The Dissertations of Vallore Arthi, Carlos Hernandez, and Vincent Geloso 2017 Allan Nevins Prize Competition of the Economic History Association

The Seeds of Divergence: The Economy of French North America, 1688–1760

Canada has been largely ignored in the debates on divergence within the Americas since the colonial era (Engerman and Sokoloff Reference Engerman and Sokoloff2002, Reference Engerman and Sokoloff2011; Allen, Murphy, and Schneider Reference Allen, Murphy and Schneider2012). Most of the attention goes to the Spanish and British colonies, and when Canada is included, it is typically blended together with the United States under the label of “North America.” This is unfortunate given its unique colonial origins. Canada’s primary basin of settlement (along the banks of the Saint-Lawrence River in the province of Quebec)Footnote ¹ shared much in common with the British North American colonies, both geographically and economically, but it was nonetheless part of the French empire and settled under a different set of institutions. This combination of distinctive features and similarities are inviting to scholars interested in the extent to which living standards varied during the colonial era (Arroyo Abad, Davies, and van Zanden Reference Arroyo Abad, Davies and van Zanden2012; Arroyo Abad and van Zanden Reference Arroyo Abad and van Zanden2016; Allen, Murphy, and Schneider Reference Allen, Murphy and Schneider2012; Challú and Gómez-Galvarriato Reference Challú and Gómez-Galvarriato2015; Lindert and Williamson Reference Lindert and Williamson2016).

The focus of this dissertation is to introduce Canada as an essential component of this literature. In doing so, my aim is to support and strengthen Canadian economic history as well as comparative work on inequality across nations throughout the early modern period. The addition of Canada is meant to document whether the colony was richer (as was the case for the American colonies relative to Britain), or poorer (as was most likely the case for Spanish colonies relative to Spain) than its motherland of France. For the purposes of comparison, this offers the collateral benefit of generating the first-ever estimates of living standards in Canada prior to the nineteenth century. And while some analysis is inherent, the main focus of this dissertation is to generate the quantitative information necessary to fuel a research agenda dedicated to exploiting Canada’s unique institutional features.

Using evidence culled from the account books of two religious congregations, both of which owned large estates across the Saint-Lawrence Valley, I construct a price index and measures of real wages and incomes that show how Canada was in fact the poorest colony in North America (and a middling colony within the Americas). In some respects, it was richer than France, but this is largely resultant from the type of “basket” used to measure real wages; from 1688 to 1760, it experienced no growth.

Technology Adoption and Product Diversification in the Brewing Industry over the Nineteenth and Twentieth Centuries

The history of American and Japanese brewing is one of adaptation and resilience to economic shocks. Breweries experienced new scientific discoveries, regulation, and changes in market access over the last two centuries. Some breweries adapted to these shocks by switching to new technologies, products, and geographical markets. Other breweries did not adapt to the shocks and closed down instead, forcing consumers to quench their thirst with beer from the more productive or resilient breweries. This process of adaptation and selection shaped the structure of the brewing industry and introduced competition and production techniques in the soft-drink and biotechnology industries. My dissertation studies the interaction of selection and adaptation as a mechanism for evolution in the brewing industry.

In particular, I use scientific discoveries and regulation as a source of historical experiments that have induced the adaptation, growth, or closure of breweries throughout the nineteenth and twentieth centuries. My contribution to the historical literature is to study the entire brewing industry using longitudinal data over multiple decades at the plant level, as opposed to focusing on particular breweries or state-level data which has been the standard approach up until now.Footnote ¹

In the first chapter, I study the rise of the U.S. Midwest as a brewing powerhouse after pasteurization was invented in the 1860s. Before pasteurization, shipping beer was very expensive due to the need for refrigeration (Plavchan Reference Plavchan1969, p. 79; Kerr Reference Kerr1998). After pasteurization was invented, breweries could ship beer in bottles without the need for refrigeration (Cochran Reference Cochran1948; Baron Reference Baron1962, p. 241; Stack Reference Stack2000). In addition, the scientific discoveries behind pasteurization allowed breweries to sell higher quality beer than ever, with a consistent, reliable taste (McGahan Reference McGahan1991; Stack Reference Stack2010). Thanks to pasteurization and its related scientific advances, bottlers were able to ship beer to previously isolated towns while charging a higher price than the local breweries in those towns.Footnote ²

I observe which breweries adopted bottling, increased output, or closed down during the following decades, as the percentage of breweries that bottled their beer increased from virtually zero in the 1860s to 21 percent in 1880. At first, bottling did not occur at the larger breweries in the more highly populated areas in the Northeast. Instead, it occurred at mid-sized breweries in the Midwest that used bottles to ship beer to regional markets.

I find that breweries in the Midwest adopted bottling earlier than breweries in the Northeast because breweries in the Midwest benefitted the most from shipping to nearby markets. In particular, breweries in the Midwest were close to (1) small and previously isolated towns where local brewers had not yet invested in quality and standardization, or (2) Chicago, where water quality was low and a fire had destroyed breweries in 1871 (Cochran Reference Cochran1948, p. 55; Baldwin Reference Baldwin1966, p. 44; Ferrie and Troesken Reference Ferrie and Troesken2008).

The adoption of bottling in the Midwest allowed for an increase in beer shipments within the Midwest that did not occur on the East Coast. As a result of the increased competition within the Midwest, the number of breweries fell substantially—in contrast with the Northeast, where the number of breweries actually grew in the absence of increased competition.Footnote ³ Furthermore, the increase in market access and quality in the Midwest allowed for an increase of 77 percent in beer output between 1874 and 1880—more than double the growth in output in the Northeast.

The increasing number of firms in the Northeast suggests that trade between regions— and hence distance to grain and other inputs—does not explain the divergence between the Midwest and the Northeast. In fact, the National Shippers—six Midwestern breweries that shipped beer from the Midwest to the Northeast—only accounted for 6 percent of the national output in 1880 and 8 percent of national output in 1898.

