International Banking

CHAPTER 7

Universal Banking

Alan D. Morrison*

Abstract: Universal banks provide under a single corporate umbrella banking, insurance, and other financial products. While universal banking has been the norm in Germany since the nineteenth century, it was uncommon until recently in other countries, and was proscribed by regulators in the United States. I examine the theory and evidence that explains historical patterns of universal banking; I discuss in detail academic work concerning the Glass-Steagall Act, which separated commercial from investment banking in the US, and I outline evidence concerning the efficiency of modern universal banks. Finally, I provide a technological explanation for the recent growth in universal banking, and I relate this explanation to the early work of Gerschenkron (1962).

Keywords: universal banking, human capital, conflict of interest, securities markets, information technology, Glass-Steagall Act.

Universal banks are institutions that combine the lending and payment services of commercial banks with a wider range of financial services. In particular, universal banks underwrite securities, and hence can offer their client firms access to a broader range of sources of funds than can specialist commercial or investment banks. While universal banks dominate the financial sector in some economies, they are relatively uncommon in others. Indeed, they were outlawed in the United States for the last two thirds of the twentieth century. Understanding this international variation has been a preoccupation of economists since Schumpeter (1939) and Gerschenkron (1962); more recently, the academic literature on this topic has examined the rationale for the American decision to separate commercial from investment banking, and discussed the extent to which universal banks require special regulation. This essay outlines Gerschenkron’s ideas, and relates them to more recent discussions of the subject. It discusses the potential for conflicts of interest within universal banks, and assesses other policy debates surrounding universal banks. Finally, it concludes with a discussion of the reasons for the recent expansion in the number and importance of universal banks, and it assesses some of the potential consequences of this expansion.

1. Universal Banking, Industrial Development, and Human Capital Formation

Numerous authors have pointed to the large-scale and well-capitalized universal banks that played an important role in the pre-World War 1 financing of German industry. These banks held the equity as well as debt securities of their clients; Schumpeter (1939) argues that this resulted in long-term relationship formation, which facilitated efficient resource direction. In contrast, the finance required for British industrialization was garnered partly from commercial banks, and also from stock market flotations, which were brought to market by small scale merchant houses, which lacked the financial capital of the commercial banks.

In a famous essay, Gerschenkron (1962) explains these differences as consequences of the ways in which the respective economies developed. He argues with reference to “economic backwardness;” this is a term that he never defines precisely, but he uses it to describe economies that were relatively late-adopters of modern methods of production and distribution. He argues that development in these economies was hampered by a number of “institutional obstacles.” First, entrepreneurs had not accumulated capital in the earlier stages of development. Second, the workforce in “economically backward” economies had little experience of new technologies and ways of doing things. In the language of modern labor economics, they lacked human capital and, in particular, they lacked tacit human capital, acquired through on-the-job experience but not easily taught at arm’s length, for example in a classroom.

Gerschenkron’s observation about human capital is central to his argument, although it has been little discussed by modern authors. Lack of labor skill renders development difficult; in some cases, as for example with the serfdom of peasants in pre-1861 Russia, he argues that it renders it impossible. In general, Gerschenkron argues, industrialization in backward economies can proceed only when technology reaches a sufficiently advanced stage to substitute for human capital: it is easier to teach a worker to operate very advanced production machinery than it is to teach him the tacit production skills he needs to obtain any benefit from a less sophisticated machine.

Gerschenkron argues that nineteenth century Germany, and also France and Russia, were “economically backward”; he contrasts them with England, which industrialized earlier, and hence was not. According to Gerschenkron, industrialization in economically backward nations relied upon technologies that rendered hard-to-transmit tacit skills sufficiently unimportant. But these technologies operated at a very large scale. They could only function effectively if an adequate infrastructure developed to support them: factories required railways to ensure adequate levels of throughput, railways required coal, and so on. Industrialization therefore had to proceed on a “broad front,” and this required capital on a scale that, by virtue of their economic backwardness, local entrepreneurs had not accumulated. Hence appropriate institutions were required to pool capital, and to direct it towards the technologies that would underpin economic development. These institutions were the universal banks.

In sum, Gerschenkron argues that universal banks arose naturally in countries that had to play economic catch-up, because a lack of human capital generated a pressure towards bigness which could only be satisfied if dispersed sources of capital were combined. And, because capital came from dispersed sources, it had to be closely watched by the banks that directed it. Gerschenkron claims that the German banks established “the closest possible relations with industrial enterprises,” and that they “accompanied an industrial enterprise from the cradle to the grave, from establishment to liquidation throughout all the vicissitudes of its existence.”

Gerschenkron’s analysis extends to what modern writers call industrial organization, but which he calls the “industrial structure” of the economy. He argues that concentration of power and relations in the universal banks served to reinforce the basic tendencies inherent in backward countries, so that attention was devoted to the heavy industry where large-scale financial capital was most useful. Moreover, Gerschenkron argues that the late nineteenth century cartelization movement in German industry was a natural result of the amalgamation of German banks which “refused to tolerate fratricidal struggles amongst their children.” In contrast, while English banks also consolidated at this time, the process was not mirrored to the same extent in industry.

Gerschenkron does not discuss the US banking sector. America industrialized later than Britain; although it lacked some of the institutional features that Gerschenkron argues were a brake on development elsewhere, one might expect America’s banking system to have developed along German lines. Indeed, while the US financial system owed rather more to the British system of financing than did the German one and, moreover, the extent of universal banking in the US was restricted by regulation, US banking had some features that, at least at the level of casual empiricism, were consistent with Gerschenkron’s stories. At the end of the first decade of the twentieth century, Redlich (1968, pp. 381-2) argues that not more than six banking firms were responsible for managing the organization of the American economy. Lamoreaux (1985) documents the merger wave that the US experienced between 1895 and 1904: it saw 1,800 firms disappear into merged entities, and many of the firms formed at this time continued to dominate their industries for the following century: examples are US Steel, DuPont, International Harvester, Pittsburg Plate Glass, American Can, and American Smelting and Refinery. As discussed by Morrison and Wilhelm (2007, pp. 182-184), these mergers were largely orchestrated by bankers, and in particular by J. P. Morgan & Co. In line with Gerschenkron’s observations, Morgan was concerned throughout his career to avoid what he regarded as destructive competition between competitors, and this concern informed his deal-making at this time.

As far as I am aware, Gerschenkron’s assertion that the rate of human capital formation affected the way in which the banking sector developed in Germany (and, arguably, in the United States) has not been subjected to a formal empirical analysis. His assertion that universal banks formed closer and longer-term relationships than their counterparts elsewhere has however been discussed.

Calomiris (1993, 1995) examines the effect of laws that prevented US banks from consolidating and branching during the second Industrial Revolution (1870-1914). He argues that these laws increased the informational and transactions costs of issuing securities, and hence he argues that there was a lower propensity to issue equities in the US than in Germany where, he argues, universal banks were better able to extract valuable information from their borrowers. Moreover, he presents evidence that the costs of financing German industrialization were lower than in the US, precisely because there were universal banks in Germany. He argues that institutional changes that increased bank concentration in the US lowered the costs of finance there.

Calomiris’ conclusions are challenged in a series of papers by Caroline Fohlin. If long-term relationships with banks eased financing conditions then one should see less contemporary evidence of credit rationing in firms with such a relationship. Fohlin (1998b) tests whether a relationship with one of the nine “Great Banks” of nineteenth century Germany eased access to credit by examining the cash flow sensitivity of investment for firms with and without such a relationship. Her approach follows Fazzari, Hubbard and Petersen (1988) and Hoshi, Kashyap and Scharfstein (1990): firms that have easy access to capital should be less reliant upon retained earnings to finance investment, and hence, after controlling for the quality of their investment opportunities, their investment levels should be independent of the cash that their operations generate. Fohlin faces an endogeneity problem, in that association with a universal bank may be related to the quality of investment opportunities. But, even after controlling for this effect, she finds, in an apparent contradiction of the relationship hypothesis, that a bank relationship actually increases the sensitivity of investment to holdings of liquid assets. In another paper (Fohlin (1998a)), she shows that universal bank affiliation in Italy did nothing to ameliorate a liquidity sensitivity of investment, and finds little support from performance data for the notion that universal banks provided screening services to investors.

