Speculative short sellers, put options and the weekend effect: a closer examination.
As documented by a voluminous research effort, equity securities show significant variation across weekdays. In particular, average Monday returns have been negative and significantly lower than other weekday returns. Recent empirical evidence, however, documents that Monday returns have become positive and larger than other weekday returns for large-firm securities only. Chen and Singal (2003) offer a hypothesis based on order flow that explains the historically low returns for Monday and a partial explanation of the shift in Monday returns for large-firm securities.
Many earlier researchers have also hypothesized that weekday variation in order flow causes negative Monday returns. These arguments include the following notions: 1) individuals use weekends as decision making time and then execute sells on Monday; 2) investors are inherently pessimistic on Mondays, making them more likely to sell and less likely to buy; and 3) institutions withdraw liquidity on Mondays, allowing sell orders to have a greater negative impact than on other weekdays. (4) Chen and Singal's novel argument attributes the weekend effect to order flow patterns caused by the trading of speculative short sellers that result in high Friday returns and low Monday returns. (5) They argue that because short selling is an inherently risky strategy, short sellers are reluctant to hold their positions over non-trading periods. Hence, speculative short sellers tend to close their positions on Friday and re-establish their positions on Mondays, (6) causing prices to rise on Fridays and fall on Mondays. This pattern, they argue, changed as less risky put options became available and bearish traders switching to this alternative no longer felt compelled to close their position over the weekend. Thus, the weekend effect disappeared for large-firm securities. For small-firm securities without available put options, however, the traditional weekend pattern continued.
Chen and Singal's argument is creative and appealing, but we identify three key factors that suggest a critical evaluation of their conclusions. First, we find it untenable that the trading pattern of bearish investors suggested by Chen and Singal, if pervasive enough to cause the weekend effect, would have gone unreported by the financial press. Likewise, if bearish investors had migrated en masse to put options from short-selling, this switch too would have been noted by the financial press. After extensive investigation we find no report of either behavior in the financial press. Nor do Chen and Singal provide any such example. Further, the financial press, as discussed below, continues to report bearish trading activity in terms of short-selling behavior.
Second, as argued below, weekend return patterns (as reported in literature extant at the time of the Chen and Singal study) would make the trading behavior assumed by Chen and Singal irrational. (7) Third, Chen and Singal argue that put buying is less risky than short selling. Although this view is widely held, we suggest that careful consideration leads to its rejection.
In the next section we report evidence of trading behavior of bearish investors as reported by the financial press and from a survey of bearish investors. We find no support for the hypothesis that short sellers routinely close their position over the weekend arguing against short-selling activity as the cause of the weekend effect. Further, we find no support for the hypothesis that bearish investors substituted put options for short selling on a wholesale basis following the introduction of put options as required by the Chen and Singal hypothesis. In Section 3 we discuss how existing empirical evidence conflicts with the Chen and Singal hypothesis and we develop the argument that put speculation is more risky than short selling, suggesting a lack of rationale for any impact on the weekend effect for individual stocks resulting from put option listing. We acknowledge that such an impact is an empirical question and in Section 4 we report results of an empirical investigation of the effect on the weekend effect from put option listing for individual equity securities. We find that on a security-by-security basis the introduction of put options has no impact on the weekend effect. Further in Section 4, we report findings that show no shift in weekday trading volume for the general market as required by the Chen and Singal argument. Section 5 of the paper contains our conclusions.
Evidence about Practitioners' Behavior
We use two sources to study trading behavior of bearish investors to search for evidence consistent with the Chen and Singal hypothesis. First, we examine the popular financial press for reports on the activities of bearish investors. Second, we directly survey a set of bearish investors and question them about their trading activity. Neither of these sources provides evidence in support of the contention that put option trading has supplanted short selling as the favored tool of risk-averse bearish investors. Nor, is any evidence provided in support of the contention that bearish investors who use short selling routinely close positions over the weekend.
Survey of the Financial Press
To examine reports from the financial press we use Pro Quest to search for references to short selling and put options in The Wall Street Journal. We search the period May 2007 through November 2008. Our search results in 238 references to short selling and 188 articles referencing put options. On an ex cathreda basis, referencing the article's title, we select 49 articles on short selling and 40 articles on put options for detail review. (8)
Reports from The Wall Street Journal referencing short selling make it abundantly clear that contrary to the Chen and Singal hypothesis professional investors have not abandoned the use of short selling in favor of put options to achieve a less risky bearish strategy. Instead, as The Wall Street Journal articles report: institutional investors rely primarily on short selling to implement bearish strategies, short-selling activity is sufficiently large to affect market prices and short-selling activity has reached record levels in recent trading activity.
A number of The Wall Street Journal articles mention the use of short selling by various bearish institutional traders. These include mutual funds, hedge funds and investment banking firms. For example, a September 2008 article indicates that Morningstar was tracking 41 bear-market mutual funds and 64 long-short mutual funds (also referred to as market-neutral funds). The article indicates that these funds use short selling and made no mention of the use of put options. This report indicates that short selling has not been abandoned by professional investors in favor of less risky put options.
Not only do these reports indicate that short selling is widespread; the reports also indicate that the volume of short selling is sufficiently large to have a quantitative impact on the market. For instance, a May 2007 Wall Street Journal article argues that a major cause of a nearly 200 point increase in the Dow Jones Industrial Average was the behavior of short sellers covering their positions. Likewise, in a September 2007 article, Browning (a long-time Wall Street Journal reporter) indicates that technical investors claim that heavy use of short selling by hedge funds has distorted a once reliable signal for a market bottom. A number of articles cite short-selling activity as impacting the price of an individual security. The most spectacular example of the impact of short-selling activity by hedge funds is provided in reports of the price gyrations of Volkswagon stock. A widespread belief that Volkswagon would drop in value led a number of hedge funds to short Volkswagon. Unfortunately for the hedge funds, as reported in October 2008 Wall Street Journal articles, Porsche created a short squeeze on Volkswagon stock. As a result when the hedge fund bears had to cover their short positions, the buying pressure was so strong that the price increase in Volkswagon stock briefly made Volkswagon the world's most valuable publicly traded company. Trading activity by institutional investor bearish on Volkswagon was sufficiently large to cause an extraordinary distortion in market valuation. This action was caused by short sellers not option buyers.
The vast majority of the articles in our sample discussed the role of short selling in the 2008 demise of a number of financial stocks. In spring of 2008 a number of articles appeared alleging that short sellers were using rumors to force down the price of financial stocks. A strategy described by SEC chairman Christopher Cox as the "distort and short" strategy. These early reports were followed by articles in which financial firms blamed not false rumors but the volume of short-selling activity of bearish investors as the cause of dramatic decreases in the prices financial securities. These claims were in turn followed by the temporary SEC ban on short selling of selected financial securities. A great deal of discussion was reported in the financial press on the impact and wisdom of such a ban, but certain features of this crisis are particularly germane to our discussion as to whether the weekend effect disappeared as a result of bearish investors switching to put options as their preferred tool. First, as reported in the financial press, bearish investors were using short sells not put options in an attempt to benefit from their predictions that the price of financial securities would fall. Second, both executives of financial firms and regulators within the SEC felt that the volume of short selling was sufficiently large to artificially depress the price of financial stocks. Indeed, bearish investors themselves do not deny the role of short selling in forcing the price of financial stocks down. Mr. James Chanos of Kynikos Associates (reported to be the country's largest short-selling fund) defends short sellers not on the basis that their activity had no effect on stock prices, but that the detective activity of short sellers exposed the true value of financial firms such as Fannie Mae and Freddie Mac. Third, the articles include the only mention in our sample of substituting put options for short selling. The cause of the substitution as reported in several articles was not, however, an attempt to reduce risk. Simply bearish investors who were prohibited from short selling adopted put options as a second best choice. Short-selling bans and the implicit recognition of the important role played by short sellers were not limited to the United States. The Journal reports short-selling bans in at least four other national markets.
Bearish investors have not surrendered the tool of short selling for an allegedly superior tool provided by put options. Indeed the use of short selling has grown to record proportions. Our sample includes three articles Wall Street Journal published in the late spring of 2008 that report record short-selling volume on the NYSE, Amex and Nasdaq, respectively. Chen and Singal (p. 691) note an increase in short-selling volume after the weekend effect disappeared, but they attribute this increase to the use of short positions to hedge put positions. This argument is inconsistent with the discussion provided in the Journal articles as to the cause of the record level of short selling. All three articles described the use of short selling as bearish bets. None of the three articles described the use of short sells to hedge put positions. Clearly bearish investors continue to use short selling as their primary tool rather than put options. This, of course, does not argue that put options are not used by bearish investors. Below we describe the discussion of the use of put options made by bearish investors as reported in The Wall Street Journal.
