The Political Economy of Status Competition: Sumptuary Laws in Preindustrial Europe

Abstract

Sumptuary laws that regulated clothing based on social status were an important part of the political economy of premodern states. We introduce a model that captures the notion that consumption by ordinary citizens poses a status threat to ruling elites. Our model predicts a non-monotonic effect of income—sumptuary legislation initially increases with income, but then falls as income increases further. The initial rise is more likely for states with less extractive institutions, whose ruling elites face a greater status threat from the rising commercial class. We test these predictions using a new dataset of country and city-level sumptuary laws.

It is unfortunately an established fact that both men and womenfolk have, in utterly irresponsible manner, driven extravagance in dress and new styles to such shameful and wanton extremes that the different classes are barely to be known apart.

—Nuremberg Ordnance of 1657, Quoted in Hunt (1996)

Sumptuary laws regulated clothing based on social status. They first emerged in Europe in the twelfth and thirteenth centuries, during the Commercial Revolution—a period of growing trade and commercialization. They were most prominent in the richest and most economically sophisticated parts of Europe. Their heyday coincided with the era of elevated per capita incomes and real wages that followed the Black Death.

Historians have extensively studied how sumptuary laws relate to patterns of consumption (see, especially, Perrot 1994; Frick 2002; Welch 2005; Stuard 2006; Muzzarelli 2009). The existence, proliferation, and disappearance of these laws directly connect to topics such as the emergence of fashion and the culture, politics, and legal systems of the Renaissance more generally; and the consumer and industrious revolutions (see Brewer and Porter 1993; Vries 2008). Historians have also investigated how sumptuary laws reinforced distinct gender norms (see, particularly, Brundage 1987b; Jackson 2010; Moran 2018; Cavallar and Kirschner 2020).

Our focus is on how the sumptuary laws were used to regulate status competition. In particular, sumptuary laws are of interest as a manifestation of the political economy of premodern states. Our argument then is an application of what Ogilvie (2007) terms the “conflict” view of institutions. Long-standing institutions, such as sumptuary ordinances, persisted not because of any efficiency properties but because they benefited well-organized and powerful elites. Sumptuary laws could act as a form of “social disciplining” (Ogilvie 2006)—targeting urban consumers and frequently women.

Sumptuary legislation can also be understood in the context of a wider set of laws, “contra laboratores” (studied by Cohn 2007). Like sumptuary laws, these proliferated after the Black Death and were a response to the economic rise of the lower classes. However, rather than intervening in labor markets, sumptuary laws focused on consumption.

Recent work in economics establishes that people will invest considerable resources in status markers, with significant consequences for inequality (Charles, Hurst, and Roussanov 2009) and savings and growth (Omer and Neeman 2012). More so than today, clothes were important markers of status. Before the Industrial Revolution, clothes were expensive, durable goods. After subsistence needs had been met, they often took up the largest single proportion of household income. Expenditures could be astonishingly high, especially for the individually designed, fitted, and ornate clothing worn by aristocrats. Fra Bernardino (1380–1444) claimed that Florentine women might spend “more than six, more than eighteen, more even than sixty” florins on a pair of shoes (quoted in O’Malley 2010, p. 61). As the annual earnings of a skilled worker were around 37 florins, this was a huge sum (Goldthwaite 1990, p. 348).¹ Frick (2002, p. 180) notes: “Inventories of personal possessions showed an inordinate portion of their total wealth was spent on clothes; 40 percent … was not unheard of.”

Clothes played a vital role in expressing identity: “Aristocrats sought social differentiation through dress codes and elaborate spending because social status depended not only on luxurious cloth but on how cloth was fashioned into garments that followed precise, often individual guidelines” (Rosenthal 2009, p. 460). Clothing was a visible manifestation of the social order.

To explain why clothes increasingly became the subject of legislation and why these regulations eventually disappeared, we combine newly collected data and qualitative evidence to formulate a general explanation of the main drivers of sumptuary legislation. We employ a formal model to derive testable predictions.

Our starting point is the view expressed by North, Wallis, and Weingast (2009) that premodern polities are best analyzed as being governed by coalitions of elites who were able to generate economic rents by limiting competition and access to markets. Seen through this lens, sumptuary laws restricted the consumption of citizens, benefiting elites at the expense of the rest of society. We consider two classes: elites and non-elites. In our historical applications, the identity of these classes varies based on the setting. The non-elites, in general, refer to the class just below the elites (though in some cases, such as the Italian city-states, this could include affluent artisans and craftsmen). In England, sumptuary laws distinguished members of the nobility (and paid attention to gradations within the nobility) from merchants and artisans. In Italian cities-states, the elite were initially nobles with feudal titles and landholdings outside the city but over time they came to include established mercantile families who held public office (Martines 1979). The non-elites in this case referred to the popolo, who included richer artisans and craftsmen.

We build on the idea that the utility from status goods is relative to the status good consumption of competing social classes.² The key idea is that status good consumption by non-elites poses a status threat to ruling elites. To maximize their “status distance” from non-elites, elites can either raise their own status good consumption or pass laws that limit the status good consumption of non-elites. We show that as incomes rise, non-elites can afford to consume status goods, but may initially find evading sumptuary laws too costly. This, then, makes enforcement worthwhile for the elites and induces them to enact sumptuary laws. As incomes rise further, however, non-elites have even more resources to evade sumptuary legislation, which increases the cost of enforcing such laws. Ruling elites then prefer to use their income to buy more status goods, rather than reduce the status good consumption of non-elites.

Our model predicts that the scale of rent-seeking of ruling elites affects sumptuary legislation and modifies the relationship between sumptuary laws and income. Specifically, we hypothesize that the relationship between sumptuary legislation and income was stronger where elites were constrained and could extract only a moderate amount of rent. In this case, the elites’ disposable income, and therefore their ability to consume status goods, was more limited, making the status threat from below more binding.

As an initial descriptive exercise, we find evidence of a non-monotonic relationship between sumptuary legislation and income at the country level. Countries with very low levels of GDP per capita were unlikely to pass sumptuary laws, as were countries with high levels of GDP per capita. It was at intermediate levels of income that we observed the highest frequency of sumptuary legislation.

