MAY/JUNE 1998

Caroline Fohlin is an assistant professor of economics at the California Institute of Technology. The author is grateful to Lance Davis, Michael Edelstein, Barry Eichengreen, John Latting, Peter Temin, and David Wheelock as well as conference participants at the St. Louis Federal Reserve Annual Economic Policy Conference for helpful comments and discussions, and to the National Science Foundation for funding this research.

The structure of the German , in Banking Systems particular, has been viewed as a key ingre- dient in ’s industrial development and Economic before World War I. Universal banking, because it combines all phases of finance Growth: Lessons in one institution, is thought by many to have yielded economies of scope and greater from Britain efficiency. Such efficiency has been argued in turn to have increased the volume and and Germany in reduced the costs of finance, thus pro- moting industrial investment.2 Furthermore, the Pre-World German banks are often assumed to have maintained close, long-term relationships War I Era with industrial firms. Equity positions are thought to have aligned the incentives of banks and firms and encouraged multi- Caroline Fohlin period optimization of their behavior. In contrast, a long line of detractors has chas- egulation of the financial system has tised the British banks for avoiding engage- occupied economists and policymak- ment with domestic industry and leaving Rers since the beginning of financial firms to seek financing from other sources. history. Such attention has been warranted Firms’ resultant recourse to securities 1 Greenwood and Smith (1997) because of the crucial role of these institu- markets is argued to have served investors’ offer what may be the most tions in economic life. There remains, short-term profit motives at the expense of reasonable compromise on the however, much disagreement over the long-term growth.3 As a result, the banks question of causality: a model ways financial and real variables interact have been blamed for the apparent under- in which financial markets arise and the extent to which financial develop- performance of the British economy since after some period of real devel- ment can promote economic growth. the late nineteenth century. opment, and in which the Modern growth theory has made strides Two lines of historical investigation expansion of those markets in modeling the relationship between may shed light on the continuing debates fuels further real growth. A logical implication of this model financial development and economic over the relative efficacy of German and is that exogenous creation of a growth, but a causal relationship is British banking systems. The first step is financial system with advanced 1 difficult to verify empirically. to determine whether the German banks features may not spur real Despite the burgeoning research on offered the advantages that have been growth. The problem then for finance and growth, the importance of ascribed to them; the second step is to implementing development financial-system structure has yet to be ascertain whether the provision of these policy is determining how to determined. Much of the debate over services by universal banks fueled economic get poor countries to the point banking reform in the United States growth. In comparing the two systems, at which financial systems will hinges on the assumption that certain however, it is important to acknowledge arise endogenously. types of financial systems allocate an econ- that the British banks were not prohibited 2 Most recently, Calomiris (1995) omy’s resources more efficiently than from combining functions or from pur- has advanced this idea. others. There is a widespread sense in suing long-term relationships with industrial 3 For a review of the literature the United States and Great Britain that firms. Thus, research on the real effects of on British banking and indus- the universal banking systems of Germany financial structure must accept that, if the trial development, see Collins and Japan have given those countries an British banks’ organization and activities (1991). Also see Capie and advantage in industrial development and were suboptimal for industrial growth, Collins (1992). For a critical economic growth over much of the past such inefficiency stemmed from market appraisal of the British banking 150 years. failures of one sort or another: rationing system, see Edwards (1987).

FEDERAL RESERVE BANK OF ST. LOUIS

37 MAY/JUNE 1998

relatively low-return or high-risk ven- past decade to formalize the link between tures or failing to perceive or act upon financial-system functioning and the growth favorable prospects. of the real economy.4 In comparison to the This study uses aggregate bank balance traditional growth models—in which output sheet data to investigate systematic differ- was seen as a function of capital, labor, and ences in the financial makeup and activities disembodied technological progress—the of universal and specialized banks. By current models provide a richer frame- explicitly comparing British and German work for interpreting the potential impact banks, it takes steps toward quantifying the of financial systems. For their motivation, possible disparity in financial-system growth nearly all appeal to the observed correla- effects over the decades leading up to World tions between financial-system develop- War I. Financial systems are thought to ment and industrial growth uncovered by influence both the quantity and quality of economic historians and development investment. Thus, this paper first measures economists during the 1960s and ’70s.5 the rate of expansion and the ultimate mag- Pagano (1993) provides a simple way to nitude of capital mobilized by British and summarize the newer models of finance and German banks. The study then investigates growth. Using several simplifying assump- the makeup of the banks’ asset portfolios tions, the model yields the growth rate of and estimates the extent of direct involve- output per capita as a function of three vari- ment by the two types of banks in ables: savings rate, return on investment, equity ownership. and costs of intermediation. Thus, financial The findings suggest that, compared institutions may enhance economic growth to British banks, German banks maintained by raising the total quantity of financial cap- at least as much liquidity relative to their ital available to entrepreneurs, improving short-term liabilities, mobilized a smaller the quality (productivity) of investments, share of the economies’ capital, and held and increasing the efficiency of intermedia- approximately the same (small) proportion tion (lowering costs) between the sources of their assets in the form of nongovernment and uses of funds. securities. Furthermore, the German banks This framework can help in comparing seem to have held only a limited number of the effectiveness of the German and British industrial equities in their portfolios and banking systems, but further refinement is often did so merely because of insufficient required to clarify the ways financial insti- markets for new issues. tutions affect the variables in the growth The results offer insights into both formula. The following sections take some differences and similarities in the organiza- first steps at comparing the impact of spe- tion of banking in Germany and the United cialized and universal banking systems on Kingdom, specifically, and into the historical the quantity and quality of investment. importance of financial structure, more gen- erally. The findings suggest that the gulf between specialized and universal banking QUANTITY OF INVESTMENT in terms of their influence on economic Banks influence the accumulation of growth and industrial development is physical capital by directing funds to entre- less than commonly believed. preneurs who wish to invest. Such capital 4 For an overview of some of mobilization proceeds in two stages: capital the literature, see Pagano collection through deposit-taking or sales (1993) and Galetovic (1996). THE LINK BETWEEN BANK of equity shares, and fund dispersal through Greenwood and Smith (1997) STRUCTURE AND GROWTH loans or advances. By repeating this process, provide more technical details. The primary purpose of banks is to the banking system multiplier expands the 5 Cameron (1967), Goldsmith mobilize otherwise idle resources for use money supply and redistributes the econo- (1969), McKinnon (1973), in productive investment. A wide array my’s capital. These banking functions and Shaw (1973) are the of theoretical models has appeared in the increase the share of resources targeted to standard references. growth and development literature in the productive investment.