We examine forty articles reporting the use of put options. Ironically, the majority of these articles discussed the use of put options by bullish investors. Most of these articles describe the use of put options as insurance for bullish investors. In some cases the reports cited the use of put insurance for an entire portfolio. For example one article mentioned Edward Hunia, CIO for the Kresge Foundation who was reported to have "bought put options to cover all his equity holdings." Other articles mention the use of put insurance on individual securities ahead of earnings or some other announcement which might adversely affect the price of the security. Several other articles mention the use of puts as an aggressive strategy by bullish investors relative to individual stocks. These bullish investors were selling puts which they assumed would expire out of the market allowing the bullish investor to pocket the entire price of the option.
Of the 40 articles selected concerning put options only 17 discuss bearish strategies. Most of these articles discuss the use of put options in taking positions concerning individual stocks such as Boeing and General Motors, or relative to an investment sector such as the financial sector. None of these articles referred to the use of put options as the standard investment strategy for any group of investors. In contrast, the articles cited above routinely ascribed short selling as the standard tool for various groups of bearish investors. None of the articles attributed the use of put options as a risk-reduction tool relative to short selling. Only one article in the sample discussed an individual bearish investor whose investment strategy relied primarily on put options. Even here the choice of put options, however, was not made with the goal of risk reduction.
A July 2007 Wall Street Journal article focuses on Nassim Taleb, author of The Black Swan, a bearish investor who places bearish bets with put options. Taleb argues that investors in general underestimate the likelihood of extreme events, especially with regard to the downside. He therefore places downside bets with the expectation of generally losing but with the expectation of winning big when the extreme event occurs. He argues that this strategy will "beat the market" because investors underweight the risk associated with an extreme event. Thus, when this particular bearish investor uses a long position in put options it is not to minimize risk but rather to acquire the large return associated with a high-risk position in underpriced put options.
The financial press clearly fails to indentify either of the main behaviors on the part of bearish investors presumed by Chen and Singal. There is no report of short sellers closing positions over the weekend. There is no report to indicate that bearish investors have adopted put options in general as their preferred trading strategy. This finding contradicts the Chen and Singal hypothesis. To further test this hypothesis, we directly survey bearish investors concerning their trading behavior. (9)
Survey of Bearish Practitioners
We began our survey by interviewing Mr. Doug Kass, general partner with Seabreeze Partners Short L. P. Mr. Kass is a frequent contributor to TheStreet.com, widely quoted in the investment community and serves as a guest host on CNBC's "squawk box." He has been senior portfolio manager with Omega Investments and has held positions with Putman and Kidder Peabody. He holds an MBA from the Wharton Business School. Mr. Kass reports that he uses both short selling and put options to establish bearish position but that he holds primarily short-selling positions belying Chen and Singal's claim that bearish investors moved en masse to put options after their introduction. Further, Mr. Kass reports that he has never terminated a short position on Friday to avoid risk over the weekend close. Thus, the central assumption employed by Chen and Singal, the assumption that short sellers routinely close positions over the weekend to avoid risk, appears to be false. Indeed, none of the practitioners in our sample indicates that they engaged in the behavior hypothesized by Chen and Singal.
All the rest of our survey participants requested anonymity but they provided support for our conclusion that bearish traders primarily use short selling and that "they do not routinely close short positions on Fridays. Sixteen participants responded to our telephone survey. The interview was open-ended, but a core set of questions (see Appendix II for the list of questions) were asked of all participants. Each of the survey participants primarily used short sales rather than put options as their tool for bearish investments. At a minimum the investors estimated that 70% of their trades were conducted using short sales. A number of survey participants indicated that their bearish trades were exclusively or almost exclusively done with short sales. Thus, our survey clearly contradicts the Chen and Singal argument that with the advent of put options bearish investors switched to the predominant use of put options. Just the opposite appears is the case. Prominent bearish investors still rely predominantly on short sales!
Our survey results also contradict the argument that short sellers routinely close their positions on Fridays. Indeed, this suggestion was met with some derision on the part of the traders as they argued that transaction costs would cause this behavior to be highly unprofitable. Two of the traders did suggest that they would close their short positions over the weekend if they were concerned about the release of a particular piece of news. This behavior certainly is not consistent with general market swings on Fridays and Mondays. One of the respondents was a day trader who held his positions for a very short period of time. He would often terminate his position at market close. And he would always terminate his position at Friday close.
We specifically inquired about the traders' view of the relative risk of short selling versus put buying. The responses were generally framed in terms other than the return distribution as described by Chen and Singal. The traders felt that their portfolio was sufficiently diversified to forgo the need for additional hedging. In comparing the risk of puts and shorting they emphasized time risk for puts. They considered themselves to be driven by fundamentals of the company which they expected to ultimately create a correction. The certainty and extent of the correction provides risk for both short selling and risk buying. The fund managers, however, felt that the use of put options involved an additional risks associated with timing. The put expiration must be timed to the market correction. They also, in general, cited the cost of the put option as a rationale for using short selling rather than put options. Thus, we find based on our survey of bearish investors evidence to dispute both facets of Chen and Singal's dual hypothesis. Bearish investors did not shift on a wholesale basis from shorting to puts when tradeable put options became available. Further, short sellers do not routinely close their positions on Fridays.
Rationale for the Continued Use of Short Selling
In the previous section we report evidence that bearish investors still rely primarily on short selling to carry out their market strategy. In this section we provide two explanations for the resilience of the short-selling strategy after the introduction of put options. Our first explanation cites the pattern in Monday returns as described in literature, extant at the publication of Chen and Singal's arguments. The composition of Monday returns described in this literature argues against the likelihood that bearish investors would close their position over the weekend. Both Rogalski (1984) and Harris (1986) show that the negative Monday effect primarily occurs between Friday close and Monday open. Rogalski reports average Friday close to Monday open sample returns of -0.08% and average Monday open to Monday close sample returns of 0.05%. Given this pattern, if short sellers behaved as suggested by Chen and Singal, they would be closing their position over a period where they would forgo, on average, substantial gains only to re-establish their position during a time period which, on average, they would lose on their position. We presume that rational short sellers would not make this mistake repeatedly.
Extant literature also questions the Chen and Singal argument that the introduction of put options relieves buying pressure at Friday close and selling pressure at Monday's open. This argument implies that, over time following the introduction of put options, Friday close to Monday open returns ought to increase relative to Monday open-to-close returns. Smirlock and Starks (1986) report findings that are exactly opposite of this prediction. In three sub periods from 1963 through 1983, Friday close to Monday open returns become progressively more negative and Monday open-to-close returns become progressively more positive.
Our second explanation, for the continued use of short selling as the primary tool to implement bearish strategies after the introduction of put options, rests on our argument that short selling is a less risky bearish strategy than buying put options. This contention directly challenges Chen and Singal's argument that put options reduce risk sufficiently to alter the behavior of risk-adverse bearish investors. Chen and Singal postulate that a put option strategy is less risky than short selling because: "the risk of put buyers is limited to put value and not unlimited as with short sales ..." (p. 695). Certainly, the lowest return that a put buyer may experience is -100%, whereas the maximum theoretical loss for a short seller is negative infinity. However, comparison of this single drawing from the return distributions of short selling and put buying is not sufficient to determine that put buying is a less risky bearish strategy than short selling.
Consistent with behavioral studies by Fredrickson and Kahneman (1993) and others we posit that investors tend to emphasize extreme negative outcomes. Thus, when comparing risk of short selling versus put options, the popular belief that short selling is more risky than put options may be due to this behavioral tendency to emphasize the worse case scenario. We postulate that if one avoids the peak-end phenomena and observes the entire return distribution faced by a bearish investor then one would find short selling to be a less risky strategy than put buying for three reasons as discussed below.
First, although -100% is the lowest return that can be suffered by a put buyer, this loss is much more likely to be achieved in put buying than in short selling. Assuming a 50% margin for a short seller, the price of the stock must rise by 50% before a 100% loss occurs. A put buyer (assuming the put is held until expiration) loses 100% if the option expires out of the money. Unfortunately for the put buyer, this is a very likely event. If the put was purchased out-of-the-money, this event would occur: if the stock price increases, remains constant or drops by an amount less than the difference between the strike price and the stock price at purchase. If the put is purchased at-the-money, the speculator loses 100% if the price of the stock increases or stays constant. (10) If the put is bought in-the-money, the price of the underlying stock must increase in order for the put buyer to lose 100% of the investment value. However, because most options are bought with strike prices close to the market price of the underlying securities, a 100% loss for an option bought in-the-money usually requires only a modest increase in stock price. Rational put buyers must invest with the realization that there is a very high likelihood of losing 100% of their investment. Surely a strategy that offers a high probability of realizing a -100% loss should not be viewed as a low-risk strategy!
