Mercantilism and Bureaucratic Modernization in Early Eighteenth‐Century France

French mercantilism is generally associated with absolutist policy&#8208;making subject to capture by rent&#8208;seeking interests. This article investigates how the Bureau du Commerce, a small state agency in charge of commerce and the supply side, handed out rents and privileges to private entrepreneurs. We coded how the Bureau investigated and decided all 267 voluntary submissions received between 1724 and 1744. It is shown that the Bureau&#8217;s formal, rule&#8208;based decision&#8208;making process could actually differentiate between alternate policy aims and target them consistently over time, with more or less powerful sets of rents. From this, a hierarchy of revealed policy preferences is derived. First comes technical innovation and diffusion, then local economic development; import substitution is only in the third position, followed by consumers&#8217; welfare. Lastly, and in contrast to a long line of authors, it is shown that the production of luxury goods was not a significant or valued objective.

T he term 'mercantilism' is widely held to refer to a broad period of European history during which absolute monarchs followed illiberal policies based on false premises: protectionism, widespread monopolies, and the confusion of wealth, money, and power. However, beyond this broad and standard definition, it is difficult to identify a clear set of intermediate objectives and policy tools that would define more precisely what exactly mercantilism was about.
A well-known part of the problem, of course, is that the theoretical foundations of mercantilism are shallow. Contrary to what Adam Smith suggested, there is no single, logically consistent 'mercantile system', so that on balance we may rather follow Landes and conclude that there was just 'pragmatism gilded by principle'. Alternatively, should we join Schumpeter, for whom mercantilism was only an 'imaginary entity'? However, even if we discount these analytical limitations and look only at the historical experience, approximations and hearsay tend to dominate. For example, many authors take for granted that the overarching aims of mercantilist policies were to substitute domestic production for imports and to absorb foreign technology-but they rarely ask whether local development or technological diffusion at the domestic level were significant objectives as well, and, if so, with what degree of priority. Another cliché is the supposed preference of many mercantilist policy-makers for the production of luxury goods. In his classic study, Heckscher argued that it was indeed a key priority, which made some sense in a country such as France, where there was much demand from the aristocracy and the court. 1 But he also added that, as a consequence, manufactures later missed the turn towards broader consumer markets and large-scale production; hence their long-term failure. This argument has resurfaced in a long series of contributions, many of which take as a standard the example of the Manufacture Royale des Gobelins, in Paris. 2 However, none of these authors ever tested the general validity of the initial proposition, nor the long-term effect suggested by Heckscher. Alongside this old discussion about mercantilist policy objectives and their hierarchy lies an entirely different, second set of questions about the resources and capabilities of emerging state bureaucracies. Schmoller was among the very first to defend the view that mercantilist policies came together with the development of modern bureaucratic states, or the Rechtstaat. 3 Beyond its origins in an age of absolutism, mercantilism is thus envisaged as a modern, forward-looking regime. The usual example cited for this is the French ministry of finance, known as the Contrôle Général des Finances. However, Ekelund and Tollison, later followed by Root, reject this modernizing view: 4 they claim that in fact this new state machinery would have only increased the overall capacity of the Crown to extract fiscal resources and redistribute them as rents, so as to maintain the socio-political equilibrium of this 'natural state'. 5 Both assessments are very general, however, in so far as they propose a very broad, undifferentiated view of this bureaucracy and its politics. Critically, they do not ask whether parts of it could escape capture and corruption, and indeed work as an agent of change, or whether different types of rents may have had different impacts. It is not actually a moot question whether progressive trends could work from within the Ancien Régime monarchy, even though the macro-picture is one of a low growth economy, slowly drowning in debt.
This article presents a novel approach to these two issues: mercantilist policy preferences and the underlying bureaucratic capabilities, with the latter being interpreted here as the capacity to align means and ends (or tools and preferences) in a time-consistent, rule-based, predictable manner. We take the case of the distribution of privilèges de manufacture; that is, franchises and rents awarded to private entrepreneurs that operated outside the guilds. This policy had its origins under the guidance of Colbert and persisted until the Revolution, though the personal will and discretion of the Contrôleur Général progressively led the way to a more rule-based process in the hands of the Bureau du Commerce: a small Paris-based department within the Contrôle Général that was in charge of trade and supply-side policies. The period under review (1724-44) was marked by well-established procedures (hence the availability of large archives) and also by supportive economic conditions thanks to, inter alia, an extended cyclical upturn and the quasi-absence of wars. The fact that a total of 267 decisions were made during these two decades, out of a total of about 450 between 1700 and 1790, also suggests that this post-Colbertist industrial policy experienced a kind of golden age during this period. Indeed, from the mid-century onwards, the influence of the mercantilist doxa would gradually decline under the onslaught of the more liberal physiocratic doctrine.
The point from which we start is that the packages of benefits handed out by the Bureau resulted from the manipulation of some 20 discrete regulatory variables, whose economic impact could not be expected to be additive. They include, for instance, tax breaks on inputs or on the entrepreneur's income, privileged access to some natural resources, or territorial exclusivity. We use this multi-dimensional pattern as an instrument to test whether this bureaucracy was able, over two decades, to align franchises and rents with the characteristics of the projects: that is, whether it was able first to identify characteristics it wanted to support and then to target them in a differentiated, time-consistent manner. If our 267 decisions had followed an entirely erratic pattern, reflecting court politics or royal caprice, then we would probably not identify any structural policy bias.
Contrary to this reasonable expectation, we show that the Bureau du Commerce in fact worked as an embryonic policy instrument in the hands of the government: decisions made on a case-by-case basis are proven to be consistent, so that over 20 years similar projects received similar support. These more or less extended packages of benefits therefore reveal a stable hierarchy of policy preferences: these early policy-makers first supported technological innovation and diffusion, then local economic development, thirdly import-substitution, and then consumers' welfare; on the other hand, and contrary to the common wisdom, supporting luxury-goods producers was definitely not a priority between 1724 and 1744.
Overall, we thus cautiously endorse the long intellectual tradition that associates mercantilism with state-building, or with an emerging practice of governmentality, rather than with systematic capture, and sterility. This conclusion only applies to the period under review, however, and it does not imply any judgment, positive or negative, regarding the actual impact of this policy on the development of the French economy under the Ancien Régime.
The rest of this article is organized as follows. It begins with a brief review of the literature on mercantilism and shows exactly where the present contribution fits (section I). The next section discusses the institutional context of the time and how the Bureau du Commerce fitted in (section II). The focus then moves to the investigation and decision-making process that underpinned the distribution of rents and franchises to firms, and the archives from which we built our dataset are described (section III). Econometric tests (section IV) confirm that the Bureau had a proven and consistent capacity to target different policy aims, or preferences, with varied packages of privileges. Section V provides a conclusion.