Is the rise of the production in the Midwest explained by a relative increase in demand due to the arrival of migrants from Europe and the Northeast? The data suggest otherwise. Most of the movement in the location of beer output occurred within the first six years of the arrival of pasteurization into the U.S. market in 1874. From 1874 to 1880, beer output moved towards the West six times faster than both total and German populations.

Instead, the rise of the Midwest is explained by the growth of regional brewers that used bottles to distribute their beer to nearby markets. Brewers adopted bottling in the Midwest, but not in the Northeast, because the benefits of shipping to nearby markets were higher in the Midwest than in the Northeast. By adopting bottling and other technological advances, these brewers increased competition, quality, and market access within the Midwest. These changes provided the initial steps in the rise of the Midwest as a brewing powerhouse.

In the second chapter of my dissertation, I study the adaptation of American brewers to the gradual enactment of prohibition during the early twentieth century. The enactment of prohibition provided an historical experiment to study how firms build capabilities as a result of demand shocks, and how these capabilities increase the resilience of firms to future, potentially larger, demand shocks.

Breweries experienced two sequential shocks. Between 1914 and 1918, many states and counties became dry.Footnote ⁴ That is, they chose to prohibit the sale and production of alcohol. Other states and counties remained wet. That is, they continued to permit the sale and production of alcohol. Breweries located in wet counties experienced a reduction in demand during local prohibition, because they could no longer ship beer to dry counties. Breweries with low transportation costs to the newly-dry counties were the most affected when the latter became dry, because they lost more customers than breweries with high transportation costs to the newly-dry counties. The second shock came with federal prohibition in 1919, when all breweries were prohibited from selling beer.Footnote ⁵ The enactment of federal prohibition required a constitutional amendment and lasted for 14 years, until it was repealed (again by constitutional amendment) in 1933. The experimental variation in this chapter arises in the era of local prohibition, when some breweries were exposed to large demand shocks, while others were insulated from these local shocks. Later, all breweries faced the common shock of federal prohibition.

I follow 1,300 breweries throughout both local and federal prohibitions. I observe the machines that breweries bought, the goods that breweries produced and the decisions that breweries made in regard to remaining in business or closing down. I link this dataset to county-level information on exposure to reductions in demand during local prohibition, and followed the adaptation and survival of breweries throughout both local and federal prohibition.

I use this dataset to test if exposure to local prohibitions in nearby counties made breweries more likely to survive federal prohibition, and how. I propose a test to disentangle the increase in survival rates due to new capabilities induced by local prohibitions (adaptation) from the increase in survival rates due to the exit of the weakest firms during local prohibitions (selection). If only selection is at work and adaptation does not occur, survival throughout the full prohibition period—from the start of local prohibition until the end of federal prohibition—does not depend on exposure to early demand reductions.Footnote ⁶ In contrast, when adaptation is also at work, survival throughout the full prohibition period can be higher among breweries exposed to early demand reductions. I test this implication of adaptation using data on brewery survival, and then corroborated my mechanism using data on machine acquisition and product switching.

I find that breweries that experienced early reductions in demand were 12 percent more likely to survive the full prohibition period—from the start of local prohibition until the end of federal prohibition—than breweries that did not experience early reductions in demand. This increase in survival occurred because a group of firms made early investments in machinery that later facilitated product switching into soda and other food stuffs. Part of this group of firms would not have survived the full prohibition period, had they not experienced an early reduction in demand. In other words, the adaptation to local prohibitions made breweries more resilient to the larger reduction in demand that occurred during federal prohibition.

The third chapter of my dissertation is coauthored with Michael Darby and Lynne Zucker. We document the entry of Japanese breweries into pharmaceutical and biochemical research after the 1980s. Scientists who were affiliated with Japanese breweries authored 81 percent of the academic articles produced by breweries all over the world between 2000 and 2005—50 percentage points more than 20 years earlier. Most of this increase in academic production occurred between 1986 and 1996, when the number of published articles affiliated with Japanese breweries increased six times.

We hypothesize that this increase in academic production is the result of product diversification towards the pharmaceutical and biochemical sectors, possibly induced by a fall of competitive barriers in the Japanese beer market that occurred in the 1980s (Craig Reference Craig1996). This diversification occurred during the early years of the biotechnology industry. At this early stage, innovation in the industry was characterized by the existence of naturally excludable knowledge whose diffusion was difficult due to its complexity or tacitness (Zucker, Darby, and Brewer Reference Zucker, Darby and Brewer1998). The practical use of this knowledge required the collaboration between academic scientists and firm scientists (Zucker and Darby Reference Zucker and Darby1996). Having done research on fermentation, Japanese breweries were in a privileged position to take part of this exchange. Necessity mixed with opportunity.

Overall, my dissertation explains the role of selection and adaptation on the evolution of the brewing industry after the invention of pasteurization, the enactment of prohibition, and the birth of the biotechnology industry. The dissertation highlights the role of economies of scale and endogenous sunk costs on the buildup of capabilities in response to these events, and the effect of these capabilities on the evolution of the industry during the nineteenth and twentieth centuries.

CARLOS EDUARDO HERNÁNDEZ, Universidad de los Andes

Human Capital Formation and the American Dust Bowl

The Dust Bowl of the 1930s was one of the greatest environmental disasters in American history (Worster Reference Worster1979). This period of extreme erosion, drought, and dust storms dealt a devastating blow to America’s Great Plains region, where the event not only posed direct hazards to health, but also threatened agricultural livelihoods (Hansen and Libecap Reference Hansen and Libecap2004; Cunfer Reference Cunfer2005; Hornbeck Reference Hornbeck2012). Motivated by the early-life health literature, which suggests that such destruction can have long-term effects on the health and wellbeing of individuals, in this dissertation, I investigate the long-term consequences of the Dust Bowl for the human capital of the children who lived through it (Almond Reference Almond2006; Almond and Currie Reference Almond and Currie2011).Footnote ¹ I focus on three main questions: first, what effects did early-life exposure to the Dust Bowl have on later-life human capital?; second, what are the sensitive periods for the development of different capabilities, and what can these findings tell us about how human capital is formed?; and third, to what extent can early damage to the child endowment be remediated?