Second, short sellers are more likely to break-even than put buyers. Short sellers break-even if the price of the stock stays constant. Since put values include a declining time-value premium, put buyers would only break-even if the price of the stock decreases enough to offset the loss in this premium. Thus, put buyers are generally at a disadvantage in terms of downside risk. Put buyers are more likely to experience a negative return and are much more likely to experience a 100% loss. Only in the case of a greater than 50% increase in stock price (assuming a 50% short-selling margin) will the percentage loss be greater for the short seller than for the put buyer.
Third, the standard deviation of returns to the typical put option strategy is much higher than the standard deviation of returns for a short-selling strategy. Chen and Singal state that: "The put options allow the short sellers to reduce their risk (because their loss is limited to the put premium) and increase leverage." (p. 695) Financial theory suggests that leverage would increase expected returns but would do so at the cost of additional risk. Put investors, as we have argued above, are much more likely to suffer a negative return than short sellers, but once put investors break even they will quickly achieve a higher return than their short-selling counterparts due to their leveraged position. The base to determine this increase in returns for bearish investors using put options is the price of the option. The base for determining returns for bearish investors who sell stocks short is 50% of the price of the stock. The possibility for large positive returns further increases the variability of returns for put buyers. These relationships clearly argue that bearish investors do not use puts to minimize risk. Rather bearish investors use puts to increase upside potential at the cost of greater risk.
We have demonstrated that when undertaking a bearish position over a particular time period the investor faces considerably higher risk with a put position than with a short position in stock. There is an additional fundamental and significant risk factor associated with a put strategy relative to a short-selling strategy. A bearish strategy assumes that a security is currently overvalued by the market and attempts to establish a position that will profit from a market correction. Because the bearish investor cannot foretell exactly when the correction will occur, the investor will face risk if the bearish position is time-limited, allowing the position to terminate before the market correction occurs. Because put options expire, bearish investors using put options face this risk much more than short sellers. For example a bearish investor who recognized technology stocks were overpriced well before the dot.com bubble burst in March 2000 could have suffered substantial losses from transaction costs incurred in rolling over expiring put positions. Taken together, the above reasons clearly argue against the Chen and Singal notion that a long position in a put option is less risky than a short position in the stock.
We have argued against the popular view that put option investing is less risky than short selling. This popular view rests on the idea that value at risk is lower with the purchase of a put option than the investment required for short selling. This argument, however, cannot stand on its own. Clearly the per unit cost of a particular investment asset does not determine the risk of a particular asset. Investment in a penny stock is not less risky than an investment in a share of Berkshire Hathaway because the potential loss per share is lower. Thus, a put investment is not less risky than short selling because of the lower per unit dollar investment. This argument must be tied to changes that the bearish investor would make to his/her total portfolio because of the availability of put options. One may assume, for instance, that the bearish investor wishes to make a certain dollar allocation to his/her bearish position in a particular stock. In this case the investor would simply be able to control more shares of the underlying stock with the use of the put option. In this case, the risk-return tradeoff would follow the arguments presented above.
On the other hand, one may argue that the new return distributions created by the availability of put options will change the investment value that the bearish investor chooses to place in a particular underlying security. The investor may wish for instance to mix a bearish position in the underlying stock with a long position in Treasury bills given the new return distributions available with the put option listings. For instance, the investor may choose to control the same number of shares with the put position as with the short position and place the excess fund in Treasury bills. Such behavior would obviously modify the risk-return tradeoff for the put strategy, possibly lowering the risk to a lower level than a short position. We conjecture that speculative short sellers are not likely to have portfolios with large investments in a risk-free strategy.
More plausibly, one might assume that the bearish investor would choose to select a risk-return tradeoff that invests in bearish positions in additional stocks with a given budget constraint. However, short sellers are often pictured as aggressively pursuing a particular bearish strategy rather than as pursuing a strategy to achieve portfolio diversification. Regardless of the theoretical argument that could be made indicating that put option buying is less risky than short selling, as reported above bearish investors do not appear to have shifted their trading behavior on this basis. Further, evidence against this proposition is reported in the next section where we measure the impact of put option listing on the weekend effect on a security-by-security basis.
Empirical Evidence Market Data
Evidence Provided by Chen and Singal
Above we report, contrary to the Chen and Singal hypothesis, that bearish investors still rely primarily on short selling to implement their strategies. We provide two explanations: First, observed patterns in Monday returns for the U. S. equity market would punish bearish investors for closing positions at Friday close; and second, we argue that buying put options provides a more-risky not a less-risky alternative to short selling as a bearish strategy. In this section we supplement our previous arguments with empirical evidence from market data testing the Chen and Singal hypothesis. We find that stocks with listed put options do not show changes in weekend return patterns as suggested by Chen and Singal. Further, we show on a macro basis that market trading volume and return patterns do not exhibit shifts implied by the Chen-Singal hypothesis. Before we examine this data we provide a brief review of the empirical evidence reported by Chen and Singal.
Chen and Singal provide empirical support for three relationships consistent with their hypothesis. First, consistent with the findings of other researchers, they show that the weekend effect disappeared for large firms but not for small firms. (11) Second, they show that on average firms with higher relative short-selling volume are associated with a stronger weekend effect. (12) Empirical tests for a third and crucial relationship between the advent of put options and the elimination of the weekend effect for individual stocks are mainly indirect. For instance, Chen and Singal study the relationship between stock volume and the weekend effect. They find that securities with high stock volume have a low weekend effect. On the basis that put volume is positively correlated with stock volume, they argue that put option trading causes a decline in the weekend effect. This indirect evidence is, however, consistent with any other cause of a decline in the weekend effect for large-firm securities. What is missing from the Chen and Singal analysis is a direct test of their hypothesis linking put option listing for individual securities with changes in the return pattern for those securities. We provide this test by examining the return behavior of individual securities that experience put option listing. The results of this micro test suggest that put listings do not cause the observed reduction in the Monday effect, let alone its reversal. On a macro basis we examine shifts in trading volume and index return for the New York Stock Exchange (NYSE). We find no shift in trading volume patterns that would be required by the Chen and Singal hypothesis. And, we find the magnitude and timing of the shifts in weekday returns to be inconsistent with their hypothesis. Below we first present the results of the micro test and then the results of the macro tests.
Changes in Weekday Return Pattern of Securities with Listed Put Options
Chen and Singal are not the first researchers to suggest a link between newly created derivatives and the elimination of the weekend effect. Kamara (1997) argues that the introduction of the S&P 500 index futures in April 1982 and the rise of institutional investors combine to encourage arbitrage activity to eliminate low Monday returns. Both the Chen and Singal and Kamara hypotheses are consistent with the empirical observation that the weekend effect was eliminated for large firms but not for small firms. Both hypotheses are subject to the limitation that some other trend factor affecting large-firm returns other than the introduction of the derivative product may have created the observed shift in the weekday returns. The Kamara hypothesis is especially vulnerable to this suggestion because the introduction of the S&P 500 futures index occurred at a particular point in time presumably affecting all 500 securities included in the index at once. However, as put options were introduced in a piecemeal fashion, we can examine the impact of put option listing on the weekend effect on a security by security basis. Such an empirical examination should find an isolated effect on individual securities after the listing occurs, while other large-firm securities that do not have a put option listing should be unaffected. This relationship allows a precise test of the Chen-Singal assertion. We develop the following three testable hypotheses: One, the relative weekend effect, as measured by the Friday minus Monday return, for the underlying security decreases following put option listing. Two, the relative Monday return for the underlying security increases following put option listing. Three, the relative Friday return for the underlying security decreases following put option listing. All three hypotheses must hold to support the Chen and Singal argument.
We collect data for put options listings from 1977 through 1981. The Securities and Exchange Commission (SEC) initially granted permission to trade put options on a trial basis beginning June 1977. On this month each of five exchanges began to trade put options on five different underlying securities. After the trial period ended the SEC allowed general listing of put options, beginning May 9, 1980. During 1980 and 1981 these exchanges added new put option listings on a total of 235 underlying companies. (13) These put listings were predominately for securities on which call options were already traded but also included securities with no previous options listed. By the end of 1981 virtually all securities with listed call options had corresponding put options and we terminate our sample at this date. This termination date provides the advantage of avoiding the confounding effect from the introduction of the S&P 500 index futures which, as noted above, according to Kamara (1997) caused the end of the weekend effect. For each of the sample securities we collect return data from the CRSP data base for 750 trading days (approximately three years) before and after the initial put option listing to search for a shift in their weekday return patterns.
