1. Understanding the formation of institutions

The view that institutions are a crucial determinant of economic and social performance has found broad acceptance in the social sciences, including in economics. Many studies have analysed the effect of institutions, and tried to capture the effects of institutional change, sometimes within short time frames, but also looking at long-run changes. The latter approach is currently reflected in a flourishing branch of economics and economic history, which aims to explain long-run divergencies in economic and social development between regions, countries or even parts of the globe through their institutional differences. A prime example is the debate on the diverging fortunes of the West and the East between the late Middle Ages and the 20th century. In the search for explanations of this Great Divergence, varying institutional explanations have been put forward, including differences in culture and values, legal organization and inheritance systems, political and public organization, household formation, market systems, and property rights (Brenner and Isett, Reference Brenner and Isett2002; Kuran, Reference Kuran2012; Landes, Reference Landes2006; Vries, Reference Vries2001).

Even though there is fierce and often productive debate about precisely which sets of institutions have enabled the favourable development of the West, and specifically of early industrializer England, many participants will agree that institutional factors have to form a vital component of any explanation of the long-run divergencies in development we observe between societies. The next step will be to better chart and explain the formation and the effects of these institutional arrangements over long time frames. Better measuring the effects – which is not the focus of this paper – will indeed be possible. Even though a direct measurement of transaction costs is not easy for historical periods with their often deficient source material, other ways to measure the quality of the institutional framework of markets have been pioneered. Among them is the reconstruction of the long-run evolution of interest rates, and of their variance between economies, in order to gauge the quality of property rights and the security they offer to actors in the capital market (North, Reference North1990: 69). Other indirect indicators used to this end, within the framework of a comparative analysis, are price variability, convergence of price movements, the skill premium and urban/rural wage differentials (examples for the pre-industrial period: Van Bavel et al., Reference van Bavel, Dijkman, Kuijpers and Zuijderduijn2012: 361–366; Van Zanden, Reference van Zanden2009: 18–31).

These indicators suggest that institutions indeed play a crucial role in the long-run performance of economy and society. This makes it urgent also to secure more insight into the formation of these institutions and the often diverse form they take; the issue that is central to this paper. It requires us not only to look at institutions as an independent variable – or as a given –, but also to examine them as a dependent variable. How can we explain the very different forms institutions take and the sharp differences we observe between periods and areas? This is where our insight is still limited. Arguably the best way to achieve progress, especially when we accept that institutional change is often a slow and path-dependent process, is by looking at history. Even if history does not allow for the control offered by field experiments or the laboratories of the natural sciences, this is the main laboratory we have available in the social sciences to analyse long-run societal developments, to test our hypotheses and to establish causal links in our search for better understanding fundamental changes, to which end we can use comparative analysis as a powerful tool (Diamond and Robinson, Reference Diamond, Robinson, Diamond and Robinson2010).

Entering the terrain of history and the formation of institutions, however, requires the economist to leave his or her comfort zone. Since the relevant institutions include social, political, cultural and religious as well as economic ones, economists need to leave their home territory and explore an area often left to political scientists or to cultural or political historians. These historians do not offer much help, however. At present, most of them choose a descriptive approach to history, often oriented to major military or political events, cultural highlights or prominent individuals. The influence of the Annales school and more structural approaches to history, which were dominant in the decades after the Second World War, has clearly waned. Social analysis at the macro level has become rare in historical studies, and the hope, or will, to analyse or reconstruct patterns and to identify underlying causes has also been abandoned, while post-modernism and post-structuralism have become popular (Stone, Reference Stone1991). In many respects, history has moved away from the social sciences. In view of the necessity to better understand how institutions take their specific forms, and the growing interest of economists in this issue, this is highly unfortunate, both for history and for economics, and for the social sciences and humanities more generally. It is only more recently that economic and social history has been picking up speed again, especially in developing a comparative and global perspective.Footnote ¹

Notwithstanding the current interest in institutional differences, explanations for the long-run changes of institutions, and the divergencies in these changes, are therefore not plentiful at present, especially as regards the institutions of the pre-industrial period, a period more difficult to penetrate, also because of the source problems. Traditionally, there are two contrasting approaches to the institutions of this early period, both of them more intuitive than informed by analysis. On the one hand, many historians are inclined to see the institutions of the pre-industrial period, and particularly those concerning the organization of politics, administration and public services, as instruments of exploitation by the elites, meant to keep ordinary people ignorant, to dominate them and to extract their surpluses. As opposed to this, there is an alternative traditional approach, found especially among economists but also among historians, that institutions develop more or less spontaneously because they provide a good answer to economic needs. This notion suggests that efficient economic and political institutions – ‘efficient’ being defined as contributing most, in a given set of circumstances, to the welfare of society – will gradually prevail over less efficient alternatives, as is believed to have happened in the more recent periods and especially in the West, either as a result of deliberate design or because they weed out the inefficient ones, by way of a kind of natural selection and competition between societies or firms.Footnote ²

These traditional approaches have recently become complemented, and superseded, by the work of several great economists, with Douglass North and Daron Acemoglu standing out among them (recent works: Acemoglu and Robinson, Reference Acemoglu and Robinson2012; North et al., Reference North, Wallis and Weingast2009). By linking economics to history from an institutionalist perspective, they have hugely enhanced our insight into the crucial importance of institutions, the diversity of institutional arrangements and their effects, and how they develop in long-run, often path-dependent processes. Their work has become a kind of standard in looking at the long-run development of institutions. Notwithstanding the brilliance of their work, for an economic and social historian focussing on the pre-industrial period there seem to be several problems with it. Most problematic are the notions, found there in an explicit or implicit way, that the Industrial Revolution has been a crucial watershed in history, that England – because it had this Industrial Revolution first – must have possessed the relevant, ‘right’ institutions, and that there has been a kind of unilinear development towards this situation. Also, in these works, changes in the institutional framework allowing for these developments are often portrayed as having been brought about from above, by political regime shifts, as most notably the Glorious Revolution in England in 1688, and to require a centralized state.Footnote ³ These notions at present exert a powerful influence in the field, and unfortunately so, as I will argue below.

