Correction Model for the Rhinegulden and Basle Pound, 1365–1426.” Working Paper, 2009.

• Neal, L. The Rise of Financial Capitalism: International Capital Markets in the Age of Reason. Cambridge: Cambridge University Press, 1990.

• Pfister, U., M. Uebele, and H. Albers. “The Great Moderation of Grain Price Volatility: Market Integration vs. Climatic Change, Germany, Seventeenth to Nineteenth Centuries.” Working Paper, University of Muenster, 2011.

• Quinn, S., and W. Roberds. “An Economic Explanation of the Early Bank of Amsterdam, Debasement, Bills of Exchange, and the Emergence of the First Central Bank.” Federal Reserve Bank of Atlanta's Working Paper, 2006.

• Studer, R. “Does Trade Explain Europe's Rise? Geography, Market Size, and Economic Development.” Working Paper 129/09, Economic History Department, London School of Economics, 2009.

• Volckart, O., and N. Wolf. “Estimating Financial Integration in the Middle Ages: What Can We Learn from a TAR Model.” The Journal of Economic History 66, no. 1 (2006): 122–39.

Monetary Policy and Credit Policy During France's Golden Age, 1945–1973

My dissertation studies the policy of the French central bank (Banque de France) from 1945 to the early 1970s, during the so-called Golden Age of European growth (Crafts and Toniolo 1995; Eichengreen 2006). Contrary to the common view which describes this policy as passive and only determined by fiscal policy objectives, I show that central bank's actions were decisive in maintaining mild inflation as well as in easing access to credit and fostering investment and production growth.

Two main features are very important to understand the policy of the Banque de France—and more generally of European central banks—during this period. First, it relied on the direct management of quantities of money and credit (quantitative controls) rather than on their prices (interest rates). Second, credit policy, banking regulation, and monetary policy were constantly intertwined.

Conclusions of this dissertation shed new light on our understanding of French and West European economic growth, policies and business cycle during the “Golden Age” and offer a new interpretation of monetary policy under the Bretton Woods system.

The first chapter, The Institution of Credit in France, from the End of World War II to the Early 1980s is a long narrative of the policy of the Banque de France. It addresses important issues in French economic and political history such as the continuity between the Vichy regime and the democratic system, the new postwar institutions and their evolution, the sources of the Great Inflation of the 1970s, and the subsequent financial liberalization. It aims to provide an institutional analysis of French monetary policy over 35 years. Rather than merely describing the Banque de France as an institution or organization, I study monetary policy in a broad perspective: the institutionalization of credit, that is the set of laws, norms, social practices, and representative committees that established credit in a new way starting 1945. This process, named “nationalization of credit” at that time, was much wider than the nationalization of the central bank and commercial banks. It organized a complete network of institutions (private and public banks, semi-public credit institutions, the Treasury, the Planning Office, the Commission of Banking control, and the National Credit Council) whose main goal was to ensure that credit was allocated in line with “national and social priorities.” In other words, credit policy was “embedded” in the social and political process. It was an essential component of what John Ruggie (1982) famously called “Embedded Liberalism” to characterize the postwar economic system.

I characterize three main dimensions of the “institution of credit” and I study how they evolved over time. The first is legal: what were the laws that enabled the state to control banks and markets, and how were they enforced? The second is collective: what were the common beliefs and practices that sustained the institution? The third is the self-constraint: how can an institution implements rules that avoid undermining its stability and efficiency?

This third dimension provides an interpretation of the role of credit controls as a tool of restrictive monetary policy. When imposing credit controls in order to fight inflation, the Banque de France protected the French interventionist system of credit allocation against price instability. In the same time, credit controls were a means to pursue “selectivity” and to maintain the fundamental principles of the dirigiste system. Thus, there was no opposition between temporary credit restrictions for monetary purposes and state intervention in the allocation of credit for growth-enhancing purposes. This policy started to be questioned only in the early 1970s when the controls were deemed insufficient and when the selectivity of credit started to be viewed as creating inefficient rents and increasing too much heterogeneity between sectors.

The paper, Monetary Policy Without Interest Rates. An Evaluation of French Monetary Policy, 1945–1973, studies the use of direct quantitative controls on credit and money. It is the core of the dissertation as it shows quantitatively that interest rates were not a policy tool of the Banque de France and proposes a general method to estimate the effect of monetary policy when quantitative tools are the main instruments of monetary policy.

The main contribution of the paper is to show and explain why standard econometric methods (VAR using interest rates) are not able to account for the stance of monetary policy when quantitative controls are the main instruments of the central bank. I use a simple model to show why quantitative controls negate the role of interest rates and why, by nature, they involve the use of many instruments that change over time to adapt to banks’ avoidance of the previous set of instruments. Using interest rates to estimate the effects of French monetary policy over the period does not provide any consistent result indeed.

Following the seminal works by Milton Friedman and Anna Schwartz (1963) and Christina and David Romer (1989), I thus implement an alternative method to measure and identify monetary policy shocks. This method, called the “narrative approach,” relies on historical research and archival records in order to identify when policymakers decided to impose monetary restrictions. I construct a dummy variable that takes the value 1 when monetary policy is restrictive. The Bank of France implemented six restrictive episodes of a rather similar magnitude between 1945 and 1973.

