«What do ®nancial intermediaries do? a,* b Franklin Allen, Anthony M. Santomero a The Wharton School, University of Pennsylvania, Room 2336, ...»
Journal of Banking & Finance 25 (2001) 271±294
What do ®nancial intermediaries do?
Franklin Allen, Anthony M. Santomero
The Wharton School, University of Pennsylvania, Room 2336, Steinberg Hall-Dietrich Hall,
Philadelphia, PA 19104-6367, USA
Federal Reserve Bank of Philadelphia, Philadelphia, PA 19106, USA
Received 29 July 1999; accepted 13 September 1999
This paper presents evidence that the traditional banking business of accepting de- posits and making loans has declined signi®cantly in the US in recent years. There has been a switch from directly held assets to pension funds and mutual funds. However, banks have maintained their position relative to GDP by innovating and switching from their traditional business to fee-producing activities. A comparison of investor portfo- lios across countries shows that households in the US and UK bear considerably more risk from their investments than counterparts in Japan, France and Germany. It is argued that in these latter countries intermediaries can manage risk by holding liquid reserves and intertemporally smoothing. However, in the US and UK competition from ®nancial markets prevents this and risk management must be accomplished using de- rivatives and other similar techniques. The decline in the traditional banking business and the ®nancial innovation undertaken by banks in the US is interpreted as a response to the competition from markets and the decline of intertemporal smoothing. Ó 2001 Elsevier Science B.V. All rights reserved.
JEL classi®cation: G2; G1; E5; L2 Keywords: Intermediation; Risk management; Delegated monitoring; Banks; Partici- pation costs * Corresponding author. Tel.: +1-215-898-3629; fax: 1-215-573-2207.
E-mail address: firstname.lastname@example.org (F. Allen).
0378-4266/01/$ - see front matter Ó 2001 Elsevier Science B.V. All rights reserved.
PII: S 0 3 7 8 - 4 2 6 6 ( 9 9 ) 0 0 1 2 9 - 6 272 F. Allen, A.M. Santomero / Journal of Banking & Finance 25 (2001) 271±294
1. Introduction Traditionally, transaction costs and asymmetric information have provided the foundation for understanding intermediaries. The emergence of interme- diaries resulting from such imperfections in the capital market has been for- malized in the contributions of Dewatripont and Tirole (1994) and Freixas and Rochet (1997). In fact, the two major reviews of intermediation theory, Santomero (1984) and Bhattacharya and Thakor (1993) illustrate how central such imperfections are to the intermediation literature of the past two decades.
The apparent implication of this view is that, if these frictions are reduced, intermediaries will become less important.
There has been a signi®cant reduction in transaction costs and asymmetric information in recent decades. Over this same period, the importance of traditional banks that take deposits and make loans has, by some measures, been reduced. However, other forms of intermediaries such as pension funds and mutual funds have grown signi®cantly. In addition, new ®nancial markets such as ®nancial futures and options have developed, as markets for intermediaries rather than for individuals. All of this seems, if not contrary to standard theory, at least inconsistent with it.
In Allen and Santomero (1997) (henceforth AS) we suggest that this is because the recent focus of intermediation theory has been too narrow. We argue that understanding these changes requires dierent theories of intermediation that stress risk trading, risk management and participation costs as the key reasons for the existence of modern intermediaries.
Our contribution has caused quite a ®restorm, with a signi®cant number of citations, discussions and debates. Recently a paper by Scholtens and van Wensveen (1999) (SW) raised a number of important issues concerning our analysis and we will take this opportunity to more fully develop our perspective on the evolution of intermediation and the theory that purports to explain it.
We choose to use this as a point of departure because their contribution raises three key issues that warrant further discussion, which will add to our understanding of the evolution of speci®c institutions within the ®nancial market.
First, it is pointed out that while it is true that the share of assets of traditional commercial banks has shrunk relative to other intermediaries in the US, it is also the case that relative to GDP banksÕ assets have increased.
Clearly, banking is not disappearing. Second, they question whether risk management is a new phenomenon. They mirror a view that argues that banks have always been in the risk management business, suggesting that the origins of banking and insurance lie in their risk transforming and management functions. Although the precise way in which risk is managed may have changed, intermediaries have always been engaged in risk management, broadly de®ned. Third, they suggest that the theory of ®nancial intermediation needs to have an understanding of the dynamic process of ®nancial innovation F. Allen, A.M. Santomero / Journal of Banking & Finance 25 (2001) 271±294 273 to adequately address the transformation of the ®nancial sector that is currently taking place globally.
In this paper we use these observations as a starting point for considering what it is that ®nancial intermediaries do. At the center, of course, ®nancial systems perform the function of reallocating the resources of economic units with surplus funds (savers) to economic units with funding needs (borrowers).
Our interest, however, is how and why this role varies across time and across countries. We will approach this question by comparing the role of intermediaries and their relationship to markets at dierent dates in a set of relevant countries, including the US, UK, France, Germany and Japan.
We will also contrast and compare our results to those in an important study by Schmidt et al. (1999). In this work, the authors have considered the issue of whether banks have been losing importance relative to markets in France, Germany and the UK. They ®nd that there is neither a general trend from intermediaries to markets, nor a decline in the importance of banks in Germany or the UK. In France, there are signs of a decline in banks and a move towards markets. However, and importantly, there is a change in the length of intermediation chains in all three systems.
We develop our argument as follows. Section 2 considers a range of appropriate measures for how the ®nancial system has changed through time in the US. Most of these suggest that the role of banks in their traditional businesses has declined. However, banks have been entrepreneurial and in many cases have changed their activities so that, by some measures, they have maintained their position.
