An economic enterprise faces two, related, problems: effectively managing its activities and communicating to outsiders that it is, in fact, well run. The credit cooperative movement that grew up in Germany in the second half of the nineteenth century had to wrestle with both. These cooperatives thrived, in part, because they adopted strategies first to obtain and then to harness the information they needed about the communities in which they were located, giving them an advantage over other lenders. Particularly effective was the tactic of using local people as managers, which helped to cement their ties with the community. Yet because few, if any, locals had banking experience and most were not even familiar with basic accounting methods, the managers created internal management problems, intensifying outside suspicion of the cooperatives as banking enterprises. The methods the cooperatives developed to overcome these problems drew on a combination of local initiative and regional assistance that was typical of the movement as a whole. The movement's ability to train its own talent suggests that it had a broader impact than has been captured by statistics on its membership or financial assets.
Historical accounts of the external audits now required of most public firms usually link the emergence of these audits to the development of corporate forms that made it possible for widely dispersed investors to hold equity in, or lend money to, enterprises they knew little about. The English Joint Stock Companies and Regulation Act of 1844, to take one important example, made it easier for an enterprise to incorporate as a joint-stock firm but at the same time stipulated that half-yearly returns of audited balance sheets be filed with the Registrar of Joint Stock Companies. The logic of the external audit is both to protect the public from fraud and to facilitate markets in stocks and bonds. In a partnership, a small number of investors had both an incentive and the ability to acquaint themselves with the firm's activities and condition. But the issuance of stocks and bonds to widely dispersed investors, many of whom held only a small share of the firm's securities, afforded the opportunity for the firm's "insiders" to use their superior knowledge of the firm's condition to abuse the "outsiders," whose trust was crucial to the development and operation of capital markets.
This conventional view of external auditing is correct so far as it goes. But the joint-stock corporation is not the only institution that can be hurt by the insiders' real or imagined misconduct. The problems German credit cooperatives and their stakeholders had to deal with during the late nineteenth century differed from those affecting the typical joint-stock firm. Every cooperative had to undergo an external audit at least every other year after passage of a new cooperatives law in 1889. German joint-stock firms, on the other hand, did not require regular audits until the 19303. The cooperatives had no traded debt or equity shares, and nearly all their members and depositors lived in close, proximity to the institutions and their managers. The outsiders were not the hands-off investors whose grievances motivated the development of external audits for joint-stock firms, and the abuses that concerned them were different from those that were likely to surface in joint-stock firms. Understanding the concerns that led to compulsory audits for cooperatives, as well as the mechanisms the cooperatives developed to perform those audits, elucidates more generally the problem enterprises face in ensuring that they are well run and convincing the public that such is the case.
This article is part of a larger project that uses economic theory and a broader perspective on the German financial system to ask why credit cooperatives prospered in the period between 1889 and 1914. I have argued elsewhere that central to their success was an ability to take advantage of the close ties among members in order to overcome, at low cost, the information and enforcement problems that made it difficult and expensive for conventional banking institutions to lend to poor people. Although Germany had both an urban and a rural credit cooperative group, I emphasize the rural institutions because their organization and methods made them most distinctive.1
After some initial background, the discussion proceeds in three stages, beginning with the structure of the local cooperatives, turning next to the performance of local managers and some of the problems they faced, and concluding with a consideration of the regional auditing associations that were erected to facilitate good management and to help the cooperatives instill confidence in their operations.
The Cooperatives
The rural credit cooperatives that thrived in Germany in the second half of the nineteenth century were managed by volunteers or part-time employees who had little banking or business experience. The institutions were individually small, and their members were poor people with similar economic fortunes. Yet, by World War I, there were about 19,000 such cooperatives, and together they had issued some 8 percent of all German banking liabilities. Key to their success was an ability to capitalize on cheap information and social sanctions to provide low-cost credit. The ability to use social ties as the basis for lending came at a steep cost, however. To make the system work, the cooperatives had to be confined to a specific area and remain small. Reliance on inexperienced local managers was one consequence of this design.
Credit cooperatives faced skepticism not only from the public but also from other financial institutions, the state, and other cooperatives. Accounts of the German financial system in the late nineteenth century give too much emphasis to the "Great Banks"-the large, universal banks made famous by Alexander Gerschenkron's analysis of economic "backwardness" and its remedies. In any case, the cooperatives did not compete with anything so large and sophisticated. Credit cooperatives did provide some competition for two other financial intermediaries. Some of the cooperatives' customers, especially those in urban areas, were good candidates for loans from private banks. Most accounts that even mention the connection, however, stress a symbiotic relationship: some private bankers provided loans and payment services to credit cooperatives.
The credit cooperatives were a direct threat to another, often ignored, financial intermediary, the savings banks (Sparkassen). The savings banks rarely extended loans of the sort that would be made by a credit cooperative. But the two institutions competed directly for deposits. Deposits at credit cooperatives lacked the public guarantee afforded those in savings banks, but the cooperatives paid higher rates of interest and were more convenient for many rural people. The evidence of competition is both direct and indirect. Some public criticisms of the soundness of credit cooperatives came directly from leaders of the savingsbank movement. Cooperatives located in villages or towns with strong savings banks also had serious difficulty attracting deposits.2
The state's attitude toward the credit cooperatives was more complex. The cooperatives predate Germany's unification in 1871. In the early days, some state (Land) governments were supportive, while others were actively hostile. A few states agreed with Hermann Schulze-Delitzsch, the leader of Germany's major urban cooperative group, and Freidrich W. Raiffeisen, who led a rival cooperative group, that the cooperative movement would be a bulwark against socialism. Both men abhorred socialism, and each viewed cooperatives as a way to help the poor and so prevent the spread of socialism in Germany. Other state officials seemed to fear that cooperation was just socialism in another guise. Some state hostility to the credit cooperatives was provoked by their internecine arguments. Over time, state governments, as well as the Reich, adopted a more supportive view of the cooperative movement. The 1889 Reich cooperatives law was largely written according to the cooperatives' blueprint, and by the 18903 even Prussia, whose hostility had been especially damaging given its size, had turned to active support of the movement.
