Showing papers in "Journal of Finance in 1985"

Does the Stock Market Overreact

Abstract: Research in experimental psychology suggests that, in violation of Bayes' rule, most people tend to "overreact" to unexpected and dramatic news events. This study of market efficiency investigates whether such behavior affects stock prices. The empirical evidence, based on CRSP monthly return data, is consistent with the overreaction hypothesis. Substantial weak form market inefficiencies are discovered. The results also shed new light on the January returns earned by prior "winners" and "losers." Portfolios of losers experience exceptionally large January returns as late as five years after portfolio formation. As ECONOMISTS INTERESTED IN both market behavior and the psychology of individual decision making, we have been struck by the similarity of two sets of empirical findings. Both classes of behavior can be characterized as displaying overreaction. This study was undertaken to investigate the possibility that these phenomena are related by more than just appearance. We begin by describing briefly the individual and market behavior that piqued our interest. The term overreaction carries with it an implicit comparison to some degree of reaction that is considered to be appropriate. What is an appropriate reaction? One class,,of tasks which have a well-established norm are probability revision problems for which Bayes' rule prescribes the correct reaction to new information. It has now been well-established that Bayes' rule is not an apt characterization of how individuals actually respond to new data (Kahneman et al. [14]). In revising their beliefs, individuals tend to overweight recent information and underweight prior (or base rate) data. People seem to make predictions according to a simple matching rule: "The predicted value is selected so that the standing of the case in the distribution of outcomes matches its standing in the distribution of impressions" (Kahneman and Tversky [14, p. 416]). This rule-of-thumb, an instance of what Kahneman and Tversky call the representativeness heuristic, violates the basic statistical principal that the extremeness of predictions must be moderated by considerations of predictability. Grether [12] has replicated this finding under incentive compatible conditions. There is also considerable evidence that the actual expectations of professional security analysts and economic forecasters display the same overreaction bias (for a review, see De Bondt [7]). One of the earliest observations about overreaction in markets was made by J. M. Keynes:"... day-to-day fluctuations in the profits of existing investments,

Dividend Policy under Asymmetric Information

Abstract: We extend the standard finance model of the firm's dividend/investment/financing decisions by allowing the firm's managers to know more than outside investors about the true state of the firm's current earnings. The extension endogenizes the dividend (and financing) announcement effects amply documented in recent research. But once trading of shares is admitted to the model along with asymmetric information, the familiar Fisherian criterion for optimal investment becomes time inconsistent: the market's belief that the firm is following the Fisher rule creates incentives to violate the rule. We show that an informationally consistent signalling equilibrium exists under asymmetric information and the trading of shares that restores the time consistency of investment policy, but leads in general to lower levels of investment than the optimum achievable under full information and/or no trading. Contractual provisions that change the information asymmetry or the possibility of profiting from it could eliminate both the time inconsistency and the inefficiency in investment policies, but these contractual provisions too are likely to involve dead-weight costs. Establishing which route or combination of routes serves in practice to maintain consistency remains for future research.

The Disposition to Sell Winners Too Early and Ride Losers Too Long: Theory and Evidence

Abstract: One of the most significant and unique features in Kahneman and Tversky's approach to choice under uncertainty is aversion to loss realization. This paper is concerned with two aspects of this feature. First, we place this behavior pattern into a wider theoretical framework concerning a general disposition to sell winners too early and hold losers too long. This framework includes other elements, namely mental accounting, regret aversion, self-control, and tax considerations. Second, we discuss evidence which suggests that tax considerations alone cannot explain the observed patterns of loss and gain realization, and that the patterns are consistent with a combined effect of tax considerations and the three other elements of our framework. We also show that the concentration of loss realizations in December is not consistent with fully rational behavior, but is consistent with our theory. IT HAS BEEN WELL-KNOWN for over thirty years that individual decision makers do not behave in accordance with the axioms of expected utility theory. The famous Allais paradoxes [1] have made this point abundantly clear. Recent work by Kahneman and Tversky [15], Machina [19], and others has sought to provide a theory which describes how decision makers actually behave when confronted with choice under uncertainty. One of the key findings by Kahneman and Tversky concerns decision makers whose recent gambling history reflects losses. They indicate that their analysis suggests that a person who has not made peace with his losses is likely to accept gambles that would be unacceptable to him otherwise (p. 287). Kahneman and Tversky's finding was obtained in a controlled experimental situation. Economists tend to treat experimental evidence with some caution and are reluctant to conclude automatically that similar features will be exhibited in real-world market settings. Indeed, it is important to look at market behavior in order to ascertain whether such behavior patterns can be discerned in actual trading. In this paper, we examine decisions to realize gains and losses in a market setting. Specifically, we focus attention on financial markets and seek to determine whether investors exhibit a reluctance to realize losses (disposition to "ride

