Introduction

1The economic development that characterizes the second half of the 19^th century goes hand in hand in most countries with an important increase of the number of bankruptcies (Lester, 1995; Mann, 2009; Hautcoeur and Levratto, 2010; Telfer, 2010). This phenomenon apparently cannot be distinguished from the other aspects of the boom of capitalism that took place during this period. The increasing role of exit is so probably a component of the Schumpeterian creative destruction which describes the “process of industrial mutation that incessantly revolutionizes the economic structure from within, incessantly destroying the old one, incessantly creating a new one” (Schumpeter, 1994, pp. 82–83). In the institutional conception of this movement, bankruptcy is thus a social mechanism designed to organize exit, giving support to contractual exchange. Indeed, bankruptcy laws and procedures provide at that time an excellent dispute settlement institution (Bignon and Sgard, 2007), and bankruptcy is at the heart of the market functioning and contribute to the reliability of the economic system (Hautcoeur and Levratto, 2007).

2Since the Ancien Regime, lawyers, scholars and legislators repeatedly proposed to change the insolvency laws in order to improve the functioning of the economy, to attract foreign firms and investors, and to increase national competitiveness (Mascret, 1863). More recently, some proponents of “law and economics” (e.g. La Porta et al., 1998) reactivated the old lawyers’ claim that the impact of bankruptcy law on national economic performance is important. Empirically, their argument is mainly based on the comparison of allegedly autonomous legal traditions, with little consideration for actual practices or for legal transformations over the long run. Our position differs from their: as in the “20^th century divergence” debate in international relations (James, 2001) or in financial systems (Rajan and Zingales, 2003), we consider that bankruptcy laws in 19^th century Europe were much more comparable among countries, than commonly assessed, and that they varied frequently in a similar manner (Sgard 2006, Hautcoeur and Di Martino, 2013). More importantly, and in accordance with Fridenson (2004), we consider that closing firms involves economic, social and human costs which go beyond their frontiers and that these externalities impose bankruptcy to be legally organized.

3A legal system is not a simple set of rules allowing for perfect and transparent contracts. Rather, it is made of several components ranging from the laws to the persons (or entities subject to the law), the lawyers, and finally the judges (within courts). This complexity cannot be tackled by looking only at the “law on the books” since it would lead to a biased view of the whole phenomenon (Macaulay, 2005). Bankruptcies do not escape this drawback. This is the reason why this research rests upon a “law in action” conception to identify the influence of economic, social, and legal contexts on the treatment of firms’ failure. It leads us to examine the interaction between legal rules and practice to explore the possibility that court practices could precede and not always follow legal changes, and that, even in a civil-law country like France, court practice can vary substantially from a direct application of the law.

4From a reconstitution of data published in the Comptes Genéraux de l’administration de la justice published by the French Ministry of Justice from 1820, we consider the different possible causes of the sharp increase in the number of corporate bankruptcies observed along the last two thirds of the 19^th century. We stop the analysis in 1913, a year frequently considered as the genuine end of the century. Our results show that the changes in the law and the judicial innovations that aimed at facilitating access to legal procedures are not always at the origin of the evolution of the number of cases. Instead, an intricate relation appears between the changes in the commercial code, the practices in the commerce courts and the financial characteristics of the insolvent companies on the one hand, and the changes in the number and results of the procedures on the other. We show thus that the continuous interactions between formal rules and practices are much more powerful to explain the evolution of the bankruptcy regime over the 19^th century than the unidirectional perspective going from the law to the practices.

5The remaining of the paper is organised as follows. Section 1 presents the evolution of French bankruptcy law during the 19^th century. Section 2 briefly describes the methodology and the dataset. Section 3 describes the evolution of the number of bankruptcies in France whereas sections 4, 5, and 6 examine alternative specific explanations (respectively local specificities, court practices and economic context) to the increase in the number of procedures. The conclusion contains some general proposals to support a renewed conception of bankruptcies.

