The Handbook of Historical Economics

By Giovanni Federico

The Handbook of Historical Economics guides students and researchers through a quantitative economic history that uses fully up-to-date econometric methods. The book's coverage of statistics applied to the social sciences makes it invaluable to a broad readership. As new sources and applications of data in every economic field are enabling economists to ask and answer new fundamental questions, this book presents an up-to-date reference on the topics at hand.

• Provides an historical outline of the two cliometric revolutions, highlighting the similarities and the differences between the two
• Surveys the issues and principal results of the "second cliometric revolution"
• Explores innovations in formulating hypotheses and statistical testing, relating them to wider trends in data-driven, empirical economics

• Language: English
• Publisher: Academic Press
• Release date: Apr 21, 2021
• ISBN: 9780128162682

Book preview

times."

Part 1: What is historical economics

Outline

Chapter 1. The economic history of economic history: the evolution of a field in economics

Chapter 2. The two revolutions in economic history

Chapter 3. History as evolution

Chapter 1: The economic history of economic history: the evolution of a field in economics

Robert A. Margo¹    Boston University, Boston, MA, United States

¹Robert A. Margo is Professor of Economics at Boston University, Boston MA, and a Research Associate in the Development of the American Economy Program at the National Bureau of Economic Research. He is the author, co-author, or co-editor of 6 books, and over 150 articles, book chapters, and book reviews.

Abstract

Like other academic disciplines, economics divides into fields. This chapter sketches the evolution of the field of economic history within economics departments in the United States from the early twentieth century to the present. Economic history is different from other fields of economics in that the subject matter is shared with another academic discipline, history. Initially, scholarly practice in economic history were similar in economics and history departments but over time, the practice of economic history in economics departments has converged to that of other empirical fields in economics. While integration has brought tangible benefits for individual scholars, there are costs, such as the loss of a generalist perspective (the big picture) and lack of communication with economic historians in history departments and related disciplines.

Keywords

Cliometrics; Economics; Economic history; History of capitalism; Integration

Acknowledgements

I am grateful to Jeremy Atack, Martha Bailey, Alberto Bisin, William Collins, Giovanni Federico, Matthew Jaremski, David Mitch, and participants at the 2019 ASSA meetings in Atlanta for helpful comments.

1.1 Introduction

Academic disciplines divide up into sub-disciplines or fields, and economics is no exception. Most fields in economics are self-contained – for example, public economics or macroeconomics. Economic history is different, however, in that it belongs to two broader academic disciplines, economics and history. This chapter reflects on the evolution of economic history as a field in economics, drawing on my previous and ongoing work (Margo, 2011, 2018a, 2018b). To be clear upfront, the analysis, particularly of recent trends, applies mainly to economic history as it is practiced in economics departments in the United States; the changes described in the chapter are considerably less dramatic elsewhere in the world (Cioni et al., 2020, 2021). However, the market for economics scholarship (and scholars) is now global, and there are good reasons to believe that the style and content of economics scholarship has and will continue to converge across countries.

In the early twentieth century, when economics was still a very young discipline, professional activity in economic history was very similar in departments of history or economics. Back then, it was one of the most popular fields in economics, accounting for the large share of economics PhDs prior to World War One. Over time, economics turned more mathematical and quantitative. Other fields grew disproportionately and economic history's share of economics PhD's declined. The trend of greater use of mathematics and statistics accelerated after World War Two, at a time when the economics profession was expanding in size.

Economic historians were swept up in these trends, and cliometrics was born after World War Two. There is an urban legend that cliometrics diffused quickly and resolutely in economic history – it was Revolution, not Evolution. The urban legend, however, is just that – a legend. Compared with the analogous human capital revolution in labor economics, the cliometrics revolution was more protracted (Margo, 2018a; Cioni et al., 2020, 2021).

The comparative slowness of the cliometrics revolution reflected the influence of two schools of thought about how so-called new economic historians should conduct their business within economics. One school argued that traditional economic history was insufficiently scientific and needed the mathematical and statistical tools of post-war economics to make it more rigorous. Here, the idea was not so much to create new questions for inquiry but, instead, to reinterpret conventional wisdom in light of new data and new methods; classic examples are Fogel and Engerman (1971) and Davis et al. (1972). Economics proper found this compelling for a time because the broader discipline was in ascendance and openly imperialistic in its goals to be the queen of the social sciences. However, to win over more traditional economic historians, cliometricians had to be respectful of scholarly norms in the history profession. It is convenient to associate these this first school of thought with one of the two economic historians who later received the Nobel Prize in Economics, Robert Fogel, although the vast majority of early cliometricians agreed with this line of reasoning.

