نیمه تاریک نوآوری مالی : مطالعه موردی از قیمت گذاری یک محصول مالی خرده فروشی

The offering prices of 64 issues of a popular retail structured equity product were, on average, almost 8% greater than estimates of the products' fair market values obtained using option pricing methods. Under reasonable assumptions about the underlying stocks' expected returns, the mean expected return estimate on the structured products is slightly below zero. The products do not provide tax, liquidity, or other benefits, and it is difficult to rationalize their purchase by informed rational investors. Our findings are, however, consistent with the recent hypothesis that issuing firms might shroud some aspects of innovative securities or introduce complexity to exploit uninformed investors.

We carry out a detailed analysis of the pricing and expected returns of 64 issues of a popular structured equity product. Our pricing results imply that, on average, across the 64 issues, purchasers of the products lock in negative abnormal returns of at least 8% per year relative to dynamically adjusted portfolios of the underlying stock and bonds with the same risk. This negative abnormal return is large enough that under the most reasonable assumptions about the expected returns of the underlying stocks, the mean of the estimates of the expected returns on the structured products is less than zero. Given that the products returns covary positively with the broad market indexes and that the products do not appear to offer tax, liquidity, or other benefits, it is difficult to rationalize their purchase by informed rational investors. These findings are difficult to reconcile with the long standing view that financial innovations help issuers and (rational) investors achieve desired goals in the presence of market imperfections. Much of the extant literature portrays financial innovations as helping economic agents achieve a desired function in the presence of one or more market inefficiencies or imperfections.1 For example, new securities may be created to enable investors to achieve desired payoffs not spanned by the previously available financial instruments, or loosely to complete markets. Ross (1976) takes this “spanning” view of financial innovation in arguing that the introduction of option contracts improves allocative efficiency when the previously existing set of securities fails to span the state space.2 Other research presents financial innovations as arising to ameliorate imperfections such as agency conflicts (Ross, 1989), to reduce transactions costs (Ross, 1989, McConnell and Schwartz, 1992 and Grinblatt and Longstaff, 2000), and to minimize the impact of taxes or regulations (Miller, 1986 and Santangelo and Tufano, 1996). A common element in these papers is that financial innovations arise in response to market imperfections or inefficiencies and provide the benefit of ameliorating at least one imperfection or inefficiency.3 This generally benign view of financial innovation is unsurprising, since, as noted by Tufano (2003), “bringing new securities to market requires the voluntary cooperation of both issuers and investors.” Financial institutions ability to create securities providing state-contingent payoffs tailored to the needs or desires of specific investors or groups of investors seems especially conducive to achieving these potential benefits. But there is a dark side to the ability to create instruments with tailored payoffs. If some investors misunderstand financial markets or suffer from cognitive biases that cause them to assign incorrect probability weights to events, financial institutions can exploit the investors mistakes by creating financial instruments that pay off in the states that investors overweight and pay off less highly in the states that investors underweight, leading the investors to value the new instruments more highly than they would if they understood financial markets and correctly evaluated information about probabilities of future events. In addition, the ability to create instruments with almost any payoffs implies that there are few limits on the complexity of financial instruments. A recent theoretical literature explores equilibria in which firms shroud some aspects of the terms on which their products are offered in order to exploit uninformed consumers (Gabaix and Laibson, 2006), and strategically create complexity to reduce the proportion of investors who are informed (Carlin, 2009). In these equilibria, prices are higher than they would be if consumers or investors were fully informed. In the context of structured equity products (SEPs), these arguments imply that markups or premiums are higher and returns lower than they otherwise would be. We provide evidence that this darker view of financial innovation is likely to apply to offerings of at least some retail financial products by focusing on the initial pricing of Stock Participation Accreting Redemption Quarterly-pay Securities (SPARQS), a subset of the U.S. publicly issued SEPs. SEPs are medium-term notes issued by financial institutions and have payments based on another company's common stock price, multiple common stock prices (e.g., baskets of common stocks), a stock index, or multiple stock indexes. They are designed