1. Legal history of bankruptcy

6The history of bankruptcy law over the 19^th century is the object of numerous research and publications (Hautcoeur and Levratto, 2010 provide a comprehensive description).

7Bankruptcy law occupies around one third of the 1807 Code du Commerce, which defines some of its major characteristics for the entire century. Since the Ancien Regime, bankruptcies are under the responsibility of specific commercial courts, composed of judges elected among notable merchants in each resort. Appeals are judged by civil courts, which makes commercial courts a specialized jurisdiction under the responsibility of civil courts. Nevertheless, commercial courts are quite independent in practice, and commercial law is quite distinct from civil law. Similarly, bankruptcy procedure is separated from the judgment of its possible delinquent or fraudulent origins. These last cases are the responsibility of penal courts under the name of banqueroute. Commercial courts are exclusively reserved to merchants (commerçants), defined as every person (natural or legal) usually buying and selling goods or services, but excluding farmers, professionals, wage earners and civil servants.

8Bankruptcy (faillite) is equivalent to the default of payment (cessation de paiements), but the procedure is started only by a judgment stating that a person or a company (société) is insolvent, i.e. is unable to repay the amount due to claimants. The procedure can be started either at the initiative of the bankrupt himself (by producing his balance sheet to the court, so the usual name dépôt de bilan), by his creditors (requête) or by the court (d’office). The bankrupt is supposed to be jailed, and the judge to name a trustee (syndic) responsible for the management of the procedure in the interest of the creditors. This supposes either finding a composition (­concordat), that is an agreement with the creditors allowing the perpetuation of the firm (usually with a restructuring of debts) or organizing the complete liquidation of the debtor’s estate (union). The procedure starts with a verification of the assets and liabilities of the bankrupt firm, which can lead to an immediate stop if assets are insufficient to cover the costs of the procedure. In such a situation the case is settled as “shortage of assets” (insuffisance d’actif). If the assets are sufficient, creditors are grouped into in secured and ordinary ones. Creditors holding mortgages, other security or special priority are reimbursed first. The remaining creditors pertain the masse, which decides on the concordat on a majority basis.

11Further changes go in the same liberal direction. In 1856, the composition by “relinquishment of assets” (concordat par abandon d’actifs) gives the possibility to the bankrupt to abandon its assets to the creditors and start a new business without having to fear that his new assets will be seized in order to compensate his creditors. In 1867, the prison for debt (contrainte par corps) is abolished (Levratto, 2009). This change is part of a political evolution favouring personal liberties which has also appeared previously in the decreasing use of jail in bankruptcy procedures (Guyot & Raffalovitch, 1901: 135). In 1903, the possibility of a full réhabilitation of the debtor is facilitated.

12More importantly, following two tentative provisory legislations under the special circumstances of the 1848 revolution and the Franco-Prussian 1870 war, a special procedure called judicial liquidation (liquidation judiciaire) is created in 1889. It allows the settlement of bankruptcy cases without the still infamous name of faillite. The new procedure is reserved to bankrupts of good faith whose failure results from unlucky exogenous circumstances, and who have initiated the procedure. Then it is supposed to end-up in a concordat, although union and insuffisance d’actifs are also possible.

13In sum, over the second part of the 19^th century, the evolution of bankruptcy law in France follows a liberalizing pattern, like in most other European countries (Sgard, 2006). The requirements of economic efficiency, especially the need to provide for the maximum repayment to creditors, for the continuity of all existing “good” firms and for the fresh start of unlucky entrepreneurs, becomes more important than the need to sanction any borrower unable to repay his debts and to exclude him from the community of merchants. Nevertheless, ex-ante incentives to repayment remain high since concordat is not a right (so that many bankrupt merchants lose entirely control over their assets) and faillite remains somewhat infamous. As asserted by Ripert (2004), French bankruptcy law has always accompanied the development of capitalism.