The second school of thought was quite different. It viewed economic history as a bulwark against narrow theorizing in economics about the underlying sources of long run growth and development. Growth theory came to the fore in the 1950s, as did development economics. Both turned to economic history for stylized facts and useful metaphors (for example, Rostow's (1960) stages of growth). Anathema to the second school was that the most popular growth model, namely Solow (1956), abstracted from institutions and institutional change. Growth would occur in the steady state if the rate of technical change was sufficiently positive, but technical change would not occur unless institutions were present in a society that encouraged and enabled it to take place. The role of the economic historian was to remind mainstream economics of its shortcomings, drawing on the historical record for telling examples. The second school I associate with Douglass North – although, again, there were many others who made similar arguments (and still do).

Cliometrics in its two flavors attracted new entrants to the field. Initially, this entry stabilized the long-term decline in the economic history share of economics PhDs. By the mid-1990s, there were many articles by economic historians (and even some economists) celebrating cliometrics, along with the intellectual validation of two Nobel Prizes. However, a more careful look reveals a field largely in isolation from the broader world of economics that no longer viewed cliometrics as novel. Even earlier, academic history had lost interest in cliometrics along with quantitative history more generally.

From the perspective of the survival of economic history as a field within economics this situation was not tenable in the medium run, let alone the long run. However, the seeds for another reinvention were in place that ultimately led to the present status, in which economic history as it is practiced in the United States is largely indistinguishable from other empirical fields within economics – a feature that I have called elsewhere the integration of economic history into economics (Margo, 2018a). The central feature of this integration is that scholarly identities are more fluid than in previous (scholarly) generations. Leading scholars who self-identify as economic historians – for example, by listing economic history as a field on their CV, or membership in the Economic History Association –publish in field journals other than those specializing in economic history and, increasingly, in mainstream, general-interest journals.

The other side of the coin is that scholars who self-identify as labor economists or macroeconomists or development economists work on topics in which historical evidence or argument is integral or, less often, on recognized issues in the literature of economic history. Among many examples, the most prominent involve research inspired, directly or indirectly, by the various papers by Daron Acemoglu and co-authors (for example. Acemoglu et al., 2001) that investigate how transient events or factors in the distant past could still shape economic outcomes in the present. This entire body of scholarship has little in common with traditional economic history or cliometrics. Aside from its substantive contribution, it is crucial in providing a (very) high profile example of the value of historical evidence in economics scholarship and, as such, has made mainstream economists accepting of the analysis of historical evidence as a routine part of the economic toolkit.

The main challenge posed by the integration of economic history into economics is the long-run viability of economic history as a separate field in economics. This may seem like an odd observation, if use of historical evidence in economics is becoming more popular. However, this new popularity does not appear to be generating an increase in the demand for specialists in the field per se. As foreseen a quarter century ago by Romer (1994), recent trends could put an end to economic history in the sense that individuals no longer specialize in the field, but instead specialize in the use of historical evidence in other fields of economics. That is, there are (or will be) labor economists whose general area of expertise might be wage inequality and whose knowledge extends to the long-run history of inequality, as opposed to just recent trends. At its extreme, this would mean no more generalists in economic history within economics, but rather only specialists in the economic history of relevant topics within fields. Sunk costs being what they are, something is lost if there are no longer generalist economic historians, because the whole is greater than the sum of the parts. For example, if there were no more generalists, it seems highly unlikely there would be generalist courses in economic history taught in economics departments, either at the undergraduate or graduate level. In turn, this may reduce the likelihood that economists will engage with scholars in other fields involved in research in economic history, to the detriment of all concerned.

1.2 Economics and economic history in the United States before World War Two

My narrative begins in the late nineteenth century, when economics had just been born as an academic discipline in the United States. Founded in the late 1880s, the American Economic Association (AEA) was slow at first to develop into a viable professional organization (see Margo, 2011). At the turn of the century, the AEA provided a variety of services to its members, but it was not until 1911 that the American Economic Review (AER) would appear as the organization's flagship journal. By the time the AER launched its first issue, the Quarterly Journal of Economics (QJE) had already been publishing for a quarter century, and the Journal of Political Economy (JPE) for just under two decades.