F EDERAL RESERVE BANK OF ST. LOUIS

38 MAY/JUNE 1998

The German universal banks are cred- Figure 1 ited with mobilizing significant amounts of capital from the public and thereby pro- Total Joint-Stock Bank Liabilities Less moting industrial growth. The British banks, Total Cash Holdings, United Kingdom by comparison, are typically presumed to and Germany, 1880-1913 have participated less aggressively in the Liabilities Less Cash/GNP accumulation of funds. Total assets of finan- 0.6 cial institutions as a share of gross national product grew substantially in both Britain 0.5 U.K. and Germany between 1860 and 1913, but 0.4 Goldsmith’s (1969) figures indicate that this ratio expanded more in the latter than in the 0.3 former. Furthermore, while Britain’s ratio 0.2 exceeded Germany’s in 1860, the British 0.1 lagged the by 1900. The gap Germany 6 grew to over 50 percent by World War I. 0 Nonetheless, the Goldsmith data indicate 1880 83 86 89 92 95 98 1901 04 07 10 13 that the British deposit banks accounted for NOTES: The British data come from the Economist series as reported in Sheppard (1971) and a greater share of their country’s GNP than include private banks, starting in 1891. The solid line represents an estimation of the joint- did the German universal banks at each stock banks' liabilities from 1880 to 1913, based on the ratio of private to joint-stock banks in 1891, but that ratio likely declined significantly between 1891 and 1913. The German data point in the pre-war era. report only joint-stock banks for the whole period. Since the private banks accounted for a In Germany, virtually all of the func- greater share of bank assets in Germany, the omission of private banks may exaggerate the British lead. Even if estimated figures for the German private banks are added, however, some tions relating to corporate finance fell gap in liabilities less cash still remained as late as 1913. Furthermore, the denominator for the under the purview of the universal banks. German series is net national product, and the ratio may therefore overestimate bank liabilities The British financial system largely sepa- as a share of GNP. The GNP/NNP data come from Mitchell (1978). rated investment banking, brokerage, and SOURCES: United Kingdom, Sheppard (1971) and Mitchell (1978); Germany, commercial services; thus, comparing the (1976) and Mitchell (1978). German universal banks to the British deposit (commercial) banks underestimates the share of corporate financing institu- of the differences in these patterns likely tions in the British economy. Nonetheless, stem from divergent timing of industrial- at 50 percent to 60 percent, deposit banks ization, part may arise from the varied and private banks accounted for twice the structure and practices of the British and share of total financial institution assets German banks. The German universal in Britain than did the universal banks banks are widely believed to have internal- (of both joint-stock and private forms) ized the secondary market in securities, and in Germany. indeed a significant portion of trading took Given the traditional emphasis on place through their offices. The continued the universal banks’ role in promoting expansion of universal banks, therefore, industrialization and economic growth may represent the expansion of the market in Germany, the universal banks’ share for securities. in both financial assets and GNP seems Another measure of the volume of cap- relatively small. Furthermore, the sharp ital mobilized by the British and German increase in the German universal banks’ commercial banks is represented by total share between 1900 and 1913—especially bank assets less cash. As a share of GNP, compared to the more gradual changes liabilities less cash in the British deposit from 1860 to 1880—raises doubts about banks exceeded that in the German universal the causal link between banks by a significant margin from 1883 expansion and industrial growth. until after the turn of the twentieth century 6 The pattern reversed after the In contrast, British deposit banks’ share (Figure 1). While the gap between the wars, and as of 1963, Britain of assets grew most rapidly between 1860 United Kingdom and Germany seems to led Germany again by a and 1880 and then leveled off. While some have virtually disappeared by the outset substantial margin.

