Navigation – Plan du site

Dealing with failure

Bankruptcy and insolvency in the English experience (1890-1939)
Gérer l’échec. Faillites et insolvabilité en Angleterre (1890-1939)
Paolo Di Martino
p. 137-165

Résumés

Cet article se penche sur le fonctionnement, dans les années 1890-1930, des deux procédures de faillite propres à l’Angleterre et au Pays de Galles. À partir de données statistiques nouvelles ou améliorées, il s’intéresse à l’arbitrage entre l’utilisation des solutions judiciaires officielles de règlement des faillites et le recours à des arrangements à l’amiable entre créanciers et débiteurs. Il montre ainsi que, pendant la période étudiée, les procédures officielles furent de moins en moins souvent choisies. D’autre part, elles étaient, semble-t-il, de moins en moins efficaces, en particulier, du point de vue des créanciers, en termes de niveau de dividendes obtenus. Le moindre recours aux procédures judiciaires peut certes avoir d’autres raisons que cette moindre efficacité ; mais ces résultats laissent au moins planer un doute sur les idées reçues qui considèrent que le régime légal anglo-saxon est forcément supérieur au régime continental en termes d’incitation à la performance économique.

Haut de page

Entrées d’index

Mots-clés :

commerce

Géographie :

Iles britanniques

Chronologie :

XXe siècle
Haut de page

Texte intégral

  • 1 . La Porta, R. & al., 1998; Levine, R., 1998.
  • 2 . The traditional “law and growth” approach stresses the former mechanism, while the latter has bee (...)

1A growing body of economic literature, known as the “law and growth” approach, argues that Anglo-Saxon legal regimes are better equipped than others to support economic performance1. The “law and growth” approach adopts a long-term perspective, suggesting that the reasons behind the relative efficiency of a country’s legal system are either factors such as colonization or invasions (determining the fact that its legal institutions belong to one or another of the various “legal families”) or mechanisms of institutional change which develop in the long term2.

  • 3 . See especially M. Bordo & P. Rousseau, 2006, A. Musacchio, 2008, and J. Sgard, 2006.
  • 4 . Di Martino, P., 2005a ; Hautcœur, P.-C. & Levratto, N., 2007.

2This “flirt” of economic analysis with the historical perspective has stimulated economic historians to turn their attention to the long-term relation between legal institutions and economic performance. This line of research has proved extremely fruitful, as historical analysis is now challenging the established wisdom provided by the “law and growth” approach, both on the actual differences between legal families and on their relative ability to foster economic performance3. In this field, the study of bankruptcy and insolvency laws and procedures is growing in popularity4.

  • 5 . Works by historians such as I. Duffy, 1985, J. Hoppit, 1987, and V. M. Lester 1995, mainly discus (...)

3In this context, our paper deals with the general functioning of these instruments in England between the 1890s and the 1930s. More specifically, it aims at assessing the level of efficiency of official devices, both in absolute and comparative terms: a perspective that, despite the abundance of the historiography, is still missing5.

4This paper is based on new or improved statistics and is structured as follows. Section 1 provides an overview of the available quantitative sources. The following three sections analyse the efficiency of English bankruptcy and insolvency laws and procedures using different perspectives: section 2 focuses on their usage, length, and level of dividends paid; section 3 looks at the structure of the productive system, and section 4 uses a comparative perspective. Some concluding remarks follow.

1. English data: An overview of the sources

5The first problem that one encounters in dealing with English statistics is the conceptual (and linguistic) difference between bankruptcy and insolvency. While in everyday language these two words are used as synonyms and, if anything, only suggest a slight difference in the seriousness of the problem (being insolvent indicates a state somehow less hopeless than being bankrupt), in the technical jargon, these two words are used with very specific meanings: bankruptcy refers to personal failure, while insolvency applies to companies. This difference remained purely theoretical until an act (53, 54 Victoria, c. 71) was passed in 1890 to deal specifically with corporate insolvency. After that, sole ownerships and unlimited partnerships remained subject to the bankruptcy regime, while joint-stock companies and limited partnerships fell under the insolvency law umbrella.

6Statistics about personal bankruptcy have been collected in England at least since the early 19th century, but only in a sporadic and unsystematic way. However, in the 1880s, profound changes in the law forced the government to take a much more active role in data collection and analysis. As a result of the 1883 Bankruptcy Act (46, 47 Victoria, c. 52), a new actor known as the official receiver became the fundamental figure in the management of bankruptcy procedures. Official receivers, instead of creditors’ trustees, were in charge of all preliminary phases of procedures up to the final stage of assets liquidation. This step could be left to creditors’ representatives, but only if courts had not decided otherwise. Because official receivers were full-time employees of the Board of Trade, this implied a massive shift from creditors to the government in the responsibility for the financial maintenance of bankruptcy apparatus.

  • 6 . The so-called deeds of arrangement consisted of different forms of friendly agreement between deb (...)

7As the 1883 act also increased the complexity and costs of dealings, and as the number of cases was on the rise, the 1880s transformation implied high expenditures for the public finances, so that the necessity arose of monitoring the efficiency and cost-effectiveness of the procedures. The very same 1883 law thus obliged the Board of Trade to collect a wide set of statistics to be published in an annual bulletin taking the name of Report by the Board of Trade under Section 131 of Bankruptcy Act (here referred to as “Bankruptcy Annual Report”). The bulletin was published yearly between 1884 and 1938, while one single, extremely summarised issue covered the whole period 1939-1953. The Bankruptcy Annual Report provides a variety of information, including the number of bankruptcies – and of other procedures such as deeds of arrangements6 – value of assets and liabilities involved (both at national level and disaggregated by court) and the debtors’ gender and field of activity. It also provides information about the length and costs of procedures and the amount of “dividends” (i.e. the percentage of liabilities paid to creditors).  