Dividends, Dilution, and Taxes: A Signalling Equilibrium

Abstract: A signalling equilibrium with taxable dividends is identified. In this equilibrium, corporate insiders with more valuable private information optimally distribute larger dividends and receive higher prices for their stock whenever the demand for cash by both their firm and its current stockholders exceeds its internal supply of cash. In equilibrium, many firms distribute dividends and simultaneously issue new stock, while other firms pay no dividends. Because dividends reveal all private information not conveyed by corporate audits, current stockholders capture in equilibrium all economic rents net of dissipative signalling costs. Both the announcement effect and the relationship between dividends and cum-dividend market values are derived explicitly. DIVIDENDS HAVE LONG PERPLEXED financial economists. Despite recent successes in constructing signalling equilibria with dividends, many important questions remain unanswered.' For example, why do corporations declare dividends and simultaneously sell new stock or, alternatively, distribute dividends and not repurchase stock? How do dividends with their dissipative costsprimarily adverse personal taxes-coexist with other presumably less costly technologies for releasing inside information, like audited annual reports? Do plausible signalling equilibria with dividends require transaction costs incurred by either corporations when issuing or retiring stock or investors when trading outstanding shares? Finally, how do the tax rates and demands for liquidity of such investors as widows, senior citizens, and financial institutions influence signalling equilibria? A satisfactory theory of signalling with dividends must also have empirical content. In particular, such a theory should provide empirically testable propositions detailing the effects of announced dividends on stock prices,2 crosssectional connections between dividends and market values, and any resulting relationships between payout ratios and rates of return on stocks.3 In addition,

International Asset Pricing under Mild Segmentation: Theory and Test

Abstract: This paper conducts a theoretical and empirical investigation of the pricing (and portfolio) implications of investment barriers in the context of international capital markets. The postulated market structure-labelled "mildly segmented"-leads to the existence of "super" risk premiums for a subset of securities and to a breakdown of the standard separation result. The empirical study uses an extended data base including LDC markets and provides tentative support for the mild segmentation hypothesis. THE QUESTION AS TO whether the international capital market is integrated or segmented appears particularly elusive. Indeed, the difficulties surrounding this important issue abound, as was made vividly clear by Solnik [20]. At the risk of tackling too ambitious a task, we undertake here to build a model and develop an empirical methodology to provide at least a partial answer to the

Optimal Release of Information By Firms

Abstract: This paper provides a positive theory of voluntary disclosure by firms. Previous theoretical work on disclosure of new information by firms has demonstrated that releasing public information will often make all shareholders worse off, due to an adverse risksharing effect. This paper uses a general equilibrium model with endogenous information collection to demonstrate that there exists a policy of disclosure of information which makes all shareholders better off than a policy of no disclosure. The welfare improvement occurs because of explicit information cost savings and improved risk sharing. This provides a positive theory of precommitment to disclosure, because it will be unanimously voted for by stockholders and will also represent the policy that will maximize value ex ante. In addition, it provides a "missing link" in financial signalling models. Apart from the effects on information production analyzed in this paper, most existing financial signalling models are inconsistent with a firm taking actions which facilitate future signalling because release of the signal makes all investors worse off. PUBLICLY TRADED FIRMS EXPEND resources to release information to outside security holders on an almost continual basis, even in the absence of regulatory requirements to do so. Firms enter into commitments to continue to release such information; e.g., such commitments are common in bond covenants. This paper develops an equilibrium model to explain such an expenditure by firms and shows that in many circumstances this public information makes all traders better off. The welfare improvement occurs in part because some traders would acquire costly information in the absence of the public announcement, while all abstain from information collection given the announcement. In addition, releasing public information is shown to improve risk sharing. There has been much analysis of mechanisms which act to signal or release information to traders. Dividend policy and capital structure policy have been analyzed as signals in, e.g., Bhattacharaya [3], Miller-Rock [20], and Ross [23]. No general explanation has been put forth which rationalizes an ex ante desire by traders for such information.' Ross [23] demonstrates that a particular observable variable can release information, but the desirability of such information release is not studied. Miller-Rock [20] presents a model which implies

An Investigation of Transactions Data for NYSE Stocks

Abstract: Using transactions data, the behavior of returns and characteristics of trades at the micro level is examined. A minute-by-minute market return series is formed and tested for normality and autocorrelation. Evidence of differences in return distributions is found among overnight trades, trades during the first 30 minutes following the market opening, trades at the close, and trades during the remainder of the day. The latter distribution is found to be normal. Unusually high returns and standard deviations of returns are found at the beginning and the end of the trading day. When the beginningand end-of-the-day effects are omitted, autocorrelation in the market return series is reduced substantially. A number of patterns in trading are reported. THIS PAPER EMPIRICALLY INVESTIGATES the nature of the return-generating process and the characteristics of trade size, price changes, trading frequency, and trading interval at the level of individual trades for a large sample of NYSE stocks. Previous studies of market indices have used daily, weekly, or monthly