At least in part, these notions result from the weakness of the interaction between the disciplines of economics and history. In the economists’ attempts to explain long-run changes in institutional arrangements, this weakness likely results in 1) unilinear reasoning and lack of historical depth; 2) a too narrow focus on the case of nation-state and early industrializer England; and 3) a too narrow focus on economic factors with a neglect of the social components that are harder to capture for non-historians. These possible drawbacks are reflected in the recent book by Douglas Allen, The Institutional Revolution (Reference Allen2012), that is central to this paper. I will refrain here from fully discussing the strengths and weaknesses of this book, since various reviews have already done so,Footnote ⁴ but rather use Allen's book to discuss these three pitfalls in the current attempts by economists to explain the pre-industrial institutions (section 2). Next, I will use the recent insights won by economic and social history, in order to see how these pitfalls can be avoided and further progress can be made. I will argue that this can be done by including the earlier periods into our analysis, by strengthening the comparative approach and extending it to various geographical levels, and by systematically integrating the social dimension into the analysis (section 3). In doing so, I will focus especially on the urbanized cores of pre-industrial Europe – the Low Countries and Northern Italy – in order to counterbalance the focus on the English experience and thus see how the experiences of these highly successful pre-industrial societies can broaden our perspective.

2. The pitfalls in Allen's The Institutional Revolution

In his compelling book The Institutional Revolution, economist Douglas Allen enters the difficult terrain of explaining the formation of institutions, loosely defined by him as the rules we live by and how life is organized. More specifically, Allen aims to explain the specific institutional arrangement of public services, including defence, administration, the judiciary, fiscality, and infrastructure, and he focuses on the case of England. He divides the history of public services in England into two parts: before the mid-19th century, when this sector was characterized by the selling of offices, patronage, aristocratic rule and other pre-modern phenomena, and after the mid-19th century, as professional, paid bureaucracies made their entrance. His explanation of this specific organization of public services, and its transition, rests on one variable: transaction costs, which he assumes in their turn to be mainly determined by technological progress. Allen argues that the very specific forms taken by this organization of public services in the period prior to the mid-19th century can be understood from the difficulties the state had in measuring the performance or ‘success rate’ of the people involved in these public services. This is still a characteristic of public services, but for this earlier period the effects of chance and the whims of nature were much greater, and as a result variability was very high. Moreover, the technical tools for measuring were weaker, so that performance could be controlled or monitored by the state to a much smaller extent or not at all. According to Allen, time measurement, for instance, was virtually impossible prior to the 19th century because of the absence of reliable clocks. As a result of these problems, the transaction costs were very high, too high to allow for a modern, professional bureaucracy to be used by the state. This only changed as the technological fruits of the Industrial Revolution reduced variability and made measurement possible.

In choosing this approach, Allen follows Ronald Coase, who attributes to transaction costs a determining role in the specific form institutions take. The societal goal in choosing certain institutions is to maximize wealth within that situation of transaction costs (Coase, Reference Coase1960). The result of this position (elaborated by Allen in note 25 on p. 230) is that another, non-societal force would be necessary, and sufficient, for a change in transaction costs and an ensuing innovation of these institutions, and this is technology. Allen thus distances himself somewhat from Douglass North (as also observed by Langlois, Reference Langlois2013), since North over the years has become more open to the idea that institutions are also linked to social organization and thus can remain even when technology or other changes would make it possible to innovate them.

Following Coase's argument, Allen argues that the solution to the impossibility of measuring, controlling and rewarding on the basis of performance in this period, and the resulting impossibility of employing a professional bureaucracy, was the choice of another system of public services, adapted to this situation of high transaction costs. One instrument in this system was patronage, in which reputation was primordial. Loss of reputation would mean loss of one's position, loss of the support of one's patron and loss of one's lucrative offices. Alternatively, the office was sold. The office-holder could then treat the office as his property, pass it on to his heirs and receive the revenues from the office instead of a wage. Selling an office has a self-selecting effect: only those who have the qualities to maximize the revenues from the office will pay the money to buy it. There is no need for the state to measure or monitor the actions of the office-holder, since this mechanism polices his behaviour in an indirect way. Still, there was also a risk that the actions of the office-holder, while maximizing his revenues, were incompatible with the goals of the state. Selling an office was therefore mainly used in cases where this risk was limited. Still, both with patronage and with sale, additional institutions were needed to generate loyalty and trust, and to safeguard the interests of the state. These institutions were obligatory for those who were qualified for an office, that is, the aristocrats, and they increased the penalty for misbehaviour and the resulting loss of the office. Among these rules and customs were otherwise incomprehensible ones, like the upkeep of big country houses in remote places and the big investments made in learning aristocratic, non-practical skills like dancing. They are all explained by Allen from this perspective, that is, as a kind of hostage capital that makes the costs or risks of default or underperformance for civil officers very high, and thus forces them to perform and to be loyal to the king/state.

Allen suggests this system existed in its clearest form in England from c. 1500 to c. 1850, where it was best-developed and where – it is at least suggested – it was most successful, militarily and economically, compared to other European nations. The end of this system came, Allen argues, after the Industrial Revolution. As a result of technological innovation in the period c. 1750–1850, the role of nature and luck was reduced, variability lessened, and performance could be better measured. Steamships and trains made the speed of transport more predictable, clocks made time measurable, and vaccination made life and diseases less variable. As a result, transaction costs went down. This ended the necessity for this pre-modern institutional arrangement of public services, and made maintaining this now unnecessary organization too costly, and it opened up the way for the modern, professional bureaucracy based on performance, directly employed by the state and paid in wages.