I: Contrasting views of mercantilism
A first striking feature of the classic literature on the Mercantilist Age is that most historians of this era are typically cautious when appraising the effectiveness and impact of post-Colbertist industrial policies. Mantoux, for instance, concludes that they had only 'a limited scope' so that no link can be established with the subsequent industrial take-off. 6 Goubert,7 like Braudel,8 states that most royal manufactures had 'an artificial character', hence a limited capacity for survival without government support. Landes 9 and Wallerstein 10 share a slightly more optimistic view and conclude that they might have had some long-term positive impact. What is thus remarkable in these conclusions is not just the care with which these authors avoid clear-cut judgments on the effects of these policies; they also resist making any strong statement about the possible imprint of the mercantilist doctrine on the policies themselves.
In stark contrast, a second large strand in the literature fully acknowledges a normative, if not partisan, investment in the subject. A first line belongs to the old statist/nationalist trend in modern French political culture that originated during the last decades of the nineteenth century, in the monarchist and then nationalist glorification of the reign of Louis XIV. 11 This view then took an economic turn during the interwar period, with Boissonnade, for instance, so that in the postwar decades Colbert's policies were widely seen as the true origin of a dirigiste economic tradition. 12 Today this romantic view still exercises a strong hold over part of the bureaucratic and political elite. The most influential body of literature, however, has German origins, first in the writing of Friedrich List, and more clearly in the work of Schmoller who, as stated earlier, was the first closely to associate mercantilism with the construction of a modern bureaucratic state. 13 The link with the experience of the German Second Reich, in his case, is most explicit. 14 However, Schmoller's best-known pupils and associates, Sombart 15 and Weber, 16 took a similar position. The theme later found a major extension in the landmark study by Heckscher, a Swedish economist, covering the whole European field, 17 which is nicely complemented by Cole, who focuses on the French experience. 18 This 'German' tradition has been contested, however, by contributions which defend a positive rather than an intellectual perspective. As argued most forcefully by Ekelund and Tollison, the proper approach to mercantilism should not start from ideas and social beliefs, but from the social and political conditions of the day, that is, the rent-seeking character of the absolutist regimes; once on that positive basis, one may account for the actual policies and for the more or less sophisticated theoretical arguments that were supposed to back them up. 19 Starting from such materialist premises, these authors reinterpret a fair part of the material collected inter alia by Heckscher and Cole and present a coherent though somewhat 6 Mantoux, Révolution industrielle. 7 Goubert, 'Le "tragique" XVII e siècle'. 8 Braudel,Civilisation matérielle,p. 288. 9 Landes,Unbound Prometheus,p. 136. 10 Wallerstein,p. 268. 11 Lavisse,Louis XIV, Boissonnade, Colbert. 13 Schmoller, Mercantile system. 14 Schmoller, Lenz, and Marcks, Zu Bismarcks. 15 Sombart, Luxury and capitalism. 16 Weber, General economic history. 17 Heckscher, Mercantilism. 18 Cole, Colbert; idem, French mercantilism. For critics of the Heckscher thesis, see Coleman,ed.,Revisions;Judges,'Idea';Viner,'Power'. 19 Ekelund and Tollison, Mercantilism; eisdem, Politicized economies; eisdem, 'On neoinstitutionalist theory'. schematic view of the Ancien Régime bureaucracy as a kind of embryonic centralplanning authority in charge of a detailed allocation of rents across the country. 20 At this point, there is of course little point in looking for economic policies in the modern sense of the word: registering and enforcing clientelistic deals across the country was all the royal bureaucracy could manage. Root offers a more nuanced and persuasive interpretation of the Ancien Régime, also based on an analysis of rent-seeking politics. 21 Specifically, he pays much more attention to the rules of the game at the provincial level and the relationships between provincialélites, the monarchy, and the bureaucrats. This work is therefore more in line with the many monographs that have analysed the role of social capital and political networks in the government of given cities or provinces; 22 it also echoes what is now known about Colbert's own clientelistic practices. 23 However, Root's mostly local or provincial perspective does not have as much to say about the build-up of administrative capabilities, procedures, or new pools of expertise at the centre. He concludes nevertheless that 'Paris and Versailles were transformed into elaborate centers of distributional play, since court politics determined what would be produced and who would benefit'; 24 he later further defends the proposition that, in practice, 'only a small coterie around each minister were thought to benefit [from rents]'. 25 Outside the public choice paradigm, and among the most recent contributions Horn develops a novel and entirely opposite interpretation of privileges: 26 rather than being a tool for systematic rent-seeking strategies, they would have worked primarily as an instrument to extricate the most entrepreneurial part of firms from the ramshackle institutional construction that was the Ancien Régime. A number of other historians have also studied mercantilist policy-making 'in action', typically at the local level. Reynard, for instance explored the evolving relationship between the state administration and the paper manufacturers of Auvergne, 27 while Thomson and Johnson dealt in similar manner with the export-oriented textile industry in Languedoc. 28 With an issue-based perspective rather than a local or territorial one, Szulman and Graber have looked at the Bureau's policies regarding transport infrastructure, which they examine from a micro-level, project-based perspective. 29 These works are complemented by Smith, who focuses on how the Bureau du Commerce gradually constructed social and administrative networks within which actors could engage in a discussion framed, relatively early on, by liberal notions of public good and open-market competition. 30 Smith on the Bureau, and Kessler on 20 According to Ekelund and Tollison,Mercantilism,p. 84, the 'French mercantile economy took on the form of a massive web of economic regulations administered by the central state'; hence, 'the French nation-state [rose] as an extension and consolidation of monopoly power in the economy' (ibid., p. 85). These authors then insist that the local bureaucracy, in the provinces, was primarily in charge of enforcing monopolies. For criticism of the rent-seeking approach to mercantilism, see Rashid, 'Mercantilism'. 21  the eighteenth-century Paris traders' court, show in particular how these themes gradually sidelined the core mercantilist doctrine and prepared the ground for the more liberal physiocratic discourse after 1750. 31 Though often close to these later works, the present article does not focus on provincial politics, on the discourse and the values of social actors, or indeed on the evolving architecture of the absolutist monarchy. Instead, we look at how the Parisian Bureau du Commerce made decisions between 1724 and 1744 and whether its formal rules and procedures endowed it with some autonomy with regard to its social and institutional environment, so that its decisions would remain consistent over time and across cases. Our point therefore is not to refute the widespread clientelistic character of the old monarchy. Archives are actually replete with personal applications for rents and favours, which are all very open and frank. The question we ask is to what extent formal bureaucratic rules and procedures were able to resist the pervasive pressure of rent-seeking politics and so provide ministers with a pliable policy instrument, strong enough to resist capture by the thousands of cronies and brokers that laid siege to their offices.
At this point, we should also remember that for Foucault, mercantilism marked the emergence of a new model of governance: against the tradition represented by Machiavelli, with its insistence on sovereignty and on the preservation of the prince's power, government would be about the needs and the welfare of the population, though also about the drive to control society, especially those who live on its fringes. 32 In his view, this trend was associated with the gradual strengthening of l'État Administratif (as opposed to the older, medievalÉtat de Justice). The experience under review here may thus be seen as a telling example where bureaucratization and mercantilist policy-making converge indeed into in an early form of governmentality.