In other papers, Fohlin presents evidence that first indicates that German banks held more liquid assets than British banks, and that, while they held a limited number of securities in their portfolios, this was often merely because they could not place new issues in their entirety (Fohlin (2001)), and second that bank affiliation in Germany was about securities issuance and stock market listings, rather than the monitoring of debt contracts and the provision of consultancy services (Fohlin (1997)).

Edwards and Ogilvie (1996) also examine the role of universal banks in German industrialization. In contrast to Gerschenkron’s claims, they find that universal banking accounted for a relatively small proportion of the total assets of financial institutions in Germany before 1914. At this time, joint stock companies never accounted for more than 20 percent of the industrial capital stock; for at least 80 percent, then, the special skills of universal banks were not relevant. In most cases, internally generated funds were the most important source of finance for joint stock companies, and much of the rest came from non-universal financial intermediaries, such as savings and mortgage banks, and credit cooperatives.

The evidence to support a close monitoring interpretation of Gerschenkron’s universal banking story therefore seems rather shaky. Interestingly, however, Ramirez (1995) finds evidence that supports it in the American context. He finds, in contrast to Fohlin’s (1998b) German analysis, that a relationship with J. P. Morgan significantly reduced the cash flow sensitivity of investment for American firms. Whether this reflects active monitoring or skilled screening is harder to establish, but it does suggest that a universal banking relationship could ease access to the credit markets. Indeed, it was concerns that the wrong types of firms might be helped into the capital markets by their investment banks that led the American authorities to separate commercial from investment banking. Their reasoning, and the evidence concerning it, is examined in the next section.

2. Universal Banking and Conflicts of Interest

Commercial banks had a significant presence in the United States securities markets of 1900. Although the Comptroller of the Currency ruled in 1902 that national banks were not permitted to engage in the securities business, the First National Bank of Chicago managed in 1903 to circumvent this ruling, by creating a securities affiliate. Securities affiliates were state banks with their own capital, owned by the shareholders of the national bank in proportion with their shares in the national bank. As state banks were not the concern of the Comptroller, affiliates were able to operate in the securities markets, and consequently the national banks functioned as de facto investment banks.

However, while commercial banks were able to operate via securities affiliates in the securities markets, their activities were viewed with some skepticism by populist regulators and legislators. A series of investigations into the governance of investment houses brought the state into conflict with the securities industry in the first quarter of the twentieth century: the Armstrong Committee of 1905 expressed concerns regarding excessively close relations between large investment banks and insurance companies, and the Pujo committee of 1912 tried but failed to prove the existence of a “money trust” that suppressed competition in finance. In the wake of the 1929 stock market crash, the investment banks were again in the line of fire, this time from the Pecora committee, established in 1932 by Herbert Hoover in an attempt to substantiate his belief that the stock market was being undermined by pools of short sellers.

Ferdinand Pecora was far from neutral: Morgan remarked at the time that “Pecora has the manner and the manners of a prosecuting attorney who is trying to convict a horse thief.” Nevertheless, he found some evidence of governance failures, most notably at National City Bank. His findings fed a public mood that demanded changes to the regulatory framework of the investment banking industry, and which found its voice in New Deal legislation that both established a regulatory framework for the securities industry, by creating the Securities and Exchange Commission, and that also profoundly altered the industrial organization of the industry.

The Banking Act of June 1933, popularly known as the Glass-Steagall Act, abolished securities affiliates by requiring a total separation of investment from commercial banking. The Act had a massive impact, since at the end of the 1920s over half of all new securities issues were sponsored by security affiliates. In the wake of the Act, all issues had to be brought to market by specialist investment houses. J. P. Morgan and Co. remained in deposit banking and hence had to leave the securities industry.

While some academic articles debate the point, the Glass-Steagall Act appears to have been motivated by concerns that commercial banks were using their securities affiliates to place low-quality securities on the market in order to avoid taking losses on their own loan portfolios. For example, the Pecora commission uncovered evidence that when the National City Bank’s securities affiliate, the National City Company, pushed Peruvian debt, it did so despite knowing that it was a poor investment. There is however evidence that commercial banks lost heavily on unsold stock when underwriting issues by their debtors (see Kroszner and Rajan (1994)).

The claim that securities affiliates pushed low-quality issues that benefited their parent firms at the expense of their investors went unchallenged in the academic literature for many years. But it is rather incredible: if securities affiliates were pushing low-quality issues then, if they were dealing with rational investors, the low quality should have been reflected in share prices. Hence, if the securities affiliates were pushing poor securities, either they were dealing with naïve investors who failed to learn from experience, or they were making no profits from their actions. Neither story is particularly convincing. Moreover, the fact that investment banks faced conflicts of interest is not necessarily evidence of institutional failure: Morrison and Wilhelm (2007, chapters 2 and 3) argue that investment banks are economically useful precisely because, by placing their reputations at risk, they are able to manage conflicts of interest.

Conflicts of interest in pre-1933 investment banking were examined carefully in the 1990s, as pressure mounted for a repeal of the Glass-Steagall Act. Kroszner and Rajan (1994) test the “naïve investor” theory by examining the performance of affiliate-underwritten securities. They find that there were fewer defaults among affiliate-underwritten securities, which mitigates against the hypothesis that these securities were of systematically lower quality.

Kroszner and Rajan also point to evidence about the pattern of securities issuance that suggests strongly that investors were perfectly aware of the conflicts that their investment banks faced. Precisely because they faced a potential conflict of interest, it was harder for securities affiliates credibly to signal the quality of their issues to the ratings agencies. Kroszner and Rajan support this assertion by showing that ratings were a less accurate predictor of default for affiliate-underwritten bonds than for those underwritten by specialist investment banks. I argue above that conflicts of interest have fewer adverse consequences within a bank that has significant reputational capital at stake. Hence, one would expect the informational problems to be particularly problematic for small affiliates with a lower reputational stake. A sophisticated investor should therefore be unwilling to buy complex and opaque securities that are underwritten by a small affiliate. Consistent with this argument, Kroszner and Rajan find first that affiliates in general underwrote larger issues where information asymmetry was less likely to be a problem, and second, that smaller affiliates underwrote more senior issues by less-risky firms than did larger affiliates.

Kroszner and Rajan’s results suggest strongly that investors were too smart to be taken in by an affiliate pushing poor quality stock. Affiliates with less to lose could not underwrite informationally sensitive issues, and hence could not make as much from their securities business as competitors with more reputational capital. It is even possible that combining commercial with investment banking improved incentives, as commercial banks strove to build reputations which would allow them to enter the lucrative securities markets.

Ang and Richardson (1994) present evidence that is consistent with Kroszner and Rajan’s. They find that bank affiliate issues had lower default rates, lower ex ante yields, and higher ex post prices than those issued by pure investment houses; moreover, they find that the relative ability of ex ante yields to predict ex post performance was no different for affiliate issues than for investment bank issues. Even issues underwritten by the National City Company and the Chase Securities Corporation, both of which were targets of the Pecora hearings, while of lesser quality than other bank affiliate issues, were no worse than those underwritten by the investment banks. Puri (1994) also presents evidence that pre-1933 bank underwritten issues defaulted less than non-bank underwritten issues.

In contrast to other papers written on this subject, Puri (1996) bases conclusions regarding the quality of affiliate issues on ex ante pricing, rather than on ex post default performance. She finds that pre-1933 investors paid higher prices for securities underwritten by banks than for those underwritten by securities houses. Puri argues that these results are indicative of a certification role for banks, which arose because banks had superior information about the firms to which they lent, and because they faced reputational risk.

In short, recent research suggests strongly that pre-1933 commercial banks in the United States did not use their securities affiliates to float securities that would repay their lowest quality loans. The Glass-Steagall Act rendered it impossible to perform precisely this type of research on contemporary US firms. However, Gompers and Lerner (1999) are able to come close, by examining the underpricing of initial public offerings (IPOs) brought to market between December 1972 and December 1992 by investment banks that held equity in the issuing firm via a venture capital subsidiary. Once again, they find no support for the “naïve investor” hypothesis; investors appear rationally to account for the quality of securities. IPOs underwritten by affiliated investment banks in their sample perform at least as well as those in which underwriters have no position. Investors demand a greater discount for investing in affiliated issues and, consistent with the evidence in Kroszner and Rajan (1994), investment bank-affiliated venture capital firms seem to invest in less information-sensitive issues.