Like that of economics, the economic history of economic history in the United States begins in the late nineteenth century (Mejia, 2015; Lamoreaux, 2016). At the time, economic history was a popular field within both economics and history, more or the less the same in content and method in both disciplines. Although always in the minority, articles in economic history appeared regularly in the AER, QJE, JPE, and the American Historical Review (AHR), and there is little, if anything to distinguish between them from a disciplinary point of view. Max Ferrand's The Taxation of Tea, 1767-1773 which was published in the AHR in the January 1898 issue, could have appeared in the JPE of the 1890s just as Ella Caroline Lapham's The Industrial Status of Women in Elizabethan England, which the JPE published in 1901, would have been right at home in the AHR of the early 1900s.

By today's standards economic history attracted a remarkably high share of PhDs in economics in the early twentieth century. In 1904 the AEA began collecting and publishing information on PhDs awarded in economics. This information included the name of the student, title of the dissertation, institution, and most important, a classification by subject matter. A convenient summary analysis of these early reports is Froman (1930), who produced several useful tables showing the distribution of economics PhDs over time by school and subject matter.

Froman's Table VI (1930, p. 241) covers 3,117 economics PhDs granted from 1904 to 1928. From the institutional perspective, production was extraordinarily unequal; just three schools (Chicago, Columbia, and Wisconsin) awarded half of the PhDs.¹ The rate of PhD production was steadily increasing over time; 1,471 PhD's were awarded from 1924 to 1928, compared with just 199 PhDs from 1904 to 1908.

Overall, economic history was the most popular subject for a dissertation in economics, accounting for 12.5 percent of all economics theses over the period (1904-28) studied by Froman. This popularity, however, was declining over time – the history share declined from 18.6 percent in 1904-08 to 9.5 percent in 1924-28. The decline occurred discretely, dropping in the early 1910s and then in the late 1920s. The falling off in popularity was relative because, while the absolute number of economic history PhDs increased, this did not keep pace with the rapid growth of economics PhDs overall.

The topics of economics dissertations do not emerge in a vacuum but rather from the coursework that students undertake along the way. To the best of my knowledge, historians of economic thought have not traced the aggregate evolution of coursework at the doctoral level – in the present case, courses in economic history. However, for one important department – the University of Chicago – Mitch (2007, 2011, 2014) has produced a time line. Here I discuss Mitch's chronology before World War Two.

Founded in 1892, the Chicago economics department immediately offered graduate instruction in economic history – narrow courses on the tariff and financial history of the United States, as well as general courses in Economic and Social History and Industrial and Economic History. Chicago was far from atypical – of the 23 US institutions that offered graduate instruction in economics in the late 1890s, 15 regularly programmed courses in economic or financial history (Collier, 2015).

There were no specific course requirements in economic history at Chicago for the degree nor, for that matter, an explicit requirement to exhibit mastery of the field in some other manner (for example, an exam). This changed in 1920, when the Chicago department declared that its PhD students were supposed to have a basic knowledge of the economic history of England and the United States. By 1925, specific courses could satisfy this knowledge requirement. By the early 1940s, PhD requirements at Chicago were refined further. Graduate students had to show competence in ten specific areas, two of which were American and European economic history.

As Chicago and (presumably) other departments were strengthening their graduate requirements, economic history continued to slide in popularity relative to other fields in attracting PhD students. The AEA's methods of collecting and classifying dissertations changed over time and it is difficult to construct a consistent time series. Nevertheless, it is clear that the decrease extended into the 1930s. Point estimates that I have made for the second half of the 1930s from the AEA's reports of dissertations granted in economics suggest that around 6 percent of dissertations were on topics in economic history, which is below the level for the late 1920s (Froman, 1930, Table VI). Continued secular erosion before the war in economic history's relative share is not surprising. The bright lights of economics – Friedman, Samuelson, among others – were embracing new methods or else confronting the pressing policy issues of the Depression and impending conflict in Europe, rather than studying economic history.

1.3 The cliometrics revolution

World War Two transformed the economics profession in the United States. Economists were deeply involved in wartime planning and their role in government expanded after the war with the passage of the Full Employment Act of 1946. Veterans took advantage of the GI Bill, and large numbers enrolled in college economics courses. The prewar trend in greater use of mathematics and statistics continued unabated, indeed, accelerated after 1945. Academic economists allocated more time to producing research, and submissions to the AER and other general interest journals increased sharply in the late 1950s and early 1960s (Margo, 2011).