Entrepreneurship under a weak, absolutist state
Together with the patrimonialist character of the French monarchy, a further, serious obstacle to the emergence of a notion of public good, and of a public policy to serve it, was the minutely fragmented institutional and social structure of the Ancien Régime. In those days, there was no such thing as a representative royal subject, endowed with an identifiable package of rights in their modern, impersonal, and abstract definition. 33 Individual statuses and franchises were defined by group affiliation, religion, guild membership, and a wide array of coutumes, or customary laws, that were highly differentiated across and within provinces. On top of that, ad hoc privilèges were the typical way for superior authorities, starting with the king, to grant resources or franchises to individual or to entire social groups, such as cities or guilds. Therefore, royal subjects were not only unequal; they were also thoroughly different, which is certainly not a good starting point for implementing government policies that would apply to all agents 31 Smith, 'Le discourséconomique'; Kessler, Revolution. 32  across the country. 34 Still, the fact that this regime presented few formal checks and that arbitrariness was a built-in option did not imply lawlessness: privileges, like those we are studying here, had a legal form and they were enforced by courts, and hence they should not be regarded as similar to a casual bribe or a pension that the king would hand out and withdraw by fiat.
The monarchy invested enormous effort, over centuries, in trying to integrate both the polity and the economy: this is what absolutism was ultimately about. It thus imposed the French language in all official occasions, unified the monetary order, constructed roads and canals, and attempted to reduce domestic customs and tolls. 35 Critically, the Ancien Régime monarchy established a common legal framework for economic activities by enacting the 1673 Ordonnance sur le commerce-in practice, a rather light formalization of the old Merchant Law. These basic, impersonal market rules were then enforced across the kingdom by elected merchants courts, or Cours Consulaires, which followed the traditional rapidpace, free-of-charge practices of the merchants' fair courts. 36 Hence the policy of handing out differentiated packages of rents to private firms unfolded against the background of an economic playing field based from a relatively early stage on principles of universality and impersonality. We focus here on a single category of bureaucratic decision: the awarding of privileges to private entrepreneurs. We do not deal, therefore, with the other pillars of this policy-the regulation of guilds by way of statutes, or the Règlements des Manufactures, which are dealt with by Thomson and Reynard. 37 The Bureau du Commerce: an emerging modern bureaucracy From the 1660s onward, the driving force behind the attempt to integrate and develop the domestic economy was the Contrôle Général des Finances, akin to today's ministers of economy and finance. While remaining anchored in the Conseil du Roi, hence the king's personal authority, it developed a small corps of civil servants as well as its own territorial network of agents and offices. 38 This is where a new model of administrative work was developed, based upon a streamlined division of labour, a top-down hierarchy, rules-based methods for investigating policy issues, well-kept files, and so on. The size of this bureaucracy remained small, however, at least by today's standards: according to Felix On the bureaucrats, see Ricommard, 'Subdélégués'; Gruder, Royal provincial intendants; Antoine, 'Notion de subdélégation'; Grevet, 'Etre subdélégué'; Mukerji, 'Jurisdiction'. On the Commis as the real embryo of modern bureaucrats, see Baxter, 'Premier commis'; Felix, 'Commis'. See also Barbiche, 'Ce que nous ne savons pas', where he asks, rhetorically, 'What we don't know about the Contrôle' and then points out that its decision-making process has not been much explored. by 1789; the local network in the provinces would have totalled some 540 persons by mid-century. 39 Whereas the Contrôle Général dealt primarily with fiscal and budgetary affairs, the Bureau du Commerce was created as a semi-autonomous department in 1700; a clear advantage derived from the fact that it was not formally a part of the Conseil du Roi, so that it could economize on the arcane rules that surrounded the exercise of the divine right to rule. 40 Until 1791 it continued to be in charge of the supply side of the economy (essentially the guilds and the manufactures) and commerce (domestic, colonial, and foreign). This is where post-Colbertist, mercantilist policies were actually debated and implemented. 41 This large portfolio was managed by a body of some 20 to 25 persons, whose head was known until 1744 as the Directeur du Commerce. A Secretary took care, inter alia, of the correspondence of the Bureau and kept its official files, the largest part of which has been preserved until today. Then, four top-level Intendants du Commerce were responsible for specific geographic areas and had authority over the lesser commis and the corps of Inspecteurs des Manufactures. 42 Together with the distribution of privileges, though not included in the present study, the Bureau also handled guilds' statutes or, for instance, the bargaining game between the Bureau and whole industries over the Règlements, as Thomson and Reynard did. 43 Submissions for privileges made by entrepreneurs to the Bureau included a detailed presentation of the project, the technique it would use, the experience and competencies of the manager, the potential contribution to the local or national economy, plus a list of privileges that were deemed necessary for success. As we explored how these applications were then processed, one of the first features that became apparent is the systematic, wide-ranging attempt by the Bureau du Commerce to collect all significant information on the conditions that would weigh locally on the development of each project. Officials thus wanted to know about the availability or shortages of natural resources and workforce, or about the possible opposition from municipal authorities, local manufacturers who already had some privileges, or most typically from guilds. The archives of the Bureau therefore encompass a wealth of reports sent by government officials in the provinces, 39 Felix, 'Commis'. 40 The royal decree of 29 June 1700 formally created a Conseil de Commerce, which should have been one of the components of the Conseil du Roi. The king, however, never chaired it, and it started to be officially recognized as a Bureau in 1715. In fact, it gradually emerged as such, especially after 1708 when a clear distinction was established between representatives of the ministries and those in charge of investigating and managing cases. With the ascent to power of Louis XV (1724), the Bureau was given its status, organization, and modus operandi that prevailed until the revolution. A Conseil Royal du Commerce was, however, established (again) in 1730, though with no more success than its antecedent. 41 The most comprehensive studies of the Bureau du Commerce are those by Schaeper, French Council, who covers the first 15 years of its existence (1700-15); see also Rothkrug, Opposition, on these early years. Parker, Bureau of Commerce, and idem, Administrative bureau, are two books on the years 1781-3 that provide a fine description of its structure and mandate, as well as of its relationship with the Ministers. Smith, '"Au bien du commerce"', and idem, 'Structuring politics', are two important contributions, together with the recent and important book by Horn, Economic development, who looks less at the bureaucracy per se than at its policies and policy tools. Wybo, Conseil de Commerce, remains a useful source. Bonnassieux and Lelong, Conseil de Commerce, is an indispensable index of the Bureau's archives from 1700 until 1790. On the Intendants du Commerce, who constituted the administrative backbone of the Bureau and should not be confused with the Intendants, who were the representatives of the Crown in the provinces, see Schaeper, French Council; Garrigues, 'Intendants'; Antoine, Coeur de l'État. 42 Minard, Fortune. 43 Thomson, Clermont-de-Lodève; Reynard, 'Early modern state'. occasional letters from local authorities, expert opinions written by well-known merchants or members of the Academy of Science, and so on. Once all this information had been collected, decision-making proceeded through two colleges, where cases were discussed on an open, horizontal, peerlevel basis. First was a group of 10 to 14 Députés du Commerce, who were wellestablished, experienced merchants, typically with a background in long-distance trade, shipping, or banking. The Députés thus offered deep expertise in matters of market practices, the law merchant, banking, or the economic geography of both the kingdom and foreign countries. The world of craft guilds and manufactures, on the other hand, was not strongly represented among the Députés, a feature that may have contributed to their readiness to seek external opinions on these matters. Also, they most often had some experience with municipal government or the local traders' courts, after which they had been elected to the Bureau by the largest trading cities, most commonly via the local Chambers of Commerce. 44 The fact that these cities paid for the living expenses of the Députés clearly supported the expectations that the latter would indeed act as lobbyists. However, the bilateral correspondences between the Députés and the local officials reflect much tension and frustration felt by city officials: 45 contrary to what their name suggests, the Députés did not speak in Paris with the authority and the legitimacy of democratic representation. Discursively, they argued, rather, in the name of 'the good of commerce', which was framed as a fully legitimate subdivision of the public good, or of the good of the kingdom. 46 A good way to describe them may thus be as a body of consultants, on de facto long-term contracts awarded with the expectation of an asymmetric allegiance: first to the royal bureaucracy, then to their city.
On a case-by-case basis, decision-making started with a thorough discussion within the college of Députés based on the submission provided by the applicant, plus the reports and opinions collected by the Bureau. A detailed and motivated collective Avis (opinion) conveyed their view and their recommendation to endorse, reject, or amend the application. The Avis and the whole file were then sent to the second deliberative college which was made up of the core officials from the Bureau du Commerce and representatives of various branches of the royal administration. It discussed the whole submission again and wrote a draft decision to be submitted to the Minister. The Députés attended these plenary meetings although, at this point, they had no say in the deliberation.
Even though he was not formally bound by this recommendation, in all cases except one the Contrôleur Général des Finances endorsed it without modification. 47 The fact that this advice came with a comprehensive, carefully worded justification further underlines the modern, impersonal character of the overall process. The Minister acted as if he trusted that this rule-based, information-driven, pluralist decision-making process offered him the best advice he could obtain. At least from 44 In principle, the Députés had to be elected on a yearly basis, but very soon after the establishment of the Bureau in 1700, new rounds of elections were abandoned and members were either renewed on a routine basis, or appointed by Parisian bureaucrats to replace the outgoing Députés. 45 See, for example, Quénet, 'Un exemple', on the Députés from Nantes; Labraque-Bordenave, 'Histoire des Députés', on the case of Bordeaux; also Pariset, 'Chambre de Commerce de Lyon'; and Fournier, Chambre de Commerce de Marseille. 46 Kessler,Revolution. 47 In the cases we investigated, we found just one example of the Ministry not endorsing the recommendation of the Bureau. a formal perspective, this bureaucratic process is quite different from the cliché of a highly centralized body, working in a few closed rooms at Versailles, a corridor away from the king and his courtiers.
In both colleges, the deliberation revolved most explicitly around the question of how to balance market competition and the protection of innovators and risktakers. This was stated rather solemnly by the Députés in 1724, when they wrote that the Bureau had always defended the principle that exclusive privileges would only be granted for new inventions. Those who requested them in order to produce things already known and produced in France had always been rejected. As the Députés once said: This rule is entirely grounded on justice and the protection that the king owes to all his subjects, as on the equality and competition that commerce demands. Indeed, the favours that would be given to some would only give them superiority over others, and would thus contribute to the destruction of commerce and manufactures. 48 Besides innovation, the development of backward regions, for instance, could justify public intervention as well, as argued in the 1729 case of a large, low-quality textile manufacture to be established in Valence, south of Lyon: the purpose of this project is to set up a new factory in a country where trade should be restored; there must be a positive example to awaken the courage of Peoples; the Entrepreneur will have to make major investments in a firm from which the King and the Province can expect great benefits, since this will prevent skilled workers [from leaving the country], will provide for the consumption of products originating in the Province . . . and finally the city of Valence will become repopulated. 49 III: The logic of granting privileges