To determine listing dates for each put option, we examine daily issues of The Wall Street Journal for June 1977 and for the period May 1980 through December 1981. We identify the option listing date by the first price reported in the Journal. (14) To ensure comparability, we require that the security must have returns available for 90% of the trading dates in each of the three-year pre- and post-listing periods. This requirement reduces the 1977 sample size from 25 to 21 and the size of the 1980-1981 sample from 235 to 223. Thus, our total sample consists of 244 put-listing events.
Methodology and Empirical Results
To assess the impact of put listings on the weekend return, we examine differences in Friday returns, Monday returns and the difference in the paired Friday minus Monday return for each security in our sample in the pre and post-listing period. (15) Data is collected for the three-year period before (pre) and the three-year period after (post) listing of put options. For each of these measures, on a security-by-security basis, we subtract the average post-listing period return from the average pre-listing period return. According to the Chen-Singal hypothesis the average difference in the Friday returns ought to be positive, as Friday returns should fall in the post-listing period. The average difference in Monday returns ought to be negative, as the Monday returns ought to increase in the post-listing periods. The separate impacts on the Friday and Monday returns ought to cause the average Friday minus Monday return differential to fall in the post-listing period. Hence when we subtract the post-listing Friday minus Monday differential from the prelisting Friday minus Monday differential, the difference ought to be positive.
Panel A of Table 1 reports the average weekend effect for the sample securities in the pre- and post-listing periods and the average difference in the weekend effect between the pre- and post-listing period. In direct conflict with the Chen and Singal hypothesis, the weekend effect for the post-listing period is a positive 0.28%. This effect is highly significant with a t-statistic of 17.31. The weekend effect does not disappear after put option listing. The size of the weekend effect is a bit smaller than in the pre-listing period, but it is more significant statistically. The weekend effect does not go away because of put option listings in the first three years following listing.
Panel B of Table 1 reports average Monday returns for the sample securities before and after put option listing. According to the Chen and Singal hypothesis the average Monday returns should increase in the post-listing period as bearish investors no longer need to re-open closed short positions on Monday. It does not appear that re-opening of short positions was the cause of negative Monday returns because after put option listing these securities continued to experience negative Monday returns. These negative Monday returns are significantly less than zero. And, indeed the Monday returns are lower in the post-listing period than in the prelisting period. These findings do not reconcile with the Chen and Singal hypothesis.
Panel C of Table 1 reports average Friday returns for the sample securities for pre- and post-listing periods. Consistent with the Chen and Singal hypothesis there is a significant decrease in the average Friday return in the post-listing period relative to the pre-listing period. However, additional results prevent the reduction in Friday's average return from providing evidence in favor of the Chen and Singal hypothesis. First, if Friday's shift resulted from short-seller's activities in the spirit of the Chen and Singal hypothesis there would need to be a corresponding change in Monday's return. As we have seen there was none. Further, the weekend effect is still significant.
As shown in Panel A, Friday's average return is significantly higher than Monday's average return. Additionally, in results not reported here but available from the authors, Friday's average returns remain higher than the average return for any other weekday in the post-listing period.
One may argue in defense of the Chen and Singal hypothesis that we find no significant change in the weekend effect, because our sample deals with the earliest put options listings. In such a period the put option market may not have been sufficiently developed to allow bearish investors to shift to their trading strategies. To test the robustness of our findings we split the sample by calendar years and examine separately the 1977 and 1980 listings as one group and the 1981 sample as the second group. If market development is an issue in our findings we might find a significant change in the weekend effect for the 1981 sample but not for the 19771980 sub-sample. Panel D of Table 1 shows that the post-listing weekend effect is still highly significant for the 1981 sample. We note, however, that the difference in the weekend shift reported in Panel A came almost entirely from the 1977-1980 sample and that the difference in the weekend effect in the post-listing and the prelisting period is significant for the 1981 sub-sample. Because of this finding, we conduct an additional test of the possibility that our results are influenced by the timing of our sample.
Test for Put Trading Volume Influence
To further investigate the concern that our results may be biased from a sample taken from the early years of put option trading we examine the relationship between trading volume and shifts in the weekday effect. One may argue that during these early listing years trading volume may have been too low for bearish investors to shift from short selling to put buying. (16) If this is the case, one should observe a differential in the put listing impact across securities based on trading volume. To examine this possibility we determine average daily put volume over the three-year period (17) following listing for the put options included in the 1977 sub-sample. For these options average daily volume ranged from 156 to 7,362 contracts a day with a median of 615 contracts. Thus, a reasonable divergence appears to exist in the trading volume of put options to identify a differential impact across listings from variation in trading volume. If inadequate trading volume explains why our findings fail to support the Chen and Singal hypothesis, we should find a stronger impact for those securities with heavy put option trading. More precisely, we should find a positive correlation between trading volume and the reduction in the Friday-Monday differential.
Across the securities in the 1977 sub-sample, the correlation coefficient between put option trading volume and the change in the market-adjusted Friday-Monday differential is -0.12 with a p-value of 0.61. (Similar results are found when we used raw returns.) There is no statistically significant correlation between put option trading volume and the change in the weekend effect. Indeed, the negative correlation implies that for the sample those securities with lower trading volume experienced a higher reduction in the weekend effect on a relative basis. Thus, we dismiss the conjecture that low trading volume explains our findings that put option listings do not impact the Friday Monday differential.
Shifts in Aggregate Trading Volume of Stocks
The Chen and Singal hypothesis explains the weekend effect as a result of order imbalance caused by short sellers and explains the demise of the weekend effect as resulting from the end of the trading imbalance. If the presumed shift in trading volume was sufficient to cause a change in weekday return patterns, we should be able to empirically observe the change in trading volume patterns. Thus, we test for a relative reduction in Friday and Monday volume after the introduction of listed put options. Support for the Chen and Singal hypothesis would result if the trading volume of both Monday and Friday were reduced relative to the trading volume of other weekdays. For sake of completeness, we also test for changes in weekday returns for the market index over the comparable period and examine if these shifts correspond to the predictions of the Chen and Singal hypothesis. Neither of these examinations provides support for the Chen and Singal hypothesis.
We gather daily composite trading volume and the value-weighted composite index return for the New York Stock Exchange from its website (www.nyse.com) for the period 1970 through 2005. We choose to track NYSE return and trading volume (18) as the shift in the weekday effect has been found for large-firm securities that are generally listed on the NYSE. We run the following OLS regression to measure weekday pattern in relative trading volume and returns over time:
[R.sub.t] = [[gamma].sub.0] + [[gamma].sub.1] * [D.sub.tu] + [[gamma].sub.2] * [D.sub.w] + [[gamma].sub.3] * [D.sub.th] + [[gamma].sub.4] * [D.sub.F]+ [[epsilon].sub.[tau] (1)
[V.sub.t] = [[gamma].sub.0] + [[gamma].sub.1] * [D.sub.tu] + [[gamma].sub.2] * [D.sub.w] + [[gamma].sub.3] * [D.sub.th] + [[gamma].sub.4] * [D.sub.F]+ [[epsilon].sub.[tau] (2)
[R.sub.t] represents the daily trading return for the value-weighted NYSE composite index.
[V.sub.t] represents the daily trading volume for the NYSE.
[D.sub.tu] is a dummy variable that takes on a value of 1 if the trading day is Tuesday and takes on a value of 0 otherwise.
[D.sub.w] is a dummy variable that takes on a value of 1 if the trading day is Wednesday and takes on a value of 0 otherwise.
[D.sub.th] is a dummy variable that takes on a value of 1 if the trading day is Thursday and takes on a value of 0 otherwise.
[D.sub.F] is a dummy variable that takes on a value of 1 if the trading day is Friday and takes on a value of 0 otherwise.
The intercept provides average Monday return (or average Monday volume for equation 2) and the slope coefficients [[gamma].sub.] through [[gamma].sub.4] measure the deviation of the mean for other weekdays relative to Monday.
To compensate for general shifts in trading volume and to examine shifts in trading volume and returns as predicted by the Chen and Singal hypothesis, we divide our data into four time periods: the 1970s, 1980s, 1990s and 2000-2005.