In the first broad section of the dissertation, I outline the data and empirical approach used throughout the study, and I present the main results on the later-life effects of early-life exposure to the Dust Bowl. The main analysis takes a difference-in-differences approach, using plausibly exogenous variation in childhood Dust Bowl exposure to explain variation in human capital in adulthood. I begin with a sample of individuals who were children around the time of the Dust Bowl, whose stable human capital and wellbeing outcomes can be observed in adulthood. Specifically, I take a sample of individuals born in 15 Great Plains and adjacent states between 1900 and 1959, and enumerated in the 1980, 1990, and 2000 censuses. These rich IPUMS microdata allow me to test for effects on a range of outcomes such as educational attainment and later-life disability (Ruggles, Alexander, Genadek, et al. Reference Ruggles, Alexander and Genadek2010). Each individual in the sample is then assigned a measure of Dust Bowl exposure in early life. The temporal component of childhood Dust Bowl exposure arises from cohort membership: of this sample, 1918–1941 birth cohorts were exposed to the Dust Bowl either as children or in utero (that is, were aged −1 to 12 at some point over the period 1930–1940), and are considered to have been treated in the baseline. This cohort variation in Dust Bowl eligibility is then interacted with the severity of the Dust Bowl in the state of birth, given by population-weighted measures of soil erosion constructed from Richard Hornbeck (Reference Hornbeck2012). One benefit of this empirical approach is that by using many overlapping age and time cohorts, I can disentangle effects of age-specific exposure from effects related to the chronology of the Dust Bowl or to cumulative exposure.

In this baseline analysis, I find that exposure to the Dust Bowl in childhood has statistically significant and economically meaningful adverse impacts on later-life outcomes, for instance, increasing the probability of disability (by 1.1–1.7 percentage points) and reducing both fertility and rates of college completion (by 0.3 children and 1.1–2.1 percentage points, respectively). Here, I do find one beneficial effect of Dust Bowl exposure: those exposed have a large and significant increase in the chance of high school completion (up to 3.7 percentage points in the main analysis). This seemingly counterintuitive finding, which appears again throughout my dissertation, likely reflects a drop in the opportunity cost of secondary schooling as the demand for child labor on family farms collapsed. I explore this result in more detail later on in the dissertation. I also show that the Dust Bowl effects presented throughout my dissertation survive adjustment for four key phenomena which might be expected to confound the interpretation of the study’s broader results: the Great Depression, migration, fertility selection into the sample, and mortality selection out of the sample.

The second broad section of the dissertation studies heterogeneity in the main results, as well as the timing and mechanisms driving the formation of specific human capital outcomes. This section begins by examining results by age at exposure. To test for sensitive periods in the formation of each outcome, I replace my original measure of treatment with a duration-weighted measure that accounts for the amount of time an individual spent treated within a given age band. Using these new specifications, I find that many of the Dust Bowl’s most severe adverse effects—for instance, for disability, poverty, and college completion—occur in those exposed prenatally and in early childhood. This suggests that congenital complications in capability development, together with low parental incomes in utero and thereafter, may be to blame for such later-life disadvantage. However, for some outcomes, such as high school completion and fertility, the sensitive period appears to be much later in childhood, implicating, for instance, time use tradeoffs between child farm labor and schooling, or adaptive responses to deprivation during the growing years. Here, the contrast between the results on high school completion (positively affected, with school-going ages mattering most) and college completion (adversely affected, with the prenatal period mattering most) are especially striking: although both are educational outcomes, the results suggest that these outcomes have fundamentally different inputs and production functions.

I then turn in more detail to the mechanisms behind the results on high school completion. Although the bulk of the Dust Bowl’s effects were detrimental, I consistently find that high school completion rates rose with childhood exposure—by up to even 6.4 percentage points. I hypothesize that this is because the destruction of farmland released older children who otherwise might have worked on the family farm to attend school. Indeed, I find evidence to support this result in the short run: looking to the 1930–1940 censuses, I find that the Dust Bowl induced higher rates of school attendance amongst teenagers, using both county- and state-level measures of exposure. I then test whether the intensity of use of child farm labor on the eve of the Dust Bowl, which I calculate from crop-specific labor requirements and exogenously given crop suitability, moderates the schooling effect in both the short term (attendance) and long term (high school completion). Here, there is evidence of heterogeneity: as in recent work on the cotton South (e.g., Baker Reference Baker2015), non-white children in high-child-labor demand regions were the Dust Bowl’s biggest beneficiaries in terms of schooling. Looking at the child labor results more generally shows that a high pre-Dust Bowl demand for child labor resulted in a less pronounced increase in childhood schooling, consistent with the notion that the Dust Bowl released children from the farm and into schools only where their labor was not critical to large-scale agricultural production.

With these results in hand, I conclude the section by discussing what the contrast between the high school and college results reveals about the underlying capabilities required for each outcome, and how they are formed. The disparity between when in a child’s development high school and college completion rates are affected highlights that secondary schooling outcomes may depend less on innate ability than on access to opportunity and questions of time use. Meanwhile, success in college appears to be more a function of intellectual capability. Indeed, the fact that higher rates of high school completion did not automatically lead to higher rates of college completion suggests that the capabilities and inputs required to achieve each may in fact be very different. While for present-day households, credit constraints have been found to be less important than child ability in determining college enrolment, it should be noted that the same low incomes that may have played a role in producing poor college outcomes through fetal malnutrition, and thus, cognitive impairments in utero, may have also constrained a Dust Bowl family’s ability to afford college (Heckman Reference Heckman2007). Importantly, neither of these issues would have been major barriers to high school completion in this setting.