Estimating the impact of monetary shocks in a VAR (vector autoregressions) using this accurate measure proves that monetary policy was an important determinant of production and prices fluctuations over the period. It explains around 40 percent of the variance of these variables. After two years, a monetary shock decreases money, production, and prices by about 5 percent. Moreover, the impulse response functions show a pattern very similar to standard VAR studies despite the sample, the country, the type of monetary policy, and the identification method being all quite different. Once monetary policy is appropriately measured, the French postwar experience with quantitative controls shows that usual results about the effects of monetary policy are obtained although interest rates were not the main instrument of the central bank.

The method and analytical framework developed in this paper can be extended to the many countries where the central bank has used or still uses quantitative controls.

The chapter, “The Quantities vs. Prices Tradeoff in Monetary Policy: An Interpretation of Credit Controls” provides a theoretical model that aims to explain why a central bank would use direct credit controls rather than relying on discount rate's changes. The model assumes that the choice of monetary policy instruments is not determined by macroeconomic disturbances but by the nature of the interaction between the central bank and the banks.

The paper shows that when the interest rate is not a market-clearing price then quantitative instruments and interest rates are not equivalent. The main intuition of the model is that the interest rates set by central banks have a dual role: they allocate credit and money but they also provide an incentive for investment. Then the interest rate set by the central bank also works as a signal that may exclude banks from the market. The main mechanism is that the central bank can run a restrictive monetary policy at a relatively low interest rate using credit controls and then exclude less banks (and then less projects) from the market. Compared to the market clearing case (first best), there are more banks receiving credit at the equilibrium but the high type bank receives a lower amount of credit. The rationale for such a choice could be either a political economy argument (maximizing the number of entrepreneurs), a policy to favor financial development through an increase in the number of banks, or a willingness to increase investment if the investment function has decreasing marginal returns to credit. Under some conditions, quantitative controls can thus be associated with financial development rather than financial repression. In an expansionary monetary policy, when the central bank commits to refinance all the banks whatever their demand (full allotment), the interest rate is fixed and does not play any role but the central bank adjusts the quantity of credit in function of the demand.

The model sheds light on French monetary policy from 1945 to 1973. First, the Banque de France did not follow a Taylor rule (the interest rate does not respond to variations in production, unemployment, inflation) but it followed a “quantity rule” (“Mc Callum rule) where the money supply is reacting to variations in production and prices. Second, the quantitative policy of the Banque de France can be interpreted as a second best equilibrium that increased investment and credit development but created rents. Rent creation explains the difficulties to reform the system in the 1970s.

The short fourth chapter, “France, the Dollar and the Bretton Woods System, 1960–1965,” offers a new explanation of the French influential proposals to reform the international monetary system in the early 1960s. This topic is widely debated by historians and political scientists, but without taking into account the actual functioning of French monetary policy. The French views on the dollar during the 1960s cannot be deemed only as a political exploitation of the U.S. balance of payments deficit by the President de Gaulle. I study the making of a consistent French views on the failures of the Bretton Woods system and explains how it led to reform proposals aimed to define new “rules of the game” for monetary discipline and central bank cooperation. The paper deals with the functioning of the gold pool, the collective reserve unit proposal and Banque de France's policy between 1960 and 1965. It demonstrates that the economic motivations of French political postures have to be found in the interwar international monetary negotiations, in the reactions towards forms of cooperation led by the United States as well as in the new policy tools used by the Banque de France. The success of credit controls in 1958/59 and 1963–1965 in France is key to understand French views on the international monetary system and foreign policies at that time.

The last paper, “Financing a Planned Economy: Institutions and Credit Allocation in the French Golden Age of Growth (1954–1974),” investigates the role of monetary policy in the allocation of credit during the highest period of sustained growth in French history. This research is motivated by the fact that whereas the importance of financial development for growth has been extensively studied and documented by many historians and economists, the recent literature about postwar European growth and the catch-up process has largely ignored the financial factors and institutions that supported this tremendous growth. Most of all, interventionist measures of the early postwar period are often associated with financial repression whereas France, as well as other European countries, actually experienced financial deepening and growth.

I especially highlight how the central bank was largely involved in the dirigiste credit allocation process. The Banque de France used its discount window and other means of credit selectivity to favor priority sectors. It regulated the banks and other credit institutions and provided both general guidelines and the legal framework for credit allocation. Credit policy was closely related to the Plan and industrial policy but pursued its owned objectives with different instruments. The government and the Banque focused their interventions on medium- and long-term credit through rediscounting medium-term bills at the central bank and through the financing and supervision of semi-public credit institutions specialized in long-term loans. Medium- and long-term loans were registered as “investment credit” (crédit d'investissement) by the central bank. Credit selectivity also worked through many exemptions of credit controls during restrictive policy.