14This increasingly tolerant and protective set of legal rules organizing firms’ default gives excellent reasons to entrepreneurs to file for bankruptcy. However, the causes of the sharp increase in the number of insolvencies observed over the second part of the 19^th century remain unclear, since it is not (as we will see below) driven by the procedures initiated by the debtors. We propose to disentangle this question considering practices of the French courts through the prism of the law. We will concentrate on long run changes from 1820 (or at least 1840) to 1913 and on the broad changes in the law previously presented.

2. Methodology and empirical strategy

16Another question emerges from a long run analysis. It deals with the endogeneity of bankruptcy law. Both political economy theory and archival records suggest that all participants in the bankruptcy procedures (mostly merchant communities organized in chambers of commerce or more specific lobbies, and the legal professions) try to modify the law in their favour. As discussed previously, the law has actually been modified, mostly in the direction favoured by the merchant community, although the timing depended on many circumstances.

17We propose to explore different possible causes of the sharp increase in the number of bankruptcies by looking at data covering a period from 1830 to 1913. The precise date at which we begin the analysis on particular topics depends on the source used.

3. The evolution of the number of bankruptcies: the law does not explain it all

20Three major points worth are worth highlighting:

• The rapid increase in the number of bankruptcies from the 1840s to the early 1860s resulted mostly from the choices of debtors, since they initiated at least 60% of the total number of new bankruptcy cases (graph 2). This probably results from the 1838 law, which aims to facilitate the recourse to the judicial treatment of insolvency instead of negotiating private settlements (unfortunately, we don’t know the proportion of debtors’ initiated procedures before the law).

• Between 1865 and 1869, a dramatic change takes place within a few years: the proportion of bankruptcies initiated by the debtor decreases from 57% (1865-66) to 45% (1868-69). According to many contemporary commentators, it results from the abolishment of prison for debt, which suppresses the incentive for debtors to file for bankruptcy (dépôt de bilan) as they did it earlier mainly because this procedure allowed them to escape it (see below).

• The 1889 law distinguishes between liquidations judiciaires, which can only be started at the initiative of the debtor, and faillites. It is followed by a drop in the proportion of faillites started by the debtor (from 40% in 1888 to 28% in 1890, see graph 2). But at the global level, the proportion of bankruptcies (faillites + liquidations judiciaries) started by the debtor increases from 40 to 50%. The introduction of the liquidation judiciaire is a major innovation since it permits an increasing number of debtors to file for bankruptcy before their creditors sue them. The best debtors then likely switched from faillite to liquidation judiciaire; the overall number of procedures increased in the short run but was stable in the long run. It may be that it had at that time become impossible to attract towards the judicial system those still preferring private settlements, or maybe there were few of them after these legal changes, something which would suggest a quite efficient judiciary organization of bankruptcy, contrary to those interpretations which argue for a belated French bankruptcy law.

Sources: Comptes généraux de l’administration de la justice (1840-1913). Authors’ computations.

4. Number of bankruptcies and local differences

21The evolution of the proportion of bankruptcies initiated by the debtors suggests that incentives brought by legislative change plays a role, but only up to a limit. Studying the regional variations of the bankruptcy data shows that legal practices are heterogeneous and that they probably differ according to the courts and the regions where they are located.

22Map 1 shows that the number of procedures initiated by the debtors scaled by the total number of bankruptcy strongly differ from one court to another. This heterogeneity characterizes the whole period under study even if one can notice a common trend, especially over the last years. Indeed, the maps show that the share of dépôts de bilans dropped all over France at the same time.