FEDERAL RESERVE BANK OF ST. LOUIS

39 MAY/JUNE 1998

Figure 2 While the commercial banks clearly represented a greater share of the economy in the United Kingdom than in Germany, Cash-to-Asset and Cash-to-Deposit Ratios, the universal banks may have expanded United Kingdom and Germany available capital at a faster rate. Indeed, 1880-1913 the faster growth rate of German bank lia- Cash/Total Assets and Cash/Total Deposits bilities compared to British bank liabilities 0.18 suggests this might have been the case. 0.16 U.K. The ultimate impact of the banks’ activ- 0.14 ities depends inversely upon the propor- 0.12 tion of the system’s assets retained as cash

0.1 reserves. In a simple model of a monetary economy, the total nominal money stock 0.08 is a function of the nominal monetary base 0.06 Germany (currency plus reserves), the ratio of bank 0.04 deposits to currency, and the cash-to-reserve 0.02 ratio.9 Financial intermediaries maintain 0 partial control over both the reserve ratio 1880 83 86 89 92 95 98 1901 04 07 10 13 and the deposit-to-currency ratio. For Germany-cash/deposit ratio for all joint-stock banks example, banks can raise the deposit-to- U.K.-cash/deposti ratio, Sheppard's series for all banks currency ratio by encouraging individuals U.K. -cash/deposit ratio, Capie and Webber series for to deposit their savings or to buy equity joint-stock banks Germany-cash/asset ratio for all joint-stock banks shares in the bank. U.K. - cash/asset ratio, Sheppard's series for all banks Two measures offer some insight into

NOTES: Sheppard's series for all United Kingdom banks includes private banks the banks’ roles in multiple expansion. The after 1890. Capie and Webber's data on U.K. joint-stock banks include all Irish money multiplier is a negative function of and Scottish banks. the cash-to-deposit ratio. Where banks SOURCES: United Kingdom, Sheppard (1971) and Capie and Webber (1987); are financed by equity or private capital in Germany, Deutsche Bundesbank (1976) addition to deposits, a more relevant ratio may be the cash-to-liabilities ratio. In comparing the British and German cases, of World War I, the series again diverged we find both ratios informative.10 7 Regression of the log of liabili- during the war and its aftermath. Germany’s cash-to-liability ratio ranged ties less cash on a time trend Figure 1 also reveals a clear differ- between 5 percent and 6 percent in the yields an estimated annual average growth rate of 8.6 ence in growth rates of liabilities less late 1880s and early 1890s, but it declined percent in the post-1894 peri- cash per capita in the two countries. considerably after 1893. This decline od, as opposed to a rate of When we take into account the shift in the universal banks’ cash-to-liability ratio 5.1 percent before1894. in coverage for the British series, the coincided with the growth of their liabil-

8 trend for that country is relatively flat. ities less cash. Over the same period, the To some extent, the apparently The German figures, in contrast, indi- British banks seem to have maintained late take-off by the universal banks is due to the switch to cate gradual expansion before 1894, considerably higher cash-to-liability ratios, 7 the joint-stock form. Private but rapid growth thereafter. The dis- and the gap appears to have widened banks were more prevalent parity in growth rates may be explained after 1893 (Figure 2). before 1894 than after. by the different patterns of industrial Theoretically, at least, the cash-to- Inclusion of the private banks development in the two countries; liability ratio affects monetary expansion, would flatten the trend some, however, the late development of the interest rates, and economic growth. Yet but it is not clear that the pri- German joint-stock banks is somewhat banks’ holdings of cash are not exogenous, vate banks provided the same surprising. Joint-stock universal banking and differences in funding methods between services in the same way as seems to have taken off more than 40 British and German banks help explain the later joint-stock banks. years after the first universal bank was part of the gap in cash-to-liability ratios. 9 See Champ and Freeman founded and after the industrialization Particularly in the nineteenth century, (1994) and sources cited there. pushes of the mid-nineteenth century.8 British deposit banks financed a much