In order to assess the effects of this dirigiste allocation of credit, I construct a new database that matches the amount of short-term credit and “investment credit” in 49 sectors to corporate tax sectoral statistics between 1954 and 1974. I find a significant positive relationship between “investment credit” (medium and long term) and the marginal product of capital (MPK). Such a correlation is not observed with short-term credit. All over the period, there was convergence in the MPKs across sectors and an increase in the access of the smallest sectors to “investment credit.” Using GMM (General methods of moments) estimations, I find that the effect of “investment credit” on investment and turnover growth is significant and positive, contrary to short-term credit's which did not play any role for investment. The planned financial system thus managed to allocate medium and long-term credit in a way that supported investment in sectors where capital was more marginally productive. I argue that this was a key factor for postwar catch-up growth. I also find that the effect of contractionary monetary policy is very different across sectors: the six biggest sectors of the economy (which were also monopolistic public sectors) did not suffer from credit restrictions over the period. I finally discuss the development of misallocations and debt overhang at the beginning of the 1970s.

ERIC MONNET, Banque de France

REFERENCES

• Crafts, Nick, and Gianni Toniolo, eds. Economic Growth in Europe Since 1945. Cambridge: Cambridge University Press, 1996.

• Eichengreen, Barry. The European Economy Since 1945: Coordinated Capitalism and Beyond. Princeton, NJ: Princeton University Press, 2006.

• Friedman, Milton, and Anna Schwartz. A Monetary History of the United States, 1867–1960. Princeton, NJ: Princeton University Press, 1963.

• Romer, Christina, and David Romer. “Does Monetary Policy Matter? A New Test in the Spirit of Friedman and Schwartz,” In NBER Macroeconomics Annual, 121–84. Cambridge, MA: MIT Press, 1989.

• Ruggie, John. “International Regimes, Transactions, and Change: Embedded Liberalism in the Postwar Economic Order.” International Organization 36, no. 2 (1982): 379–415.

Essays in Education and Immigration Throughout the Twentieth Century

My dissertation addresses two issues in the history of human capital accumulation: education and immigrant assimilation. I focus on historically disadvantaged groups, including blacks in the U.S. South, ethnic white immigrants arriving in the United States during the Age of Mass Migration, and black South Africans after the imposition of Apartheid policies.

My dissertation consists of three papers. My first paper analyzes the relationship between access to schooling in childhood and criminal behavior later in life for black children born in the U.S. South and educated between 1920 and 1940. Specifically, I use the introduction of Rosenwald schools which increased access to schools for black children.

The second paper,Footnote ¹ which is jointly authored with Ran Abramitzky and Leah Boustan, concerns the extent of immigrant assimilation in the U.S. labor market during the Age of Mass Migration, a period in which the United States maintained an open border to all European immigrants and experienced large immigrant in-flows. We construct panel data sets to estimate the occupational trajectories of immigrants relative to natives between 1900 and 1920.

Finally, the third paper is interested in language of instruction in education and looks at a natural experiment in South Africa in the 1950s. The question of whether to instruct students in their mother tongue is a question that is relevant to immigrant schooling in the United States today.

In the following sections, I outline the three papers.

New Deal Policies and Recovery from the Great Depression

The U.S. recovery from the Great Depression was nearly as exceptional as the Depression itself. When Franklin Roosevelt took office in March 1933, the unemployment rate was 21 percent, real GNP per capita was lower than it had been in 1907, and much of the banking system had collapsed.Footnote ¹ Seven years later, in 1940, the unemployment rate had fallen to 9.5 percent and GNP per capita had grown 54 percent. Productivity had advanced at an unprecedented pace (Field 2011). But recovery had not been uniform. It was interrupted by a severe recession in 1937/38 in which the unemployment rate rose more than 3 percentage points.

Economic historians have long been interested in what caused the 1929 to 1933 U.S. downturn, but comparatively little work examines the recovery that followed.Footnote ² This dissertation helps fill this gap. It adds to our understanding both of the rapid recovery and of the double-dip recession. I consider three episodes that illuminate the role of fiscal policy, unionization, and devaluation in this period. The first chapter examines the effects of a 2 percent of GDP fiscal stimulus program, the veterans’ bonus of 1936. I document that fiscal policy worked: veterans spent a large share of their bonus. The second chapter examines the causes of the severe recession that followed the boom year of 1936. I show that a supply shock in the auto industry helps to explain this recessions’ severity as well as its timing and geographic incidence. Finally, in the third chapter, I consider the sources of recovery during Franklin Roosevelt's first months in office. Just as a shock to the relative price of autos helped cause the 1937-38 recession, I suggest that an increase in the relative price of farm products helped cause the rapid recovery in spring 1933.

In addition to its focus on a common time and place, two themes unify this work. First, all three chapters show the importance of non-monetary factors during the 1930s. GDP growth of 13 percent in 1936 and −3 percent in 1938 can only be explained by the interaction of monetary factors with fiscal policy and exogenous developments in the auto industry. Chapter three suggests that rapid recovery in spring 1933 was due in part to the relative price effects on farmers of dollar devaluation. To be clear, none of these chapters disputes the primary importance of monetary factors. Rather, I show that at crucial times, non-monetary factors mattered.