The big merit of Allen's book is that he tackles with a very clear hypothesis an issue which historians have mostly left to description. His story is highly coherent and compelling. He shows that historical arrangements are not strange phenomena we can marvel at, or laugh about, or the fruit of irrational people, but they developed in response to certain needs and they fulfilled a practical function. Allen does not limit himself to describing these institutions, but tries to explain them, and to understand how they develop over time, by way of a coherent framework of interpretation. Despite its merits, however, there are several weaknesses in the book. Some of them have already been noted by other reviewers, who pointed at the absence of any measurement of the supposed causes and effect in the book, for instance, and also at the absence of any substantiation of the supposed effect of this system on private prosperity and economic growth (McCloskey, Reference McCloskey2013: 364–366). I will not further discuss these issues here, but focus on three more general pitfalls, of which Allen's book to my mind forms a clear illustration.

First, there is the neglect of the social dimension. Allen assumes that the arrangement of public services is mainly, or even fully, driven by economic logic (transaction costs) and that transaction costs are determined by technology. In these assumptions, the social context is left out as a factor, and this is a serious omission. Even if these institutions can indeed fulfil economically useful roles, their formation and use is always linked to the societal position of those people and groups forming and using them.

This can be shown with regard to measurement, which plays a large part in the argument made by Allen. Measurement is not an objective act – as he seems to suggest –, and its development is not solely enabled and determined by technological advancement, but is also directly linked to social relations, authority or even dominance, as already noted by Witold Kula many decades ago (Kula, Reference Kula1986: 18–23 and 114–119). Kula, who is not even mentioned by Allen, shows how the major European waves in standardization of measures were all part of political unification attempts, such as by the Carolingian kings who ordered the standardization of coins, weights and measures, and later with the rulers of the Renaissance and the absolutist kings. Even if in some cases this standardization of measures was carried out also for economic reasons, or enabled by technological innovations, it was always and predominantly driven by political goals, that is, as an instrument to subordinate others, or to equalize them, and to determine what is measured, in what was an intensely contested process (also: Scott, Reference Scott1998: 24–33). This political impetus mostly decided whether technological and administrative means, that often had already been available for long, would be used to this end.

The socio-political dimension of measurement is also reflected, for instance, in the introduction of public clocks in the 14th century. These clocks, and the associated replacement of the unequal canonical hours by precise, regular hours, were employed particularly for measuring and regulating the working-hours and the times of selling in the market. They were thus introduced earliest in the towns where labour and commodity markets were important, in Northern Italy and the Low Countries (Dohrn van Rossum, Reference Dohrn van Rossum, Lepetit and Hoock1987; Landes, Reference Landes1983: 70–78). This process, however, was a contested one. Especially the big entrepreneurs pushed the town governments to install these clocks, have a bell rung at the beginning of the working day, and to force labourers to start work at these fixed times on penalty of a fine, as ordered in Aire-sur-Lys in Artois in 1355 (Le Goff, Reference Le Goff1963). Rebellious workers, on the other hand, often tried to silence the work bells, or at least to regulate them in order to safeguard their interests. Technological innovation, and the resulting decrease of transaction costs, can therefore not alone explain the diffusion and use of this new technology. This diffusion also required an economic rationale – the importance of markets and wage labour to be regulated – and a social one – in this case: the powerful position of entrepreneurs in local politics, allowing them to build and use these clocks. The process was, therefore, not a direct result of technology alone, but intrinsically tied up with the social and economic make-up of the localities in question and the incentives it offered to use this technology.

The second deficit of the book is the lack of historical depth and chronological precision, which results in incorrect assumptions about causality. Allen presses the relevant developments into a simple mold, allowing him to tell a simple, unilinear story. He divides history into two parts: pre-modern (16th to 19th centuries, with its pre-modern institutions) and modern (19th century and after). In passing, he mentions a preceding period, indicated with the label ‘feudalism’, which would have ended in England around 1500, but is not discussed in the book.

This chronology, with the rise of the pre-modern system of public services in the 16th century and its disappearance in the 19th, is not convincing. Not only is an analytical basis for the chronological starting point of the book lacking, but when we look more closely at the relevant historical literature it is clear that the proposed chronology is plainly wrong. This system of public services in many parts of Europe had developed several centuries earlier, and also, as I will elaborate below, disappeared several centuries earlier than suggested in the book. This is not nit-picking by an historian, or at least not solely so, but has direct implications for the proposed causality.

To start with, any explanations of the assumed rise of the pre-modern system of public services in the 16th century are lacking. There are some vague hints that in the military sector this rise is connected to changes in military technology at that time, ‘which increased the white noise and made the measurement of soldier inputs more difficult’ (p. 166). No further proof of this is offered, and for the other segments of the public sector this issue is not discussed at all. However, a look at the literature shows that many parts of Western Europe saw the rise of this system some three centuries earlier. The practice of purchasing positions in the army started in the 13th century. Also, mercenary armies are not a new phenomenon of the 16th or later centuries, but saw their rise in the 14th century, as documented for Northern Italy (e.g., Caferro, Reference Caferro1998). Likewise, the sale of public offices reached its high point in the 14th century, as in many parts of the Low Countries and the German Rhineland most high administrative offices were leased out, pawned or sold by the princes. They did so, as well as pawning of tolls, taxes and revenues, in order to raise the liquid capital they needed for wars and state building (Tewes, Reference Tewes1987; Kuys, Reference Kuys1994: 71–78 and 129–131). In many parts of Western Europe, the ‘classical age’ of this pre-modern institutional system was thus not the 16th to 19th centuries, but rather the 13th to 15th centuries, and the causes for its rise are therefore likely found in that and the immediately preceding period. I will come back to this issue in the next section.

