Are Banks Special

Are Banks Special? Introduction The recent evolution of the financial structure in the United States has produced two competing points of view regarding the proper direction for further change. On the one hand, there is the view that the "financial services industry" — encompassing banks, thrifts, brokers, investment banks, and insurance companies—should be looked at as a single entity. According to this view, efforts to distinguish among kinds of institutions are both futile and unnecessary. This view of the financial services industry is based on the belief that many financial services offered by various classes of institutions are so complementary to (or such close substitutes for) one another that institutional distinctions are rendered useless. Implicit in this view is the assumption that banks are not special. * The competing, if not opposing, view is that banks are indeed special. This view holds that specialization of financial institutions has worked well and, at least in some cases, specialization may still be more efficient and also better serve the public interest. This view is associated with the historical separation of banking from commerce and from investment banking. In general, this "separation doctrine" in banking grew out of concerns about concentration of financial power, possible conflicts of interest, and the appropriate scope of risks banks should incur in the face of the special trusteeship falling on institutions that engage in the lending of depositors' money. In a shorthand way, as pertains to banks and the banking system, these concerns are typically captured by the phrase, "safety and soundness." These two points of view do not necessarily represent mutually exclusive approaches to financial market structure. For example, in the context of a large financial services holding company, banks could be legally separated from nonbanks, but it is not clear that such separation would necessarily provide the kinds of protections that are currently built into federal banking laws. Thus, assessing the merits of these two competing views must start with some very basic questions: Are banks "special" or are they simply another provider of financial services? Does it matter what kinds of risks banks incur? Does it matter who owns banks? Is "safety and soundness" a cliche, or should it have genuine and substantial meaning for banks, for bank regulators, and for the public at large? While banking practices have naturally evolved over time, recently a combination of events has shifted that process to one of an almost revolutionary character. Amidst this process of rapid change, with market innovation and new sources of competition, there is a perception that banks' competitive position—and presumably their market share—has slipped. Casual observation of 1 the growth of the commercial paper market, the thrift industry money market mutual funds (MMMFs), and the de facto trend toward ownership of banks by securities firms and commercial enterprises, tends to support that perception. Indeed, there are numerous instances in which nonbanks have been able to provide "bank-like" services at a lower cost (or a higher rate of return) to the individual or corporate customer, thereby drawing business away from banking institutions. High on the list of reasons that are cited for this perceived shift of market position from banks to nonbank competitors is the extra burden of regulation on banks. The fact of a heavy regulatory burden on banks is beyond dispute, but in some cases it is also true that regulation— relating to, for example, deposit insurance or access to the discount window—provides powerful incentives for individuals and businesses to maintain relationships with banks. While it is difficult to judge the net competitive results of differing degrees of regulation, it does seem clear that of all the regulatory burdens on banks, there have been two that stand out in terms of their impact on banks' competitive position over time: Regulation Q and limitations on the scope of bank services. This is not to suggest that other regulations on banks—ranging from reserve requirements to community reinvestment—have not been costly. But, at the cutting edge of market position or market share, it is Regulation Q and service line restrictions that have been the most critical restraints on banks. Despite these regulatory restraints, banks have not stood still in the face of changing financial markets and new sources of competition. By using the flexibility provided by the Bank Holding Company Act, by developing sophisticated liability management techniques, by major expansions abroad, and by creative and innovative adaptations of "conventional" banking services, banks have actually fared rather well in terms of preserving their overall market position. While it is not easy to measure what has happened to the relative position of banks over time, the appendix to this report makes such an effort. Allowing for the inherent measurement problems in such an exercise to say nothing of the data limitations the analysis simply does not bear out the perception noted earlier that banks have lost ground in the domestic marketplace over the past three decades. (While not captured by the data, banks have, of course, made major expansions abroad during this period.) The analysis does not, however, imply that heavy regulation has not constrained the growth of banks and their market share, for it is quite possible that absent such regulations, banks' position would have risen rather than essentially held steady. Nor does the analysis indicate whether a rising or falling bank share is good, bad, or indifferent from the perspective of the public interest. To some extent these issues depend upon whether, in fact, there is something special about banks that is worth preserving. Indeed, if banks are special it would not be in the public interest for the features or functions that make banks special to be eroded by competitive, regulatory or legislative forces. By the same token, if what is special about banks dictates a relatively heavy dose of regulation, public policymakers should not be goaded into eliminating necessary regulation simply because bank market share might grow to some higher level without that regulation. What Makes Banks Special? Reduced to essentials, it would appear that there are three characteristics that distinguish banks from all other classes of institutions—both financial and nonfinancial. They are: 2 1. Banks offer transaction accounts. 