The second theme of this dissertation is the role of the auto industry as an amplifying mechanism for demand and supply shocks. The veterans’ bonus had large effects because veterans used their money to buy cars. Unionization in 1937 interacted with price-setting practices in the auto industry to create a large macro shock. And in spring 1933, auto production grew far more rapidly than industrial production as a whole, in part spurred by the car purchases of farmers enjoying higher crop prices.

Typically such industry-specific factors are absent from economists’ accounts of this period. The existing literature emphasizes either the economy's automatic recovery mechanisms or the importance of monetary policy. Some authors argue that output growth after 1933 reflected the disappearance of temporary negative shocks (DeLong and Summers 1988) or the economy's strong self-correcting mechanisms (Bernanke and Parkinson 1989; Friedman and Schwartz 1963). Other authors dispute that there was anything natural or inevitable about rapid recovery post-1933. Barry Eichengreen and Jeffrey Sachs (1985) do not focus on the U.S. experience, but their finding that across countries devaluation was positively correlated with recovery suggests that monetary factors were important. Christina Romer (1992) forcefully articulates the case for a monetary explanation of U.S. recovery. She complements earlier work by Peter Temin and Barrie A. Wigmore (1990) who argue that the departure of the United States from the Gold Standard in April 1933 was a monetary regime change.Footnote ³

Just as expansionary monetary policy is often thought to explain rapid growth from 1933 to 1937, contractionary monetary policy is the most popular explanation for the 1937/38 recession.Footnote ⁴ The Federal Reserve raised reserve requirements in August 1936 and again in March and May 1937. And in December 1936 the Treasury began sterilizing gold inflows. There is an emerging consensus that the key shock was the

Treasury gold sterilization rather than the increase in reserve requirements. Douglas A. Irwin (2012) estimates that absent gold sterilization, in 1937 the monetary base would have been as much as 10 percent larger.

This brief tour of the literature suggests a unified monetarist interpretation of the 1930s: when money growth was rapid, as it was from 1933 through 1936, and again after 1938, growth was rapid. When money growth slowed, as it did in 1937/38, so did growth. We shall see that this story is too simple.

My first chapter considers fiscal policy. Economists have generally accepted E. Cary Brown's (1956) statement that, “Fiscal policy … seems to have been an unsuccessful recovery device in the ‘thirties-not because it did not work, but because it was not tried”’ (pp. 863–66). In fact, this chapter demonstrates that fiscal policy, though inadvertent, was tried in 1936, and a variety of evidence suggests that it worked. In June 1936, a deficit-financed veterans’ bonus provided 3.2 million World War I veterans with cash and bond payments totaling 2 percent of GDP; the typical veteran received a payment equal to annual per capita personal income, enough money to buy a new car.

This chapter uses cross-sectional data to identify the effects of the bonus. I examine four sources of quantitative evidence: a detailed household consumption survey, cross-state and cross-city regressions, aggregate time-series, and a previously unused American Legion survey of veterans. The evidence paints a consistent picture in which veterans quickly spent the majority of their bonus. The point estimate of the marginal propensity to consume is 0.7, with at least half this spending on cars and housing. Combined with a reasonable estimate of the multiplier, this suggests that the bonus added 2.5 to 3 percentage points to 1936 GDP growth.

This chapter contributes both to the economic history of the 1930s, and to the literature on the consumption effects of fiscal transfers. In particular, I show that the marginal propensity to consume can be high even when transfer payments are very large. I argue that features of the 1936 economy, some unique to the time, some generally present in economic recoveries, were conducive to large effects from a transfer payment. Liquidity constraints were pervasive and were made binding by expectations of higher future income. Stocks of autos and housing were low, so many households chose to buy a car or house rather than save their bonus. Indeed, receiving the bonus made veterans 22 percentage points more likely to purchase a car. Finally, since interest rates were at the zero lower bound, the spending multiplier is likely to have been high.

Thus the effect of the bonus was not independent of the state of the economy. In particular, it mattered that the share of the population that owned cars was large and growing.Footnote ⁵ In chapter two, I argue that the size of the auto industry also played a central role in the 1937/38 recession. The conventional wisdom blames the double-dip recession on monetary contraction, and to a lesser extent on the backside of the veterans’ bonus and the instatement of social security taxes. But while these factors were undoubtedly important, I show that there are in fact timing, geographic, and sectoral anomalies in the recession not easily explained by macro policy. Specifically:

• The decline in industrial production from September to December 1937 was the largest three-month decline in the history of the Federal Reserve and Miron-Romer series going back to 1884.

• Manufacturing employment rose in eight states despite a 12 percent decline nationwide.

• There was a lack of comovement between components of GDP. Durables consumption fell over 17 percent while nondurables consumption rose.

