Senate Banking, Housing and Urban Affairs Committee

Subcommittee on Financial Institutions and Regulatory Relief

Prepared Testimony of Mr. Charles W. Calomiris
Professor of Finance and Economics
Graduate School of Business, Columbia University

The Federal Reserve Proposals to Modify the ‘Firewalls'
that Separate Commercial Banks and Their Securities Affiliates
10:00 a.m., Thursday, March 20, 1997

Mr. Chairman, members of the Subcommittee on Financial Institutions and Regulatory Relief, it is a pleasure to appear here today. The last decade has seen dramatic and beneficial bank deregulation. Of particular importance have been the relaxation of geographical and activity limitations on bank holding companies, achieved respectively by Congressional action and new regulatory interpretations of existing law. Recently, both the Federal Reserve Board and the Comptroller of the Currency have sought to further enhance the flexibility of bank operations by reducing limits on bank activities, and there is a push for Congressional action to repeal outright the separation of commercial and investment banking imposed in the Banking Act of 1933.

While deregulation has been dramatic, it is worth noting that it has sparked little controversy, or even surprise, among scholars of banking or banking history. There is a remarkable degree of agreement among these scholars - supported by an extensive body of research - that historic limitations on bank locations and activities have served no legitimate economic purpose. Rather, these costly limits - especially the separation of commercial and investment banking in 1933 - are artifacts of specific historical political battles in which facts and economic logic have often taken a back seat to populist passions or special interest politics.

Competitive pressure - especially the loss of U.S. banks' domestic and international market shares in the l98Os - helped to push the Federal Reserve Board in 1987 and subsequently to use its authority to relax restrictions on bank underwriting activities. Initially, the Fed proceeded with extreme caution. It allowed very limited underwriting of private debt, and later equity, but attached to those activities a host of "firewalls" intended to serve two different sets of purposes - (l) to prevent the transfer of loss from a Section 2O underwriting affiliate to a chartered bank, and (2) to prevent "conflicts of interest" from arising from the mixture of commercial and investment banking. The first concern was largely that of economists who correctly worried about the abuse of deposit insurance and the discount window - the possibility of government subsidization of risk in new activities. The second concern was that banks might use their new powers to coerce client firms or cheat purchasers of securities (despite clear and strong economic incentives not to do so). This second concern reflected the unsubstantiated and misguided arguments that underlay the Pecora Hearings of 1932, and the limitations imposed on the mixing of commercial and investment banking in 1933.

The Federal Reserve Board has recently circulated a set of proposals to relax or eliminate many of the "firewalls" that limit connections between banks and the non-bank affiliates of bank holding companies. It is worth noting that if the Fed proceeds with these proposed changes, the consequences will stretch beyond the holding company affiliates regulated by the Fed, since the Office of the Comptroller of Currency, which has invited banks to channel their "non-banking" activities through national bank subsidiaries, is almost sure to follow the Fed's lead on firewalls, at least in the near term.

The principal question I will address today is the desirability of the Fed's recent proposals to reduce firewalls. More broadly, however, I will assess the desirability of expanding bank activities into securities markets and related areas, and thee importance of relaxing firewalls to allow efficiency gains from such an expansion. Mr. Chairman, at the end of my remarks, I will also provide responses to the specific written questions you posed in your letter of March 14.

To summarize, I support enlarging the range and size of permissible "non-bank" activities within banks or bank holding companies, and scaling back the firewalls that the Federal Reserve Board established in 1987 and 1989, as the Federal Reserve Board now proposes to do. Doing so will allow banks to compete more effectively for corporate relationships, and will benefit bank customers and their stockholders at least as much as it will benefit banks. I will also argue that outright repeal of Glass-Steagall prohibition~,s and further relaxation of firewalls should be considered by Congress and bank regulators.