To deepen our understanding of this phenomenon, we use another, city-level, dataset.³ Studying city-level sumptuary laws allows us to leverage a source of exogenous variation in per capita income: local outbreaks of the bubonic plague following the initial Black Death outbreak (1348–1352). As medieval European economies were Malthusian, plague outbreaks generated upward wage pressure and higher incomes per capita (Galor 2005; Ashraf and Galor 2011; Voigtländer and Voth 2013; Fochesato 2018; Jedwab, Johnson, and Koyama 2022, 2024). Following the Black Death, episodes of bubonic plague continued to reoccur at the local level. Florence was hit by plague in 1430, 1437–38, 1449–50, and 1478–79 (Morrison, Kirshner, and Molho 1985). These plagues had an impact on labor supply; those of 1478–79 recorded a mortality of 20–25 percent.

We test whether, at intermediate levels of income, increases in per capita income raise the probability of elites enacting a sumptuary law. We find that cities that were affected by the plague in the prior two decades were more likely to enact a sumptuary law in the next period. This relationship between plagues and sumptuary laws is strongest in states where rent-seeking is likely to be moderate, as our theory predicts.

A concern is that the initial Black Death outbreak (1347–1352) had a widespread impact on culture and society. However, the subsequent outbreaks of plague had little discernible impact on culture or institutions. For example, plague outbreaks are uncorrelated with changes in political regimes (a finding confirmed for German cities by Dittmar and Meisenzahl 2020).⁴ For robustness, we exclude: (i) the years surrounding the initial outbreak; (ii) the entire period 1347–1400; and (iii) the cities that were most severely affected by the Black Death.

An extensive literature has studied sumptuary laws. This includes older monographs focused on England (Hooper 1915; Baldwin 1926); Venice (Newett 1902); and Nuremberg (Greenfield 2018). Freudenberger (1963) and Miskimin (1964) discussed these laws in the context of wider economic developments. Since the 1980s, as attention focused on issues of consumption, fashion, gender, and the law, new scholarship has come to the fore. Recent studies have documented the extent to which sumptuary laws were updated and enforced (e.g., Kovesi 1993, 2016; Frick 2002; Killerby 2002; Pittman 2021). This literature has also drawn attention to highly gendered aspects of the laws.

Renaissance Italian sumptuary laws have attracted a lot of scholarship (e.g., Hughes 1983; Brundage 1987b; Killerby 2002; Muzzarelli 2002, 2018, 2019, 2022). Scholars of Italian sumptuary laws have paid particular attention to the incentives of policymakers and to issues of enforcement (e.g., Kovesi 1993; Mola and Riello 2019). Sumptuary laws have also been studied in the Ottoman Empire (Quataert 1997; Dunn 2011), Eastern Europe (Kizik 2010), and ancient Rome (Zanda 2011).

In contrast to this voluminous historical literature, less research on sumptuary legislation has been conducted by economic historians. One exception is Ogilvie (2010), who studies the enforcement of sumptuary legislation in early modern Germany. Drawing on archival evidence from Wildberg, she finds that violations were frequently sanctioned and that sumptuary laws had a real bite.⁵

SUMPTUARY LEGISLATION IN HISTORY

First, let us consider the definition of sumptuary laws, the reasons behind them, and their enforcement. Sumptuary laws regulated dress and extravagant expenditure and were common across pre-modern societies, particularly in societies where dress was an important signifier of status, as in Republican Rome (see Online Appendix C) and in late medieval Europe.

These laws typically forbade members of the lower orders from wearing luxurious, more expensive items of clothing and also specified what rulers and elites could exclusively wear. They proliferated in the Italian city-states of the twelfth century during a period of increased trade and prosperity. Ostracized groups, such as Jews, were typically forced to wear special garb (Roth 1928). They were common in both China and Japan (Shively 1964). Harte (1976) described them as an “identifying characteristic” of the pre-industrial world.

Why Pass Sumptuary Legislation?

Sumptuary laws varied across time and place. But they shared common features. They sought to control consumption; they targeted specific classes of individuals, designating some clothes or expenditures prohibited for one class, but permitted for their social betters.⁶ In agrarian societies, this meant the landed nobility. In medieval cities, the elites who benefited from these might be merchants (see Hughes 1983; Brundage 1987a; Killerby 2002). Sumptuary laws were thus distinct from both luxury taxes and the mercantilist encouragement of domestic industry.

There has been widespread disagreement about the motivation for sumptuary legislation. The proclamations that accompanied sumptuary legislation justified these laws in terms of the common good (Howell 2010, pp. 234–36). These justifications could be religious—for example, an ordinance in Palermo in 1420 stated that “All that is immoderate, excessively vain and overly lavish is prohibited and censured by Holy Writ and judged ever to be noxious to the Republic” (quoted in Bridgeman 2000, p. 212). Rulers also defended sumptuary legislation because it would reduce crime; because money spent on luxuries could not be used to pay taxes; to reduce budget deficits; and because luxury was a source of immorality and decadence.⁷

Nonetheless, despite pious or altruistic rhetoric, these laws generated private benefits for elites. Rulers were active in crafting sumptuary legislation, and they were supported by members of the political elite. Alfonso X of Castile, Edward IV, Henry VIII, and Elizabeth I of England were personally involved in designing sumptuary codes.⁸ In the late fourteenth century, English MPs lobbied for the reinstatement and vigorous enforcement of sumptuary laws, suggesting “a strong desire on the part of at least one vocal segment of society to see these controls enacted and enforced” (Doda 2014, p. 177).

We build on the work of historians, who have linked sumptuary laws to a desire to maintain status distinctions. Killerby (2002, p. 7) argues that sumptuary laws were imposed because it was felt that “luxury could be used to erode class distinctions.” For Freudenberger (1963, p. 37), the “power of money unquestionably posed a threat to the established privilege.” Sumptuary laws benefited existing elites; they “worked to bolster or define existing hierarchies so that visual signs matched the rank of the wearer” (Lemire 2018, p. 88). According to one scholar: “One of the most important means through which rising classes displayed their claims to higher status was outward appearance, especially clothing … The threat of changes to the prevailing social hierarchy prompted responses from the nobility and the monarchy in the attempt to regulate displays of false status” (Shulman 2007, p. 73).