Public concern about the soundness of credit cooperatives was most sharply aroused by comments the cooperatives made about each other. The debates, which were complex, began with the publication of a series of articles by Schulze-Delitzsch, pronouncing the rural credit cooperatives organized by Raiffeisen as fundamentally unsound. He was partly motivated by reservations about specific features of the Raiffeisen cooperatives, including their internal structure, their methods, and their regional organizations. Schulze-Delitzsch was perhaps most worried by the Raiffeisen practice of funding long-term loans (often extending ten years or more) with deposits that had, at most, a three-month recall provision. Raiffeisen also shunned the large capital contributions that were typical of Schulze-Delitzsch's own cooperatives, which meant the new ones relied almost entirely on deposits until they built up reserves out of retained earnings. This aspect of Schulze-Delitzsch's critique was just a recitation of banking orthodoxy. By accusing the rural credit cooperatives of being insufficiently orthodox, his comments resonated with bankers and made his own cooperatives seem sound.3
However sincere Schulze-Delitzsch's views, he was driven to criticize Raiffeisen's cooperatives (which applied, by extension, to most other rural credit cooperatives) by other motives as well. Schulze-Delitzsch was disgusted by the overtly Christian tone of Raiffeisen's group, viewing it as at best woolly-headed and at worst a danger to his own efforts to promote cooperation as a sensible, pragmatic way to ameliorate the conditions of the working classes.4 Schulze-Delitzsch probably also feared that the rural movement would grow to eclipse his own in size and importance. This in fact happened.
During Schulze-Delitzsch's life, and continuing even more intensely after his death in 1883, leaders in the several branches of the cooperative movement carried on extensive, vituperative, and often misleading debates about the strengths and dangers of different types of credit cooperatives. Each branch of the movement had its national and regional publications, which seldom missed the opportunity to extol the virtues of their own type while casting aspersions on others. The criticisms both reflected and contributed to a public sense that the various branches were like "brands." Problems in one Raiffeisen cooperative could be exploited to suggest that all were weak. The most instructive articles concerned troubled cooperatives in towns with common names. For example, on one occasion the Raiffeisen newspaper was at pains to clarify that a recently failed cooperative was not that in Gunfimgen (Hesse-Cassel), one of theirs, but in Gunfmgen (Rheinhesse), someone else's.5 This sensitivity to confusion about the identity of failed institutions reflected long years of public efforts to advance one brand over another. The most public critics of credit cooperatives were leaders of cooperative associations. Could the public then be blamed for wondering about the soundness of the entire idea?
The consequences of worries about credit cooperatives were serious. Bad news could scare away potential depositors, who provided most of the loan funds in rural cooperatives. The greater concern, however, was that negative claims would deter potential members, making it harder to form new cooperatives and to expand existing ones. All members bore some liabilities for the cooperatives' debts (primarily deposits). In an unlimited-liability cooperative, this obligation theoretically amounted to all of the member's assets. In limited-liability cooperatives the risk was less, although it was still substantial, as they tended to have large share values. As a practical matter, few cooperatives failed. In those that did, the liability contributions required to make good the institutions' debts were small. But the possibility that membership in a cooperative could wipe out a family's assets was real. The aspersions cast on the various branches of the cooperative movement were motivated in part by ideological differences and personal jealousy, but they had the potential to inflict real damage on the movement. Thus, it became necessary to find ways to counteract the claims.
The Cooperative Movement and Its Regional Institutions
By the 1880s there were three dominant groups of credit cooperatives in Germany, reflecting the activities of the three men most responsible for this movement. Schulze-Delitzsch (1808-83) founded several cooperative associations during the 1840s and 1850s. By 1861 there were 364 Schulze-Delitzsch credit cooperatives, with nearly 49,000 members.6 Raiffeisen (1818-88) was at first an imitator of Schulze-Delitzsch. Raiffeisen's first credit cooperative was founded in Neuwied (Prussian Rheinland) in 1864. The number of Raiffeisen cooperatives at first grew rapidly, but they were later eclipsed by cooperatives affiliated with a group formed by Wilhelm Haas in the 1870s. At the outbreak of World War I the Schulze-Delitzsch group numbered some 945 credit cooperatives, with 620,000 members; the Raiffeisen group had 4,400 cooperatives, with 485,000 members; and those affiliated with the Haas group, 11,165 cooperatives, with more than l million members.7 Ordinarily only one person per household could join a credit cooperative, so these membership figures imply that millions of Germans were part of the movement.8
Schulze-Delitzsch thought credit cooperatives should be primarily urban, should discourage the poor as members, and should have some paid, full-time management. When it became legal in 1889 (after his death), his followers advocated that credit cooperatives convert to a limited-liability basis. Raiffeisen, on the other hand, worked in rural areas, wanted cooperatives to be open to everyone in the community, and thought that only a part-time treasurer should be paid. His group insisted on unlimited liability for credit cooperatives, even after 1889. Haas and other cooperative leaders were less doctrinaire, but most credit cooperatives outside Schulze-Delitzsch's group were rural and resembled Raiffeisen's.
Most groups developed two regional institutions to help manage the problems their local cooperatives faced. A regional bank (called, confusingly, a "Central" bank) took deposits from and made loans to its member cooperatives. The Haas group had many regional Centrals, while the Raiffeisen group had a single institution with several branches. Members of the Schulze-Delitzsch group did not develop Centrals in the same way, and they relied on commercial banks for liquidity services. The other regional organizations were the auditing associations.9 At the national level there was a single cooperative bank, founded by the Prussian government, that acted as a sort of Central's central. Each of the large cooperative groups was organized within a national federation that combined all their constituent institutions, including cooperatives engaged in activities other than credit.
Surviving sources allow a close look at both individual cooperatives and the regional associations that worked with them.10 Throughout the article, these are called "study" cooperatives and are referred to by the name of the village in which they were located. I also draw on the annual reports of auditing associations and on a lively cooperative and academic press.
The Local Cooperatives. Credit cooperatives shared internal organizational features, regardless of their type, in part because of legal requirements. The management committee (Vorstand) represented the cooperative legally and made most important decisions on matters such as accepting new members and granting loans. The supervisory committee (Aufsichtsrat) met less frequently to oversee the management committee. The supervisory committee's primary role was to review and approve the management committee's decisions and to act as a sort of internal auditor.11 Members of these two committees served terms of fixed length, and no single person could be on both committees at once. The cooperative's entire membership (Generalversammlung) met annually to elect the members to the two main committees and to make decisions on basic policies, such as interest rates. Membership in the cooperative was not automatic, but once accepted all members participated on an equal basis in elections for management positions and on issues put to a general vote.