Managerial Incentives for Short-term Results

Abstract: Of late, concern has been expressed that American managers tend to make decisions that yield short-term gains at the expense of the long-term interests of the shareholders. In this paper, we have attempted to investigate managerial incentives for such decisions. We find that, when the manager has private information regarding his or her decisions, there exist situations wherein the manager has incentives to make decisions which yield short-term profits but are not in the stockholders best interests. This incentive for suboptimal decisions arises because the manager, by taking decisions yielding short-term profits, hopes to enhance his reputation earlier, thus boosting his wages. We also find that this incentive is inversely related to her experience, the duration of her contract, and the risk of the firm.

The Week‐End Effect in Common Stock Returns: The International Evidence

Abstract: This paper examines the daily stock market returns for four foreign countries. We find a so-called "week-end effect" in each country. In addition, the lowest mean returns for the Japanese and Australian stock markets occur on Tuesday. The remainder of the paper answers four questions. Are seasonal patterns in foreign stock markets independent of those previously reported in the U.S.? Do Japan and Australia exhibit a seasonal one day out of phase due to different time zones? Do settlement procedures across countries bias week-end effects? Does the seasonal pattern in foreign exchange offset the week-end effect in stocks for Americans investing overseas? SOME OF THE MOST ANOMALOUS empirical findings in finance are associated with the sample distributions of daily common stock returns. Cross [1], French [2], Gibbons and Hess [3], and Keim and Stambaugh [4] have documented that the average return on Friday is abnormally high, and the average return on Monday is abnormally low. To our knowledge, this so-called "day-of-the-week effect" or "week-end effect" has yet to be explained. Because this anomaly has been reported primarily for U.S. stock returns, it is appropriate to investigate whether similar results occur for other countries. Positive findings would strongly support the proposition that the weekly seasonal effect is a general, world-wide phenomenon rather than the result of a special type of institutional arrangement in the U.S. To shed more light on this proposition, our paper examines stock market returns in the U.K., Japan, Canada, and Australia. Since we find a week-end effect in each country, we can examine a set of interesting questions. For example, are these seasonals independent of the previously reported seasonal in the U.S.? Due to different time zones, do Far Eastern countries exhibit a seasonal one day out of phase? Do different settlement procedures across countries influence weekend effects? Does the seasonal in foreign exchange fluctuations (see McFarland, Pettit, and Sung [7] and Levi [6]) offset the week-end effect in stock market returns for Americans investing overseas?

Introducing Recursive Partitioning for Financial Classification: The Case of Financial Distress

Abstract: The purpose of this study is to present a new classification procedure, Recursive Partitioning Algorithm (RPA), for financial analysis and to compare it with discriminant analysis within the context of firm financial distress. RPA is a computerized, nonparametric technique based on pattern recognition which has attributes of both the classical univariate classification approach and multivariate procedures. RPA is found to outperform discriminant analysis in most original sample and holdout comparisons. We also observe that additional information can be derived by assessing both RPA and discriminant analysis results.

Option Pricing and Replication with Transactions Costs

Abstract: Transactions costs invalidate the Black-Scholes arbitrage argument for option pricing, since continuous revision implies infinite trading. Discrete revision using Black-Scholes deltas generates errors which are correlated with the market, and do not approach zero with more frequent revision when transactions costs are included. This paper develops a modified option replicating strategy which depends on the size of transactions costs and the frequency of revision. Hedging errors are uncorrelated with the market and approach zero with more frequent revision. The technique permits calculation of the transactions costs of option replication and provides bounds on option prices.

On Jumps in Common Stock Prices and Their Impact on Call Option Pricing

Abstract: The Black-Scholes call option pricing model exhibits systematic empirical biases. The Merton call option pricing model, which explicitly admits jumps in the underlying security return process, may potentially eliminate these biases. We provide statistical evidence consistent with the existence of lognormally distributed jumps in a majority of the daily returns of a sample of NYSE listed common stocks. However, we find no operationally significant differences between the Black-Scholes and Merton model prices of the call options written on the sampled common stocks.