Another opportunity to perform research on modern data was provided by a partial relaxation of the Glass-Steagall Act in 1987, under which some banks were allowed to set up subsidiaries (“Section 20 subsidiaries”) to underwrite corporate securities. The subsidiaries were subject to firewalls that limited information flows, and they limited in size: initially to 5 percent of the gross revenues of the parent bank, and ultimately to 25 percent. Gande, Puri, Saunders and Walter (1997) examine the operations of section 20 subsidiaries. Their findings are in line with all of the research cited above, in that they find no evidence of malfeasance. They control in their work for the use to which the proceeds of the issue are put. When the securities are issued for purposes other than debt repayment, spreads for sub-investment grade issues are 42 basis points lower than for investment houses; when the stated purpose is refinancing, the spreads are statistically indistinguishable. Moreover, and in contrast to some of the earlier papers cited, Gande et. at. find that section 20 subsidiaries tend to underwrite smaller issuers than investment houses. The evidence of this work is therefore that, if anything, the informational advantage of lending banks serves to attract investors, rather than to repel them: Puri (1999) presents a model along these lines, in which the information that commercial banks acquire through lending allows them to obtain better prices for securities. Gande, Puri and Saunders (1999) find moreover that the entry of section 20 subsidiaries lowered fees for security underwriting, particularly among lower-rated and smaller issues, where section 20 subsidiaries were particularly active. Evidence largely consistent with the results of these papers is presented by Roten and Mullineaux (2002), who find that section 20 subsidiaries charged lower fees than investment bank underwriters, who were able to capitalize upon their stronger reputational capital, but that there was no significant overall difference in yield spreads between the two types of underwriters.

While a substantial body of evidence suggests first that the market accounts for conflicts of interest when US banks underwrite securities, and second that such conflicts are seldom a significant concern, there is little comparable evidence in other countries. However, a paper by Ber, Yafeh and Yosha (2001) generates results for the modern Israeli market that are somewhat at variance with those for the United States of the 1930s. The Israeli banking industry is highly concentrated, and it is universal, with banks managing investment funds as well as controlling subsidiaries that specialize in underwriting. While most pre-Glass Steagall data is for bond issues, Ber et. al. focus on straight equity issues. They find that the post-issue accounting performance of firms underwritten by their lender is significantly better than average. However, they find that the same firms exhibit negative stock excess returns in the first day and year after issuance, which suggests that these issues are systematically overpriced. If buyers are not naïve, we must look elsewhere for an explanation for this persistent mis-pricing. The authors suggest that it arises because the buyers are investment funds controlled by issuers. Hence, they argue that, at least in the Israeli market, the combination of bank lending, underwriting, and investment fund management in a single institution is potentially harmful.

Ber et. al.’s finding are worrisome. They suggest that, while managed funds are controlled by entirely rational agents, they are able to find and to exploit naïve retail investors; hence, financial infrastructure needs to be designed so as to ensure that fund managers’ incentives are properly aligned with their investors. Arguably, then, the efficiency consequences of allowing universal banking in one economy could be different to those in another, which has different institutional and legal features. While research into the pre-Glass Steagall US economy helped to justify the repeal of the Banking Act, it should be applied to other economies with caution. The next section examines the policy arguments that surrounded the introduction of universal banking into the US

3. Universal Banks and Economic Efficiency: The Repeal of the Glass-Steagall Act

The November 1999 Gramm-Leach-Bliley Act dismantled the barriers to universal banking that had been erected in the United States by the Glass-Steagall Act. The Gramm-Leach-Bliley Act responded to an increasing commercial need for universal banking that had already been recognized by the Federal Reserve Board’s 1998 approval of the merger of Citicorp and Travelers. The same pressures were apparent in Europe where, in the absence of Glass-Steagall-type legislation, financial conglomeration had been taking place for at least a decade. The most important pressure for the Gramm-Leach-Bliley Act was therefore commercial, but it did not pass without considerable discussion. In this section I briefly outline and comment upon some of the sources of debate.

Universal banking could be introduced to the United States only when several concerns had been assuaged. First, the conflicts of interest that had motivated the passage of the Glass-Steagall Act had to be addressed. Second, regulators and legislators had to be convinced that universal banking would not create new systemic risks that threatened the stability and efficiency of the financial sector.

As discussed in the previous section, a growing body of research in the 1990s suggested that the deleterious consequences of the conflicts of interest facing universal banks may have been more perceived than real. The systemic risks fell into several categories. First, there were concerns that the formation of large universal banks would be too central to the operation of the economy to be allowed to fail, as a result of which a moral hazard problem would exist between the shareholders and managers of these banks. The danger that some banks might be treated as too big to fail was reflected in market prices after the Comptroller of the Currency acknowledged in testimony to Congress that eleven of the largest US national banks could expect to receive the sort of $1 billion bailout extended in 1984 to the insolvent Continental Illinois Bank: Avery, Belton, and Goldberg (1988) show that subsequently, bank bond spreads were barely related to ratings, and Boyd and Gertler (1993) find that large banks took on bigger risks than smaller commercial banks.

When large banks are systemically important, could better regulation assuage the too big to fail problem? The Federal Deposit Insurance Corporation Insurance Act (FDICIA) of 1991 was landmark legislation partly intended to accomplish this goal. It requires regulators to take prompt corrective action against distressed banks, and places checks and balances upon the decision to declare a bank too big to fail. Stern and Feldman (2004) argue that FDICIA did little to resolve the too big to fail problem, claiming that regulators still have the incentive and the ability to bail out insolvent banks. Some evidence does indicate that FDICIA did not entirely resolve the too big to fail problem: Morgan and Stiroh (2005) find that the spread-rating relationship for banks identified in the mid-1980s as too big to fail was little changed by FDICIA, although they find more sensitivity than did Avery et. al., and Brewer and Jagtiani (2007) show that banks are prepared to pay a premium for acquisitions that will push them over perceived too big to fail boundaries. Nevertheless, Mishkin (2006) argues in an essay reviewing Stern and Feldman’s book that the weight of evidence does not support their assertion: Ennis and Malek (2005) find no evidence in the wake of FDICIA of the excessive risk taking documented in large banks by Boyd and Gertler (1993), and Flannery and Soresco (1996) find stronger market discipline in the subordinated debt market for banks in the post-FDICIA period. Hence, even if large universal banks are systematically so important that the regulator cannot credibly commit to deny them access to the government safety net, a case can be made that the concomitant incentive problems can be counteracted by well-designed regulatory institutions. Mishkin (1999) makes this case, arguing that universal banking should be accompanied by greater regulatory vigilance, coupled with some constructive ambiguity regarding bailout policy.

A further systemic cost of universal banking may arise if the securities arm of a universal bank is able to access the deposit insurance safety net provided to the commercial banking arm: this would be likely to result in risk-shifting, as securities firms take excessive risks, for which, by virtue of the deposit insurance scheme, their depositors do not charge them. Furthermore, as Boyd, Chang and Smith (1998) note, banks that hold equity stakes in their borrowers have strong incentives to take advantage of the deposit insurance safety net. Benston (1994) discusses this point. He argues that there is no evidence that universal banking is more risky than specialized banking. Cornett, Ors and Tehranian (2002) support Benston’s assertion, finding that bank riskiness around the introduction of a section 20 subsidiary does not change. In any case, several authors suggested in the 1990s that the diversification that universal banking would bring would more than outweigh any risk-shifting dangers. Mälkönen (2004) and Allen and Jagtiani (2000) both perform simulations using portfolios of commercial bank loans and insurance company investments, and show that combining the two generates inter-divisional diversification. This work is however subject to a Lucas-style critique: Freixas, Lóránth and Morrison (2007) show that the non-bank divisions of financial conglomerates could take more risk in order to profit from the deposit insurance put option than they would have done as standalone firms. Whether or not the diversification effect outweighs the enhanced risk-shifting incentive is context specific. With the appropriate capital adequacy policy, Freixas et. al. demonstrate that optimal regulation forces the deposit taking and non-deposit taking arms of the bank to maintain separate balance sheets: although this reduces diversification opportunities, it enhances market discipline sufficiently to compensate.