In the final substantive section of the dissertation, I investigate contemporaneous coping strategies, and ask whether remediating investments were effective in aiding recovery. First, I examine two potential margins of adjustment by households and individuals: fertility and migration. Although voluntary and selective restriction of fertility has been documented in response to events such as famines and recessions (see Ó Gráda Reference Ó Gráda2011), I find no evidence of this in the Dust Bowl setting, using a wealth of approaches, including analysis of vital statistics, of cohort size, and of birth history reconstructions by maternal characteristics. Similarly, using my baseline definitions of Dust Bowl exposure, I find no systematic interstate migratory response to the Dust Bowl, whether during childhood or thereafter. For those in the Dust Bowl cohort that did move, however, I find that outcomes are better than for their non-migrant counterparts. I interpret this result as reflecting the contributions of the destination-state environment, since these migrants do not appear to have been positively selected, nor is there evidence that they moved early in childhood, thus cutting short the most crucial period of treatment. In short, although migration was a successful mitigation strategy when it was undertaken, it does not appear to have been a systematic or selectively taken strategy.

The remainder of this section of the dissertation examines more systematic attempts at recovery. Because household-level compensatory investment in human capital is endogenous, one way to cleanly test for remediation is to look for a second, positive and plausibly exogenous shock to human capital, to interact with Dust Bowl exposure. In this case, I look to data on state-level spending on New Deal programs, which I find was largely unrelated to Dust Bowl severity or to state-level measures of civic engagement or public goods provision (Fishback, Kantor, and Wallis Reference Fishback, Kantor and Wallis2003). I find that public spending attenuates the adverse effects of the Dust Bowl, and that relief and loan programs were especially effective, with a 1-standard deviation increase in spending, for instance, offsetting roughly a third of the adverse college completion effect, and anywhere from one-fourth to one-third of the disability effect. Only a handful of other studies have examined resilience from early-life shocks (see Adhvaryu et al. (Reference Adhvaryu, Molina and Nyshadham2016), Gunnsteinsson et al. (Reference Gunnsteinsson, Adhvaryu and Christian2016), and reviews in Almond and Mazumder (Reference Almond2013), Duque, Rosales, and Sanchez (Reference Duque, Rosales and Sanchez2017), and Rossin-Slater and Wüst (Reference Rossin-Slater and Wüst2017)). What little evidence that arises from this emerging literature on remediation is mixed, and to my knowledge, mine is the first of these studies to show these gains persisting into adulthood. In this context, the findings in this dissertation provide striking empirical evidence of the capacity for remediation through compensatory investments. Furthermore, they suggest both that the New Deal was especially long-reaching through its beneficial effects on individuals’ life-course wellbeing, and that it helped relieve household-level constraints on corrective investment.

Although these results indicate that some measure of recovery may in fact be possible, the fact that recovery is incomplete, and that those exposed in earlier stages suffer worse outcomes than those struck later on in life, nevertheless suggests that investments in human capital cannot be perfectly substituted across periods in a child’s development. In this way, the empirical results are consistent with dynamic models of capability formation such as those outlined by Flavio Cunha and James Heckman (Reference Cunha and Heckman2008) and Heckman (Reference Heckman2007). Indeed, in the conclusion of the dissertation, I discuss my results when taken as a whole, and place them in the context of these theoretical frameworks.

Throughout this dissertation, I find strong evidence that early-life exposure to the Dust Bowl causes meaningful damage to human capital and wellbeing in adulthood. I find that these effects are generally worst in those exposed in utero and in early childhood, but that these insults to the child endowment can be substantially remediated in the presence of timely intervention. Thus, this dissertation makes two main contributions. First, it fills a gap in the literature regarding the human costs of the Dust Bowl. It suggests that by failing to account for its long-term effects on human capital, we may currently underestimate the damage caused by this important event in American history. It also provides new evidence on the lasting impact of the New Deal through a channel which has perhaps been somewhat overlooked. Second, it contributes to our understanding of human capital formation. It sheds light on the capabilities underlying key outcomes, and on the sensitive periods for their development—insights which could help policymakers better target childhood investments. Importantly, this dissertation adds to a relatively new literature on remediation, providing strong empirical evidence demonstrating the capacity for recovery from early-life crisis.

VELLORE ARTHI, University of Essex

The Dissertations of Michela Giorcelli, Trevor Jackson, and Craig Palsson: 2017 Alexander Gerschenkron Prize Competition

Twenty PhDs in economics, economic history and history competed for the Alexander Gerschenkron Prize this year. Their research covered Africa, China, European and Latin American countries, India, and Japan.

Choosing three among these dissertations was a difficult and worrisome task. Did I make a mistake and underrate some of the non-finalists? As my reading progressed, every new dissertation created an increasingly cornelian dilemma, and caused an internal struggle in me. Life is about choices!

The three panelists of the 2017 edition are Michela Giorcelli, Trevor Jackson, and Craig Palsson.

In their dissertations the finalists have done a great job in assessing the role of economic history and cliometrics for the development of our knowledge in economics and history. Their message is clear: economic history can proceed only if the nature of causal explanation, that is, what causes what and why it matters, has been resolved. In other words, the study of methods or the methods themselves employed by the three finalists become central: procedures, working concepts, rules, postulates, analytical techniques, etc. The crux of the philosophical problem: can any force be experienced directly or not? If it can, are extrapolations ever justified? Or, are extrapolations the result of personal beliefs concerning the nature of direct experience and the extrapolations justified by that experience? The result will determine which of two very different perspectives will be the best. Is causation only correlation and causal analysis the search for the maximum of consistence of correlations? Or is causation observed correlations, explained in terms of forces or mechanisms at work assuring that similar causes must be followed by similar effects?

