# Guo's dummy speculation: blind investments on rising or falling stocks.

ABSTRACTThis paper simulates the stock market, during the bull market period, and attempts to speculate and profit from buying stocks with a sharp drop in excess of 10% and those with a sharp increase of 10% in daily stock prices. Using the descriptive analysis, coefficient of correlation and optimal scaling of regression, the results statistically showed that speculating in stocks with a daily drop of more than 10% will have an overall positive return, while speculating in stocks with a daily rise of more than 10% will have an overall negative return. It also proved a significant correlation between the stocks' earnings per share (EPS) and their returns. The significant regression models, between the EPS and returns, showed that EPS is an important indicator for the return in this speculation, under such a scenario. Although, this study needs to be replicated during the bear market period, it does provide some very interesting results.

INTRODUCTION

Investment strategies can be broadly divided into two types: hedging and speculating. Hedging is an investment made in order to reduce the risk of adverse price movements in a security, by taking an offsetting position in a related security, such as an option or a short sale. Speculating includes taking large risks, especially with respect to trying to predict the future; it is similar to gambling, in the hopes of making quick, large gains (Investorword.com, 2004).

Advocates of hedge funds point to their superior absolute returns as well as superior risk-adjusted returns, and their low correlation with stock market return. Also, hedge funds depend on the selection skill of managers to produce performance (Jahnke, 2004). Whether the decision is made to hedge or speculate, the aim is to make a profit thus avoiding a loss.

This paper will proceed from a speculative perspective in the real stock market, attempting to make a profit in the stock market in a short term period--of not more than one week--by buying two major types of stocks: stocks with a daily drop of more than 10% in stock price and stocks with a daily increase more than 10% in stock price.

LITERATURE REVIEW

AMW and others from the technical and quantitative school confronted the Efficient Market Hypothesis and showed that prices do not always adjust quickly to information shocks as financial theory would suggest. They also showed that a deficiency in the quality of information and its effect on price in specific markets for limited periods of time can offer a trading window using simple technical trading methods. In short, an analysis of past prices can earn abnormal profits for the investor (Tam, 2002). Therefore, there are always opportunities for speculators to profit from the markets by the analysis of past prices, and by the use of technical trading methods. Technical analysis (Investword.com, 2004) tries to evaluate securities by relying on the assumption that market data, such as charts of price, volume, and open interest, can help predict future (usually short-term) market trends. And the use of technical analysis to predict security price movements from past price series has been supported by a number of academic research studies (Summers, Griffiths, and Hudson, 2004).

A cornerstone of technical analysis is the Dow Theory which states that in a true bull or bear trend, both the Dow Jones Industrial Average (DJIA) and Dow Jones Transportation Average (DJTA) must be moving in the same direction. In general, Dow Theory adherents will buy when the market moves higher than a previous peak and sell when it goes below the preceding valley (Moneycentral.msn.com, 2004). Also, Castle in the Air theory (Malkiel, 2003) suggests that investors will be attracted to high priced stocks, due to the psychic behaviors of investors, who believe stock prices will go even higher.

On the other hand, it is also suggested that bullish investors buy a stock in the hope it will go up, buy low and sell high (Sands, 2004), therefore realizing a profit. Basing speculations on the Dow Theory, if the stocks have gone higher than the previous peak, the Dow Theory adherents should buy, because they are likely to profit. Basing speculation on the rule of buy low and sell high under bullish conditions, one would buy stocks when they drop lower than their values, and sell it at a higher price in order to earn a profit.

To profit from the Dow Theory of buy high, sell higher and the rule of buy low, sell high, one needs to determine whether the market is a bull or bear market. Fisher suggests that 2004 will be a good year for equities (Fisher, 2004). Reviewing the past year, the Dow Jones average has gone up from 7,416.64 to 10,753.63 (DJ Industrial Average, 2004). NASDAQ has increased from 1,253.22 to 2,153.83(NAS/NMS Composite, 2004). Therefore, it appears the bull markets have returned.