The early postwar years were a defining moment for economic history. As discussed earlier, the field had been losing market share among economics PhDs before the war but there were hopeful signs on the horizon. The field had its own flagship outlet in the United States, the Journal of Economic History (JEH). Established in 1941, the JEH was a joint venture of the AEA and the Organization of American Historians, acknowledging the interdisciplinary character of the field. Regular issues of the JEH appeared three times a year, with a fourth devoted to the The Tasks of Economic History consisting of papers given at the annual convention of the Economic History Association (EHA), along with discussant comments. A typical regular issue had three main articles; a similar number of shorter notes, comments, or review articles; and book reviews – roughly 100 pages in length, including front and back matter.

Another hopeful sign was the emergence of a specific intellectual demand from economics for historical evidence on development and growth. This demand had three sources.

The first was the National Bureau of Economic Research. Established in 1920, the Bureau immersed itself in the development of national income accounting, including the extension of national accounts back in historical time for the United States. This research was under the direction of Simon Kuznets. In the very first issue of the JEH, Kuznets (1941) published an important article arguing for closer collaboration between economic historians and statistical economists (Kuznets' phrase), in order to achieve what he called the final goal of economic study. Kuznets trained many PhD students – most famously, Robert Fogel, but also Richard Easterlin and Robert Gallman, who, like Fogel, front and center of the cliometrics revolution and who were to become towering figures in economic history as their respective careers unfolded. Easterlin's early work provided estimates of regional national income in the United States for various census years beginning in 1840. Gallman's dissertation and early papers provided detailed figures on aggregate output and its components and the capital stock, also beginning in 1840 (Gallman and Rhode, 2019). Later researchers (see, for example, Atack and Margo, 2019) would revise these early estimates but, to this day, they remain part of the data infrastructure used daily by American economic historians.

The second source was the Cold War, which created a pressing need in the West to provide policy advice to developing nations lest they fall into the Soviet orbit. Here, as noted earlier, the classic example is Rostow (1960). Today, Robert Fogel's dissertation-cum-monograph (Fogel, 1964), is remembered mainly for his estimate of the social savings of the railroad but his Chapter 4 and related Chapter 5 offered deep and compelling critiques of Rostow's take-off hypothesis. In doing so, Fogel demonstrated the value of cliometrics in validating – or, in this case, not – the use of historical evidence in economic argument in contemporary development economics.

The third source was growth theory. This originated before the war but received an enormous intellectual boost with the publication of Solow's (1956, 1957) fundamental papers. The assumptions of Solow's model as well as the application to growth accounting opened up major topics of research in economic history that, to this day, attract scholarly attention.

Another hopeful sign was that, despite the long run downward trend in the field's share of new economics PhDs, a numerically significant share of graduate students took courses in economic history ca. 1950 and a solid majority of graduate faculty felt that knowledge of economic history should be a required part of graduate training. We know this from an important, if neglected study of graduate education in economics, which I refer to as the Bowen (1953) study.

The impetus for the Bowen study was a prior report on undergraduate education in economics, commissioned by the AEA in the late 1940s. Impressed by the undergraduate study as it was nearing completion, the AEA's Executive Committee decided that a second study focusing on graduate education would also be valuable. An ad-hoc committee of four individuals, one of whom was Milton Friedman, was appointed to study the feasibly of a report on graduate education. The ad-hoc group was in favor, so the next step was for the Executive Committee to seek funding, which they did successfully from the Rockefeller Foundation.

The principal investigator of the AEA study was Howard R. Bowen. Born in 1908, Bowen received his BA and MA degrees from Washington State University and his PhD in economics from Iowa in 1935, followed by post-doctoral work at Cambridge and LSE. He began his career teaching at the University of Iowa but quickly moved into government service and the private sector. After World War Two, he became dean of the College of Commerce at the University of Illinois, but was forced to resign in a dispute over his efforts to reform the business curriculum to include discussion of social responsibility. Bowen was a faculty member at Williams College when the AEA hired him to conduct the study. He later moved back to Iowa where he served as president of Grinnell College and the University of Iowa. During his academic career Bowen published 14 books, 12 on the economics of higher education, on which he was a leading authority. He spent the 1951-52 academic year working on the AEA study.