A major challenge facing the scholars on this panel was how to organize reality into homogeneous sets or entities. This is possible only by abstraction, that is by ignoring as relevant some of the characteristics of a specific event. Even if all of the associated problems can be solved, even if reality can be organized into sets generally viewed as being homogeneous, the remaining difficulty would be to find, for a given type of effect, a list of causes known to be complete. Causal generalizations are incomplete. The reason is simple. All previous experimental situations may have included an unknown but relevant cause, and there is no way to prove that this was not the case. In other words, causal explanation consists of subsuming specific facts under generalization. Finally, it appears that the construction of causal explanations is conditional to the knowledge, experience, and beliefs of the individual attempting such an exercise. Biases unavoidably govern causal explanations… Wertfreiheit doesn’t exist! One therefore finds such words as axiom, postulate, premise, assumption, and hypothesis to designate the starting point of most research. Whatever terminology is preferred, it is the beginning of statements accepted as true without in any way establishing their validity (Diebolt Reference Diebolt2016).

This being said, my problem is as much psychological as methodological, that is to understand the mental rules being employed by the three panelists to come to a satisfactory causal explanation. This brings me immediately to a semantic puzzle, especially in terms of theories and models. The possible meaning of theory ranges all the way from an unproven assumption to a coherent set of hypothetical, conceptual, and pragmatic principles forming the general frame of reference for a field of inquiry. In history and in human and social sciences in general, theory is frequently used to designate almost all general statements. Causal analysis in history is not causal analysis in economics. The historian is concerned with the unique, not the average; with understanding, not prediction; with clarification, not control. When economists try to be more precise, they often incorporate an empirical bias and use the two words theory and model as synonyms. Paul Samuelson’s definition is here very illustrative: theory is as a set of axioms, postulates, or hypotheses to stipulate something about observable reality (1963, p. 233). But, do we need to refer to observable reality? More contemporary research in economics shows that theory can also be defined as a set of laws that are deductively inter-related and arranged logically into an axiomatic system. Of course, in so doing, this abstract way of thinking necessarily generates uncertainty in the formulation of the conclusion, as all the forces are not necessarily captured in the premises. Another viewpoint is to consider theory and model as slightly different. That is, models are a simpler and more accurate determinable state of affairs, facilitating deduction of further consequences, which can be reapplied to real systems. Theory, on the other hand, is, ceteris paribus, a more precise specification of causal explanations or laws (factual or counterfactual) enunciated in models.

Let’s turn to the three finalists, each of whom exhibits a clever and unique take on how to address these issues.

Economic Recovery and the Determinants of Productivity and Innovation: Evidence from Post-WWII Italy

My dissertation is an empirical analysis of the economic development of Italy that, between the late 1940s and the early 1970s, moved from being a war-ravaged agricultural country to the seventh most industrialized power of the world.Footnote ¹ I first analyze the process of Italian post-WWII physical capital reconstruction and the extent to which it was affected by the Marshall Plan—the program of economic and financial aid sponsored by the United States to help European recovery between 1948 and 1952. Second, I examine the long-term impact of the transfer of managerial knowledge and technology on productivity of Italian firms between 1952 and 1970, using evidence from the Marshall Plan Technical Assistance and Productivity Program. Finally, I investigate the effects on innovation of a series of reforms to the college system that Italy implemented between 1961 and 1969 with the goal of increasing STEM (science, technology, engineering, math) human capital accumulation.

This work contributes to our understanding of the role of foreign aid, management and technology transfers, and education reforms on historical economic development and innovation.Footnote ² More in general, it answers to broad economic questions that, despite being currently relevant, are hard to tackle with modern data.

Markets of Exception: An Economic History of Impunity in Britain and France, 1720–1830.

Since 2008, many scholars have turned their attention to the darker aspects of economic history, following a long stretch of relative optimism. Economic historians previously studied the origins of economic growth through trust and reputation mechanisms (Greif Reference Greif2006), efficient and rational institutions, credible constraints on sovereign power (North and Weingast Reference North and Weingast1989), and market integration (Findlay and O’Rourke Reference Findlay and O’Rourke2003). Now a great deal of cutting-edge economics research focuses more on fraud (Akerlof and Schiller Reference Akerlof and Schiller2015), instability (Galbraith Reference Galbraith2012; Calomiris and Haber Reference Calomiris and Haber2014), and of course, the persistence of inequality (Piketty Reference Piketty2014; Atkinson Reference Atkinson2015, inter alia). At the same time historians have begun to return to questions about the history of capitalism (Hilt Reference Hilt2017), and new approaches to the history of inequality.Footnote ¹

Most of the vibrant economic history work on institutions approaches them from the perspective of individual participants in normal circumstances, assuming that the “rules of the game” apply to everyone, and at all times. To put it differently, scholars often analyze economic actors as though they are price-takers in an institutional environment. But there is enough historical and contemporary evidence to indicate a need to approach these questions again, from the perspective of institutional exceptions: that is, from the perspective of a monopolist who gets to set the costs of violating rules and norms.

To address this gap, my dissertation adapts the concept of impunity from recent scholarship on international law and applies it to the economic history of European finance.Footnote ² It argues that impunity is a function of three components acting with the sphere of the economy, and that the capacity for impunity changed as the relative weight of each component changed over time. The first component is prosecutorial discretion, whether contingent and corrupt or institutionalized in the limits of jurisprudence such that certain actions are unregulated or decriminalized. The second is technical knowledge, as financial instruments became increasingly esoteric and economic theory became increasingly formalized across the long eighteenth century, ensuring that financial practices were consistently more sophisticated than securities regulation. The third is international mobility, of both capital and its owners, since European capital markets integrated sooner and more thoroughly than markets for land, labor, or commodities, which produced the problems of regulatory arbitrage and of the asymmetry between national regulations and international economic activity.