In a bull market, one can draw out some of the real stock data to test whether the Dow Theory of buy high, sell higher is profitable, or whether buy low, sell high is preferred. According to the Central Limit theorem, the fundamental result that the sum (or mean) of independent identically distributed random variables with finite variance approaches a normally distributed random variable as their number increases. And in particular, if enough samples are repeatedly drawn from any population, the sum of the sample values can be thought of, approximately, as an outcome from a normally distributed random variable (Word reference, 2004). Therefore, if the data, which is a sample of the stock market population, is significant, then it should prove whether the Dow Theory of buy high, sell higher is profitable, and if buy low, sell high brings earnings in a speculative perspective.

[GRAPHIC OMITTED]

GENERAL HYPOTHESES

According to the Castle in the Air theory, stocks will continue to go up. Therefore, it is expected the stocks with a daily dramatic increase greater than 10% will have a positive return performance in the following days, so:

H1: Ho: R [less than or equal to] 0 Ha: R > 0

It is also expected that the stocks with a daily dramatic decrease greater than 10% will have a positive return performance in the following days according to the buy low, sell high rule, so:

H2: Ho: R [less than or equal to] 0 Ha: R > 0

Based on the Fundamental Analysis (Brown and Riley, 2004), stocks with better earnings will have better performance than stocks with lower earnings. Therefore, it is predicted that the randomly selected stocks with a daily dramatic increase greater than 10% and a positive EPS will have a better return performance than the stocks with a negative EPS:

H3: Ho: R (+EPS) = R (!EPS) Ha: R (+EPS) > R (- EPS)

If the Fundamental Analysis holds, there must be a correlation (~) between EPS with the return performance of the stocks; therefore, it is predicted that if the randomly selected stocks have a daily dramatic increase greater than 10% then their EPS is correlated to their returns:

H4: Ho: R ~ (+ EPS) Ha: R x~ (+ EPS)

Under the same situation, based on Fundamental Analysis for stocks with a dramatic decrease greater than 10%, stocks with better earnings will perform better than stocks with worse earnings. Therefore, it is predicted that, stocks with a daily dramatic decrease greater than 10% and a positive EPS will have a better return performance than the stocks with a negative EPS:

H5: Ho: R (+EPS) = R (! EPS) Ha: R (+EPS) > R (- EPS)

Also, it is predicted that, if the randomly selected stocks have a daily decrease greater than 10% then their EPS is correlated to their return:

H6: Ho: R ~ (+ EPS) Ha: R x~ (+ EPS)

If there is correlation between the EPS and the return performance for the stocks, there must be a regression model that can represent the relationship, because EPS determines the return performance; therefore, it is predicted that randomly selected stocks with a daily dramatic increase greater than 10% will have a regression model of:

H7: Ho: R = EPS + e Ha: R [not equal to] EPS + e

It is also predicted that the randomly selected stocks with a daily dramatic decrease greater than 10% will have a regression model of:

H8: Ho: R = EPS + e Ha: R [not equal to] EPS + e

DATA SOURCE AND METHODOLOGY

Data were exclusively and randomly collected daily from the Yahoo! Finance stock screener, from Feb 3, 2004 to Feb 20, 2004. One hundred forty-seven stocks with a daily increase of more than 10% were randomly selected and 76 stocks with a daily decrease of more than 10% were randomly selected.

Dow Jones and NASDAQ stocks with a daily dramatic increase or decrease more than 10% in stock prices were collected along with their EPS. This provided two sample pools. One sample pool with stocks that have a daily increase greater than 10% in stock price and one with stocks that have a daily decrease greater than 10% in stock price. The return performance of these stocks in the following 4 days was traced; their geometric returns in the following 2 days, 3 days and 4 days were calculated and used for statistical analyses together with their individual EPS.

The raw data was analyzed with SPSS to evaluate their descriptive statistics. These data were categorized and run to compare their means according to their EPS and to test the correlation between the EPS and the returns. Also, and evaluation was conducted as to whether a regression model exists between EPS and the return by running the Optimal Scaling of Regression. The Optimal Scaling examines the underlying metrics of data gathered from different a priori known populations (Mullen, 1995). Essentially, Optimal Scaling takes the data as ordinal or nominal and re-scales it in order to uncover the optimal relationship between predicting dependent and independent variables.

FINDING AND ANALYSES

Table I shows the result that stocks with an increase greater than 10%, will not have a positive mean return performance in the following days. The mean returns in Table 1 are all less than zero in the following one-day to four-day returns. Therefore, we fail to reject Ho of Hypothesis 1, for stocks with a daily increase greater than 10% will not have a positive return performance in the following days.