My analysis points strongly to a role for the auto industry in explaining the unusual features of the recession: manufacturing employment fell most in Michigan, and auto production fell by nearly 50 percent. Furthermore, auto prices rose while the economy-wide price level fell. In the second half of the chapter, I ask what explains the peculiar behavior of the auto industry, and how exogenous developments in this industry affected the rest of the economy. I argue that a cost shock and consequent auto price increases explain the collapse of car sales and production. Unionization and an increase in raw material costs led auto manufacturers to raise prices in fall 1937. Equally important, the increase in costs combined with nominal rigidity to make the price increase predictable. Expectations of price increases brought auto sales forward and thus sustained sales during the summer and early fall of 1937, despite negative monetary and fiscal factors. When auto prices did increase in fall 1937, sales plummeted. Narrative evidence confirms this story.

To quantify the effect on auto sales, I perform a forecasting exercise. Using criteria that penalize overfitting, I choose the model that best matches the monthly pre-recession behavior of auto sales given fiscal and monetary factors and overall economic conditions. As expected, the model underpredicts sales in the summer of 1937, when consumers expected auto prices to soon increase, and overpredicts sales in 1938. This exercise implies that without the expected increase in auto prices, the fall in auto sales between 1937 and 1938 would have been 0.8 million rather than the actual 1.6 million. Using data on the price of cars sold and an estimate of the multiplier, I find that positive price expectations added 0.3 percent to GDP in 1937, while the subsequent drop-off in sales subtracted 1 percent from GDP in 1938: absent the shock, the output decline in 1938 would have been more than a third smaller.

This chapter contributes to our understanding of the severe 1937/38 recession and provides an unusually clear example of how supply shocks can affect the economy. Perhaps surprisingly, the 1937/38 recession may also help us understand the 1933 recovery. My third chapter considers the causes of U.S. growth in the immediate aftermath of Roosevelt's inauguration. From its low point in March 1933, seasonally adjusted industrial production rose 57 percent in four months.Footnote ⁶ The existing literature has emphasized two explanations for the economy's emergence from the Depression in spring 1933: first, the effects of Franklin Roosevelt's words and actions on inflation expectations, and second, the direct effects of the dollar's devaluation on farmers. Both explanations can in theory explain the economy's turnaround and the concentration of the recovery in durable goods industries, particularly autos.

It has been difficult to know whether there was a distinct channel leading to recovery through agriculture absent high frequency geographically disaggregated data. We have not known whether recovery proceeded more or less rapidly in North Dakota or in New York. The contribution of this chapter is to provide such data and to use it to test whether recovery proceeded more quickly in agricultural states. I exploit a previously unused data set of monthly, state-level auto sales to examine whether the share of a state's population engaged in farming was correlated with auto sales recovery in spring 1933. I find that it was. A one-standard-deviation increase in the share of a state's population living on farms is associated with roughly a 25 percentage point increase in auto sales growth from late winter to early summer 1933. My results imply that the recovery, at least as measured by auto sales, might have been as much as 50 percent slower if farm demand had not grown more rapidly than demand in the economy as a whole.

As in 1937 price expectations may have been an important transmission channel. I hypothesize that although higher crop prices only directly benefited the 25 percent of the population in agriculture,Footnote ⁷ they may have contributed to consumer expectations of future price increases. This in turn could have boosted consumer demand in both urban and rural areas.

This dissertation is unified both by themes and by methodology. Throughout I use disaggregated data to shed light on macro questions. I examine the micro implications of macro hypotheses: if the veterans’ bonus was effective, households with veterans ought to have spent more; if an auto shock contributed to the 1937/38 recession, employment ought to have fallen most in Michigan; and if higher crop prices contributed to recovery in spring 1933, recovery ought to have proceeded most rapidly in agricultural states. Of course, such micro tests are no panacea. The challenge is often to interpret what the micro results say about the macro question. What does a household level marginal propensity to consume reveal about the aggregate effect of the veterans’ bonus? What might the geographic dispersion of auto sales say about aggregate recovery in spring 1933? The mapping from the micro data to such macro questions often requires further assumptions. But while inevitably subject to debate, this approach has great promise for future research.

JOSHUA HAUSMAN, University of Michigan

REFERENCES

• Bernanke, Ben, and Martin Parkinson. “Unemployment, Inflation, and Wages in the American Depression: Are There Lessons for Europe?” American Economic Review 79, no. 2 (1989): 210–14.

• Brown, E. Cary. “Fiscal Policy in the ‘Thirties: A Reappraisal.’” American Economic Review 46, no. 5 (1956): 857–79.

• Cole, Harold L., and Lee E. Ohanian. 2004. “New Deal Policies and the Persistence of the Great Depression: A General Equilibrium Analysis.” Journal of Political Economy 112, no. 4 (2004): 779–816.

• Darby, Michael R. “Three-and-a-Half Million U.S. Employees Have Been Mislaid: Or, an Explanation of Unemployment, 1934–1941.” Journal of Political Economy 84, no. 1 (1976): 1–16.

• DeLong, J. Bradford, and Lawrence H. Summers. “How Does Macroeconomic Policy Affect Output?” Brookings Papers on Economic Activity 1988, no. 2 (1988): 433–94.