Rajan (1996) expresses concern that universal banks may use power derived from their informational monopoly to suppress competing institutions and markets. While good regulation can probably counteract this danger, he suggests that concentration of economic power in a few universal banks could act as a brake on economic progress in developing countries. As we saw in section 1, a similar point was made by Gerschenkron, who identified a tendency within late nineteenth and early twentieth century universal banks to suppress competition in the real sector of the economy. Unlike Rajan, of course, Gerschenkron argued that universal banks aided development, and hence that the danger of anti-competitive behavior was worth accepting. In any case, Benston (1994) argues that modern universal banks serve such a broad constituency that they are unlikely to favor one interest group over another, and hence that they are less likely to be a source of damaging rent-seeking than more specialized institutions.

Boot and Thakor (1997) identify another way in which universal banks may reduce the beneficial effects of competition. They argue that borrowers choose between bank and market finance by weighing up the relative benefits of bank monitoring, which attenuates moral hazard, and more informative price signals, which facilitate efficient resource allocation. Financial innovations that increase price informativeness result in a shift from bank to market finance. These innovations raise welfare, but their effect within a universal bank is to transfer revenues from one part of the business to another. Hence, Boot and Thakor argue that the incentive to innovate in a universal bank is lower than in an investment bank, which can hope to attract new customers by innovating.

A further concern is that, by combining depository institutions with other types of financial firms, universal banks may open new channels for financial contagion, so that instability outside the banking sector, for example in the insurance market, could be transmitted to banks via universal firms that encompass both sectors. The evidence on this point is mixed. In work that to some extent anticipated Kroszner and Rajan, White (1986) finds no evidence of greater instability in universal institutions at the start of the 1930s: while 26.3% of national banks failed between 1930 and 1933, only 6.5% of the 63 banks that had security affiliates in 1929 and 7.6% of the 145 banks with large-scale bond operations failed. Logit regressions on White’s data confirm that the presence of a security affiliate reduced the probability of a bank failure. Colvin (2007) argues that the Netherlands experienced in the 1920s its only traditional banking crisis since 1600; he presents evidence that the relatively large difficulties that the Rotterdamsche Bankvereeniging experienced relative to its rival Amsterdamsche Bank were attributable to its universal status. Franke and Hudson (1984) find no evidence that universal banks were behind any of the major twentieth century financial crises to affect West Germany in the twentieth century. Canals (1997) cites Cuervo (1988) on the effect of the European recessions of the late 1970s and early 1990s upon Spanish banks. In both cases, the banks that experienced the biggest losses were universal banks with major stakes in the industrial sector.

The aforementioned evidence suggests that the dangers associated with universal banking are less than was believed in the United States for much of the twentieth century. Indeed, there is a body of evidence that indicates that universal banks are positively efficiency-enhancing. For example, Barth, Brumbaugh and Wilcox (2000) point to technological advances that open new economies of scope in large banks. Berger, De Young, Genay and Udell (2000) discuss economies of scale: universal institutions can share offices, computers, information systems, investment departments, account service centers, or other operations; they can economize on the fixed costs of raising capital, and they can re-use information about a client in several business lines. On the other hand, like any other organization, universal banks may experience diseconomies of scale (see Winton (1999) for a model incorporating this effect): the extent to which universal banks can realize economies of scale and scope is of course an empirical question. The second Banking Co-Ordination Directive of 1989 made universal banking the norm in the European Union by introducing a single banking license valid throughout the EU, and limiting product mix restrictions to those imposed by home regulators. Hence one would hope for evidence for or against scope efficiencies in the European market. However, the few studies that exist are rather inconclusive: Allen and Rai (1996) and Vander Vennet (1999) find only limited evidence of scope economies in European universal banks; Cyberto-Ottone and Murgia (2000) find evidence that scope-expanding mergers in European banking markets increase shareholder wealth.

Gorton and Schmid (2000) use data from 1975 to 1985 to examine the consequences of universal banking for the real economy in Germany. They account for control rights, voting restrictions, and the effects of co-determination. They find that banks affect firm performance beyond the effect they would have as non-banks, and that the concentration of control rights in banks improves firm performance. A number of authors have suggested that introducing universal banking into other countries would bring benefits that mirrored the German experience. Indeed, in an analysis of 60 countries, Barth, Caprio and Levine (1999) find that restricting securities activities reduces bank efficiency and raises the likelihood of a banking crisis. Their data contains no evidence that restricting financial firms assists financial development, or that it increases industrial competition. Nevertheless, Rajan (1996) argues that one should be careful of drawing strong conclusions from this type of work: universal banking exists within a broader institutional framework, and it need not follow that the benefits associated with the entire framework can be achieved simply by embracing only universal banking. For example, Rime and Stiroh (2003) find no evidence to suggest that any efficiency benefits are being derived from the trend towards universal banking in Switzerland.

Notwithstanding the institutional caveats expressed by Rajan, the arguments of this section provide only weak support for regulation that prevents universal banking. Financial markets appear rationally to discount conflicts of interest within universal banks. Hence, as Kanatas and Qi (1998) argue, borrowers will choose to deal with universal banks only if the costs of conflict are outweighed by the scope economies that the universal banks can realize. Only if universal banking generates an unpriced social cost is there a case for restricting it. Kanatas and Qi suggest that this cost might arise because conflicts of interest give rise to a soft budget constraint: they argue that, because borrowers from a universal bank anticipate that they will be bailed out via a stock issue in the event of poor performance, they choose lower quality investments. However, the empirical evidence of section 2 suggests that, in fact, ex post conflicts are relatively small. In advanced economies, one can arguably deal with other potential problems, such as anti-competitive behavior and abuse of the deposit insurance safety net, through careful regulation.

In light of the previous paragraph, the case for proscribing universal banking seems rather weak, and the decision to repeal the Glass-Steagall Act seems justifiable. But, although the academic case for repeal was strong, the Gramm-Leach-Bliley Act was also a response to intense commercial pressures. Several authors have suggested that these pressures reflected the enhanced benefits of scale and scope made possible by advances in information technology. The following section discusses this point, and relates this argument to the early ideas of Gerschenkron, discussed in section 1.

4. Scale and Scope in Twenty First Century Banking

The investment banking industry became increasingly reliant upon financial capital in the second half of the twentieth century. Morrison and Wilhelm (2008) report data for the US: on a CPI-adjusted basis (1983 dollars), the combined capitalization of the top ten investment banking firms rose at an increasing rate from $821 million in 1955 to $2,314 mn in 1970 , $6,349 mn in 1980, $31,262 mn in 1990 and $194,171 mn in 2000. Over the same period the industry became increasingly concentrated, with the capitalization of the 11 to 25th largest investment banks as a proportion of that of the top ten dropping from 80% to 10%. Moreover, it appears that the importance of financial capital significantly increased relative to human capital over this period: while the average number of employees in the largest five banks quadrupled between 1979 and 2000, the mean capitalization per employee in these banks increased by a factor of more than 15. I will argue in this section that the imperative for universal banking at the end of the twentieth century was created by the same economic forces that increased both concentration and capitalization in the investment banking industry.

Starting from its origins in the nineteenth century Atlantic trade, investment banks provided services over which it was very hard to contract: while clients may be able to distinguish a well-priced IPO from a poorly-priced one, good advice from bad, or a well-executed security transaction from a botched deal, making this distinction stick in court is very hard. It is precisely for this reason that investment banks depended upon their reputations: because clients would pay a significant premium to a trustworthy bank, investment bankers would work hard to retain their reputations, so that a strong reputation could underpin agreements that were not enforceable under black-letter law. The need for a reputation created a substantial barrier to entry into the business, and, arguably, explained the very long-lived super normal profits that the early investment bankers made.

When investment bankers relied upon reputation to underpin tacit agreements with their counterparties, their business was inevitably based upon close relationships. Many of the skills that investment bankers needed were tacit: that is, they were best learned on-the-job, through a close mentoring relationship with a senior banker. Morrison and Wilhelm (2004) argue that partnership firms provide the strongest possible incentives to maintain these relationships, and hence the early investment banks were constituted as partnerships. While partnership status assisted in human capital formation, it limited the size and capitalization of investment banks (Morrison and Wilhelm (2008)).