Sumptuary legislation often contained explicit provisions for enforcement.⁹ Fines were the most common, but not the only punishment. Fines could be complemented with confiscation of the goods in question, loss of employment, excommunication, or even corporal punishment and, in rare cases, imprisonment.¹⁰ In Milan, imprisonment was used against tailors who sold forbidden goods (Kovesi 2019, p. 191). In Venice, shoemakers who made shoes with a higher platform than specified in the law were subjected to both a fine and imprisonment. In Italian cities, violations of sumptuary laws were often handled by the podestà. In Florence, the tailors who made the offending garment were also fined, indeed at a higher rate than the owner of the good (Frick 2002, p. 182). Sumptuary laws restrict the size and extravagance of weddings and baptisms, requiring that notification be given days earlier with a list of invited guests. French sumptuary laws encouraged denouncements by members of the public who received part of the fine as a fee for informing (Moyer 1996, p. 252). Male heads of household were often responsible for their wives and daughters. In Florence, the fines could be deducted from dowries. Tudor sumptuary legislation was enforced by the Star Chamber as well as local courts. Early modern French sumptuary laws required notaries to report violations, and police officers or judges who failed to enforce the law could also be punished (Moyer 1996, p. 255).

Enforcement could be costly. Cities “appointed special prosecutors” to investigate sumptuary violations (Hughes 1983, p. 96). Special boards were also established for this purpose, although sometimes, as in Florence in 1427, citizens could not be found who would serve on them, so other government officials had to be reassigned (Frick 2002, p. 186). In Elizabethan London, individuals—who had to be men of substance—were appointed in each ward to police sumptuary offenses, but this number proved insufficient and soon had to be increased (Pittman 2021, pp. 80–81). Enforcement was always patchy, given the limited state capacity of premodern states (see Wilson 2016). The difficulty of enforcement, however, did not mean that sumptuary laws were unimportant; recent research concurs that “to simply dismiss the laws as ineffective would be incorrect” (Mola and Riello 2019, p. 239).

Guilds were involved in the enforcement of sumptuary laws. Legislation in Faenza in 1410 specified that guilds were to “place the sumptuary statute among their own guild statutes and have a member of their guild announce the law in the assembly at least every six months ‘in a loud voice so that everyone can understand it well”’ (Killerby 2002, p. 206). In Florence, in 1406, to improve the enforcement of sumptuary laws, a committee of four citizens was established to enforce sumptuary laws, and at least one of these had “to be a member of one of the lesser guilds” (Rainey 1985, p. 435). In 1433, these measures were strengthened so that at least five members elected to the committee to enforce sumptuary laws were to be members of the seven major guilds (Killerby 2002, p. 216).

Guilds may have been further motivated to lobby for sumptuary laws that favored their own products at the expense of imported luxuries. Guilds certainly lobbied for protection from foreign imports (Ogilvie 2019, p. 346). Hunt (1996, p. 212) notes that in England, the Merchant Taylors regulated the dress of their members and “actually committed members to jail for excess apparel.”¹¹

Detailed evidence for enforcement comes from Ogilvie (2010). For the community of Wildberg, she obtained a unique source, a single surviving booklet of “‘Polizei-Tax-Rechnungen,’ which lists all the clothing fines levied over a 12-month period between February 1713 and February 1714 … Over this 12-month period, 110 individuals in a community of only about 1,300 inhabitants were fined for wearing forbidden garments, most of them small items or calico” (Ogilvie 2010, p. 308).¹²

The Data

We construct two new datasets of sumptuary legislation: a country-level dataset for all of Europe and a city-level dataset for late medieval Italy.

The country-level dataset contains information on 1,181 sumptuary laws across Europe, spanning the medieval and early modern periods. We use this dataset to track the number of sumptuary laws at the country level from 1100 to 1850. We collected information by building on the work of numerous historians who have worked on specific countries or cities. While the resulting listing does not capture the universe of sumptuary laws, we believe it is representative; it is certainly the largest list of sumptuary laws available. Important sources include Hooper (1915) and Baldwin (1926) for England; Moyer (1996) for France; Guarinos (1788) and Wunder (2019) for Spain; as well as numerous local and city-level sources for Germany, the Low Countries, and other parts of Europe. The city-level dataset compiles sumptuary laws in medieval and Renaissance Italy, which were originally collected by Killerby (2002). Details of data construction are provided in Online Appendix 1.

The information we have typically only lists the passing of a sumptuary law and not the content of each law. Nor does it distinguish between laws on new products or laws that update or renew existing legislation. We view it as measuring sumptuary legislation on the extensive rather than the intensive margin.

Figure 1 depicts the rise and fall of sumptuary legislation in France, England, the Holy Roman Empire, Italy, the Low Countries, and Spain.¹³ Across Europe, sumptuary legislation followed similar patterns. They first appeared in the rapidly developing and commercializing Italian city-states. The earliest medieval sumptuary law was passed in Genoa in 1157. The number of laws increased dramatically in the late Middle Ages and continued to be high in the sixteenth and seventeenth centuries before declining.¹⁴ In the seventeenth century, mercantilistic laws proliferated. We discuss their relationship to sumptuary laws in Online Appendix 2.

Figure 1 SUMPTUARY LEGISLATION OVER TIME

Data Sources: See text and Online Appendix 1.

Next, we propose a model that sheds light on the proliferation and decline of sumptuary legislation and derive testable predictions.

A MODEL OF SUMPTUARY LEGISLATION

In our model, ruling elites E and non-elites B consume ordinary goods, and goods, that is, clothing, that express status. However, the elites can limit the status good consumption of non-elites by enacting sumptuary laws. Elites incur fixed costs of enforcing the law, as well as variable costs that increase with the value of status goods that non-elites consume above the regulated limit. Elites allocate their disposable income between their consumption of status goods and ordinary goods and the enforcement of sumptuary legislation. Meanwhile, if non-elites flout the law, they face a penalty if caught, which increases with their consumption. Non-elites allocate income between their consumption of status and ordinary goods, as well as the costs of evading the law if they disobey it.

We abstract from the process by which total income Y is produced and only assume that Y can be below or above a threshold level $$\bar Y$$ . This threshold captures the level of income at which elites and non-elites start valuing a particular good as one that conveys status.