Bowen used three principal sources in researching and writing his report. The first was the extant literature on graduate education, almost all of which was general rather than specific to economics and not much at that. The second, and the heart of the study, was a series of twelve questionnaires on a variety of issues sent to deans, department chairs, faculty members, and former and current graduate students. Bowen reported the statistical results from the questionnaires in simple one-way tables throughout the book. Lastly, Bowen conducted personal interviews.

The fundamental goal of Bowen's study was to make recommendations about whether PhD instruction in economics should include a core – that is, required courses in the first and second year – and if so, what those courses should be. Indeed, one of Bowen's questionnaires asked faculty what they thought on this very point – 72 percent were in favor of a core and another 9 percent were willing to consider one (depending on the details). Of those in favor of a core, there was near universal agreement – 98 percent – that there should be courses in microeconomic theory. Beyond micro-theory, however, there was less consensus but a solid majority – 55 percent – felt that economic history should be part of the core, slightly higher than the fraction that favored econometrics (53 percent).

How did faculty attitudes correspond with practice at the time? Bowen's (1953, p. 105) Table 25 gives the proportion of programs requiring, for example, courses in economic theory as well as the proportion in which students usually electing said courses if not required. A little more than two-thirds (65 percent) of programs had required courses in economic theory, but even if these were not required, the courses were routinely offered and almost always taken by students, for a total proportion of 95 percent. At 49 percent, almost half of the programs had a course requirement in history of thought; of the 51 percent that did not have a requirement, students usually took such courses in slightly more than half (53) of the programs for a combined total of 76 percent. Thus, around the middle of the twentieth century, Bowen's survey results suggest that PhD students in economics had universal exposure to doctoral level microeconomic theory and a large majority received instruction in the history of thought.

On the remaining elements of Bowen's prospective core, there was more variation. A little more than a fifth (22 percent) of programs had a course requirement in statistics. Among the 78 percent that did not, students rarely took it up on their own (7 percent of the time). Regarding economic history, 14 percent of programs had a history requirement but of the 86 percent that did not, students routinely took economic history in 22 percent of programs. The combined proportion is 33 percent – among PhD students in economics in the US in the early 1950s, one-third took some graduate instruction in economic history. As I discuss later, this is approximately twice as high as my best guess of the proportion today.

The circumstances just described set the stage for the cliometrics revolution. Cliometrics is an invented word combining Clio (for the Greek goddess of history) with metrics, meaning measurement. Conventional wisdom dates the revolution from professional meetings held in the late 1950s and various manifestos published at the time. The movement picked up steam in the 1960s and, allegedly, generated a full-blown takeover by the early 1970s, leading to a marked, immediate change in intellectual orientation within economic history towards economics, including regular and widespread use of formal models, econometrics, and hypothesis testing.

As with all conventional wisdom, there is some truth but the revolutionary aspect is overblown. Margo (2018a) conducted a text analysis of articles in five economics journals from the early 1950s through 2009. Two of these are economic history journals, the JEH and Explorations in Economic History (EEH), two are labor economics, Industrial and Labor Relations Review (ILRR) and the Journal of Human Resources (JHR); and one is a top-five, the AER. The JEH and ILRR were already in existence at the end of World War Two; EEH and JHR came into existence in the 1960s. My text analysis looks for instances of econometric language (e.g. regression) in empirical articles (defined as articles in which one or more tables appear). The comparison of economic history and labor economics is relevant because both were effectively non-quantitative in the early 1950s. Economic history had its cliometric revolution and labor economics, its human capital revolution. Which came quicker?

The AER was already routinely publishing articles with econometrics in the 1950s but there was essentially no econometrics in either the JEH or the ILRR during the decade. Then, in the 1960s, the content of ILRR begins to change and it rapidly converges on the AER in terms of usage of econometrics. The JHR, which began publishing during the second half of the 1960s, embraced modern empirical methods almost immediately. By intellectual standards, the human capital revolution in labor economics happened overnight.

By contrast, economic history was much slower to adopt modern econometric techniques. Change came faster in EEH but even so, it was not until the 1980s that the journal caught up with labor economics or the economics profession more broadly (for a similar conclusion, see also Cioni et al., 2020, 2021). The extent of econometrics in the JEH was always below EEH and the pace of change slower, although consistent with a weak revolutionary metaphor, the trend was definitely upward.