Starting in the early 1960s, a number of factors undermined the traditional structure of the investment banking firm. First, the advent of transistor-based mainframe computers in the early 1960s rendered cost-effective the overnight batch processing of the large-scale repetitive tasks associated with settlement. This type of processing was particularly valuable to “retail” firms like Merrill Lynch, which performed high volumes of small-value transactions. Mainframe computing was extremely costly, but retail firms that failed to adopt it found it impossible to cope with a massive increase in trading volumes at the end of the decade: they ultimately failed, or were absorbed by larger institutions (see Morrison and Wilhelm (2007, pp. 235-238)). The retail firms acquired the capital needed to acquire mainframe computers by floating in the early 1970s (see Morrison and Wilhelm (2008)).

Further advances in information technology were more applicable to investment banks that specialized in wholesale business and, ultimately, to universal banks. The cost of computing started to plummet in the late 1970s, as microcomputers found their way into banks, and allowed traders and relationship managers to interrogate databases and to perform complex pricing calculations in real time. For example, the ability rapidly to create spreadsheet-based financial models revolutionized the operation of the LBO market, and made it far easier to price new offerings. At the same time, advances in financial economic theory were transforming the financial market place. The Black-Scholes-Merton framework for financial options valuation became a practical tool rather than an academic exercise when it could be implemented with a desktop computer; risk management practices could be hard-coded into computers, rather than based upon judgment and recruitment practices; trading and hedging strategies could be driven by computer algorithms rather than by humans.

Unlike mainframe computers, microcomputers were cheap, and they substituted for a great deal of human expertise. One might expect them to lower the minimum scale at which investment banks could operate. Indeed, Rajan (1996) makes this point, arguing that there is no a priori reason to assume that better information technology should increase the optimal scale of a bank. But better information technology not only automated tasks that previously were the province of the human expert; it also changed the nature of investment banking skill. Activities that could be expressed in the formal language of financial economics could be taught in a classroom. Trading results that could be captured with computers and analyzed using portfolio theory could be contracted upon. As a result, businesses that previously were the preserve of a few specialists operating in businesses with reserves of human capital and reputation started to be open to any firm that could hire a smart financial engineer. Precisely because information technology and the codification of skills combined to render market entry easy for any firm, large or small, financial markets became extremely competitive. In the end, precisely because the financial markets had become so contestable by small firms, they could no longer sustain small-scale trade: bid-ask spreads narrowed to such an extent that participation in the markets became cost-effective only for firms that could operate at a large scale. The consequence was the massive increase in investment bank capitalization and concentration that I highlighted in the opening paragraph of this section.

In short, distributed micro computers and advances in financial economics lowered the value of tacit skill relative to technical, codifiable skill in many investment banking activities. It also facilitated entry, and hence lowered the minimum scale at which these activities were economically viable. Morrison and Wilhelm (2008) argue that these effects combined to cause the demise of the traditional investment banking partnership. They also opened the door to commercial bank entry into investment banking. Commercial banks had greater reserves of capital than the investment banks. Where they were legally allowed to underwrite, they could bundle their services with lending business in a way that investment banks could not. Particularly when underwriting bond issues, whose prices are most susceptible to codification, commercial banks therefore had advantages that were denied to investment banks. Similarly, commercial banks were playing to their strengths when they invested in derivatives trading partnerships in the late 1980s: derivatives trading was a technical, computer-oriented activity that required capital on a huge scale.

Gerschenkron argued in the 1960s that “economically backward” economies relied for development upon codified knowledge that was embedded in large-scale and capital-intensive production technologies; it was for this reason that he believed that universal banking was common in economies that historically had developed from a backward state. The arguments of this section suggest that something similar is afoot in the modern banking sector, where production techniques have been revolutionized by new computer-based technologies that formalize many formerly tacit skills. As in Gerschenkron’s work, the new technologies require very high levels of capital investment, which arise in this case because they generate competitive pressures that significantly raise the minimum operating scale in banking. The commercial pressures for universal banking seem unsurprising in the light of this argument; the steady erosion of the Glass-Steagall Act, starting in 1986 with the Fed’s approval of an application by Banker’s Trust to underwrite commercial paper and culminating in the passage of the Gramm-Leach-Bliley Act, was perhaps inevitable.

The immense scale and scope of the modern universal bank does not come without challenges, however. It may be very hard for an institution to run large-scale codified businesses side-by-side with those that rely upon more traditional tacit skills. When universal banks build systems and procedures around “hard” codifiable information that can be fed into a computer, their decision-making becomes increasingly remote from the loan officers who forge relationships with their customers. As a result it becomes hard for them to accommodate lending based upon “soft” relationship-based information that cannot easily be computerized. Stein (2002) argues that, as a result, loan officers in banks that rely upon formal systems to make decisions may be less inclined to gather information at all. Berger, Miller, Petersen, Rajan and Stein (2005) find evidence consistent with Stein’s hypothesis, stating that “large banks are less willing to lend to informationally ‘difficult’ credits, such as firms with no financial records.” Of course, whether or not information is hard is to some extent a decision variable: Petersen (2004) argues that ratings agencies emerged in the nineteenth century as ways of hardening previously soft information about borrowers. But there are presumably limits to this process, and it may prove difficult in general to reconcile small-scale relationship lending with the needs of the universal bank.

5. The Crisis of 2008

The causes of the financial crisis of 2008 are still a matter for discussion, and its consequences are still unfolding. It is too early to draw categorical conclusions from the financial crisis about banking in general, and about universal banking in particular. Nevertheless, it is possible to make some tentative observations.

First, and notwithstanding the remarks of section 4, it has become clear that, even in the most complex derivatives markets, tacit knowledge and non-codifiable information still have an important role in finance. The crisis appears to have started in the market for securitised sub-prime mortgage debt, and its ramifications have been keenly felt in the so-called “shadow banking sector,” which relies upon complex debt securitisations. The mathematical models upon which these markets rested have been shown to be rather less precise than was previously believed. The consequence has been a loss in confidence in ratings agencies, and in the most complex securities that they rated: this loss of confidence is having profound effects upon real economic activitity.

The loss of confidence highlighted in the previous paragraph has highlighted the importance in complex markets of reputation. Many users of credit ratings relied upon them to assess the quality of assets that they did not fully understand. Similarly, purchasers of complex securities underwritten by blue chip investment banks relied to some extent upon the reputation of their underwriter for quality certification. I have argued in this essay that reputation is most effectively fostered in small, focused institutions, where conflict-of-interest problems are least likely to impair incentives to maintain reputation. If reputation and the non-codifiable skills upon which it rests are more important that we thought at the start of 2008, then the challenges that universal banks face in providing hard-to-quantify services are greater than appeared in the last decade.

Universal banks have come in for more direct criticism, too. Some commentators have suggested that they are able to extend the reach of the deposit insurance fund; others have argued that universal banks create systemic problems because their businesses are so complex that no-one, least of all their regulators, can understand them. The latter point is closely related to the reputational one, of course: reputation should substitute for transparency but, as argued in the previous paragraph, it appears not to have done in recent years.

If universal bank scale does indeed present a problem to service provision, what is the appropriate response? Some commentators have suggested that the right approach would be to re-introduce something like the Glass Steagall Act. This, they argue, would improve investment banker incentives by taking away any access to the deposit insurance safety net, and would ensure that they could fail. At the same time, it would reduce bank complexity, and so enhance systemic stability.

It is too early to say whether bank scope restrictions will be enacted. But such restrictions would come at a cost. The economies of scale and scope that have been discussed in this article remain important. Universal banks achieve diversification that ought to increase financial stability. And, finally, if reputation is an important basis for profitable business, one would expect participants in a free market to evolve institutions and procedures for maintaining it. This process may already have started: Citigroup announced at the start of 2009 that it planned to divest itself of its consumer finance operations, private-label credit card businesses and parts of its investment banking business, thus partially reversing the move into universal banking that began with its 1998 acquisition of Travelers. Nevertheless, a properly focused universal bank can generate significant economies of scale and scope. The universal banking model seems highly unlikely to vanish.

6. Conclusion

Historically, universal banking was common in some economies, but not in others. Gerschenkron (1962) argued that this variation could be explained with reference to the way in which development occurred: economies that had to play “catch-up” did so by adopting technologies on a broad front that could compensate for the lack of a deep human capital pool, and the institution that collected and directed capital into these technologies was the universal bank.