2. Banks are the backup source of liquidity for all other institutions. 3. Banks are the transmission belt for monetary policy. These three essential bank characteristics and the interrelationships between them are discussed below. Of necessity, the discussion treats each factor separately. However, it is clear that these essential characteristics are highly complementary and furthermore that it is the relationship among them that best captures the essence of what makes banks special. Issuance of Transaction Accounts Only banks issue transaction accounts; that is, they incur liabilities payable on demand at par and are readily transferable by the owner to third parties. The owner of a transaction account can demand and receive currency in the face amount deposited in the account; write a check in the full amount of the account; or perhaps most importantly the owner of the account can transfer the full amount of the account to a third party almost instantaneously by wire transfer. The liquidity, mobility, and acceptability of bank issued transaction accounts permit our diverse economic and financial system to work with the relative ease and efficiency to which we are accustomed. Moreover, in periods of financial stress, the capacity to quickly move transaction account balances to third parties takes on special significance by providing elements of flexibility and certainty in making and receiving payments that help to insure that financial disruptions do not spread. Individual banks can also create these highly liquid and mobile balances through their lending function. The capacity to "create" liabilities with these characteristics is vital to the ongoing needs of commerce, but it takes on special significance in periods of financial stress. Because of the peculiarities of law and regulation, not all classes of transaction accounts have the same precise legal or regulatory characteristics. The "demand deposit" is the purest form of transaction account, since, for example, negotiable order of withdrawal (NOW) accounts and some share drafts at mutual organizations have restrictions on the extent to which they are payable on demand. However, from the perspective of both the issuing institutions and their customers, these differences appear to be without substance since the accounts are perceived and treated as transaction accounts both by the issuing institution and by the public. For this reason, a contemporary definition of "transaction" accounts—at least for purposes of identifying and defining special characteristics of banks—should focus on functional characteristics rather than existing legal or regulatory distinctions. If a financial asset satisfies the functional test of being payable on demand at par and readily transferable to a third party, it should—for those purposes—be a "transaction" balance. A case can be made that nonbank financial institutions incur liabilities that appear to have some or all of the characteristics of a transaction account issued by a bank. However on close inspection it appears that such instruments whether—MMMFs, retail repurchase (RPs) agreements, customer credit balances with brokers, sweep accounts, etc.—do not, at least in a technical sense, in fact possess the characteristics associated with the bank issued transaction account. However, as is discussed later making the distinction is particularly difficult in the case 3 of MMMFs. In all of these cases, including money market mutual funds, instruments which appear to have bank transaction account characteristics take on those characteristics in part because the acquisition or disposition of such assets involves, at some point, the use of a transaction account at a bank. However, technology makes it possible to manage these financial assets in a way in which their ultimate dependence on a bank account is not apparent to the individual holder of the asset. As long as banks issue transaction accounts they, by definition, incur a form of "term structure" risk. That is, the presence of transaction balances on the books of a bank makes it difficult, if not impossible, to match the maturities of assets and liabilities, particularly in a contemporary setting in which bank holdings of liquid assets have shrunk and in which some assets, traditionally considered as liquid, may not, in fact, be all that liquid. Indeed, the asset side of the balance sheet for at least some banks provides a small margin of functional liquidity that can readily be brought to bear to meet large and sudden deposit outflows. In this setting, the inherent term structure mismatch on the books of banks is one of the realities that gives rise to concerns about strains on bank liquidity and sudden drains on bank deposits. Banks and bank regulators have long since recognized the importance of banks acting in ways that preserve public confidence in banks' capacity to meet their deposit obligations, thereby minimizing the likelihood of large, sudden drains of bank deposits. Deposit insurance and direct access to the lender of last resort are uniquely available to banks to reinforce that public confidence. Indeed, deposit insurance and access to the lender of last resort constitute a public safety net under the deposit taking function of banks. The presence of this public safety net reflects a long-standing consensus that banking functions are essential to a healthy economy However the presence of the public safety net uniquely available to a particular class of institutions also implies that those institutions have unique public responsibilities and may therefore be subject to