Efficiency and the Bear: Short Sales and Markets around the World

Transcription

1 Efficiency and the Bear: Short Sales and Markets around the World Arturo Bris Yale School of Management and ECGI William N. Goetzmann Yale School of Management and NBER Ning Zhu University of California Davis

2 What is a Short Sale S.E.C.: the sale of a security that the seller does not own or that the seller owns but does not deliver. In order to deliver the security to the purchaser, the short seller will borrow the security, typically from a broker-dealer or an institutional investor

3 In the U.S. As of June 2001, the NYSE, NASD, and AMEX reported short interest with a market value in excess of $260 billion, just over 1.7% of total market capitalization (D Avolio, 2002) In February 2004, the NYSE reported that the short interest was equal to 2.0 percent of the total shares on the Exchange.

4 Overview Question 1: Short sales and efficiency. R-Squared Cross-autocorrelation Question 2: Short sales and crashes. Conditional skewness. Probability of crashes Question 3: Arbitrage Opportunities Auto-correlation Momentum and contrarian profit

5 Results Short-sales constraints associated with lower efficiency and slower price discovery at the individual security level No conclusive evidence that short-sales-constrained markets have more negative skewness or greater crash probabilities. Cross-market arbitrage opportunities

6 0.6 Peru Malaysia Colombia 0.4 Pakistan Spain Indonesia Chile Venezuela Greece Japan Portugal Italy Ireland New Zealand South Korea UK Australia Thailand Denmark Sweden Brazil Finland Hungary Luxembourg Norway -0.2 USA Switzerland Zimbabwe Netherlands MalaysiaSingagore Canada Hong Kong South Africa Mexico Thailand Taiwan Turkey Germany Argentina Israel China Norway Poland Belgium Czech Hong Republic Kong Philippines -0.4 Austria Sweden Comovement

7 History in the Netherlands Isaac Le Maire stock scandal short-sales restrictions and They delve into the deepest state and business secrets and do not hesitate to attack even the Government and excite the masses in order to profit more Widows and orphans are seriously hurt by speculators a la baisse. Dutch jurist, Muys van Holys, 1680 s Short-sales taxed 1689

8 Great Britain and France Crash of The Company of the South Seas, , Sir John Barnard s Act. Outlawed futures and short-sales till Napoleon, 1802, 1 year imprisonment for short-sellers.

9 Pujo Commission, 1913 Untermyer: Under what circumstances would you regard short selling as legitimate and proper Sturgis: I should regard it so if there was a panic raging over the country and it was desirable to protect interests which could not be sold. I think it would be a perfectly legitimate thing to do. Untermyer: Let us see about that. If there was a panic raging over the country and a man sold stocks short, would not that simply add to the panic Sturgis: It might. Self preservation is the first law of nature.

10 Fear of Bear Raids European capitalists had supplied much of the cash needed to engineer the greatest bear raid in history. These proverbially open-handed and trusting gentleman had accepted the leadership of New York's adroit Democratic financier, Bernard Baruch, Philadelphia Public Ledger, Short-selling,'is really an expression of opinion, subject to personal risk. It cannot determine value, but only estimate what prospective values really are and will be. Edward Meeker, NYSE Economist, Uptick rule. Had been voluntary from 1931.

11 Data Sources Goldman and Morgan Stanley sources. Direct , telephone and mail contact with regulators. Manuals and published sources. Worldwide Directory of Securities Lending and Repurchase. International Securities Services Association (ISSA) Handbook

12 Short Sales Regulation 45 countries 35 allow shorting 12/ changed in the sample period. 22 allowed it in 1/1990 or before. 10 prohibit shorting over the entire period Malaysia is tricky Allowed in 1996, Prohibited in 8/1997, 2/2001 Allowed, not practiced.

13 Short-Sales Practice Allow and practice short-sales. Australia, Austria, Belgium, Canada, Czech Republic, Denmark, France, Germany, Ireland, Italy, Japan, Luxembourg, Mexico, Netherlands, Portugal, Singapore, South Africa, Switzerland, the U.K., and the U.S. Prohibit short sales and do not practice. China, Colombia, Greece, Indonesia, Pakistan, Peru, South Korea, Venezuela, and Zimbabwe. Allow short sales but it is rarely practiced. Argentina, Brazil, Chile, Finland, Hungary, Israel, New Zealand, the Philippines, Poland, Spain, Taiwan, and Turkey. Practice changed in the sample period. Hong Kong, Norway, Sweden, Malaysia, and Thailand.