As reported in Table 2, the results from estimating regression (1) confirm the existence of a weekend effect in returns for the early years of the sample and the disappearance of the weekend effect in the later years of the sample period. In the period 1970 through 1979 the average Monday return is negative and significantly less than the return for any other weekday and the average Friday return is higher than the average return for any other day. This return pattern is consistent with buying pressure on Friday and selling pressure on Monday predicted by the Chen and Singal hypothesis. During the 1980s the Monday returns remain significantly lower than all other weekdays and the weekend effect continues despite the widespread availability of listed put options. The weekend effect does disappear in the decade of the nineties and is also absent in the first six years of the Twenty-First Century. However, the shifts are a bit tardy to be ascribed to an impact from listings of put options, and are also too large. In both the latter two periods Monday's average return is the largest of all weekdays. The Chen and Singal hypothesis does not predict unusually high returns on Monday. Nor does the hypothesis explain the shift in Friday's returns which becomes the lowest of any of the weekday returns for the last sample period. Clearly the shift in the weekday return pattern is driven by something other than the actions of bearish investors substituting put options for short selling. (19) Examination of weekday volume patterns across time strengthens this conclusion.
Evidence from estimating equation (2) directly conflicts with the Chen and Singal hypothesis attribution of the weekday return pattern to unusual buying activity on Friday and unusual selling activity on Monday by risk-averse bearish investors. In the decade before listed put option trading, trading volume is smallest on Monday of all weekdays. Indeed, trading volume on Monday is significantly less than any other weekday. Furthermore, Friday has the second lowest trading volume. Albeit, these results may obscure the importance of the trading by risk-averse bearish investors on light trading days at the start of the week and at the end of the week. However, there is no shift in the trading patterns that would explain the observed shifts in returns. Despite important changes in the weekday return patterns, the pattern in trading volume remains unchanged across sample periods. In each of the four sample periods Monday has the lowest average volume, significantly lower than other weekdays and Friday has the next lowest average volume. Nothing in this pattern suggests a shift in weekday order flow that is large enough to cause the observe shift in the weekday pattern or that is even consistent with the shift in observed weekday patterns. Both daily volume and daily return patterns for NYSE securities contradict the Chen and Singal hypothesis.
Explanations for both the negative Monday returns and their reversal for large-firm securities are plentiful but generally incomplete. Recently, Chen and Singal (2003) offer a comprehensive explanation for both the negative returns on Monday and the reversal of this pattern for large-firm securities. The explanation is appealing both in its completeness and its reliance on the rational behavior of market participants. The weekend effect according to this argument occurs because risk-averse bearish investors avoid exposure to risk over the weekend as they close out short position. The weekend effect disappears as the listed put options allow risk-averse bearish investors a less risky alternative that no longer requires them to close their position over the weekend. Market behavior in this explanation remains totally rational and the seemingly anomalous seasonal pattern is dependent entirely on market friction.
The explanation, however, does not withstand close scrutiny. The predicted behavior of market participants depends on a risk comparison between put option buying and short selling that is not carefully defended. Indeed we develop arguments that indicate that put buying not short selling is the more risky alternative for risk-averse bearish investors. We support this conclusion with evidence from a review of the financial press and a survey of bearish market participants that indicate that bearish inverse still rely primarily on short selling to implement their strategies. Further, we show that on a security-by-security basis put option listing does not affect the existence of the weekend pattern in security returns. Our findings along with other recent critical examination of this hypothesis cast serious doubts on its validity.
These findings beg the question then as to what caused and reversed the tendency for Monday returns to be significantly negative. Behavioral explanations of the Monday effect argue that investor behavior affected as it is by sentiment will tend to create low returns on Monday simply because investor behavior is affected by "Blue Mondays." The disappearance of this effect during the 1990s may be explained by the generally buoyant attitude of investors during that period. It may even be reconcile with Monday returns being higher than other weekdays, if investors tend to make investment decisions over the weekend. This view, however, is not easily reconciled with the relatively high Monday returns that occur in the bear market following the Dot.com crash.
Another explanation consistent with rational investor behavior in a market impeded by friction offers, perhaps, a more viable explanation for the disappearance of the Monday effect. According to Chan, Leung and Wang (2004) the rise of institutional investors causes the disappearance of the weekend effect as they supplant the activity of individual investors. Consistent with this argument, Gu (2004) links the disappearance of negative Monday returns to the rise of more sophisticated investors creating a more efficient market. Gu further suggests that the reversal of the weekend effect may represent over trading on the part of investors attempting to exploit the Monday effect. Both of these explanations would be more appealing if they offered suggestions as to the original cause of the Monday effect. The behavior of Monday returns in future years may allow researchers to more fully address this question.
List of References from The Wall Street Journal
References Indicating Various Types of Professional Investors Engaging in Short Selling
Anand, S., Dec 27, 2007. 'Long-Short' Funds Labor to Thrive; Smattering Are in Red For Year, Despite Gains in Dow Industrials, S&P, The Wall Street Journal 210, C.1.
Anand, S., August 7, 2007. Long-Shorts Fall ... Short, The Wall Street Journal 210, C.1.
Bogle., J., Feb 9, 2007. 'Value' Strategies," The Wall Street Journal 209, A.11.
Cui, C., Nov 27, 2007. How Social Conscience Hooks Hedge Funds; Push for Human Rights, Environmental Standards Attracts More Big Investors, The Wall Street Journal 210, C.13.
Gongloff, M., May 19, 2008. Ahead of the Tape, The Wall Street Journal 211, C.1.
Gullapalli, D., Jan 19, 2008. Green Thumb: Cash-Rich Mutual Funds May Battle Bear, The Wall Street Journal 211, B.1.
Laise, E., Sep 24, 2008. Short-Sale Ban is Hitting Mutual Funds; Some Managers Say They Aim to Avoid Big Cash Amounts, The Wall Street Journal 212, C.3.
Lauricella, T. and D. Gullapalli, Mar 17, 2007. Shorting Out: Fast-Money Crowd Embraces ETFs, Adding Risk for Individual Investors, The Wall Street Journal 209, A.1.
Lobb, A., Sep 25, 2008. Short Selling: Bearish Bets Firmed Ahead of Ban, The Wall Street Journal 212, C.8.
Salisbury, I., Aug 31, 2007. Why "Quant" Strategy Has Ripple Effects; Managers Say Popularity Of Growth-Oriented Stock Fueled a Dangerous Cycle, The Wall Street Journal 210, C.11.
Whitehose, K., August 11-12, 2007. How Stocks' Unexpected Moves Had 'Quant' Managers on Run, The Wall Street Journal, 210, C.3.
References Indicating that Short-Selling Activity Has a Large Market Impact
Browning, E.S., June 14, 2007. Dow Rebounds, Climbing By 187.34 Points, The Wall Street Journal 209, C.1.
Browning, E., Sep 10, 2007. New Rules For Picking A Bottom? The Wall Street Journal 210, C.1.
Brulliard, N., Mar 5, 2008. Stock Sales at Google Send Shivers, The Wall Street Journal 211, C.3.
Esterl, M. and G. Zuckerman, Oct 30, 2008. Porsche Move Helps Send VW Shares Back to Earth, The Wall Street Journal 212, C.1.
Greenberg, H., Nov 10, 2007. The Buzz-Market Watch Weekend Investor: How a Bear Paws at Mr. Coffee, The Wall Street Journal 210, B.3.
Jenkins, H., Jun 11, 2008. Jr. Business World: How a short Sale Went Wrong, The Wall Street Journal 211, A.21.
Lahart, J., May 30, 2007. Short Story: Bearish Bets Lose Bullish Bias, The Wall Street Journal, 209, C.1.
Lahart, J., August 11-12, 2007. Why the Weak Rise, the Strong Fall, The Wall Street Journal, 210, C.1.
Richardson, K., and S. Ng, Moody's July 10, 2007. Faces the Storm, The Wall Street Journal, 210, C.1.
Zuckerman, G., J. Strasburg and M. Esterl. VW's 348% Two-Day Gain Is Pain For Hedge Funds, The Wall Street Journal Oct 29, 2008 212, C.1.
References Discussing Short-Selling Activity and the 2008 Financial Crisis
Corrections & Amplifications, Jul 23, 2008. The Wall Street Journal 212, A.2.
For Every Short Sale, There Is Commitment to Buy, Sep 25, 2008. The Wall Street Journal 212, A.19.
Short on Common Sense, Sep 25, 2008. The Wall Street Journal 212, A.20.
The Afternoon Report: Carrying On; Online Edition, Jul 15, 2008. The Wall Street Journal 212.
Bums, J., Aug 13, 2008. Stock Under 'Short' Order Fell During Protection Period, The Wall Street Journal 212, C.6.
Chanos, J., Sep 22, 2008. Short Sellers Keep the Market Honest, The Wall Street Journal 212, A.23.
Cox, C., Jul 24, 2008. What the SEC Really Did on Short Selling, The Wall Street Journal 212, A.15.
Gordon Crovitz, L., Sep 22, 2008. Information Age: Information Haves and Have-Nots, The Wall Street Journal 212, A.21.