Production is exogenous. Income is above or below a threshold level $$\left( {\bar Y} \right)$$ . In keeping with our understanding of the institutions of premodern Europe, the elite appropriate a share τ = [0, 1] of income Y. We refer to this as a rent. Depending on context, this could refer to income from office holding and funded by taxes or to actual land rents (in which case it is effectively paid by wage laborers). It determines the distribution of income between elites and non-elites. The elites’ disposable income is thus: y _E = τY. While non-elites disposable income is y _B = (1 – τ)Y.

Both elites and non-elites derive utility from the consumption of ordinary good x and, after some threshold level of income, from the consumption of status good l. Elites, however, are threatened by the status good consumption of non-elites and thus seek to maximize their “status distance’ (l _E − l _B) from non-elites. The payoffs for elites and non-elites are:

(1) $${U_E} = \left\{ {\begin{array}{*{20}{c}} {u({x_E},({l_E} - {l_B}))}{{\rm{if}}\,{y_E} \ge \tau \bar Y} \\ {u({x_E})}{{\rm{if}}\,{y_E} < \tau \bar Y} \\ \end{array}} \right.$$

(2) $${U_B} = \left\{ {\begin{array}{*{20}{c}} {u({x_B},{l_B})}{{\rm{if}}\,{y_B} \ge (1 - \tau )\bar Y} \\ {u({x_B})}{{\rm{if}}\,{y_B} < (1 - \tau )\bar Y} \\ \end{array}} \right.$$

These utility functions are very general and depict simply the notion of status distance à la Akerlof (1997).¹⁵ The only restriction is that preferences are non-homothetic above the threshold.¹⁶

We analyze the following game S.¹⁷

1. 1. Nature decides whether income reaches a threshold . With probability . In turn, Y determines the disposable income y _i of player i {E, B}, where E denotes the elite and B the non-elites (the number of each is normalized to 1).

2. 2. Elites E choose to enact a sumptuary law (η = 1) or not (η = 0).

3. 3. If η = 0, then, simultaneously, each player i allocates their disposable income y _i over bundle (x _i, l _i), where denotes ordinary goods and status goods. Let l _i be priced at ρ, while x _i is the numeraire.

4. 4. If η = 1, B chooses whether to obey the law (ω = 1) or not (ω = 0). If ω = 0, E incurs fixed cost C _F and variable cost C for enforcing the law. If ω = 1, E incurs only fixed cost C _F. Let denote the maximum amount of status goods that E permits B to consume. Then C _F and C are decreasing in —the more lenient the sumptuary law, the lower the costs of enforcement. Variable cost C is also increasing in B’s status good consumption l _B, as the more status goods non-elites consume, the more goods have to be inspected for infractions of the law.

   If ω = 0, B is caught violating the law with probability χ, and incurs fine F if caught. Denoting the odds-ratio of being caught as $$\theta = \frac{\chi }{{1 - \chi }},$$ thus incurs expected cost of evasion θF, which decreases with the leniency $${\bar l_B}$$ of the sumptuary law, and increases with B’s status good consumption l _B.

5. 5. Simultaneously, B chooses bundle (x _B, l _B), while E chooses $$({x_E},{l_E},{\bar l_B}).$$

Whether a sumptuary law is enacted (η) depends on income (Y), and whether non-elites evade the law (ω) depends on Y and on η. The consumption choices of both elites and non-elites ((x _i, l _i)), in turn, depend on Y, η, ω, on the choices of the other group (x _−i, l _−i) and, if η = 1, on the level of status goods proscribed for non-elites $$({\bar l_B})$$ which, in turn, depends on Y, ω, and (x _i, l _i).

Thus, a strategy profile for the ruling elites E is $${\sigma ^E} = \{ \eta ,\,\,({x_E},\,\,{l_E},\,\,{\bar l_B})\},$$ while a strategy profile for non-elites B is σ ^B = {ω, (x _B, l _B)}. Let Σⁱ be the set of all such strategy profiles for player $$i \in \{ E,B\} ,\sum { = \sum {^E} \times \sum {^B}},$$ and the payoff to player i. Game S is thus defined as .

The game is solved by backward induction. In Steps 3 and 5, player i maximizes her utility subject to her budget constraint; i spends their disposable income on ordinary goods, priced at 1, and status goods, priced at ρ; E spends some of their income on enforcement if a sumptuary law is enacted; while B spends on evasion if they disobey the law.

Denote x _i,case, l _i,case as the optimal values of (x _i, l _i) chosen by player i, the optimal value of chosen by E, and U _i,case the indirect utility function obtained by I under case = {a, b, c, d, e, f}, where:

1. a: income below threshold, no sumptuary law enacted:

2. b: income at/above threshold, no sumptuary law enacted:

3. c: income below threshold, sumptuary law enacted and obeyed:

4. c: income at/above threshold, sumptuary law enacted and obeyed:

5. d: e : income below threshold, sumptuary law enacted and disobeyed:

6. f: f : income at/above threshold, sumptuary law enacted and disobeyed:

We thus have the following result.

Proposition 1 Game S has a unique equilibrium in which elites enact a sumptuary law (η = 1) if U _E,f ≥ U _E,b, or if U _E,f < U _E,b and U _B,d ≥ U _B,f ; otherwise, if U _E,f < U _E,b and U _B,d < U _B,f, they do not enact it (η = 0).

Proof See Online Appendix 3.

We now use our model to inform our empirical analysis. We are interested in the effect of income Y and the extent τ of elite rent-seeking on the probability that elites enact a sumptuary law.

Proposition 1 states that the decision η of elites to enact a sumptuary law depends on the relative values of U _E,f, U _E,b, U _B,d, U _B,f. These, in turn, depend on the exogenous variables in the model, notably income Y and the extent τ of rent-seeking. For a given population of units, where each unit has particular values of Y, τ, and, therefore, U _E,f, U _E,b, U _B,d, U _B,f, the probability that a unit in the population passes a sumptuary law is:

(3) $${Pr}(\eta = 1) = \upsilon + (1 - \upsilon )w,$$

where G is a cumulative distribution function, v ≡ G[(U _E,f − U _E,b)] denotes the probability that (U _E,f − U _E,b) ≥ 0, and w ≡ G[(U _B,d − U _B,f)] the probability that (U _B,d − U _B,f) ≥ 0.

