State formation

Jørgen Møller - State Formation, Regime Change, and Economic Development (Routledge, 2017)

Representative institutions redux

In recent decades, the ‘Why Europe?’ question has attracted overwhelming interest in the social sciences. From having been a dusty, forgotten issue, primarily studied by historical sociologists, it is now a part of mainstream political science and economics. For example, over the last decade, the leading journal in political science, American Political Science Review, has published a series of articles that, often with the help of sophisticated statistical methods, address different aspects of the question (e.g., Acemoglu and Robinson 2006; Stasavage 2010; 2014; Hariri 2012; Woodberry 2012; Blaydes and Chaney 2013; Kokkonen and Sundell 2014). Similarly, the question has received attention from economists, an issue to which we return in Chapter 13.

It is thus hardly an exaggeration to say that much of contemporary social science – as in the late nineteenth and early twentieth centuries – revolves around the causes of the emergence of modernity in the West and the implications of this for the rest of the world. This new interest in old questions primarily seems to have been prompted by the following insight: only a historical perspective can enable us to understand the contemporary variation in economic prosperity and levels of democracy, particularly outside of Europe and the European settler colonies (e.g., Acemoglu et al. 2001; 2002a; Acemoglu et al. 2008; Hariri 2012). The more specific claim is that a series of medieval political institutions contributed to putting a leash on European monarchs and limiting their arbitrary exercise of power. These institutions were subsequently transplanted to some of the European colonies, where they facilitated economic growth as well as democratization.

What kinds of institutions have been attributed these tantalizing effects? As we shall see in Chapter 13, which addresses economic and political development outside of Europe, the answer to this question is not always clear. But it should be obvious that the representative institutions of the Middle Ages – Estates or parliaments and diets – form part of the core. Here, we can repeat Ertman’s (1997, 19) observation from Chapter 2 that these institutions alone were able to limit the monarchy’s exercise of power in a systematic manner.

The following pages are premised on this point. The purpose is to address a new body of literature on the origins and character of representative institutions. I have already reviewed Ertman’s (ultimately, Otto Hintze’s) account of why medieval representative institutions were stronger in some areas than others, and why they therefore had a varying impact on state formation and regime change. However, new research has been burgeoning in recent years. Most importantly, English political scientist David Stasavage has followed up on Hintze’s and Ertman’s analyses of the character and impact of the representative institutions in medieval Europe. Stasavage’s work has thus far resulted in two books (2008; 2011) and four scientific articles (2007; 2010; 2014; 2016). This chapter discusses this work and sets the stage for the next chapters, which use this medieval legacy to explain contemporary variations in economic and political development.

Surveying representative institutions

To explain a phenomenon, one first has to capture it. With this in mind, Stasavage compiles a dataset tallying representative institutions in twenty-four European states in the five centuries between 1250 and 1800 – the period to which Myers (1975) refers as the Age of the Polity of Estates. Instead of distinguishing between the number of chambers and whether or not the representatives were representing estate groups or localities, as Hintze and Ertman do, Stasavage (2010) maps the following:

(1) Does a representative assembly exist?
(2) Does the representative assembly have a veto on taxes?
(3) Does the representative assembly audit government spending?

The first condition is fulfilled if a collective assembly is found on the national level, convenes with some regularity, and at a minimum is consulted by the monarch. We find assemblies of this kind throughout Western (or Latin) Christendom in the Middle Ages. They emerge at different points in time: relatively early in Western Europe and relatively late in Scandinavia and East-Central Europe (Poland and Hungary). The only states in Stasavage’s dataset that do not live up to the first condition at any time in the period 1250–1800 are the duchies of Milan and Tuscany and, hardly surprisingly, Russia, which is the only state outside of Western Christendom that is tallied.

The second condition is more demanding as it requires that the assemblies have a veto on taxation. This was the core prerogative of most representative bodies, and this condition is fulfilled in the vast majority of the twenty-four states. In fact, among the states with representative institutions, Denmark and Naples are the only ones without. 1 The third condition is met if the representative institutions had a direct right to audit the monarch’s expenditure and possibly even decide over public spending. There are very few cases in which we find this prerogative (eight of Stasavage’s twenty-four states). In addition to city-states such as Siena and Florence, England after the Glorious Revolution in 1688 is an instance. Finally, Stasavage codes how frequently the representative institutions convene, ‘annually’ being the highest value and ‘never’ the lowest. 2

Geographic barriers for representation

What explains the variation captured in Table 10.1? And what explains why the frequency of assemblies was so different across this universe: from the city-states, where the assemblies met many times annually, via the almost annual meetings in states such as Württemberg, Austria, and England, to the extremely rare gatherings in states such as Denmark and France?

Stasavage’s attempt to answer this question begins with an apparent paradox. As we know from previous chapters, it is widely assumed that representative institutions made it easier to impose and collect taxes for the purpose of financing warfare, and they have also been seen as easing public borrowing and promoting economic growth. If only some of these postulates are correct, then why did the representative institutions not win out throughout Europe? And why were many states so slow to introduce them? These are the questions raised by Stasavage (2010) in the article ‘When Distance Mattered’.