An even more fundamental chronological problem for the argument made by Allen concerns the end, or dissolution, of the ‘pre-modern’ institutional system. He places this end in the 19th century, and he takes great pains to argue how it is brought about by the technological changes of the Industrial Revolution. But many of these changes, and the disappearance of this system, actually occurred many centuries earlier. For instance, contrary to what Allen asserts about the impossibility of telling the time and of measuring the length of the working day, we can see clear advances in hydraulic timekeeping and a growing emphasis on punctuality already in the 11th and 12th centuries. This trend was followed, especially in the towns, by a growing role of clocks and time in the 14th century, associated with the diffusion of public clocks, noted above (Landes, Reference Landes1983: 68–83). Even though clocks only slowly became more accurate, their loud and highly visible presence allowed the urban population direct access to fairly reliable time measurement. Allen bends the historical evidence when he says that people had to wait another 500 years, that is, up to the start of the 19th century, for mechanical clocks to become accurate, and that the real use of clocks only started once they were cheap enough to be privately owned. This is incorrect: the drive to quantify and measure, and the technical means to do so in a fairly accurate way, had their start in the Middle Ages (see also Mokyr and Espín-Sánchez, Reference Mokyr and Espín-Sánchez2013: 378–379).

This earlier chronology also applies to the monitoring of public servants. The rise of record-keeping, not really discussed by Allen, and the hiring of public servants selected at least in part for their abilities, skill and education, may both be considered features of ‘modern’ bureaucracies, but these developed from as early as the 12th/13th centuries on. The use of writing, book-keeping and making cross-records were exactly the main instruments used by the town and village populations to monitor the actions of administrators and to control their actions and avoid arbitrariness. It is not surprising that in the communal revolts in 13th-century Italy, northern France and the Low Countries record-keeping and access to, and inspection of, the records was a main demand of the rebellious populations and the rising guilds (Van Bavel, Reference van Bavel2010: 118–119).Footnote ⁵ In their demand for greater control of urban finances, they were sometimes supported, as in late 13th-century Flanders, by the princes or king, who had the same desire for better control, but also wanted to use this in order to break the rivalling power of the urban elites and nobility. This, again, shows that measurement is not only a technological issue but linked to political power and social relations (the point made above).

Similarly, social-political changes, and not technological ones, were instrumental in the disappearance of the pawning and selling of offices, taking place in the 15th century. The princes now could turn to the capital markets, which had developed, in order to raise the needed cash. They were stimulated in doing so also by the growing influence exerted by the representative bodies, the Estates, who wanted more control over the offices and disliked having them pawned out (Tewes, Reference Tewes1987). Both elements – the growth of the market and the growth of political representation – offer firm, alternative explanations of the disappearance of the pre-modern system, and they sealed the end of this system here in the 15th century, not in the 19th. This example, again, suggests that both the chronology (much earlier) and the causality (more driven by societal changes) are different from the ones proposed by Allen, and also that the big geographical variances in this process do not allow for one, unilinear story in which the Industrial Revolution plays the decisive role.

To be sure, Allen's reasoning does not hold in the light of the experiences of the present either. Even though technological innovation has proceeded and even further sped up ever since the 19th century, and may have allowed the reduction of transaction costs, the present situation in the provision of public services does not conform to his picture. Allen presents the fact that public services like the post, roads and policing in his period were sold and/or provided through the market as a distinctively pre-modern situation. He considers the full responsibility taken by the state for these services, and the dominant role of professional, state bureaucracies, as the rational outcome of the growth of technology and the reduction of transaction costs, and thus of modernity. However, since the 1980’s, the state and public offices in most western countries are rather retreating and give up many public services to semi-public or private parties again, in roads, postal services, water, electricity but even where public security or armed surveillance are concerned (Avant, Reference Avant2005). Policy makers and governments assume that the performance by these private parties is adequately measured through monitoring and competition through the market, and works better than providing these services by professional bureaucracies, state officials or workers directly employed by the state. This development, not even mentioned by Allen, takes place even while technology is much further advanced than in the 19th century. Better elaborating the historical dimension would easily show that Allen's unilinear and technology-driven story is not correct.

A third problem is in the geographical comparison, or rather its absence, which – again – impedes a proper search for underlying causes. The book essentializes the English experience, without placing it in context by way of comparison with other societies, while even a basic comparison would have shown there are fundamental problems with the mono-causal explanation of institutional formation offered here. For example, based on the English situation, Allen suggests that clearly scheduled, regular transport services would be impossible in his period. However, in the Dutch Republic already from the 15th century a large-scale, highly organized, reliable, regular and cheap transport system was developed, which was fully in place in the mid-17th century (De Vries, Reference de Vries1981). By that date, the problems arising from unpredictability of speed and cost of transport had largely been solved. As English travellers noted to their surprise, there were detailed, published schedules, strict measurement with hourglasses and high penalties for skippers failing to observe the schedule, all resulting in the barges actually operating on time. The statement made by Allen, therefore, is wrong, but even more relevant is the question why these regular transport services were possible in the Dutch Republic, but apparently not in contemporary England. Technology cannot be the sole answer. More important is the role of the fundamentally different organization of society, with a far bigger role played in Dutch economic life by the market, in combination with that of the local public authorities in developing infrastructure and regulation, already from the late Middle Ages onwards. Local authorities, using the financial markets to raise the capital needed for investment, and skippers, who were something between private entrepreneurs and public functionaries, together made this complex and advanced system work. The sector was managed and supervised by commissioners, who were not members of the elite but men chosen for their qualities and paid a regular wage (De Vries, Reference de Vries1981: 102–120), again in contrast to the picture sketched by Allen. Comparison would thus show that but the presence of active local authorities and dynamic markets for labour and capital, not technology, was the determining factor in the institutional organization used here.