The effect of income on this probability is obtained by taking the derivative of Equation (3) with respect to Y. Online Appendix 3 shows that there exists a threshold level of income at and below which Pr(η = 1) is increasing with income and above which it is decreasing with income; that is, a non-monotonic inverted-U relationship between Pr(η = 1) and Y:

Hypothesis 1: Inverted-U-Shape Relationship The probability that a sumptuary law is enacted rises with income at lower levels of income, but eventually falls at higher levels of income.

What drives this non-monotonic, inverted-U, relationship? As Equation (3) implies, the hypothesized relationship between Pr(η = 1) and Y is a result of the change in v and in w with respect to income Y.

Consider the following illustration of how the comparative statics of v and w can generate non-monotonicity.¹⁸ Initially, an increase in Y generates effective enforcement of sumptuary laws—non-elites can consume more status good consumption but cannot afford to consume more than what would be prescribed by elites, which induces elites to enact such laws. As income further increases, however, non-elites are able to afford much higher levels of status good consumption. On one hand, if elites did not enact a law and non-elites had free rein, their status distance from non-elites could decrease. On the other hand, if they enacted a law, non-elites could now afford to disobey it. This, then, would make enforcement ineffective—to enact the law would be to incur enforcement costs without preventing the erosion of status distance. Thus, as income continues to increase, there comes a point at which elites cease sumptuary legislation.

Hypothesis 2: Rent-Seeking The non-monotonic, inverted-U relationship between income and the probability that elites enact a sumptuary law is more likely to hold when the ability of elites to extract rents is lower.

Online Appendix 3 provides a formal discussion, but the result is intuitive. When elites appropriate less from total income, there is less income inequality between elites and non-elites. This intensifies status competition. As incomes rise, non-elites are better able to close their status distance from elites by consuming more status goods, which then prompts the latter to limit such consumption through sumptuary laws. This means, though, that elites incur enforcement costs, which leaves less income for their own status good consumption. As incomes continue to rise, non-elites buy even more status goods, even to the point of defying sumptuary laws. This further increases enforcement costs and leaves much less income for elites’ status good consumption. At some point, enforcement becomes so ineffective and costly that elites cease enacting sumptuary laws.

We also consider an extension in which the fines collected from violators of sumptuary laws are a source of revenues for the elites that can offset the costs of enforcing the laws (see Online Appendix F.1 for the formal analysis). This changes the budget constraint of the elites, increasing their disposable income by the amount of fines F(·) whenever non-elites violate sumptuary laws, and generates the following result.

FINES DEFRAY ENFORCEMENT COSTS

When sumptuary laws are enforced through fines that offset the costs of enforcement, this moderates the inverted u-shape relationship between income per capita and sumptuary legislation.

Muzzarelli (2019) provides several examples of how sumptuary laws could be used as fiscal devices.¹⁹ Bridgeman (2000) also provides examples where sumptuary law fines had similar properties to luxury taxes.

We find that if the fines imposed on violators of sumptuary laws were significant and offset, to a degree, the costs of enforcement, this weakens the relationship between income and sumptuary legislation but does not, however, overturn our main results.

COMMERCIAL ELITES

Finally, our benchmark model captures status competition between an elite class who derive incomes from either landownership or from their control of the state and the middle class. However, in some Renaissance city-states, notably Florence, power was eventually captured by individuals whose wealth came from trade and banking. To explore how the rise of such a new type of elite might affect status competition, we consider an extension of the model where we allow the elite to “produce” their own income (see Online Appendix F.2). We show that if the new set of elites have a much larger share of income, then status competition is less intense, as elites can easily outspend non-elites. Sumptuary legislation is then less responsive to changes in income.

EMPIRICAL ANALYSIS

Our country-level panel data provides suggestive evidence that there is an inverted-U relationship between the probability that a sumptuary law is enacted and income. Next, we investigate whether a similar relationship holds in a panel of Italian city-states. This dataset includes more than 300 sumptuary laws collected by Killerby (2002), spans the period between 1100 and 1500, and covers more than 40 cities. The drawback, however, is that Killerby collected sumptuary laws only until 1500, preventing us from analyzing the subsequent decline of sumptuary legislation in these city-states. Thus, when using this data, we are only able to test the relationship between the probability of enacting a sumptuary law and income at lower levels of income, which, according to Hypothesis 1, should be positive. We also test whether this positive relationship at lower levels of income is more likely to hold when elites are less rent-seeking²⁰ (Hypothesis 2).

Inverted-U Relationship Between Sumptuary Laws and Income: Country-Level Analysis

Our country-level data provide rich information on the timing of sumptuary legislation at the country-level. We match this information to estimates of historical income per capita from Bolt and Zanden (2020).²¹

Historical estimates of GDP are noisy and subject to revision. The underlying data quality varies by country. A particular challenge is that GDP estimates are systematically missing for earlier historical periods. For example, there are no GDP estimates for pre-1500 Germany. The absence of medieval GDP in many countries reduces our sample significantly. Due to these limitations, we first provide a qualitative exploration of the relationship between income and sumptuary legislation.

We code each datapoint (country/century) as corresponding to a high- or low-tension sumptuary regime, where, according to our preferred definition, high-tension regimes have > 5 sumptuary laws per century. We then assign data points based on whether their estimated income levels were low, medium, or high.²² The resulting categorization in Figure 2 is highly suggestive and motivates a more formal analysis.

Figure 2 HIGH INTENSITY SUMPTUARY LAW REGIMES BY COUNTRY-CENTURY PAIRS

Sources: See text and Online Appendix 1.

Figures 3a and 3b report the bivariate relationship between income and sumptuary legislation. Both figures suggest the existence of an inverted u-shaped relationship between per capita income and the number of sumptuary laws passed.

Figure 3 POOLED COUNTRY-LEVEL ANALYSIS: SUMPTUARY LAWS AND PER CAPITA GDP

Sources: See text and Online Appendix 1.