FEDERAL RESERVE BANK OF ST. LOUIS

40 M AY /JUNE 1998 greater share of their operations with same time, banks do retain significant deposits than did the German universal control over their investment portfolios, banks, and United Kingdom provincial and the riskiness of those portfolios banks (at least to some extent) also issued must also affect the banks’ assessment notes. Universal banks were prohibited of their need for cash reserves. A bank’s from issuing their own notes. structure and activities may measurably Given the divergent liability structures influence the composition of its asset of German and British banks, the cash-to- portfolios, and the different levels of deposit ratio offers greater insight into the bank specialization may therefore partly banks’ participation in maturity transforma- explain the somewhat divergent patterns tion.11 Among the German universal banks, of cash ratios—and thus of capital expan- cash-to-deposit ratios followed a similar, sion—of British and German banks. though more extreme, pattern than cash-to- liabilities ratios, rising over the late 1880s and declining after 1893. Until the last QUALITY OF INVESTMENTS years of the nineteenth century, German Banks’ role in mobilizing capital is inti- cash-to-deposit ratios exceeded the United mately tied to their involvement in the uti- Kingdom ratios, and the gap reached as lization of funds. Through decisions about 10 The data for the two ratios come much as 10 percentage points around 1891. how to lend and invest funds, banks can from two sources: Sheppard’s The variability of cash ratios is also influence the quality of capital formation. (1971 [1873]) compilation of important. The largest German banks kept As with capital mobilization, the structure The Economist’s series and Capie aggregate ratios as low as 7 percent and as of the German universal banks is thought and Webber’s (1985) newer high as 22 percent during this period. The by many to have offered advantages over estimates. The latter only pro- British deposit banks, by comparison, held the British system in promoting the efficient vide cash and deposit figures, so relatively steady cash ratios throughout the use of financial capital. The literatures on the cash-to-liabilities ratio cannot end of the nineteenth and the start of the German and British banking have suggested be calculated from this source. twentieth centuries. As a group, the British that the British banks invested rather conser- 11 That is, changing short-term lia- joint-stock banks maintained cash balances vatively, while the German banks opted for bilities into longer-term assets. between 10 percent and 15 percent riskier strategies. Such risky investment, it is 12 Most proposals recommend- of deposits. argued, channeled funds into high-growth ed required cash holdings of Though neither the British nor the and high-return industries and helped pro- 1 percent to 5 percent of the 13 German banks were bound by minimum mote Germany’s industrialization. sum of deposits and current reserve requirements in the pre-World War I For influencing the quality of invest- account balances to be held at period, cash ratios often still depended on ment, the crucial organizational advantage the Reichsbank. Defenders of factors outside the banks’ control. Even in of the German banks is their supposed long- the German joint-stock banks the absence of regulations, central monet- term direct participation in industrial firms. claimed that the British banks ary authorities may tacitly impose a ratio By holding industrial shares, the banks are held much slimmer reserves on commercial banks. In Britain, the thought to have monitored and even con- than the Germans. The British apparent floor at 10 percent, while cer- trolled the firms they financed. The British banks were accused of padding tainly not proof of the central bank’s role, banks, in contrast, are traditionally accused their reserves for their semian- nual statements of account. See is consistent with the notion that the Bank of having little to do with industry and are Riesser (1910, 1911) and of England held some sway over the banks’ criticized for taking a short-term, arms- sources cited there. Goodhart minimum cash ratio. The ratios for length approach to industrial lending. (1972) also discusses the report- Germany, however, suggest no successful There are several theoretical reasons ing practices of British banks. suasion by the Reichsbank. Bankers, politi- why bank equity holdings may increase 13 cians, and economists often debated the the efficiency of investment. Many of these Tilly (1986) points out that, given that the main clientele of need for a required reserve, but little was hypotheses originate in the idea that asym- the universal banks appears to done toward imposing regulations like metric information between borrowers and have been large, older, publicly 12 those enacted in the United States. lenders poses extra costs and creates inef- traded enterprises, the banks Clearly, many of the factors involved in ficiencies in the selection and funding of may not have been actively determining the cash-to-deposit ratio fall investment projects. Cost reductions involved in risky, innovative outside the purview of the banks. At the may result from imposing discipline on investment in general.

FEDERAL RESERVE BANK OF ST. LOUIS

41 MAY/JUNE 1998

Figure 3 resolving agency problems associated with multiperiod financial contracts. Explicitly Nongovernment Securities as a Share motivated by modern perceptions of the of Total Assets, United Kingdom and German and Japanese banking systems, Germany 1884-1913 this model shows that the efficiency of Securities/Assets inside investing hinges on the use of fixed fraction contracts. In such arrangements, 0.14 the investor receives a fixed percentage of England 0.12 Germany project returns and finances that same proportion of future investments.15 0.1 According to these theoretical argu- 0.08 ments, banks that hold equity stakes in firms improve the firms’ investment effi- 0.06 Germany: nongovernment/nonsyndicate ciency. Thus, the relative extent of equity 0.04 holdings in the portfolios of British and German firms offers one way to assess 0.02 banks’ direct involvement in raising 0 investment quality.16 Figure 3 compares 1884 87 90 93 96 99 1902 05 08 11 nongovernment securities for the two

NOTES: Figures for the United Kingdom include England and Wales only and include private countries. Because of uncertainties about banks after 1890. Figures for Germany include great banks only (large Berlin banks). valuation and reporting, these figures See explanation in text. should be viewed as approximations.17