Table II shows stocks with a daily dramatic decrease greater than 10% and their mean returns. The mean returns for these stocks are all positive. Therefore, Ha of Hypothesis 2 stands. The day following a decrease gives the greatest return among all the different day returns. In addition the longer the time to invest under this scenario, the less the mean returns.

In the rescaling version of this test, the deciles are used instead of the raw data versions in Table I and II. Table III shows stocks with a daily dramatic increase greater than 10% and their deciles of returns in 1-day, 2-day, 3-day, 4-day returns and their EPS. It statistically suggests that 1-day, 2-day, 3-day and 4-day returns are significant at a 95% confidence level, reject Ho of Hypothesis 3. Therefore, the stocks with a positive EPS will have a better return than those with a negative EPS

Table IV shows the stocks with a daily increase greater than 10%, their correlation (~) between the deciles of EPS and the deciles of return performance in 1-day, 2-day, 3-day and 4-day. The return performances are significant at a 95% confidence level. So, stocks with a daily increase greater than 10% and their deciles of EPS are correlated to their return performance.

Table V shows stocks with a daily decrease greater than 10%, their deciles of 1-day, 2-day, 3-day and 4-day returns and their EPS. Since T-test for all is not significant at 95% confidence level, we fail to reject Ho of Hypothesis 4, or we failed to say that for the stocks with a daily decrease greater than 10%, the returns for positive EPS stocks perform better than negative EPS stocks.

Table VI shows the stocks with a daily decrease greater than 10%, their correlation (~) between the deciles of EPS and the deciles of return performance in 1-day, 2-day, 3-day and 4-day. Since all correlations are not significant at a 95% confidence interval, the Ho of Hypothesis 6 cannot be rejected. For stocks with a daily decrease greater than 10%, their deciles of EPS are correlated to their deciles of returns.

Tables VIIa, VIIb, VIIc and VIId show the outputs of running Regression's Optimal Scaling in SPSS, of stocks with a daily increase greater than 10%, with a regression model for their deciles of EPS and deciles of return performance in 1-day, 2-day, 3-day and 4-day, respectively. Table 8.1, 8.2, 8.3 and 8.4 show the outputs of running Regression's Optimal Scaling, of stocks with a daily decrease greater than 10%, with a regression model for their deciles of EPS and deciles of return performance in 1-day, 2-day, 3-day and 4-day, respectively.

Table VIIa shows stocks with a daily increase greater than 10%, a regression model for 1-day deciles of return and their deciles of EPS. The regression model is significant at a 99% confidence level, which means the regression model is valid; The R square of 0.073 says the model explains 7.3% of the 1-day returns, and even with only 144 observations, this relationship is statistically significant. In Chart I, the deciles of 1-day return (shown in chart as VAR00001) and the deciles of EPS are both showing monotonic upwards trends. The implication for the rescaling of 1-day return is that there are three important levels, the lowest level for the first, second, and third deciles, a middle level for fourth to eighth deciles, and a high level for the top two deciles. Approximately, the EPS variable is also rescaled as having three levels, lowest level for first to fourth deciles, a middle level for fifth to seventh deciles and a high level for the top three deciles. When this done, the resulting correlation is maximized.

The result of Optimal Scaling indicates the lowest level of 1-day return is correlated to the lowest level of EPS, the middle level of 1-day return is correlated to the middle level of EPS and the high level of 1-day return is correlated to the high level of EPS.

[GRAPHIC 1 OMITTED]

Table VIIb shows stocks with a daily increase greater than 10%, a regression model for 1-day deciles of return and their deciles of EPS. The regression model is significant at a 99% confidence level, which means the regression model is valid. The R square of 0.073 indicates the model explains 17.3% of the 2-day return. In Chart II, the deciles of 2-day return (shown in chart as Return2D) and the deciles of EPS are both showing monotonic upwards trends. It implied five important levels--the lowest level for the first deciles, second lowest level for deciles from two to four, a middle level for deciles five and six, second highest level for deciles seven to nine, and the highest level for tenth deciles. EPS also rescaled as having five levels, the lowest level for the first deciles, second lowest level for deciles from two to four, a middle level for deciles five to eight, second highest level for deciles nine, and the highest level for tenth deciles. And these different levels in 2-day return are correlated to the counterpart levels in EPS.