8In 1890, the Parliament passed the Company (winding-up) Act and the Board of Trade started to collect statistics on corporate insolvency too. They took the form of the annual Report by the Board of Trade under Companies (winding-up) Act (here “Insolvency Annual Report”), which started in 1892. Up to 1938, the bulletin was published yearly, but one single issue covered the 1939-1945 period. Data provided by the Insolvency Annual Report are less informative and detailed than what one can find in the Bankruptcy Annual Report, especially regarding the value of capital (assets and liabilities) involved in procedures. The problem arises from the fact that the English law allowed three types of company liquidation (winding-up): compulsory, voluntary, and voluntary under court supervision. The numbers of cases for each different sort of winding-up procedure were collected, but the recording of assets and liabilities was only mandatory for compulsory liquidations. As compulsory liquidations were only a small minority (on average about 10% of the total number of cases), this means that complete information about capital involved in all sorts of procedures is not directly available.

  • 7 . Lester, V. M., 1995; Di Martino, P., 2005a.

9Statistics provided by the Bankruptcy and Insolvency Annual Reports have already been used by researchers7. Lester’s book is undoubtedly the most important reference on insolvency and bankruptcy during the Victorian period. Although the author’s main interest is in the reconstruction of the political and parliamentary process leading to the approval and dismissal of various laws, he also offers an accurate analysis of statistical material. In particular, he provides data on bankruptcy and insolvency at national level (number, assets and liabilities), including an estimate of capital involved in voluntary liquidations, something which was not available from the official Reports. Lester also analyses data on failures in various cities and trades, and their relation to the business cycles. The only limit of Lester’s extremely valuable contribution is that it is confined to the pre-World War i period.

2. General patterns and efficiency

  • 8 . Hautcœur, P.-C. & Levratto, N., 2007.

10In order to assess the efficiency of the English bankruptcy and insolvency laws and practices, the very first step is to consider statistics on the evolution of their usage over time. In approaching the topic this way, we follow the methodology developed by Hautcœur and Levratto to study the functioning of bankruptcy institutions in France during the 19th century8. According to this approach, in circumstances of de facto insolvency (i.e. the inability to repay all debts), the actors involved (creditors, debtors, or courts) face a strategic choice between making recourse to official devices or simply seeking friendly agreements. This choice depends on the relative convenience of the two options and it is influenced by a number of elements, inclunding the efficiency of the formal procedures themselves. Creditors and debtors have incentives to use those laws and procedures that are able to allow “good” entrepreneurs (or sound companies) to restart, avoid the shame inherent to the “bankrupt” condition, and/or guarantee fast and cheap recovery of debts. In other cases, extra-judicial solutions, such as friendly agreements with creditors, are favoured. Using this approach, we study the number of insolvency and bankruptcy procedures as an indirect measure of their efficiency, keeping in mind that we are always comparing them to the (unobservable) level of extra-judicial remedies.

11Graph 1 plots the number of official procedures in England and Wales alongside a measure of the size of the economy (gdp at constant prices). Data are diseggregated between insolvency, ordinary bankruptcy, and deeds of arrangement, a solution available to personal debtors, consisting of court-supervised friendly agreements between creditors and debtors. The graph shows that between 1891 and 1939, while the economy experienced growth at a macro level, the total number of procedures not only did not increase at the same pace, but actually declined. This trend, however, is not common to all series: bankruptcy – strictly speaking – and deeds of arrangements declined, but the number of insolvencies increased. This simple piece of evidence already suggests that the structure of production, and especially the rise of limited-liability companies vis-à-vis other forms of governance, might be the main explanation of this pattern.

Graph 1. Number of official procedures in England and Wales (1891-1939)

Graph 1. Number of official procedures in England and Wales (1891-1939)

Source: Bankruptcy, insolvency, and deeds of arrangement: see Appendix; gdp: Mitchell, B., 1988.

12Before exploring this direction, however, it is useful to go back to the original argument: might formal procedures have became progressively less appealing – which, conversely, would have led to a growing appeal to extra-judicial agreements?

  • 9 . Di Martino, P., 2005b.

13If this were the case, such evidence would contradict the historical accounts of the evolution of formal practices in England and Wales during the period under examination. According to the historiography, between 1891 and 1939, no major change in the legislation occurred, but during the 19th century, the use of official procedures had already become easier and more convenient. In the English system, the difference between extra-judicial agreements and formal procedures was gradually fading. Deeds of arrangement, introduced in the 1830s, were directly managed by debtors and creditors, limiting the court’s involvement to ex post approval. On the other hand during the course of the 18th century, ordinary bankruptcy procedures had already largely lost their unpleasantness because of the introduction of debt discharge. In different paper, we already showed9 that in the period analysed here, both formal criteria and the actual implementation of debt discharge ensured that sound debtors filing for bankruptcy could be confident that their discharge from past debts, and consequently their rehabilitation, only depended on their behaviour and technical competences, leaving almost no impact to arbitrary elements such as class or gender.

14Despite the apparent efficiency of these various instruments, however, there is no doubt that the number of both deeds and insolvencies declined: the hypothesis that something in their actual implementation made them progressively less appealing therefore deserves investigation. Increasing length of procedures and/or rising costs are the first suspects.

15Graph 2 shows the evolution of the average length of procedures, desegregated between bankruptcy and deeds of arrangement, between 1905 and 1937. The graph clearly shows an overshooting in the early 1920s, but no long-term increase in the average length able to discourage creditors and debtors. If anything, in terms of duration, more complex procedures such as bankruptcies converged towards the length of “simpler” devices such as deeds of arrangements.

Graph 2. Average lenght of procedures (in months)

Graph 2. Average lenght of procedures (in months)

Source: Elaboration from P. Di Martino, 2005a.

  • 10 . This is a valuable contribution in itself because, as mentioned in section 1, some of these data (...)
  • 11 . V. M. Lester, 1995 uses data from C. Feinstein, 1976, while we use figures from B. Mitchell, 1988 (...)
  • 12 . Lester’s reconstruction is based on two estimates provided by the Board of Trade. Lester, V. M., (...)