SOURCES: United Kingdom, Sheppard (1971) ; Germany, Deutsche Bundesbank (1976). The German banks show no consistent tendency toward higher securities holdings than the British banks. Indeed, according to these estimates, the range was nearly identical 14 Myers and Majluf (1984) ana- management, overseeing investment in the two countries (7 percent to 12 percent lyze many of these theoretical planning and outcomes, optimizing for the German banks and 8 percent to 12 issues and provide a formal risk-taking by firms, and aligning percent for the British). The figures, it should model of the potential subopti- mality of investment under banks’ and firms’ incentives for long- be noted, provide as conservative an estimate 14 asymmetric information. term benefits. as possible of the German and British non- John, et. al. (1994) models the effects government securities holdings. The Scottish 15 Repeated interaction naturally of equity ownership on firms’ risk-taking and Irish deposit banks held higher levels of adds the problem of renegotia- and shows that investment efficiency investments than did their English and Welsh tion. Admati and Pfleiderer increases with the proportion of bank counterparts, and the largest of the German (1994), Persons (1994), and Dybvig and Zender (1991) all financing held in the form of equity. universal banks held more of their assets in address this question. Imperfect oversight of investment choices securities than did the provincial banks. and outcomes creates incentives for firms Therefore, the fact that Figure 3 still shows 16 The remainder of this section to use borrowed funds in an excessively the British banks’ securities holdings on par borrows heavily from Fohlin risky manner. When banks maintain veto with the Germans’ provides a strong indication (1997c). power over the use of funds, pure debt that, despite the measurement difficulties, 17 Disaggregation for the German holdings induce them to minimize their the British banks held a position in non- figures begins only in 1912. risks in order to guarantee a fixed return. government securities similar to that of the The figures for the years before Equity holdings, in contrast, encourage German banks. that are estimated on the basis banks to seek higher firm valuation. Such a finding would fall in line with of the lowest holdings of govern- Thus, the greater the banks’ equity expectations if one thought that the two ment securities between 1912 holdings in the firm, the higher the types of banks were roughly similar. The and 1920 as well as on the detailed account of one of the banks’ incentives for efficient tradeoff predominance of underwriting and bro- great banks between 1896 and between risk and expected return. kerage functions among the universal banks, 1899. The proportion for great In related work, Admati and Pfleiderer however, should have led to higher levels banks ranged from 17.6 percent (1994) also demonstrate the potential of securities holdings at German banks, to 28.6 percent of total securi- importance of equity stakeholders in compared to British commercial banks.

FEDERAL RESERVE BANK OF ST. LOUIS

42 MAY/JUNE 1998

Thus far, the numbers for the German Table 1 banks have included securities holdings resulting from their underwriting and Bank Holdings of Non-Government brokerage business.18 A significant por- Securities/GNP tion of the universal banks’ total invest- 1880 1900 1913 ments arose out of their involvement in Germany 0.022 0.027 0.040 underwriting consortia (or syndicates). These participations therefore include Britain 0.044 0.063 0.058 some shares that remained on the banks’ books only temporarily and because of SOURCE: Calculated from Deutsche Bundesbank (1976) and Goldsmith (1969). the banks’ inability to place the shares. It is useful to compare the estimates for the largest German banks to the securities the banks’ holdings of nongovernment secu- ties held between 1912 and holdings of British institutions engaged rities accounted for a very small share of the 1920. Given that this period in investment banking. Cottrell (1985) economy. The German banks’ share did covers World War I, it would be shows that, at least in some cases, British increase after 1880, but their holdings of natural to expect that govern- investment banks held more than half their nongovernment securities still amounted to ment securities might comprise a higher proportion of securities assets in the form of illiquid investments. By only 4 percent of GNP by 1913. Further- than they did in the preceding contrast, the German universal banks more, the biggest part of the increase came years. In the one detailing of reported liquidity coefficients (the ratio of after the major push of industrialization bank securities holdings that I immediately available or quick assets to in Germany. could find for the period before total liabilities) of 85 percent in 1893. The British banks’ holdings of non- 1900 (Bank für Handel und These figures gradually declined by more government securities were also low rela- Industrie, a great bank), govern- than 20 points over the ensuing 15 years.19 tive to GNP; in contrast to the German ment securities amounted to 24 Naturally, these banks cannot be compared case, however, the banks’ securities share percent to 55 percent of total directly with the German universal banks, of GNP rose between 1880 and 1900 and securities (in the period 1896- but the forgoing examples do support the then leveled off. Given the measurement 99). Thus, since I am trying to notion that the great banks in particular, difficulties already discussed, and the like- err on the side of finding high rates of nongovernment securi- because of their active engagement in lihood that securities accounted for a greater ties holdings, 17 percent seemed investment banking, should be expected share of the economy in Britain than in a conservative enough estimate to have held a significantly greater share of Germany, it is best not to overemphasize of the proportion of all great their assets in the form of securities than the differences between the German and bank securities held in the form did the British deposit banks. Comparison British numbers.20 Nonetheless, these cal- of government securities. with the British investment banks also culations cast doubt on the idea that the 18 According to Riesser (1910), underscores the potential inconsistency in banks’ holdings of securities provided a the largest universal banks the idea that universal banks could hold significant stimulus to either the German earned approximately half of substantial long-term (illiquid) engagements or the British economies during the last their gross profits from under- with industrial firms and still operate a com- half of the nineteenth century. writing and brokerage services. mercial business on the order of the British It is often claimed that the British 19 deposit banks. banks held only gilt-edged securities in Riesser (1911), p. 655, discusses the banks’ liquidity To understand how important the banks’ their portfolios, and that the German at length. direct investment in industrial compa- banks participated more actively and nies may have been for the growth of the directly in risky, start-up ventures. The 20 Banks’ equity holdings may economy, it is useful to combine the data on official figures, however, do not allow actually represent a greater bank investments with those on bank assets specific types of securities to be distin- proportion of share capital in relative to the economy as a whole. Table 1 guished. Such distinctions, unfortu- Germany than in the United Kingdom. Such figures are inter- reports the results of this calculation and nately, depend on spottier evidence from esting from the point of view of indicates that the nongovernment securities individual banks. German bank records the banks’ involvement in the holdings of universal banks ranged between for the pre-1880 period are generally secondary market for shares, but 2 percent and 4 percent of GNP for the three unavailable. Nonetheless, some details the banks’ ultimate impact must decades preceding World War I. Even if the are available for two of the earliest German be measured against the estimates are only approximately correct, joint-stock universal banks, the Disconto- economy as a whole.