Karnitschnig, M., and S. Craig, Aug 21, 2008. Fed Acted on Lehman Rumor; Amid Market Speculation in July, Credit Suisse Was Urged to Maintain Ties, The Wall Street Journal 212, C.1.
Mollenkamp, C., A. MacDonald and A. Lucchetti, Mar 20, 2008. True or False? Rumor Mill Banks in a Tough Spot, The Wall Street Journal 211, C.1.
Reilly, D., T. Barker, L. Denning and M. Curtin, Sep 22, 2008. Heard on the Street / Financial Analysis and Commentary, The Wall Street Journal 212, C.10.
Rogow, G., Jul 22, 2008. Market Beat / Market Insight from WSJ.com, The Wall Street Journal 212, C.5.
Rogow, G., Aug 14, 2008. Stone Energy, Olympic Steel Walk Tall; Tween Brands Thumbles, The Wall Street Journal 212, C.5.
Scannell, K., and G. Zuckerman, Apr 25, 2008. See Accuses Ex-Trader of Blackstone Ruse, The Wall Street Journal 211, C.1.
Scannell, K., Sep 23, 2008. The Financial Crisis: SEC Quickly Revises Short-Selling Rules; Shift on Financials, Hedge Sends Traders Scrambling, The Wall Street Journal 212, A.3.
Strasburg, J., K. Scannell and R. Smith, Jul 28 2008. SEC Intensifies Efforts to Rein In Short Selling; Wall Street Readies For Longer Limits; Are Curbs Working?, The Wall Street Journal 212, C.1.
Tracy, T., Jul 17, 2008. SEC's Limit on Short-Selling May Exempt Market Makers, The Wall Street Journal 212, C.5.
Wilson, A., in Melbourne and L. Santini in Hong Kong, Sep 4, 2008. International Finance: Fortescue's Short Story; Harbinger, Caught Unaware, Seems Stock Source for Bears, The Wall Street Journal 212, C.2.
References Discussing Academic Research on Short Selling and Discussing the Substitution of Put Options for Short Selling during the Period that Short Selling Was Banned
Burns, J., Jul 22, 2008. Short Selling Jumps Ahead of CEO Sales, The Wall Street Journal 212, C.2.
Jenkins, H. W., Jul 23, 2008. Jr. Business World; Washington Loves Bank Investors, The Wall Street Journal 212, A.15.
Scannell, K., and G. Zuckerman, Sep 16, 2008. Crisis on Wall Street: SEC to Trackle Short Sales; As a Crisis Unfolds, New Urgency Is Seen In Shoring Up Rules, The Wall Street Journal 212, C.6.
Tracy, T., Aug 13, 2008. As Short-Sale Rule Expires, Some Trading Likely to Fall Off, The Wall Street Journal 212, C.6.
Zuckerman, G., and K. Scannell, Sep 20, 2008. Short Selling Ban May Have Loopholes, The Wall Street Journal 212, B.1.
References to the Importance of Short-Selling in International Markets
Pakistan Engineers Rally; Index Surges 8.6% on 'Short' Freeze, Daily Loss Limit, Jun 25, 2008. The Wall Street Journal 211, C.2.
Batson, A., Aug 25, 2007. Shorting China: There's a Will, But Which Way? The Wall Street Journal 210, B.1.
Scannell, K., Sep 22, 2008. Short-Sale Ban Spreads Around Globe; Australia, Taiwan, Netherlands Join Push To Keep Investors From Betting on Decline, The Wall Street Journal 212, C.1.
References Discussing Record Levels of Short Selling
Lobb, A., Jun 20, 2008. Short Selling on NYSE Sets Record, The Wall Street Journal 211, C.7.
McKay, P., Apr 25, 2007. Short Selling: Bearish Bets Hit a Record on Nasdaq; Positions Topped Eight Billion Shares In Mid-April Period, The Wall Street Journal 209, C.6.
McKay, P., Apr 5 2008. NYSE and Amex Short-Selling: Bearish Bets on the NYSE Hit a Record, The Wall Street Journal 211, B.6.
References Citing the Use of Put Options as Insurance on Bullish Strategies
Ball, Y., Aug 8, 2007. Cisco Sees Active Trading Ahead of Earnings Report, The Wall Street Journal 210, C.6.
Blumenthal, K., Jun 18, 2008. When Protecting An Investment Raises Risks, The Wall Street Journal 211, D.1.
Hadi, M., May 26, 2007. Glaxo Drug Concerns Fuel Activity in Puts, The Wall Street Journal 209, B.5.
Hadi, M., Jun 2, 2007. Investors Buy Puts to Protect Against Pullback, The Wall Street Journal 209, B.14.
Hadi, M., July 5, 2007. Merger Arbitragers Use Options For Protection as Spreads Widen, The Wall Street Journal 210, C.5.
Hadi, M., May 7, 2007. Microsoft Speculation Ignites Yahoo Trading, The Wall Street Journal 209, C.4.
Hadi, M., Jul 10, 2007. National City, Comerica See Heavy Put Trade, The Wall Street Journal 210, C.6.
Hadi, M., Apr 14, 2007. Put-Option Strategy May Be Salve For Investors Nervous About Bears, The Wall Street Journal 209, B.5.
Hadi, M., Jul 11, 2007. Sears' Woes Traders Seeking Cover, The Wall Street Journal 210, C.6.
Hadi, M., Apr 11, 2007. Washington Mutual Reflects Bearish Bias; Puts Trade Heavily on Earnings Fears In Mortgage Sector, The Wall Street Journal 209, C.8.
Hadi, M., and Y. Ball, Jul 27, 2007. Cost of Insuring Portfolio Soars as Stocks Slump, The Wall Street Journa1 210, C.2.
Karmin, C., Aug 30, 2008. Small Kresge Foundation Teaches Big Lessons in Investment Strategy, The Wall Street Journal 212, B.1.
Laise, E., Jul 15, 2008. Market's Swoon Prods Investors To Try Options, The Wall Street Journal 212, D.1.
Laise, E., Apr 4, 2007. Mutual Funds Add Exotic Fare to the Mix; Many Are Making Increasing Use of Complex Derivatives To Enhance Performance, but the Strategies Carry Risks, The Wall Street Journal 209, D.1.
McKay, P. A., Jun 16, 2007. How to Hedge Wisely; Amid Higher Volatility, Investors Seek Shelter In Options, New ETFs, The Wall Street Journal 209, B.1.
Zuckerman, G., Oct 17, 2007. Lessons From Black Monday: The More Hedges the Better, Right?; Blind Spots, Bottlenecks Lurk Among New Ways Of Preventing a Crash, The Wall Street Journal 210, C.1.
References to Selling Put Options as a Bullish Strategy
Hadi, M., Feb 9, 2007. Harrah's Trading Is Preplexing, The Wall Street Journal 209, C.2.
Hadi, M., Feb 7, 2007. Selling Puts on Lear Is a Way to Play, The Wall Street Journal 209, C.6.
Hadi, M., Jun 6, 2007. Takeover Buzz Creates Thirst For Anheuser, The Wall Street Journal 209, C.5.
Hadi, M., Apr 10, 2007. Trading in BlackBerry Maker Is Heavy Ahead of Earnings, The Wall Street Journal 209, C.3.
Ovide, S., and M. Hadi Feb 21, 2007. OfficeMax Puts And H-P Calls See Brisk Trade, The Wall Street Journal 209, C.5.
Smith, R., Jul 17, 2008. Street Gears Up for Short Changes; Brokerages Parse New Rules' Details; The 'Naked' No-No, The Wall Street Journal 212, C.1.
References to Bearish Strategies Using Put Options for Individual Securities
The Buzz: Best of WSJ.com's Money Blogs / From Deal Journal, Market Beat and Wealth Report, May 24, 2008. The Wall Street Journal 211, B.3.
Ball, Y., Oct 11, 2007. Boeing Warning On Jet Deliveries Triggers Trading, The Wall Street Journal 210, C.5.
Ball, Y., Sep 19, 2007. Rate Cut, Lehman Earnings Spur Traders to take Action, The Wall Street Journal 210, C.6.
Ball, Y., Oct 26, 2007. Trading Swings From Puts to Calls On AIG Worries, The Wall Street Journal 210, C.5.
Curran, R., Feb 15, 2008. Bearish Traders Hedge Bets As Financial Worries Return, The Wall Street Journal 211, C.4.
Curran, R., Sep 15, 2008. Traders in Lehman, AIG Held Out Hope (Friday), The Wall Street Journal 212, C.3.
Curran, R., Jan 10, 2008. Options Bears At Countrywide, The Wall Street Journal 211, C.5.