Notwithstanding the success of universal banking, it was regarded with suspicion in the United States for much of the twentieth century, where it was outlawed by the 1933 Banking Act. Contemporary evidence suggests that this suspicion was largely misplaced. Moreover, commercial pressures in the final decades of the twentieth century were for large, complex financial intermediaries that offered services that encompassed security market business as well as traditional commercial banking. I have argued that these pressures contained an echo of the forces studied by Gerschenkron: simultaneous advances in information technology and financial economies codified traditional knowledge and created massive pressure for scale. These pressures resulted ultimately in the 1999 repeal of the Banking Act.

Universal banking creates challenges for some traditional commercial banking activities. As banks adopt hierarchical structures and rely increasingly upon hard quantifiable data, it will be harder for them to act upon the tacit knowledge and skills of their relationship managers. If large universal banks struggle to supply credit to small opaque businesses then a long-term role for small, specialist commercial lenders will remain.

References

   Allen, L. and A. Rai (1996). Operational Efficiency in Banking: An International Comparison. Journal of Banking and Finance,  20, 655–672.

   Allen, L. and J. Jagtiani (2000). The Risk Effects of Combining Banking, Securities, and Insurance Activities. Journal of Economics and Business,  52,  485 – 497.

    Ang, J. S. and T. Richardson (1994). The Underwriting Experience of Commercial Bank Affiliates Prior to the Glass-Steagall Act: A Re-Examination of Evidence of Passage of the Act. Journal of Banking and Finance,  18,  351–395.

Avery, R., T. Belton and M. Goldberg (1988). Market Discipline in Regulating Bank Risk: New Evidence from the Capital Markets. Journal of Money, Credit, and Banking, 3, 597–619.

   Barth, J. R., G. Caprio, Jr and R. Levine (1999). Banking Systems Around the Globe: Do Regulation and Ownership Affect Performance and Stability? Policy Research Working Paper 2325, World Bank, Washington, DC.

    Barth, J. R. , R. D. Brumbaugh, Jr. and J. A. Wilcox (2000). The Repeal of Glass-Steagall and the Advent of Broad Banking. Journal of Economic Perspectives,  14,  191–204.

   Becker, G. S. (1964). Human Capital: A Theoretical and Empirical Analysis, with Special Reference to Education. Chicago: University of Chicago Press.

    Benston, G. J. (1994). Universal Banking. Journal of Economic Perspectives,  8, 121 – 143.

   Ber, H., Y. Yafeh and O. Yosha (2001). Conflict of Interest in Universal Banking: Bank Lending, Stock Underwriting, and Fund Management. Journal of Monetary Economics,  47, 189–218.

   Berger, A. N., N. H. Miller, M. A. Petersen, R. G. Rajan and J. C. Stein (2005). Does Function Follow Organizational Form? Evidence From the Lending Practices of Large and Small Banks. Journal of Financial Economics,  76, 237 – 269.

   Berger, A. N. , R. De Young and G. F. Udell (2001). Efficiency Barriers to the Consolidation of the European Financial Services Industry. European Financial Management,  7, 117–130.

   Berger, A. N., R. De Young, H. Genay and G. F. Udell (2000). Globalization of Financial Institutions: Evidence from Cross-Border Banking Performance. Brookings-Wharton Papers on Financial Services,  3, 23 – 125.

   Black, F. and M. Scholes (1973). The Pricing of Options and Corporate Liabilities. Journal of Political Economy,  81,  637–654.

   Boot, A. W. A. and A. V. Thakor (1997). Banking Scope and Financial Innovation. Review of Financial Studies,  10, 1099 – 1131.

   Boyd, J. H., C. Chang and B. D. Smith (1998). Moral Hazard Under Commercial and Universal Banking. Journal of Money, Credit and Banking,  30,  426–468.

Boyd, J. H. and M. Gertler (1993). U.S Commercial Banking: Trends, Cycles, and Policy. NBER Macroeconomics Annual 1993 ed. Blanchard, O.J. and S. Fisher. Cambridge, Mass: MIT Press, 319-368.

Brewer, III, E. and J. Jagtiani (2007). How Much Would Banks be Willing to Pay to Become “Too-Big-to-Fail” and to Capture Other Benefits? Research Working Paper 07-05, Federal Reserve Bank of Kansas City.

Brunnermeier, M. (2009) Deciphering the Liquidity and Credit Crunch of 2007-08. Journal of Economic Perspectives, 23, 77-100.

   Calomiris, C. W. (1993). Corporate Finance Benefits from Universal Banking: Germany and the United States, 1870-1914. Working Paper 4408, NBER, Cambridge, Mass.

   Calomiris, C. W. (1995). Universal Banking and the Financing of Industrial Development. Policy Research Working Paper 1533, World Bank, Washington, DC.

   Canals, J. (1997). Universal Banking: International Comparisons and Theoretical Perspectives. New York, NY: Oxford University Press.

   Carosso, V. P. (1970). Investment Banking in America: A History. Cambridge, Mass: Harvard University Press.

   Chandler, A. D. (1990). Scale and Scope; The Dynamics of Industrial Capitalism. Cambridge, Mass: Harvard University Press.

   Colvin, C. L. (2007). Universal Banking Failure? An Analysis of the Contrasting Responses of the Amsterdamsche Bank and the Rotterdamsche Bank to the Dutch Financial Crisis of the 1920s. Economic History Working Paper 98, London School of Economics, London, UK.

    Cornett, M. M., E. Ors and H. Tehranian (2002). Bank Performance around the Introduction of a Section 20 Subsidiary. Journal of Finance,  57,  501–521.

   Cuervo, Á. (1988). La Crisis Bancaria en Espana, 1977-1985, Ariel, Barcelona.

    Cyberto-Ottone, A. and M. Murgia (2000). Mergers and Shareholder Wealth in European Banking. Journal of Banking and Finance,  24, 831–859.

   De Long, J. B. (1991). Did J. P. Morgan’s Men Add Value? An Economist’s Perspective on Financial Capitalism. in P. Temin, ed., Inside the Business Enterprise: Historical Perspectives on the Use of Information, University of Chicago Press, Chicago, IL, 205 – 236.

   Dierick, F. (2004). The Supervision of Mixed Financial Services Groups in Europe. Occasional Paper 20, European Central Bank, Frankfurt.

   Edwards, J. and S. Ogilvie (1996). Universal Banks and German Industrialization: A Reappraisal. Economic History Review,  49, 427–446.

    Fazzari, S. M., R. G. Hubbard and B. C. Petersen (1988). Financing Constraints and Corporate Investment. Brookings Papers on Economic Activity,  1988, 141–195.

Flannery, M. J. and S. M. Sorescu (1996). Evidence of Bank Market Discipline in Subordinated Debenture Yields, Journal of Finance, 51, 1347–1377.

   Fohlin, C. (1997). Universal Banking Networks in Pre-War Germany: New Evidence From Company Financial Data. Research in Economics,  51, 201–225.

   Fohlin, C. (1998a). Fiduciari and Firm Liquidity Constraints: The Italian Experience with German-Style Universal Banking. Explorations in Economic History, 35,  83–107.

   Fohlin, C. (1998b). Relationship Banking, Liquidity, and Investment in the German Industrialization. Journal of Finance,  53, 1737–1758.

   Fohlin, C. (2001). The Balancing Act of German Universal Banks and English Deposit Banks, 1880-1913. Business History, 43, 1–24.

   Franke, H.-H. and M. Hudson (1984). Banking and Finance in West Germany, St. Martin’s Press, New York, NY.

    Freixas, X., G. Lóránth and A. D. Morrison (2007). Regulating Financial Conglomerates. Journal of Financial Intermediation, 16, 479–514.

   Gande, A., M. Puri and A. Saunders (1999). Bank Entry, Competition, and the Market for Corporate Securities Underwriting - A Survey of the Evidence. Journal of Financial Economics,  54, 165–195.

   Gande, A., M. Puri, A. Saunders and I. Walter (1997). Bank Underwriting of Debt Securities: Modern Evidence. Review of Financial Studies,  10, 1175–1202.

   Gerschenkron, A. (1962). Economic Backwardness in Historical Perspective, Cambridge, Mass: Harvard University Press.