23After the enactment of the 1838 law, the regional pattern of the proportion of bankruptcies initiated by the debtors suggests that if this law introduces major changes, it mainly reflects practices already existing in Paris (map 1). Paris is a first mover in having a high proportion of bankruptcies initiated by the debtor. Actually, during the early 1840s, Paris is almost the only place with, to a lesser extent, Bordeaux and Colmar, where such bankruptcies dominate (77% of new bankruptcies in 1840). This Parisian specific pattern may result from the characteristics of the regional economy, but it may also result from the specific behaviour of the judges in the Tribunal de commerce de la Seine, who are considered as pioneers by several contemporaries. Indeed, Parisian judges use more than in other regions the contrainte par corps allowing for the imprisonment of debtors at justice or creditors’ request (see map 2), so much indeed that creditors from all over France try to bring their cases to Paris when possible. But the same judges also provide widely sauf-conduits (provisory liberation from jail) once a case is started (especially if the debtor initiates the procedure or is considered reliable or as having few reasons to escape) so that at the end few debtors stay in jail (see maps 3).

24Actually, some consider that the 1838 law was probably passed in order to “legalize” and extend the Parisian model i.e. more bankruptcies, and a higher proportion of them initiated by the debtor and conducing to more satisfactory public settlements through concordats instead of private settlements. In the following decades, the model was adopted in most French regions. This convergence then explains a substantial part of the nation-wide growth in the number of bankruptcies. This convergence was not a naturel process, but it resulted from the dissemination of “good practices” to the entire territory under the control of the Ministry of Justice and the Parisian Appeal court (which covers about one third of the national territory as measured by number of bankruptcies). The change introduced by the 1867 law does not substantially alter this growing homogeneity (table 1): it causes a decrease in the number of dépôts de bilan in every region, and especially in Paris where their share drops from 63 to 46% in four years, and the standard deviation remains at its low 1865 level.

Table 1. Proportion of dépôts de bilan in bankruptcies (27 judiciary ressorts)

	1840	1857	1865	1869
France	0,47	0,57	0,57	0,44
Average of regions	0,30	0,51	0,51	0,44
Standard deviation/average	0,47	0,24	0,18	0,17
Correlation with previous year		0,50	0,73	0,46
Minimum	0,08	0,30	0,38	0,28
Maximum	0,77	0,76	0,68	0,56

Sources: Compte général de l’administration de la justice.
Authors’ computations.

26This suggests that, if the law probably pushes up the number of bankruptcies, the heterogeneity among regions tells also a lot about the dissemination of the new procedures and the choices made by entrepreneurs, creditors and judges. Indeed, commercial courts’ specificities and the reaction of the persons subject to the law play a key role in the explanation of local differences. These geographical and institutional differences lead us to consider that, even in a centralized and civil law country like France, judges are important drivers of both the global evolution of bankruptcies and some of their spatial discontinuities.

5. Evolution of the practices and change in the procedures

27Another potential explanation for the rise of bankruptcies lies in the decrease in procedural costs. These costs are considered along the century (and in the historiography) as the most important obstacle to the recourse to the courts by debtors and creditors. They are mostly independent from the legal evolution itself. However, for many creditors, the most important barrier to resort to the commercial courts does not consist in excessive financial cost. Instead, it consists in the time needed to get a judgement. The long delay between the start of the procedure and the final decision entails uncertainty and loss of time, which in many cases threatens the very existence of the business under investigation. Various measures of the evolution of this delay can be computed from the available data, which does not provide the average length of the courts’ decisions.

29The second indicator of court efficiency is the age structure of the opened files at a given date. As shown in graph 3, the proportion of the pending cases at the end of a year opened less than 6 months before is fairly stable from the late 1870s to 1913. The creation of the liquidation judiciaire helps to keep the length of the procedure identical whereas the number of cases tremendously increased. Indeed, this procedure is more rapid than the faillites, which is not surprising since reducing the workload of the courts is one of the main reasons for creating this procedure.

30For these two indicators, the leadership of Paris and a few large commercial courts also appears when one looks at the geography of bankruptcies. The convergence in efficiency among courts (map 5), which is also at work throughout the century, may thus be part of the convergence in the number of bankruptcies discussed above.