implicit codes of conduct or explicit regulations that do not fall on other institutions. Experience suggests rather strongly that public confidence in a bank— with or without deposit insurance and the Feds' discount window— is ultimately related to public perceptions about the financial condition of banks and specifically about the quality of banking assets, liquidity capital, and the capacity to absorb short-run shocks. Sudden drains on bank deposits occur when depositors conclude that loan losses or other circumstances might jeopardize a banks' ability to meet its deposit obligations. The evidence is overwhelming, for example, that most "problem" bank situations in recent years involved concerns growing out of losses or perceived losses associated with lending, securities activities, foreign exchange activities, and/or poor management. In this regard, it should be noted that even when "problem" bank situations have been resolved with a minimum of costs to the individual institution, these situations have, on occasion, involved high costs in terms of generalized financial market disruption. Thus, while deposit insurance and access to the lender of last resort may rightly be viewed as the public policy safety net under banks' deposit taking function, the integrity of the deposit taking process and therefore the strength of the public safety net process depend to a substantial degree on the prudent management and control of risks on the part of the banking system as a whole. 4 Looked at in this perspective, the critical difference between banks and other classes of financial institutions rests with the capacity of banks to incur (and to create) liabilities that are payable on demand at par and that are readily transferable to third parties. The resulting mismatch of the maturities of assets and liabilities makes banks particularly vulnerable to sudden drains on deposits that can jeopardize their solvency. In practice, depositors—reinforced by the public policy safety net—have demonstrated tendencies to drain deposits from particular banks only when confronted with the reality or the perception of losses growing out of asset management problems and/or poor management of banking organizations. Thus, while the deposit taking function of banks is what makes them unique, the integrity of that process depends upon the risks, real and perceived, associated with the lending and related activities of the banking system as a whole and its capacity to absorb shocks in the short run. Backup Sources of Liquidity As discussed above, the fact that banks issue transaction deposits is the key factor that distinguishes them from other classes of financial and nonfinancial institutions. However, experience also suggests that public confidence in the ability of banks to meet their deposit obligations is ultimately related to the quality of bank assets and to the overall financial condition of the bank. This relationship takes on additional importance when it is recalled that banks can also create, through their lending activities, transaction deposits. Indeed, in a very real way banks are the primary source of liquidity for all other classes and sizes of institutions, both financial and nonfinancial. The extent to which banks play this role cannot be judged simply by looking at the number and value of loans on the books of banking organizations. For these purposes, contingent credit obligations of banks, such as loan commitments and standby letters of credit, must be considered in virtually the same light as direct loans. These standby credit facilities are, for example, the arrangements which permit most financial markets and institutions to function as they do. It is highly unlikely that the commercial paper market would function very well were it not for the presence of standby bank credit facilities obtained by those corporations that issue commercial paper. Similarly, it is very difficult to imagine that even the best managed and capitalized broker/dealers could handle their day-to-day business with the efficiency that is now so common without ready access to bank lines of credit. The same, of course, applies to nonfinancial corporations. Indeed, while all such institutions may over time, have access to a wide variety of funding sources, direct or standby bank credit facilities are the cornerstone upon which these alternative sources of credit rest. If there are problems in one segment of the credit network, institutions will simply shift their borrowing activities elsewhere in the network. However, if the problem is in the banking sector, banks must either turn to each other or to the central bank. Even in the "normal" course of events, the direct and standby credit facilities provided by banks are the foundation upon which other credit markets depend for their vitality. This relationship takes on special significance, however, in periods of selective or generalized financial stress. For example, in virtually every case of "selective" financial shock in the 1970s and early 1980s, troubled institutions—financial and nonfinancial, bank and nonbank—turned to the banking system to provide at least a bridge until more lasting solutions to the problem could be worked 5 out. At the very least, these bridging arrangements helped to contain problems and prevent them from spreading to other institutions or to the financial system generally. Banks' ability to supply credit and liquidity, particularly in situations where other institutions or markets may be unwilling or unable to do so, arises because the deposit creating function of banks (in tandem with banks' relationship with the central bank) provides an element of credit and liquidity elasticity which is not immediately available to other institutions. In point of fact, the extent and frequency with which banks have had to directly rely on extraordinary funding by the central bank (either through the discount window or via open market operations) have been quite limited. In the normal course and even in periods of stress, individual banks and the banking system as a whole are able to provide necessary liquidity because of their ability to quickly fund loans through a variety of market sources including the domestic and foreign interbank market, RPs, the issuance of large certificates of deposit (CDs), and so on. For many banks, access to these markets has become the primary source of bank liquidity. Banks' access to these markets—and by extension, banks' ability to function as backup sources of liquidity—occurs in a context in which individual suppliers of such funds—whether federal funds, CDs, Eurodollars, etc.