16Duration, however, is only one part of the problem. The convenience in use of formal procedures also depends on how much payment they are able to guarantee. One way of assessing this problem is to calculate the so-called “loss to creditors”. This technique and this concept were developed in the early 20th century by the Board of Trade, in order to analyse the aggregate cost for the whole economy that resulted from company insolvency and personal bankruptcy. This aggregate measure is constructed by subtracting “net” assets from the level of liabilities and dividing the result by gdp at current prices. “Net” assets are estimated as one third of “gross” assets, in order to account for the cost of procedures. Lester has constructed an index of the loss to creditors for the period up to 1913, taking into account both personal bankruptcy and insolvency10. Using data from Bankruptcy Annual Reports, we provide an estimate of the “loss to creditors” ratio for the whole period 1890-1939; this series is plotted in Graph 3. Our data are slightly different from Lester’s figures: the series from 1890 to 1913 has been recalculated using a different estimate of gdp11 as deflator, although differences turned up to be minimal (see Appendix 2). More importantly, our data only include bankruptcies (and deeds of arrangement), as figures for corporate insolvency are not directly available and an indirect estimation has been made complicated (if not impossible) by the absence of criteria such as the ones used by Lester to produce his pre-1914 series12.

17Graph 3 shows that loss to creditors generated by personal bankruptcy between the 1880s and the Second World War tended to decline over time. At first glance, this can be interpreted as the sign that, in the long run, the increasing accessibility of English bankruptcy institutions had attracted a growing mass of “ordinary” failures, and not only the desperate cases, thus augmenting the average quality of applicants. Two considerations, however, make this interpretation problematic. Firstly, the improvement in the trend is not homogenous. During the First World War, for instance, the decline of the loss to creditor is remarkable, but the early 1920s and the 1930s crises are paralleled by a clear increase in the ratio. This shows that even if we accept the hypothesis of a long-term improvement due to institutional reasons, the system was to some extent still prone to the influence of short-time crises on the average quality of debtors. Secondly, as shown in another section of this paper, bigger firms have gradually tended to choose the incorporated form and thus moved under the insolvency umbrella. This means that the loss to creditor generated by bankruptcy data more and more reflected the condition of smaller companies: the ones that possibly had a better assets/liability ratio. In other words, the apparent improvement in the series might be due to a bias in the data sample.

Graph 3. Loss to creditor (% of gdp)

Graph 3. Loss to creditor (% of gdp)

Source: See Appendix.

18In addition, any conclusion based on the “loss to creditor” suffers from two severe limitations in the methodology itself. Firstly, the “loss to creditors” measures the impact of insolvency at aggregate level by reference to gdp, but it does not say much about the direct consequences to creditors, for example in terms of percentage of liabilities actually paid at the end of the procedure. The general process of growth of the economy in fact could hide a worsening ratio between the amount of assets and liabilities involved in procedures. Secondly, the loss to creditors is based on no real calculation of costs: they are simply considered as a fixed percentage of assets.

19In order to control for the first problem and provide a better assessment of the actual impact of bankruptcy on creditors, it is useful to look at the average dividend (as a percentage of liabilities) paid at the end of procedures. Graph 4 plots it, distinguishing between bankruptcies managed by the official receiver, bankruptcies administered by the creditors’ trustee, and deeds of arrangement.

Graph 4. Average level of dividends (% of liabilities)

Graph 4. Average level of dividends (% of liabilities)

Source: Elaboration from P. Di Martino, 2005a.

20Graph 4 shows that between 1891 and 1934, the average dividend served to non-secured creditors declined, contradicting the evidence provided by the “loss to creditors” index. Interestingly, no difference in the general pattern can be found between bankruptcies of any kind and deeds of arrangement, although the latter were supposed to be used by the best cohort of debtors, those who were able to pay relatively higher share of their debts.

  • 13 . This analysis is based on the assumption that the share of secured credits does not change. An in (...)

21It is evident that over time, the efficiency of English official procedures in ensuring the payment of a satisfactory dividend declined: this supports the idea that the appeal of extra-judicial agreements might have increased. The problem of diminishing dividends can have had two causes, both of them hard to detect using the “loss to creditors” ratio: an increase in the costs and/or an actual deterioration of the assets/liability ratio, for example because of a decreasing “quality” of debtors filing for official procedures13.

Graph 5. Percentage of liabilities not covered by assets

Graph 5. Percentage of liabilities not covered by assets

Source: Elaboration from P. Di Martino, 2005a.

22Graph 5 plots the difference between liabilities and assets as a percentage of the liabilities themselves. It shows that, as far as bankruptcy procedures are concerned, the percentage of liabilities not covered by assets increased over time, indicating a decline in the aggregate quality of debtors. This, however, was not the case for deeds of arrangement.

23The quality of debtors thus explains why bankruptcy procedures provided diminishing dividends, but not why the same happened in the case of deeds of arrangement. An increase in costs might be the solution to this puzzle. Table 1 shows that costs varied much among different kinds of procedures: deeds of arrangement had the lowest, and bankruptcy managed by trustees was cheaper than the same procedures in the hands of official receivers. Over time, the cost of bankruptcies run by trustees slightly decreased, while an increase can be noticed in the other two cases.

Table1. Costs of procedures (percentage of gross assets)

Period (Average)

Official receivership

Trustee

Period (Average)

Deeds

1884-1888

35.52

1889-1893

33.55

26.51

1891-1895

18.12

1894-1898

33.15

26.09

1896-1900

19.12

1899-1903

31.54

24.59

1901-1905

20.13

1904-1908

35.02

27.67

1906-1910

19.75

1909-1913

34.38

25.42

1911-1915

20.96

1914-1918

29.62

26.05

1916-1920

21.10

1919-1923

29.82

26.48

1921-1925

18.76

1924-1928

39.00

29.84

1926-1930

20.67

1929-1934

39.69

25.30

1931-1934

21.24

Average

34.13

26.05

19.91

Source: Di Martino, P., 2005a.

24In conclusion, it seems that the functioning of official mechanisms explains, at least to some extent, why they became less and less appealing. The length of procedures did not increase, but, contrary to what was suggested by the “loss to creditors” ratio, dividends declined. In the case of deeds of arrangements, growing costs was the problem, while bankruptcy procedures managed by creditors’ representatives suffered from the worse-ning quality of debtors. Bankruptcy dealings run by receivers were exposed to both problems.

3. Efficiency and the structure of the productive system

  • 14 . See F. Carnevali, 2005, and L. Hannah, 1983.