The cooperative's daily business activities, as well as its bookkeeping, were handled by a treasurer, who in some cases was a designated member of the management committee, and in others, a cooperative member selected by the management committee. In the Raiffeisen group, and in many Haas cooperatives, the only paid employee was the part-time treasurer. Some other rural cooperatives paid their managers a small fee for attending meetings.12
During the period from the 1870s through the outbreak of World War I, the credit cooperatives were simple banking operations. Most of their assets were loans. Rural cooperatives were highly leveraged; as much as 90 percent or more of their liabilities were deposits, whereas urban credit cooperatives typically had lower leverage because of their larger member shares. Most transactions involved cash. (Some credit cooperatives established checking and other payment services in the early twentieth century, but this practice remained rare prior to World War I.) Managing this institution entailed three problems: The first was fraud, as there was ample opportunity for a manager to steal from the cooperative by taking cash or falsifying records. Embezzlement by treasurers was rare but remained a constant concern. The second problem was record-keeping, as, in an era that predated automatic information systems, the mechanics of tracking the inflows and outflows of funds was an enormous undertaking. The cooperatives also were legally required to maintain promissory notes, minutes of meetings, lists of members, and other internal information. A final, and important, class of problems arose out of honest mistakes. Admittedly a residual category, these problems were the consequence of decisions that were not attempts to take advantage of the cooperative and did not simply represent failure to keep adequate records. They are best thought of as reflecting the managers' imperfect grasp of the problems their institution might face.
The Managers. Upon election, most members of the two committees were taking on a position that involved them in formal recordkeeping and reporting for the first time. How were they recruited, and how long did they have to learn the job? Lists of all members of the management committees and the dates of their service appear in some cooperative Festschriften, permitting us to construct measures of turnover. Nine Festschriften have complete information on the management terms in the period from their formation up to 1914.13 These histories suggest that the members of both committees served for very long times. In four of nine cooperatives, 100 percent of all terms on the management committee were for at least five years. In all nine cases there was no turnover at all on the management committee in 80 percent of the years between the cooperative's founding and 1914 (the youngest cooperative in this group was formed in 1899, but most date to the 1880s and earlier). One of the few individuals who left the Leer management committee died in office and was replaced by his son. Although the duties of the supervisory committee were more varied, and terms tended to be shorter, the median term for a member of either of the leadership committees in these cooperatives was about fifteen years. Given that, of the two, the responsibilities of the management committee were more time consuming, it seems unlikely that the minimal turnover recorded was due to the burdens of office.
There are several interpretations for these long, unbroken terms in office. The first is that the managers worked out well and the cooperative saw no reason to replace people when elections were held. A second interpretation would be that so few people in a given village were willing and able to serve in such a position that turnover was impossible. The references to elections contained in the minutes of the Generalversammlung meetings do not support this latter view. Some elections for the management and supervisory committees were contested. Even people who were elected by significant majorities faced real opposition. To take one fairly typical example, in 1912, in the Huttersdorf (Schmelz) cooperative, Peter Gross was elected treasurer with 158 out of 271 votes. But he had four opponents, one of whom garnered 64 votes.14
A third interpretation of these low turnover rates would be that the small group of people running the cooperative were using the organization for their own ends. This view, which is a theme in some contemporary criticisms of the cooperative movement, is hard to reconcile with contested elections and with a lack of evidence of managerial abuse. I have found no instances of managers using their positions to secure the nineteenth-century equivalent of a corner office. In some instances, cooperatives paid for their managers to travel to a training session in a city, but the amounts reimbursed hardly suggest a lavish holiday. For example, the management committee of the Aussen (Schmelz) cooperative was granted three Marks each to cover travel costs associated with attending a cooperative school.15 There is also little evidence that managers received a disproportionate share of loans. Indeed, to be given a loan, management committee members had to receive the approval of the supervisory committee. Of all the study cooperatives, only the Leer management committee received many loans, but in that cooperative a large share of deposits was forwarded to the Central as a matter of course. Thus, loans to the managers did not deprive other members of funds.
The membership in credit cooperatives was remarkably stable. A few members were ejected for failure to repay a loan or for some other reason, but this was rare. Most members who left a cooperative did so because of death, and a handful resigned when they moved from the area in which the cooperative operated. Thus the annual meetings were attended by individuals with ever-growing experience in the workings of the institution. On the other hand, the minutes of the annual meetings show that attendance could be poor-as little as one-half the current membership. The Aussen (Limbach) cooperative tried several times, and failed, to change its statutes, because it could not get the required quorum to attend the annual meeting. In one year, only 44 of in members showed up.16 Cooperatives took several approaches to the problem. One was exhortation: articles in the cooperative press on the importance of the general meeting were common. Another was to provide financial incentives. Some cooperatives paid members to attend the annual meeting or fined those who did not attend. The Gersbach (Maulburg) cooperative, for example, fined members forty Pfennig for neglecting this obligation.17
Nearly all accounts, both critical and supportive, of the cooperative movement agree that the treasurer was the most important manager and that problems in individual cooperatives largely reflected this person's failings. The treasurer was assigned a considerable range of responsibilities. He had to keep track of all deposits and loans, manage the cooperative's accounts at the regional Central, prepare promissory notes (and mortgages if the cooperative was making mortgage loans), and compile monthly and annual reports for the internal managers and the auditing association. All this was done without the assistance of any mechanical calculation devices.
Most cooperatives required their treasurer to post some kind of bond or other security commitment against fraud or embezzlement. For treasurers with assets, such as a farm, this was not a serious problem. But for the many who were teachers it was a serious issue. The most common solution revealed in the study cooperatives was for someone else to sign a note, the same kind of commitment a cosigner would make. In the Adelhausen (Maulburg) cooperative, for example, the treasurer's honesty was guaranteed by a thousand-Mark note of this sort.18 In his cooperative handbook, Raiffeisen says of the treasurer: "It is important to select a man of reliable character, who has as much business ability as possible, who enjoys widespread trust, and, where possible, has means. That last point will, in addition to his personal qualities, reassure the cooperative."19
Cooperatives recognized the importance of paying a salary high enough to attract and retain a treasurer with the right qualities. Higher pay meant, implicitly, reduced reserves for the cooperative, higher interest charges for borrowers, and lower interest payments to depositors. Cooperatives often devoted half or more of their net interest income to this purpose. Many treasurers were schoolteachers. In one study cooperative, the schoolteacher-treasurer's cooperative income amounted to about one-quarter of his earnings from teaching school. Being a treasurer was a part-time job but represented significant income.20
Cooperative manuals advised against paying the treasurer according to a formula that gave him an incentive to increase the number of loans or other transactions. Anton Quabeck, an influential leader in the Westfalian cooperative association, prepared a handbook in which he wrote that if the treasurer's pay was a fixed percentage of the cooperatives' income, then "he would have an interest in the cooperative earning as much as possible, and could pressure the management committee to this end." Whatever the advice, there are examples of explicit formulas of the sort Quabeck feared. The Gersbach (Maulburg) and Maulburg cooperatives both set their treasurers' pay by linking them explicitly to the cooperative's turnover.21
Treasurers were the central characters in most cooperative horror stories. Few rural treasurers had any business experience at all. Cooperative manuals introduced double-entry bookkeeping as if it were an unknown concept (and why should it be familiar to a schoolteacher?). More than a few record books are filled with the red corrections of an auditor. There were also cases of extreme laxness. Sometimes an auditor would find that a treasurer had not been keeping an entire set of required books or had failed to provide the right report at the specified time.