[GRAPHIC 2 OMITTED]

Table VIIc shows stocks with a daily increase greater than 10%, a regression model for 1-day deciles of return and their deciles of EPS. The regression model is significant at a 99% confidence level, which means the regression model is valid. The R square of 0.141 indicates the model explains 14.1% of the 3-day return. In Chart III, the deciles of 3-day return (shown in chart as Return3D) and the deciles of EPS are both showing a monotonic upward trend. The rescaling of 3-day return tells that there are three important levels- the lowest level for the first to fourth deciles, a middle level for fifth to sixth deciles, and a high level for the top four deciles. Approximately, the EPS variable is also rescaled as having three levels, lowest level for first deciles, a middle level for second to ninth deciles and a high level for tenth deciles. The result of Optimal Scaling tells us that the lowest level of 3-day return is correlated to the lowest level of EPS, middle level of 3-day return is correlated to the middle level of EPS and the high level of 3-day return is correlated to the high level of EPS.

[GRAPHIC 3 OMITTED]

Table VIId shows stocks with a daily increase greater than 10%, a regression model for 1-day deciles of return and their deciles of EPS. The regression model is significant at a 99% confidence level, which means the regression model is valid. The R square of 0.111 indicates the model explains 11.1% of the 4-day returns. In Chart IV, the rescaling of 4-day return shows there are two important levels, a lowest level for first to fifth deciles and a top level for sixth to tenth deciles. And EPS can be approximately rescaled into 2 levels as well, a lower level from first to seventh deciles and a top level from eighth to tenth deciles. These different levels in 2-day return are correlated to the counterpart levels in EPS.

[GRAPHIC 4 OMITTED]

Table VIIIa shows stocks with a daily decrease greater than 10%, a regression model for 1-day deciles of return and their deciles of EPS. The regression model is significant at a 99% confidence level, which means the regression model is valid. The R square of 0.074 indicates the model explains 7.4% of the 1-day returns. In Chart V, the deciles of 1-day return (shown in chart as Var00002), the rescaling of 1-day return shows that there are two important levels, a lower level from first to ninth deciles and a high level of tenth deciles. For EPS, there are two important levels, lower level from first deciles to fifth deciles and a top level from sixth deciles to tenth deciles. And these different levels in 2-day return are correlated to the counterpart levels in EPS.

[GRAPHIC 5 OMITTED]

Table VIIb shows stocks with a daily decrease greater than 10%, a regression model for 1-day deciles of return and their deciles of EPS. The R square of 0.063 indicates the model explains 6.3% of the 2-day returns. Chart VI shows the deciles of 2-day return (shown in chart as Dretrn2D), the rescaling of 2-day return shows three important levels, the lowest level consists of first to eighth deciles, a middle level of ninth deciles, a high level of tenth deciles. For EPS, it also shows 3 important levels, the lowest level for the first deciles, a middle level for second to ninth level, and a high level for the top deciles. These different levels in the 2-day return are correlated to the counterpart levels in EPS.

[GRAPHIC 6 OMITTED]

Table VIIIc shows stocks with a daily decrease greater than 10%, a regression model for 1-day deciles of return and their deciles of EPS. The regression model is significant at a 99% confidence level, which means the regression model is valid. The R square of 0.103 indicates the model explains 10.3% of the 3-day returns. In Chart VII, the deciles of 3-day return (shown in chart as Dretrn3D), the rescaling of 3-day return shows three important levels, a lowest level of the first deciles, a middle level for second and third decides, and a high level for fourth to tenth deciles. For EPS, it also shows 3 important levels, a lowest level for the first to third deciles, a middle level for fourth to ninth level, and a high level for the top deciles. And these different levels in 2-day return are correlated to the counterpart levels in EPS.