25The analysis conducted so far suggests that we have reasons to believe – or at least to suspect – that the official bankruptcy procedures progressively lost their appeal. However, this result is based on the assumption that the expansion of the English economy caused an increase in the number of firms, so that the decline in the use of official devices can be directly interpreted as a sign of the parallel increase in the recourse to friendly agreements. In fact, the positive link between macroeconomic growth and number of firms requires qualifications. Firstly, the period under examination is considered as a phase of industrial consolidation and concentration, suggesting that the number of firms did not necessarily increase while the economy was expanding14. Secondly, it might well be the case that, even if the total number of businesses augmented, some types of companies declined in number while others increased.

26In order to appreciate the relevance of these problems, we have to go back to the data plotted in Graph 1, which shows that the total decline in the use of procedures is the net result of two opposite trends: an increase in the number of insolvency dealings and a parallel decline in the number of bankruptcy cases. To consider this pattern as the sign of a declining efficiency of official devices, we have to assume that the size of the sample of businesses subject to bankruptcy increased or remained constant over time, while the number of cases that fell under the insolvency law increased more than proportionally with the rise of procedures.

  • 15 . Lester, V. M., 1995.

27Unfortunately, we lack data about the number of sole-ownerships ope-rating in England and Wales during this period, and any guess-estimate implies a high degree of speculation. The same must be said about partnerships, even if, in this case, a more solid ground can be found by considering the relation between unlimited partnerships and limited-liability companies. Although no statistics are available, the historiography supports the hypothesis that partnerships above a given size might have gradually turned into limited-liability companies, and that new firms were incresingly created as limited-liability companies instead of partnerships15. This would increase the sample of firms subject to insolvency and decrease the number of businesses exposed to the bankruptcy regime. This could provide an explanation of the decreasing number of bankruptcy procedures, that would be alternative or complementary to the one based on their increasing cost and declining dividends.

28However, the total number of procedures (bankruptcy and insolvency) still declined over time. If we assume that no major change in the population of sole-ownerships took place and that a shift only occurred from partnerships (subject to bankruptcy) to companies (subject to insolvency), the question remains open: why did the increase in the number of winding-ups not compensate for the decline in the number of bankruptcy dealings? A simple answer could be that, even if partnerships progressively turned into joint-stock companies, the total number of firms subject to insolvency did not increase in proportion, for example because of concentration and increasing size of firm. As a consequence, we would not expect the number of insolvencies to grow in the same proportion as the decline in the number of bankruptcies.

  • 16 . For details, see Appendix.
  • 17 . It is unclear whether or not these figures include the number of limited partnerships registered (...)

29In order to test this hypothesis, an estimate of the number of companies existing in England between 1891 and 1937 is provided16 and data on numbers of insolvencies are plotted against this variable. It shows that the ratio between the number of insolvencies and the number of companies declined from 1891 to 1937 (Graph 6)17.

Graph 6. Ratio number of insolvency/number of companies 1891-1937 (%)

Graph 6. Ratio number of insolvency/number of companies 1891-1937 (%)

Sources: Number of insolvencies: see Appendix; number of companies: Insolvency Annual Report, various issues.

30Having dropped the possibility that the evolution of the number of insolvency procedures might be explained by the number of companies subject to it, we now have to decide whether this result must be connected to a declining appeal of these procedures and an increasing recourse to extra-judicial agreements. This would be an extremely surprising result, as the very nature of the English corporate insolvency was already very much oriented towards friendly agreements in the form of voluntary liquidations, even if this solution, which implied a very limited role for official bodies, was usually confined to still solvent companies. Unfortunately, data about the functioning of insolvency procedures, such as length and dividends, are not available and any assessment of their efficiency remains a matter of speculation.

  • 18 . Only companies with nominal capital have been included. It must be noticed that a negligible amou (...)
  • 19 . Because of the positive average rate of inflation during the period under investigation, the decr (...)

31As an alternative, we could discuss whether the declining recourse to insolvency in fact was just the result of an increasing stability of companies, because of better governance, improved risk-assessment, increasing size, and so on. This approach would require the collection of an enormous amount of data and, even if this were possible, the limited scope of this paper would make a comprehensive study of such a complex topic unfeasible. However, something can be said at least about the companies’ size, thanks to data on the average amount of nominal capital provided by Insolvency Annual Reports18. Graph 7 shows that, parallel to the phenomenon of industrial concentration, the average capitalization (in nominal terms) of newly-created English companies declined over time19. These two phenomena are not in conflict, as whilst few big companies became bigger, a vast number of relatively little concerns entered the market.

Graph 7. Average nominal capital of English companies 1891-1937 (Pounds)

Graph 7. Average nominal capital of English companies 1891-1937 (Pounds)

Source: Insolvency Annual Report, various issues.

32Although this is far from being a conclusive piece of evidence, the hypothesis of an English system progressively becoming more stable, and therefore the expectation that the number of failures in proportion to the number of companies would decline, is not supported, as smaller firms tend to be more prone to failure than the bigger ones.

4. Comparative approach and relative efficiency

33Sections 2 and 3 discussed the efficiency of the English law and practices in a national perspective. The risk implicit in this kind of approach is that, because no system is perfect, weaknesses might be overemphasized. The use of a comparative methodology allows us to put the strengths and weaknesses of the English system in perspective and provides a more nuanced assessment of its efficiency.

  • 20 . Di Martino, P., 2006.
  • 21 . Sgard, J., 2006.
  • 22 . Di Martino, P., 2005a.
  • 23 . Di Martino, P. & Vasta, M., 2007.

34In this regard, Italy offers a good measure of comparison, as the two countries had different approaches to the problem of insolvency. It is interesting to notice that these differences do not seem to have been the unavoidable consequences of belonging to a different legal family, as predicted by the “law and growth” approach, rather the product of idiosyncratic patterns of development of bankruptcy and insolvency laws in the two legal systems20. Although during the 20th century, as correctly pointed out by Sgard21, the attention shifted everywhere from the issue of punishing “immoral” debtors to the problem of finding a set of devices minimising the collective costs of insolvency and bankruptcy, the results of this process varied significantly from country to country. In England, the search for solutions inducing debtors’ cooperation, rather then using pure coercion, was pursued with more efficiency than everywhere else in Europe. This led to the provision of specific “soft” devices (for example, debts discharge was available in England, but not in Italy), as well as easier formal conditions to file for alternatives to bankruptcy (like deeds of arrangement in England and concordati preventivi in Italy)22. Qualitative evidence suggests that in the Italian case, formal conditions made a big difference on the actual possibility for debtors to use alternative procedures23.

