«Journal of Banking & Finance 21 (1998) 1461±1485 The theory of ®nancial intermediation Franklin Allen, Anthony M. Santomero * The Wharton School, ...»
Journal of Banking & Finance 21 (1998) 1461±1485
The theory of ®nancial intermediation
Franklin Allen, Anthony M. Santomero *
The Wharton School, University of Pennsylvania, Philadelphia, PA 19096, USA
Traditional theories of intermediation are based on transaction costs and asymmetric
information. They are designed to account for institutions which take deposits or issue
insurance policies and channel funds to ®rms. However, in recent decades there have been signi®cant changes. Although transaction costs and asymmetric information have declined, intermediation has increased. New markets for ®nancial futures and options are mainly markets for intermediaries rather than individuals or ®rms. These changes are dicult to reconcile with the traditional theories. We discuss the role of intermediation in this new context stressing risk trading and participation costs. Ó 1998 Elsevier Science B.V. All rights reserved.
JEL classi®cation: G2; G1; E5; L2 Keywords: Intermediation; Risk management; Delegated monitoring; Banks; Participation costs
1. Introduction In this paper we review the state of intermediation theory and attempt to reconcile it with the observed behavior of institutions in modern capital markets. We argue that many current theories of intermediation are too heavily focused on functions of institutions that are no longer crucial in many develCorresponding author. Tel.: +1 215 898-7625; fax: +1 215 573-8757; e-mail: santo@®nance.wharton.upenn.edu.
0378-4266/97/$17.00 Ó 1997 Elsevier Science B.V. All rights reserved.
PII S 0 3 7 8 - 4 2 6 6 ( 9 7 ) 0 0 0 3 2 - 0 1462 F. Allen, A.M. Santomero / Journal of Banking & Finance 21 (1998) 1461±1485 oped ®nancial systems. They focus on products and services that are of decreasing importance to the intermediaries, while they are unable to account for those activities which have become the central focus of many institutions.
In short, we suggest that the literature's emphasis on the role of intermediaries as reducing the frictions of transaction costs and asymmetric information is too strong. The evidence we oer suggests that while these factors may once have been central to the role of intermediaries, they are increasingly less relevant.
We oer in its place a view of intermediaries that centers on two dierent roles that these ®rms currently play. They are facilitators of risk transfer and deal with the increasingly complex maze of ®nancial instruments and markets.
Risk management has become a key area of intermediary activity, though intermediation theory has oered little to explain why institutions should perform this function. In addition, we argue that the facilitation of participation in the sector is an important service provided by these ®rms. We suggest that reducingparticipation costs, which are the costs of learning about eectively using markets as well as participating in them on a day to day basis, play an important role in understanding the changes that have taken place.
The paper proceeds as follows. In Section 2, we oer a review and critique of the usual views of intermediation found in the literature. This critique is supported by data presented in Section 3, which outlines the changes in ®nancial systems that have occurred over the recent past. In Section 4 the current justi®cations for one of the growth areas of intermediary services, namely risk management, are presented, while Section 5 discusses the risk reduction activities that intermediaries should take. Section 6 then outlines the importance of participation costs as another rationale for intermediation and assisting in risk management. Finally, Section 7 contains concluding remarks.
2. Review and critique of current intermediation theory
In the traditional Arrow±Debreu model of resource allocation, ®rms and households interact through markets and ®nancial intermediaries play no role.
When markets are perfect and complete, the allocation of resources is Pareto ecient and there is no scope for intermediaries to improve welfare. Moreover, the Modigliani±Miller theorem applied in this context asserts that ®nancial structure does not matter: households can construct portfolios which oset any position taken by an intermediary and intermediation cannot create value (see Fama, 1980).
A traditional criticism of this standard market-based theory is that a large number of securities are needed for it to hold except in special cases. However, the development of continuous time techniques for option pricing models and the extension of these ideas to general equilibrium theory have negated this F. Allen, A.M. Santomero / Journal of Banking & Finance 21 (1998) 1461±1485 1463 criticism. Dynamic trading strategies allow markets to be eectively complete even though a limited number of securities exist.
Such an extreme view ± that ®nancial markets allow an ecient allocation and intermediaries have no role to play ± is clearly at odds with what is observed in practice. Historically, banks and insurance companies have played a central role. This appears to be true in virtually all economies except emerging economies which are at a very early stage. Even here, however, the development of intermediaries tends to lead the development of ®nancial markets themselves (see McKinnon, 1973).
In short, banks have existed since ancient times, taking deposits from households and making loans to economic agents requiring capital. Insurance, and in particular marine insurance, also has a very long history. In contrast, ®nancial markets have only been important recently, and then only in a few countries, primarily the UK and the US. Even there, banks and insurance companies have played a major role in the transformation of savings from the household sector into investments in real assets.
Our understanding of the role or roles played by these intermediaries in the ®nancial sector is found in the many and varied models in the area known as intermediation theory. These theories of intermediation have been built on the models of resource allocation based on perfect and complete markets by suggesting that it is frictions such as transaction costs and asymmetric information that are important in understanding intermediation. Gurley and Shaw (1960) and many subsequent authors have stressed the role of transaction costs.