— make judgments about the strength and vitality of individual banks and the banking system as a whole. Experience is clear for example, that individual banks experiencing problems with classified assets, earnings, and so on, often see that phenomenon first manifest itself in the form of having to pay a risk premium over the "going" rate for federal funds and large CDs. Similarly when concerns about the banking system arose in 1974-1975 and more recently in 1982, an early manifestation was a widening of the interest rate spread between bank and treasury liabilities of comparable maturities. In the extreme cases of severe problems with individual banks, widening spreads ultimately result in these sources of funding being cut off, with a consequent need to either contract the size of the bank, borrow from the Feds' discount window, or in some cases, close or merge the bank. The point is, of course, that the ability of a bank to fulfill its role as a backup supplier of liquidity to the financial and business communities depends on easy access not only to traditional sources of deposit liabilities, but also to markets for nondeposit sources of funding. The same applies to the banking system as a whole, because while one or a few banks can turn to the London market to fund themselves in times of adversity, it is clear that the banking system as a whole cannot. Thus, as with the preservation of the integrity of the deposit taking function described earlier, experience clearly suggests that the ability of banks to provide the essential function of a backup source of liquidity is ultimately dependent on market judgments as to the quality of the banks' assets and overall financial strength. Looked at in this light, the ability of banks to fulfill their role as standby sources of liquidity and credit rests importantly on the quality and consistency of credit judgments made by banks. This is particularly true in periods of stress when banks may be called on to supply credit to borrowers who, for one reason or another, temporarily do not have access to other sources of funds or to make the even more difficult decisions as in which borrowers are experiencing problems of a fundamental or irreparable nature. It is in these particular circumstances that banks must be in a position to make rigorous, impartial, and objective credit decisions, because it is precisely in such circumstances that the potential for compromise in the impartiality of the credit decision 6 making process is greatest and the potential for asset quality deterioration is the largest. It is in this light that considerations about the commingling of banking and other interests and concerns about the ownership and control of banks become compelling. To summarize, virtually all other financial markets and other classes of institutions are directly or indirectly dependent on the banking system as their standby or backup source of credit and liquidity. Banks can fulfill this function for a variety of reasons, including their relative ease of access to deposit and nondeposit sources of funding. However, experience suggests that the capacity to provide this function or more directly, to provide access to these markets and sources of funding—like the integrity of the deposit taking function—is ultimately related to the overall financial strength of banks and the quality of bank assets. This role of banks as a standby source of liquidity takes on special significance in periods of stress and in this light underscores the importance of rigorous and impartial credit judgments by banks. This, in turn, provides a particularly relevant context in which concerns about the commingling of banking and other interests should be evaluated. Transmission Belt for Monetary Policy As the preceding discussion suggests, there is a direct link between banks and the central bank arising in part from the central banks' lender of last resort function. More broadly the fact that banks are subject to reserve requirements places the banking system in the unique position of being the "transmission belt" through which the actions and policies of the central bank have their effect on financial market conditions, money and credit creation, and economic conditions generally. To put it somewhat differently, the required reserves of the banking system have often been described as the fulcrum upon which the monetary authority operates monetary policy. The reserves in the banking system also serve the complementary purpose of providing the working balances which permit our highly efficient financial markets to function and to effect the orderly end-of-day settlement of the hundreds of billions of dollars of transactions that occur over the course of each business day. Some have argued that neither monetary policy nor the payments mechanism are dependent on the relationship between reserves and the banking system. There have been, or are, schemes for conducting monetary policy and operating a payments mechanism that do not use bank reserves and the banking system in the way the U.S. system currently operates. However it is also true that any of these alternative arrangements would entail major institutional changes and run the risk that they might not work as efficiently as the current framework or the possibility that they might not work at all. In short, to justify departure from the current arrangement the weight of evidence should be overwhelming that the current system is not working or that some alternative system