But the real fear was embezzlement. The cooperatives operated primarily on cash. Only in the early twentieth century did they begin to use checks and other payment means, so a clever treasurer could steal, under the right conditions, thousands of Marks. Embezzlement by cooperative treasurers was rare, although it did not have to be common to pose a threat to the movement. In examining the sources on this issue, it is difficult to separate real concerns about embezzlement from two rhetorical points. The Schulze-Delitzsch group often referred to the combination of unpaid managers and part-time treasurers in rural credit cooperatives as an invitation to trouble and thus tended to exaggerate the embezzlement problem in rural cooperatives. For their part, the rural auditing associations stressed the value of their services by highlighting cases in which they had detected embezzlements before serious harm was done. Whatever their rhetorical import, embezzlements were real. One took place in the Maulburg study cooperative. At a joint meeting of the management and supervision committees in 1899, the auditor reported that the late treasurer had helped himself to about 1,275 Marks of the cooperative's money, and he recommended several steps to help cover the loss. At a general meeting the same day, the auditor reminded the membership that the management committees and current treasurer should not be held accountable. Since no members had objected to the (fraudulent) presentation of the previous year's report, they were equally responsible for the shortfall. Several days later, the widow and children of the deceased treasurer signed an agreement to repay the embezzled funds.22 This is a simple tale with a happy ending. More lurid accounts in the cooperative press suggest that treasurers in larger cooperatives got away with a lot more. The Schulze-Delitzsch organ reported one case of a 7,500-Mark embezzlement in a rural cooperative in 1901 and suggested that the failure of another, representing losses of nearly 100,000 Marks, reflected the relatively great power of treasurers in cooperatives whose other managers were not paid.23
An article published in 1909 by Quabeck in the Haas group's organ, Genossenschaftspresse, gives a generous and revealing picture of how embezzlements affected the cooperative movement. He notes that cooperatives from every group faced this problem, and that the damage to an affected institution went far beyond financial matters. Embezzlements contributed to mistrust of cooperatives by outsiders and destroyed trust within the affected institution. Echoing a common complaint, he noted that although other banking institutions suffered the same problem, for some reason embezzlements elsewhere did not receive as much attention. He concluded with a call for vigilance, but refrained (unlike many others) from claiming that cooperatives of any one type were more or less prone to embezzlement.24
Embezzlements were the most dramatic and well-publicized management problem in credit cooperatives. But they were rare, and few cooperatives closed as a result. Most of the funds stolen seem either to have been recovered or to have been so small that the cooperative was able to make good the loss through retaining earnings. This ability to contain embezzlements no doubt reflects the same kind of information that enabled the cooperatives to lend on good terms. Treasurers lived in a small community, making it hard to conceal the fruits of embezzlements. The cooperatives also had the benefit of outside assistance in uncovering these crimes.
Help from the Regional Level. The local cooperatives' managers required four forms of assistance: First, they needed someone to correct errors they made. Second, they needed education in business record-keeping and the specifics of cooperative practice. Third, because the structure of the local cooperative apparently could not provide the right incentives for the managers to perform their jobs properly all the time, coming up with mechanisms that improved those incentives would further the cooperatives' growth and development. Finally, the cooperatives needed a way to signal to others, including potential members and depositors, that they were safe and well run. The auditing associations established by the cooperative groups were an attempt to meet all four of these needs.
Cooperatives were the first business enterprises in Germany to face any legal requirement for an external audit.25 Regular auditing was an outgrowth of the way the cooperative movement was first formed and developed. From their earliest days, the leaders of the cooperative movement promoted regional associations of cooperatives to share information, lobby governments for legal changes to make it easier to form and run cooperatives, and to launch new ones. In most groups, these regional associations encouraged an informal audit of member cooperatives. As the movement grew and public awareness of criticisms leveled against cooperatives became a more serious concern, some associations began to make periodic audits a condition for membership. Most of the regional associations in the Schulze-Delitzsch group instituted this requirement in the 1870s. Some rural associations had followed suit by the early 1880s. But before 1889 these audits had no legal status.26
External auditing was closely tied to nineteenth-century developments in corporate forms that separated ownership from the control of an enterprise; the auditor's role was to guarantee to stockholders and others that the managers were reporting the firm's true condition. The English Joint Stock Company Act of 1844 required external auditors, although not until later were sufficient details added to such provisions to make the external audits satisfactory.27 During his lifetime, Schulze-Delitzsch successfully resisted compulsory audits for cooperatives. Calls for compulsory auditing of German cooperatives typically came from the movement's enemies, and they took the form of demanding examination by state officials. Several bills to this effect had been introduced in the Reichstag in the early 1880s.28 Schulze-Delitzsch was aware of English auditing practice and impressed by the auditing requirements of England's friendly societies. He opposed state audits for practical and political reasons. The practical objection was a concern about moral hazard, as the cooperative members and the supervisory committee might consider a compulsory audit a substitute for their own vigilance. On a political level, he had a well-grounded suspicion of state involvement in credit cooperatives and viewed compulsory audits by state officials as the first step in making them de facto arms of the government.29
As pressure grew for compulsory auditing, the several cooperative branches relented in their opposition. Even Schulze-Delitzsch's group came around to supporting compulsory audits after his death. The cooperatives' support for compulsory audits reflected the leadership's judgment that problems in individual cooperatives, even if rare, could seriously undermine them all. The Schulze-Delitzsch group thought the rural cooperatives were flawed and saw the imposition of this legal requirement as one way to prevent them from doing damage to the entire movement.30 Many cooperative leaders also feared that if they did not propose a constructive solution to the problem, cooperatives would be subjected to close state oversight, which few in the movement wanted. The 1889 cooperatives law, which was the first such law that applied to the entire Reich, was thus a partial victory for the cooperative movement. Article 51 of this Act required that every cooperative be audited at least once every other year. A cooperative could satisfy the requirement by asking the local court to appoint an auditor for that cycle only, or it could have the audit performed by someone who was an employee of the regional association. Thus the 1889 law transformed the old regional groups into recognized auditing associations. The law, in fact, represented the government's judgment that the auditing procedures in use already were sufficient, and the remaining problem was to encourage all cooperatives to join the auditing associations.31