14 Foreign Listing and Short Sales Dual-Listed Stocks are shortable in the foreign market even if short sales are prohibited in the domestic market Example: Nokia Short Sales are prohibited in Finland. Nokia accounts for 2/3 of the market capitalization. Nokia is shortable in NYSE since July 2004 We classify stocks within a country as shortable when they are dual-listed in the U.S. or the U.K., even if the domestic country prohibits short sales.

15 All Countries Argentina Finland New Zealand Spain Shortable $158,923 $933,524 $1,991,153 Non-Shortable $407,657 $2,812,947 $3,573,801 Ratio 28.05% 24.92% 35.78% Shortable $810 $102,431 $62,115 Non-Shortable $22,995 $154,631 $61,188 Ratio 3.40% 39.85% 50.38% Shortable $253 $82,214 $571,223 Non-Shortable $53,510 $224,591 $241,469 Ratio 0.47% 26.80% 70.29% Shortable $10,176 $46,808 $21,588 Non-Shortable $20,497 $112,522 $60,186 Ratio 33.18% 29.38% 26.40% Shortable $145,574 $375,643 $562,408 Non-Shortable $291,732 $665,280 $646,334 Ratio 33.29% 36.09% 46.53% Table 2. World Market Capitalization and Short-Sales Restrictions. Countries where Short Sales are Not Allowed / Not Practiced This table classifies the World Market capitalization into shortable and non-shortable, for countries where short sales are not allowed / not practiced. Market Capitalization is in US$ millions. To calculate the numbers in these columns we have taken into account firms in countries where short sales are not allowed / not practiced, that list in markets where short sales are allowed and practiced, in particular the U.S. (NYSE and Nasdaq) and the U.K. (LSE). Pakistan, Venezuela, and Zimbabwe are not included in the table because we do not have data available on dual-listed firms from these countries.

16 The Identification Problem It is difficult to identify the pure effect of regulation on country-wide measures. We only have five countries that have changed restrictions over the sample period. The short-sales dummy is therefore a country-fixed effect Ideally we would like to find both shortable and non-shortable stocks within a country, even if short sales are allowed / prohibited. We did find them.

17 Short Sales and World Market Capitalization 97% 96% 95% 94% 93% 92% 91% 90% 89% 88% 87% ,000,000 30,000,000 25,000,000 20,000,000 15,000,000 10,000,000 5,000,000 0 $ Billion % Shortable Shortable Non-Shortable Bris, Goetzmann, and Zhu, Short Sales in Global Perspective, in Frank Fabozzi, ed., Short Selling: Strategies, Risks, and Rewards, Wiley, 2004.

18 SHORT SALES AND WORLD MARKET CAPITALIZATION (Only Countries with both Shortable and Non-Shortable Stock) 50% 1,400,000 45% 40% 1,200,000 35% 1,000,000 30% 25% 800,000 20% 600,000 15% 400,000 10% 5% 200,000 0% Jan-90 May-90 Sep-90 Jan-91 May-91 Sep-91 Jan-92 May-92 Sep-92 Jan-93 May-93 Sep-93 Jan-94 May-94 Sep-94 Jan-95 May-95 Sep-95 Jan-96 May-96 Sep-96 Jan-97 May-97 Sep-97 Jan-98 May-98 Sep-98 Jan-99 May-99 Sep-99 Jan-00 May-00 Sep-00 Jan-01 May-01 Sep-01 Jan-02 May-02 Sep-02 Ratio Shortable Non-Shortable 0 Bris, Goetzmann, and Zhu, Short Sales in Global Perspective, in Frank Fabozzi, ed., Short Selling: Strategies, Risks, and Rewards, Wiley, 2004.

19 SHORT SALES AND WORLD MARKET RETURN 400% 350% 300% 250% 200% 150% 100% 50% 0% Jul-02 Jan-02 Jul-01 Jan-01 Jul-00 Jan-00 Jul-99 Jan-99 Jul-98 Jan-98 Jul-97 Jan-97 Jul-96 Jan-96 Jul-95 Jan-95 Jul-94 Jan-94 Jul-93 Jan-93 Jul-92 Jan-92 Jul-91 Jan-91 Jul-90 Jan-90 Shortable Non-Shortable

20 In cross-sectional regressions, we regress measures of market efficiency on controls, and three dummy variables: 1. The short sales indicator is a dummy variable that equals one whenever short selling is allowed and practiced in a given country and year, and zero otherwise. 2. ADR0 equals 1 if the observation corresponds to duallisted stocks in countries where short sales are not allowed or not practiced, zero otherwise. 3. ADR1 equals 1 if the observation corresponds to duallisted stocks in countries where short sales are allowed and practiced, zero otherwise.