31However, it may be misleading to consider all bankruptcy cases as similar, as the procedures they enter are quite heterogeneous in terms of potential duration so that composition effects may explain an apparent improvement in efficiency, especially given the rise of rapid judgments recognizing insufficient assets (insuffisance d’actif) throughout the period. One may fear that the decent aggregate performance of the courts reflect mostly that rise, and that the duration of the two “normal” procedures, i.e. concordats and unions, actually increases, discouraging firms from going to court. Actually, an ongoing complementary investigation based on individual bankruptcy files from the Parisian commercial court’s archives suggests a somewhat more optimistic view than these global statistics: the length of the median procedure decreases by almost half (from 7 to 4 months) from 1850 to 1899. Furthermore, if the time necessary to reach a concordat does not decrease (it actually slightly increases), the one for settling entire liquidations (unions) decreases sharply (from 16 to 7 months), which suggests the reduction in duration was not only the result of the rise in the number of insuffisances d’actifs. Taken together, these indications suggest that the decrease in the duration of the procedure attracts some debtors or creditors from private to judicial settlements, especially before the end of the 1880s. It likely contribute to the rise in bankruptcies over that period.

32A significant share of financial costs is related with the taxation of the official documents required at the various stages of the procedure, and with the payment of the syndic. In response to a constant demand from economists, from the Chambers of commerce and other business lobbies, the level of the taxes decreases several times in the early 1830s and during the second half of the century. In the opposite direction, some contemporaries complain about the increasing court’s expenses (Tessier du Cros, 1906). If no aggregate statistic is available on that subject, our sample confirms the decrease of the average value of the expenses of the syndics (including their own wage) scaled by the income they are able to bring in the bankruptcy purse. In sum, all procedural costs go down throughout the century, which can help explaining the surge in bankruptcy cases.

33Our analysis suggests that if the legal changes play a role in the evolution of the number of bankruptcies over the 19^th century, it does not explain this entire growing trend. Instead, it has to be coupled with other features, such as the convergence of courts’ practices, and probably with a reduction of procedural costs. The increase in the number of dépôts de bilan (as expected by the legislator) and a rise in the number of bankruptcies started at the creditors’ initiative result from various factors having to do with formal and informal rules. We also suggest that the various needs of different regions or types of firms require a more diversified menu of legal rules. This may explain why, although in the most important economic regions the 1838 law allows to reach a satisfying result, a new law in 1889 is considered as useful, and is maybe necessary in some places.

6. The composition of bankruptcy outcomes

34If the legal changes and the reduction of procedural costs were the only explanation of the rise in the number of bankruptcies, and if, simultaneously, the proportion of firms facing liquidity shortages were stable over time, one would expect the outcomes of bankruptcy procedures to reflect only the increasing recourse to the courts. This section aims to explore this possibility.

35The statistics on bankruptcy outcomes give a fairly pessimistic view of the evolution of bankruptcies (graph 4). The ratio of concordats to the total number of bankruptcies decreases sharply over the period. Paradoxically, at the end of the period they are not even the most common outcome for liquidations judiciaires. Whereas they represent more than 50% of the total number of bankruptcies in the 1840s, the share of concordats decrease to less than 20% in the 1870s and 1880s. Thanks to the liquidation judiciaire, it reaches about 20% in the 1900s (around 10% for faillites and 30% for liquidations judiciaires). At the same time, closures for shortage of assets (insuffisance d’actif), which represent 20% of all bankruptcies in the 1850s, rise up to almost 50% in the 1880s, a trend only briefly interrupted by the enactment of the liquidation judiciaire procedure. On average, they still represent the majority of bankruptcies in the 1900s (60% for faillites only). The other procedures end up either with the “relinquishment of assets” (abandon d’actifs) procedure or with the liquidation of all assets (unions). These two outcomes usually provide quite small dividends to the creditors.