Let me begin with some historical perspective on the origins of the separation of commercial and investment banking, which is essential to the debate over bank activity limits and firewalls. In the United States, banks became interested in underwriting and holding corporate debt and equity in the 192Os. The relaxation of branching laws during that period created new opportunities for banks to become large. As they became large, they turned increasingly to industrial as opposed to only commercial finance, and they found it profitable to mix bank lending with underwriting. Doing so made it possible to better serve their corporate clients' funding needs, provide better access to investments for their trust customers, and reduce banks' costs of funds, since the new activities were highly diversifying.5

The mixing of lending and underwriting activities had produced important synergies in German universal banking around the turn of the 20th century, which were well understood in the United States by the 192Os. German banks offered very rich portfolio opportunities to their trust customers, charged very little (even by late 2Oth-century standards) for equity underwriting, and managed to promote high-tech, large scale industrialization in Germany at an unprecedented speed.

Both: the American and German historical experiences support the view that allowing banks to lend, underwrite, and manage securities accounts substantially reduced bank costs of intermediation. The synergies from mixing lending, equity holding, and underwriting in the United States and Germany historically are present and visible in the United States today, even under the current limited incarnation of universal banking. Evidence from the current U.S. experience suggests that Section 20 (underwriting) affiliates increase average returns and reduce risk for their bank holding companies. Venture capital finance is another area in which there are strong corporate finance synergies, both with lending and with underwriting, since venture capital finance precedes public stock offerings.

The primary synergies are the repeat use of information, the enhancement of corporate control, and the diversification of bank income.8 By lending and underwriting, banks are able to use knowledge about the performance and risk of their corporate borrowers to reduce the cost of underwriting and marketing securities. Furthermore, informed banks' involvement as holders of equity (mainly via trust account proxies) can enhance their control over firms and helped them to protect trust customers' interests, which also makes it easier to market equity. Finally, because lending and underwriting tend to be imperfectly correlated risks, combining the activities can help banks to diversify their assets, which economizes on bank capital and reduces banks' costs of funds.

It is important to keep in mind that in a competitive banking system these reduced costs of intermediation accrue mainly to bank customers (that is, banks' clients' stockholders and bank depositors). It is also important to note that these synergies are undermined by firewalls.

If firewalls require separation of types of activities into different entities (that is, entities with separate corporate identities, liabilities, and capital) then cost of funds advantages from diversification across activities will be severely limited. If firewalls require that lending and underwriting by the bank not be allowed for the same client firm, then any potential information synergies across activities will be foregone. If firewalls prevent banks from being actively involved in controlling the firms they finance (via boards of directors, or trust account and mutual fund holdings of stock), then control synergies will be limited.

The current Fed proposals to repeal firewalls would improve the efficiency of the financial system mainly by increasing the multiple use of information within the bank and by enhancing bank involvement in corporate control. These proposals reflect a growing realization that concerns over conflicts of interest are largely unfounded (as they were in 1933) because dishonest behavior is contrary to the interests of the bank, especially in the highly competitive financial services industry, and is already proscribed by other laws and regulations.

It is worth noting that the Fed is not proposing to remove all firewalls. The Fed proposes to retain the limitations on the mixing of "bank" and "non-bank" risks (specificially, Sections 23a and 23b of the Federal Reserve Act, which limit the extent of lending from banks to non-bank affiliates, and require that it be fully collateralized and priced as an arms-length loan). Chairman Greenspan argues in his testimony of February 13, 1997 that Sections 23a and 23b provide essential protection against the abuse of deposit insurance protection by bank holding companies.

I share Chairman Greenspan's concern that under the current deposit insurance system the absence of financial firewalls (like 23a and 23b) could lead to abuse of deposit insurance. But I do not agree with his solution to the problem. A better alternative would be to remove the possibility that banks would receive subsidized insurance, which would thereby remove any incentive problem from mixing the risks of lending and underwriting, and thus allow banks and their customers to reap the gains from relaxing Sections 23a and 23b. Such reform is feasible and the relative merits of various approaches for doing so - including subordinated debt requirements and other deposit insurance reforms - are currently being debated among bankers, academics, and regulators.