35According to this picture, we would expect the English data to show less fluctuations than Italian figures: English bankruptcy and insolvency laws are supposed to have become devices used under normal circumstan-ces as well as during crises. In Italy, on the contrary, these procedures were painful for debtors: we expect rational economic agents to have tried to avoid them by any means. Insolvency, in a sense, was only for the desperate. During crises, however, their number would have increased and, at the same time, escaping bankruptcy (for example by further recourse to credit) would have become more difficult. For these reasons, we expect the correlation between macroeconomic fluctuations and the number of insolvencies and bankruptcies to be stronger in Italy than in England. Graph 8 plots the number of bankruptcy and insolvency cases in Italy and England (and Wales) between circa 1890 and 1939.

Graph 8. Total number of bankruptcies and insolvencies in England (and Wales) and Italy, c. 1891-c.1939

Graph 8. Total number of bankruptcies and insolvencies in England (and Wales) and Italy, c. 1891-c.1939

Source: See Appendix.

36The two series show common aspects and remarkable differences. During World War I, insolvency massively declined in both countries. Two possible explanations for this pattern can be provided. Firsty, the war imposed peculiar conditions to the economy, inflated by an artificially increased demand; in a sense, anyone who was able to sell something useful for the conflict could find a market. Secondly, during wars, the functioning of legal institutions might be so affected as to made them unable to process all cases. This would explain the sudden rise in the number of cases in the immediate after-war years.

37Long-term stability is the second pattern shared by the two series. Despite a clear growing trend in the number of cases noticeable in Italy up to the early 1930s, at the end of the period, the total number was close to what it was at the beginning of the period. The situation is similar in the English case, while, as noticed in section 2, the number slightly declined over time.

38When we look at short-term fluctuations, however, significant differences surface, especially during years traditionally associated with short-term crises. In Italy, the 1911 crisis, the reintroduction of the gold standard in 1926, and the Great Depression are all paralleled by sudden increases in the number of insolvencies and bankruptcies. England provides a different picture: neither the Baring crisis in 1890, nor the return to the gold standard in the 1920s is associated to remarkable peaks in the number of business failures. The early 1920s contraction sees an increase in the number of cases, but this seems to be related to a general growing trend that followed the war, rather than to a specific crisis (in fact, the number is in line with the average pre- and post-war levels). Only the dramatic years of the great Depression represent an exception. Also from this point of view, however, the picture of England as compared to Italy is one of stability. Table 2 shows the difference between the average number of cases during the depression years and the total average in the two cases. In England, the average number of insolvencies between 1930-1932 is only 13% higher than the total sample average, a figure that goes down to 9% when 1933 (although hardly a year of crisis in England) is included. By contrast, in Italy, the 1930-1933 average is 184% higher than the sample average.

Table 2. Average of total sample versus depression years

Italy

England

Total average vs. average 1930-1933

183.72%

9.4%

Total average vs. average 1930-1932

13.2%

Source: See Appendix.

  • 24 . The coefficient of variation is constructed by dividing the standard variation by the means. This (...)

39The analysis of the standard deviation of the two series provides a further way of measuring the relation between macroeconomic fluctuations and the number of insolvencies. Using this procedure, we focus on “ordinary” fluctuations linked to the business cycle, rather than on the sensitivity to short-term exceptional shocks. In order to make the comparison meaningful, data are turned into coefficients of variations shown in Table 324. This confirms the above picture: as compared to Italian figures, English data have a much lower measure of dispersion, especially when the exceptional war years are not included in the sample.

Table 3. Coefficient of variation of insolvencies series

Italy

England

WWI years included

0.82

0.24

WWI years not included

0.78

0.15

Source: See Appendix.

  • 25 . Di Martino, P. & Vasta, M., 2007.

40These results are exposed to two possible criticisms. The first one refers to the fact that because of the inaccessibility of official procedures in Italy, the recourse to unofficial agreements might have been higher than in England. Because different types of firms might have found it easier to reach extra-judicial agreements, the two series would include different kinds of firms. In particular, it might be argued that the recourse to informal deals is easier for smaller companies, because of lower coordination problems. Therefore, the English data might include a lower number of corporate insolvencies, which could influence the relative stability of the two series. Unfortunately, the Italian data do not allow us to differentiate between insolvency and bankruptcy, but qualitative evidence suggest that official procedures (including concordati preventivi) were as hard to obtain for joint-stock companies as for partnerships or sole-ownerships25.

41On the other hand, the higher volatility in the number of bankruptcies in Italy, as compared to England, might be the mere result of a less stable macroeconomic environment. This hypothesis can be tested by looking at the variance of series of gdp at constant prices in the two countries (Table 4).

Table 4. Coefficient of variation of gdp (constant prices) series

Italy

England

Including WWI

0.34

0.18

Excluding WWI

0.35

0.19

Source: Italy: Rossi, N., Sorgato, A. & Toniolo, G., 1993; England: Mitchell, B., 1988.

42Table 4 shows that the coefficient of variation of gdp in Italy is roughly two times higher than the English one. However, the coefficient of variation of the bankruptcy and insolvency series is about four times higher in Italy than in England. If we assume a stable proportional relation between variation in the number of insolvency/bankruptcy cases, variation of gdp, these results suggest that the higher turbulence of the macroeconomic environment only partially explains the difference in the number of failures in the two countries.

43Descriptive statistics thus seem to confirm the expected view about the relative degrees of stability of the two legal regimes. The picture obtained from the analysis of standard deviations goes in the same direction, although the last test relies on the rigid assumption of a stable, strictly proportional correlation between growth and variation in the number of cases of insolvency. On the whole, these results support the view that England had found better solutions than Italy to the problem of providing trust-worthy and efficient insolvency and bankruptcy laws and procedures.

  • 26 . Gerschenkron, A., 1962.

44Scholars using the findings of the so-called “law and growth” paradigm as the conceptual framework for historical research had to give credit, once again, to the wise words of the old master Alexander Gerschenkron, who claimed that Clio is a messy housewife26. This paper is no exception and shows that even in England, the motherland of common law, the functioning of legal institutions might not have been quite as smooth as commonly believed.

