Financial_Innovation

Frame W S and White L J (2002) “Empirical Studies of Financial Innovation: Lots of talk, little action “WP 2002-12 Federal Reserve bank of Atlanta DIFFUSION We are aware of five studies of the diffusion of financial innovations, three of which focus on ATM deployment by banks. These studies generally use hazard models that estimate the adoption pattern of the innovation under study conditional on firm- and market-specific effects. Hannan and McDowell (1984) find that -- consistent with the Schumpeterian hypotheses -- larger banks and those operating in more concentrated local banking markets registered a higher conditional probability of ATM adoption. This study also found bank product mix, bank holding company affiliation, urban location, branch banking restrictions, and the area wage rate were all positively related to ATM adoption. In a subsequent study, Hannan and McDowell (1987) find that the conditional probability of ATM adoption is positively related to a rival's adoption and that firms in less concentrated markets react more strongly to rival precedence than do their counterparts in concentrated markets. Consistent with their previous results, bank size and local market concentration were positively related to ATM adoption. Similar results were found for bank 19 holding company affiliation, branch banking restrictions, and market deposit growth.20 Using the same data, Saloner and Shepherd (1995) find that the expected time to adoption of ATMs declines in both the number of users (deposits) and locations (branches), indicating the presence of network externalities. For limited branching states, market concentration is positively related to ATM adoption speed, while depositor growth is negatively related. For unrestricted states, the area bank wage rate is positively related to ATM adoption speed. Molyneux and Shamroukh (1996) examine the diffusion of the underwriting of junk bonds and of note issuance facilities (NIFs) during the 1978-1988 and 1983-1986 periods, respectively. 21 The authors find that exogenous factors, such as regulatory or demand changes, played a significant role in the diffusion of junk bond underwriting. Conversely, the diffusion of NIFs underwriting appeared to be motivated by bandwagon effects. Molyneux and Shamroukh argue that banks (commercial and investment) are more likely to respond to competitive and institutional bandwagon pressures by adopting an innovation when it threatens an existing business, rather than when it represents new business opportunities. However, for both underwriting innovations, the authors find that adoption by one bank makes it more desirable for other banks to follow suit – and this effect increases in the number of adopters. More recently, Akhavein, Frame, and White (2001) examine the diffusion of small business credit scoring (SBCS) by large banking organizations in the mid-1990s. Estimates from a hazard model indicate that larger banking organizations and those located in the New York Federal Reserve district adopted this technology sooner. A tobit model confirms these 20 Interestingly, in this follow-up paper (and using the same data) the author’s did not include the product mix and urban location variables that were significant in their initial paper. 21 A note issuance facility is an arrangement by which a bank or group of banks agree to act as managers underwriting a borrower's issue on short-term paper as and when required and to back the facility with medium-term credit should the note not find a market (Molyneux and Shamroukh 1996, 513). 20 results and also finds that organizations with fewer separately chartered banks, but more branches, introduced innovation earlier, which is consistent with theories stressing the importance of bank organizational form on lending style. A. What is financial innovation' "The primary function of the financial system is to facilitate the allocation and deployment of economic resources, both spatially and across time, in an uncertain environment." (Merton 1992, p. 12) This function, in turn, encompasses a payments system with a medium of exchange; the transfer of resources from savers to investor-users of the resources (and the eventual repayment to the savers); the gathering of savings for the purposes of pure time transformation (i.e., deferral/smoothing of consumption); and the reduction of risk through insurance and diversification. The operation of a financial system involves real resource costs, such as labor, materials, and capital employed by financial intermediaries (e.g., banks, insurance companies, etc.) and by financial facilitators (e.g., stock brokers, market makers, financial advisors, etc.). Further, since multiple time periods are an inherent characteristic of finance, there are also uncertainties about future states of the world that generate risks. For risk-averse individuals, these risks represent costs. The possibility of new financial products/services/instruments that can better satisfy financial system participants' demands is always present. Viewed in this context, a financial innovation represents something new that reduces costs, reduces risks, or provides an improved product/service/instrument that better satisfies participants' demands. Financial innovations can be grouped as new products (e.g., adjustable rate mortgages; exchange-traded index funds); new services (e.g., on-line securities trading; Internet banking); new "production" processes (e.g., electronic record-keeping for securities; credit scoring); or new organizational forms (e.g., a new type of electronic exchange for trading securities; Internet-only banks). Of course, if a new intermediate product or service is created and used 4 by financial services firms, then it may become part of a new financial production process. There are close analogies with familiar forms of innovation in non-financial contexts. There