• Eggertsson, Gauti B. “Great Expectations and the End of the Depression.” American Economic Review 98, no. 4 (2008): 1476–1516.

• Eichengreen, Barry. Golden Fetters. New York: Oxford University Press, 1992.

• Eichengreen, Barry, and Jeffrey Sachs. “Exchange Rates and Economic Recovery in the 1930s.” The Journal of Economic History 45, no. 4 (1985): 925–46.

• Field, Alexander J. A Great Leap Forward: 1930s Depression and U.S. Economic Growth. New Haven, CT: Yale University Press, 2011.

• Friedman, Milton, and Anna Jacobson Schwartz. A Monetary History of the United States, 1867–1960. Princeton, NJ: Princeton University Press, 1963.

• Irwin, Douglas A. “Gold Sterilization and the Recession of 1937/38.” Financial History Review 19, no. 3 (2012): 249–67.

• Romer, Christina D. “The Prewar Business Cycle Reconsidered: New Estimates of Gross National Product, 1869–1908.” Journal of Political Economy 97, no. 1 (1989): 1–37.

• _____. “What Ended the Great Depression?” The Journal of Economic History 52, no. 4 (1992): 757–84.

• Roose, Kenneth D. The Economics of Recession and Revival: An Interpretation of 1937–38. New Haven, CT: Yale University Press, 1954.

• Ruggles, Steven, J. Trent Alexander, Katie Genadek, Ronald Goeken, Matthew B. Schroeder, and Matthew Sobek. Integrated Public Use Microdata Series: Version 5.0 [Machine-Readable database]. Minneapolis: University of Minnesota, 2010.

• Temin, Peter, and Barrie A. Wigmore. “The End of One Big Deflation.” Explorations in Economic History 27, no. 4 (1990): 483–502.

• Velde, Francois R. “The Recession of 1937—A Cautionary Tale.” Economic Perspectives 33 (2009): 16–37.

From Memory to Mastery: Accounting for Control in America, 1850–1880

In 1855 G.W. Fisher published the fourth edition of Scribner's Ready Reckoner. Ready reckoners—the pocket calculators of antebellum America—typically contained a short introduction, a table of contents, and pages upon pages of simple mathematical tables designed to aid in daily computations. Most were tiny: despite being “enlarged and improved” from prior editions, Scribner's measured just 5 inches tall, conveniently sized for a shopkeeper, a merchant, a butcher, or even a workman to slip into his coat pocket. If Fisher's sales figures can be trusted, Scribner's was a bestseller. The 1855 edition touted sales in excess of 100,000; the 1866 edition 300,000; and the 1872 edition more than 450,000. Scribner's was only one of dozens of similar books published around the country, and ready reckoners are but one artifact of the ascendance of commercial calculation in eighteenth- and nineteenth-century America (Scribner 1855, 1866; New Hampshire Register and Business Directory, p. 172).

From Memory to Mastery: Accounting for Control in America charts the development of commercial arithmetic and accounting in America and the English-speaking Atlantic world between 1750 and 1880. The dissertation draws on a wide array of quantitative and partially quantitative documents—from manuals and textbooks to time books, journals, and ledgers, to trial balances and casual calculations scribbled on scraps of paper. These sources survive in incredible numbers, occupying many thousands of feet of shelf space in archival collections. Together they offer a window into the reasoning of eighteenth- and nineteenth-century economic actors, revealing how they used numbers to make sense of the increasingly complex world around them. Accounting, bookkeeping, and commercial arithmetic offered a language for weighing values—both moral and monetary. An extraordinary range of people used this numerical language, condensing, classifying, and ordering information in new ways. They strove to make their businesses more profitable, to stretch their wages to cover their rent, to extract more labor from slaves, and to better understanding almost every aspect of the increasingly complex economy around them.Footnote ¹

From Memory to Mastery makes two important contributions to our understanding of the role of commercial calculation in early America. First, it adds to research that explores the shift from accounting as recordkeeping (an aid to “memory”) to accounting as a tool for an analysis and negotiation (an instrument of “mastery”). Throughout the project, I explore the ways Americans not only collected data but also began to put it to use. Second, the study reorients the history of quantitative business practices to focus attention on slave plantations. Many histories of this period continue to emphasize northern factories over southern plantations, which Alfred Chandler described as a “fundamentally ancient form” of production (Chandler 1977, pp. 63–64). The dissertation compares practices on slave plantations with those in free factories, showing that in some cases plantation bookkeeping techniques advanced ahead of those in northern factories. Planters’ control over their laborers made data both easier to collect and easier to put to use. The harsh power relations of slavery enabled sophisticated planters to adopt strict labor regimes, allocating and reallocating labor from task to task and speeding up production. By contrast, the mobility of workers in free northern factories made detailed recordkeeping necessary for tracking wages but relatively futile for detailed benchmarks and comparisons. Unlike slaves, free workers confronted with speedups and intrusive management practices could quit, and high levels of turnover thwarted efforts to collect and use productivity data.Footnote ²

The dissertation is structured in three parts, each of which explores a different aspect of the development of commercial numeracy. Part 1 (chapters 1 and 2) describes the development of quantitative practices on southern and West Indian plantations. Part 2 (chapters 3 and 4) contrasts this history with the development of numerical management in northern factories, where a wider array of men and women made use of commercial calculation, but few matched consistency and fervor evident in the most comprehensive southern account books. Finally, Part 3 (chapter 5) closes the dissertation with a statistical analysis of accounting practices among Massachusetts corporations. This section makes use of a unique data set to both clarify and complicate the qualitative close readings of account books elsewhere in the project.