would work decidedly better. In fact, the current system seems to work rather well, although recent developments may have introduced elements of slack into the transmission belt. For example, the proliferation of close substitutes for bank-issued transaction accounts narrows the effective scope of reserve coverage. The narrowed reserve coverage can introduce more slippage into the process of monetary control, and it also means that a relatively smaller reserve base is supporting a larger flow of payments. Similarly the deregulation of the liability side of banks' balance sheets seems to imply 7 that, in order to achieve a given degree of monetary restraint, a higher level of market interest rates is required than might otherwise have been the case. Further, increased leverage of banking organizations may work in the direction of introducing slippage into the monetary control process, in that a larger volume of credit flows may be associated with some given rate of growth of "money". Finally, higher leverage and greater risk exposure may weaken the capacity of the banking system to adjust to and to absorb the changes in credit market conditions that must accompany periodic monetary restraint. As suggested above, these and other forces may already be working to introduce a larger margin of slack into the transmission belt. While the slack evident today is of manageable proportions, the future design of the banking and financial system must leave intact a strong yet adaptable mechanism through which monetary policy and the payments mechanism can function. This imperative underscores the case for attempting to segregate essential banking functions into an identifiable class of institutions and seeking to ensure that these institutions have the financial strength and vitality to perform their essential functions and to absorb changes in the credit market and economic conditions associated with periods of monetary restraint. Defining a Bank From the previous discussion, it should be clear that there are in fact certain special and unique functions of banks and that they are essential to the functioning of an efficient and safe financial and economic system. However, it also seems likely that if "banks" did not provide these essential functions, someone else would—just as it is abundantly clear that the process of market innovation has already produced services which are close substitutes for essential bank services. Given these considerations, the threshold question that arises is whether it is still desirable, from a public interest point of view, to attempt to segregate essential banking functions into an identifiable class of institutions and, if that is the case, whether it is possible to define a bank in a manner that is both functionally and intellectually satisfactory Putting aside for the moment practical problems of definition, it would seem that the case for segregating essential banking functions into an identifiable class of institutions is every bit as powerful today as it was in the 1930s. If anything, concerns regarding financial concentration, conflicts of interest, and the fiduciary responsibilities associated with lending depositors' money may be more relevant today than they were 50 years ago. To be sure, the lines of distinction may not have to be drawn in the same way and in the same place that they were in the past, but the earlier discussion of the essential functions of banks serves as a powerful argument for separation at some point. Indeed, to reject the notion of separation would—as a matter of logic— require that deposit insurance and access to the lender of last resort, together with the associated supervisory and regulatory apparatus, either be done away with altogether or be made universally available to any institution that provides essential banking functions—irregardless of what other types of business or commerce it might be engaged in. However as a practical matter, the case for separation is only viable if we are able to provide a satisfactory definition of a bank. Over time, a variety of tests have been used for the purpose of defining a bank. These tests ranged from a charter test to the functional test of issuing demand deposits and making 8 commercial loans. At one time, each of these tests was satisfactory. However, currently neither existing statutes nor regulations seem to contain a definition that is satisfactory. A satisfactory definition of a bank must start with a clear recognition of the essential functions provided by such institutions. From the earlier discussion, it is clear that the single characteristic of banks that distinguishes them from other classes of institutions is that they issue transaction accounts; that is, accounts that in law, in regulation, or in practice are payable on demand at par and are readily transferable to third parties. A powerful case can be made that the definition of a bank should stop right there: a bank is any organization that is eligible to issue transaction accounts. If an institution meets this test, it would (1) be eligible for government deposit insurance; (2) have direct access to the discount window; (3) be subject to the Feds' reserve requirements; and (4) have direct access to the Federal Reserves' payments services, particularly the wire transfer system. For these purposes, an appropriate statute would have to redefine transaction accounts. At a minimum, such a definition would have to include conventional demand deposits, NOW accounts, and share drafts. It might also include the new money market deposit accounts (MMDAs) and, depending on the standards of definition, perhaps even MMMFs or other nonbank institutional arrangements that provide "check" writing capabilities. On the surface, this definition of a bank may seem inadequate because it contains no corollary asset or lending test; it focuses only on the liability side of the balance sheet. This seeming inadequacy arises in part because the current Bank Holding Company Acts' definition requires that a bank issue demand deposits and make commercial