Collected data
In order to analyse how privileges were granted to private entrepreneurs in practice, we built a database that includes 267 submissions made to the Bureau between 1724 and 1744 out of a grand total of 284. We thus systematically collected and coded the main characteristics of each submission throughout the whole deliberation and decision-making process. 50 We rely primarily on the Avis des Députés, the contributions of the provincial Intendants, and the minutes of the deliberations of the Bureau. In most cases, these sources allow us to identify the content of the initial application; the position and arguments put forward by the other voices involved in the investigation; the conclusions reached at the plenary meeting of the Bureau; the final decision with the list of privileges that were actually granted; 48 Archives Nationales, Paris (hereafter AN), F 12 / 73, Avis des Députés, 22 Dec. 1724. 49 AN, F12/77, Avis, Intendant du Dauphiné, 17 May 1729. 50 Our database has been built entirely on the archives of the Bureau du Commerce, which are kept at the AN (Fond F 12 ); see Bonnassieux and Lelong, Conseil de Commerce, for a most useful inventory. We relied on three types of documents: first, the minutes of the Bureau (F 12 51-91), with at least three entries for each case (deliberation and recommendation, decision by the Minister, draft of the official decree); second, the files and the personal notes left over by the Secrétaire of the Bureau or its members (F 12 662-670 and 685-704); and third, the Avis des Députés (F 12 693-724) that provide a precise description of each project, and add a cross-regional or sectoral perspective, and references to similar cases submitted earlier to the Bureau. and the official justification that backed them up. In addition, we also collected additional variables that we used for various controls: each project is identified in terms of localization and jurisdiction, date, and industry. Lastly, we identify with a dummy variable the ventures that targeted luxury goods and those where a foreign entrepreneur or a nobleman was a party. 51 Table 1 summarizes the main features of the applications in our sample. Close to 10 per cent of them are requests based on innovation. The remaining cases are industrial ventures in a broad sense-they cover natural resource exploitation projects, such as mines or new crops; all kinds of factories, mainly in textiles, metals, and earthenware; and some services (for example, transportation and warehousing). Out of this total almost two-thirds are new ventures; the remaining third are submissions for a renewal or an extension of an already granted privilege.