FEDERAL RESERVE BANK OF ST. LOUIS

43 M AY /JUNE 1998

Gesellschaft (DG) and the Darmstädter Bank it can hardly be argued that even the early (Bank für Handel und Industrie, or BHI).21 activities of the great banks included The DG held no securities in its first extensive, direct involvement in indus- four years, but the proportion of securities trial companies. holdings rose to around 12 percent of Economic historians can pick up assets in 1856 and grew rapidly over the this story in the 1880s by using evidence following few years. The bank seems to from BHI.23 Holdings of industrial shares have unloaded securities during the boom amounted to less than 1 percent of BHI’s years of the early 1870s, but it then took assets for most of the 1880s and ’90s, and on extremely high shares of securities even at its peak, the proportion of indus- during the middle of that decade. While trial shares to assets reached only 1.3 per- the bank’s holdings continued to fluctuate cent (in 1882).24 Moreover, BHI reported throughout the remainder of the nine- substantial holdings of only 12 different teenth century, the proportion of securi- companies between 1882 and 1897 and ties followed a generally downward trend no more than seven firms in any one year. toward the end of the period. Together, these data provide further sup- Between 1856 and 1865, two mining port for the notion that the great banks companies accounted for the vast majority invested a relatively small portion of of DG’s industrial holdings, averaging their portfolios in long-term stakes in around 11 percent of bank assets during industrial firms.25 this period. Direct participation arose out As for the securities holdings in Britain, of the bank’s intention to convert the firms Goodhart provides some details for three into joint-stock companies, but because of British commercial banks (Metropolitan the thin market for the securities, DG was Bank, London and Midland, and Union forced to hold these companies’ shares Bank). Nearly all of the investments until the bank could extricate itself in the reported consisted of British, colonial, or more favorable market of the late 1860s foreign-government securities or railway and early 1870s. stocks and bonds. Given his warnings The remainder of DG’s securities about the banks’ desire to hide any invest- portfolio was held in relatively conserva- ments in industrial firms, it is impossible tive investments: government debt, railway to tell for sure what industrial shares the shares and bonds, and other priority bonds banks may have held. Edelstein, however, and shares. With the exception of a few has provided more general estimates of U.K. 21 See Däbritz (1931) on DG and minor holdings of shares, the DG con- securities holdings, and those results indi- Saling’s Börsen-Jahrbuch on the fined its participation in industry to three cate an expansion of industrial holdings Darmstädter Bank. companies (the two already discussed between 1871 and 1913. Industrial con- 22 Däbritz (1931). plus another mining concern). Indeed, cerns and railways, both foreign and

23 the bank’s holdings of industry stocks domestic, accounted for 37 percent of all BHI published unusually amounted to between zero percent and securities holdings in 1871 and 62 percent detailed accounts of its securi- ties holdings, and until 1899, 3 percent of its assets for the years in by World War I. Home company holdings Saling’s reproduced the infor- which disaggregated data are available alone increased from 4 percent to 17 percent 22 26 mation in its series on Berlin- (1852-65). of all U.K. holdings over the period. listed companies. Tilly (1967) shows in his discussion For the period between 1883 and 1907,

24 of the early industrial promotion activi- Davis and Huttenback (1986) find that the Fohlin (1997c). ties of the Bank für Handel und Industrie, financial community owned around 5 per- 25 While the experiences of two another of the great banks, that while the cent of U.K. share value and averaged 4 banks may not necessarily be bank was energetic in such activities in percent stakes in those companies. In generalized to the population its first four years, it had difficulty placing addition, public companies, some of which as a whole, these two banks shares at reasonable prices. By the early may have been banks, held nearly 4 percent do represent a significant pro- 1860s, BHI had extricated itself from this of domestic share capital. The banks might portion of the great banks. side of the business and had turned to be expected to have participated to some 26 Edelstein (1982), p. 48. railway and government finance. Thus, extent in these investments, though firm