In the dissertation I argue that there was a gradual diffusion of economic impunity in Europe across the very long eighteenth century, and that the significant moments of transition were three international financial crises: 1720, 1793–1797, and 1825. Structuring the dissertation around financial crises as transitions has meant focusing on how each crisis was unique rather than part of a general pattern. Each crisis after 1720 has been informed by the distorted historical memory of previous crises, and the resolution of each crisis has defined the institutional parameters in which the next crisis took place. Each time, popular anger at perceived injustice and improvised policy responses produced efforts to eliminate the capacity for impunity, but succeeded only in changing who could act with impunity, and how.

The first chapter analyzes the disparity between the increasing complexity of international financial instruments and the simultaneously limited scope of securities regulation in Britain and France to argue that the Financial Revolution witnessed the first expansion of economic impunity from the sovereign to managers of capital, culminating in the world’s first international financial crisis in 1720. The burst of financial innovation and institutional changes between about 1680 and 1720 that scholars now call the Financial Revolution (Dickson Reference Dickson1967; Neal Reference Neal1990) marked an abrupt change in the technical complexity and international reach of capital markets, such that speculators, stockjobbers, and financiers were acting in uncharted spaces, beyond the reach of formal laws. The frantic activity in Exchange Alley and the Rue Quincampoix was connected by flows of capital, information, and personnel which were beyond the capacity of either the French or British government to regulate. When the fraud of the South Sea directors was revealed, they were not prosecuted in the criminal court of the Old Bailey—rather, they were brought before the House of Commons to be tried on four new laws drawn up for that express purpose, which meant that the Commons constituted itself as an emergency inquest to try other members of Parliament for a scheme that the Commons had itself authorized. In France, the wreckage of John Law’s famous System was cleared away by his political enemies (Murphy Reference Murphy1997; Neal Reference Neal2012). Law himself died in exile in a villa in Venice, surrounded by an extensive art collection, but the institutional damage had been done—central banks and active monetary policy were politically impossible in France through 1789. For the first time, financial instruments and techniques existed which were beyond the understanding of the educated dilettante, and were powerful enough to provoke wide-ranging economic disorder.

The second chapter shows how eighteenth century economic theorists tried to solve the problems raised by this first international financial crisis. The most important books on political economy written before the 1760s were all authored by people who were personally involved in the 1720 crisis, and all were attempts to understand both what had happened, and who had been at fault. Richard Cantillon, Nicolas Dutot, the Pâris brothers, Jean-François Melon, John Trenchard and Thomas Gordon, and even David Hume developed their general economic theories in the course of writing the specific economic history of 1720. They all read each other and argued with each other, and these three decades of pre-Physiocratic and pre-Smithian economic thought were overridingly concerned with the threat of economic despotism implied by the rise of new finance.

The third chapter uses the financial records of the speculator Étienne Clavière to illustrate the normal workings of the eighteenth century financial system and how that system came apart during the French Revolution, turning impunity into a political characteristic. International capital markets came apart after 1789. Changes to the legal category of property rights and the debate over nationalizing the royal debt prompted uncertainty and capital flight from France, which was exacerbated by issuing the assignats in 1791. The sequester of foreigners and foreign property under the Terror of 1793 broke the mechanisms of international trade, and the combination of hyperinflation in France in 1795–1797 with the suspension of gold convertibility in England also in 1797 upended the international monetary system. This separation allowed for about 15 crucial years in which British and French institutions developed along different and isolated tracks, which in turn ensured that when the international financial system was re-cast after 1815, it was on a different basis from what had obtained before 1789. This breakdown in Paris and London marked a nationalist politicization of impunity, driven by shifts in prosecutorial discretion. Once again, state sovereignty decided on economic exceptions, but now driven by the nationalist exigencies of war. This was especially true in France, but the immunity given to the Bank of England by the government when gold payments were suspended in 1797 indicates that it was not limited to the Revolutionaries in Paris. The Revolution was preoccupied with eliminating exceptional jurisdictions—the nobility, famously, but also the Church, foreigners, and autonomous cities and regions. The Terror, and especially the suspension of the Constitution of 1793 in favor of rule by penal code, marked the emergence of a new kind of purely political exception.

The fourth chapter investigates this revolutionary interregnum by comparing case studies in Dublin and Strasbourg. Both were major provincial cities where a small ruling elite attempted to govern a population with a different religion and a rising sense of ethno-linguistic nationalism. Centralizing and reforming governments attempted to exert more direct control over both cities and increasingly restricted the autonomy of their local elites, mainly through new mechanisms for the control of property. In Dublin this meant reforms against the practices of Old Corruption (Harling Reference Harling1996), which had best been exemplified by politicians like John Claudius Beresford and property institutions like the Wide Streets Commission. In Strasbourg, the Revolutionary government was overwhelmingly concerned with creating and enforcing a rigid border against foreign influence and foreign sabotage, especially regarding counterfeit assignats. The destruction of the old elite during the Terror and their replacement with people who governed under conditions of military and political exception marked the end of Strasbourg’s exceptional local status. In both Dublin and Strasbourg, the pressures of wartime reforms ended the capacity for local impunity and corruption, replacing it with exceptions determined by the central government.

The final chapter shows how international private banks like Barings, Rothschilds, B.A. Goldschmidt’s, and Parregaux-Laffitte reconstituted the European financial system after 1815, culminating with their efforts to contain the first international financial crisis of the nineteenth century in 1825. Beginning with the French reparations loans of 1817–1819, accelerating with loans to other European powers, and culminating in the lending boom to newly-independent Latin American countries after 1821, new practices in capital markets reconfigured the relationship between national regulation and international capital. London banks played a central role in funneling British savings into sovereign debt and into speculative mining stocks which ultimately proved to be foolish or fraudulent (Flandreau and Flores Reference Flandreau and Flores2009). The limit case here is of course Gregor MacGregor and his fictitious country of Poiyas. MacGregor was able to persuade a London bank to underwrite a loan for £200,000 and convinced hundreds of colonists to buy land in a country that did not exist. He was only discovered when news of the colonists’ fate reached England, but he escaped and lived a full life on a pension in Venezuela.