and issued by banks and investment banks and typically marketed to retail investors. SPARQS, created and marketed by Morgan Stanley, are the most popular of the publicly offered SEPs with payoffs based on the prices of individual common stocks. We find that the primary market investors pay, on average, an 8% premium for these securities, where the premium is defined as the difference between the offering price and an estimate of the fair market value. This is a large premium for a product that is callable after about six months and has a maximum maturity of slightly more than one year, as it implies that the purchaser locks in a negative abnormal return of at least 8% per year relative to a dynamically adjusted portfolio of the underlying stock and bonds with the same risk. Examination of the behavior of the secondary market price premiums to the model values over time indicates a gradual adjustment toward the model values. These premiums are large enough, and the expected lives of the SPARQS short enough, that under the objective measure the expected returns on the SPARQS are less than the risk-free rate of interest. In a standard model of portfolio choice, such a security is rationally purchased by an investor only if its returns covary positively with the investor's marginal utility (Merton, 1982). The returns of the SPARQS covary positively with the broad market indexes, and for the vast majority of investors, almost certainly covary positively with consumption and negatively with marginal utility. Thus, it is difficult to rationalize primary market purchases of SPARQS by investors who hold portfolios that are positively correlated with or uncorrelated with the broad market indexes. With such SPARQS investors would have been better off investing in bank certificates of deposit.4 It is unlikely that SPARQS satisfy any hedging needs of retail investors. The payoffs of SPARQS are qualitatively similar to those of covered calls, but SPARQS are callable by the issuer at a time-varying schedule of call prices, complicating their use as hedging instruments. Even for positions such as naked purchased put options for which SPARQS at first glance seem like a reasonable hedge, hedges involving ordinary exchange-traded options and the underlying stock seem more natural. The SPARQS also do not provide tax advantages, are not particularly liquid, and do not appear to help investors avoid transactions costs. In a complementary paper, Bergstresser (2008) presents evidence on the abnormal returns of a broad sample of SEPs that is consistent with our findings of overpricing on the offering date. Other researchers who have found that investors pay more than the fair market values for innovative securities include Rogalski and Seward (1991) and Jarrow and O’Hara (1989), who examined foreign currency warrants and primes and cores, respectively. These researchers argue that the securities they consider provide hedging benefits to investors, so their results do not have the negative implications for the products that ours do. (Benet, Giannetti, and Pissaris, 2006) consider the valuation of U.S. reverse exchangeable securities, and suggest that credit enhancement due to the positive correlation between the payoffs and the issuer financial performance and possible tax benefits might explain investor demand. Outside the U.S. markets, Szymanowska, Horst, and Veld (2009) estimate the values of reverse exchangeable securities issued in The Netherlands by ABN AMRO with payoffs based on the prices of other common stocks, while Baule, Entrop, and Wilkens (2008) calculate margins in the German retail structured products market using a sample of German “discount certificates” based on the prices of individual common stocks. Burth, Kraus and Wohlwend (2001) provide evidence about the pricing of Swiss structured products based on individual equities, which also have sizable markups. None of these researchers estimates the expected returns under the objective measures, so again their results do not have the negative implications about the products that ours do.5 The next section of the paper briefly describes the market for SEPs and the SPARQS that are the focus of our analysis. Section 3 analyses the initial pricing of the SPARQS and their post-issue performance. Section 4 departs from the risk-neutral measure used for valuation and shows that reasonable estimates of expected returns under the objective measure are less than the riskless rate. Section 5 explores the determinants of SPARQS issuances, and Section 6 discusses the implications of our findings and briefly concludes

