Economically relevant human capital or multi-purpose consumption good? Book ownership in pre-modern Württemberg

Abstract

We investigate books as an indicator of human capital using extraordinary, individual-level data on book ownership and signature literacy for a population of German women and men between 1610 and 1900. Although book ownership was very high from an early date, it was associated with signature literacy, gender, urbanization, and wealth in ways inconsistent with its having registered economically relevant human capital. The books people owned were overwhelmingly religious, as elsewhere in pre-modern Europe. People consumed books for multifarious purposes, many of them non-economic. In this pre-modern economy, books were not a good indicator of economically relevant human capital for the population at large, which creates doubt about their use for this purpose more generally.

Introduction

Between c. 1500 and c. 1900, most European economies made the transition to sustained economic growth and also increased their overall levels of educational human capital.¹ There is some evidence of a relationship in this period between economic development and “upper-tail” human capital, that of people at the very top of the distribution of skills. But studies of overall human capital, relating to the population at large, have struggled to establish that increases in education over these four centuries were investments that caused economic growth, rather than forms of consumption that resulted from growth or from underlying sociopolitical changes associated with it (Mitch 2005; Ogilvie and Küpker 2015; A'Hearn, Delfino and Nuvolari 2019; Cvrček and Zajiček 2013, 2019a, 2019b).

A fundamental question that arises in the analysis of the historical relationship between overall human capital and economic growth is how to measure human capital in past societies. Signatures are the conventional historical indicator, since they were widely recorded for purposes unrelated to educational testing, were produced by individuals themselves instead of relying on subjective reports by others, and registered outcomes rather than inputs. However, the historical association between signatures and economic growth is uneven, due to the existence of many European societies, such as those in Scandinavia and wide swathes of German-speaking Central Europe, in which signature literacy was high and rising for centuries, while economic performance remained poor and stagnant. This has led some to advocate books as a superior measure of overall human capital (Van Zanden 2004, pp. 2, 22-3; Baten and Van Zanden 2008; Buringh and Van Zanden 2009). First, books are viewed as having transmitted “useful knowledge” of science, engineering, technology, and commerce required for productivity-enhancing innovations. Second, books are believed to demonstrate advanced literacy which enhanced people's broader economic capacities to form contracts or adopt new techniques. We follow the literature in using “economically relevant human capital” to refer to both of these mechanisms.

The existing literature focuses on the production of books rather than their consumption. In a sufficiently general framework, of course, books produced must have equalled books consumed. However, in early modern Europe, books produced in one place were often consumed in others. The largest producer of Spanish books in the 1550s was Antwerp in the Southern Netherlands (Wilkinson 2018, pp. 283-4). Dutch publishers produced half the books in Europe around 1700 and sold to consumers in Germany, France, Switzerland, England, Sweden, Denmark, Poland, the Baltic, and America (Pettegree and Der Weduwen 2019; Hoftijzer 2015; Goodfriend 2011). Seventeenth-century Dutch readers consumed books produced in Germany, Switzerland, France, and the Spanish Netherlands (Pettegree and Der Weduwen 2019, p. 270), and eighteenth-century French readers relied heavily on producers abroad who supplied them with the many books prohibited by the French state (Popkin 1984, pp. 443-4). One estimate suggests that 35% of all titles produced in the Dutch Republic in the seventeenth century were “directed towards the international market” (Rasterhoff 2017, p. 66). Focusing on the production of books as an indicator of human capital cannot, therefore, reveal where and how books were used, which is a key question in assessing their potential economic impact. To understand whether books can be regarded as an indicator of economically relevant human capital, it is necessary to complement information about book production with data on book consumption.

Another question concerns whether advanced, economically relevant literacy is better measured by books or by signatures. On the one hand, historians of education argue that reading was the more elementary skill, since schools in pre-modern Europe typically taught reading in the first two or three years, followed by writing after pupils could read words written by others; girls and poor boys usually attended school for shorter periods, and were thus able to read but not to sign their names (Schofield 1968, p. 325; Spufford 1979, pp. 408-9; Furet and Ozouf 1982, p. 167; Houston 1985, pp. 189-90; Graff 1987, p. 226; Schad 1997, pp. 129-80). On the other hand, scholars analyzing trends in book production in different parts of Europe hold that writing was the more elementary and formulaic skill, whereas reading measured advanced literate comprehension that could influence economic performance (Baten and Van Zanden 2008). In principle, both arguments could contain elements of truth: reading ability might register less advanced and less economically relevant literacy than writing at the lower range of the educational distribution (e.g. where the books consumed were simple and formulaic) but more advanced and more economically relevant literacy than writing in the upper part of the distribution (e.g. where the books consumed contained technical knowledge and difficult ideas). This paper seeks to resolve the puzzle using two distinct approaches. First, we analyze book consumption and signature literacy for the same population of individuals, exploring how book ownership was associated both with signing ability and with other individual characteristics. Second, we analyze the books which those individuals consumed, focusing on the quantitative prevalence of books that were religious (and thus normally more simple and formulaic) compared to those that were secular (and hence more likely to contain economically relevant information and ideas).

We make use of an extraordinarily rich compilation of micro-data on thousands of women and men across several centuries. Individual-level quantitative indicators in pre-industrial economies are rare, and our study breaks new ground in analyzing such data on the human capital indicators and other characteristics of more than 5,000 people, including women and villagers alongside the more frequently studied urban males. We focus on a zone in the poor and stagnant hinterland of central Europe, which was typical of much of the continent in the pre-industrial period and which needed to be transformed to unleash economic expansion. We cover not merely a cross-section or a brief growth phase, but rather nearly three centuries from 1610 to 1900, enabling us to illuminate human capital in a backward economy during most of the European “Little Divergence”. For each individual, we were able to determine both the number of books they owned and whether they could sign their name, making it possible to assess the relationship between the two indicators. Women make up over half our sample, enabling us to analyze economic agents whose human capital is central to modern growth theory but whose experience often eludes economic historians. Our data also record a number of socioeconomic attributes of each individual at the same life-cycle stage, making it possible to analyze the association between human capital indicators and other characteristics at the level of individual decision-makers. Finally, we are able to mobilize historical evidence to understand how books were used by economic agents in this economy across these three centuries.

The uneven historical association between economic growth and indicators of overall human capital has led a number of recent studies to emphasize the role in historical growth of “upper-tail” human capital – the capabilities of people at the very top of the distribution of skills (Mokyr 2005; Squicciarini and Voigtländer 2015; Cantoni, Dittmar, and Yuchtman 2018; van der Beek, Mokyr and Sarid 2019). In this paper we are concerned with books as a measure of human capital in the general population, and our evidence leads us to conclude that books owned by the population as a whole are not a good measure of human capital. This finding is wholly compatible with arguments such as those in Dittmar and Seabold (forthcoming), for instance, who find that publications of merchant manuals were significantly related to city growth in sixteenth-century Germany, and thus that a small number of specific types of book owned by a small sliver of the population measure economically relevant human capital. Books may well be a good measure of upper-tail human capital, but this does not imply that they are a good measure of economically relevant human capital in the population at large, the question investigated in this paper.