loans. More substantially, the absence of a lending test seems to fly in the face of arguments made earlier concerning the critical link between the deposit taking function and the lending or asset acquisition functions of banks. However, it is precisely because of the nature of the relationship between deposit taking and asset acquisition that the essential definition of a bank should be couched in terms of its deposit taking function—without regard for the particular distribution or classification of its loans and/or investments. Taken by itself, there is nothing unique or special about the asset side of a banks' balance sheet, except for the limits on the scope of asset acquisition powers discussed below. Concerns about the nature and risk characteristics of bank assets arise in the context of the unique nature of bank liabilities, the need to preserve the integrity of the deposit taking function, and the special trusteeship growing out of that function. Thus, while it may be appropriate from the standpoint of public policy to limit the asset powers of banks to certain less risky activities, the definition of a bank need only deal with the liability side of the balance sheet. The absence of an asset test might, however, create a definitional loophole. That is, "banks" could conceivably refrain from issuing transaction deposits while funding their asset acquisition activities with insured time and savings deposits. However, this problem could be minimized by reliance on such an institutions' eligibility to issue transaction accounts. If so eligible, it would be defined and regulated as a bank even though, in practice, it refrained from issuing transaction accounts. An institution that was not eligible to issue transaction accounts would not be a bank and would not be eligible for deposit insurance, access to the Fed, and so on. By this definition, existing commercial banks, thrifts, and credit unions would be considered "banks." Similarly most of the "non bank" banks formed in recent years under the Bank Holding Company Act (by not engaging in commercial lending) would be banks, as would, depending on 9 state laws, some "industrial" banks. Treating thrifts and certain other institutions as "banks" raises a host of difficult and politically charged issues relating to regulatory treatment, tax status, divestiture, and grandfathering arrangements. However, for purposes of this discussion, the fact that certain "nonbank" financial institutions are, for a variety of reasons, banks does not require immediate or perhaps even parallel regulation. Rather, the suggestion would be that there is an essential core of regulation that should apply more or less equally to this broader class of institutions which provides essential banking functions. The issue of whether money market mutual funds fit the definition of a bank—even at a conceptual level—is not so easy to deal with. Many such funds certainly appear to have all the characteristics of bank transaction accounts. In the case of the money market mutual fund, the critical distinction relative to a bank transaction account appears to be the extent to which the liabilities in question are payable at par. In the case of a bank deposit, deposit insurance, the capital of the bank, and the banks' access to alternative sources of short-run funding provide assurances that a depositor can withdraw dollar-for-dollar from the bank the principal amount deposited—even when changes in interest rates may have reduced the market value of bank assets. In the case of the money market mutual fund the ability to pay out dollar-for-dollar the amount of the initial "deposit" is less certain. The fund itself does not have capital as such, and in the short-run it cannot easily tap alternative sources of liquidity to pay out to some shareholders thereby buying time for assets to mature or for interest rates to reverse course. As a related matter, the fund is not insured so that even though the risk of loss to the individual shareholder is small, it does exist. The fact that in recent months a number of money market mutual funds have taken steps in the direction of securing some form of private insurance would suggest that some fund managers perceive that there is an important distinction to be drawn between the fund shares and bank deposits. The irony of this, of course, is that to the extent funds obtain insurance, they come even closer to possessing bank-like characteristics. From a competitive viewpoint, the question of whether a money market mutual fund is a bank is far less important today than it was before the introduction of MMDAs at banks. Indeed, if being a "bank" is equated with deposit insurance, access to the Feds' discount window and payments services—the costs of reserve requirements notwithstanding —some money funds might not object at all to being called a bank in the current market setting. Moreover, if the power of banks or bank holding companies was expanded to permit such institutions to offer mutual funds, the question, from a competitive point of view, would be even less pressing. However, in terms of intellectual consistency, the question of whether money market mutual funds (or similar arrangements which permit "check" writing) should fall within the definition of a bank does not disappear simply because current competitive conditions render the issue less compelling. On technical grounds, it would seem that the distinction arising from the payment at par principle could justify treating money funds as nonbanks. On functional grounds, however, and particularly from the perspective of the shareholder the check writing features of some funds simply may create too much of a "look alike" situation to make a meaningful distinction on the technical grounds of payment at par. It may therefore be necessary to place certain restrictions— such as limits on the number of third-party transfers (as with bank-issued MMDAs) and/or 10 reserve requirements—on "non bank" financial instruments or institutions that provide check writing features. Of course, if MMDAs