Differential Information and Performance Measurement Using a Security Market Line

Abstract: An uninformed observer using the tools of mean variance and security market line analysis to measure the performance of a portfolio manager who has superior information is unlikely to be able to make any reliable inferences. While some positive results of a very limited nature are possible, e.g., when there is a riskless asset or when information is restricted to be "security specific," in general anything is possible. In particular, a manager with superior information can appear to the observer to be below or above the security market line and inside or outside of the mean-variance efficient frontier, and any combination of these is possible. MEAN-VARIANCE THEORY PREDICTS that if we plot expected returns against beta coefficients, all securities plot on a single line, known as the security market line (SML). Since fpi, the beta coefficient for security i, is interpreted as a measure of the riskiness of security i, the market line is a graphical representation of the linear relationship between risk and return. What are we to make of deviations from the SML, i.e., of securities or managed portfolios that do not plot on such a line? If we can retain our intuitive interpretation of fj as the riskiness of asset i and the market line as the appropriate expected return needed to "reward" agents taking on various degrees of risk, then vertical deviations from the market line represent abnormal returns that differ from what is merited. This paper and the companion piece, Dybvig and Ross [10], explore in detail the validity of this intuition and the question of what SML deviations really measure. The companion piece analyzes SML deviations caused by the choice of an inefficient market or reference portfolio. In this paper, we analyze SML deviations caused by superior performance based on superior information. The intuitive explanation of deviations from the SML has motivated many

Corporate Combinations and Common Stock Returns: The Case of Joint Ventures

Abstract: The gain to stockholders from mergers is well documented. However, there is little evidence as to whether the source of the gain is due to synergy or management displacement. Merger is just one of an almost limitless variety of ways in which firms combine resources to accomplish some objective. A joint venture is another. In addition to being of interest as an independent phenomenon, because the original managements of the parent firms remain intact under a joint venture, investigation of wealth gains from joint ventures provides an opportunity to isolate the management displacement hypothesis from the synergy hypothesis as the source of gains in corporate combinations. Our results are 1) there are significant wealth gains from joint ventures, 2) the smaller partner earns a larger excess rate of return while the dollar gains are more equally divided, and 3) the gains, scaled by resources committed, yield "premiums" similar to those in mergers. We are inclined to interpret our results as supportive of the synergy hypothesis as the source of gain from corporate combinations. THIS PAPER IS AN empirical investigation of the common stock returns of U.S. companies that enter into joint ventures with other U.S. companies. We consider joint ventures in the context of what is by now a burgeoning literature examining the security returns of companies involved in intercorporate mergers and tender offers.1 In this context, our objective is to provide evidence as to the source of the well-documented gains to stockholders of firms involved in mergers and tender offers. In addition, joint ventures are an interesting economic phenomenon in their own right. They are undertaken by many of the largest U.S. companies, involve billions of dollars in capital expenditures annually, and are subject to considerable scrutiny in the popular press. In this regard, our analysis provides data on what is as yet a relatively unexplored area of economic activity.2 A corporate merger represents the joining together of all the resources of two companies under a single management to accomplish some set of objectives. Empirical studies have found that mergers generally result in significant increases in the total market values of the merging firms with the stockholders of the

On the Relevance of Debt Maturity Structure

Abstract: In this paper, we present a tax-induced framework to analyze debt maturity problems. We show that under some modifications of the existing U.S. tax code, debt maturity is irrelevant even in the presence of taxes and bankruptcy costs that yield an optimal capital structure. If this restrictive structure is relaxed, and assuming the Miller [15] equilibrium does not prevail, tax reasons would usually imply the existence of an optimal debt maturity structure. If there exists a gain from leverage, then an increasing term structure of interest rates, adjusted for default risk, results in long-term debt being optimal. A decreasing term structure, under similar circumstances, renders short-term debt optimal. In the absence of agency costs, a Miller [15]-type result emerges at equilibrium and irrelevance prevails. We also argue that agency costs could again reverse the irrelevance and imply a firm-specific optimal debt maturity structure. THERE HAS BEEN EXTENSIVE discussion in the literature concerning the existence of an optimal debt maturity structure. Kraus [13] states, without proof, that efficient capital markets preclude the existence of an optimal debt maturity. Stiglitz [20] demonstrates the irrelevance of debt maturity, but in an economic environment in which the entire financing decision is irrelevant due to the absence of taxes and bankruptcy costs. In contrast, Morris [16] argues that the issuance of short-term debt can reduce the risk to stockholders and thereby increase equity value, if the covariance between the net operating income and future interest rates is positive. This result by Morris is obtained even in the absence of taxes and when there is no probability of default. The discrepancy between Morris and Stiglitz arises because Morris uses the Bogue and Roll [31 multiperiod Capital Asset Pricing Model to incorporate uncertain future interest rates, which implicitly assumes that investors cannot diversify away intertem