36In this case, Paris is not leading the movement. Actually, the proportion of concordats in Paris in the 1840s is one of the highest in France, and it falls in line with other regions.

37One can wonder if this sharp deterioration of the bankruptcies’ outcomes is consistent with the facts we observed earlier and with the legal theory of the evolution of bankruptcies In comparative statics, such an analysis suggests that a more liberal bankruptcy law, especially when reinforced by a decrease in procedural costs, leads to the choice of a legal settlement by firms presenting a better (less distressed) financial situation, other things being equal. On the other hand, over time a more liberal law may lead to more risky behaviour and borrowing by firms, leading to a decrease in bankruptcies’ outcomes for creditors. The combination of an increase in the number of bankruptcies and a deterioration of their outcomes suggest that the second effect is dominant. This may also explain why the proportion of debtors’ initiated bankruptcies (dépôts de bilan) did not increase despite the liberalization. It is also consistent with the fact that the decrease of concordats and the rise of unions and insuffisances d’actifs appeared even among the category of bankruptcies initiated by the debtors (we cannot measure this directly but it results mechanically from the fact that dépôts de bilan still represent 45% of the total number of bankruptcies in the 1900s whereas the proportion of concordats remains below 20%).

38This transformation of the outcomes in a context characterized by a more and more intensive use of judicial procedures to close businesses is confirmed by our archival sample of bankruptcies. It shows that almost all dépôts de bilan ended up in concordat in 1850, when this is only the case for a third of them in 1899. At the very end of the century, the number of concordats scaled by the number of cases initiated by the creditors or the court decreases to almost zero.

39Except if the commercial judges tend to be more and more severe over the period, a behaviour that would be in contradiction with the liberalization of the law (not impossible but unlikely given the identity of the merchant-judges), the declining financial outcomes of the procedures (whatever their origin) can only be explained by a rise in the number of firms suffering liquidity shortage or a worsening of their financial problems. This suggests a study of the link between increased indebtedness and the rise in bankruptcies, which we cannot provide in this paper.

7. Conclusion

40This paper sought to explain the increasing use of judicial procedures to close businesses over the 19^th century and the growing number of possibilities provided by the legislator to debtors. We have examined different sources of these changes in the legal rules and practices. Thanks to the collection of data published along the 19^th century by the Ministry of Justice, we show that the number of bankruptcy cases reflects not only a change in the legal architecture organising the closing of businesses but also extra-judicial transformation of the French economy. Our main conclusion is that the rapid rise in bankruptcies observed is not only fuelled by the liberalization of bankruptcy law and the decrease in the procedure’s costs. Instead, we strongly support the idea that the changes in courts’ practices and firm functioning played a substantial role, most clearly observed through the 1838 and 1867 reforms.

41Our main conclusions are the following:

• The evolution of bankruptcy law partially entails the rise in bankruptcies over the major part of the 19^th century: the 1838 law has an important impact; the 1867 law on contrainte par corps is also influential, even if it was against the wishes of its promoters and maybe only in the short run. However, it is of no help to explain their stabilization after 1889. Indeed, the judicial liquidation authorized by the law of 1889 only has a regional impact on bankruptcy practices. This result contrasts with an idea largely resting upon the examination of aggregated data.

• Court practices strongly differ among regions, and the differences in courts’ efficiency and practice are a necessary complement in order to understand the variations in the number of bankruptcies.

• The evolution of the legal context is not enough to explain major changes in the composition of bankruptcies. This is particularly the case as far as the decrease in the proportion of concordats and the increase in the number of insuffisances d’actif are concerned. As a result, the change in the law cannot be separated from a change in the global behaviour of firms. These changes not only reflect their individual choices when facing liquidity shortage. Instead, they likely reveal a change in their financial structure and strategy, primarily an increased recourse to financial debt.

42Further work should allow to precise the respective roles of legal evolution and real economic changes, as well as to understand better the reverse causality going from economic and judicial changes toward legal evolution.