In our regression analysis, we use data at the country-century level and the country-50-year level. Specifically, we estimate:

(4) $${\rm{Sumptuary}}\,{\rm{La}}{{\rm{w}}_{i,t}} = \alpha + {\beta _1}{\rm{GDP}}\,p{c_{i,t}} + \beta {\,_2}{\rm{GDP}}\,pc_{i,t}^2 + {\gamma _t} + {\lambda _i} + {\varepsilon _{i,t}},$$

where γ _t are century or 50-year fixed effects and λ _i are country fixed effects. Our main results are presented in Table 1. The coefficients on both GDP per capita and GDP per capita squared suggest a non-monotonic relationship between income and sumptuary legislation consistent with our theory. The magnitude of the coefficients suggests that an increase in per capita GDP of $100 (1990) is associated with 2–2.5 more sumptuary laws being passed per century. Naturally, these regressions should be interpreted with caution, as we do not have a causal research design. The coefficients shrink somewhat and lose precision when we employ country fixed effects, reflecting a lack of statistical power. The sign and magnitude of the coefficient estimates remain consistent with the hypothesized non-monotonic effect of income.²³

The Rise of Sumptuary Laws at Lower Levels of Income: City-Level Analysis

Real wages rose after the Black Death. For the following 150 years, recurrences of the plague repeatedly reduced populations and put upward pressure on real wages. We exploit plague outbreaks as a source of exogenous variation in income to study the effect of income on sumptuary legislation in Renaissance Italy.

The relationship between plague outbreaks and income is well-established for the period between 1300–1600.²⁴ Urban artisans and craftsmen, in particular, saw their incomes grow as demand for their services increased (Dyer 2005). In England, as the total population halved, “each household could afford to buy more goods, global consumption fell by much less than a half, and in cases such as meat or cloth the total may well have increased … A reduced number of traders and artisans were kept busy supplying the demand, and their increased workload brought them higher incomes” (Dyer 2005, p. 132).²⁵

The quality of the dress of ordinary people improved after the Black Death. Before, peasants wore simple tunics and “economized on the dyeing of their clothes” by “wearing ‘white,’ the natural color of the fleece, or russet, a shade of grey” (Dyer 1989, p. 176). Afterward, this changed as ordinary people had the income to purchase showy clothing. Peasants “could indulge in the greater comfort of the linings, and dress more colourfully, with blues and greens to some extent replacing the old whites and russets.” The number of clothes owned by ordinary people also increased. Similar developments are attested to across Europe (see Online Appendix 2).²⁶

Across Europe, migration increased following the Black Death (see Gottfried 1983, p. 135). Higher wages attracted workers to cities from the countryside. But any downward pressure from inward migration did not offset the overall effect of the plague, which was to raise real wages. We confirm this empirically for Florence. Online Appendix A.4 uses data from Fochesato (2018) to show that plague outbreaks in Florence were strongly positively associated with population falls and real wage growth.

Historians have speculated about why sumptuary laws increased following the Black Death (see Scheidel 2017). Rublack (2019) writes that “sumptuary legislation reflected a growing concern for social regulation from the fourteenth century in response to plague, famine, extended warfare and greater mobility.” For Muzzarelli (2009, p 608):

On the one hand, the tendency toward unnecessary spending could be understood as an expression of vitality unrestrained by the fear of death that the plague perpetuated everywhere. On the other, the high percentage of deaths … would have increased, for some people, the possibility of spending lavishly on clothing and jewelry.

In contrast, our model provides an economic explanation for the increase in sumptuary laws. Increased incomes for peasants, craftsmen, and merchants intensified status competition. One response was the attempt by elites to crack down on status competition through sumptuary legislation.

While the intensity of the initial outbreak of the Black Death was largely exogenous to city characteristics (Jedwab, Johnson, and Koyama 2024), one may wonder whether subsequent outbreaks of the plague were correlated with city-level characteristics such as size, trade links, and transport infrastructure. The consensus among both historians and social scientists, however, is that the timing of these subsequent outbreaks was largely idiosyncratic (Siuda and Sunde 2021). Dittmar and Meisenzahl (2020, p. 984), for instance, note that “the short-run distribution of plague outbreaks was random, conditional on observables, and was geographically localized.”

We use data on plague outbreaks from Biraben (1975) and data on city-level sumptuary laws from Killerby (2002).²⁷ We exploit variation in the timing of a plague reoccurrence and conduct a panel-data analysis to take into account city-specific unobservables and common time trends. Specifically, we estimate the following difference-in-differences equation:

(5)

where N. Sumptuary Laws is the number of sumptuary laws in city i and decade d. Our explanatory variable, N. Plagues_i,d–j is the sum of the number of plague outbreaks in the prior two decades.²⁸ Γ_i are city fixed effects, Λ_d are decade fixed effects, and X is vector of city-level controls. Specifically, we include controls for geographic characteristics that might make a city more suitable for commerce and trade, such as latitude, longitude, Roman roads, whether a city has access to the sea or is on a river, its elevation, ruggedness, and soil quality. We also control for its population size in 1300 and institutional characteristics that might influence the decision to introduce sumptuary legislation, such as the presence of an archbishopric, bishopric, university, or commune.²⁹

Equation (5) identifies the effects of the plague on the number of sumptuary laws under the assumption that in the absence of the plague, the unobserved differences between cities affected and unaffected by the plague would be the same over time. To explore whether the parallel trends assumption holds, we exploit the fact that there were no plagues prior to the outbreak of bubonic plague in 1348. Figures 4a and 4b report the number of sumptuary laws in cities that had above the mean or median number of plague outbreaks after 1348 against those cities that had below the mean or median number of plague outbreaks. There are no discernible differences prior to 1348: cities above and below the mean or median number of plague occurrences exhibit the same trend. This lends credence to our difference-in-differences research design.

Figure 4 PARALLEL TRENDS: SUMPTUARY LAWS IN CITIES MORE OR LESS AFFECTED BY POST-1348 PLAGUE.

Sources: See text and Online Appendix 1.

Table 2 reports the estimated impact of the plague on the number of sumptuary laws. The sample includes all cities in Italy for which there is information on plague occurrences in Biraben (1975). Our preferred specification includes the interactions of geographic and institutional controls with decade fixed effects. Due to concern that the data might be over-dispersed, we report results using both OLS and negative binomial specifications. The corresponding incidence ratio (in square brackets) implies that one occurrence of the plague in city i in the previous two decades increased the number of sumptuary laws by 1.2.

One concern is that plagues could have affected city-level institutions through mechanisms other than their effect on income. While we cannot rule this out entirely, we do not believe that it is driving our results. Indeed, we find no relationship between plagues and city-level institutions in Italy between 1300–1500 (Online Appendix Table A.16).³⁰ Overall, the historical literature suggests that post-Black Death outbreaks of bubonic plague affected per capita income and real wages, but did not have a decisive impact on political institutions (see Cohn (2004, 2010) for an analysis of popular rebellions following the Black Death).