Divergence of Opinion in Complete Markets: A Note

Abstract: We consider an Arrow-Debreu model with agents who have different subjective probabilities. In general, asset prices will depend only on aggregate consumption and the distribution of subjective probabilities in each state of nature. If all agents have identical preferences then an asset with "more dispersed" subjective probabilities will have a lower price than an asset with less dispersed subjective probabilities if risk aversion does not decline too rapidly. It seems that this condition is likely to be met in practice, so that increased dispersion of beliefs will generally be associated with reduced asset prices in a given Arrow-Debreu equilibrium. THERE HAVE BEEN SEVERAL recent investigations concerning the effect of heterogeneous probability beliefs on asset prices. Most of these investigations have taken place in the context of CAPM-like mean-variance models; see, for example, Lintner [5], Miller [7], Williams [10], Jarrow [4], and Mayshar [6]. Comparatively little has been done in the analysis of differences of opinion in an Arrow-Debreu contingent claims context. Rubinstein [8], Breeden and Litzenberger [2], and Breeden [1] have presented nice analyses of asset price determination in an Arrow-Debreu model, but most of their results have assumed commonly held probability beliefs. In this paper we analyze the impact of divergence of opinion on asset prices in an Arrow-Debreu economy. The results serve to generalize the findings of Rubinstein and Breeden-Litzenberger to the case of different probability beliefs. Among the results we establish are: 1. In equilibrium, asset prices depend only on aggregate consumption and the distribution of subjective probability beliefs. 2. Asset values are an increasing function of any one individual's probability beliefs. 3. An increase in the "spread" of the probability beliefs of investors may increase or decrease equilibrium asset values depending on the value of a parameter of the utility function. However, the most likely effect is to decrease the asset values. These results are especially interesting in light of the recent empirical work of Cragg and Malkiel [3]. They studied the relationship between ex post return and various measures of risk for common stocks and found that the measure of risk that performed best in their analysis was a measure of divergence of opinion

The Effect of Voluntary Sell-off Announcements on Shareholder Wealth

Abstract: Sell-off activities arise when a firm sells part of its assets (e.g., a segment, a division, etc.) but continues to exist in essentially the same form. This study investigates the effect of voluntary sell-offs on stock returns. From a sample of over 1000 sell-off events (first public announcements), the evidence shows that both sellers and buyers earn significant positive excess returns from these transactions. The excess returns earned by buyers are smaller than those earned by sellers. There is also evidence that sell-off announcements are preceded by a period of significant negative returns for the sellers which suggests that the sellers, on average, performed poorly prior to their sell-off activities. SUBSTANTIAL RESEARCH HAS BEEN undertaken in recent years on mergers, acquisitions, tender offers, capital structure changes, and other financial planning topics. This research has considerably expanded our knowledge about the association of these decisions with the welfare of the stockholders involved in these transactions.' Selling a portion of the firm's assets is another important financial planning decision.2 This paper investigates the effect of voluntary sell-off announcements on shareholder wealth. A sell-off is defined as selling part of a firm's assets (e.g., a segment, a division, etc.), while the firm continues to exist in essentially the same form as that prior to the sell-off. This study does not examine the effects of involuntary (ordered by the government) sell-offs and other similar activities such as spin-offs, leveraged buy-outs, liquidations, reorganizations, etc.3

Determinants of Corporate Leasing Policy

Abstract: The existing finance literature assumes the real operating cash flows from leasing or owning are invariant to the ownership of the asset and focuses on tax-related incentives for corporate leasing policy. Our analysis suggests that taxes are important in identifying potential lessees and lessors, but are less important in identifying the specific assets leased. We provide a unified analysis of the various incentives affecting the lease-versuspurchase decision. We then show how these incentives explain the use of contractual provisions such as maintenance clauses, deposits, options to purchase the asset, and metering. WHEN A FIRM BUYS AN ASSET, it obtains both the right to the services of that asset over the period it is owned plus the right to sell the asset at any future date. With a lease, the firm acquires only the right to the asset's services for a period specified in the contract. The existing finance literature analyzing corporate leasing policy concentrates on the tax-related incentives to lease or buy (e.g., see Miller and Upton [19], Myers et al. [21], Lewellen et al. [13], Franks and Hodges [11], and Brealey and Young [4]). We believe that taxes play an important role in explaining certain dimensions of leasing policy, for example, the choice between manufacturer and third-party leasing. However, taxes provide only a limited explanation of why specific assets are leased rather than owned, and for the choice of provisions in lease contracts. In this paper, we want to accomplish two things: (1) to provide a unified analysis of the various incentives affecting the lease-versus-buy decision and (2) to employ that analysis to explain observed variation in corporate leasing policy. In deriving testable hypotheses about the characteristics of lessees, lessors, the assets leased, and provisions in lease contracts, we extend the standard leaseversus-buy analysis found in corporate finance textbooks (e.g., Brealey and Myers