In Section 3 we consider whether the way in which intermediaries manage risk diers across dierent countries by looking at the portfolios of assets ultimately held by households in each country. In the US and UK, households hold a large proportion of risky assets, such as equity, and bear considerable risk. However, in Japan, France and Germany, they mostly hold their assets in safe instruments, such as bank accounts and ®xed-income securities, and therefore hold very few risky assets. How can the risk borne by households dier so much across countries?
Using the framework of Allen and Gale (1997, 1999a) we suggest that there is a fundamental dierence in the way that intermediaries in the two sets of countries manage risk. In Japan, France and Germany competition from ®nancial markets has been muted, and as a result intermediaries are able to manage risk through intertemporal smoothing. This involves building up reserves in safe low yielding assets when returns are high. When returns are low the intermediaries can draw on these reserves and shield their customers from risk. However, when there is signi®cant competition from ®nancial markets, as in the US and UK in recent years, intertemporal smoothing is no longer possible. Households will withdraw their funds from banks that try to build up reserves and use them to buy other assets that allow the households to consume 274 F. Allen, A.M. Santomero / Journal of Banking & Finance 25 (2001) 271±294 all the current payos. In this case it will not be possible for banks to manage risk through the accumulation and depletion of reserves. Instead, they must use derivatives and other similar techniques for managing risk.
Section 4 considers how the transformation of the US banking system documented in Section 2 can be understood using the framework developed in Section 3. We suggest that, before the surge in ®nancial innovation in the last 25 years in the US, intermediaries such as banks were able to manage risk through intertemporal smoothing. However, as ®nancial markets became increasingly more competitive, banks were forced to move away from using intertemporal smoothing as their main risk management technique. Instead, they were forced to develop alternative ways of managing risk that did not involve holding large amounts of funds in low yielding safe assets. These techniques involved using instruments such as options and swaps and dynamic trading strategies. As traditional banking businesses began to dry up, the management of those institutions was forced to become entrepreneurial and develop new businesses in order to survive. Some of these new businesses involved catering to the demand for new risk management services associated with the growing need of their customers to deal with the risks previously absorbed by the intermediaries themselves. In short, the evolution of intermediaries is part and parcel of the evolution and competition that is accelerating in the markets they serve.
Finally, Section 5 contains a summary of our view and raises some puzzles that remain unanswered in our attempt to understand dierences in ®nancial systems. Many of these appear to be measurement issues, but they mask substantive issues and gaps in our knowledge of this important area of inquiry.
2. Measuring dierences in ®nancial systems
As is widely acknowledged US depository institutions have had a falling share of the ®nancial assets of intermediaries for many years. Scholtens and van Wensveen (1999) use Fig. 1 to illustrate this point. Although this declining share is often assumed to be a recent phenomenon, in fact the trend has been apparent since the 1920s, as shown in Fig. 2. Indeed, the 1920s was an era much like the last two decades, in which the share of assets held by banks declined and proportion of ®nancial assets held in the form of pension funds, trusts and investment companies grew. In the broader historical context the anomaly may have been the relative stability of the bank share of total assets from the 1940s through to the mid-1970s.
However, Scholtens and van Wensveen (1999) also observe that relative to total ®nancial assets as a percentage of GDP the assets in the banking sector have not declined markedly. The diagram they use to illustrate this point is given here as Fig. 3.
F. Allen, A.M. Santomero / Journal of Banking & Finance 25 (2001) 271±294 275
Fig. 1. Distribution of US ®nancial assets by the main types of ®nancial intermediaries (Source:
Barth et al., 1997, and updated tables from Barth).
Fig. 2. Relative shares of total ®nancial intermediary assets, 1900±1995:IV (Source: Federal Reserve Bulletin).
Fig. 3. Relative size of the US ®nancial sector and the banking industry (Source: Barth et al., 1997, and updated tables from Barth).
276 F. Allen, A.M. Santomero / Journal of Banking & Finance 25 (2001) 271±294 Fig. 4. The decline in the role of banks as intermediators of credit risk (Source: Board of Governor of the Federal Reserve System, ``Flow of Funds Accounts'', 1998).
How can these data be reconciled with the data shown in Fig. 1, where it is clear that banks are losing ground relative to other intermediaries? The answer lies in recognizing that ®nancial assets can be divided into three mutually exclusive subsets. The ®rst is those held by banks or more properly depository institutions; the second is those held by nonbank intermediaries such as pension funds and mutual funds; and the third is those that are directly held assets, such as stocks and bonds. If banks are shrinking relative to other intermediaries, but are stable relative to total ®nancial assets then this implies that there is a switch from directly held assets to nonbank intermediaries. This is consistent with the long-term trend of the decline in the individual ownership of corporate equity which was documented in AS.
The evidence presented in Figs. 1±3 is not the only way to measure how the ®nancial system is changing. Other changes have taken place in the US banking system which are often associated with the widespread view that banks are becoming less important. The decline in the role of banks as intermediators of credit risk has been most pronounced in a US context with regard to business ®nance as Fig. 4 indicates. Banks have lost ground to other intermediaries such as ®nance companies and to securities markets, especially the commercial paper and high yield securities market.
The decline in business lending is also mirrored in consumer lending as shown in Fig. 5. Banks have lost market share to nonbanks such as AT&T, GMAC, GE and Morgan Stanley Dean Witter. Twenty years ago, banks completely dominated the entire credit card business. Now, banks hold no more than 25% of receivables, and close to 80% of credit card transactions are processed by nonbanks such as First Data Resources. 1
Fig. 5. Bank market share of credit card receivables, 1986±1998 (Source: Faulkner and GrayÕs Card Industry Directory, various years).
Increasingly, single-purpose providers have successfully competed for some of the most pro®table traditional bank products provided by full service banking institutions. In addition, the development of securitization techniques has transformed the way in which many kinds of credit transactions ± which would previously have been conventional bank loans ± are structured.