Comparison would also disqualify the statements made about the outcome and effects of this institutional system. England from c. 1500 to c. 1850 is presented by Allen as the clearest example of this system, where it was best-developed and where it was most successful, militarily and economically. The link to economic success is also suggested by Barry Weingast and by Gary Libecap, in the backflap text, saying that Allen's book ‘increases understanding of how changes in measurement and institutions can increase trade and welfare’. Missing in the book, however, is any reference to the private sector, which would be essential to establish this causality (McCloskey, Reference McCloskey2013: 365–66). No evidence for a causal link between this institutional organization of public services and economic success is offered. Hard figures within a full, systematic comparison are not supplied. Such a geographical comparison would have made it hard to sustain any claims. What about the Netherlands/the Dutch Republic, for instance, which had a fundamentally different system of public services, but was militarily and economically at least as successful as England in the early modern period (De Vries and Van Der Woude, Reference de Vries and van der Woude1997)? And what about Northern Italy, which in the late Middle Ages had a system of public services similar to the Netherlands, organized by local authorities and bottom-up associations making ample use of factor markets, but in the early modern period became more similar to England, with a growing role of aristocrats, sale of offices and use of hostage capital (Jones, Reference Jones1997)? Italy, however, in this period was not at all successful and rather saw its economic leadership turn into backwardness (Malanima, Reference Malanima2011). And why did the pre-modern system last longer in England than on the European Continent? This applied, for instance, to the purchasing of positions in the army, a practice prohibited in France and Prussia already in the late-18th century, but persisting in England up to c. 1870, and to many other public services, as we have seen above. Was this, if we would be following the general argument of the author, because England lagged technologically? This would turn the argument upside down, in view of his attempt to link the English system to the precocious industrial development in that country. A brief comparison would already suffice to disqualify the causal links suggested here.

3. Toward a better theory of institutional change

The risk is that Allen's book, despite its merits, will leave a teleological, one-dimensional view of institutional changes, a view which lacks historical and geographical depth and is too detached from the societal context in which these institutions were formed and changed. Bringing the historical and social context more into the picture would help us to better understand whether, why and how these institutions changed, and why the geographical differences between societies in their institutional arrangements often persisted that long. I will discuss the three relevant elements (the historical, social and geographical dimensions) here in turn, making use of recent insights from economic and social history, and showing that the three are closely related to each other and that all three are needed to arrive at a full, multi-dimensional picture.

The first point, pertaining to historical depth, is that the development towards institutional ‘modernity’ was not a unilinear, recent one, with the Industrial Revolution as a crucial shift, and England as the only role-model, but one with various accelerations, and also stagnation and decline. Also, it started much earlier than often assumed in conventional economics. Both insights, which require us to rethink the possible causes of this process, can be demonstrated especially by looking at two other Western European cases, those of Northern Italy and the Low Countries. These were the two areas in Western Europe where changes were most pronounced in the 11th to 13th centuries, as the institutional organization of societies and their systems of exchange and allocation underwent drastic changes. Both areas saw a wave of revolts in this period, which largely broke feudal, aristocratic rule and opened the way for urban and rural communes and for political representation for broad layers of society. Under the pressure of these bottom-up movements, public offices were not sold or dominated by aristocrats anymore, and neither were they organized or dispensed by central government, as in the England described by Allen. Rather, public services were mainly provided by local and regional authorities, already from the 13th century on. Offices were held by merchants, representatives of the craft guilds, or substantial farmers, who acted as officers in courts, aldermen or members of water management boards (for northern Italy: Jones, Reference Jones1997: 370–386 and 403–419). If officials were hired, they were often employed by local or regional authorities, they might well be of non-aristocratic descent and were selected at least in part for their abilities. They were held accountable to the authorities and the representative bodies. The workers within these public sectors are, from the 14th century on, mostly wage labourers, often hired through contractors or entrepreneurs, as shown for infrastructural services within the town of Bruges or for water management works in the Holland countryside, for instance (Sosson, Reference Sosson1977: 167–178 and 189–201; Van Dam, Reference van Dam, Hoppenbrouwers and van Zanden2001). Crucial for this development, just as in the sector of passenger transport described in section 2 above, were the volume and the opportunities offered by labour and capital markets, combined with the active role played by local authorities.

Also, in the same period, there were big changes in the institutional organization of the private sector. Here, measurement and monitoring problems were often as extensive as in the public sector, for instance in long-distance trade, with its characteristic principal-agent issues and high variability caused by the vagaries of nature, but these were solved ever more effectively (Greif, Reference Greif2002). These solutions were initially often based on ethnicity, peer pressure and reputation, but from the 13th century on increasingly on registration, legal instruments and third-party enforcement by local and regional authorities, with a decisive role for the governments of the urban trade centres (Gelderblom, Reference Gelderblom2013). This period also saw the emergence of larger, more durable commercial organizations and partnerships.