The title hints at his answer. Boiled down to a single sentence, there were significant geographical barriers to representation. A myriad of researchers have linked democracy with the size of the political unit. This point is best illustrated by the direct democracies of antiquity, which required that all (male) citizens were able to participate in the popular assembly. But scholars have also been pointed out that many island states have been able to maintain democracy in the period following World War II, an observation that has been linked to their limited size, which has made it easier to create a sense of political community (Dahl and Tufte 1974).

The representative institutions of the Middle Ages were not particularly democratic (see Møller and Skaaning 2013, Chapter 4). However, Stasavage argues, this does not alter the fact that geographical barriers had at least as much significance for how they worked. In Europe of the High Middle Ages, traveling great distances was associated with exorbitant costs – measured in time as well as money. The network of Roman roads had fallen into disrepair in the Early Middle Ages. In fact, it was mostly the memory of them that remained by the year 1200, at which time the representative institutions were emerging. Some improvements were made around this time, but it was only really after 1800 that large-scale advances of European infrastructure began to occur. In other words, transport was extremely cumbersome throughout the period Stasavage analyses. 3

Stasavage argues that this affected the impact of representative institutions on public borrowing. Here, he invokes economic theories about corporate finance. One of the key insights of these theories is that a common condition for being able to raise new funds for investments is that the borrower accepts external control or at least monitoring. However, such control can be so costly that potential investors pull back for this reason alone (2010, 625–626). Stasavage draws an analogical inference about the medieval representative institutions, the point being that the cities’ representatives would only go along with raising funds for the monarch if they were allowed to monitor how he spent them, but that such ex-post control could easily become too costly if geographical barriers prevented the representatives from meeting with some regularity. 4

Stasavage accordingly reasons that a state’s geographical size will have an impact on (i) the existence of representative institutions, (ii) how frequently they were called, and (iii) whether or not they had the right to oversee public spending. Conversely, he does not expect to find a relationship between geographical size and a veto on taxation, as this does not require a high meeting frequency. Stasavage applies these expectations in a series of statistical analyses in which he supplements the information from the dataset above with a number of control factors, including indicators measuring the presence of external threats and the size of the population, respectively. The intuition behind the first control variable is that the threat of war can trigger the summoning of representative institutions – and for that matter bring about an expansion of the prerogatives of these assemblies. The reasoning behind the second control variable is that the per capita expenses related to warfare will be greater in areas where the population is smaller. Finally, Stasavage controls for urbanization by rerunning his analysis without the city-states that are included in the dataset – on the basis of the potential objection that the representative institutions in these states were peculiar.

Stasavage measures geographic barriers in several different ways, including the average distance any representative would have to travel – ‘as the crow flies’ – in order to participate in an assembly. More specifically, he tests his expectations in two different ways. The first analysis operates on the ‘national’ level – that is, across the national units described in Table 10.1 above. As a next step, Stasavage shifts the level of analysis to the regional level by repeating his analysis on the French regions that had separate representative institutions. That is, Stasavage investigates whether there were also geographical barriers to representation within France.

What do the analyses show? Stasavage finds that geographical size has a very consistent, statistically significant, and substantial effect: first, in terms of whether or not a representative assembly existed; second, whether it exercised oversight over public spending; and third and finally, on how frequently it met. For example, the assembles in the quartile of the smallest states in the dataset met on average more than once every second year, while the assemblies of the corresponding quartile of the largest states met less than once every third year. Conversely, as expected, there is no correlation between geographical barriers and whether or not the representatives have a veto on taxation (2010, 636–637).

Stasavage uses this baseline to predict how often the French regional Estates were convened. The model provides a reasonably good prediction of the meeting frequency in eleven of the thirteen regions. The two exceptions are Normandy and Brittany, where the regional Estates convened surprisingly often considering the large size of these units. Finally, Stasavage repeats his national analysis but substitutes the French regions for France. The results are again robust, which is reassuring in as much as it is rather artificial to work with ‘national’ units in the medieval world. France is thus not the only country with regional Estates. The same applies to the Holy Roman Empire, which had a single representative assembly (the Imperial Diet – Reichstag), but where Stasavage has instead coded the assemblies in the individual states (Landtage).

A final objection is that, over time, the political institutions might have an impact on how large a given state is (or, rather, becomes). If a certain kind of representative institution made it easier to borrow or charge money for warfare, this should make it possible to swallow up neighbouring states that did not have the same advantage – as money is the fuel of warfare. Conversely, one might imagine that authorities that are not accountable to a representative assembly would be able to engage in ambitious foreign policy more easily, which might provide opportunities to increase the size of the state in question. In both cases, this would undermine Stasavage’s test of the relationship between geographical barriers and the characteristics of the political institutions. However, Stasavage also takes this into account in his test and dismisses this objection.

Representative institutions and public borrowing