We present evidence that the initial offering prices of SPARQS, the most popular of the publicly offered U.S. retail structured equity products, include “markups” or premiums large enough that the SPARQS have expected returns less than the riskless rate. What conclusions should one draw from this? Due to marketing costs, the offering prices of virtually all retail financial products must include premiums over estimates of their fair values. The equal- and value-weighted average commissions on the sample of SPARQS are 1.71% and 1.76%, respectively, so commissions alone would imply average premiums of this magnitude. In addition to these marketing costs, Morgan Stanley incurs transactions costs in hedging its liabilities on the SPARQS, pays fees associated with registering the securities and listing them on an exchange, and must also recover the staffing and other costs that go into designing, modeling, and valuing the SPARQS. Thus, conditional on existing, the SPARQS must be sold at considerable premiums over their model values. The interesting finding is that the markups are large enough, and the times to maturity or call short enough, that reasonable estimates of the expected returns on the SPARQS are less than the riskless rate. For the estimates of expected returns on the underlying stocks that seem most reasonable, the average expected return on the SPARQS is actually negative. In a standard model of portfolio choice, securities with expected returns less than the riskless rate are rationally purchased by investors only if their returns covary positively with the investor's marginal utility (Merton, 1982). The returns of the SPARQS covary positively with the broad market indexes, and for the vast majority of investors, almost certainly covary positively with consumption and negatively with marginal utility. Thus, the finding that SPARQS expected returns are less than the riskless rate makes it difficult to rationalize purchases of SPARQS by investors who hold portfolios that are positively correlated or uncorrelated with the broad market indexes. In light of this, the fact that investors buy SPARQS, and thus the existence of these securities, is a puzzle. It seems unlikely that SPARQS satisfy any hedging needs of retail investors. The payoffs of SPARQS are qualitatively similar to those of covered calls and, with the exception of naked puts, it is difficult to think of a position likely held by many retail investors for which a purchased SPARQS is a reasonable hedge. SPARQS are callable by the issuer at a time-varying schedule of call prices, decreasing their usefulness in hedging even those positions such as naked puts for which at first glance they might seem to be useful hedges. Setting aside specific arguments about what positions are or are not naturally hedged by SPARQS, in order to believe that SPARQS are designed to satisfy investors hedging needs, one would need simultaneously to believe that the investors who purchase SPARQS (i) are sophisticated enough to understand that SPARQS hedge the position that they have, and to manage the call risk; but (ii) are too unsophisticated to be aware that the underlying common stock and exchange-traded options on it provide alternative fairly priced hedging vehicles. SPARQS also appear not to be motivated by any desire to benefit from a preferential tax treatment, circumvent regulations, reduce agency conflicts, or reduce transactions costs. Thus, we cannot identify benign reasons for Morgan Stanley to structure and issue SPARQS. Because SPARQS expected returns are less than the risk-free rate, to the extent that SPARQS investors have marginal utilities that covary negatively with the SPARQS, they would have been better off investing in bank certificates of deposit. In fact, because the most plausible estimates of the SPARQS expected returns are negative, most SPARQS investors would likely have been better off investing in non-interest bearing accounts. Such strong statements cannot be made about many other retail financial products. For example, SEPs with longer times to maturity might not have expected returns less than the riskless rate. SPARQS issuances are, however, consistent with recent theoretical literature advancing the idea that firms shroud some aspects of the terms on which their products are offered in order to exploit uninformed consumers (Gabaix and Laibson, 2006), and deliberately create complexity to reduce the proportion of investors who are informed (Carlin, 2009). In these equilibria, prices are higher than they would be if consumers or investors were fully informed, implying that returns are lower than they otherwise would be. The overpricing of SPARQS is also consistent with the hypothesis that banks and investment banks design financial products to exploit investors misunderstandings of financial markets, cognitive biases in evaluating probabilistic information, and framing effects. For example, Shefrin and Statman (1993) argue that narrow framing and loss aversion19 can cause investors to demand positions that are similar to covered calls, and also make investors value a single security that provides these payoffs more highly than a portfolio of other securities that produces an equivalent payoff. This dark side of financial innovation is in stark contrast to the generally benign view of financial innovation that has been a traditional focus of the literature. The demand for SPARQS is also consistent with overconfidence, if an overconfident investor is one who believes that future returns are distributed narrowly around their expected outcome (Kyle and Wang, 1997; Daniel, Hirshleifer and Subrahmanyam, 1998; Odean, 1999). Such an investor will overvalue SPARQS because he or she will undervalue the embedded written call. Overconfidence by itself, however, does not provide an explanation of why investors might prefer SPARQS to covered call positions created using exchange-traded call options.