A deep dive into the JEH articles published in the 1960s that used any econometrics confirms the tepid pace of change evident in the formal text analysis. By deep dive I read every relevant article closely – quite a bit less work than it might be imagined, as, there were only 16 of them, or roughly 5 percent of the articles, notes, comments, and so on published in the JEH in the decade. It is not simply that there are few articles. Rather, the use of econometrics is defensive, as if authors feel they have to apologize to some portion of their audience. Results are discussed briefly and often only in footnotes.

The relatively slow pace of adoption of econometrics suggests that, while cliometrics might have been a revolutionary in a certain sense, its goals were not entirely consonant with the analogous revolution in labor economics. Margo (2018a) confirms this through an analysis of the publication histories of successive PhD cohorts of prominent economic historians. A remarkable feature of those who received their degrees in the late 1960s and early 1970s is that many did little or no publishing whatsoever outside of economic history journals in the first decade of their professional careers; included in this statement are such prominent economic historians as Claudia Goldin and Joel Mokyr. This is not the case with similarly prominent labor economists, who not only published in, say, the JHR or ILRR, but routinely published in top mainstream journals like the AER, QJE, or JPE.

What accounts for the different publication strategies? In Margo (2018a) I argue that one should view intellectual academic revolutions through the lens of an overlapping generations model of academic labor markets. New entrants to a field, even those who arrive at a time of intellectual formant, are evaluated for tenure and promotion by an older generation of scholars. New methods and new questions may diffuse fairly quickly or may not, depending on the equilibrium of the model.

The model suggests that there are good reasons to expect the cliometrics revolution to be slow. Although the initial demand for a new economic history may have originated in economics, the revolution quickly morphed into providing general reinterpretation of past work in the field informed by economic methods. In part, this reflected a reaction to the prevailing content of graduate courses in economic history in the early 1950s, as documented in the AEA report. Although a majority of graduate faculty felt that instruction in economic history was necessary, students were less favorable, complaining that the courses emphasized dates and places – a common criticism of history courses at the time – and made insufficient use of economic reasoning. Indeed, when the first generation of cliometricians took a careful look at the economic history literature they realized there were intellectual holes, often profound, wherever they looked. The pickings were ripe, not slim.

At the same time, however, the early cliometricians were not in a position to obliterate the traditional opposition, even if that had been their public goal (Diebolt and Haupert, 2018). Pointedly, it was not – rather, the idea was to supplement the traditional tools of the economic historian with the tools of economics. This required much more than lip service to professional norms. Early cliometricians were expected to learn foreign languages if they studied a non-English speaking country, read professional history journals, write books in addition to journal articles. Most important, they were to wear their economics lightly when engaging with professional historians –hence, the character and tone of the econometrics articles in the JEH in the 1960s.

Although the attempt to make traditional economic history more like economics is understandable in context, with the benefit of hindsight it is far from obvious that it deserved to go on for as long as it did. Ultimately, the majority of economists did not care very much if cliometricians were successful in proselytizing among traditional economic historians. Adherence to scholarly norms in history, such as writing monographs to gain tenure, ultimately had little or no import in economics. For a while, historians played along with quantification but began to lose interest, especially after the divisive controversies over economics scholarship on slavery (Fogel and Engerman, 1974) but also because, by the 1980s, articles and books written by economic historians in economics departments were too technical for the average historian.

The impulse in early cliometrics to make traditional economic history more scientific is usually associated with Robert Fogel although, as pointed out earlier, it was routinely accepted by other cliometricians. However, there was another school of thought that was also present. This school of thought saw economic history as a way to discipline economists in theorizing about development and growth.

The notion that economic history places significant limitations on economic modeling about growth and development is often associated with the other Nobelist, Douglass North, but it has a long intellectual tradition in the field. North was a card-carrying cliometrician in the early years of the revolution but by the late 1960s and 1970s his views had changed, evolving towards the work on institutions and institutional change for which he is remembered today. North certainly believed that economics had powerful tools, but he also believed that the prevailing models omitted first order causal factors, largely because economic theory lacked the tools to describe these adequately. He was also frequently critical of his fellow economic historians for playing too nice with the rest of economics. For a long time North was a voice crying in the wilderness. Some of this reflects that fact that disciplinary criticism is rarely successful in real time. But a deeper reason is that North was not able to come up with technical or econometric tools on his own that would have persuaded the rest of economics. Of course, economics did eventually buy into the idea that institutions matter and North's contributions were celebrated in his Nobel. But the requisite tools to truly further the research agenda did not come from North, his students, or elsewhere in economic history but rather from outside the field. These tools were dynamic games and the development and application of causal identification strategies from econometrics.