Chapter 1 opens on a mid-sized plantation in Mississippi, following an individual planter as he refined his accounting techniques, eventually adopting Thomas Affleck's “Plantation Record and Account Book.” First published in the late 1840s, Affleck's journals contained every record necessary for a year of cotton or sugar planting. With the appropriate version of this fill-in-the-blanks journal, anyone who could write and add could keep sophisticated records. The book contained fifteen distinct forms, ranging from accounts of daily productivity, to records of crop yields, inventories of slaves, and an annual balance sheet. Affleck and other accounting advocates popularized a range of sophisticated techniques, including the evaluation of depreciation, the grading of slaves into fractional hands, and the measurement of productivity. They also shared the results of their efforts in the lively southern agricultural press.Footnote ³

Plantation accounting in the American South did not develop in isolation, and chapter 2 situates antebellum bookkeeping practices in a longer Atlantic story. Eighteenth-century networks of numerical expertise connected England and Scotland to the West Indies and North America. Here I draw on manuscript account books to reconstruct the management structure of sugar plantations in Jamaica and Barbados. Large sugar plantations might employ as many as six or seven bookkeepers who kept a wide array of records, including daily work logs, balance sheets of births and deaths, records of boiling, inventories of livestock, and complex abstracts to be sent back to England. They were middle managers in complex hierarchies: above them were overseers and attorneys, and below them enslaved drivers and the men who toiled in the fields. As planters, agents, overseers, and bookkeepers moved through the Atlantic world, they carried accounting practices with them. Jamaica and Barbados became the center of an exchange of ideas just as important as the trade in commodities.

Chapter 3 turns to numerical practices in free northern factories, focusing on New England and the Mid-Atlantic. Here too, the early to mid-nineteenth century was a time of innovation and experimentation, and entrepreneurs applied accounting practices to the operations of their businesses. As textile mills and iron forges grew, managers adapted a wide array accounting methods to meet their needs: they tinkered with the form of the ledger, the process of closing the books, the management of labor turnover, and the setting of piece and wage rates. These helped them to continue to know about their workers even when they no longer knew them. In many aspects northern accountants kept more complex and extensive records than elite southern planters, but in the analysis of labor productivity they sometimes lagged behind. I argue that their more rudimentary practices reflected workmen's ability to negotiate terms and quit when they were dissatisfied. The type of control possible on southern plantations was simply impossible in textile mills and iron forges that often struggled to maintain staff. Further, with capital invested in machinery rather than men, stoppages were costly. Therefore many managers and accountants focused on keeping enterprises running rather than pushing down labor costs or speeding up production.

Chapter 4, “Schools, Books, and Book-keeping,” explores the dissemination of accounting knowledge. The expanding scope for commercial calculation required a growing class of clerical workers. These men (and increasingly women) learned accounting both in school and independently using a wide array of textbooks and manuals. Common schools increasingly taught commercial arithmetic to children, and older students could attend commercial colleges to learn accounting. The spread of for-profit commercial colleges was particularly dramatic: hundreds of these schools opened across the United States, typically offering short courses in various business skills. Armed with certificates and skills from these colleges, a growing numbers of men and women went to work for large corporations, helping to organize unprecedented quantities of business information. The expansion of recordkeeping accompanied escalating division of responsibilities among clerical and information workers. Where previously the categories of clerk, bookkeeper, and accountant had been overlapping and sometimes interchangeable, firm professional boundaries came to divide highly skilled accountants from bookkeepers and clerks, the least of whom often only needed to be able to write and add.

Finally, chapter 5, “Effectiveness and Unevenness,” steps back from the close readings of individual account books elsewhere in the dissertation to assess bookkeeping practices in the aggregate. How common were bookkeeping techniques? And did entrepreneurs’ attempts to be more “rational” actually improve the performance of their businesses? Though no national data is available, in 1875 the state of Massachusetts required corporations to report their assets and earnings on a preformatted balance sheet. Analyzing how well firms completed these forms and where they committed errors offers a window into their overall level of accounting expertise. In 1875 approximately one-third of corporations failed even to balance their balance sheets, and two-thirds did not consider depreciation. The proportion balancing their books increased steadily from year-to-year, while the use of depreciation remained stable. Interestingly, those firms correctly completing the forms survived longer than their counterparts who did not. This finding—that firms using proper accounting survived longer—is robust to controls for size and industry, though difficulties in identification prevent any causal conclusions. What is clear is those firms that used “correct” accounting were among the most important and long-lived corporations in Massachusetts Together with the close readings of account books elsewhere in the dissertation, this data suggests that accounting practices were incredibly diverse but also fundamental to firm survival (Massachusetts 1876).