    Gompers, P. and J. Lerner (1999). Conflict of Interest in the Issuance of Public Securities: Evidence from Venture Capital. Journal of Law and Economics, 42, 1–28.

    Gorton, G. and F. A. Schmid (2000). Universal Banking and the Performance of German Firms. Journal of Financial Economics, 58, 29–80.

   Hoshi, T., A. Kashyap and D. Scharfstein (1990). Corporate Structure, Liquidity and Investment: Evidence from Japanese Industrial Groups. Quarterly Journal of Economics, 106, 33–60.

   Kanatas, G. and J. Qi (1998). Underwriting by Commercial Banks: Incentive Conflicts, Scope Economies, and Project Quality. Journal of Money, Credit and Banking,  30, 119 – 133.

   Kroszner, R. S. and R. G. Rajan (1994). Is the Glass-Steagall Act Justified? A Study of the U.S. Experience with Universal Banking Before 1933. American Economic Review, 84, 810–832.

   Lamoreaux, N. R. (1985). The Great Merger Movement in American Business, 1895 – 1904. Cambridge, UK: Cambridge University Press.

   Langevoort, D. C. (1987). Statutory Obsolescence and the Judicial Process: The Revisionist Role of the Courts in Federal Banking Regulation. Michigan Law Review,  85,  672 – 733.

   Leuchtenburg, W. E. (1963). Franklin D. Roosevelt and the New Deal, 1932-1940. New York, NY: Harper and Row.

   Lown, C. S., C. L. Osler, P. E. Strahan and A. Sufi (2000). The Changing Landscape of the Financial Services Industry: What Lies Ahead?. Federal Reserve Bank of New York Economic Policy Review,  6, 39 – 55.

   Macey, J. R. (1984). Special Interest Groups Legislation and the Judicial Function: The Dilemma of Glass-Steagall. Emory Law Journal,  33, 1 – 40.

   Mälkönen, V. (2004). Capital Adequacy Regulations and Financial Conglomerates. Discussion Paper 10.2004, Bank of Finland, Helsinki.

   Merton, R. C. (1973). Theory of Rational Option Pricing. Bell Journal of Economics and Management Science,  4, 141–183.

   Milbourn, T. T., A. W. A. Boot and A. V. Thakor (1999). Megamergers and Expanded Scope: Theories of Bank Size and Activity Diversity. Journal of Banking and Finance,  23, 195 – 214.

   Mishkin, F. S. (1999). Financial Consolidation: Dangers and Opportunities. Journal of Banking and Finance,  23, 675–691.

Mishkin, F. S. (2006). How Big a Problem is Too Big to Fail? A Review of Gary Stern and Ron Feldman’s Too Big to Fail: The Hazards of Bank Bailouts, Journal of Economic Literature, 91, 988–1004.

Morgan, D. and K. J. Stiroh (2005). Too Big to Fail After All These Years. Staff Report No. 220, Federal Reserve Bank of New York.

   Morrison, A. D. and W. J. Wilhelm, Jr. (2004). Partnership Firms, Reputation and Human Capital. American Economic Review,  94, 1682 – 1692.

   Morrison, A. D. and W. J. Wilhelm, Jr. (2007). Investment Banking: Institutions, Politics and Law, Oxford University Press, Oxford, UK.

   Morrison, A. D. and W. J. Wilhelm, Jr. (2008). The Demise of Investment Banking Partnerships: Theory and Evidence. Journal of Finance (Forthcoming).

   Petersen, M. A. (2004). Information: Hard and Soft. Working paper, Kellogg School of Management, Northwestern University, IL.

   Polanyi, M. (1966). The Tacit Dimension. Garden City, NY: Doubleday.

   Puri, M. (1994). The long-term default performance of bank underwritten security issues. Journal of Banking and Finance, 18, 397–418.

    Puri, M. (1996). Commercial Banks in Investment Banking. Conflict of Interest or Certification Role? Journal of Financial Economics,  40, 373 – 401.

   Puri, M. (1999). Commercial banks as underwriters: implications for the going public process. Journal of Financial Economics,  54, 133–163.

   Rajan, R. G. (1996). The Entry of Commercial Banks Into the Securities Business: A Selective Survey of Theories and Evidence. in A. Saunders and I. Walter, eds, Universal Banking: Financial System Design Reconsidered, Richard D. Irwin, Chicago, 282–302.

   Ramirez, C. D. (1995). Did J.P. Morgan’s Men Add Liquidity? Corporate Investment, Cash Flow. and Financial Structure at the Turn of the Twentieth Century. Journal of Finance, 50, 661–678.

   Redlich, F. (1968). The Molding of American Banking: Men and Ideas. New York, NY: Johnson Reprint Corporation.

    Rime, B. and K. J. Stiroh (2003). The Performance of Universal Banks: Evidence from Switzerland. Journal of Banking and Finance,  27,  2121–2150.

   Roten, I. C. and D. J. Mullineaux (2002). Debt Underwriting by Commercial Bank-Affiliated Firms and Investment Banks: More Evidence. Journal of Banking and Finance, 26, 689 – 718.

   Schumpeter, J. A. (1939). Business Cycles: A Theoretical, Historical and Statistical Analysis of the Capitalist Process. New York : McGraw-Hill.

   Seligman, J. (1982). The Transformation of Wall Street: A History of the Securities and Exchange Commission and Modern Corporate Finance. Boston, Mass: Houghton Mifflin Company.

   Stein, J. C. (2002). Information Production and Capital Allocation: Decentralized versus Hierarchical Firm. Journal of Finance,  57, 1891 – 1921.

Stern, G. H. and R. J. Feldman (2004). Too Big To Fail: The Hazards of Bank Bailouts. Washington, DC: Brookings Institution Press.

   Vander Vennet, R. (1999). The Effect of Mergers and Acquisitions on the Efficiency and Profitability of EC Credit Institutions. Journal of Banking and Finance,  20, 1531–1558.

   White, E. N. (1986). Before the Glass-Steagall Act: An Analysis of the Investment Banking Activities of National Banks. Explorations in Economic History, 23, 33–55.

   Winton, A. (1999). Don’t Put All Your eggs in One Basket? Diversification and Specialization in Lending. Working paper, Carlson School of Management, University of Minnesota, MN.

* Saïd Business School, University of Oxford. Many of the ideas concerning human capital in this essay arose during joint work with Bill Wilhelm, to whom I am extremely grateful for numerous conversations and insights. I am also grateful to Alexander Gümbel and Dimitri Tsomocos for comments on an earlier draft.

� Becker (1964) discusses human capital; its importance to development has been stressed by many authors. Tacit skill is discussed by Polanyi (1966).

� One might expect serfdom to provide the owners of peasants with incentives to invest in human capital. Gerschenkron argues that serfdom was symptomatic of a social sclerosis that undermined any tendency towards innovation.

� More recently, � HYPERLINK ".//C:/Academic/ResearchProjects/UnivBank/UniversalBanks.html.LyXconv/UniversalBanks.html" \l "XCharndlerAD:scacdi"��Chandler� (� HYPERLINK ".//C:/Academic/ResearchProjects/UnivBank/UniversalBanks.html.LyXconv/UniversalBanks.html" \l "XCharndlerAD:scacdi"��1990�) stresses the importance to the development of industrial capitalism of a large-scale infrastructure that can service expensive capital.

� They were J. P. Morgan & Co., First National and the National City Bank of New York, Kuhn, Loeb & Co. and, to a lesser extent, Kidder, Peabody & Co., and Lee, Higginson & Co.

� � HYPERLINK ".//C:/Academic/ResearchProjects/UnivBank/UniversalBanks.html.LyXconv/UniversalBanks.html" \l "XCarossoVP:invbah"��Carosso� (� HYPERLINK ".//C:/Academic/ResearchProjects/UnivBank/UniversalBanks.html.LyXconv/UniversalBanks.html" \l "XCarossoVP:invbah"��1970�, p. 276) discusses at some length the operation of securities affiliates

� � HYPERLINK ".//C:/Academic/ResearchProjects/UnivBank/UniversalBanks.html.LyXconv/UniversalBanks.html" \l "XMorrisonAD:invbip"��Morrison and Wilhelm� (� HYPERLINK ".//C:/Academic/ResearchProjects/UnivBank/UniversalBanks.html.LyXconv/UniversalBanks.html" \l "XMorrisonAD:invbip"��2007�, pp. 196-215) discuss the hearings, and their consequences, in some detail.