[GRAPHIC 7 OMITTED]

Table VIIId shows stocks with a daily decrease greater than 10%, a regression model for 1-day deciles of return and their deciles of EPS. The regression model is significant at a 99% confidence level, which means the regression model is valid. The R square of 0.134 indicates the model explains 13.4% of the 4-day returns. From the Chart VIII the rescaling of 4-day return (shown in chart as Dretrn4D) shows two important levels, a lower level of first to eighth deciles, and a high level for ninth to tenth deciles. For EPS, it also shows 2 important levels, a lowest level for the first to ninth deciles and a high level for the top deciles. And these different levels in 2-day return are correlated to the counterpart levels in EPS.

[GRAPHIC 8 OMITTED]

DISCUSSION AND IMPLICATIONS

The above results show that stocks which have a daily increase exceeding 10%, will generate overall negative returns. This could be explained by the Efficient Market Theory; price has effectively reflected their price when there is information available. So, in the first day of dramatic increase, stocks have truly reflected their price.

On the other hand, stock speculations with daily decreases more than 10%, will generate overall positive returns, and furthermore, return in the following day after the dramatic decrease will give the highest returns. This phenomenon can be explained by the investing rule of buy low and sell high in the bull markets. Once the investors see the stocks price drop, the investors believe the market will go up in the future. Investors will try to buy the stocks at the low price, in which they believe the stock prices will rise soon. Therefore, the stock price should rise after a sharp decrease.

The relationship between EPS and the returns, from Table III in the Finding and Analyses Section, indicates for stocks with a daily increase greater than 10%, investing in the stocks with positive EPS should give a better return than those with a negative EPS. Further statistics also show in Table V, Table VIIa, VIIb, VIIc, and VIId, that there is a significant correlation between the EPS and the returns. It can be expressed in the regression model: R= EPS + e. This finding gives an important indication for investment during the Bull market. When speculating in stocks under such a scenario, the EPS should be an important indicator. If the stocks have a positive EPS, the returns it generates under such a scenario should be more favorable.

In stocks with a daily decrease greater than 10%, both with negative EPS and positive EPS, there is not enough statistical evidence that the returns will favor either side. This could be explained by the Expectation Theory and the rule of buy low, sell high. In the Bull market, investors' expectation for the future markets are bullish, and they will arbitrarily buy those stocks that have gone down, with the expectation that they will go up and profit from it soon. Therefore, the Expectation Theory dominates the others. Under this investing scenario, no matter how good or bad the stocks are, investors will buy low and sell high.

This paper shows the correlation between 1-day return, 2-day return, 3-day return, 4-day return of stocks with a daily decrease greater than 10%, and their EPS are not significant in Table VI. However, in Regression Optimal Scaling, Table 8.1, 8.2, 8.3 and 8.4, it shows that the correlation between 1-day return, 2-day return, 3-day return, 4-day return, and the EPS are significant. Regression Optimal Scaling examines the underlying metrics of data gathered from different a priori known populations; it is a much more complex model than Correlation. Regression Optimal Scaling considers the underlying metrics of the data. The output covers dimensions and areas that Correlation does not . Therefore, the Regression Optimal Scaling results should prevail under the contradicting situation.

LIMITATIONS AND APPLICATION

This paper assumes that the stock buy and sell orders can be executed at a specific time. That is, buying the day after a daily increase or decrease greater than 10% in daily stock price and then selling right after1-day, 2-day, 3-day or 4-day. Also, this paper assumes that there is no transaction fee incurred during the buy and sell. Investing in stocks with a daily decrease more than 10% in a bull market, the next day will generate the overall highest daily return of 2.948%. If annualized to the 260 days for Exchange Index working days, it yields a 766.48% return in a year. This scenario might be very profitable, yet it is also very risky as well.

SUMMARY AND CONCLUSION

When speculating in stocks in a bull market, stocks with a daily increase or decrease greater than 10% would be preferable as the following day's purchase should give the overall best returns. Additionally, the selection should include a review of their EPS, because the EPS is related to the stock's performance. Generally, speculating in the stocks with daily dramatic increase more than 10%, a positive EPS stocks will have a better overall return than those with a negative EPS.

REFERENCES

Dow Jones Industrial Average. Finance.yahoo.com. Retrieved from http://finance.yahoo.com 1y

Fisher, Kenneth L. (2004). Bush's Bull Market, Forbes; Mar 2004

Investorword.com--Retrieved from http://www.investorwords.com/2122/ fundamental_ analysis.html

Investorword.com--Retrieved from http://www.investorwords.com/4925/technical_analysis.html

Jahnke, William. (2004). Hedge Funds Aren't Beautiful, Journal of Financial Planning; Feb 2004.