Perhaps social and institutional changes were greatest in the exchange and allocation of land, labour and capital. Up to the 13th century this exchange and allocation was organized within families, kin, manors, communities and other non-market systems. But from this period onward, at least in some regions, markets started to take over this role. In contrast to the conventional view, sometimes based on an uncritical reading of Karl Polanyi's work, that factor markets are a modern phenomenon and became powerful only from the 19th century, studies in recent decades have demonstrated those in many Western European regions were already growing in the 13th century, not only in the towns but also in the countryside (a succinct statement: Hejeebu and McCloskey, Reference Hejeebu and McCloskey1999). The next step, taken in recent years, was to better chart and measure the volume and functioning of these markets. The resulting studies show that, by the 16th century, most of the agricultural land in Northern Italy and the Low Countries was regularly exchanged in the land market, with annual turnover rates at 1–3 %, which are not dissimilar to modern rates (Van Bavel, Reference van Bavel2010:178–181). Moreover, large shares of the land were leased out in competitive lease markets. In many regions, half to three-quarters of the land was leased out. Also, a substantial part of labour was performed for wages; in the 16th century this was between a quarter and half of all labour, with regional peaks up to 60 % (Van Bavel, Reference van Bavel2007). At that point, most of the land, labour, capital and commodities in these economies was exchanged and allocated through competitive markets. This also affected the provision of public services, which by now could rely on the liquidity offered by the capital market and the labour hired in the labour market.

The same medieval period also saw radical technological innovation and a rise of economic complexity, most conspicuously in the areas where political changes and the rise of open, accessible factor markets had been most pronounced. Both the institutional and the physical framework for trade in areas like Northern Italy and the Low Countries rapidly developed, including a massive wave of building of bridges, ferries, roads, from the 11th century onward. The effects of this medieval transport revolution on the speed, regularity and cost of transport, and the associated rise in commercialization of the economy, division of labour and regional specialization, were momentous. Moreover, the resulting Smithian growth was accompanied with a rise in labour productivity as the result of heavy investments in labour-saving technology and fixed capital goods, especially in industries (Van Bavel, Reference van Bavel2010: 343–371). Competition in commodity and factor markets, and credit available through capital markets, made it necessary and possible to implement technological innovations, especially aimed at reducing labour inputs, as can most clearly be observed in the late medieval Low Countries. These insights gleaned from recent research have discredited the older idea that ‘real’ technological progress and economic growth before the Industrial Revolution were impossible.

The effects of these structural changes of the economy are reflected by the recent calculations of GDP per capita. The old, Malthusian view, that assumed that before the Industrial Revolution societies were unable to generate economic growth and were stuck at subsistence level, was already discredited by the reconstructions of GDP per capita made by Maddison (Reference Maddison2003: 249; 2006: 244–265). These suggest that in the pre-industrial period Western European societies already achieved some growth. Around 1300, GDP per capita in Western Europe, according to Maddison's estimates, was about $ 600, that is, about 150 percent of the bare subsistence level, and this figure gradually rose to $ 800 around 1500, and to $ 1,000 around 1700. In recent years, Maddison's estimates for the pre-industrial period have been substantially revised upward. Around 1300 most Western European economies would rather seem to have been at GDP per capita levels of $ 800, and in Italy even higher, and many of them witnessed quite substantial growth in the following centuries, to levels of $ 1,000 – 1,500 in the 16th century (Broadberry et al., Reference Broadberry, Campbell, Klein, Overton and van Leeuwen2011: table 24). Growth was even more pronounced in Holland, the western part of the Netherlands, where GDP per capita rose from $ 800–900 around 1300, to $ 1,500 around 1500 and no less than $ 2,500 around 1600 (Van Zanden and Van Leeuwen, Reference Pomeranz2012).Footnote ⁶ The rise of total factor productivity in these cases was sustained and substantial. These calculations thus push back the start of ‘modern’ economic growth in Western Europe by several centuries, long before the Glorious Revolution and the Industrial Revolution (van Zanden, Reference van Zanden2009: 232–260, thus contradicting Pomeranz, Reference Pomeranz2000, who presents the Industrial Revolution as the decisive turning point). Also, we can now see better that this process was not uniform and unilinear, but a localized process that displayed sharp regional differences and alternations in economic success. The economic leader of the 13th and 14th centuries, Italy, for instance, had become a laggard in the 18th century (Malanima, Reference Malanima2011).

The picture becomes even more variegated, and the idea of a unilinear, recent development towards modernity even further discredited, by broadening our view to developments outside Europe, as economic and social history has increasingly done over the past few years. Many of the elements we consider part of institutional ‘modernity’ already existed much earlier in history there, and in some pockets of development were found already in Antiquity. This applies to the provision of public services by big bureaucracies, staffed by professionals remunerated with money salaries and using written records extensively, but perhaps even more conspicuously for the market as a main system for exchange and allocation, not only of goods and products, but also of land, labour and capital.

This picture contrasts with the noted absence or weakness of factor markets outside Western Europe in the early modern period, as observed for areas like India and Southeast Asia, Africa and the Ottoman empire, but also for highly-developed societies like China and Japan.Footnote ⁷ A more optimistic picture of factor markets in early modern China is offered by Kenneth Pomeranz, who stresses the favourable institutional framework of these markets (Pomeranz, Reference Pomeranz2000: 69–82; Pomeranz, Reference Pomeranz2008). And indeed, in China, and perhaps even more in Japan, market exchange grew in importance in the early modern period (for Japan: Saito, Reference Saito2009), but the pace of this development was much slower than in Western Europe. Factor markets outside Europe remained weak and small almost universally in the early modern period, up to the 19th century.