Moral Hazard and Information Sharing: A Model of Financial Information Gathering Agencies

Abstract: We propose a theory of information gathering agencies in a world of informational asymmetries and moral hazard. In a setting in which true firm values are certified by screening agents whose payoffs depend on noisy ex post monitors of information quality, the formation of information gathering agencies (groups of screening agents) is justified on two grounds. First, it enables screening agents to diversify their risky payoffs. Second, it allows information sharing. The first effect itself is insufficient despite the risk aversion of screening agents and the stochastic independence of the monitors used to compensate them. THIS PAPER DEVELOPS A model that explains the formation of financial institutions which acquire and process information for the purpose of certifying asset qualities, but do not get involved in funding. The model makes two basic assumptions. The first is that there is an informational asymmetry in the capital market, involving "insiders" possessing more accurate information about the true economic values of their firms than "outsiders", and the second is that those engaged in the production of information to rectify this asymmetry can gain from sharing their information. Examples of the information gathering agencies (IGA's) we study are creditrating agencies, financial newsletters, Standard and Poor's Value Line, investment counselors, credit bureaus, etc. IGA's are distinct from funding financial intermediaries (FFI's)-like banks and credit unions-that, in addition to acquiring and processing information, also issue claims against themselves, fund their customers' needs, and transform asset characteristics. From a modeling standpoint, a key distinction between IGA's and FFI's is that, in the case of the latter, "outsiders" (depositors, for instance) do not necessarily have to assess the quality of the intermediary's information since the intermediary can be made "accountable" for this quality through the positions it takes in the screened assets. Thus, models that explain FFI's focus on how the contracts offered by the intermediary ensure information reliability without reliance on explicit monitoring. On the other hand, because of the absence of loan- and deposit-type contracts to dissipate informational problems, a model explaining IGA's must rely on monitoring-related mechanisms.

Returns to Speculators and the Theory of Normal Backwardation

Abstract: A nonparametric statistical procedure is employed to examine the returns to speculators in wheat, corn, and soybeans futures markets. We find that the theory of normal backwardation is supported. Moreover, the presence of the risk premiums to speculators tends to be more prominent in recent years than in earlier years. We also find that large wheat speculators as a whole possessed some superior forecasting ability. The evidence is inconsistent with the hypothesis that commodity futures prices are unbiased estimates of the corresponding future spot prices. THE CLASSIC ECONOMIC RATIONALE for futures markets is that they facilitate hedging. That is, futures markets allow those who deal in commodities to obtain contracts through which the risk of price changes can be transferred to those who are willing to assume it. A side benefit of the market is that a publicly known, uniform future value for a commodity is created. Therefore, all commodity market participants can make production, storage, and processing decisions by looking at the pattern of futures price, even if they don't take positions in the futures market. Many believe that the current price of a futures contract equals the market consensus expectation of the spot price on the delivery date.' Keynes [10], however, in his theory of normal backwardation, suggests that it is unlikely that the above two functions can be fulfilled simultaneously. He argues that hedgers use the futures market to avoid risks, and that they pay a significant premium to speculators for this insurance. He concludes implicitly that the futures price is an unreliable estimate of the spot price prevailing on the date of expiration of the futures contract. Keynes' conclusion is based upon the argument that the long (short) speculator realizes the premium by refusing to purchase a contract from the short (long) hedger except at a price below (above) that which the futures price is expected to approach.2 Over the years, various studies have sought to confirm the existence of such a risk premium to speculators in the

An Unbiased Reexamination of Stock Market Volatility

Abstract: Recent work demonstrates serious statistical problems with standard volatility tests. This paper proposes new tests that are unbiased in small samples and that do not require assumptions of stationarity. The new tests continue to find evidence against the model positing rational expectations and a constant required rate of return on equity. IN HIS SEMINAL WORK on the volatility of the stock market, Robert Shiller [10] examines the view that changes in the level of the stock market reflect news about future dividends. In particular, he considers a model in which stock prices are the present discounted value, using a constant discount rate, of expected dividends.1 He provides what appears to be overwhelming evidence against this model. Dividends appear a smooth time series, while stock prices are volatile. The standard deviation of stock prices is, he concludes, about five times greater than would be expected from the volatility of dividends. Recent work by Flavin [2], Marsh and Merton [6], and others calls into question the validity of Shiller's test for a variety of reasons. Flavin examines the small sample properties of volatility tests and shows that they are extremely biased toward finding excessive volatility. Marsh and Merton argue that the excess volatility of stock prices relative to dividends is not surprising since firms, consciously smoothing dividends, make dividends a nonstationary series. They show that under a plausible dividend pay-out rule which makes dividends a random walk, the standard volatility test finds excess volatility in every sample, even though the efficient markets model is correct. In this paper, we propose new volatility tests. Our goal is to examine the same question Shiller addresses: can stock prices, a very volatile series, be the expected present discounted value of dividends, a relatively smooth series? The new tests, however, do not suffer from the problems Flavin and Marsh and Merton discuss. In Section I we briefly review Shiller's volatility test and then present our new ones. In Section II we consider the objections to Shiller's test and show that our tests are immune to these problems. We apply our tests to the data in Section III. We find that our new unbiased tests continue to provide evidence contra