21

22 Residual Risk Measure of Efficiency R-square in CAPM Improving disclosure, increasing residual risk. Time-series analysis. Co-movement of stock returns GDP explains differences. Quality of government index. Upside vs. downside movement

23 Comovement Index: 1990 to Argentina Belgium Brazil Canada Finland Greece Japan U.S.A. U.K.

24 Short sales constraints impede the market s ability to rapidly impound valuerelevant information. If only price adjustment to bad news is constrained, then idiosyncratic risk will be smaller when market returns are negative. Response to positive/negative factors should be different

25 r mt Our Measures of Efficiency 1. R-Squared We first estimate for every stock and year in our sample, and using weekly data: ijt ij M i mjt W i r r r wt ijt r mjt M W r r r ijt ij i mjt i wt ijt where equals the market return when it is either positive or zero, and r mjt equals the market return when it is negative

26 We then calculate the R-Squared in each regression and aggregate it across stocks in a year, and within the group of shortable / non-shortable stocks, as follows: R 2 jtd 2 ijtd R i i SST SST ijtd ijtd R 2 jtd 2 ijtd R i i SST SST ijtd ijtd Finally, we compute R 2Diff jtd R 2 jtd R 2 jtd

27 MYY Procedure Gain/loss ratio R-square scaled ijt ij M i mjt W i r r r wt ijt

28 Domestic Stocks R-Squared Upside R- Squared Downside R- Squared Difference Downside minus Upside Short Sales Not Allowed / Not Practiced Number of Observations Mean 19.38% 12.73% 14.57% 1.84% Median 16.99% 11.33% 12.14% 1.17% Short Sales Allowed And Practiced Number of Observations Mean 12.87% 9.09% 11.26% 2.17% Median 9.74% 7.28% 8.88% 2.08% Difference (p-value) (0.0000) (0.0000) (0.0000) (0.0053) Whole Sample Number of Observations Mean 16.01% 10.85% 12.85% 2.01% Median 13.35% 9.07% 10.56% 1.73%

29 2. Cross-Autocorrelation We calculate ijtd corr( rijt, rmjt 1) and ijtd corr( rijt, rmjt 1), for all stocks i in group D (={Shortable, Non-Shortable}) and country j, using weekly observations in each year T. We then average the cross-autocorrelations across stocks and calculate: jtd i N ijtd jtd jtd i N ijtd jtd Diff jtd jtd jtd

30 Domestic Stocks Cross- Autocorrelation Upside Cross- Autocorrelation Downside Cross- Autocorrelation Difference Downside minus Upside Short Sales Not Allowed / Not Practiced Number of Observations Mean 3.66% 2.89% 4.51% 1.61% Median 3.48% 3.06% 4.92% 1.79% Short Sales Allowed And Practiced Number of Observations Mean 2.69% 2.03% 3.36% 1.34% Median 2.49% 2.04% 3.10% 1.11% Difference (p-value) (0.0854) (0.9690) (0.0496) (0.6012) Whole Sample Number of Observations Mean 3.16% 2.44% 3.91% 1.47% Median 2.94% 2.44% 3.92% 1.38%

31 Issues and Controls Number of firms GDP Country and industry Herfindahl indices Size Volatility of earnings growth Earnings co-movement Good government LLSV Year and country fixed effects Liquidity: % days with zero return (Bekaert et al., 2003)

32 R-Squared - Overall R-Squared Difference Downside Minus Upside Downside R- Squared Upside R- Squared Short Sales Allowed and Practiced *** * *** ** [3.64] [1.86] [3.05] [2.57] ADR * [0.93] [1.80] [1.16] [0.54] ADR ** [1.45] [1.15] [2.38] [1.51] Days with Zero Return (%) *** *** *** *** [10.26] [3.06] [5.92] [9.01] Log (Number of Stocks) *** *** *** [3.93] [1.06] [4.58] [3.22] Mean of Dependent Variable St. Dev. Of Dependent Variable Observations Number of Countries Test ADR0 =ADR1 (p-value) (0.3900) (0.0290) (0.0680) (0.5900) R-squared within R-squared between R-squared total Year Fixed Effect YES YES YES YES Country Fixed Effect YES YES YES YES Country Random Effect NO NO NO NO * significant at 10%; ** significant at 5%; *** significant at 1%