F EDERAL RESERVE BANK OF ST. LOUIS

44 MAY/JUNE 1998 proof of such a contention is apparently participation yields the same kind of incen- unavailable. Yet even if the British banks tive effects as direct ownership. In theory, at held no industrial shares, the evidence for least, systems in which banks exert control DG and BHI suggests that the German uni- over investment decisions but do not align versal banks were not far ahead on this count. their incentives with those of the firms It is important to note that the banks’ through equity stakes force firms into exces- ownership of shares, at the margin, may have sively safe and thus inefficient investment provided important injections of liquidity or programs. So the German system of proxy signals of quality for newly public firms. In a voting and interlocking directorates may thin market for industrial securities, and in have increased bank control and oversight, cases of lumpy investments, such holdings but it may have led to more internal may permit firms to invest when they other- financing and fewer risky investments. wise would not have. Thus, small and tran- sient equity stakes may increase the quantity of investment, even if they do not have the CONCLUSION qualitative, efficiency effects that long-term The financial system may promote real holdings are thought to have. Since such growth of the economy by enhancing the equities may not have made it onto the banks’ quantity, quality, or efficiency of invest- books, though, it is difficult to estimate the ment. Using evidence on bank financial ultimate impact of transient holdings. structure, this study has compared the Share ownership represents only the contributions of the British and German most direct kind of involvement in indus- banking systems in the first two of these trial firms. The banks may have also par- areas. The analysis yields no compelling ticipated indirectly in companies, either evidence that one system consistently or sig- through proxy voting of customers’ shares nificantly outperformed the other in raising or through positions in the firms’ super- the quantity or quality of investment. visory boards.27 Because of their combi- The findings indicate that the German nation of underwriting, brokerage, and universal banks, despite their broader commercial services, the German banks involvement in corporate finance, accounted probably obtained greater control of indus- for a markedly smaller proportion of the trial shares than did the British banks. economy than did the British banks. The Since shares taken as collateral or simply gap of the 1880s, much of which may have held as a service to customers would not been due to the later onset of industrializa- appear in the banks’ balance sheets, and tion in Germany than in Britain, only began since firms did not have to reveal their to diminish after 1894 and never fully dis- shareholders, it is virtually impossible to appeared. The universal banks may have, quantify the extent of proxy voting by however, expanded their available capital the German banks. at a faster rate, since they invested or lent It is possible to quantify board posi- a greater share of their total liabilities than tions, and such data suggest that the bank did the British banks. The disparity in cash- directors held positions in relatively few to-liability ratios, however, stems from the companies. Approximately 23 percent of heavily deposit-based financing of the British German joint-stock companies had a private banks. Until the late 1890s, the German 27 The German supervisory banker or bank manager on their supervi- banks actually maintained more conserva- board comprises shareholders’ sory boards, but only half of these attached tive coverage of short-term liabilities than representatives. Currently, this companies received representation from the did the U.K. banks. Only with the serious body must also represent the great banks.28 onset of the deposit business in the mid- firm’s workers. Proxy votes and supervisory board posi- 1890s did the German cash-to-deposit ratios 28 Fohlin (1997a, b) discuss the tions may have enabled banks to monitor begin their steady decline. prevalence, sectoral distribution, their investments and even control the use of The German banks are frequently and determinants of interlocking bank funds. From a theoretical perspective, credited for their active participation in directorates between banks however, it is unclear whether such indirect industry, and bank equity positions in and firms.

FEDERAL RESERVE BANK OF ST. LOUIS

45 M AY /JUNE 1998

firms are thought to improve the quality systems; and it indicates quite strongly that of investment. The findings show, how- without a significant period of real develop- ever, that the universal banks held only a ment, financial institutions can offer only small share of their portfolios in the form limited benefits for economic growth.29 It of industrial equities. Evidence from two may still be true that the German economy of the largest universal banks suggests that has outperformed its British counterpart the universal banks held stakes in only a over the past century, but this study suggests few firms and often did so for lack of that differences in banking structure are demand for their shares. Based on the probably not the cause. Such findings may theoretical work on bank equity stakes, prove useful for policymakers both in fore- this article also argues that if the German casting the effects of banking deregulation banks wielded greater control over firms in the United States and in formulating (through board positions, for example) development programs in other parts of than did the British banks but took no the world. greater equity stakes in those firms, then the German system of relationship bank- REFERENCES ing may actually have led to relative under- investment in risky projects. Admati, Anat R., and Paul Pfleiderer. “Robust Financial Contracting and the Role of Venture Capitalists,” The Journal of Finance This study has raised the possibility that (June 1994), pp. 371-402. the German banks’ choices of investment and reserve holdings were constrained by the Calomiris, Charles. “The Costs of Rejecting Universal Banking: extent of the secondary market in securities. American Finance in the German Mirror, 1870-1914,” Coordination It is possible that the dominance of universal and Information: Historical Perspectives on the Organization of banking in Germany may have hampered the Enterprise, Naomi Lamoreaux and D. Raff, eds., University of growth of complementary financial institu- Chicago Press, 1995. tions. In the German case, however, regulation Cameron, R. Banking in the Early Stages of Industrialization, that encouraged the growth of universal Oxford University Press, 1967. banking also may have inhibited the devel- opment of securities markets. Thus, the Capie, Forrest, and Michael Collins. Have the Banks Failed British existence of universal banking, per se, may Industry? London: Institute of Economic Affairs, 1992. not curtail the functioning of securities markets. Capie, Forrest, and Alan Webber. A Monetary History of the United Clearly, further work on this subject Kingdom, 1870-1982, London: George Allen & Unwin, 1985. is required. But if specialized and uni- versal systems of finance generally provide Champ, Bruce, and Scott Freeman. Modeling Monetary Economies, similar quantities and qualities of invest- John Wiley & Sons, 1994. ment, then cost-efficiency may prove to be Collins, M. Banks and Industrial Finance in Britain 1800-1939, the crucial determinant of the relative London: MacMillan, 1991. growth effects of the two systems. Univer- sal banking may yield economies of scale Cottrell, P. L. Investment Banking in England, 1856-1881, Garland, 1985. or scope compared to a specialized system, Däbritz, Walther. Gründung und Anfänge der Disconto-Gesellschaft but these economies may also lead to exces- Berlin, Leipzig: Verlag Von Duncker & Humblot, 1931. sive concentration, market power, and inef- ficiency in the banking sector. In addition, Davis, Lance, and Robert A. Huttenback. Mammon and the Pursuit of the internalization of the secondary securi- Empire, Cambridge University Press, 1986. ties market within the banking system may Deutsche Bundesbank. Deutsches Geld- und Bankwesen in Zahlen hamper both the efficient distribution of 1876-1975, Frankfurt: Fritz Knapp Verlag, 1976. financial capital and the market for corpo- rate control. Such factors bear directly on Dybvig, Philip, and Jaime Zender. “Capital Structure and Dividend the costs of finance, and such costs influ- Irrelevance with Asymmetric Information,” Review of Financial Studies, (4(1) 1991), pp. 201-19 ence economic growth. 29 Tilly (1967), pp. 114-15, This study narrows the perceived gulf Edelstein, Michael. Overseas Investment in the Age of High Imperialism: argued similarly 30 years ago. between the British and German banking The United Kingdom, 1885-1914, Columbia University Press, 1982.