A Sequential Signalling Model of Convertible Debt Call Policy

Abstract: In this paper we attempt to resolve two puzzles concerning convertible debt calls. The first is that although it has been shown that conversion of these bonds should optimally be forced as soon as this is feasible, actual calls are significantly delayed relative to this prescription. The second is that common stock returns are significantly negative around the announcement of the call of a convertible debt issue. Our purpose is to simultaneously rationalize managers' observed call decisions and the market's reaction to them in a framework in which managers behave optimally given their private information, compensation schemes, and investors' reactions to their call decisions. Moreover, investors' reactions are rational in the sense of Bayes' rule given managers' call policy. In equilibrium, a decision to call is (correctly) perceived by the market as a signal of unfavorable private information. In addition to rationalizing observed call delays and negative stock returns at call announcement, several other testable implications are derived.

Reformulating Tax Shield Valuation: A Note

Abstract: Standard financial theory (in the absence of agency costs and personal taxes) implies that each dollar of debt contributes to the value of the firm in proportion to the firm's tax rate. To derive this result, incremental debt is assumed permanent. This paper shows that when the firm acts to maintain a constant market value leverage ratio, the marginal value of debt financing is much lower than the corporate tax rate. Since Hamada's [2] unlevering procedure for observed equity betas was derived under the assumption of permanent debt, we derive an unlevering procedure consistent with the assumption of a constant leverage ratio. IMPLICIT IN THE MODIGLIANI-MILLER [4] hypothesis regarding the effect of leverage changes on firm value is that the firm maintains a constant level of debt. The purpose of this paper is to examine the implications for tax shield valuation of maintaining a constant market value leverage ratio instead of a constant debt level. Assuming a constant leverage ratio instead of a constant debt level does not imply any change in the normative implications of capital structure theory. However, we show that when the firm's financing strategy is to maintain a constant market value leverage ratio, the marginal value of a change in debt level resulting from a change in this leverage ratio is much lower than the corporate tax rate. We also derive the relationship between the firm's equity beta and its unlevered beta under the assumption of a constant leverage ratio.

Implications of the Discreteness of Observed Stock Prices

Abstract: Stock prices on the organized exchanges are restricted to be divisible by 1/8. Therefore, the "true" price usually differs from the observed price. This paper examines the biases resulting from the discreteness of observed stock prices. It is shown that the natural estimators of the variance and all of the higher order moments of the rate of returns are biased. An approximate set of correction factors is derived and a procedure is outlined to show how the correction can be made. The natural estimators of the "beta" and of the variance of the market portfolio, on the other hand, are "nearly" unbiased. THE BEHAVIOR OF STOCK PRICES has been an issue of interest to the financial economist for many years. This interest resulted in a growing number of empirical studies which attempt to estimate this behavior (e.g., Blattberg and Gonedes [2], Fama [6], Fama and Roll [7, 8], Barnea and Downes [1]). To date, stock price behavior is estimated under the assumption that the observed trading price is the "true" equilibrium price. However, observed stock prices and stock price changes on the organized exchanges are restricted to multiples of 1/8 of a dollar.1 Therefore, if the "true" distribution of stock prices is continuous, an observed trading price can be different from the "true" price. This paper examines the biases in estimating the moments of stock price changes caused by the discreteness of observed stock prices. The major focus of the paper is in noting this problem, providing a model which explains the source of these biases, and quantifying their size. Section I demonstrates that due to the discrete nature of observed stock prices the natural estimators for the variance and for the higher order moments of the rate of returns are biased upward. This bias is larger for stocks with lower prices and smaller standard deviation. For instance, assuming that the standard deviation, a, is 0.001, the stock price is one dollar, the "true" probability distribution of stock prices is lognormal, and the observed prices are as close as possible to the "true" prices, then the natural estimator of r has expectation 0.01400; hence, it is biased upward by 1300%. Significant biases have important implications in option pricing. We derive an approximate set of correction factors which can be applied to the

Upper and Lower Bounds of Put and Call Option Value: Stochastic Dominance Approach