In contrast to 1720 or 1793, the bubble of the 1820s was generated endogenously by the financial system itself: the new expansion of the banking system both domestically and internationally, the Bank of England’s monetary policies, and the specific structure of corporate finance and sovereign lending practices produced the bubble without any need for malfeasance or exogenous shocks. The new practices of sovereign debt and joint stock companies both privileged the projectors and underwriters—the financial intermediaries required to make the new systems work. Their special position in turn made them a conduit through which sovereign defaults and Latin American mining disasters could provoke bank failures in Devon and textile bankruptcies in Lancashire.

In the moment of extreme peril, the decision about priorities and interests was taken not by a political sovereign or a regulatory legal institution, but by a private bank. My work in the archives turned up the surprising discovery that when the Bank of England was denied permission to suspend cash payments in December 1825 and was threatened with outright bankruptcy, they were bailed out by Rothschilds, who opened a secret account for them called “Account H” and obtained some £7.4 million in gold from all over Europe in order to keep the bank on convertibility. The resolution of the 1825 crisis began to answer who was competent to act, when existing institutions failed: central banks were the primary actors, responsible for maintaining the monetary system, price stability, and to a lesser extent, economic growth. But when they failed, it fell to systemically powerful private banks. They did not necessarily violate any laws, rather, they operated on their own initiative, at their own discretion, and decided on how to handle exceptional cases in legally indistinct environments.

This dissertation accomplishes three things: It develops a new interdisciplinary approach to the history of inequality by using the legal concept of impunity to illustrate how exceptions in political institutions allow for the frequent but disavowed episodes of economic dispossession that accompanied the rise of modern finance. It examines how the historical separation of economic activity and political justice in legal jurisdictions has diverged from and responded to public perception of elite impunity. Finally, it attempts to historicize financial crises, rather than taking to be eternal, inevitable, and natural events that are either the result of human psychology or impossibly contingent causes.

TREVOR JACKSON, University of California, Berkeley

Institutions and Economic Development: Lessons from Haiti’s Economic History

Haiti holds the distinction of being the poorest country in the Western Hemisphere. The comparison with its island neighbor is astounding: the Dominican Republic’s GDP per capita in 2016 was $6,722, while the same statistic in Haiti was a paltry $740. Yet economists have largely ignored Haiti’s trajectory and the lessons it provides for economic development. There is a clear culprit for this oversight: Haitian data are hard to find. In this dissertation, I use new data to elucidate Haiti’s history and show how Haiti’s experience teaches important insights about development today.

In the first chapter, I examine the origins of Haiti’s farm-size distribution. Around the world, agriculture makes up a significant share of developing countries’ economies, but the cultivation is usually performed on very small farms. Moving from small farms to larger farms has the potential to significantly improve agricultural and labor productivity (Adamopoulos and Restuccia Reference Adamopoulos and Restuccia2014). Property rights should adjust to changes in the economy (Demsetz Reference Demsetz1967), and their failure to adjust suggests something is preventing them.

Like developing countries today, Haiti missed a development opportunity at the turn of the twentieth century that favored large farms. At a time of export demand growth, many countries similar to Haiti developed large-scale, export-oriented farming. Haiti was a world leader in these exports during colonial times, yet its agricultural exports did not change in response to this demand growth. In fact, it never adjusted; in 1970, less than 2 percent of Haiti’s cultivated land were on farms larger than 20 hectares, while the same statistic for its island neighbor the Dominican Republic was 57 percent. Today Haiti is the poorest country in the western hemisphere, and its failure to take advantage of this opportunity and switch to large-scale agriculture seems to be the starting point for that divergence (Bulmer-Thomas Reference Bulmer-Thomas2012, p. 190).

I test whether transaction costs over property rights impeded the Haitian economy from taking advantage of changes in the world economy. I build on Paul Moral (Reference Moral1961), who argues that Haiti’s property rights institutions resulted in high transaction costs and caused Haiti to not develop plantation agriculture, but his case rests mostly on anecdotal evidence. I deepen this argument by showing how Haiti’s historical institutions differed from similar countries and led to higher transaction costs. Then I use a model to show how transaction costs can explain many of Haiti’s experiences in the nineteenth and twentieth century, and I introduce new data on land adoption patterns that I use to test the model’s predictions. The evidence demonstrates that transaction costs significantly impaired Haiti’s agricultural development, suggesting that today’s developing countries may face similar challenges to increasing agricultural productivity.

Three historical property rights institutions led to Haiti’s high transaction costs. First, right after independence in 1804, the Haitian government redistributed land held by the French. Second, the farmers who obtained property then passed it to heirs through partible inheritance and gave multiple people claim to the same land. This complex property rights system prevented farmers or investors from reallocating land because any claimant could veto alienation of the property. Third, the 1805 constitution banned foreigners from owning land in Haiti. Investors in the nineteenth century might have acquired the land before it became too costly, but the constitutional ban cut off this key source of capital. Even though Haiti’s factor endowments were favorable to sugar cultivation and other high-value crops, getting the land to farm was difficult because of high transaction costs. In contrast, institutions in the Dominican Republic and Jamaica, countries with similar endowments and histories, did not create the same transaction costs and allowed these countries to take advantage of the growth in export demand.

A simple model shows how differences in transaction costs can cause the patterns observed in Haiti during this time period. The model allows for trade between two small countries and the world, where labor is free to move between the two countries, but acquiring land faces transaction costs. In equilibrium, workers migrate to the country with higher productivity or lower transaction costs. When an exogenous force halts migration, the net-exporter of labor will adopt more plantation land, similar to how protectionist tariffs spur industrial development, unless transaction costs prevent them.