45Despite sustained economic growth between the 1890s and the 1930s, the use of official bankruptcy procedures in England declined over time. Assuming a stable level of aggregate risk in the economy, and a positive correlation between macro growth and the number of businesses operating in the country, this picture casts doubts on the actual efficiency of these procedures, suggesting that recourse to unofficial, extra-judicial agreements might have risen over the period. The decline of dividends extracted via official devices, either because of growing costs and/or a diminishing quality of debtors, supports this view. A conclusive assessment of this hypothesis, however, is made impossible by the fact that the actual number of firms subject to bankruptcy procedures (sole ownerships and unlimited partnerships) is unknown.

46This problem does not exist in the case of insolvency, as an estimation of the stock of joint-stock companies registered in England can be provided. Data show that during the period under study, their number increased more than proportionally to the volume of official winding-up procedures. This might suggest that insolvency also declined in popula-rity, to the advantage of extra-judicial deals. Unfortunately, statistics on the efficiency of official insolvency procedures are not available, and an exhaustive analysis of all elements able to shift the companies’ preference from the use of judicial solutions to friendly agreements requires a wider scope than the one of this paper. The study of one of these elements (the new firms’ size) reveals that the average English start-up became smaller over the period studied. This evidence, however, is inconclusive, because smaller companies are, on the one hand, more prone to failure, but, on the other, in a better position to obtain friendly agreements.

47Finally, comparative analysis shows that English procedures were less sensitive to macroeconomic fluctuations than Italian ones. This suggests that, despite possible limitations and problems, the English system still managed to attract a relatively stable cohort of average firms, while under the Italian legislation, only the very worse segment of debtors used official remedies.

Haut de page

Bibliographie

Bordo, Michael D. & Rousseau, Peter L., “Legal-Political Factors and the Historical Evolution of the Finance-Growth Link”, European Review of Economic History, vol. 10, n° 3, 2006, p. 421-444.

Carnevali, Francesca, Europes advantage. Banks and Small Firms in Britain, France, Germany, and Italy since 1918, Oxford, Oxford University Press, 2005.

Di Martino, Paolo, “Approaching Disaster: A Comparison between Personal Bankruptcy Legislation in Italy and England (c.1880-1939)”, Business History, vol. 47, n° 1, 2005(a), p. 23-43.

—, “Debt Discharge and Entrepreneurship in Britain, c. 1880-1939”, Paper presented at the 2005 Economic History Society Annual Conference, 2005(b).

—, “Lobbying, Institutional Inertia, and the Efficiency Issue in State Regulation: Evidence from the Evolution of Bankruptcy Laws and Procedures in Italy, England, and the US (c. 1870-1939)”, Paper presented at the Annual Conference of the European Association of Banking and Financial History, 2006.

Di Martino, Paolo & Vasta, Michelangelo, “Size, Structure, and Strategies: Insolvency and ‘The Nature of the Firm’ in Italy, 1920s-1970s”, Department of Economics, University of Siena, Working Paper n° 515, 2007. http://www.econ-pol.unisi.it/quaderni/515.pdf

Duffy, Ian P. H., Bankruptcy and Insolvency in London during the Industrial Revolution, New York and London, Garland Publishing, 1985.

Feinstein, Charles H., Statistical Tables of National Income, Expenditure and Output of the U.K. 1855-1965, Cambridge, Cambridge University Press, 1976.

Gerschenkron, Alexander, Economic Backwardness in Historical Perspective. A Book of Essays, Cambridge, Cambridge University Press, 1962.

Hannah, Leslie, The Rise of the Corporate Economy. London and New York, Methuen, 1983.

Hautcœur, Pierre-Cyrille & Levratto, Nadine, “Bankruptcy Law and Practice in 19th century France” Paris School of Economics Working Paper, n° 29, 17 p. http://www.pse.ens.fr/document/wp200729.pdf

Hoppit, Julian, Risk and Failure in English business, 1700-1800, Cambridge, Cambridge University Press, 1987.

La porta, Rafael, Lopez-de-Silanes, Florencio, Shleifer, Andrei & Vishny, Robert W., “Law and Finance”, Journal of Political Economy, vol. 106, n° 6, 1998, p. 1113-1155.

Levine, Ross, “The Legal Environment, Banks, and Long-Run Economic Growth”, Journal of Money, Credit and Banking, vol. 30, n° 3, 1998, p. 596-613.

Lester, V. Markham, Victorian Insolvency. Bankruptcy, Imprisonment for Debt, and Company Winding-up in Nineteenth-Century England, Oxford, Clarendon Press, 1995.

Mitchell, Brian R., British Historical Statistics, Cambridge, Cambridge University Press, 1988.

Musacchio, Aldo, “Can Civil Law Countries Get Good Institutions? Lessons from the History of Creditors Rights and Bond Market in Brazil”, Journal of Economic History,vol. 68, n° 1, March 2008, p. 80-108.

Rajan, Raghuram G. & Zingales, Luigi, “The Great Reversal: The Politics of Financial Development in the 20th Century”, Journal of Financial Economics, vol. 69, July 2003, p. 5-50.

Rossi, Nicola, Sorgato, Andrea & Toniolo Gianni, “I conti economici italiani: una ricostruzione statistica”, Rivista di Storia Economica, vol. 10, n° 1, 1993, p. 1-47.

Sgard, Jérôme, “Do Legal Origins Matter? The case of Bankruptcy Laws in Europe 1808-1914”, European Review of Economic History, vol. 10, n° 3, 2006, p. 389-419.

Shannon, Herbert A., “The Limited Companies of 1866-1883”, The Economic History Review, vol. 4, n° 3, 1933, p. 290-316.

, “The Coming of the General Limited Liability”, in Eleonora M. Carus-Wilson (ed.), Essays in Economic History, London, Edward Arnold, 1954.