CAITLIN ROSENTHAL, University of California, Berkeley

REFERENCES

• “Advertisement for a New Edition of the Plantation Record and Account Book by Thomas Affleck.” Weld & Company, 1847. State Library and Archives of Florida, Florida Memory Project, Tallahassee, FL. Available online at http://edocs.dlis.state.fl.us/fldocs/FloridaPrint/BR0009.jpg.

• Chandler, Alfred. The Visible Hand: The Managerial Revolution in American Business. Cambridge, MA: Belknap Press, 1977.

• Clark, George R., and Thomas Affleck. “Eustatia Plantation Account Book.” Issaquena County, Mississippi, 1861 1860. Vol. 649. Ohio Historical Society. Available online at http://dbs.ohiohistory.org/africanam/det.cfm?ID=13902.

• Cohen, Patricia Cline. A Calculating People: The Spread of Numeracy in Early America. Chicago: University of Chicago Press, 1982.

• Cooke, Bill. “The Denial of Slavery in Management Studies.” Journal of Management Studies 40 (December 2003): 1895–1918.

• Fleischman, Richard K., David Oldroyd, and Thomas N. Tyson. “Plantation Accounting and Management Practices in the United States and the British West Indies at the End of Their Slavery eras1.” The Economic History Review 64, no. 3 (March 2011): 765–97

• Levenstein, Margaret. Accounting for Growth: Information Systems and the Creation of the Large Corporation. Stanford, CA: Stanford University Press, 1998.

• Massachusetts. “Public Document 10.” In Public Documents of Massachusetts, 1876.

• New Hampshire Register and Business Directory. Claremont Manufacturing Co., 1872.

• Previts, Gary John, and Barbara Dubis Merino. A History of Accountancy in the United States. Ohio State University Press, 1998.

• Scribner, John Marston. The Ready Reckoner: For Ship Builders, Boat Builders, and Lumber Merchants. 4th ed. Rochester: G.W. Fisher, 1855.

• ________. The Ready Reckoner: For Ship Builders, Boat Builders, and Lumber Merchants. Rochester: G.W. Fisher, 1866.

• Smith, David Eugene, and Frank Swetz. Capitalism and Arithmetic: The New Math of the Fifteenth Century, Including the Full Text of the Treviso Arithmetic of 1478, Translated by David Eugene Smith. La Salle, IL: Open Court, 1987.

• Van der Linden, Marcel. “Re-Constructing the Origins of Modern Labor Management.” Labor History 51, no. 4 (2010): 509.

• Yates, JoAnne. Control through Communication: The Rise of System in American Management. Baltimore, MD: Johns Hopkins University Press, 1989.

Comments on Gutberlet, Li, and Monnet

Comments on Eriksson, Hausman, and Rosenthal

All of the theses submitted for consideration for the Allan Nevins Dissertation Prize are drawn from a self-selected group. Research that is not of a quality that has already far surpassed the expectations of the authors’ supervisors, external examiners, colleagues and classmates, and in many cases, journal editors and referees, is never sent to the dissertation session conveners. This makes the job of selecting a short list of just three dissertations tremendously difficult. If one consults the conveners’ comments published in the The Journal of Economic History from the recent past, one can identify a set of fairly consistent criteria that have been used to assess the work submitted for consideration. Authors selected for inclusion on the short list are expected to have provided proof that they are exemplary historians, carefully accumulating new evidence, recognizing that it matters where data comes from, and worrying about context across both time and space. But that is not enough. We also want those included on the short list to show us that they are also exemplary economists, deploying the tools of theoretical and empirical economics in sensible ways, acknowledging the limitations that economists’ assumptions impose, all while working to meticulously distinguish causes from effects. Given the job of selecting just three theses for the short list for the Allan Nevins Prize, one quickly discovers that dissertations that do not meet these criteria are simply never submitted for consideration. In short, to be invited to participate in the dissertation session authors must do more than demonstrate their potential as exciting, energetic, and thoughtful historians and economists.

Naomi Lamoreaux aptly described the “extra something” that all of the short-listed theses have as a “spark” of originality and creativity. These sparks show themselves in strong narratives that carry through entire dissertations, and in the authors’ ability to link their narratives to those originating in other contexts, thereby making connections and drawing readers in to their specific time and place. And as if these criteria do not set the bar high enough, I was also looking for evidence that the authors were not blinded by their own strengths and creativity. I wanted the caveats and the weaknesses of their work to be recognized and the reader's skepticism to be anticipated and met head-on, even if this meant establishing the conditions under which one should question the authors’ conclusions. Katherine Eriksson, Joshua Hausman, and Caitlin Rosenthal were not the only authors who submitted their work for consideration who successfully told compelling stories, found connections, probed their own work with a critical eye, and showed off the skills and tools of a talented economic historian, but theirs are the three dissertations that time and again met and surpassed these demanding expectations.