Malkiel, Burton G. (2003). A Random Walk Down Wall Street: The Time Tested Strategy for Successful Investing, New York: W. W. Norton.

Moneycentral.msn.com--Retrieved from http://moneycentral.msn.com/investor/glossary/ glossary.asp?TermID=515

Mullen, Michael R. (1995). Diagnosing Measurement Equivalence in Cross-National Research, Journal of International Business Studies.

NAS/NMS Compsite. Finance.yahoo.com. Retrieved from http://finance.yahoo.com

Sands, Eric. Experts: Beginner Investing. Retrieved from http://experts.about.com/q/3253/ 2188358.htm

Summers, Barbara; Griffiths, Evan & Hudson, Robert. (2004). Back to the future: an empirical investigation into the validity

of stock index models over time, Applied Financial Economics. March 2004.

Tam, Fred.(2002). Analysis of past price trends pays, Nov 2002.

Wordreference.com--Retrieved from http://www.wordreference.com/english/ definition.asp?en=central+lim

Guoliang He, Sam Houston State University

Bala Maniam, Sam Houston State University

Table I: Stocks with an increase greater than 10% and their returns Std. N Minimum Maximum Mean Deviation return1d 147 -2.45000 .33230 -.0201892 .21162339 return4d 142 -.07482 .10778 -.0029510 .02377460 return3d 147 -2.50602 .14169 -.0360473 .2696371 return2d 145 -.14311 .14752 -.0049578 .03739231 Valid N 142 (listwise) Table II: Stocks with a decrease greater than 10% and their mean returns Std. N Minimum Maximum Mean Deviation retrn1d 76 -.34070 1.63000 .0294776 .20128841 dretrn4d 76 -.06938 .24174 .0068428 .03788476 dretrn3d 76 -.10727 .36166 .0070437 .05398998 dretrn2d 76 -.13931 .56897 .0087518 .07691867 Valid N (listwise) 76 Table III: Stocks with a daily increase greater than 10%--With a positive EPS vs. with a negative EPS Levene's Test t-test for for Quality Equality of Variance of Means F Sig t NTILES of VAR00001 Equal variance 2.356 .127 -2.268 Assumed Equal variance -2.329 Not assumed NTILES of RETURN2D Equal variance 0.617 .434 -3.05 Assumed Equal variance -3.105 Not assumed NTILES of RETURN3D Equal variance 0.073 .787 -3.127 Assumed Equal variance -3.134 Not assumed NTILES of RETURN4D Equal variance 0.002 .968 -2.582 Assumed Equal variance -2.579 Not assumed t-test for Equality of Means df Sig. Mean (2-tailed) Difference NTILES of VAR00001 Equal variance 145 .025 -1.096 Assumed Equal variance 120.098 .022 -1.096 Not assumed NTILES of RETURN2D Equal variance 143 .003 -1.467 Assumed Equal variance 114.551 .002 -1.467 Not assumed NTILES of RETURN3D Equal variance 145 .002 -1.487 Assumed Equal variance 111.665 .002 -1.487 Not assumed NTILES of RETURN4D Equal variance 140 .011 -1.270 Assumed Equal variance 103.370 .011 -1.270 Not assumed t-test for Equality of Means 95% Confidence Std. Error Interval of the Difference Difference Lover Upper NTILES of VAR00001 Equal variance 0.483 -2.052 -0.141 Assumed Equal variance 0.471 -2.028 -0.164 Not assumed NTILES of RETURN2D Equal variance 0.481 -2.417 -0.516 Assumed Equal variance 0.472 -2.402 -0.531 Not assumed NTILES of RETURN3D Equal variance 0.476 -2.428 -0.547 Assumed Equal variance 0.475 -2.428 -0.547 Not assumed NTILES of RETURN4D Equal variance 0.492 -2.242 -0.297 Assumed Equal variance 0.492 -2.246 -0.293 Not assumed Table IV: Stocks with a daily increase greater than 10%-- Correlation (~) between EPS and return performance Correlations NTILES of NTILES of NTILES of VAR00001 RETURN2D RETURN3D NTILES of VAR00001 