Chairman Greenspan also argued that the holding company structure is a superior mechanism for administering firewalls, and that thus it is essential to restrict Section 20 activities (and other "non-bank" activities) to bank holding company affiliates. Thus he does not support the Comptroller's "Part Five" initiative to allow new activities to occur within subsidiaries of national banks. I do not share his view. There is no reason I can discern why any particular firewall cannot be administered as well within a bank subsidiary structure as it is within a bank holding company affiliate structure. Also, it is my understanding that there is little merit to the argument that legally, the "corporate veil" of a subsidiary is more easily pierced than that of an affiliate.

The argument for retaining a Federal Reserve monopoly on regulating Section 2O activities seems, then, to rest on the view that there is a need to have an "umbrella" regulator (namely, the Fed) that oversees all holding company risks. That is only necessary, logically, if two conditions are satisfied: First, if potential subsidies from deposit insurance cannot be eliminated, and second, if firewalls cannot insulate banks from the risks undertaken by non-bank subsidiaries or affiliates. Since l believe neither of these necessary conditions is true, I do not support the Fed's view that it should retain a regulatory monopoly over non-bank activities.

An advantage to allowing the Comptroller to "compete" with the Fed over nonbank activity regulation is that such competition will promote a healthy tendency for both regulators to properly balance the costs and benefits of firewalls, and thus produce neither excessive caution nor imprudent deregulation. Competing regulators jealous of each other's "turf" will not be imprudent because a financial collapse of a bank under their authority would risk a significant loss of their powers. At the same time, excessive caution would produce an immediate loss of regulatory power, since banks would gravitate toward the more liberal regulator. Thus competing regulators are likely to consider and balance both costs and benefits when making regulatory decisions.

In conclusion, I strongly support the Fed's proposal to roll back firewalls. I also encourage the Fed, the Office of the Comptroller, and Congress to consider reforms to deposit insurance that would permit additional rolling back of firewalls. Finally, I encourage Congress to retain the current healthy competition that exists between the Office of the Comptroller of the Currency and the Federal Reserve Board.

If Congress and bank regulators succeed in freeing banks from costly and unwarranted limitations on their powers, the advantages for banks and their customers will be large. I would emphasize that the banks that will gain from such expansion of powers will increasingly include super-regional and regional banks, in addition to moneycenter banks. As the technology for placing securities improves, and barriers to low-cost public issues of securities are removed for many firms, it seems likely that more and more regionals and super-regionals will find it advantageous to establish Section 20 affiliates or subsidiaries to publicly place their clients' debt and equity in securities markets. Those innovations and the curtailment of firewalls will also encourage regionals and super-regionals to enter the highly profitable area of venture capital finance via SBICs (Small Business Investment Companies) or Regulation Y venture capital affiliates. Venture capital finance is an important middle ground in the financial life cycle of firms between an early "bank lending stage" and a late "public securities financing stage." Thus bank involvement in venture capital will magnify synergies between bank lending and public securities underwriting.

Not only will regional and super-regional banks benefit from those changes, but so will their corporate clients. Many more firms will enjoy lower costs of finance from access to private equity (in venture capital affiliates), which leads naturally to public debt and equity offerings (underwritten by Section 20 affiliates). The social gains of that expanded access to private and public equity finance could be large.

Answers to Specific Questions

Below I provide answers to the questions posed in Chairman Faircloth's letter of March 14, 1997.

#1. [Provide] a historical perspective on the separation of investment and commercial banking and the development of the firewalls in 1987.