were defined as transaction accounts, then the case for treating MMMFs as banks would become powerful. Bank Powers and Structure If a bank can be satisfactorily defined along the lines suggested above, there are three related questions which must be answered in order to sketch out a reasonable approach to the future scope and structure of banking activities and banks. They are: (1) What kinds of subsidiary powers should banks have? (2) What restraints, if any, should be placed on the ownership or control of banks? (3) Is it important, from a public policy perspective, whether the subsidiary activities of banks are performed in the bank, a subsidiary of the bank, or in a subsidiary of a bank holding company? The answers to each of these questions must be guided by the earlier discussion of what it is that makes banks special and the relationship between the integrity of the deposit taking function, the financial strength of the bank, and ultimately the strength of the financial system. That discussion implied that in thinking about asset powers, ownership, and the organizational structure of banks, substantial weight needed to be given to safety and soundness considerations, the special trusteeship of banks and the objectivity and impartiality of the credit decision making process. This is not to suggest that other factors such as concentration and public convenience and need are not important from the perspective of public policy. Indeed, these things may be very important, but their importance— in the context of questions relating to banking powers, ownership, and structure—is secondary to the safety and soundness factors. Having said that, a case can be made that whatever weight safety and soundness and related criteria have been given in the past, these factors should be given less weight in the future. Better information and management systems, more efficient markets, greater disclosure, improved supervision, and the presence of the public safety net, all seem to work in the direction of reducing public policy concerns about the safety and soundness of banks. However, there are strong forces working in the opposite direction. Financial affairs generally are much more complex and more interdependent than they once were. One consequence of this is that when problems arise they are more difficult to isolate and contain than in the past. Perhaps more importantly, the combination of liability management techniques and deregulation has significantly altered the overall liability structure of banks. Stable and low cost core deposits are virtually a thing of the past. These developments have,in combination with more sophisticated and interest-rate conscious corporate treasurers and individuals, increased the term structure risk at banks and made banks more susceptible to sudden deposit shifts. At the same time, "spread management"—whereby banks attempt to float the rate of return on assets in some reasonably fixed relationship to changes in the cost of funds—may, subtly but insidiously, be working to undermine the traditional disciplines of both borrowers and lenders. Finally, the far-flung international activities of banks have introduced new elements of risk into the equation. While it is a matter of judgment as to whether this cross current of events is working to reduce or to increase the risks associated with the activities of 11 banks, it does seem prudent to conclude that they are working in the direction of creating greater risks. Bank Subsidiary Powers As suggested earlier, to preserve and protect the essential functions of banks, banks must be competitively viable institutions. This means, among other things, that banks must be able to offer a sufficiently wide and competitive range of services to maintain profitability, attract capital, and preserve a de facto monopoly on the transaction account business. Without delving into the specific types of powers banks should have, the preceding discussion is suggestive of the general criteria which should be used in making judgments about the scope of banking powers. While a number of factors may be relevant in this regard, the essential functions of banks as described earlier suggest the primacy of two general criteria. They are: subsidiary banking activities should not entail excessive risk of loss and should not impair the impartiality of the credit decision making process. This dual criteria, while conceptually useful, is operationally ambiguous. To some extent, it becomes more clear in a context in which secondary criteria relating to competition/concentration considerations are introduced. Similarly, as a practical matter, defining the extent of appropriate subsidiary banking powers can be guided by policies governing bank ownership. That is, logic would seem to dictate that a particular set of powers be vested in banks only to the extent that there is a willingness to permit another institution engaging in those activities to own and/or control banks. For example, if we are willing to permit banks to engage in commerce generally (that is, the acquisition, manufacture, or distribution of goods and nonfinancial services), then we should be prepared to say that firms engaged in such business, whether oil companies or shoe stores, can own and control banks. The converse also should follow: if we are unwilling to permit banks to engage in such activities, then logic would seem to dictate that such commercial firms should not own banks. The symmetry of this argument is important, for it lends weight to the apparent consensus that the separation of banking from commerce generally is appropriate and should be maintained in both directions. However, even in the realm of so-called financial services, the risk/impartiality criteria do not provide unambiguous insights as to how far banking powers should be extended. For example, if there is a consensus that the risk/impartiality test should not preclude banking organizations from engaging in the sale and distribution of mutual funds shares or in the distribution and brokerage of securities, it