we see new products (e.g., DVD players; self-stick postage stamps); new services (e.g., Internet-based retail shopping); new production processes (e.g., improved processes for manufacturing computer chips) and new organizational forms (e.g., the "M-form" decentralized corporate structure).3 And innovations in producer goods are often essential for innovations in production processes. Much of the research attention to innovation focuses on the new idea. But at least as important is the adoption and spread of an innovation -- its diffusion -- across an industry. Indeed, faster diffusion means a higher societal return on the underlying investments in the innovation. B. Why is financial innovation important' Innovation is clearly an important phenomenon in any sector of a modern economy. Although standard microeconomic theory (rightly) focuses much of its attention on the issues of static resource allocation and economic efficiency, there is nevertheless general appreciation that performance over time is driven by a variety of dynamic factors,4 including innovation.5 The centrality of finance in an economy and its importance for economic growth (e.g., Levine, 1997) naturally raises the importance of financial innovation. Since finance is an input for 3 The M-form structure was originally discussed by Drucker (1946), Sloan (1964), Chandler (1962), and Williamson (1975). 4 "Making the best use of resources at any moment in time is important. But in the long run, it is dynamic performance that counts" (Scherer and Ross, 1990, p. 613). 5 From Solow (1957) onward, there has been a widespread realization that expansions of the capital stock of an economy are responsible for only a modest fraction of economic growth. The remainder, or residual, is due to a number of factors, of which research and development and the resultant innovations are a major component (Scherer and Ross 1990). 5 virtually all production activity and much consumption activity, improvements in the financial sector will have positive direct ramifications throughout the economy. Further, since better finance can encourage more saving and investment and can also encourage better (more productive) investment decisions, these indirect positive effects from financial innovation are yet greater still. III. What Motivates Innovation in General and Financial Innovation in Particular' Profit-seeking enterprises and individuals are constantly seeking new and improved products, processes, and organizational structures that will reduce their costs of production, better satisfy customer demands, and yield greater profits. Sometimes this search occurs through formal research and development programs; sometimes it occurs through more informal "tinkering" or trial and error efforts. When successful, the result is an innovation.6 If the search-and-success were a relatively constant phenomenon, innovations would tend to appear in a roughly continuous stream. However, since the observed streams of innovations do not appear to be uniform across all enterprises, across all industries, or across all time periods, the general innovation literature (see Cohen and Levin 1989; Cohen 1995) has sought to uncover the environmental conditions that may encourage greater (or lesser) search efforts and a larger (or smaller) stream of innovations. That literature has focused on hypotheses concerning roughly five structural conditions: (1) the market power of enterprises; (2) the size of enterprises; (3) technological opportunity; (4) appropriability; and (5) product market demand conditions. We will first briefly sketch these general conditions and then will focus on financial innovation and the environmental factors that the descriptive literature (cited above) suggests may encourage financial innovation; we will also relate these latter factors back 6 We will not here try to delve deeper into the "microstructure" of specifically how and why a "flash of inspiration" arises, generating an idea that eventually becomes an innovation. 6 to the general conditions. A. General structural conditions. 1. Market power. This hypothesis originates with Schumpeter (1950), who argued that market power is necessary to permit firms to generate sufficient returns from innovation. This is because of: (1) the inherent public good/free rider problems associated with new ideas, and (2) the difficulties of obtaining the finance for the sizable and uncertain investment in research and development (R&D) that is often required for successful innovation. 2. Enterprise size. This hypothesis also is identified with Schumpeter (1950). A larger size of an enterprise implies that the sale of the product embodying the innovation is likely to be large, yielding a greater return on the investment in the innovation. Also, greater size is necessary to allow the firm to accommodate the economies of scale inherent in R&D facilities, which are necessary to yield innovations. Finally, greater size is more likely to accommodate a wider range of activities and products, which may allow the firm to capture more of the unexpected spin-offs of the uncertain R&D process.7 3. Technological opportunity. Some industry technologies seem inherently more susceptible to innovation. For the past few decades, for example, computer chips, hardware, and software have experienced rapid technological progress; in earlier decades, the chemical industry seemed to have this susceptibility. 4. Appropriability. As mentioned above, information has the properties of a public good. In the absence of some protection or frictions, a productive new idea will be rapidly copied by rivals (who, in a competitive marketplace, will price their output at marginal cost), 7 Implicit in this discussion is the notion that innovations – expected and unexpected – are difficult to license or sell to other enterprises (because of asymmetric information problems), so that a firm must rely on its own capabilities. This is, of course, related to the appropriability issues discussed below. 