Each group had its own specific practices, but in every case auditing associations were either restricted to a region or had "subassociations" that were assigned to each region. In the rural cooperative groups, auditing associations usually had a formal relation with a Central cooperative bank, but the two institutions were distinct, and in theory a cooperative could belong to the auditing association and have no connection to the related Central. Member cooperatives paid annual dues plus a charge for each audit. The auditing association also earned income from the sale of books, printed forms, and other items. Most auditing associations did not cover their full costs through these devices; they made up the deficit with gifts from their affiliated Central and through modest grants from government agencies. These government grants were the only direct state support for the cooperatives in the period prior to World War I, and they were trivial in comparison to the total costs faced by the cooperative movement. Many state governments would make one-time grants to various cooperative organs, but regular contributions went almost exclusively to the auditing associations. The Saarland's cooperative association is fairly typical in this regard. In the period 1900-08, it took in 65,935 Marks as direct charges for audits and another 95,465 Marks in membership fees. Only 11,400 Marks came from state sources. This left a deficit of about 23,000 Marks, which was made up from profits generated by the Central cooperative bank.32 The total state grants to cooperatives made by Prussia for the period from 1896-97 through 1903 amounted to some 36,400 Marks, or about 4,500 Marks per year.33
The auditors at first were individuals with experience in the cooperative movement. By the late 1890s, the several cooperative groups were experimenting with more formal ways of training auditors. In the early twentieth century, the Haas group established its own school for cooperators, which ran several different training programs, including courses designed for local managers and treasurers. The Raiffeisen group started with an apprenticeship system, whereby experienced auditors acted as mentors for prospective auditors.34 But they gave up this approach and in 1913 began a regular course for new auditors. In that first year, the course ran from October 2, 1913, through February 14, 1914, and trained a total of fourteen men. The youngest was twenty and the oldest thirty-five, suggesting that these new auditors planned a career in the movement. The topics were revealing. The 516 hours of instruction were broken down into the history and organization of the cooperative movement (60 hours), cooperative law (96 hours), bookkeeping (90 hours), general law (68 hours), and auditing (42 hours).
Many cooperative leaders were concerned that the regional organizations not imply any liability across cooperatives. In the case of auditing associations, there was no such liability. Even if an auditing association failed to detect problems that later led to losses, the association bore no responsibility. The imperial court strengthened this principle with a ruling in 1912.35
There were two kinds of audits. One took place via the mail: the treasurer sent all his records to the auditing association, which checked them and sent them back. Although this method was cheaper than a site audit, several observers noted that lack of direct personal contact made it less desirable than a personal visit. In the more common site audit, the auditor spent a few days at a cooperative. He checked all the cooperative's records, making sure that the arithmetic was correct, that all forms had been filled out correctly, and that legal documents (such as promissory notes) were in order. He ordinarily met with the treasurer and the two committees to discuss his findings. Later he sent a formal report to the auditing association, and the cooperative had to submit a response to his report.
Auditing associations stressed that these audits were rigorous and thorough, which of course they had to say in order to justify the expense. The amount of time that went into each audit supports the claim that they were not pro forma. Most auditing associations published summary tables that included the average number of days spent at cooperatives audited during the year. The Raiffeisen federation's report for 1906 is fairly typical. In that year, the average audit took 3.7 days. But this figure conceals a great deal of regional variation. Each audit of the Nurnberg regional subassociation took an average of 1.7 days; in Strassburg the auditors averaged 5.4 days. As the report noted, several factors contributed to these differences. In some areas, the auditors spent more time on travel to and from cooperatives. In other regions, cooperatives were more intensively involved in marketing and sales, which implied a more time-consuming audit. Most reports also noted that a cooperative that was audited every year was easier to audit each time.36
Only a few manuscripts of auditors' reports survive, but these share common traits. Auditors found and corrected simple arithmetic mistakes without comment. They noted common errors in record-keeping, such as discrepancies between the cooperative's membership list and the list filed with the court at which the cooperative was registered. More serious problems were also common. Cooperatives sometimes violated their own rules by, for example, granting an unusually large loan without the required agreement of the supervisory committee. Equally serious oversights were neglecting to have the required meetings of the supervisory committee or failing to keep records of meetings that were held. Even these more serious problems did not elicit much comment from the auditor if dealt with promptly.
In other cases, the auditor's report and the cooperative's response could be testy. Formally the auditor was someone hired by the cooperative to perform a legally mandated audit. Informally he was an envoy of the regional association, sent, potentially, to criticize respected local people. Cooperative leaders complained that auditors were too petty (kleinlich). Some local leaders charged that the auditor had to find something wrong, however trivial, to justify his performance to his superiors.37 The auditing associations and other cooperative leaders were sensitive to the fact that the external audits might be viewed as implicit criticisms of managers. Thus the auditors were often portrayed as "advisors," and the point of the audit was described less as a check than as an opportunity for an experienced professional to discuss management problems. Quabeck's introduction to audits in his manual is typical of the turns of phrase applied to these matters:
If the management committee, supervision committee, and treasurer endeavor to the best of their ability to deal conscientiously with every aspect of their responsibilities, they can only make minor mistakes. But sometimes we have to take into account that the management committee, supervision committee, and treasurer are not experts, that in some cases they do not have business experience, and that in other cases this is only a part-time occupation. This is why many have viewed it as desirable to have an expert, independent advisor stand at the cooperative's side, and at regular intervals conduct a thorough accounting and management audit.