Haut de page

Annexe

Appendix 1.Stock of English joint-stock companies: sources and methodology

H. A. Shannon provides the stock of existing companies in 1884 (944)27. In another paper28, he also provides the number of companies registered, sold, merged, and liquidated in London in 1866-1874 and 1875-1883. Using these data, we calculated the percentage of companies that disappeared in London during these two periods. Assuming that this percentage is a good approximation of the data at the national level, we calculated the number of companies lost in England by applying the relative to the first period percentage to the cumulative sum of companies registered in the country in 1845-187429, and the relative to the second period percentage to the cumulative sum of companies registered in England in 1875-1883. We then simply calculated the stock of companies in 1874 and the number of companies added until 1883 as the difference between the cumulative sum of registered companies and the number of companies lost. The actual stock of companies in 1883 is the sum of these two numbers, plus the original estimate for 1884. The stock created between 1884 and 1890 has been constructed with a similar methodology, but we only used data for merged companies30, rather than the number of sold or merged companies provided by Shannon in his 1933 paper. This procedure underestimates the number of companies lost by about 10%, but has the advantage of relying on more solid statistics.

After 1890, we calculate the stock of companies for each year by adding the number of companies registered in that year to the previous year’s stock and subtracting the wound-up and merged companies.

Appendix 2. Numbers of failures in England and Wales

Bankruptcy

Deeds

Insolvency

Total

1891

4242

3008

873

8123

1892

4657

3333

1091

9081

1893

4901

3938

1042

9881

1894

4794

3894

998

9686

1895

4415

3462

1060

8937

1896

4170

3271

1262

8703

1897

4098

3208

1587

8893

1898

4310

3246

1745

9301

1899

4111

2974

1793

8878

1900

4410

3354

1804

9568

1901

4244

3369

1618

9231

1902

4202

3305

1629

9136

1903

4286

3622

1632

9540

1904

4546

4085

1600

10231

1905

4764

3839

1693

10296

1906

4436

3641

1730

9807

1907

4111

3488

1797

9396

1908

4306

3822

2120

10248

1909

4070

3491

1944

9505

1910

3880

3364

1960

9204

1911

3742

2950

2061

8753

1912

3581

2770

2067

8418

1913

3358

2411

1946

7715

1914

2867

1776

2022

6665

1915

2379

1652

1648

5679

1916

1557

1050

1354

3961

1917

1089

612

1233

2934

1918

628

198

1113

1939

1919

745

165

1908

2818

1920

1565

451

3158

5174

1921

3472

1368

2928

7768

1922

4733

1847

2834

9414

1923

5048

1808

2859

9715

Bankruptcy

Deeds

Insolvency

Total

1924

4794

1901

2810

9505

1925

4708

1877

2730

9315

1926

4239

1763

2737

8739

1927

4314

1856

2703

8873

1928

4167

2054

3064

9285

1929

3929

1971

2707

8607

1930

4133

2154

2760

9047

1931

4403

2415

2804

9622

1932

4645

2676

2712

10033

1933

4110

2102

2600

8812

1934

3623

1861

2568

8052

1935

3523

1635

2711

7869

1936

3249

1598

2693

7540

1937

3123

1630

2671

7424

1938

3105

1663

2810

7578

1939

2635

1414

2980

7029

Sources: 1891-1913: Lester, V. M., 1995; 1914-1939: Bankruptcy Annual Report and Insolvency Annual Report(various issues). Data on numbers of bankruptcies have already been published by B. Mitchell, 1988, but our series, following Lester’s indication, takes the receiving orders into account, rather than the adjudicated bankruptcies collected by Mitchell. This allows the inclusion of all cases ending up with any kind of composition, together with the actual bankruptcies. For more details, see V. M. Lester, 1995, p. 214-215.

Appendix 3. Loss to creditor (% of gdp)

Lester

Our calculations

Bankruptcy

Deeds

Winding-up

Total

Bankruptcy

Deeds

Total

1888

0.5

0.26

1889

0.43

0.23

1890

0.39

0.21

0.40

0.26

0.65

1891

0.53

0.22

0.54

0.29

0.83

1892

0.57

0.3

1.37

2.24

0.56

0.36

0.92

1893

0.48

0.4

1.26

2.14

0.49

0.47

0.96

1894

0.44

0.32

1.27

2.03

0.43

0.37

0.81

1895

0.4

0.23

0.97

1.6

0.41

0.28

0.69

1896

0.34

0.2

1.24

1.78

0.34

0.24

0.58

1897

0.31

0.18

1.38

1.87

0.32

0.22

0.54

1898

0.36

0.16

1.55

2.07

0.37

0.20

0.57

1899

0.3

0.13

1.62

2.05

0.30

0.16

0.46

1900

0.3

0.15

1.43

1.88

0.31

0.19

0.50

1901

0.3

0.14

1.24

1.68

0.30

0.17

0.47

1902

0.25

0.15

1.31

1.69

0.25

0.19

0.44

1903

0.24

0.15

1.28

1.67

0.24

0.19

0.43

1904

0.32

0.17

1.04

1.53

0.32

0.22

0.54

1905

0.26

0.13

1.02

1.37

0.27

0.16

0.42

1906

0.25

0.14

0.98

1.38

0.26

0.18

0.44

1907

0.25

0.16

0.9

1.31

0.25

0.21

0.46

1908

0.24

0.19

1.16

1.59

0.24

0.24

0.48

1909

0.25

0.13

1.24

1.62

0.25

0.16

0.41

1910

0.35

0.11

0.89

1.35

0.35

0.14

0.49

1911

0.28

0.12

0.8

1.2

0.28

0.15

0.44

1912

0.19

0.09

0.99

1.29

0.20

0.12

0.31

1913

0.19

0.07

0.75

1.01

0.19

0.10

0.29

1914

0.25

0.15

0.40

1915

0.13

0.07

0.20

1916

0.08

0.04

0.12

1917

0.04

0.02

0.06

1918

0.01

0.01

0.02

1919

0.03

0.01

0.04

Lester

Our calculations

Bankruptcy

Deeds

Winding-up

Total

Bankruptcy

Deeds

Total

1920

0.08

0.03

0.11

1921

0.28

0.15

0.43

1922

0.39

0.19

0.58

1923

0.27

0.12

0.39

1924

0.31

0.10

0.41

1925

0.25

0.09

0.34

1926

0.23

0.10

0.32

1927

0.19

0.09

0.28

1928

0.18

0.09

0.27

1929

0.20

0.11

0.30

1930

0.22

0.11

0.33

1931

0.27

0.12

0.39

1932

0.25

0.14

0.40

1933

0.16

0.09

0.25

1934

0.15

0.08

0.22

1935

0.13

0.11

0.24

1936

0.10

0.05

0.15

1937

0.11

0.06

0.17

1938

0.13

0.06

0.19

1939

0.12

0.04

0.16

Haut de page

Notes

Sources: Lester data: Lester, V. M., 1995; Our data: our calculation based on various issues of Bankruptcy Annual Report (assets and liabilities) and B. Mitchell, 1988 for gdp.

