PII: S0378-4266(99)00085-0 A critique on the theory of ®nancial intermediation Bert Scholtens a,*, Dick van Wensveen b,c a Department of Finance, University of Groningen, P.O. Box 800, 9700 AV Groningen, Netherlands b Erasmus University, Rotterdam, Netherlands c University of Amsterdam, Amsterdam, Netherlands Received 25 September 1998; accepted 22 April 1999 Abstract This comment discusses the review by Franklin Allen and Anthony Santomero of the theory of ®nancial intermediation in the 20th anniversary special issue of the Journal of Banking and Finance. We do not fully agree with their view that risk management is only of recent importance to the ®nancial industry and with putting central the concept of participation costs. We suggest how the theory of ®nancial intermediation might be developed further in order to understand present-day phenomena in the ®nancial ser- vices sector. Ó 2000 Elsevier Science B.V. All rights reserved. JEL classi®cation: E51; G10; G20; L23 Keywords: Banks; Financial intermediation; Financial system; Risk management; Transformation 1. Introduction In a recent paper, Allen and Santomero (1997) review the state of inter- mediation theory and attempt to reconcile it with the observed behavior of Journal of Banking & Finance 24 (2000) 1243±1251 www.elsevier.com/locate/econbase * Corresponding author. Tel.: +31-50-363-7064; fax: +31-50-363-7207. E-mail address: l.j.r.scholtens@eco.rug.nl (B. Scholtens). 0378-4266/00/$ - see front matter Ó 2000 Elsevier Science B.V. All rights reserved. PII: S 0 3 7 8 - 4 2 6 6 ( 9 9 ) 0 0 0 8 5 - 0 institutions in modern capital markets. They argue that current theory of ®- nancial intermediation too heavily focuses on the functions of ®nancial insti- tutions that are no longer crucial in mature ®nancial systems. They suggest that the emphasis on the role of intermediaries as reducing the frictions of trans- action costs and asymmetric information is too strong; while these factors may once have been central to the role of intermediaries, they are increasingly less relevant. Allen and Santomero suggest a view on ®nancial intermediaries that centers on two of their roles. First, they are the facilitators of risk transfer and deal with an increasingly complex maze of ®nancial instruments and markets. The key area of intermediary activity therefore has become risk management, whereas traditional intermediation theory o�ers little to explain why institu- tions should perform this function. Second, ®nancial intermediaries reduce participation costs ± the costs of learning about e�ectively using markets as well as participating in them on a day-to-day basis ± and this plays an im- portant role in understanding the changes that have taken place. We welcome their e�ort to bring the theory of ®nancial intermediation further. We basically agree with Allen and Santomero but we think that their analysis is incomplete. In this comment, we go into some points which, in our opinion, are missing in their theory of ®nancial intermediation. We question whether risk management is something that is undertaken only recently in the ®nancial industry, we focus on their concept of participation costs, we argue that some of their critique on the existing literature goes too far and we argue that some of their critique does not go far enough. We hope to bring the fundamental discussion about ®nancial intermediation a small step further by suggesting elements for a theory that is helpful in understanding and explaining the daily operations of ®nancial institutions and markets and their function and signi®cance in the real economy. 2. Do ®nancial intermediaries face extinction? We wholeheartedly agree with Allen and SantomeroÕs (hereafter AS) anal- ysis in terms of a functional perspective, rather than an institutional perspec- tive. Countries use di�erent de®nitions of various types of ®nancial intermediaries. Furthermore, regulations in some countries still forbid the formation of certain types of ®nancial institutions that are an essential part of the economy elsewhere. The paradigm used in the current theory of ®nancial intermediation is the famous classical idea of the perfect market, introduced by Marshall and Walras and since then the leading principle, the central point of reference in the theory of monopolistic competition and the neo-classical growth theory, the portfolio investment theory and the theory of ®nancial intermediation. This paradigm is formalized in the traditional Arrow±Debreu model of resource allocation. 1244 B. Scholtens, D. van Wensveen / Journal of Banking & Finance 24 (2000) 1243±1251 Financial intermediaries, according to that theory, have a function only be- cause ®nancial markets are not perfect. They exist by the grace of market imperfections. As long as there are market imperfections, there are interme- diaries; as soon as markets are perfect, intermediaries are redundant: they lose their function as soon as savers and investors have the perfect information to ®nd each other directly, immediately and without any impediments, so without costs. Thus, in a world with a tendency towards greater market transparency and e�ciency, ®nancial intermediaries are an endangered species. However, despite globalization, the information revolution and a much more prominent role of public markets, ®nancial intermediaries appear to survive. A tendency towards a relative reduction of certain activities of some ®nancial intermediaries is going on, most clearly in the US. For example, Fig. 1 indicates that the relative size of depository institutions (commercial banks, savings institutions, credit unions) in the ®nancial system diminishes. The same applies to insurance companies. In contrast, the relative size of pension funds (both private and government funds) as well as that of mutual funds (including money market funds) increases considerably. The overall size of the assets held by US ®nancial institutions in relation to gross domestic product is depicted in Fig. 2 (left-hand scale). It rose from 100% in 1950 to 120% in 1960, it remained 120% in 1970 but rose to 210% in 1980. In the eighties, it became somewhat smaller, namely 200% in 1990. However, in the nineties, the increase continued, and in 1995 the combined assets of the ®nancial intermediaries in the US stood at 250% of GDP. 1 Banks, both uni- versal and investment banks, are under pressure, not only from competition amongst themselves but increasingly also from other intermediaries such as life Fig. 1. Distribution of US ®nancial assets by the main types of ®nancial intermediaries (Source: Barth et al., 1997). 1 Based on Barth et al. (1997) and the IMFÕs international ®nancial statistics. B. Scholtens, D. van Wensveen / Journal of Banking & Finance 24 (2000) 1243±1251 1245 insurance companies, investment funds, leasing companies and other ®nance companies, such as specialized daughters of big industrial or trade companies, merger and acquisition specialists and advice departments of auditor or lawyer ®rms. And large companies are entering the money market directly instead of dealing with a bank. The growth of securitized assets re¯ects the upcoming role of stock exchanges at the expense of traditional banking. But it is a miscon- ception to interpret the relative declining role of banks in Fig. 1 as a general process of disintermediation. Fig. 2 (right-hand scale) shows that the ratio of depository institutionsÕ assets to GDP was 0.53 in 1950. Thereafter, it fell to 0.45 in 1960 and it was 0.46 in 1970. The 1970s saw a huge increase in the growth of bank assets and in 1980 the ratio had become 0.73. The savings and loans crisis of the 1980s was very much responsible for a decline to 0.56 in 1990. In the 1990s we witnessed a renewed increase in the ratio of bank assets to GDP as it was 0.63 in 1995. Thus, the relative size of depository institutions in the US in the mid-1990s was more than one third larger than that in the mid- 1960s. Note that Fig. 2 does not take into account the rise in o�-balance sheet items that took place in the banking industry in the 1980s and 1990s. From Fig. 2, we conclude that ®nancial intermediaries are not fading away. Neither is the banking industry. Only an institutional perspective might conclude (on the basis of Fig. 1) to disintermediation of the banking industry. The functional Fig. 2. Relative size of the US ®nancial sector and the banking industry (Source: Barth et al., 1997). 1246 B. Scholtens, D. van Wensveen / Journal of Banking & Finance 24 (2000) 1243±1251 view, however, reveals that ®nancial intermediaries are of increasing impor- tance to the modern economy. Why is it that the relative importance of ®nancial intermediaries increases? Financial intermediaries perform gradually more sophisticated functions in the modern ± more and more complex ± economy. Despite the ongoing perfection, indicating a declining price of information, asymmetric information and transaction costs seem to be still important elements in intermediation pro- cesses. This suggests that there is something extra that is relevant for ®nancial intermediation. But what is that ÔextraÕ? We agree with AS that risk manage- ment has become a prominent function of ®nancial intermediation. However, in our opinion, it is not the only factor that can be held responsible for the seemingly steady rise of the ®nancial industry within the modern economy. In the next section, we evaluate ASÕs suggestion of risk management as responsible for modern ®nancial intermediation. In Section 4, we put forward ideas that may complement the justi®cation of the growth of the ®nancial in- dustry as given by AS. 3. Risk management and participation costs In our opinion, AS are right in suggesting a central role of risk in the in- termediation process and proposing that risk management become the main item in the research agenda. Risk analysis is, since the emergence of the modern portfolio theory, fully incorporated on the micro level in pricing models and plays the central role in the research on securities and derivatives. But the risk/ reward relation has not yet been analyzed on an industry level or on a macro- economic level. What is the remuneration of the industry for its risk trans- forming activity? How is the relation of this part of its revenues to total revenues? Is the remuneration adequate in view of risk losses (bad debtors, interest rate ¯uctuations, stock price movements, mortality)? Is risk well paid for at the industry level? Furthermore, is risk management really a new phenomenon? Something that emerged in the 1960s or 1970s as AS argue? We very much doubt it. Risk management has not become important only in the recent past. In contrast, we see risk as the root of ®nancial intermediation and as its main raison dÕêtre. The origins of banking and insurance lie in their risk transforming and risk man- aging functions. The merchant bankers in the Italian Renaissance already managed the ®nancial risks not only of kings and popes but also of merchants. Insurers took over the risks of merchants sending their goods overseas. In the Dutch Republic in the eighteenth century, true investment