FEDERAL RESERVE BANK OF ST. LOUIS

46 M AY /JUNE 1998

Edwards, George T. The Role of Banks in Economic Development, Sheppard, David K. The Growth and Role of UK Financial Institutions, London: Macmillian, 1987. 1880-1962, London: Methuen, 1971. Fohlin, Caroline. “Universal Banking Networks in Pre-war Germany: Tilly, Richard. “Germany (1815-1870),” Banking in the Early Stages New Evidence from Company Financial Data,” Research in Economics of Industrialization, Cameron, R., ed., Oxford University Press, 1967. (September 1997a), pp. 201-25. ______. “German Banking, 1850-1914: Development Assistance ______. “The Universal Banks and the Mobilization of Capital in for the Strong,” Journal of European Economic History (Spring 1986), Imperial Germany,” Social Science Working Paper No. 1008, pp. 113-51. California Institute of Technology, 1997b. Forthcoming in Finance and Verlag für Börsen- und Finanzliteratur A.-G. (various years), Saling’s the Making of Modern Capitalism, P. Cottrell, G. Feldman, and J. Reis, Börsen-Jahrbuch, Leipzig. eds., Scolar Press. ______. “Bank Securities Holdings and Industrial Finance before World War I: Britain and Germany Compared,” Business and Economic History (Winter 1997c), pp. 1-13. Galetovic, Alexander. “Finance and Growth: A Synthesis and Interpretation of the Evidence,” BNL Quarterly Review (March 1996), pp. 59-82. Goldsmith, Raymond. Financial Structure and Development, Yale University Press, 1969. Goodhart, Charles. The Business of Banking, 1891-1914, London: Weidenfeld & Nicolson, 1972. Greenwood, Jeremy, and Bruce D. Smith. “Financial Markets in Development, and the Development of Financial Markets,” Journal of Economic Dynamics and Control (January 1997), pp. 145-81. Harris, Milton, and Artur Raviv. “The Theory of Capital Structure,” The Journal of Finance (March 1991), pp. 297-355. John, Kose, Teresa John, and Anthony Saunders. “Universal Banking and Firm Risk-Taking,” Journal of Banking and Finance (March 1994), pp. 307-23. McKinnon, Ronald. Money and Capital in Economic Development, The Brookings Institution, 1973. Mitchell, Brian R. European Historical Statistics 1750-1970, Columbia University Press, 1978. Myers, Stewart, and Nicholas Majluf. “Corporate Financing and Investment Decisions When Firms Have Information That Investors Do Not Have,” Journal of Financial Economics (June 1984), pp. 187-221. Pagano, Marco. “Financial Markets and Growth: An Overview,” European Economic Review (April 1993), pp. 613-22. Persons, John. “Renegotiation and the Impossibility of Optimal Investment,” The Review of Financial Studies (Summer 1994), pp. 419-49. Riesser, Jakob. Die deutschen Großbanken und ihre Konzentration, Jena: Verlag von Gustav Fischer, 1910. English translation: The German Great Banks and their Concentration, The National Monetary Commission, Washington: Government Printing Office, 1911. Shaw, Edward. Financial Deepening in Economic Development, Oxford University Press, 1973.

FEDERAL RESERVE BANK OF ST. LOUIS

47