7 thereby depriving the originator a return on his original investment in the innovation. The property rights regimes embodied in patents, copyrights, and trademarks provide some protections for innovators. Trade secrets and proprietary know-how provide additional protections, even where formal intellectual property protections are not available. 5. Product market demand conditions. Market size and growth are the main features capturing product market demand conditions. Specifically, a larger market will provide a greater return to a successful innovative effort, while a growing market is likely to provide the rents (profits) that can both entice and finance innovation. Other market characteristics might also be influential, such as the variability of demand, general macroeconomic conditions, tax regimes, regulatory regimes, etc. It should be emphasized that these conditions are hypotheses, to which counterhypotheses have sometimes been offered. For example, in contrast to the Schumpeterian hypotheses that suggest that monopoly and giant size are conducive to rapid innovation, Scherer (1984) suggests that smaller firms, with (at most) only modest levels of market power, may be more likely to be rapid innovators, because of the competitive pressures that are absent in the "quiet life" world of monopoly.8 B. The conditions that influence "equilibrium" rates of financial innovation. The descriptive literature that we cited above has suggested a number of environmental factors that have encouraged financial innovation. The list provided by Campbell (1988, ch. 16) is the most inclusive, and we will draw heavily on it. But, as good as Campbell's list is, it is seriously incomplete, because it focuses only on the levels of environmental factors and 8 And in response, Schumpeter (1950) likely would have argued that the possibilities of entry by one giant firm into another’s field was always enough of a possibility that the ”quiet life” was not a likely prospect. 8 neglects changes in environmental factors, as we will explain below. 1. Underlying technologies. The basic underlying "physical" technologies of finance are those of telecommunications and data processing, which permit the gathering of information, its transmission, and its analysis. Increasingly, these technologies allow financial market participants to measure and manage their risk exposures more efficiently and effectively. For example, with respect to lending, asymmetric information problems imply that lenders have difficulties determining who is a creditworthy borrower (adverse selection) and also have difficulties monitoring borrowers after a loan has been made (moral hazard). Accordingly, better (more advanced, faster, lower cost) physical technologies have permitted more innovations (e.g., credit and behavioral scoring) that allow lenders better to overcome those asymmetric information problems. Similarly, in terms of market risk, the use of value-at-risk and portfolio stress testing provide useful risk measures that can be used internally to set risk tolerance levels or allocate capital and externally to provide investors with a sense of overall exposure. Better physical technologies may also permit organizational innovations (e.g., electronic securities exchanges) that would not be possible with less advanced technologies. There is another technological dimension that is important for finance: intellectual technologies, such as the Black-Scholes option pricing model or the capital-asset-pricing model (CAPM). Advances in these technologies will, again, permit a wider range of innovations (e.g., computer programs that will readily compute option values). This category of environmental conditions for financial innovation has a direct parallel to the "technological opportunity" category of the general list above. 2. Macroeconomic conditions. Unstable macroeconomic conditions -- e.g., fluctuating prices, interest rates, exchange rates -- create uncertainties and risks and thus are likely to spur more innovation (to alleviate those risks) than would be true in a stable macroeconomic environment. Greater instability is likely to be associated with a faster pace of innovation. 9 This environmental condition seems best categorized as a parallel of "product market demand conditions" in the general list. 3. Regulation (legal environment).9 Regulation is a two-edged sword. On the one hand, some forms of regulation must inhibit innovation. For example, if regulation prevents commercial banks from owning insurance companies (and vice-versa), then whatever innovations might arise from joint ownership and operation will not occur. But, on the other hand, it is also clear that innovation can arise from efforts to circumvent regulation. To continue with the bank/insurance example, if cross-ownership is prevented, then banks will have the incentive to create insurance-like products and services (but, of course, will avoid labeling them as insurance), while insurance companies will have an incentive to create banklike products. Accordingly, it is impossible a priori to assign a positive or negative sign to the connection between the stringency of regulation (however measured) and the pace of financial innovation. Also, the innovation that arises as a consequence of regulation may be a socially positive or negative phenomenon. This depends on whether one sees the regulation itself as socially worthwhile (so that innovative evasion is a waste of resources or may even have socially deleterious consequences) or as a social waste (in which case the innovative evasion is a secondbest improvement). Again, this environmental condition seems best categorized as a parallel of “product market demand conditions” in the general list. 4. Taxes. To the extent that a tax system levies differential taxes on different streams of income or on different categories of assets, the higher taxed parties will seek ways of reducing their taxes. Financial innovation will follow. Higher levels of taxation will likely yield a larger 9 A more extensive discussion of the interaction between regulation and financial innovation can be found in White (2000). 