� See � HYPERLINK ".//C:/Academic/ResearchProjects/UnivBank/UniversalBanks.html.LyXconv/UniversalBanks.html" \l "XLeuchtenburgWE:fradrn"��Leuchtenburg� (� HYPERLINK ".//C:/Academic/ResearchProjects/UnivBank/UniversalBanks.html.LyXconv/UniversalBanks.html" \l "XLeuchtenburgWE:fradrn"��1963�, p. 59).

� Charles E. Mitchell, the president and board chairman of National City Bank, was paid a salary of $25,000, but awarded himself bonuses of $1 million in 1927 and 1928. � HYPERLINK ".//C:/Academic/ResearchProjects/UnivBank/UniversalBanks.html.LyXconv/UniversalBanks.html" \l "XSeligmanJ:trawsh"��Seligman� (� HYPERLINK ".//C:/Academic/ResearchProjects/UnivBank/UniversalBanks.html.LyXconv/UniversalBanks.html" \l "XSeligmanJ:trawsh"��1982�) discusses the hearings, and their legislative consequences, in detail.

� A year later, partners from Morgan and from Drexel founded the new firm of Morgan Stanley & Co. as an investment bank. See � HYPERLINK ".//C:/Academic/ResearchProjects/UnivBank/UniversalBanks.html.LyXconv/UniversalBanks.html" \l "XCarossoVP:invbah"��Carosso� (� HYPERLINK ".//C:/Academic/ResearchProjects/UnivBank/UniversalBanks.html.LyXconv/UniversalBanks.html" \l "XCarossoVP:invbah"��1970�) for a discussion of the industry changes that the Act caused.

� For example, � HYPERLINK ".//C:/Academic/ResearchProjects/UnivBank/UniversalBanks.html.LyXconv/UniversalBanks.html" \l "XMaceyJR:spcigl"��Macey� (� HYPERLINK ".//C:/Academic/ResearchProjects/UnivBank/UniversalBanks.html.LyXconv/UniversalBanks.html" \l "XMaceyJR:spcigl"��1984�) argues that the Act was intended to protect investment bankers at the expense of commercial bankers; � HYPERLINK ".//C:/Academic/ResearchProjects/UnivBank/UniversalBanks.html.LyXconv/UniversalBanks.html" \l "XLangevoortDC:staojp"��Langevoort� (� HYPERLINK ".//C:/Academic/ResearchProjects/UnivBank/UniversalBanks.html.LyXconv/UniversalBanks.html" \l "XLangevoortDC:staojp"��1987�) argues that Carter Glass believed that his bill would encourage banks to channel money towards small companies, rather than into the securities markets.

� See � HYPERLINK ".//C:/Academic/ResearchProjects/UnivBank/UniversalBanks.html.LyXconv/UniversalBanks.html" \l "XLownCS:chalfs"��Lown, Osler, Strahan and Sufi� (� HYPERLINK ".//C:/Academic/ResearchProjects/UnivBank/UniversalBanks.html.LyXconv/UniversalBanks.html" \l "XLownCS:chalfs"��2000�) for a survey.

� For related discussions, see � HYPERLINK ".//C:/Academic/ResearchProjects/UnivBank/UniversalBanks.html.LyXconv/UniversalBanks.html" \l "XMilbournTT:megest"��Milbourn, Boot and Thakor� (� HYPERLINK ".//C:/Academic/ResearchProjects/UnivBank/UniversalBanks.html.LyXconv/UniversalBanks.html" \l "XMilbournTT:megest"��1999�) and � HYPERLINK ".//C:/Academic/ResearchProjects/UnivBank/UniversalBanks.html.LyXconv/UniversalBanks.html" \l "XDierickF:supmfs"��Dierick� (� HYPERLINK ".//C:/Academic/ResearchProjects/UnivBank/UniversalBanks.html.LyXconv/UniversalBanks.html" \l "XDierickF:supmfs"��2004�).

� See � HYPERLINK ".//C:/Academic/ResearchProjects/UnivBank/UniversalBanks.html.LyXconv/UniversalBanks.html" \l "XBergerAN:effbc"��Berger, De Young and Udell� (� HYPERLINK ".//C:/Academic/ResearchProjects/UnivBank/UniversalBanks.html.LyXconv/UniversalBanks.html" \l "XBergerAN:effbc"��2001�) for a discussion of this directive, and of the consolidation of financial services in the European Union.

� Co-determination gives German workers a right of representation on the Board of all but the very smallest companies. Gorton and Schmid find that it worsens firm performance.

� See for example � HYPERLINK ".//C:/Academic/ResearchProjects/UnivBank/UniversalBanks.html.LyXconv/UniversalBanks.html" \l "XBarthJR:repgsa"��Barth et al.� (� HYPERLINK ".//C:/Academic/ResearchProjects/UnivBank/UniversalBanks.html.LyXconv/UniversalBanks.html" \l "XBarthJR:repgsa"��2000�).

� � HYPERLINK ".//C:/Academic/ResearchProjects/UnivBank/UniversalBanks.html.LyXconv/UniversalBanks.html" \l "XDeLongBJ:didjpm"��De Long� (� HYPERLINK ".//C:/Academic/ResearchProjects/UnivBank/UniversalBanks.html.LyXconv/UniversalBanks.html" \l "XDeLongBJ:didjpm"��1991�) argues that the impossibility of matching J. P. Morgan’s reputation gave the firm a strong competitive position in the nineteenth century. � HYPERLINK ".//C:/Academic/ResearchProjects/UnivBank/UniversalBanks.html.LyXconv/UniversalBanks.html" \l "XMorrisonAD:invbip"��Morrison & Wilhelm� (� HYPERLINK ".//C:/Academic/ResearchProjects/UnivBank/UniversalBanks.html.LyXconv/UniversalBanks.html" \l "XMorrisonAD:invbip"��2007�, chapters 4-8) trace the evolution of the modern investment bank.

� The reason for this is two-fold: first, partnership capital is provided by the partners, who have limited resources; and second, the number of partners is limited by a free-rider problem amongst partners.

� � HYPERLINK ".//C:/Academic/ResearchProjects/UnivBank/UniversalBanks.html.LyXconv/UniversalBanks.html" \l "XMorrisonAD:invbip"��Morrison and Wilhelm� (� HYPERLINK ".//C:/Academic/ResearchProjects/UnivBank/UniversalBanks.html.LyXconv/UniversalBanks.html" \l "XMorrisonAD:invbip"��2007�, pp. 238-249) discuss the phenomena outlined in this paragraph.

� This argument is given in greater detail, and with more supporting statistics, in � HYPERLINK ".//C:/Academic/ResearchProjects/UnivBank/UniversalBanks.html.LyXconv/UniversalBanks.html" \l "XMorrisonAD:demibp"��Morrison and Wilhelm� (� HYPERLINK ".//C:/Academic/ResearchProjects/UnivBank/UniversalBanks.html.LyXconv/UniversalBanks.html" \l "XMorrisonAD:demibp"��2008�).

� The most prominent derivatives trading partnerships were O’Connor, CRT, and Cooper Neff, which were acquired by Swiss Bank, Nations Bank and BNP, respectively: see � HYPERLINK ".//C:/Academic/ResearchProjects/UnivBank/UniversalBanks.html.LyXconv/UniversalBanks.html" \l "XMorrisonAD:invbip"��Morrison and Wilhelm� (� HYPERLINK ".//C:/Academic/ResearchProjects/UnivBank/UniversalBanks.html.LyXconv/UniversalBanks.html" \l "XMorrisonAD:invbip"��2007�, p.279).

� See Brunnermeier (2009)

� For example, Both Will Hutton and Jon Moulton argued in evidence to the UK's House of Commons Treasury Select Committee on 13 January 2009 that the repeal of Glass Steagall fanned the flames of the credit boom and, ultimately, that it was a contributing factor to the credit crunch. Geoffrey Wood made a similar point in evidence to the House of Lords Economic Affairs Committee on 20 January 2009, stating that complexity allows for rapid bank failure, which in turn creates systemic problems.

� See “Universal model dies as bank goes back to basics,” Greg Farell, Financial Times, 14 January 2009