Rubber-stamping: can we trust the data?
Before we move forward and analyse how this industrial policy was conducted, however, we have to consider the possibility that the description of individual projects might have been endogenous to the final decision of the Bureau. Indeed, we cannot directly observe the intrinsic characteristics of projects, neither can we rely on the descriptions provided by the entrepreneurs themselves: they typically envisaged a grand and glorious future for their firm, while also harbouring preconceptions about what the Bureau wanted to read and support. Last but not least, the Bureau itself could have been only a façade behind which rents would have been handed out 'as usual', following external factors such as court politics or local clientelistic deals. However, for the following three reasons, we do not think that our database is strongly biased, and hence we believe that there is no endogeneity flowing backwards from the final decisions to the project descriptions.
First, as stated, the whole process at the Bureau was structured to gather as much information as possible on each project, a priority that inevitably came at a cost. The decision procedure was thus 'open' in the sense that, in both colleges, 52 the parties were provided with all the available information and all competing opinions: they thus had to argue publicly given the arguments brought forward by all other participants in the process. Rule-based, collegial deliberations thus worked de facto 51 To our knowledge, this article is the first ever systematic analysis of the distribution of privileges for manufactures by the Bureau. Deyon and Guignet, 'Royal manufactures', explicitly denied that any quantitative approach might be viable on this subject. The closest approach to the one adopted here, though applied to a very different institutional and political context (namely eighteenth-century England), is in Bogart, 'Turnpike trusts', and idem, 'Glorious Revolution '. 52 This was also true for the local representatives of the Ministry of Finance in the provinces (the Intendants): when their reports were deemed incomplete or biased by the colleges they were asked to deepen their investigations and to provide more informed opinions.
as an instrument of mutual control. At the same time the diversity of professional and social backgrounds (noble vs. commoner, Parisian vs. provincial) did not leave much room for a powerful esprit de corps that would have arisen from the collective, implicit adhesion to a given coherent set of legitimate policy aims. In another investigation of these data we have actually shown that the final decision was indeed aligned with (or predicted by) the opinions voiced by the key participants during the deliberations inside the Bureau, and on the actual, substantive arguments they raised (for example, the actual degree of innovation represented by the project). 53 Second, this rather complex procedure operated in an environment in which rent-seeking politics were played out in full daylight. Discounting the intrinsic value of the formal rules of the Bureau on a prima facie basis would thus beg the question of why this sprawling patrimonialist power system would have invented a machinery so entirely at odds with its own modus operandi. 54 In a regime devoid of parliamentary or media oversight, there was no compelling reason to demonstrate accountability.
Lastly, we systematically checked in the archives of the Contrôle Général for possible signs of outside influence or for signs that some entrepreneurs could actually bypass the Bureau. First, there is no trace in the archives of the Bureau of any demands or pressure by external stakeholders, or patrons, on its officials. Neither did we find any written instructions from the Ministry of Finance, or any other Ministry, pertaining to individual claims under investigation. Second, and more importantly, since most applications were addressed to the Contrôleur Général, that is, the principal of the Bureau, we have systematically checked its remaining correspondence, either with other officials or with private persons. 55 While we found hundreds of applications, acceptances, and rejections of personal rents and appointments each year, only four letters concerned manufactures: in each of those cases, the Minister refused to intervene in the process.
In 1739, for instance, the Contrôleur Général received a letter from the aristocratic Princesse d'Armagnac, certainly a person with easy access to the Court, who soon received a most elegant answer: I received the letter you did me the honour to write in favour of the glassware entrepreneurs in Sevres. These men have long been serving badly the interests of the King, those of the public as well as their own ones. It is therefore very desirable that the Company can pass into the hands of people who would make more out of it. In addition, this case being now in the hands of the commissaries of the Conseil [du Commerce], I cannot decide anything by myself and would not do anything that would oppose the free course of justice. I beg you to be convinced of the respect with which I have the honour to be, Madam, your very humble servitor.
We do not argue of course that the Bureau was entirely immune from pressure and that attempts at manipulating it were always foiled. There are many examples of 53 Beuve, Brousseau, and Sgard, 'Modern bureaucracies'. 54 Antoine, coeur de l'État; Parker, Bureau of Commerce; Schaeper,French Council. 55 To check for pressures that might have been exercised during the process, we reviewed the remaining correspondence from the Ministry of Finance over the period (AN, G 7 31-61). We also checked its incoming correspondence: first, from the provincial Intendants (index G 7 71-531); and second, from private persons (G 7 601-602). We also checked the official requests sent to the Ministry by individuals (that is, the 'Placets'; G 7 675-680). bribes being handed out, from the very bottom to the very top of the Contrôle. 56 We only claim that decision-making regarding privileges to firms was not systematically captured, so that the de jure rules of decision-making were reasonably aligned with the de facto rules, that is, that privileges were granted on the basis of the collectively acknowledged characteristics of each real-world project. Moreover, if this rule had been systematically violated, that is, if privileges had been handed out on other grounds, say princely caprice or court politics, then there is good reason to believe that econometric testing would not reveal any stable, significant set of policy preferences.