The two schools of thought carved out an intellectual pace in which economic historians with PhDs in economics could function professionally – teach, do research, publish, and gain tenure. All fields have such spaces but economic history was different. The publication records of successive cohorts of economics PhDs in their first ten years of professional life reveal these differences. Economic historians with economics PhDs who began in the 1970s were far more likely to write books, articles in economic history or interdisciplinary history journals than articles in economics journals. 1980s and 1990s PhD cohorts began to move towards economics but the changes were relatively modest (Margo, 2018a).

By the early to mid-1990s cliometrics was at a crossroads. On the one hand, cliometricians believed their revolution to be successful, especially after the awarding of the Nobel Prizes to Fogel and North. Cliometrics can also claim credit for stabilizing the long-run decline in share of economic history PhDs in economics, albeit at a very low level; according to estimates I have made from dissertations reported to the AEA, the economic history share of new PhDs in economics fluctuated between 1 and 2 percent from the mid-1960s to mid-1990s, On the other hand, and ominously, the field functioned largely in isolation from the rest of Economics. In one of his most remarkable articles, George Stigler (1995, published posthumously with Stephen J. Stigler and Claire Friedland as co-authors) measured the extent to which different sub-fields in economics cited each other. In contrast to other fields, economic history was essentially an island. It imported results from other fields to some extent – that is, cited articles in economics outside of economic history – but rarely exported its findings – economic history articles were cited rarely outside the field.

However, with the arrival of the post-2000 PhD cohorts arrive on the scene, there was the stirring of a revival. Evidence is found in a strong divergence in publication patterns (Margo, 2018a). Post-2000 cohorts are much less likely to publish books, and more likely to publish articles in economics journals outside of economic history and in top-five mainstream economics journals.

Why does the structural break happen? One straightforward explanation is cohort succession. By the early 2000s the earliest cliometricians were already leaving the research scene through life-cycle declines in productivity, retirement, and death. Even earlier, there were second or third generation PhDs who began to engage directly with other fields in economics. One of the most successful is Claudia Goldin, whose earliest work was totally within cliometrics but whose second book (Goldin, 1990) resonated strongly with a broader audience in labor economics. Additional examples include Barry Eichengreen, Christina Romer, Alan Taylor, among many others.

While I believe this role model story to be important, it is not the full story. Intellectual developments outside of economic history have dramatically raised the profile and acceptance of use of historical evidence in economic analysis.

There are three principal developments. The first and arguably most important is research on the (very) long roots of contemporary differences in per capita income around the world. The idea is that, in the distance past, events led to differences in institutions with very different growth potentials that, because of sunk costs, persisted for a very long time or even to the present. As mentioned earlier, this type of argument is quite alien from the point of view of traditional economic history or cliometrics but fits naturally into economics, creating a clear reason for studying the economic past for those not pre-disposed to do so (Cioni et al., 2021). The well-known and astonishingly well-cited papers by Daron Acemoglu and his various coauthors (e.g. Acemoglu et al., 2001) essentially jump-started this area of research some two decades ago, quickly went into orbit and continues to attract new researchers to this day.

The second and third are related to advances in modeling in growth theory and macroeconomics. Growth theory was re-invigorated by endogenous growth models that quite naturally invoke certain kinds of historical evidence (see Galor and Weil, 2000). Macroeconomists realized that dynamic general equilibrium models were a powerful tool for exploring long run changes in aggregate economic outcomes, such as the shift of labor out of agriculture, long run changes in fertility, and changes in labor force participation. By itself, the use of general equilibrium models in economic history was not new. But the dynamic models were far more sophisticated and, perhaps most importantly, internally consistent and, therefore, more persuasive as theoretical constructs, even as debates arose over particular applications.

What is perhaps equally critical is that the various economists leading the charge had little or no interest in persuading card-carrying economic historians by publishing their research in economic history journals. Rather, they sought to persuade economists by targeting and succeeding in publishing in top-five journals. In turn, this opened up space for historical argument in field journals outside of economic history (for example, in development or macroeconomics). The post-2000 cohort of PhDs in economic history were paying attention because they too began to craft articles intended for publication in top-five and non-economic history field journals – collectively, a much larger audience than just economic history (Margo, 2018a).