is by no means clear that such a consensus would extend to activities relating to the underwriting of stocks and corporate bonds generally or to taking positions in commodities. The point is, of course, that while it is a fairly easy matter to conclude that a continued separation of banking and commerce makes sense, it is not nearly so easy to conclude—as a matter of public policy—that the full range of financial services should be fair game for banking organizations. At the very least, the risk/impartiality criteria suggested above and the bank ownership/control questions discussed below suggest that we should not be indifferent to the scope of financial services offered by banking organizations. Bank Ownership 12 If there is some agreement (1) that the segregation of essential banking functions into identifiable classes of institutions makes sense; (2) on the definition of a "bank"; and (3) on the appropriate scope of powers to be housed within banking organizations, then dealing with the question of bank ownership becomes fairly easy That is, nonbanking organizations would be permitted to own banks only insofar as the activities of such entities match the activities in which banking organizations would otherwise be permitted to engage. For example, a securities firm whose activities did not go beyond the activities directly permissible to banks and bank holding companies could own a bank, but in the process that organization would become a bank holding company. On the other hand, financial or nonfinancial firms could not own a bank unless they were willing to divest those activities which fall outside the list of permissible activities for banks and bank holding companies. Thus, depending on the determination of the scope of banking powers—which, as noted earlier, should be undertaken primarily within the context of the risk/impartiality criteria—this approach would require that a number of existing situations involving the ownership of "banks" by financial and nonfinancial firms would have to be grandfathered or, perhaps in some cases, divestiture arrangements would have to be worked out over a period of time. Banking Structure Finally in this context, questions will inevitably arise as to whether it matters, from the perspective of public policy, if particular subsidiary activities of banks are carried out in the bank, in a subsidiary of the bank, or in a subsidiary of the banks' holding company. Given the earlier discussion about the importance of segregating essential banking activities and the importance of the risk/impartiality criteria for purposes of evaluating the appropriate scope of banking activities, it would seem to follow that there is a powerful case for placing some subsidiary activities of banking organizations into affiliates of bank holding companies. This case is reinforced by the protections against self-dealing, which are made possible by certain provisions of the Bank Holding Company Act and by the de facto segregation of capital that is made possible by the holding company structure. However it does not follow from the above that we can be indifferent as to the degree of risk associated with such activities simply because they may be housed in a separately organized and separately capitalized subsidiary of a bank holding company. To the contrary, experience suggests rather clearly that in times of peril it may not be possible to insulate the bank from the problems of its sister organizations—even when such problems arise in affiliated organizations, including subsidiaries of bank holding companies. While there are good and sufficient public policy reasons for conducting that at least some "nonbank" activities of banking organizations should be housed in subsidiaries of bank holding companies, such organizational arrangements are not likely to produce a situation in which the bank is immune from the problems, risks, or losses that might develop in such subsidiaries. In short, the holding company structure is neither a substitute for prudent management nor a fail-safe device for containing risk. In Conclusion 13 This essay started out with a seemingly straightforward question: Are banks special? Having answered that question in the affirmative, it does seem appropriate that the current debate about the powers and structure of banks be framed in a context that gives greater weight to the underlying issues of what banks are, and what, from the perspective of public policy, we want them to be. Looked at in that light, and with a firmer grasp on what it is that makes banks special, it becomes somewhat easier to grapple with the very difficult questions relating to the definition of a bank, the scope of banking powers, the ownership and control of banks, and the structure of banking organizations. This approach—entailing as it does an element of going back to square one—can help to ensure that bankers, regulators, and legislators approach successive steps in the reshaping of our financial system in a manner which helps to preserve the unique functions and characteristics of banks while at the same time encouraging those elements of competition and innovation that will permit the banking system and the financial system more generally to safely and efficiently meet the needs of a growing and stable domestic and international economy. * In this essay, the term "bank" is used in a generic way that makes no effort to distinguish commercial banks from thrifts and other "depositors institutions". This is done merely to simplify the discussion however, in considering the essential functions of "banks" in light of the Depository Institutions Deregulation and Monetary Control Act of 1980 and the Garn-St Germain Depository Institutions act of 1982, it is clear that in substance there are no longer meaningful differences. To be sure differences in powers, in regulatory treatment, and in tax status remain, but the basic characteristics that distinguish banks from other classes of financial and nonfinancial enterprises now seem to apply to thrifts as well as to commercial banks. 14