1 . La Porta, R. & al., 1998; Levine, R., 1998.

2 . The traditional “law and growth” approach stresses the former mechanism, while the latter has been developed, in particular, by R. Rajan & L. Zingales, 2003.

3 . See especially M. Bordo & P. Rousseau, 2006, A. Musacchio, 2008, and J. Sgard, 2006.

4 . Di Martino, P., 2005a ; Hautcœur, P.-C. & Levratto, N., 2007.

5 . Works by historians such as I. Duffy, 1985, J. Hoppit, 1987, and V. M. Lester 1995, mainly discusses the political economy of bankruptcy and insolvency laws, while a previous work by the author (Di Martino, P., 2005a) focuses on length of procedures and amount of dividends, but did not provide a more general picture.

6 . The so-called deeds of arrangement consisted of different forms of friendly agreement between debtors and creditors requiring the supervision and approval of a court.

7 . Lester, V. M., 1995; Di Martino, P., 2005a.

8 . Hautcœur, P.-C. & Levratto, N., 2007.

9 . Di Martino, P., 2005b.

10 . This is a valuable contribution in itself because, as mentioned in section 1, some of these data were not available in the various bulletins.

11 . V. M. Lester, 1995 uses data from C. Feinstein, 1976, while we use figures from B. Mitchell, 1988.

12 . Lester’s reconstruction is based on two estimates provided by the Board of Trade. Lester, V. M., 1995, chapter 6.

13 . This analysis is based on the assumption that the share of secured credits does not change. An increase in this share would make the problem of the worsening of the assets/liabilities ratio even more serious for unsecured creditors, as a bigger part of a declining amount of resources would be devolved to the payment of higher-ranked debts. Conversely, less secured credits would make the relative position of unsecured creditors easier and soften the impact of declining assets/liabilities ratio. Unfortunately, to our knowledge, data on secured credits are not available.

14 . See F. Carnevali, 2005, and L. Hannah, 1983.

15 . Lester, V. M., 1995.

16 . For details, see Appendix.

17 . It is unclear whether or not these figures include the number of limited partnerships registered in England. However, this number was negligible.

18 . Only companies with nominal capital have been included. It must be noticed that a negligible amount of company without capital existed in England during this period.

19 . Because of the positive average rate of inflation during the period under investigation, the decrease in real terms is even more remarkable.

20 . Di Martino, P., 2006.

21 . Sgard, J., 2006.

22 . Di Martino, P., 2005a.

23 . Di Martino, P. & Vasta, M., 2007.

24 . The coefficient of variation is constructed by dividing the standard variation by the means. This procedure makes series comparable when data are of different orders of magnitude.

25 . Di Martino, P. & Vasta, M., 2007.

26 . Gerschenkron, A., 1962.

27 .Shannon, H. A., 1954.

28 .Shannon, H. A., 1933.

29 .Data until 1862 have been taken from H. A. Shannon, 1933, and for the following phase from various issues of the Insolvency Annual Report.

30 .Taken from L. Hannah, 1983.

Haut de page

Table des illustrations

Titre Graph 1. Number of official procedures in England and Wales (1891-1939)
Légende Source: Bankruptcy, insolvency, and deeds of arrangement: see Appendix; gdp: Mitchell, B., 1988.
URL http://journals.openedition.org/histoiremesure/docannexe/image/3083/img-1.jpg
Fichier image/jpeg, 200k
Titre Graph 2. Average lenght of procedures (in months)
Légende Source: Elaboration from P. Di Martino, 2005a.
URL http://journals.openedition.org/histoiremesure/docannexe/image/3083/img-2.jpg
Fichier image/jpeg, 128k
Titre Graph 3. Loss to creditor (% of gdp)
Légende Source: See Appendix.
URL http://journals.openedition.org/histoiremesure/docannexe/image/3083/img-3.jpg
Fichier image/jpeg, 208k
Titre Graph 4. Average level of dividends (% of liabilities)
Légende Source: Elaboration from P. Di Martino, 2005a.
URL http://journals.openedition.org/histoiremesure/docannexe/image/3083/img-4.jpg
Fichier image/jpeg, 176k
Titre Graph 5. Percentage of liabilities not covered by assets
Légende Source: Elaboration from P. Di Martino, 2005a.
URL http://journals.openedition.org/histoiremesure/docannexe/image/3083/img-5.jpg
Fichier image/jpeg, 176k
Titre Graph 6. Ratio number of insolvency/number of companies 1891-1937 (%)
Légende Sources: Number of insolvencies: see Appendix; number of companies: Insolvency Annual Report, various issues.
URL http://journals.openedition.org/histoiremesure/docannexe/image/3083/img-6.jpg
Fichier image/jpeg, 152k
Titre Graph 7. Average nominal capital of English companies 1891-1937 (Pounds)
Légende Source: Insolvency Annual Report, various issues.
URL http://journals.openedition.org/histoiremesure/docannexe/image/3083/img-7.jpg
Fichier image/jpeg, 144k
Titre Graph 8. Total number of bankruptcies and insolvencies in England (and Wales) and Italy, c. 1891-c.1939
Légende Source: See Appendix.
URL http://journals.openedition.org/histoiremesure/docannexe/image/3083/img-8.jpg
Fichier image/jpeg, 156k
Haut de page

Pour citer cet article

Référence électronique

Paolo Di Martino, « Dealing with failure », Histoire & mesure [En ligne], XXIII - 1 | 2008, mis en ligne le 01 juin 2011, consulté le 10 décembre 2017. URL : http://journals.openedition.org/histoiremesure/3083

Haut de page

Auteur

Paolo Di Martino

University of Manchester, Oxford Road, Manchester, M13 9PL, Royaume-Uni.

Haut de page

Droits d’auteur

© Éditions de l’EHESS

Haut de page