This may be true, but still recent years have seen a change of perspective, thanks to the latest research on much earlier periods. A case in point is Iraq in the neo-Babylonian period. The long 6th century BC, which is relatively well-documented, showed strong demographic growth and urbanization, monetization of exchange and the growth of land and especially labour markets, with an associated rise of cash crop production, and of regional and occupational specialization (Jursa, Reference Jursa2014). Another relevant case is Iraq in the 7th to 10th centuries, or the early Islamic period. Sales of land were widespread then, and leasing and later sharecropping leasing also became important. In the towns, but to a lesser extent also in the countryside, wage labour was ubiquitous (van Bavel et al., Reference van Bavel, Campopiano and Dijkman2014). The rural and especially the urban economy in this period was highly monetized and market-oriented, also induced by highly dynamic credit and capital markets. Perhaps the Lower Yangtse delta in China in the period between c. 1000 and 1400 is a similar case, as all kinds of restrictions on market activities were removed and factor markets flourished (Shiba, Reference Shiba1969: 127–130). Relevant to the chronological issue under consideration here is also that in all these cases after a few centuries these developments were reversed. In Iraq, for instance, after the 9th century the factor markets increasingly became skewed, and they shrank, elements of unfreedom became stronger and professional bureaucracies and fiscal systems made way for the sale of offices and the leasing out of tax-rights. This reverse development, again, makes it impossible to see the rise of professional bureaucracies and markets as a unilinear or irreversible process. Rather than assuming a path towards modernity, each institutional formation should be carefully placed in its historical context, while allowing for diversity and regression. It also shows that in institutional development there is not a single underlying cause, and certainly it is not technology, and the Industrial Revolution thus does not form the single watershed.

The second dimension needed is the societal one. Institutional arrangements, even when having a clear economic function, are fully embedded in society as a whole, and often fulfil multiple functions, of which the economic one does not need to be the one that wholly determines the form of that institution. In this sense, the recent book by North et al. (Reference North, Wallis and Weingast2009) is more comprehensive than Coase’s, Allen's or North's earlier work, by bringing the whole of society into the picture. This enables them to include the effects of societal changes on institutional arrangements and to link institutions to the whole logic of society. They discern two varieties: the natural or limited access societies and the open access societies, with the latter ones in essence being the modern, Western ones, which they assume to have developed over the past two centuries. This is a more fruitful, holistic approach to institutional formation, although it possesses a teleological undertone and their chronology places the relevant changes too late in time (see the point made above).

Further disentangling the role of societal formation, and analysing institutions as the outcome of contention between social groups, would be a next step (Acemoglu et al., Reference Acemoglu, Johnson and Robinson2005: esp. 427–439; Ogilvie, Reference Ogilvie2007: 662–665). Such analyses will likely show, and have shown, that, even though some institutions were in part shaped in response to economic challenges, they were also formed by social bargaining, and dictated by the interests and preferences of persons and social groups, because of their distributional effects. In order to explain institutional change, we thus have to analyse what social groups are involved in the formation and use of the institutions under study, and what their position is in society, as defined by their grip on resources, revenues and political leverage. This can be done most effectively through an empirical test within a comparative framework. An example is the recent investigation of the differences in the institutions regulating the accessibility of common wastelands in early modern Northwestern Europe. These rules, in part informal ones, and their use, are shown to have been shaped according to the different degrees of equity of bargaining power within the rural communities in question. Where a broader group of stakeholders, and a bigger variety, was involved in political decision-making, these institutions were more inclusive (De Keyzer, Reference de Keyzer2013).

As this example shows, the relevant processes were not only found within the pinnacles of centralized states, and relevant groups were not always operating from above, as seems to be suggested in the accounts by Acemoglu and North, but also at the local and regional level and operating from below. The most conspicuous examples for these bottom-up processes in the pre-industrial period are found in the towns, as in Northern Italy, starting in the 10th century, and later in the Low Countries, where merchants and middling groups successfully revolted against aristocratic rule (Violante, Reference Violante1974: 173 ff.; De Moor, Reference De Moor2008). One of the main aims of the insurgents was the removal of all feudal arbitrariness, and changes thus often went along with the introduction of new forms of administration, financial control and accountability (Jones, Reference Jones1997:103–120). Likewise, the merchants from the 12th century proved highly successful in developing their own institutional framework for long-distance trade and solving principal-agent problems, resolving conflicts and increasing contract enforceability, often in concert with local authorities (Gelderblom, Reference Gelderblom2013: 102–140). In the towns of Italy and the Low Countries, where these merchants often became politically dominant, their rules and methods were also used in the organization of the political domain and public services, which in its turn further sped up economic growth, as analyzed for the trade metropolis of Genoa around 1200 (Greif, Reference Greif1994).

This intertwining of social and institutional structures displayed geographical patterns with sharp differences. For instance, in the late medieval period there was a stark contrast between Central England, with a dominance of aristocrats, who held by far most of the land – the most important factor of production - and Holland, where a free peasantry with full control over the land predominated, and Northern Italy, with its powerful urban elites and the political grip they held over the surrounding countryside. These societies offered radically different incentives, opportunities and obstacles, for instance, for the organization of markets for land, labour and capital and thus for their development, volume and degree of dynamism, with the absence of lordly restrictions in Holland allowing for a high degree of openness and accessibility and a high level of integration of factor markets there (Van Bavel et al., Reference van Bavel, Dijkman, Kuijpers and Zuijderduijn2012). This social context, and the development of markets it allowed, or disallowed, in its turn also determined the options available for the institutional organization of public services.