Curran, R., and Y. Ball Nov 20, 2007. GM Puts Are Active Amid Sales Worries, The Wall Street Journal 210, C.5
Hadi, M., Jul 19, 2007. Puts on Lehman Trade Heavily On Mortgage-Relayed Worries, The Wall Street Journal 210, C.4.
Hughes, S., Sop 29, 2007. SEC Files Insider Charges Tied to Trading in Dell, The Wall Street Journal 210, A.10.
Scannell, K., Mar 20, 2008. Credit Crisis: SEC's Bear Stearns Probe Zeroes In 'Put' Trades; Options to Sell Stock Spiked Before Fall; 'Very Unusual' Bets, The Wall Street Journal 211, C.2.
Tracy, T., Apr 21, 2008. Moving The Market--Option Report; Traders Got Pummeled as Bearish Bets Fell Flat, The Wall Street Journal 211, C.2.
Winstein, K. J., May 2, 2008. Herbalife President Resigns; Departure Comes After False Claims of M.B.A. Degree, The Wall Street Journal 211, B.5.
Other References to Put Option Strategy including both Bullish and Bearish Strategies
Ball, Y., Sep 18, 2007. Lehman Attracts Earnings-Eve Puts, The Wall Street Journal 210, C.6.
Laise, E., Mar 11, 2007. How to Play a Shaky Market Like a Pro, The Wall Street Journal 210, 1.
Substitution of Put Options for Short Sales because of Restrictions on Naked Short Sales
Smith, R., J. Strasburg and K. Scannell Jul 17, 2008. Street Gears Up for Short Changes; Brokerages Parse New Rules' Details; The 'Naked' No-No, The Wall Street Journal 212, C.1.
The Use of Put Options by Nassim Taleb
Patterson, S., July 13, 2007. Mr. Volatility and the Swan, The Wall Street Journal, 210, C1.
Questions Asked of Practitioners
1. Are you mainly a bearish investor?
2. Do you use short sells or put options? If you use both what is the split?
3. Do you consider one a substitute for the other?
4. Do you consider puts to be less risky than short selling?
5. Do you close short positions on Friday and re-open on Mondays?
6. Do you hedge short positions?
Brooks R. and H. Kim, 1997. The Individual Investor and the Weekend Effect: A Reexamination with Intraday Data, Quarterly Review of Economics and Finance 37, 725-737.
Blau, B., B. Van Ness and R. Van Ness, 2009. Short Selling and the Weekend Effect for NYSE Securities, Financial Management (forthcoming).
Brusa, J., P. Liu and C. Schulman, 2000. The Weekend Effect, "Reverse" Weekend Effect, and Firm Size, Journal of Business Finance and Accounting 27, 555-574.
Chan, H., W. Leung and K. Wang, 2004. The Impact of Institutional Investors on the Monday Seasonal, Journal of Business 77, 967-986.
Chen, H. and V. Singal, 2003. Role of Speculative Short Sales in Price Formation: Case of the Weekend Effect, Journal of Finance 58, 685-705.
Christophe, S., M. Ferri and J. Angel, 2009. Short Selling and the Weekend Effect in Nasdaq Stock Returns, The Financial Review 44, 31-57.
Fields, M., 1931. Stock Price: A Problem in Verification, Journal of Business 4, 414-418.
Fredrickson, B., and D. Kahneman, 1993. Duration Neglect in Retrospective Evaluation of Affective Episodes, Journal of Personality and Social Psychology 65, 45-55.
Ma. T., P. Gao and I. Kalcheva, 2008. Short Sales and the Weekend Effect--Evidence from Hong Kong, Working Paper, State University of New York at Binghamton.
Gondhalekar, V. and S. Mehdian, 2003. The Blue-Monday Hypothesis: Evidence Based on Nasdaq Stocks, 1971-2000, Quarterly Journal of Business and Economics 52, 73-90.
Gu, A., 2004. The Reversing Weekend Effect: Evidences from U.S. Equity Markets, Review of Quantitative Finance and Accounting 22, 5-14.
Harris, L., 1986. A Transaction Data Study of Weekly and Intra Daily Patterns in Stock Returns, Journal of Financial Economics 16, 469-487.
Kamara, A., 1997. New Evidence on the Monday Seasonal in Stock Returns," Journal of Business 70, 63-84.
Kelly, F., 1930. Why You Win or Lose: The Psychology of Speculation, (Houghton Mifflin, Boston).
Lakonishok, J. and E. Maberly. 1990, The Weekend Effect: Trading Patterns of Individual and Institutional Investors, Journal of Finance 45, 231-243.
Mehdian, S. and M. Perry, 2001. The Reversal of the Monday Effect: New Evidence from U. S. Equity Markets, Journal of Business Finance and Accounting 28, 1043-1065.
Miller, E., 1988. Why a Weekend Effect? Journal of Portfolio Management 14, 43-48.
Pettengill, G., J. Wingender, and R. Kohli, 2003. Arbitrage, Institutional Investors and the Monday Effect, Quarterly Journal of Business and Economics 42, 49-64.
Rogalski, R., 1984. New Findings Regarding Day-of-the-Week Returns Over Trading and Non-Trading Periods: A Note, Journal of Finance 39, 1603-1614.
Rystrom, D. and E. Benson, 1989. Investor Psychology and the Day-of-the-Week Effect, Financial Analysts Journal 45, 75-78.
Sias, R. and L. Starks, 1995. The Day-of-the-Week Anomaly: The Role of Institutional Investors, Financial Analysts Journal 51, 58-67.
White, H., 1980. A Hetroskedasticity-Consistent Covariance Matrix Estimator and a Direct Test of Hetroskedasticity, Econometrica 48, 817-838.
Glenn Pettengill *
Grand Valley State University
Grand Valley State University
(4) Miller (1988), Lakonishok and Maberly (1990) and Kelly (1930) argue that weekend investment decisions by individual investors result in sell orders coming to the market on Mondays. Rystrom and Benson (1989) and Gondhalekar and Mehdian (2003) argue that investors are inherently more pessimistic on Mondays than other weekdays resulting in more sell orders and less buy orders on Mondays. Brooks and Kim (1997) and Sias and Starks (1995) argue that institutional investors withdraw liquidity on Mondays fearing trades with informed investors.
(5) The literature on variations in weekday returns has centered on low Monday returns, although significant variation in returns have occurred for other weekdays. For small-firm securities Friday returns have been historically high, but for large-firm securities Wednesday returns have historically been higher than Friday returns. Ironically, Chen and Singal's explanation for the shift in returns would be more applicable to small-firm securities than to large-firm securities in terms of Friday's returns. The literature documenting the change in weekday returns has centered on the shift of Monday returns from negative to positive. Because Chen and Singal's argument refers to both Friday and Monday returns, we refer to the weekend effect as we examine their hypothesis.
(6) Chen and Singal (2003) are not the first authors to suggest a connection between variation in weekday returns and trading pressure from speculators closing out stock positions. Fields (1931) made a similar argument. However, he postulated that Saturday returns would be low due to day traders closing out long positions. His empirical investigation found instead low returns on Monday.
(7) Additional research following the publication of this study has presented evidence to argue against the proposition that short-selling activity caused the weekend effect in U. S. equity markets. Gao, Kalcheva and Ma (2009) show the existence of a weekend effect in the Hong Kong market during a period when short selling was prohibited, casting doubt on the suggestion that short selling causes the weekend effect in the U. S. equity market. Blau, Van Ness and Van Ness (2009) find that short selling is more abundant during the middle of the week rather than on Monday, casting doubt on the Chen and Singal hypothesis that heavy short selling on Monday causes low Monday returns. Finally, Christophe, Ferri and Angel (2009) find insufficient speculative short selling on Fridays and Mondays to explain the weekend effect.
(8) In the interest of brevity we do not provide citations for the individual articles referenced in this section. In the interest of academic integrity we provide a full list of the articles reviewed from The Wall Street Journal in Appendix I. The articles are grouped by topic in roughly the order of presentation.
(9) We are indebted to Clay McDaniel of D&L Partners of Fort Lauderdale, Florida for providing us a list of 25 bearish investors. We contacted all 25 via telephone and 16 agreed to respond.
(10) Given that stock prices generally trend upward over the long run, one may argue that puts purchased at the money should more frequently than not result in a 100% loss. On the other hand, one might argue that rational put buyers more frequently purchase puts on stocks that will decrease rather than increase. However, to infer this rational behavior for put buyers, one must infer that put writers irrationally write puts more frequently on stocks that will decline rather than increase in price.