Pearson Correlation 1 .616 ** .575 ** Sig. (2-tailed) .000 .000 N 147 145 147 NTILES of RETURN2D Pearson Correlation .616 ** 1 .825 ** Sig. (2-tailed) .000 .000 N 145 145 145 NTILES of RETURN3D Pearson Correlation .575 ** .825 ** 1 Sig. (2-tailed) .000 .000 N 147 145 147 NTILES of RETURN4D Pearson Correlation .576 ** .675 ** .835 ** Sig. (2-tailed) .000 .000 .000 N 142 142 142 NTILES of EPS Pearson Correlation .182 * .323 ** .289 ** Sig. (2-tailed) .029 .000 .000 N 144 143 144 Correlations NTILES of NTILES RETURN4D of EPS NTILES of VAR00001 Pearson Correlation .576 ** .182 * Sig. (2-tailed) .000 .029 N 142 144 NTILES of RETURN2D Pearson Correlation .675 ** .323 ** Sig. (2-tailed) .000 .000 N 142 143 NTILES of RETURN3D Pearson Correlation .835 ** .289 ** Sig. (2-tailed) .000 .000 N 142 144 NTILES of RETURN4D Pearson Correlation 1 .206 * Sig. (2-tailed) .014 N 142 141 NTILES of EPS Pearson Correlation .206 * 1 Sig. (2-tailed) .014 N 141 144 ** Correlation is significant at the 0.01 level (2- tailed) * Correlation is significant at the 0.05 level (2-tailed) Table V: Stocks with a daily decrease greater than 10%-- With a positive EPS vs. with a negative EPS Independent Samples Test t-test for Levene's Test for Equality Quality of Variance of Means F Sig t NTILES of VAR00001 Equal variance .302 .584 -1.269 Assumed Equal variance -1.278 Not assumed NTILES of RETURN2D Equal variance .094 .761 -1.003 Assumed Equal variance -0.979 Not assumed NTILES of RETURN3D Equal variance 1.018 .316 -1.003 Assumed Equal variance -0.959 Not assumed NTILES of RETURN4D Equal variance .087 .769 -1.815 Assumed Equal variance -1.785 Not assumed t-test for Equality of Means df Sig. Mean (2-tailed) Difference NTILES of VAR00001 Equal variance 74 .208 -.904 Assumed Equal variance 42.574 .208 -.904 Not assumed NTILES of RETURN2D Equal variance 74 .319 -.717 Assumed Equal variance 39.700 .334 -.717 Not assumed NTILES of RETURN3D Equal variance 74 .319 -.717 Assumed Equal variance 37.976 .344 -.717 Not assumed NTILES of RETURN4D Equal variance 74 .074 -1.278 Assumed Equal variance 40.352 .082 -1.278 Not assumed t-test for Equality of Means Std. 95% Confidence Error Interval of the Difference Difference Lover Upper NTILES of VAR00001 Equal variance .712 -2.323 .515 Assumed Equal variance .707 -2.331 .523 Not assumed NTILES of RETURN2D Equal variance .715 -2.142 .708 Assumed Equal variance .732 -2.198 .764 Not assumed NTILES of RETURN3D Equal variance .715 -2.142 .708 Assumed Equal variance .748 -2.231 .797 Not assumed NTILES of RETURN4D Equal variance .704 -2.681 .125 Assumed Equal variance .716 -2.725 .169 Not assumed Table VI: Stocks with a daily decrease greater than 10%--Correlation (~) between EPS and return performance Correlations NTILES of NTILES of NTILES of VAR00002 DRETRN4D DRETRN3D NTILES of VAR00002 Pearson Correlation 1 .592 ** .626 ** Sig. (2-tailed) .000 .000 N 76 76 76 NTILES of DRETRN4D Pearson Correlation .592 ** 1 .844 ** Sig. (2-tailed) .000 .000 N 76 76 76 NTILES of DRETRN3D Pearson Correlation .626 ** .844 ** 1 Sig. (2-tailed) .000 .000 N 76 76 76 NTILES of DRETR24D Pearson Correlation .774 ** .696 ** .777 ** Sig. (2-tailed) .000 .000 .000 N 76 76 76 NTILES of EPS Pearson Correlation .028 .007 -.012 Sig. (2-tailed) .809 .952 .919 N 75 75 75 Correlations NTILES of NTILES DRETRN3D of EPS NTILES of VAR00002 Pearson Correlation 1774 ** .028 Sig. (2-tailed) .000 .809 N 76 76 NTILES of