Elsewhere he refers to the auditor as a "friend and advisor."38
When the auditor did not like what he found, the language could be direct, even harsh. Sometimes an auditor noted that a cooperative had taken no note of recommendations he had made the previous year. This is a common theme in the auditor's reports for the Limbach (Schmelz) cooperative in the period from 1901 to 1915. In an admittedly extreme case, an auditor named Schwamborn addressed himself to severe problems in a fairly new credit cooperative in Miel (now Rheinbach, Prussian Rheinland). In 1906 this cooperative had only twenty-four members and only one loan outstanding. In his report dated March 17, 1909, Schwamborn noted that the management committee had met only once in 1908, and the general membership not at all. Most deposits were in very large amounts (one for 30,000 Marks), and most of those were deposited at the Central cooperative bank. As the auditor put it, "[T]he cooperative does not fulfill any of its obligations." A letter from the Cologne auditing association (dated May 18, 1909) gave the cooperative four weeks to report back on how it would address these problems, commenting how "regrettable" it was that the cooperative's managers were not doing their jobs. In this case, the harsh language had the desired effect, because a second audit (June 15, 1909) produced a satisfactory result. Miel presented an especially dangerous situation, as a cooperative with lax management and a lot of money in deposits was an inviting target for embezzlement.39
Several auditing associations published statistical summaries of results from their audits. Most associations used consistent categories to rate the performance of various management organs within a cooperative, which allowed them to give a clear idea of what they thought were the weak points in the institutions they audited. In 1907 the Berlin auditing association of the Raiffeisen group reported results as "very good," "good," "passable," or "unsatisfactory." Its treasurers received good ratings: 57 percent were called good or very good. About 40 percent of management committees were rated good or very good. Only 35 percent of supervisory committees were evaluated this positively.40 In 1904 an auditor for the Raiffeisen group called the supervisory committee the "problem child" (Schmerzenskind) of the cooperatives, complaining in particular that such committees did not perform their internal audits thoroughly.41 Similar reports for other years and other areas tell a similar story. Treasurers were considered strong on average, while lax supervisory committees were a constant source of worry. In some cases, the only effective internal supervision came from a single individual: "It is a matter of difficulty, as the reports of many unions [auditing associations] bear testimony, to secure in the local societies [cooperatives] the effective interest of the boards of supervision, and the committees are sometimes inclined to leave all matters to the decision of their chairman."42
Some auditing associations took to publicizing their roles in detecting embezzlements. Starting in the early twentieth century, Raiffeisen annual reports included a description of the number and type of embezzlements their auditors detected. In 1906, for example, they found seven (out of 3,301 cooperatives audited). The rhetoric was important: by admitting that 1.8 per thousand cooperatives suffered embezzlement that year, they highlighted the fact that the problem was rare enough to offset the charges made by their critics but frequent enough to warrant the auditing fees. As they put it, "In seven cases the cooperative, because of the immediate action of the auditing association, suffered no or only small losses. If the embezzlement had been discovered later things might not have turned out so well."43 In 1907 they only found four embezzlements in 3,339 cooperatives audited, but the lesson was drawn more starkly: "If it were not for the auditing association, who knows how serious the damage would have been for the cooperatives affected."44
The associations had to walk a fine line in claiming credit. Auditors thought that some problems were beyond their detection but well within the control of the supervisory committee. The Raiffeisen group's report for 1908 makes this point with an account of embezzlement in a cooperative in the Berlin area. The treasurer had stolen money by withdrawing money from the Central on the cooperative's behalf and then presenting false receipts to the management committee. The supervisory committee had discovered this problem, presumably because it was aware that the treasurer had been asked to withdraw the money. The report concluded that the external auditor would have been fooled as well.45
The auditors had no direct way of forcing cooperatives to follow their recommendations. Auditing associations could, and did, in drastic circumstances, expel member cooperatives from their group, but they did not publish their findings on acceptable cooperatives, nor did they issue scores of the types used by modern credit-rating firms.46 Auditing associations that were connected to a Central cooperative bank had a more subtle tool at their disposal: the Central ordinarily asked auditors for a recommendation on loans to local cooperatives, and a negative report could result in a cooperative's being denied credit. Private auditors had no power to compel cooperatives to follow their recommendations; they could only inform the court of their findings.
Auditing associations performed services that have no counterpart in public auditing today. For example, they offered assistance in the form of monthly publications, advice books, and printed forms. These aids are included in the archives of most of the study cooperatives, and given that they were not free (their sale was, in fact, a source of income for the auditing associations), the cooperatives must have found them of some use. The monthly magazines contained a variety of material that was little more than propaganda for cooperatives. But they also carried advice columns, warnings about the dangers of lax management, and detailed responses to questions that arose in the course of audits. In 1882, for example, the Raiffeisen organ ran a long article explaining promissory notes and the obligations of cosigners.47
Many organizations published manuals that provided legal, accounting, and managerial advice for local cooperatives. On specific legal issues, such as how to write a promissory note, they were an inexpensive alternative to a lawyer. Some advice was practical and reflected experience with other cooperatives. For example, several books cautioned that frequent changes of interest rates created work for the treasurer and thus should be avoided. These advice books covered a range of topics, such as the kind of people the cooperatives could hire as treasurer and the difficulty of accounting without calculation devices; every book had detailed, step-by-step instructions on how to calculate interest payments. Others contained tables that could be used to shorten computations.48
The printed materials of most use to the treasurer were probably the model record books and forms provided by the auditing associations. Some of these were little more than a standard ledger with the cooperative's logo stamped on the cover. But others were created specifically to handle cooperative needs. A cooperative could buy preprinted promissory notes, forms for quarterly reports to the supervision committee and annual reports to the auditor, and preprinted constitutions (with blanks for local decisions, such as the length of committee terms). Reliance on these forms enforced some discipline, as they told the treasurer just what information was needed.
Auditing associations also sponsored courses to train the leaders of individual cooperatives. These programs appear to have been a late development, as they were first mentioned in the late 1890s. In 1914, the Raiffeisen federation claimed that "next to auditing, the instruction of managers is the most important task of the auditing association" (p. 60). In promoting these courses, the association noted that most treasurers were either farmers or schoolteachers with little background in bookkeeping and related matters. The yearbook gave some details on courses offered in that year. The longest, designed for treasurers, lasted four to six days and concentrated on bookkeeping. One-day courses were also offered for members of the management or supervision committee.
Advantages of the Auditing Association. The 1889 law did not require cooperatives to belong to an auditing association; they could instead hire a private auditor. This raises the question of why cooperatives joined auditing associations. Most did, as Table 1 shows. This table summarizes a series of official reports that began in the late 1890s, representing the only official attempt to provide even basic information about all credit cooperatives in Germany. More detailed information is available on a large subgroup from the reports of the auditing associations, but these reports by definition did not attempt to contrast cooperatives that belonged to associations with those that did not.