Abstract: Applying stochastic dominance rules with borrowing and lending at the risk-free interest rate, we derive upper and lower values for an option price for all unconstrained utility functions and alternatively for concave utility functions. The derivation of these bounds is quite general and fits any kind of stock price distribution as long as it is characterized by a "nonnegative beta." Transaction costs and taxes can be easily incorporated in the model presented here since investors are not required to revise their portfolios continuously. The "price" that is paid for this generalization is that a range of values rather than a unique value is obtained. IN THE LAST DECADE, there has been an increasing interest in the options market by both academicians and practitioners. The most well-known valuation models are the Black-Scholes [4] model, the jump stochastic process model (see Cox and Ross [6]), the binomial model (see Cox, Ross, and Rubinstein [7], Rendleman and Bartler [19], and Sharpe [22]), and Rubinstein's discrete model [21]. Recently, Perrakis and Ryan [18], using Rubinstein's approach, have derived upper and lower bounds for option prices.' The relative pricing theory of contingent claims differs in the continuous time and discrete time models. In a continuous time framework, the instantaneous expected rate of return of the option and that of the underlying stock must follow a certain market equilibrium condition. In this case, the equilibrium condition is easily determined by a continuous hedging strategy. In the discrete time case, there is no hedging strategy to provide a simple relationship between the option and the stock price. However, for a given value distribution of the stock, the value distribution of the option is uniquely determined by the option exercise price, and this suggests some possible relationships between the stock and option values.

On Option Pricing Bounds

Abstract: The purpose of this article is to compare the Perrakis and Ryan bounds of option prices in a single-period model with option bounds derived using linear programming. It is shown that the upper bounds are identical but that the lower bounds are different. A comparison of these bounds, together with Merton's bounds and the Black-Scholes prices in a lognormal securities market, is presented. THE BLACK-SCHOLES [1] OPTION pricing model is preference-free. Its simplicity derives from the continuous-trading assumption which enables a portfolio of stock and bonds to be constructed so as to replicate the payouts of the option. In an incomplete market with only finite opportunities to revise portfolios, no preference-free option price can be constructed [3, 6]. For this case, preferencedependent prices can be established. Rather than place restrictive assumptions on preferences, Perrakis and Ryan (PR) [5] have recently derived upper and lower bounds for a single- (and multiple-) period call option price. The development of these bounds is based on the single price law and arbitrage arguments and requires modest assumptions on preferences. The bounds exist as equilibrium values given a consensus on stock price distributions. In this article, we shall develop bounds for option prices based on primitive prices in incomplete markets and compare these bounds to the single-period bounds derived by Perrakis and Ryan. Towards this goal, we first review the pricing approach for primitive securities and proceed to re-express the PR bounds as derived from their arbitrage portfolios. Then, using linear programming, bounds on option prices are derived. Linear programming approaches for obtaining relationships among option prices have been investigated. Garman [2], for example, established an algebra for evaluating hedge portfolios. Using this algebra, a one-period linear program can be established that provides a way to search for arbitrage opportunities over an infinite number of portfolio strategies. One of the features of the algebra is that only one linear program need be solved to determine if riskless arbitrage strategies can be established. By investigating the dual of the linear program, Garman obtained necessary and sufficient conditions for rational option pricing in a one-period setting. These bounds were shown to be consistent with the bounds determined by Merton [4]. Unlike Garman's approach, the linear programming approach presented here places constraints on time preferences, risk aversion, and state probabilities. Under specific sets of assumptions, linear programs are formulated where the objective function to be optimized is the call price. Minimization and maximi

The Rule 415 Experiment: Equity Markets

Abstract: Rule 415 allows a firm to register all the securities it reasonably expects to sell over the next two years and then, at the management's option, to sell those securities over these two years whenever it chooses. This paper examines whether equity offerings made under Rule 415 (shelf offerings) differ in issuing costs from equity offerings not sold under this rule. We find that shelf offerings cost 13% less for syndicated issues and 51% less for nonsyndicated issues. We also investigate the empirical relevance of the market overhang argument which suggests that shelf registrations depress the price of the registering firm's shares more than traditional registrations. Our data does not support the market overhang argument. THE SECURITIES AND EXCHANGE Commission's (SEC) Rule 415, which was permanently adopted for large firms in November, 1983, allows a firm to register all the securities that it reasonably expects to sell over the next two years and then sell those securities whenever it chooses. This procedure is known as shelf registration. Investment bankers, securities regulators, and corporate issuers vigorously debate the effect of Rule 415 on firms' issuing cost for new securities. Some (mostly investment bankers) argue that Rule 415 increases new issue costs while others (mostly issuers) argue that Rule 415 decreases new issue costs. In addition, many investment bankers argue that placing common stock "on the shelf" depresses the price of a firm's outstanding shares of stock. Most issuers,