Another concern is that the Black Death affected consumption patterns through another channel, such as discount rates. Richardson and McBride (2009) suggest that the Black Death reduced incentives to cooperate in the future. They suggest that this was why late medieval craft guilds acquired religious functions. Indeed, the Black Death was such a dramatic shock that it could have affected sumptuary legislation through various channels.

To address this, we show that our results are not driven by the period of the Black Death (1347–1352) itself. In Tables A.11 and A.12, we systematically exclude first the immediate years following the plague (Columns 1–2); second the half century after the plague (Columns 3–4); and finally, all cities that experienced more than 50 percent population loss during the Black Death according to Christakos et al. (2005) (Columns 5–6). The coefficient on plagues remains strong and precisely estimated.

Plague outbreaks following the Black Death were not spatially correlated. There was little direct city-to-city transmission. Nonetheless, as a further check, in Online Appendix 4, we consider spatial autocorrelation using Conley standard errors of varying radiuses (Tables A.9 and A.10). None of these exercises significantly modify our baseline results. Table 3 demonstrates that our results are highly robust when we exclude cities in the north or south of Italy, cities in Sicily, and coastal or riverine cities. Our results similarly hold for both large and small cities.

Next, we consider how this relationship between sumptuary legislation and income would have been influenced by political and economic institutions, specifically the degree of rent-seeking in a society.

Institutions and the Extent of Rent-Seeking

As North, Wallis, and Weingast (2009) argue, the degree and character of rent-seeking within natural states vary considerably. The most advanced late medieval Italian city-states were on the “doorstep of the transition” from natural states to what North, Wallis, and Weingast (2009) call open access orders. Belloc, Drago, and Galbiati (2016, p. 1877) study the formation of the commune movement in Italy as “a radical change toward broader-based political institutions.” Acemoglu and Robinson (2019) consider them key case studies in the “shackling of Leviathan.” We expect that in cities with communes or in cities with republican institutions, the degree of rent-seeking by elites (as captured by τ) was lower.³¹ In contrast, in despotic regimes—cities dominated by a single ruler, a despot, or signorie in the terminology of contemporaries—the was a narrower group of elites.³²

We find evidence that the relationship between sumptuary legislation and income is indeed stronger for cities in which elites’ rent-seeking was low. We use three different codings to capture differences in τ. First, we consider despotisms, or signoria, as coded by Killerby (2002). These were city-states ruled by a single individual as opposed to a republic. In the thirteenth century, many Italian cities transitioned from being governed as republics or communes to being ruled by a single individual. In cities under despots, such as Milan, ruled by the Visconti family, or Ferrara by the Este family, we expect τ to be high.

Second, we use information from Bosker, Buringh, and Zanden (2013) to ascertain whether cities established communes in the eleventh or twelfth centuries. This measure has the advantage of being determined prior to our period of study. We expect τ to be lower in cities with communes.

One concern is that plague outbreaks might impact τ as well as Y. Here, we are reassured by recent research indicating that while the Black Death did reduce income and wealth inequality, this was not necessarily true of later outbreaks (Alfani 2021, 2022). Rather, the relationship between demographic shocks and inequality was mediated by initial institutions. Moreover, in the Italian cities that we study, elite incomes in the fifteenth century rose because Italian elites were merchants as well as landlords, and their incomes were buoyed by the growing demand for luxury goods.³³

We also construct our own measure of republican institutions. This allows us to exploit variation over time in a city’s institutions.³⁴

Table 4 reports the effect of plagues on sumptuary legislation when we split the sample according to the degree of rent-seeking by elites τ. The relationship between income and the number of sumptuary laws is strongly positive for cities that were non-despotic, had historical communes, or were republics. For cities that were despotisms, non-communes, or non-Republicans, however, we find either no effect or a smaller effect. This does not reflect differences in the number of observations for despotic cities, non-commune cities, and non-Republican cities. These results support Hypothesis 2: “despotic regimes within Italy passed fewer laws than their republican counterparts” (Killerby 2002, pp. 33–34). Sumptuary legislation was more intense in cities with more inclusive institutions.

The Decline of Sumptuary Laws

Our city-level dataset does not extend beyond 1500. We therefore cannot explore the decline in sumptuary laws at the city level. But the evidence we have is consistent with Hypothesis 1 and suggests that evasion and the costs of enforcement help to explain the post-1600 decline in sumptuary laws.

First, we can investigate the relationship between sumptuary legislation and income in France quantitively, as Ridolfi (2019) provides yearly estimates of per capita GDP for 1280–1779. Figure 5 plots the number of sumptuary laws by decade against per capita GDP by decade. There is a clear inverted-U relationship consistent with Hypothesis 1.

Figure 5 FRENCH SUMPTUARY LAWS (BY DECADE) AND GDP PER CAPITA

Source: Data described in Online Appendix 1.

Second, the qualitative evidence from eighteenth-century France provides insight into their decline. Opponents of strict sumptuary laws like Jean Francois Melon argued that many so-called luxury goods, such as cheap silks, were already widely available to the lower classes.³⁵ They further noted that enforcement was increasingly difficult with a wider range of goods available: “individuals ignored the spirit of the laws by substituting another type of luxury for the forbidden item” (quoted in Moyer 1996, p. 355).

Over time, greater resources had to be spent on enforcement. Moyer observes that “As the seventeenth century wore on, sumptuary law mandated more frequent searches of workshops and boutiques, usually by officers of the local police accompanied by guild jurés” (p. 254). Inventories and records were also required to be kept by merchants to ensure that they were not selling prohibited products. Nevertheless, and despite these police measures, the “royal and municipal officials entrusted with sumptuary law execution were beset with problems” (Moyer 1996, p. 255).