Privileges: a menu of franchise and rents
The 'menu' of privileges from which individual applicants could draw can be grouped into four main blocks: a legal right to operate a business, a territorial exclusivity, personal tax reductions, and custom duties exemptions. On top of this, there was a further set of miscellaneous franchises that were rarely requested or granted, and so they are ignored in the rest of this article. 57 This classification has the advantage of allowing for a form of ordering within each of these four blocks, thanks to the relative homogeneity of the respective privileges. We shall thus use this ordering as a 'quasi-metric', or a proxy for a measure of the 'intensity' of the privileges being requested and/or obtained by each project. This 'intensity' reflects the number of privileges granted when they are cumulative (for example, tax cuts) or the scope of the privilege when they can be ranked (see Table 2).
Block One submissions first included the application for a legal status, allowing him to establish the firm; that is, to begin producing and selling goods. Then, the investor could ask for a 'privilège exclusif' which provided more institutional leverage but generally implied compliance with statutory regulations (especially in the textile industry); his firm would often be called a 'Manufacture'. Finally, there was the status of 'Manufacture Royale', which added a degree of grandeur in a society where symbolic display was a tangible resource. 58 Hence in this first block the 'intensity' measure takes (quite intuitively) the form of an ordinal variable equal to 0 when no specific right of production was granted, 1 when the Bureau authorized production of the good or the service (RightToProduce), 2 when the Bureau awarded the status of manufacture (Manufacture), and 3 in the case of a Manufacture Royale (RoyalManufacture).
Block Two addresses territorial exclusivity. This is the monopolistic feature that has attracted a lot of attention, primarily from Ekelund and Tollison,and Root. 59 However, these exclusivities came with two important caveats. First, their spatial 56 Schaeper,French Council. 57 Among these miscellaneous benefits are, for instance, the possibility for noblemen to take part in the project without forfeiting their nobility; subsidies or loans (very rarely) granted by the Crown; direct access to the consumer market, thus de facto bypassing the merchant guilds; or when there was a perceived risk that opposition to the project could coalesce around the local or provincial jurisdictions, the Bureau could decide that any future dispute would be settled by the central administration.
58 See Bondois, 'Organisation industrielle', for a presentation of these types of firms: the Manufactures Royales, which were private businesses, and a limited number of Manufactures d'État (these state-owned enterprises belonged in fact to the royal domain and were not operating under market constraints). Les Gobelins are the best-known example. See Horn, Economic development, for a broader and more analytically developed discussion of the broader institutional context in which firms and entrepreneurs operated. 59 Ekelund and Tollison, Mercantilism; eisdem, Politicized economies; Root, Fountain.  extent was generally limited to a perimeter of less than 20 or 40 kilometres (that is, five to 10 lieues, according to the old units) around the firm; sometimes it covered a given province, and most rarely a monopoly franchise over the whole kingdom. In this latter case, the privilege concerned a technological innovation (possibly imported) and so resembled a modern-day patent. The second caveat is that (except for these quasi-patents) exclusivity did not imply market closure, but only covered the ability to produce: outside competitors could enter the local market, so that the competitive advantage for the protected producer was a function of transportation costs and internal tariffs. Besides, the way in which exclusivity provisions were discussed by the Bureau reflects the clear assumption that they were exceptions to the rule of market competition: their aim was either to guarantee to the entrepreneur a given income stream, or to limit the pressure on scarce local resources, such as wood or fresh water. It should also be noted that the monopolies that were granted by the Bureau had no fiscal counterpart: the privileges did not serve as an instrument for the administration to extract revenue indirectly from the economy, as in the case of tax farms, salt mines, or the tobacco monopolies. The derived intensity variable Exclusivity again takes the form of an ordinal variable, set to 0 when no territorial exclusivity was granted, 1 when the producer's monopoly was strictly local, 2 when it extended to the whole province, and 3 when it covered the whole country. Block Three consists of personal tax exemptions targeted at the entrepreneurs and their business partners, occasionally skilled workers-or ouvriers. These exemptions could apply to taxes in kind, such as serving in the local militia or hosting soldiers from the royal army (intensity variable Taxes take value 1); or, more rarely, breaks could also extend to monetary taxes on income flows, such as the taille (value 2).
Block Four includes tax exemptions that directly affected the competitive position of the firm; hence they presented a potentially much more distortionary character than the previous categories. This block includes custom duties on imports and exports and, more importantly, internal tariffs and tolls that proliferated between provinces, at the entrance of cities, or along roads and rivers. We differentiate here between taxes on inputs (the ordinal variable CustomTax takes value 1) and on outputs (value 2).

No easy access to privileges
The first evidence that emerges from our dataset is that success was in no way guaranteed: a significant proportion of submissions were rejected (28.5 per cent) and an equivalent share (29.2 per cent) received less than initially requested. The balance (42.3 per cent) corresponds to fully successful submissions. Figure 1 also shows that entrepreneurs who submitted applications to the Bureau had little chance of obtaining benefits in all four blocks of possible privileges. There was a clear pattern of only granting two 'components': a status and either a territorial exclusivity or a tax cut on income and/or custom duties. Only 4.8 per cent of the claimants were able to obtain privileges pertaining to all four categories. This observation is also reflected in the table of correlation (see online appendix table A1) which shows that there is only a weak correlation among the four categories of advantages granted, ranging from 0.200 to 0.498. This should be taken as a first indication that the different sets of privileges were indeed granted on the basis of alternative justifications given the heterogeneity of the projects. For instance, figure 2 highlights the fact that the title of Manufacture was granted in only half the cases, and most projects that applied for a territorial exclusivity were simply not granted any such benefit. When they did receive one, it was mostly local in nature, and national exclusivity was clearly exceptional (7.9 per cent of cases), as already mentioned. The personal tax exemptions that were most frequently awarded were the less costly ones (civic duties), and even these were granted to only one-fifth of the projects. Finally, these tax cuts were more common than exemption from (internal) customs duties (more than 80 per cent of applications were rejected).