bankers emerged who not passively managed assets but actively assumed underwriting risk themselves. And even the seemingly dull business of savings and loans asso- ciations and credit unions in the US in the 1950s and 1960s is risk management, B. Scholtens, D. van Wensveen / Journal of Banking & Finance 24 (2000) 1243±1251 1247 as they manage interest rate risk, credit risk, and liquidity risk. Therefore, dealing with risk is ± and always has been ± the bread and butter of ®nancial intermediaries. By specializing in information production and processing, and by diversifying individual credit and term risks, they have been able to absorb risk. AS associate the risk management function of ®nancial intermediaries with the growing importance of new ®nancial instruments and markets (fu- tures, options, swaps) and characterize their role as facilitators of risk transfer as a new role (p. 1462). As such, AS appear to overlook the traditional role of banks in the process of risk transfer, taking deposits from savers and extending credit to borrowers with risky business. We agree with AS that intermediation theory has o�ered little to explain why institutions should perform this func- tion (pp. 1462, 1465), the more so because risk management, contrary to what AS say, has always been Ôa key area of intermediary activityÕ. Informational asymmetries and transaction costs do not fully explain why savers deposit money in banks and do not select investors themselves. A second key concept in ASÕs theory of ®nancial intermediation is Ôpartic- ipation costÕ. This is the cost of participation in a ®nancial market. AS argue that participation costs are crucial to understanding the current activities of intermediaries and in particular to their focus on risk management. Trading costs have fallen dramatically in the 1980s and mid-1990s which would have encouraged the direct participation of households and reduced the role of mutual funds. However, AS argue, the value of peopleÕs time has increased signi®cantly in the last 15 years (p. 1482), which promoted the role of mutual funds, whose participation costs are low and thus are an e�cient means to invest for individuals whose costs of direct participation have risen. Further- more, there has been a spreading of the income distribution and a resulting increase in the value of time at the high end of this distribution. This argument is rather weak as we ®nd that the growth of wages was much higher in the 1970s than in the 1980s and 1990s. 2 And not only has the size of mutual funds risen rapidly (see Fig. 1), but also the direct participation of the public in the stock market has increased substantially. In all, ASÕs explanation on the basis of lower participation costs is not very convincing. AS also point to the fact that the use of derivatives by ®nancial institutions for risk management has increased enormously. AS argue that the concept of participation costs sug- gests that intermediaries create products with a relatively stable distribution of returns. However, this also happened before the widespread and intense use of 2 For example, the average increase in the unit labor costs in the business sector in the US in the 1970s was 7.0% per year. In the 1980s it fell to 4.0% per year, and for the years 1991±1997 it averaged 2.5% per year. Furthermore, the compensation per employee in the business sector stood at 7.9% in the 1970s. In the 1980s, it was 5.1%. For 1991±1997 it was 3.4% per year (OECD, Economic Outlook 63, 1998, 236±237). 1248 B. Scholtens, D. van Wensveen / Journal of Banking & Finance 24 (2000) 1243±1251 ®nancial derivatives, as it is an important rationale for savers to put their money into a bank (see Bryant, 1980). Therefore, we conclude that although participation cost can be relevant in understanding new roles of the ®nancial intermediary, it does not seem to be able to explain drastic changes in the ®- nancial industry such as the dramatic rise of mutual funds and the widespread use of ®nancial derivatives. The key word for their growth is again ÔriskÕ. AS discuss the current rationales for the interest in risk management that is evident in the market. They can be divided into four cases: (1) managerial self- interest, (2) the non-linearity of taxes, (3) the costs of ®nancial distress, and (4) the existence of capital market imperfections. The plausibility of these four explanations varies especially if not only the bene®ts of hedging but also the costs of risk management are taken into account, as AS rightly state. It should be clear that the bankruptcy costs as discussed by AS should be viewed as the total economic costs of the ®nancial distress and/or bankruptcy. We think that the most important rationale of ®nancial risk management is the prevention of bankruptcy of a company induced by monetary and ®nancial factors. Financial risk management aims to protect the companyÕs balance sheet against severe losses of a monetary nature (e.g. exchange rate shocks) and the companyÕs operational cash ¯ow against serious ®nancial uncertainties (interest rate and exchange rate ¯uctuations, credit risk). 