The consumer revolution of the seventeenth to eighteenth centuries made sumptuary laws increasingly difficult to enforce. Opportunities for luxury consumption could not easily be restricted to the elite (McKendrick, Brewer, and Plumb 1982; Brewer and Porter 1993; Vries 2008; Koyama 2012). Consumer aspirations diffused among the middle and lower classes (see Styles 2007). New opportunities for fashionable clothing rose with the emergence of retail shops, where individuals could purchase ready-made clothes, as opposed to purchasing whole cloth (Mui and Mui 1989). The real cost of clothing also fell (Shammas 1990). By the eighteenth century, Lemire (1991, p. 97) observes that even servants could save up the “eight shillings for a ready-made gown,” creating “a potentially vast market among working-women, for whom these prices meant perhaps one week’s wages or less.” All of this made it more costly to limit the expenditures of non-elites.

With economic growth, evasion became easier and enforcement more difficult. It became increasingly challenging to differentiate between those who were permitted to wear particular items of clothing and those who were not.³⁶ Medieval guilds played an important role in regulating economic activity (Ogilvie 2019). They often cooperated with local authorities in the enforcement of sumptuary laws (Moyer 1996). But after 1600, guilds went into decline, especially in Western Europe. As their hold on the economy weakened, it became harder to punish vendors who violated sumptuary legislation. Finally, at the end of our period of analysis, the rise of ready-to-wear clothes made it easier for non-elites to emulate the clothing of elites. As a consequence, in the nineteenth century, luxurious clothing ceased to be as important a signifier of social status as it had been previously.

CONCLUDING COMMENTS

Sumptuary laws were denounced by Adam Smith (1776) as the “highest impertinence and presumption.” As rulers are “the greatest spendthrifts in the society,” he cautioned them to “look well after their own expence, and they may safely trust private people with theirs. If their own extravagance does not ruin the state, that of their subjects never will.” But Smith did not investigate why sumptuary laws were passed or why they fell out of favor, and, while they have attracted attention from historians and sociologists, sumptuary laws have been neglected in economics.

We introduce a model of sumptuary legislation that generates a non-monotonic, inverted-U relationship between per capita income and sumptuary legislation. As incomes rise, non-elites can spend more on status goods, obliging elites, on the one hand, to also spend more and, on the other hand, to enact sumptuary laws in order to restrict the consumption of non-elites. As incomes rise still further, however, non-elites are better able to evade these laws. Elites, moreover, also benefit from economic growth and can spend more on status consumption. At a certain point, therefore, sumptuary laws cease to be worthwhile for elites.

We consider two extensions. First, as sumptuary laws could be used a de facto luxury taxes (i.e., Muzzarelli 2019), we allow for the fines from sumptuary violations to offset the costs of enforcement. This encourages elites to pass more sumptuary laws and thus slows down the decline in sumptuary laws (muting the inverted-U-shaped relationship). Second, the baseline model considered status competition between the landed elite and ordinary citizens. We extend the model to allow for the elites to also be merchants. The results of the model carry over: commercial elites also have an incentive to maintain status distance from those below them in the social hierarchy. The main difference is that if the commercial elite is extremely rich (as was the case, for example, with the Medici family in Florence), this lessens the status threat that they face, reducing, at the margin, their incentive to employ sumptuary laws.

The non-monotonic relationship between income and sumptuary laws is steeper for states when elite rent-seeking is lower and more moderate, where the elites capture a larger share of income. Overall, we argue that this framework can account for much of the otherwise puzzling rise and fall of sumptuary laws in premodern Europe.

Empirically, we document a non-monotonic, inverted-U relationship between sumptuary laws and per capita GDP at the country level, and at the city level, we provide evidence linking increased incomes to sumptuary laws in Italy.

Table 1 SUMPTUARY LAWS AND GDP

Notes: This table reports the relationship between estimates of per capita GDP and the number of sumptuary laws per century. All coefficients are multiplied by 1000 for readability. Standard errors clustered at the modern country level are reported in parentheses.

* p < 0.10, * p < 0.05, *** p < 0.01

Source: Authors’ calculations, see text.

Table 2 THE EFFECT OF PLAGUE SHOCKS ON THE NUMBER OF SUMPTUARY LAWS: CITY-LEVEL PANEL ANALYSIS, DECADE LEVEL

Notes: This table reports difference-in-differences estimates of the impact of the plague. Columns (1)–(3) report OLS results. Columns (4)–(6) report results obtained from a negative binomial specification. The unit of observation is a city-decade. All specifications include city and decade fixed effects. Geographical controls include longitude, latitude, elevation, whether a city is on a river or the sea, and soil quality. Institutional controls include the presence of universities, bishoprics, and communes. The number of observations drops as we lack control variables for some smaller cities. We report incidence ratios in square brackets. Robust standard errors clustered at the city level are reported in brackets.

Source: Authors’ calculations, see text.

Table 3 THE EFFECT OF PLAGUE SHOCKS ON THE NUMBER OF SUMPTUARY LAWS: CITY-LEVEL PANEL ROBUSTNESS, DECADE-LEVEL

Notes: This table reports our main robustness checks for our city-level analysis. The unit of observation is a city-decade. All specifications include city and decade fixed effects. Geographical controls include longitude, latitude, elevation, whether a city is on a river or the sea, and soil quality. Institutional controls include the presence of universities, bishoprics, and communes. Column (1) reports our baseline estimates. The largest cities excluded in Column (2) are Venice, Florence, and Milan. Column (3) drops cities in Sicily. In Column (4), we focus on northern Italian cities (defined as those above 41.9028, the latitude of Rome). Column (5) focuses on cities below this latitude. Column (6) excludes coastal cities. Column (7) drops cities on rivers. In Column (8) we exclude cities with elevation greater than 122 meters. Column (9) drops cities were on the interaction of a Roman road. Column (1) drops cities with medieval universities. Columns (10)–(11) exclude larger and smaller cities. Robust standard errors clustered at the city level are reported in brackets.

Source: Authors’ calculations, see text.

Table 4 THE EFFECT OF PLAGUE SHOCKS ON THE NUMBER OF SUMPTUARY LAWS: CITY-LEVEL PANEL ANALYSIS BY CITY INSTITUTIONS

Notes: This table reports difference-in-differences estimates of the impact of the plague on the number of sumptuary laws by state type. The unit of observation is a city-decade. All specifications include city and decade fixed effects. Geographical controls include longitude, latitude, elevation, whether a city is on a river or the seas, and soil quality. Institutional controls include the presence of universities and bishoprics. Robust standard errors clustered at the city level are reported in brackets.

Source: Authors’ calculations, see text.