For example, ®xed costs of asset evaluation mean that intermediaries have an advantage over individuals because they allow such costs to be shared. Similarly, trading costs mean that intermediaries can more easily be diversi®ed than individuals.
Looking for frictions that relate more to investors' information sets, numerous authors have stressed the role of asymmetric information as an alternative rationalization for the importance of intermediaries. One of the earliest and most cited papers, Leland and Pyle (1977), suggests that an intermediary can signal its informed status by investing its wealth in assets about which it has special knowledge. In another important paper, Diamond (1984) has argued that intermediaries overcome asymmetric information problems by acting as ``delegated monitors.'' Many others followed, expanding on these two contributions and advancing the literature in substantive ways (e.g., see Gale and Hellwig, 1985; Campbell and Kracaw, 1980; Boyd and Prescott, 1986).
Bhattacharya and Thakor (1993) have provided an excellent survey of the current state of the literature on banking, building on an earlier review of the banking literature published in Santomero (1984). Dionne (1991) contains a set of surveys of the literature on insurance. Readers wishing detailed accounts of particular literatures should consult these papers. Our contribution here will not be a duplication of these eorts. Rather, it will attempt to contrast 1464 F. Allen, A.M. Santomero / Journal of Banking & Finance 21 (1998) 1461±1485 the traditional view of the role and functions performed by intermediaries with the evolution of these institutions over the last few decades. It is an attempt to confront the literature with a view of the practice to see if the literature adequately addresses the reasons that these institutions exist in the ®nancial markets, and how they perform value added activity.
The reality is that the ®nancial systems in many countries have undergone a dramatic transformation in recent years. Financial markets such as the stock and bond markets have grown in size using nearly any metric, such as the value of companies listed or any other conceivable measure of their importance. At the same time, there has been extensive ®nancial innovation acceleration in the 1970s and 1980s. This includes the introduction of new ®nancial products, such as various mortgage backed securities and other securitized assets, as well as derivative instruments such as swaps and complex options. These have all had a virtual explosion in volume. At the same time, new exchanges for ®nancial futures, options and other derivative securities have appeared and become major markets.
Interestingly, this increase in the breadth and depth of ®nancial markets has been the result of increased use of these instruments by ®nancial intermediaries and ®rms. They have not been used by households to any signi®cant extent. In fact, the increased size of the ®nancial market has coincided with a dramatic shift away from direct participation by individuals in ®nancial markets towards participation through various kinds of intermediaries.
The importance of dierent types of intermediary over this same time period has also undergone a signi®cant change. The share of assets held by banks and insurance companies has fallen, while mutual funds and pension funds have dramatically increased in size. New types of intermediary such as non-bank ®nancial ®rms like GE Capital have emerged which raise money entirely by issuing securities and not at all by taking deposits. In short, traditional intermediaries have declined in importance even as the sector itself has been expanding.
Perhaps in response, but clearly contemporaneously, the activities of traditional institutions such as banks and insurance companies have also changed.
Banks which used to take deposits and make loans found that the possibilities for securitizing loans meant that they did not need to keep on their balance sheet all the loans they could originate. At the same time, insurance ®rms realized that their actuarial function was but a minor part of their asset management capabilities and these ®rms too innovated and broadened their products and services.
Some of these changes in the volume of ®nancial activity, along with the relative importance of some institutions and the changes in others, can be explained using traditional theories which are based on transaction costs and asymmetric information. But, others cannot. For example, the standard explanation for the existence of mutual funds is that, while diversi®cation is desirF. Allen, A.M. Santomero / Journal of Banking & Finance 21 (1998) 1461±1485 1465 able, the high costs of trading for individuals makes it expensive for individuals to achieve this. Mutual funds can trade at signi®cantly lower cost and so can achieve diversi®cation much more cheaply. Given this explanation it might be expected that if individuals' trading costs were lowered the share of ownership of mutual funds would fall. However, we have not observed this behavior.
Although with the introduction of competition for brokerage fees on the NYSE in the early 1970s trading costs for individuals fell dramatically, the share of assets invested in mutual funds has risen signi®cantly. Likewise, the advent of the technological revolution has substantially reduced the cost of information and reduced information asymmetry. Yet it did not reduce the need for intermediary services and encourage direct lending by households. In fact, the data suggest the opposite. In short, the decline in frictions which were allegedly the market imperfections that led to a need for intermediation services has not reduced the demand for them. Intermediation is growing and prospering even as the frictions decline.
The contrast between theory and reality is perhaps most apparent in the area of risk management. Arguably the most important change in intermediaries' activities that has occurred in the last thirty years is the growth in the importance of risk management activities undertaken by ®nancial intermediaries. As we noted above, the change in the breadth of the markets that are available for hedging risk has not led very many individual or corporate customers to manage their own risk. Rather, it has meant that risk management has now become a central activity of many intermediaries. Most current theories of intermediation have little to say about why risk management should play such an important role in the activities of intermediaries.
In some cases, theories explaining why both ®nancial and non-®nancial ®rms should undertake risk management have been added on to our understanding of ®rm level decision making. However, these descriptions of why they undertake hedging activities are almost an afterthought in the literature.
Little is oered as a cogent argument as to why intermediaries should be the ones oering these services, and what value they bring to the activity. In short, the intermediation literature is noticeably quiet as to why these institutions should be engaged in one of their central areas of activity.