The separation of commercial and investment banking in 1933 reflected two factors. The first was the Depression-era view (seemingly confirmed by the Pecora hearings) that securities underwriting by banks had been "dishonest" and had crippled banks during the Depression. Recent research (noted above) has shown that securities underwriting by banks prior to 1933 was at least as honest as securities underwriting outside of banks, and that securities activities helped to stabilize banks because they were highly diversifying. The second, and most important, factor leading to the passage of Glass-Steagall was Carter Glass's fervent commitment to separating commercial and investment banking. The Banking Act of 1933 was a compromise between Steagall (who advocated deposit insurance as a palliative for small banks in distress) and Glass (who pushed for the separation of commercial and investment banking, and for Regulation Q). Glass was an original architect of the Federal Reserve Act of 1913, and an advocate of the "real bills doctrine," which held that banks should mainly finance commercial, rather than industrial activity. Regulation Q was designed largely to undermine reserve pyramiding in New York, and thus remove banks from financing activities in the call money market. Similarly, the separation of commercial and investment banking would encourage banks to avoid securities dealings, and industrial finance more generally. The real bills doctrine was viewed as a means to preserve the money supply from unwarranted supply fluctuations unrelated to real economic activity. Obviously, this debate is not relevant to current bank regulatory policy, since Section 20 affiliates' activities have no effect on the money supply. Furthermore, to my knowledge, the real bills doctrine has no current advocates.

#2. Can changes be made to firewalls while preserving the safety and soundness of the banking system and the integrity of the insured depository institution?

To reiterate, I believe that there is no threat to the safety of banks or the banking system from relaxing the firewalls the Fed is now proposing to relax. If the Fed were considering relaxing Sections 23a and 23b without making additional changes to deposit insurance law, that might pose a problem, but that is not what the Fed is proposing to do.

#3. Can changes be made to firewalls while adequately protecting customers of the bank or securities affiliate from coercion?

Evidence suggests that allowing closer linkages between commercial bank and underwriting operations would improve banks' information about their customers and their ability to control their financial decisions, but that does not mean it would increase "coercion" (which I would define as the power of banks to extract "rents" from their customers). Whether beneficial discipline or non-beneficial coercion results from expanded bank powers depends on how competitive the banking system is. In today's highly competitive system, discipline rather than coercion would result from greater bank powers. Indeed, one could argue that because expanding powers via relaxed firewalls would increase competition in financial services broadly defined, it would actually reduce the potential for coercion. It is important to note that corporate bank customers do not see any threat to their well being from expanding bank powers. Indeed, corporate clients of banks were among the earliest advocates of allowing banks to underwrite securities. The Fed's decision to permit banks to engage in underwriting specifically cited survey results of corporate managers, who overwhelmingly recommended doing so.

#4. Will the firewalls give securities firms affiliated with an insured depository institution an advantage over other firms?

The fact that there are synergies between commercial banking and investment banking implies that banks' competitive positions would be improved relative to investment banks by relaxing firewalls. But that is not to say that banks will be better off because of their access to deposit insurance, per se. So long as Sections 23a and 23b limit the mixing of risks between banks and non-bank affiliates (or subsidiaries) there is little potential for abuse of deposit insurance, even if deposit insurance were subsidized, and thus the presence of deposit insurance cannot provide a competitive advantage.

#5. If the firewalls are modified what other protections exist in statute and regulation that will serve as a substitute?

The Fed's "Review of Restrictions in the Board's Section 20 Orders" (January 9, 1997) reviews a large range of remaining safeguards, including capital requirements, Section23a and 23b restrictions, SEC regulations, and other laws and regulations. I agree with the Fed's view that these remaining safeguards are adequate.

#6. Please offer general comments on other recent modifications made by the Board to Section 20 activities, including the raising of the cap from 10% to 25% and the easing of cross-marketing restrictions.

I am in favor of both. Obviously, synergies will be larger if the volume of equity underwriting is larger, since this is one of the most important sources of the gains discussed above. I would be in favor of eliminating the revenue cap altogether. With respect to cross-marketing, it is important to note that banking is a "relationship-based", not a "product-based" business. cross-marketing is essential to the logic of the efficiency gains from universal banking. Recent research has argued against the view that dishonest behavior would result from permitting cross-marketing.

Home | Menu | Links | Info | Chairman's Page