4. How further? How must we move further in constructing a theory of ®nancial interme- diation that can understand and explain the day-to-day operations of ®nancial institutions and markets and their role within the real economy? It will be obvious that ÔriskÕ will need to get a central place in that theory. But there is more. In order to place risk transfer as an entrepreneurial activity in its proper context, the theory of ®nancial intermediation should move beyond its present borders. It should leave its paradigm of static perfect markets and assume a more dynamic concept in which new markets are developed for new products, where ®nancial institutions do not act as ÔagentsÕ who intermediate between savers and investors and thus alleviate Ômarket imperfectionsÕ like asymmetric information and participation costs, but are independent market parties that create ®nancial products and whose value added to their clients is the trans- formation of ®nancial risk, term, scale, location, and liquidity. Table 1 gives in key words an overview of how the stylized present theory might be amended to come nearer to this goal. Stylized present theory, as already clearly summarized by AS, has the perfect market as its benchmark. Financial intermediaries emerge to make do with the market imperfections that mainly stem from informational asymmetries. They may reduce the information and transaction costs within the economy, but B. Scholtens, D. van Wensveen / Journal of Banking & Finance 24 (2000) 1243±1251 1249 they still have to make do with agency problems and with moral hazard and adverse selection. In all, the ®nancial intermediary is a more or less passive agent who intermediates between ultimate savers and investors. The process of disintermediation threatens the agent, as public ®nancial markets promote a more e�cient and transparent handling of the allocation of scarce resources in the economy, thanks to deregulation and information technology. In line with AS, we think the traditional approach should be amended. The right-hand column of Table 1 shows the main elements that the theory of ®- nancial intermediation, in our opinion, should include to explain the present- day ®nancial industry. Instead of the static case of the perfect market that is hampered by incidental imperfections, the theory of ®nancial intermediation needs to have the dynamic process of ®nancial innovation and market di�er- entiation at its basis. We need to explain changes instead of describing com- parative statics. Furthermore, as in our vision, the ®nancial intermediary provides consumer and business households with a variety of services that ful®ll their di�erent needs, the ®nancial intermediary is involved in a complex process of ®nancial transformation. In the course of qualitative asset trans- formation ± with respect to maturity, liquidity, risk, scale, and location ± it adds value for ultimate savers and investors. This active role contrasts sharply with the passive intermediating of savings to investments within the economy, a thought that prevails in the traditional theory of ®nancial intermediation. Value-addition appears to be a major drive of the modern ®nancial interme- diary. Therefore, value-addition should be the focus of intermediation theory. This may be accomplished through participation cost reduction as AS suggest, and/or through an expansion of the set of services in the ®nancial sector. This means that though asymmetric information is relevant in understanding certain policies of the ®nancial intermediary, customer orientation is his general de- vice. His business is selling ®nancial services to customers and making a pro®t Table 1 Stylized and amended theory of ®nancial intermediation Stylized (present) theory Amended theory Static: perfect market Dynamic: market di�erentiation Market imperfections Product innovation Financial intermediary is an agent between savers and investors Financial intermediary is an entrepreneurial provider of ®nancial services E�cient allocation of savings Financial transformation Costs Value Asymmetric information Customer orientation, both to borrowers and savers Adverse selection, moral hazard, credit rationing Risk/reward optimization and risk management Disintermediation Dynamics of intermediation (new markets, new products) 1250 B. Scholtens, D. van Wensveen / Journal of Banking & Finance 24 (2000) 1243±1251 on it. Reducing costs and informational asymmetries may be part of this process, but it occurs as a by-e�ect. We suggest that the analysis of policies of optimization of risks and rewards by ®nancial intermediaries helps to better understand the essence of its business than does paying attention to the po- tential e�ects of asymmetric information, i.e. adverse selection, moral hazard, credit rationing. As an illustration, the modern ®nancial intermediation theory would see the rise of the mutual fund industry not as disintermediation, but as the involvement of a new type of ®nancial intermediary in providing ®nancial transformation services. Of course, we cannot present here a complete modern theory of ®nancial intermediation yet. The aim of our critique was only to illuminate some issues not dealt with by AS and some suggestions on how to proceed further. We hope the discussion started by AS will result in a better understanding of the ®nancial world. Acknowledgements The authors wish to thank one anonymous referee for helpful comments and suggestions on a previous version of this paper. References Allen, F., Santomero, A.M., 1997. The theory of ®nancial intermediation. Journal of Banking and Finance 21, 1461±1485. Barth, J.R., Nolle, D.E., Rice, T.N., 1997. Commercial banking structure, regulation, and performance: An international comparison. Working Paper 97-6, March, Comptroller of the Currency Economic, Washington, DC. Bryant, J., 1980. A model of reserves, bank runs, and deposit insurance. Journal of Banking and Finance 4, 335±344. B. Scholtens, D. van Wensveen / Journal of Banking & Finance 24 (2000) 1243±1251 1251