In thinking about these changes it is essential to keep in mind that it is not simply a matter of who does economic history and where economic history is published. The motivation of research in the field has also fundamentally changed. Certainly it is still possible to write a paper that is traditional cliometrics, on a topic that would not have seemed unusual in the 1970s or 1980s (see, for example, Atack and Margo, 2019). However, to a significant extent, papers being written today in economic history routinely follow the logic of modern empirical economics especially the emphasis on identification of causal parameters. This is true whether or not the paper is written originally for an economic history journal, or finds its way to an economic history journal after being submitted to a mainstream journal or a field journal outside of economic history.

1.4 Not your father's economic history

Economic history appears to be enjoying a professional rebirth. Compared with just a few decades ago scholarly identity in the field falls on a continuum. Well known economic historians are visible in other sub-fields of economics and there are many economists who, from time to time, pursue historical evidence or historical topics. Young scholars can work on research projects in economic history and be confident that their work will get a hearing. Overall, these are welcome changes. At the end of the day, there are only two types of empirical evidence in economics – experimental or observational. All observational data in economics sit in an historical context. A central feature of economic method is, when necessary, to abstract from historical context, often to argue in favor or against a particular theoretical model of behavior. The special gift of economic history to economics is the constant reminder that historical context is always present, either in the background or front and center. Broader recognition of the fundamental role of historical evidence and argument in economics can only improve and enrich the discipline overall.

At the same time, there are costs as well as benefits to the recent changes in the field. All fields in economics have specialized knowledge and, within fields, some economists are more specialized than others. For economic history, the issue is whether the specialized knowledge can be reproduced within other sub-fields or whether there is some benefit to keeping it in-house. There are two reasons why this issue is important. First, within economics, the number of new PhDs specializing in economic history in the United States is, once again, dwindling sharply. The AEA's records indicate that, between 2006 and 2015, the economic history share of new PhDs in economics in the United States averaged around 0.8 percent per year. In absolute terms, the number is tiny. For example, in the year 2015, 1,080 PhDs in economics were awarded in the United States. The AEA classified just 7 of these into economic history. Sustaining a field with just 7 new entrants per year seems impossible.

How did this happen, one might wonder, if economic history is increasing in popularity? The answer is that individuals who do historical work increasingly identify with the topical subfield, rather than economic history per se. A young scholar who writes articles on topics in historical labor economics, in other words, calls herself, first and foremost, a labor economist rather than an economic historian. This possibility was foreseen by Romer (1994) – instead of economic historians for whom the Great Depression was a topic to be mastered along with others in American economic history, there are macroeconomists who write articles on the economics of the 1930s, along with other topics in the field.

The second reason is related to the first. As I documented earlier in this paper, approximately a third of PhD students in economics took classes in economic history in the early 1950s. In Margo (2018b; see also Jaremski, 2020) I make an estimate of this number for today's graduate students. In the United States currently there are approximately 120 active PhD programs in economics. Just 12 of these have an economic history requirement (loosely defined) as part of the core and perhaps another 8 or so offer PhD courses in the field but do not have a formal requirement (my own department, Boston University, is an example). As a consequence, students in perhaps only one sixth (17 percent) of today PhD programs are at risk of ever taking a PhD course in economic history in their home department. Earlier in the paper I estimated that ca. 1950 one-third of economics PhD students took a graduate course in economic history as part of their training. Today's fraction, therefore, is half that at the mid-twentieth century. Taking a PhD-level course is neither necessary nor sufficient to ultimately doing professional level research in any field of economics but it would difficult to argue that, on average, the treatment effect of such a course is anything but positive.

As a thought experiment, imagine a counterfactual discipline in which economics is still divided into fields but economic history is not one of them. Instead, within other fields, there are economists who specialize in historical topics – public finance economists whose role is to study the economic history of taxation, macroeconomists who study the economic history of business cycles, and so on. In this counterfactual, there is a tradeoff between general and specific. On the assumption that, on average, specialization is a good thing, we might learn more about relatively few, but important (to economics) topics in economic history, as opposed to a little about a lot. It is not obvious which is better, but if we lose the economics take on the generalist perspective in economic history there are implications for scholarship, not necessary