Testing the relationship between transaction costs and land adoption in Haiti has been difficult because of the paucity of data; however, I have collected new data to examine this question. The data come from a reform implemented by American officials during the 1915–1934 United States occupation of Haiti and consist of over 5,700 plots adopted from 1928 to 1950. The data reveal four empirical patterns. First, under favorable rental conditions, farmers did not adopt much land. In a 22-year period, farmers adopted less than 30,000 hectares of the nearly one million hectares available. Second, farmers adopted small plots. Third, farmers chose small plots even when they had the chance to adopt more land. Finally, farmers adopted land in a pattern that indicates large contiguous government land did not exist.

Combining the model and the data allows me to test the relationship between transaction costs and land adoption. First, the model predicts that in the absence of transaction costs, a labor supply shock should increase the amount of land adopted under large-scale farms relative to small-scale farms. However, transaction costs in the land market will distort adoption towards smaller farms. I test this prediction using the Trujillo Massacre of 1937. The massacre shut off the Dominican Republic as a source of labor demand and pushed refugees into Haiti’s border districts. I show that the massacre increased land adoption in the whole country, but disproportionately so in border regions, as we would expect with a labor supply increase. Yet practically no new plantations started in response to the shock, and three-quarters of the land was in subsistence farm sizes. This result is consistent with transaction costs impeding the land market.

Second, I examine how settlement patterns affect land adoption. In a Coasean world, settlement patterns have no effect on transaction costs. But in a world with transaction costs, whether someone is located on the land or not makes a big difference. I show that how households are distributed across the land has specific implications for land adoption. I test these predictions using settlement patterns derived from a geodetic map. The regressions confirm the predictions, providing further evidence that transaction costs play a significant role in the land market.

Small farms dominate Haiti’s economy because transaction costs prevented farmers from creating plantation-sized farms. Haiti’s experience suggests developing countries today that feature similar small farm agriculture may face the same transaction cost barriers transitioning to a different agricultural structure.

The second chapter in my dissertation asks why developing countries do not improve their property rights institutions. Secure property rights improve economic outcomes on many dimensions; the theory and evidence resoundingly agree on this point. In the realm of policy recommendations that could quickly improve the lives of many of the world’s poorest, property rights reforms are sold as low-hanging fruit. Yet many developing countries have not implemented the property rights reforms that would capture this potential economic growth. To improve the lives of the poor, we must understand why these countries have not already implemented them. In this paper, I argue that the policy recommendations have failed to consider the political economy impetus behind key historic reforms.

I argue that property rights proponents have focused too much on the efficiency gains and not on the political gains. Reviewing the literature, I show that political concerns motivated many of the cases used to prove the efficiency gains from strengthening property rights. The Peruvian titling program studied by Erica Field (Reference Field2007) was the government’s attempt to woo citizens away from the Shining Path. West Bengal’s tenancy reforms studied by Abhijit Banerjee, Paul Gertler, and Maitreesh Ghatak (Reference Banerjee, Gertler and Ghatak2002) were implemented by the new Left Front government to cement political control. The relevant question for policymakers is not whether property rights reforms improve economic efficiency but whether they increase political support.

I test the relationship between property rights reforms and political support using the 2014 cow registration program in Haiti. Leading up to the 2015 election, the incumbent government established a national cow registry and tagged about two-thirds of the nation’s cows. The program intended to improve cattle values and reduce theft through clear property rights. I show that at the 2015 election areas with relatively more cows also voted relatively more for the incumbent party. Using an instrumental variables strategy, I establish that the effect is not because of fraud or because the government rewarded voters for prior support.

In the final chapter of my dissertation, I again use new historical data to examine the barriers developing country governments face to keeping the rule of law. Economic development and the rule of law are closely related: the rule of law captures 63 percent of the variance in log GDP per capita across countries. Is the rule of law weak in poor countries because corrupt governments have little ability to control opportunistic behavior, or does the government intend to keep the law, but lacks the resources to do so?

In this paper, I examine a public land rental program in Haiti to see how the government honors its own contracts with citizens. The Direction General des Impots (DGI) administers the public land rental program and has branches throughout the country. Because the 2010 earthquake destroyed DGI headquarters in Port-au-Prince, collecting the data requires visiting local branches. The archives in local branches are nothing more than closets or filing cabinets, so it is difficult to obtain enough records to get enough data to understand rental patterns. Nevertheless, I was able to recover rental payments recorded in several books at these local archives.

From the books, I am able to create a panel data set of rental payments on 710 plots across three districts. The books contain limited but important information on the contract: the plot’s size, the tenant, and how much he or she paid each year. The records also note whether the tenant paid on time, and if the tenant paid late, how large of a penalty the tenant paid.

Under the program, citizens rent land from the government, and the law dictates how the contract should evolve. First, I look at whether the government follows the contract terms prescribed by law. I find that the government violates its contract. Second, I look at what drives the government to deviate from the contract. The government does not appear to target plots according to their characteristics but rather implements across-the-board changes exogenous to the plot and tenant. The government’s actions are consistent with costly contract enforcement, and the deviations match a government trying to maintain the contract’s spirit while minimizing costs. Third, I show that the government’s ability to extract rents from its tenants is limited by the tenants’ outside option. Tenants demand for state-owned land is relatively elastic, and when the government raises rents tenants leave.

In conclusion, property rights in Haiti bring new insights on economic development. Development policy says we should strengthen property rights, but Haiti’s experience suggests property rights can be too strong. Development policy says better property rights improve economic efficiency, but Haiti suggests improving property rights might improve political support. Development policy says poor countries struggle to protect property from opportunistic behavior, but Haiti suggests that even if governments want to respect property rights they might not have the resources to do so. Haiti’s history contains many unexplored puzzles that once researched will provide significant insight to development theory and policy.

CRAIG PALSSON, Naval Postgraduate School