DRETRN4D Pearson Correlation .696 ** .007 Sig. (2-tailed) .000 .952 N 76 75 NTILES of DRETRN3D Pearson Correlation .777 ** -.012 Sig. (2-tailed) .000 .919 N 76 75 NTILES of DRETR24D Pearson Correlation 1 .028 Sig. (2-tailed) .809 N 76 75 NTILES of EPS Pearson Correlation .028 1 Sig. (2-tailed) .809 N 75 75 ** Correlation is significant at the 0.01 level (2-tailed) Table VIIa stocks with a daily increase greater than 10% --A regression model for their deciles of 1-day return and deciles of EPS Model Summary Adjusted Multiple R R Square R Square .270 .073 .066 ANOVA Sum of Mean Squares df Square F Sig Regression 10.471 1 10.471 11.136 .001 Residual 133.529 142 .940 Total 144.000 143 Table VIIb stocks with a daily increase greater than 10%--A regression model for their deciles of 2-day return and deciles of EPS Model Summary Adjusted Multiple R R Square R Square .415 .173 .167 ANOVA Sum of Mean Squares df Square F Sig Regression 24.687 1 24.687 29.421 .000 Residual 118.313 141 .839 Total 143.000 142 Table VIIc stocks with a daily increase greater than 10% --A regression model for their deciles of 3-day return and deciles of EPS Model Summary Adjusted Multiple R R Square R Square .376 .141 .135 ANOVA Sum of Mean Squares df Square F Sig Regression 20.312 1 20.312 23.319 .000 Residual 123.688 142 .839 Total 144.000 143 Table VIId stocks with a daily increase greater than 10% --A regression model for their deciles of 4-day return and deciles of EPS Model Summary Adjusted Multiple R R Square R Square .333 .111 .104 ANOVA Sum of df Mean F Sig Squares Square Regression 15.634 1 15.634 17.334 .000 Residual 125.366 139 .902 Total 141 140 Table VIIIa stocks with a daily decrease greater than 10% --A regression model for their deciles of 1-day return and deciles of EPS Model Summary Adjusted Multiple R R Square R Square .271 .074 .061 ANOVA Sum of Mean Squares df Square F Sig Regression 5.525 1 5.525 5.805 019 Residual 69.475 73 .952 Total 75.000 74 Table VIIIb stocks with a daily decrease greater than 10% --A regression model for their deciles of 2-day return and deciles of EPS Model Summary Adjusted Multiple R R Square R Square .252 .063 .051 ANOVA Sum of df Mean F Sig Squares Square Regression 4.752 1 4.752 4.938 029 Residual 70.248 73 .962 Total 75.000 74 Table VIIIc stocks with a daily decrease greater than 10% --A regression model for their deciles of 3-day return and deciles of EPS Model Summary Adjusted Multiple R R Square R Square .320 .103 .090 ANOVA Sum of Mean Squares df Square F Sig Regression 7.697 1 7.697 8.348 005 Residual 67.303 73 .922 Total 75.000 74 Table VIIId stocks with a daily decrease greater than 10% --A regression model for their deciles of 4-day return and deciles of EPS Model Summary Multiple R R Square Adjusted R Square .366 .134 .122 ANOVA Sum of Mean Squares df Square F Sig Regression 10.022 1 10.022 11.260 001 Residual 64.978 73 .890 Total 75.000 74

Printer friendly Cite/link Email Feedback | |

Author: | He, Guoliang; Maniam, Bala |
---|---|

Publication: | Academy of Accounting and Financial Studies Journal |

Article Type: | Statistical table |

Date: | May 1, 2008 |

Words: | 5943 |

Previous Article: | The relationship between dividend payouts and systematic risk: a mathematical approach. |

Next Article: | Can trading on federal funds rate change announcements produce above normal stock market returns?: A test of market efficiency. |

Topics: |