10 flow of innovation. Again, whether one sees this innovation as a socially positive or negative phenomenon will depend on the social interpretation that one puts on the differential taxation scheme. Again, this environmental condition seems best categorized as a parallel to "product market demand conditions" in the general list. 5. Other influences' It is noteworthy that Campbell’s list does not include appropriability (and the intellectual property considerations that are associated with appropriability) or the Schumpeterian hypotheses of firm size and market power. Traditionally, the intellectual property protection system (i.e., patent, copyright, trademark) has not been considered important for financial innovation; patents for financial innovations were a rare phenomenon before 1980 and only became noticeable and significant in the late-1990s (Lerner 2002). Since Tufano (1989) shows that imitation of some innovations is rapid, a regime of intellectual property protection could encourage greater innovation.10 As for the Schumpeterian hypotheses, the absence of formal R&D facilities in financial services firms has probably been a significant factor in the relative neglect of the size and market-power considerations with respect to financial innovation. Also, neither the general innovation literature nor the financial innovation literature has satisfactorily addressed how the presence of network externality effects (Rohlfs 1974) influences the type and pace of innovation.11 With networks, the benefits to membership increase as more members join the network. Also, economies of scale and compatibility among members are usually important features of networks. The implications for innovation are cloudy, but 10 It is also possible, however, that excessively broad protection for ideas could discourage follow-on innovation, because of the transaction costs of negotiations between the original innovator and the follow-on innovators. 11 The exception in the financial innovation literature, which we note below, is Saloner and Shepard (1995). 11 potentially important. Incremental innovations within the compatibility confines of a network are clearly possible. But the scale-related problems of creating new networks may discourage such “large” innovations. 6. Interactive effects. The categories discussed above are not mutually exclusive. There may well be interactions among them. For example, regulations that are non-binding under one set of environmental conditions may be binding under another and may inspire circumventing innovations in the latter state, provided that the technological capabilities are present. For example, it is clear that the greater macroeconomic fluctuations of the late 1960s and the 1970s caused a tighter binding of the Federal Reserve's Regulation Q (which limited the payment of interest on bank deposits). This, in turn, inspired innovations such as money market mutual funds, "sweep accounts" for bank deposits, and Merrill Lynch’s “cash management account”; but these innovations would not have been possible without the improved computer and telecommunications capabilities of the 1970s. Some summary characteristics are in order: - Only five studies precede the 1990s – and 14 have appeared since 2000! - Only two studies address the environmental conditions that encourage financial innovations. Thus, the hypotheses advanced by the broad descriptive literature on innovation remain largely untested. Five studies address the characteristics of the customers for and users of financial innovations. Six studies address the diffusion of financial innovations. The remaining studies examine consequences and (explicitly or implicitly) welfare effects. - Only one study covers financial patenting; five cover innovations that pertain to securities or securities underwriting; the remaining 18 studies apply to banking. - Two financial organizational innovations (the establishment of Section 20 subsidiaries by BHCs and Internet-only banks) are covered by six studies. Some studies cover financial product/service innovations (e.g., debit cards); some studies cover financial process innovations (e.g., small business credit scoring); and some studies covered innovations that could be described as a process or as a product/service depending on the perspective taken (e.g., the offering of Internet banking). Taken together, these relatively few studies are suggestive (but not definitive) of some broader conclusions: Regulation does spur financial innovation, and this consequence should be considered before new policies are implemented. The adoption and diffusion of new technologies by banks is related to institution size. The use of new financial technologies by consumers is related to age, income, and population. 27 There are two studies of the characteristics of banks that offer Internet banking; two studies of start-up Internet-only banks; three studies of the diffusion of ATMs; four studies of the adoption of SBCS by large banks; and four studies of BHCs that established Section 20 subsidiaries. 25 The welfare effects of financial innovation appear to generally be positive. However, for organizational innovations this is less concrete. It seems clear that considerably more empirical work is possible, especially for testing the hypotheses concerning the conditions that encourage innovation. Further, some of the results that have been established in one area (e.g., banking) could be expanded to others. Why this has not already happened will be the topic of the next section.