Policy objectives
As we coded all the substantive arguments brought forward during deliberations, we identified up to 30 different arguments (see online appendix table A2). Some of them are raised frequently, and others hardly ever. We take the 10 arguments most commonly acknowledged by the Bureau as reflecting, in practice, the main policy objectives that actually drove its decisions (see table 3).
These arguments, ranked by their frequency of occurrence, are dummy variables set equal to 1 when the Bureau recognized the quality of products (Quality), the potential to reduce imports or contribute to exports (TradeBalance), the contribution of the projects to the development of the local economy (LocalEconomy), the benefits that could be expected for consumers in an economy characterized by supply shortages (Consumer), the requested capital to start the project (FixedCosts), the contribution to technological development-through invention or the importing of a foreign technology that was not domestically available-(TechnicalInnovation), the ability of the projects to develop local resources such as commodities and minerals (ValLocalResources), the positive impact on employment of the poor (SocialEmployment), precedents (PriorPrivileges), and the need to attract specialized labour, often from abroad (AttractLaborForce).
It should be noted, however, that these 'policy objectives' were not stated by the administration ex ante: there was no 'call for projects' that would have invited entrepreneurs to submit projects targeting any specific policy objective. Entrepreneurs selected from an informal list of potential outcomes that had been de facto recognized as legitimate motives for public support. The fact that intermediaries might have helped them to prepare their submission, as suggested by Schaeper, is not problematic for our perspective, as long as it is not established that these were in fact patrons. 60 We can thus see in table 3 that this list of revealed policy aims is larger and far more diverse than is usually recognized by the common discourse on mercantilism. Quality of products and import substitution are in fact the most frequent, though development of the local economy is the third most common policy goal at this point. An interest in local employment (typically the poorer part of the population) or improving exploitation of scarce natural resources is also regularly mentioned as a valid argument, both in the submissions and the appraisals of the Bureau. A simple cross-section analysis of how privileges were distributed among classes of projects allows for a first, simple measure of where the Bureau concentrated its resources: that is, which characteristics (such as Quality or TradeBalance) attracted more support, as a whole, than others.

Identification strategy
We now estimate different versions of the following equation: where PRIVILEGES i is the dependent variable for each individual application i, ARGUMENTS is our vector of independent variables that includes the 10 60 Schaeper, French Council. 17.6 -3.9 11.8 -9.7 Regional exclusivity 9.7 13.6 11.  1 0 -5.8 1.7 -4.1 11.4 5.6 5.7 -0.1 6.9 1.1 3.9 -1.9 23.5 17.7 17.6 11.8 Notes: This table provides descriptive statistics about differentiated frequencies of granted privileges depending on the arguments brought forward by the Bureau to justify its decision (compared to granted requests; the differences are provided in the columns in italics). For instance, when the argument Quality is mobilized (that is, in 81 out of 267 cases), the probability of obtaining the status of Manufacture Royale is equal to 44.4. This corresponds to 11.4 points more than the frequency of similar privileges over the total sample of granted privileges (that is, 191 out of 267 cases). As a second example, the probability of obtaining national exclusivity when the Bureau recognized the innovative character of the project (TechnicalInnovation, 35 out of 267 cases) is equal to 45.7, that is, 34.7 points higher than in the sample of granted requests. Source: As for tab. 1. most frequent positive arguments put forward by the Bureau to justify granting advantages; and Z i is a vector of control variables. Z i contains dummy variables reflecting whether the firm intends to produce luxury goods (Luxury) and whether the firm is to be managed by foreign entrepreneurs (Foreigner), as well as the date of the decision (Year), in order to check whether our results might be driven by a time trend. Controls also take sector-group dummies into account to control for possible effects due to sector preferences. Ideally, we would also like to consider fixed effects due to regional dimension, but the requests for privileges are from 38 different provinces, and thus we cannot introduce region dummies without prohibitively reducing the degrees of freedom of the (ordered) logistic models. To limit the potential impact of the problem and to account for potential heteroskedasticity and auto-correlation of the error terms within geographical areas, we cluster standard errors at the regional level in all of our regressions. While keeping our set of explanatory variables unchanged, we run estimates using different dependent variables so as to extract maximum information on the way policy-makers adjusted means (privileges) and ends (their preferences). As a first step, we investigate the likelihood that the applicant will obtain a privilege pertaining to one of the four blocks, depending on the character of his business project. Hence, we construct a first dependent ordered variable, Obtained, which takes values 0 to 4 and stands for the number of different blocks of privileges among which the claimant obtained at least one individual benefit. This corresponds to model 1 in table 4. Then, we look separately at each of the blocks and assess the determinant of the 'intensity' of the granted privileges, as described above. This leads to models 2, 3, 4, and 5 in table 4. Finally, to observe if some features of the projects have a specific impact, we run logit estimates for each dummy variable corresponding to each individual advantage that could be granted by the Bureau du Commerce. This results in models 11 to 13 for the status of producer (table 6); 14 to 16 for territorial exclusivity (table 7); 17 and 18 for per capita tax cuts; and 19 and 20 for custom duties exemptions (table 8).

Empirical results
The results of model 1 in table 4 go a long way towards revealing the hierarchy of policy objectives of this post-Colbertist, mercantilist policy. Project features that received the broadest support were, in declining order, those contributing to: first, technical innovation and diffusion; second, local economic development; third, a stronger trade balance; and fourth, increased supply and consumers' welfare. This is confirmed by the results of model 6 in table 5, which provides marginal effects. Indeed, it indicates that the probability of not obtaining any privileges (Pr(Obtaining) = 0) is 30.4 per cent lower when a project is considered innovative (TechnicalInnovation = 1), when all other independent variables are set at their mean value. This probability is respectively 24.4 per cent, 20.3 per cent, and 16.1 per cent lower for projects that favour local development (LocalEconomy), import substitution (TradeBalance), and consumers' welfare (Consumer). Further to these four main policy objectives, we observe, though with a less significant coefficient and lower marginal effects, a positive impact of product quality, and possibly the presence of high fixed costs or the exploitation of local resources. Our results then have implications for two common statements of the literature on mercantilist policies. First, there is no indication that attracting foreign entrepreneurs or technicians to the kingdom was a priority during the period under review (see Foreigner, and AttractLaborForce). Second, and more important, the control variable Luxury remains remarkably silent in those first regressions, as well as in all subsequent ones. Hence, this post-Colbertist industrial policy of the early eighteenth century was not driven by a preference for luxury good producers, contrary to what Heckscher and a long line of authors have argued. If anything, this factor has a negative impact on the probability of receiving privileges. In fact, this policy line is not only revealed by econometric testing, but surfaces quite often in the minutes of the Bureau. For example, in 1731, a submission by a silk producer from Lyon was rejected on the ground that it would be better for the manufactures of Lyons if workers with superior qualifications were to exercise their talents in ways that would help them to meet more readily the dominant tastes in France, where people preferred fabrics from Holland or India to French ones, because they were softer, lighter, and less expensive than those produced in Lyon. 61