Table 1 suggests some regularities. "Wild" cooperatives were certainly larger than the average cooperative that was affiliated with an auditing association. The average credit cooperative in 1914 had 149 members. The average wild cooperative had 328. The table also suggests that cooperatives with limited liability were less likely to belong to an auditing association. Nearly all unlimited-liability cooperatives belonged to an auditing association, but this was not true of those with limited liability. (This is an interesting difference, but given the limited information available I cannot push the point very far.) Most credit cooperatives with limited liability were in cities, and the difference may reflect either Schulze-Delitzsch's reservations about compulsory auditing or the greater ease of finding a private auditor in a city.
The comparative ages of the affiliated versus the unaffiliated cooperatives as of 1906 can be calculated most simply by considering the percentage of all member cooperatives that were formed after 1890, the date at which auditing associations acquired their semiofficial status. Seventy-seven percent of unaffiliated institutions, accounting for 62 percent of members, had been formed after 1890. The corresponding figures for the Raiffeisen group were 89 percent of institutions and 82 percent of members, and for the Haas group, 87 percent of institutions with 81 percent of members.49
Hans Cruger, Schulze-Delitzsch's successor as leader of the urban credit cooperatives, claimed that, after 1889, many credit cooperatives that did not want to be audited reorganized themselves as joint-stock corporations, for which there were no auditing requirements.50 It is true that, in 1889, 159 credit cooperatives liquidated voluntarily, and 80 of these transformed themselves into joint-stock firms. Cruger attributed this worry about audits to a fear that long-concealed problems would be revealed to members and the public. His point is instructive in two ways. First, it suggests that the number of unaffiliated cooperatives listed in Table 1 is an underestimate, because many institutions simply switched their form. Second, it reinforces other complaints that auditors were too expensive, and strict, to suit some cooperatives.
This admittedly inadequate information, and the comments repeated in the cooperatives' press, supports two generalizations. The first is that most "wild" cooperatives were institutions that thought they did not need an auditing association. They were larger and more likely to have limited liability than cooperatives that belonged to an auditing association. Older cooperatives would have established their own procedures and would have built up their own reputations. They had less need of an auditing association to reassure a nervous public. A second type of wild cooperative was one that had been ejected from an auditing association for refusing to live up to the association's standards. This presumably would account for the claim by some observers that wild cooperatives were overrepresented in bankruptcy proceedings.51
We can infer more about the private auditors from debates about compulsory membership in auditing associations. The 1889 law had not entirely satisfied the cooperative movement, because it left open mechanisms for audit that were outside the control of any part of the movement. Discussion of various alternative proposals reveals some interesting nuances on the question. Cruger, who at that time headed the Schulze-Delitzsch cooperative federation, argued in 1913 against compulsory membership in auditing associations. He noted that most of the auditing associations in his group insisted on an audit before a cooperative could join, a process that led some cooperatives to be rejected. Compulsory membership (which implies that all auditing associations would be compelled to accept some undesirable cooperatives as members) would, he argued, reduce the quality of the auditing associations and the audits themselves. In a move typical of the lively debate within the Schulze-Delitzsch group, its main organ, Blatter fur Genossenschaftswesen, published a dissenting view by a Mr. Petersen in 1914. Petersen argued that it was important for all cooperatives to be audited by some association, and he thought that the element of choice present in the ability to switch from one auditing association to another was enough to preserve the only good features of being able to go to a private auditor. He noted that some cooperatives preferred a wild status because the supervision committee feared excessive interference from the auditor, implying that auditors unaffiliated with cooperatives took a more hands-off approach. He also alluded to a frequent complaint about the auditing associations, which was that the combination of membership fees and auditing fees cost the cooperative more than a private auditor. In the end, Petersen's proposal was that all cooperatives be required to belong to an auditing association, but that they be allowed to choose their association and switch from one to the other.52
The auditing associations and other cooperative groups stressed several advantages of their auditors over the private auditors. We know from complaints that auditing associations were usually more expensive (in part because they were not content with the every-other-year audits required by law).53 We can also surmise that the auditing associations were stricter than private auditors. Little is known about private auditors and how much they charged. One source says that most non-association auditors were clerks of large banks, notaries, mayors, or business managers.54 The cooperative movement argued that its own auditors were better than private auditors because of more experience and better incentives. Auditors hired by auditing associations dealt with nothing but cooperatives. They were experts in a way a notary might not be. However simple the credit cooperatives were, it was unlikely that most auditors with another background would, in their other line of work, see the full range of issues that confronted these tiny banks. The association's auditors also had a stake in the future of the cooperative movement. In his handbook, Quabeck contrasted the auditing association to another auditor who might treat the affair as a bureaucratic necessity. The failure of a credit cooperative could damage the group and limit the advancement opportunities for an auditor employed by that group. The career prospects of cooperative auditors depended directly on their ability to head off trouble among their association's members.55
The auditing associations were also aware that policemen themselves required policemen. In the period prior to World War I, all the cooperative groups were experimenting with mechanisms to provide oversight for their own auditors, including reaudits and the installation of "superauditors," who would check the work of primary auditors. The association publications do not elaborate on the kinds of problems these monitors encountered.
To appreciate fully the potential advantages of the association auditor, it is necessary to step back and consider the larger problem facing credit cooperatives. They were the subject of considerable criticism. Leaders of other banking institutions resented them and considered their entire operations unsound. The public (that is, potential members and depositors) were aware of this criticism. Perhaps more important, the quarrels within the movement created, at least in some quarters, a strong sense that all cooperatives of a given type (such as Raiffeisen and Haas) had characteristics in common, whether strengths or weaknesses. These shared perceptions imply a reputational externality within the cooperative movement. When cooperative A decided to make a loan, it considered the impact of that loan on its own future. But the management committee of cooperative A had no reason to take into account the damage that bad news created by their cooperative could inflict on other cooperatives. The auditing association helped to internalize that externality. One reason for complaints about the association being too tough was that, in making recommendations, it took into account the costs of damage to the entire group that could result from bad news in one cooperative. There is a sense in which each cooperative wanted other cooperatives to belong and to be subject to the association's discipline, while preferring to remain exempt from the discipline itself.
Another advantage of auditing associations was their ability to take care of problems quietly and internally, a feature of auditing practice that reflects the concern with reputation. In the 1899 embezzlement case in the Maulburg cooperative noted above, the auditor worked out an agreement with the late embezzler's widow to repay the cooperative for what he took. The cooperative did not fail, and the cooperative did not have to become involved in a public confrontation to get its money back. While a private auditor might have handled the matter similarly, he would not have had the same incentive to solve the problem quietly. One could argue, in fact, that a private auditor might have wanted the incident publicized in order to trumpet his own competence and garner additional business from other clients.