Including the social dimension also helps to better understand the seemingly inefficient nature of some institutions, or their lack of change, even if more efficient, ‘modern’ alternatives were available. The peasants in the sandy, infertile Dutch region of Drenthe who succeeded up to the 19th century in defending their common rights and communal forms of agriculture against privatization and the switch to absolute, exclusive property in private hands, and against the associated commercialization of agricultural land use, were not irrational or ignorant people who failed to recognize the economic growth to be generated by this switch. Rather, the peasant holders of family-sized holdings, who were the dominant segment of this society, clearly understood that they would be the ones who would lose out in this switch (van Zanden, Reference van Zanden1999). The technology to make this switch economically worthwhile had long been available in many cases, but its precise timing and location depended on the power balances in the region in question, or the shifts in these, such as the growing power of large landholders who were able to push through the division of the commons. In order to understand institutional changes, or the lack thereof, and the big regional differences in the timing and form of institutional changes, that occurred even if available technology was more or less similar, we therefore have to include an assessment of the structural characteristics of that society, and most notably of the distribution of property and decision-making power within that society.

A third point to be included in the research into the formation of institutions is the awareness of the sharp geographical diversity in these developments. Connected to the unilinear views of history is the idea of convergence, where different institutions or societies will sooner or later converge towards the better, or more rational and efficient ones. These are mostly assumed to be the ‘modern’ institutions we know in the Western societies (a view implicit in the work by Douglass North, Douglas Allen and many others). However, the recent studies in economic and social history show us that divergence is as prominent as convergence. Even under growing political and economic integration and interaction, the divergences between societies and their institutional arrangements can persist or even increase. This is related to the previous point. Since societies and the make-up of groups within them are very different, their choices, goals and interests are also different, resulting in different or even divergent paths of development.

In pre-industrial history, these divergent paths can most clearly be observed at the regional level. I have tried to demonstrate this by way of a long-run, comparative analysis of the social and economic development of the twenty or so regions in the Low Countries, from c. 500 to 1600, used as a testing ground to this end (van Bavel, Reference van Bavel2010). Contrary to what one would expect in more unilinear accounts of history, the sharp differences between these regions were not eradicated by the late medieval emergence of markets, industrialization, and political unification. Rather, the extent to which regions responded to growing markets, for instance, varied according to the social context. Small- or medium-scale peasant producers, such as those in Drenthe, who possessed viable, alternative systems of exchange and allocation, such as associations, commons and families, were often less inclined to enter the market, as observed above. As opposed to this, tenant farmers in regions dominated by leased out large landholdings, such as the Dutch river area, were compelled to orient themselves towards the market and to commercialize their production, because they needed to maximize profits in order to remain successful in obtaining land in the highly competitive lease markets.

Growing inter-regional trade even sharpened the regional distinctness in economic development. Trade, by integrating regions within one economic market system, rather stimulated regional divergence, with each region using its comparative advantages, in part geographical and in part social-institutional ones. The growing complementarity of regions did not flatten out institutional differences, except in a few sectors. In long-distance trade, for instance, with a high mobility of traders, flexibility of trade streams and numerous potential centres of trade, there was institutional convergence (Gelderblom, Reference Gelderblom2013: 200–203). On the other hand, institutional arrangements in most spheres, including the organization of landownership, leasing, labour markets, marriage and family, and public services, instead showed persisting or sometimes even widening differences, even in the face of growing economic integration.

The medieval history of the Low Countries is a clear example of how great the degree of path dependency in socio-institutional structures at the regional level was. As such, these structures also determined the paths of industrialization in the early modern period. Even as trade and technology improved, and capital flows and trade in raw materials grew, industrial development still remained a regional phenomenon and specializations assumed a regional character. This can be observed most clearly in the process of proto-industrialization, which was tied in with the formulation and distribution of rights to land, with the prevailing household patterns and with the social-institutional make-up of the region more generally, and this even applied to the ensuing Industrial Revolution (Hudson, Reference Hudson and Hudson1989). The region thus remains the most relevant geographical unit for investigating and understanding economic and social history, and the formation of institutions, perhaps into the 19th century.

From the 16th century onwards, however, and even more clearly with the formation of nation-states in the 19th century, social-institutional differences at the national level came ever more clearly to the fore. Differences in societal organisation between England, the Netherlands/Dutch Republic and northern Italy were striking, as observed above, and can be assumed to have contributed greatly to the economic divergence between these areas in the early modern and modern periods (for this economic divergence: Allen, Reference Allen2001). This observation, in combination with the more obvious political role of the nation-state in the modern era, has made the national level the favoured arena for investigating path-dependent developments (as reflected, for instance, in Acemoglu et al., Reference Acemoglu, Johnson and Robinson2005). Still, this focus on the nation-state may contribute to a teleological view and lead to anachronistic reasoning. For long periods in history, and even in the modern period, the region is probably more relevant as the unit of analysis. Strikingly, and perhaps also induced by the present, weakening role of nation-states within Europe, in economic geography the region as the unit in investigating developmental paths is fully back again, as reflected by a growing number of regional studies (Henning et al., Reference Henning, Stam and Wenting2013; an early example: Storper, Reference Storper1997). Progress is also made in the other direction, looking at units at a bigger geographical scale. It is becoming clear how aggregating socio-institutional characteristics for even larger areas can offer striking insights and help us better understand long-run divergencies at the global level, such as those between China and Western Europe, as in the above-mentioned Great Divergence debate, or between North America and Latin America (Acemoglu and Robinson, Reference Acemoglu and Robinson2012: 7–44). Future comparative analysis should thus encompass three geographical levels: the regional, the national and the global, without favouring one because at present it seems to be the dominant one.

Comparative analysis and a multidimensional, not economically deterministic view, combined with an open eye for long-run patterns, will help us in the coming years to better understand and explain the formation and evolution of institutions. Along these lines, and using the insights of recent economic and social history, we could thus retain some of the strengths of Allen's book – most notably the clarity of its hypothesis, the clear conceptual framework and the boldness of its approach – while making the research less teleological and one-dimensional, more analytical and more testable. Thus, we would be able to use the full potential of history as the laboratory of the social sciences, including economics.