(11) Mehdian and Perry (1987), Brusa, Liu and Schulman (2000), Pettengill, Wingender and Kohli (2003) and Gu (2004) all document that for large stocks Monday returns in recent years are positive and higher than on other weekdays while the returns on small stocks are negative and lower than on other weekdays. The latter study argues against the hypothesis that arbitrage activity of institutional investors eliminated the Monday effect on the basis that the returns on Monday became significantly higher than returns for some other weekdays. The Chen and Singal hypothesis suffers a similar question in that the significantly greater return for Monday relative to other weekdays in recent years requires a hypothesis to explain why, after the short selling influence was eliminated, a significantly positive Monday effect would appeared. The Chen and Singal hypothesis is also challenged by empirical evidence indicating a reduction in the Monday effect for small-firm securities. In a recent study, Gondhalekar and Mehdian (2003) document that Monday returns on Nasdaq stocks for the period 1990-1994 although not negative are lower than on some of the other weekdays, but for the period 1995-2000 the Monday returns are not only non-negative but also not different from the returns on other weekdays.
(12) This relationship is especially strong for IPO securities where arbitrage short selling is unlikely. Although, this evidence supports a connection between short selling and a weekend effect, the results tend to contradict the Chen-Singal hypothesis because the empirical analysis showing a connection between short selling and a Monday effect was conducted during the period 1988 through 1999, a period after put options were introduced.
(13) During this period several exchanges added some put option listings for securities that were listed on other exchanges. These additional listings were not included in our sample.
(14) The Journal only lists those options that trade on a particular day. This date may not be the exact listing date. Causal empiricism suggests that once the put option is initially listed trading occurs on the vast majority of the days. If an option were listed some time before any trade takes place, we would prefer the sample date to correspond with the advent of trading because it is the trading that would provide the impact cited by Chen and Singal. Toward the end of 1981 the Journal began to report all listings regardless of whether a trade took place or not. And we then used that date as the first listing day.
(15) Measuring the weekend effect using the paired Friday minus Monday returns eliminates holiday weekends where the market was closed either on Monday or Friday. Separately we tested for a shift in the weekend effect including Friday minus Tuesday returns for Monday holiday closings and Thursday minus Monday returns for Friday holiday closings. We find no material difference using these alternative definitions of the weekend effect. These results are available from the authors. The difference reported in Table 1 for the Friday minus Monday differential does not match the difference in the Friday differential and Monday differential because the Friday minus Monday differentials do not include Fridays preceding Tuesdays and Mondays following Thursdays. The Friday and Monday differentials include all Friday and all Monday trading days.
(16) Although we feel that it is important to test for this possibility, we note that Chen and Singal indicate an immediate impact of put options on the weekend effect. They state (page 686) "Thus, migration of speculative short sellers to the options market should coincide with a reduction in the weekend effect. Indeed, we find that the weekend effect weakens significantly after 1977 and disappears in the 1990s ..."
(17) As the only source available for put option volume was daily copies of The Wall Street Journal the labor cost of collecting the data was extensive. Hence, we collected data once a week from Friday's trading activity. We have no reason to presume that over the three-year period our collection procedures would bias the relative volume across securities.
(18) We conduct a similar study including Nasdaq data and come to identical conclusions.
(19) Chen and Singal's hypothesis argues for an equal impact on Friday and Monday returns. Observed differences in the weekday effect result primarily from a relative increase in Monday returns rather than an equivalent leveling of both Monday and Friday returns relative to other days of the week. The uneven change observed in Friday and Monday returns, in of itself, provides a challenge to the Chen and Singal hypothesis.
* Corresponding author: Seidman College of Business, Grand Valley State University, 438C DeVos, 401 W. Fulton Street, Grand Rapids, MI; Tel: 616.331.7430; Fax: 616.331.7445; email@example.com
Table 1--Weekend Effect Before and After the Listing of Put Options on Stocks The sample covers all the stocks on which put options were introduced between 1977 and 1981 (the SEC allowed put options for the first time in June 1977 and allowed additional listings starting May 1980). The three-year period before (after) the date of put listing is taken as the pre (post) listing period. The weekend effect is taken to be the Friday return minus the return on the following Monday. The standard pair-wise parametric t-test provides the significance for the difference between the pre minus post returns. Alphabets a and b represent statistical significance at the l% and 5% levels respectively. Panel A Weekend Effect Fri-Mon Fri-Mon Fri-Mon (Pre Put Listings) (Post Put Listings) (Pre-Post) Mean Return (%) 0.0032 0.0028 0.0004 T-Statistic 15.6 (a) 17.31 (a) 1.72 (c) Sample Size 244 244 244 Panel B Monday Return Pre Put Listing Post Put Listing Pre--Post Mean Return (%) -0.0010 -0.0012 0.0001 T-Statistic -8.48 (a) -9.57 (a) 1.04 Sample Size 244 244 244 Panel C Friday Return Pre Put Listing Post Put Listing Pre--Post Mean Return (%) 0.0025 0.0015 0.0010 T-Statistic 19.82 (a) 13.34 (a) 6.41 (a) Sample Size 244 244 244 Panel D Weekend effect (sub-periods) Fri-Mon (Pre Put Fri-Mon Fri-Mon Listings) (Post Put Listings) (Pre-Post) 1977-1980 Mean Return (%) 0.0031 0.0030 0.0001 T-statistic 12.68 (a) 15.13 (a) 0.40 Sample Size 166 166 166 1981 Mean Return (%) 0.0034 0.0024 0.0010 T-Statistic 9.06 (a) 8.62 (a) 2.57 (a) Sample Size 78 78 78 Table 2--NYSE Composite Return and Volume: Day of the Week Effect The daily return on the NYSE composite index and the daily NYSE volume (number of shares traded) are regressed against zero-one dummy variables for days of the week other than Monday. The intercept provides the mean for Monday and the slopes provide the difference between the mean for Monday and the mean for other days of the week. The t-statistics (in parenthesis) are adjusted for hetroskedasticity according to White (1980). Alphabets a and b represent statistical significance at the 1% and 5% levels respectively. The daily return is in percentage and the share volume is in millions of shares. Panel A 1970-1979 Intercept Tue Wed Thu Composite -0.13 0.11 0.20 0.18 Return (-3.1) (a) (3 .0) a (3.5) (a) (3.3) (a) Volume 17.94 1.84 2.42 2.13 (ml shares) (49.2) (a) (3.6) (a) (4.8) (a) (4.2) (a) Panel B 1980-1989 Intercept Tue Wed Thu Composite -0.12 0.21 0.27 0.16 Return (-2.7) (a) (3.4) (a) (4.2) (a) (2.6) (a) Volume 100.12 10.43 14.43 13.45 (ml shares) (38.8) (a) (2.9) (a) (4.0) (a) (3.7) (a) Panel C 1990-1999 Intercept Tue Wed Thu Composite 0.08 -0.03 -0.01 -0.11 Return (2.1) (b) (-0.3) (-0.2) (-1 .8) (b) Volume 349.57 43.51 53.62 48.40 (ml shares) (34.8) (a) (3.l) (a) (3.8) (a) (3.4) (a) Panel D 2000-2005 Intercept Tue Wed Thu Composite 0.07 -0.07 -0.09 -0.005 Return (1.1) (-0.8) (-1.0) (-0.06) Volume 1256.8 103.36 165.04 163.67 (ml shares) (76.1) (a) (4.7) (a) (7.1) (a) (7 .0) (a) Panel A 1970-1979 Fri Sample Adj. [R.sup.2] F-Stat Size (%) Composite 0.21 2523 0.67 5.2 (a) Return (3.6)8 Volume 1.22 2523 0.96 7.1 (a) (ml shares) (2.4) (b) Panel B 1980-1989 Fri Sample Adj. [R.sup.2] F-Stat Size (%) Composite 0.22 2526 0.67 5.3 (a) Return (3.5) (a) Volume 9.97 2526 0.63 5.0 (a) (ml shares) (2.8) (a) Panel C 1990-1999 Fri Sample Adj. [R.sup.2] F-Stat Size (%) Composite -0.02 2528 0.06 1.4 Return (-0.4) Volume 33.61 2528 0.57 4.6 (a) (ml shares) (2.4) (6) Panel D 2000-2005 Fri Sample Adj. [R.sup.2] F-Stat Size (%) Composite -0.09 1508 0.00 0.53 Return (-1.0) Volume 94.26 1508 3.47 14.5 (a) (ml shares) (3.4) (a)
|Printer friendly Cite/link Email Feedback|
|Author:||Pettengill, Glenn; Gondhalekar, Vijay; Wingender, John|
|Publication:||Quarterly Journal of Finance and Accounting|
|Date:||Jan 1, 2011|
|Previous Article:||Problems in performance attribution.|
|Next Article:||Options expected returns: variation by moneyness and maturity.|