It is not possible to test for the effectiveness of the auditing associations based solely on material they published. Little is known about wild cooperatives, and thus it is not possible to form meaningful comparisons. But the associations' effects were in all likelihood far-reaching and subtle. I have stressed the importance of oversight and discipline, but there were other features. One author argues that effective external supervision helped promote cooperation by making limited liability more workable. He argues that unlimited-liability structures signaled to potential depositors that the cooperative members were serious about their institution, but that this same unlimited liability scared off potential members. External auditing provided the feeling of safety depositors wanted, while permitting cooperatives to switch to limited liability and thus enjoy a wider membership.56
Summary and Conclusions
Rural credit cooperatives in the nineteenth century were run, by necessity and intent, by people with little formal business experience and almost no bookkeeping training or experience. Reliance on these managers presented the possibility of fraud and the less dire consequences of faulty bookkeeping. The institutions dealt with these potential problems through a combination of local and regional responses that were characteristic of the system. First, they paid their most important manager, the treasurer. Second, they tended not to replace members of the management committee at elections. There are several possible interpretations of this behavior, but the low turnover of managers undeniably meant that the cooperative built up, over time, a cadre of experienced people in these positions. Finally, the cooperative system developed regional auditing associations that helped to check and train local managers. Some of the auditor's responsibility was simple auditing in the sense that we understand that term today, but many took a more active role, providing the local cooperatives with helpful materials and eventually setting up formal training courses for cooperative managers. The solution to the problems posed by local managers is typical of the system: the cooperatives drew from the strength of local knowledge and social ties and used regional organizations to offset some of the local managers' limitations.
The auditing associations contributed directly to the growth and stability of the credit cooperatives by allowing them to overcome serious problems implicit in their local design. The associations thus helped to promote the primary goal of cooperative banking, which was the provision of credit to people who would otherwise have had to turn to informal lenders. The auditing associations also promoted a serious secondary goal of the cooperative movement, which was the education and training necessary for its members to become practical businessmen. The local cooperatives themselves trained, informally, many thousands of managers and treasurers. There is no way to ascertain how much impact these people had on other aspects of their communities and the German economy more broadly, but it is impossible to believe this training did not matter. The auditing associations also offered more formal instruction to their own employees, giving the movement a practical and educational depth that surely helped it to weather the storms of the twentieth century.
The auditing associations were part of a solution to a problem at first created largely by the cooperative movement itself. Skepticism about these small banks came from many quarters, but the strongest and most damaging criticism arose from the squabbles within the cooperative movement. Schulze-Delitzsch's withering attacks on Raiffeisen's cooperatives, and the many replies in kind, created an atmosphere in which any bad news about any credit cooperative could be used to create a fear of credit cooperatives as an institution. The informal audits that preceded the formal auditing system created in 1889 were designed in part to deny the enemies of each cooperative group the bad news that fueled their ability to cast aspersions on cooperative practices. And here lies an important irony: many cooperative leaders at the time decried the Systemstreit as a waste of time and energy and worried about its impact on the larger view of cooperatives among the German public. But it may well be that one reason the German cooperative movement thrived was that this competitive mechanism forced everyone in the movement to take seriously the need to train and oversee the institutions' managers. The German cooperative system in the nineteenth century had an unusual degree of competition between institutions that provided very similar services. The attacks on each other's methods, and the institutional responses those attacks provoked, stand as testimony to the value of competition, even for cooperative institutions and even when that competition took the form of claims that were often wild and uninformed.
The cooperative system faced serious challenges and surmounted them to become a vibrant, safe part of the German banking system. Their auditors played, to use modern terminology, the role of both auditors per se and management consultants, a practice that today, when occurring in a single firm, is viewed as leading to conflicts of interest. In the accounting scandals of 2002, some auditors were accused of covering for their clients in order to secure more lucrative management-consulting business. What was so different about the German cooperatives? Cooperative auditors were not competing with other auditors for the "consulting" business; these services were bundled with the auditing, and a cooperative could not easily shop. No auditor had a reason to neglect his auditing responsibilities. At the end of the day, the cooperative auditing associations were so firmly tied to their members' health and reputation that no amount of non-auditing income could be worth risking damage to public confidence in the local cooperatives. The German system worked not because the auditors were more independent-the remedy often proposed for public accountants today-but because they were so firmly tied to their clients that they had no reason to engage in behavior that might endanger their movement. That a relative lack of independence can be salutary may well speak to the peculiarities of this institution and period, but it may also hold larger lessons for the problem of ensuring that auditors do their job.
A Note on Sources
The local cooperatives that supplied the manuscript business records used in this paper do not maintain formal archives. Rather, the material is kept in cartons on the premises of the successor credit cooperative. Thus no formal cataloguing numbers can be provided. Given the small amount of material available for each cooperative, it is sufficient to identify the documents, as I have here, by their name and date.
The most important documents for purposes of this paper were the minutes of the Vorstand, or management committee, and those of the membership's annual meeting. Formally entitled Protokollbuch fur den Vorstand, the former is here abbreviated PB Vor and identified by the name of the credit cooperative. Thus Protokollbuch fur den Vorstand, Maulburg, 28.7.1902, is the minutes of the management committee for July 28, 1902, for the credit cooperative in Maulburg. Gersbach (Maulburg) indicates records from the now defunct credit cooperative in Gersbach that has been absorbed into the survivor in Maulburg. The minutes of the general annual meeting, the Protokollbuch fur die Generalversammlung, are abbreviated PB Gen and are identified in the same way.
[Author Affiliation]
TIMOTHY W. GUINNANE is professor of economics and history at Yale University.
An earlier version of this paper appeared, under a different title, in the Yale Economic Growth Center's Discussion Paper Series (No. 824, 2001). I appreciate the financial support of the National Science Foundation and the German Marshall Fund of the United States, as well as the kind assistance of several institutions associated with Germany's credit cooperatives today. Participants in workshops at UCLA, the Bank of Italy, the University of Arizona, and University College Dublin and All Souls College, Oxford, provided useful feedback, as did three anonymous referees, Bruce Carruthers, Robert Evans, Naomi Lamoreaux, Alessandro Lizzeri, Carolyn Moehling, Stephen Morris, Steven Nafziger, Cormac O Grada, and Jochen Streb. This paper was written while I was a Visiting Scholar at the Russell Sage Foundation and revised while I was the Pitt Professor at the University of Cambridge.
Business History Review 77 (Summer 2003): 235-264. (C) 2003 by The President and Fellows of Harvard College.