Blocks of privileges and their 'intensity'
On top of these policy preferences was, first, evidence of a bureaucratic capacity to align policy tools and ends in a consistent manner. Models 2 to 5 in table 4 show in greater detail how the four main blocks of privileges were targeted at differentiated types of projects; this information is then confirmed and extended by tables 6-8, which evaluate how the determinants of the 'intensity' of support, within each block of privileges, were affected by the underlying preferences of policy-makers. We may thus move from a more quantity-based approach of the 61 AN, F12/80, Bureau, 15 Feb. 1731. 'volume' of privileges granted to a more nuanced understanding of how they were differentially mobilized by the Bureau. TechnicalInnovation is typically supported by two instruments: the right to produce (though only rarely with the status of Manufacture Royale), and national exclusivity. 62 As highlighted in model 8 (table 5), the probability of receiving a privilege of territorial exclusivity was 68.7 per cent lower when the Bureau did not recognize the presence of technical innovation. However, it should be noted that these two benefits were awarded jointly, so that this class of project was not 62 Technological innovation came in two contrasting forms: either as a stand-alone submission, or in the context of an industrial venture. In the former case, the applicant claimed for an exclusive and tradable right to exploit. When the enforcement of this right appeared to be difficult, the applicant could receive a grant from the Crown and his innovation would fall in the public domain (10% of the innovations in our sample; see Scotchmer, Innovation, on this practice in premodern Europe). In the case of an industrial venture, the claimant usually received a right to produce, though not a statute of manufacture, combined with a national territorial exclusivity. Often, a privilege of direct distribution was added as a way to protect his innovator's rent from intermediaries. Hilaire-Perez, 'Invention'; idem, L'invention. associated with an intellectual property right in the modern, impersonal sense. National exclusivity required the establishment of a factory. LocalEconomy should be understood as the overall development of productive capabilities in a given district, with its multiple impacts (for example, on the demand for labour), or in terms of providing intermediary goods or fostering demand for existing business. This important (and often neglected) dimension of mercantilist policies is typically supported by, inter alia, the status of manufacture (2) and some territorial exclusivity (6), though only at the local level (models 14 to 16, table 7). 63 Supporting the trade balance-typically considered the key mercantilist targetoften motivated an award of the status of Manufacture, a regional (or provincial) production monopoly, and substantial support in the form of custom duty relief, which directly affected the production function (respectively the investors' personal revenue). Here again, the marginal effects indicate that the probability of not benefiting from custom duties exemptions was 14.6 per cent lower when there was an expectation that the project would reduce imports. Although this is not the most highly valued policy objective of the Bureau du Commerce, it motivated the types of support that were already perceived, in those times, as most distortionary. Quality of production was clearly valued by the Bureau du Commerce, though this characteristic did not attract a typical set of privileges; discussions at the Bureau indicate that it was only a factor when awarding the status of Manufacture Royale.
This series of elements thus confirms both the presence of a clear hierarchy of policy preferences in the decisions made by the Bureau, and a bureaucratic capacity to align tools with preferences in a consistent manner over a 20-year period. Furthermore, we also see that some policy instruments were designed to attain a more diffuse set of goals whose impact does not show up econometrically. Concern for the development of local resources, for example, crops up regularly in the minutes of the Bureau and in the Avis des Députés, though support for this policy objective was rather diffuse in terms of the instruments used, and is thus not captured by our tests.

Fine-tuning
We now focus on policy objectives that were less important than the ones discussed in the previous two sections. What the analysis highlights is the fine-tuning of the alignment of the granted privilege with the objectives. Second-tier goals bring less crucial benefits, and again one can see how well they are aligned. As an example, the recognized quality of the output was the essential argument for obtaining the status of Manufacture Royale (see model 13 in table 6), but it did not bring additional benefits. Note also that the contribution to trade is not crucial for this distinctive status, confirming both the error of the dominant vision of the goal of these establishments, and the apparent high marketing value of the government's endorsement of the quality of production.
Other illustrations of the capacity to fine-tune policy targeting are also seen, for instance, in the determinants of per capita tax cuts. They are driven by the need to attract skilled labour and by the existence of precedents. In the latter case, precedents play essentially on non-monetary taxes (CivicDuties; see model 17 in table 8). This is a further econometric illustration of the willingness of the Bureau to limit distortion of competition by preferring to grant privileges with a low impact on competition, and to grant them on a fair basis-that is, giving similar bonuses to substantially similar projects.

Endogeneity issues
Technically, our empirical strategy defines the outcome of the decision process as a dependent variable and characteristics of the projects for which the entrepreneurs sought privileges, plus characteristics of the entrepreneur and his business, as the explanatory variables. The positive aspect is that we do not have endogeneity issues related to possible reverse causality: the outcome of the decision derives directly from the decision-making process, once the various arguments by the parties on the basis of prospective assessments concerning the nature of the project have been put forward. However, a problem may arise from the fact that all explanatory variables are probably endogenous. Therefore, the variables on the right-handside of the equation would be both explanatory and explained, which may lead to misleading results. For instance, a technical innovation can be the source of more productivity and differentiation, enabling lower costs optionally associated with higher quality. Therefore, our variable TechnicalInnovation would be positively and significantly correlated with our variables Consumer, Quality, and TradeBalance. However, the statistical observation shows that the arguments used to justify the final decision are only weakly correlated (see online appendix table A1). The average correlation between the arguments is 0.096 and the highest correlation between two arguments is 0.381 (between Quality and TradeBalance). Moreover, the most frequent number of arguments used is two (23.6 per cent of cases) and no cases used all 10 arguments (the highest number of arguments put forward in a single case is six and this affects less than approximately 2 per cent of the sample, that is, five cases). This suggests that the arguments were mobilized only when they were considered real characteristics of the project and that, de facto, endogeneity issues are not particularly acute.

V: Conclusion
We have reconstituted and coded the bureaucratic process by which the Bureau du Commerce issued 267 decisions out of a total of 281 requests for privilèges de manufacture submitted by private entrepreneurs between 1724 and 1744. On that basis, it has been shown that this rule-based, information-driven open process of deliberation could indeed differentiate between policy aims and consistently target them with different packages of franchises and rents. A hierarchy of revealed policy preferences was thus identified. These mercantilist policy-makers were targeted in the following order: technical change, local development, the trade balance, and consumer welfare. We take these results as an indication that decision-making was relatively free from capture and that it also avoided the symmetric risk of bureaucratic confusion and erratic decision-making.
The way the Bureau du Commerce supported individual manufacturers thus reflects the attempt, by an emerging rule-based, impersonal bureaucracy, to weigh as 'from the outside' the economy's spontaneous development in directions deemed conducive to the common good; whether defined as industrial catch-up, import substitution, or otherwise. In other words, the king's men relied on signals and incentives to individual economic agents, who were expected to respond in ways that indirectly contributed to the ultimate objectives of the policy-makers. In this attempt the Bureau relied on instruments that are typical of modern governments: information, technical competencies and procedural consistency, and collegial decision-making.