33 Cross- Autocorrelation Overall Cross- Autocorrelation Downside minus Upside Downside Cross- Autocorrelation Upside Cross- Autocorrelation (I) (I) (I) (I) Short Sales Allowed and Practiced * *** [0.82] [1.84] [2.87] [0.30] ADR ** ** ** [2.49] [2.03] [2.30] [1.33] ADR ** * *** [2.34] [0.76] [1.74] [2.79] Days with Zero Return (%) 0.035** [2.55] [0.46] [1.26] [0.61] Log (Number of Stocks) ** * [2.32] [0.66] [0.92] [1.81] Mean of Dependent Variable St. Dev. Of Dependent Variable Observations Number of Countries Test ADR0 =ADR1 (p-value) (0.4000) (0.0160) (0.9500) (0.0600) R-squared within R-squared between R-squared total Year Fixed Effect YES YES YES YES Country Fixed Effect YES YES YES YES Country Random Effect NO NO NO NO * significant at 10%; ** significant at 5%; *** significant at 1%

34 Skewness Skewness and short sale Regulators believe prohibiting short sales can prevent extreme negative returns Restraining negative information by short sale can indeed aggravates skewness Do we find support for each argument Weak evidence on market-level relation between constraints and skewness. Insignificant relationship at the individual stock level.

35 Crashes Regulators change short sales restrictions in response to market crash (Malaysia, Thailand) For each country (and group of shortable / nonshortable stocks), we calculate the probability of a crash as: Pr[ R 2 mt R m If returns are distributed log-normal, then this probability equals m ]

36 Domestic Stocks Frequency of Extreme Negative Returns Market Skewness Mean Skewness of Individual Stock Returns Mean Skewness of Individual Residual Returns Short Sales Not Allowed / Not Practiced Number of Observations Mean 8.58% Median 6.17% Short Sales Allowed And Practiced Number of Observations Mean 6.00% Median 4.93% Difference (p-value) (0.0001) (0.0041) (0.0976) (0.8058) Whole Sample Number of Observations Mean 7.25% Median 5.41%

37 Dependent Variable: Skewness of Individual Stock Dependent Variable: Skewness of Individual Stock Raw Dependent Variable: Skewness of Market Indices Abnormal Return Return (I) (IV) (I) (IV) (I) (IV) Short Sales Allowed and Practiced *** [3.01] [0.95] [0.28] ADR [1.56] [0.05] [0.79] [0.46] [0.92] [0.98] ADR ** [2.23] [0.20] [0.13] Days with Zero Return (%) 0.116*** 0.168*** * ** [3.34] [3.59] [0.22] [0.49] [1.68] [1.99] Log (Number of Stocks) [0.69] [1.11] [0.55] [0.25] [1.13] [1.56] Detrended Turnover (lagged) [0.95] [0.00] [0.86] [0.57] [1.30] [0.46] Market Return (lagged) ** 0.079* [0.94] [0.27] [1.35] [2.49] [1.77] [1.20] Mean of Dependent Variable St. Dev. Of Dependent Variable Observations Number of Countries Test ADR0 =ADR1 (p-value) (0.1000) (0.6500) (0.3000) R-squared within R-squared between R-squared total Year Fixed Effect YES YES YES YES YES YES Country Fixed Effect YES YES YES YES YES YES Country Random Effect NO NO NO NO NO NO * significant at 10%; ** significant at 5%; *** significant at 1%

38 Distributional Properties of Stock Returns We estimate the kernel density for the market returns, both for the total sample and for the sub-sample of countries with regulatory changes, depending on whether short sales are allowed or not, and depending on whether stocks are dual-listed or not

39

40 Event Study In the final part of the paper, we focus on the countries that have changed regulation between 1990 and Hong Kong, Malaysia, Norway, Sweden, Thailand Efficiency improves but no conclusive evidence on stability

41 20.00% 15.00% 10.00% 5.00% 0.00% % % R-Squared Difference Cross-Autocorrelation Difference

42 16% % 12% 10% 8% 6% 4% 2% 0% Probability of Extreme Negative Return Market Skewness

43 Conclusions Efficiency Individual Asset Returns Markets