• No results found

Overreaction in America : an analysis of the short-term overreaction in the American Stock Market during and after the financial crisis

N/A
N/A
Protected

Academic year: 2021

Share "Overreaction in America : an analysis of the short-term overreaction in the American Stock Market during and after the financial crisis"

Copied!
26
0
0

Bezig met laden.... (Bekijk nu de volledige tekst)

Hele tekst

(1)

Overreaction in America

An analysis of the short-term overreaction in the American Stock Market during and

after the financial crisis

Bachelor Thesis Economics and Business: Finance & Organization

University of Amsterdam

Maarten van Lith

(2)

2

Statement of Originality

This document is completely written by Maarten van Lith, who declares to take full responsibility for the contents of this document.

I declare that the text and the work presented in this document is original and that no other sources other than those mentioned in the text and its references have been used in creating it.

The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.

(3)

3

Abstract

Research in historical prices suggest that people tend to overreact negatively when dramatic and unexpected news become public. This study investigates if this behavior creates the opportunity for investors to gain abnormal returns. By using CRSP daily stock data, empirical evidence is found that supports the overreaction hypothesis and the hypothesis that there is an opportunity to gain abnormal returns from this. Therefore, evidence for weak form market inefficiencies has been found.

(4)

4

Table of Contents

Statement of Originality ...2 Abstract ...3 Introduction ...5 1. Literature Review ...6 1.1 Overreaction Hypothesis ... 6

1.2 Evidence for Overreaction in the Short-Term ... 8

1.3 Explanations for the Overreaction Phenomenon ... 9

2. Data ...11

3. Methodology ...12

4. Results and Analysis ...13

4.1 Descriptive Statistics ... 13

4.2 Overreaction Hypothesis ... 19

4.3 Crisis versus Post-Crisis ... 20

5. Conclusion...21

(5)

5

Introduction

Ever since Fama (1970) came with his research about the Efficient Market Hypothesis there have been numerous researches testing this hypothesis. The efficient market hypothesis argues that all public information and expectations are incorporated in a stock’s price and that it is impossible to consistently earn abnormal returns, however, numerous researches show that there might be factors that lead to inefficient markets. Among these studies there were psychological and behavioral motives as the reason for the inefficient market. In 1985 De Bondt and Thaler proved that prior “loser portfolios” outperformed prior “winner portfolios”. With this study they introduced the overreaction hypothesis. The overreaction hypothesis says that the market overreacts to certain new information which would lead to a larger increase or decrease in a share’s price than what should be reasonable. In turn, this would lead to a reversal of the stock price.

The debate about the hypothesis is still not finished. More recently Piccolo, Chaudhury, Souza and Vieira da Silva (2017) confirmed the overreaction hypothesis as described by De Bondt and Thaler in 1985. However, they still failed to adjust their research for the question whether investors are able to profit from the overreaction hypothesis because they did not look into transaction costs or held the bid-ask into account. This is important because if there are no corrections made for real life situations, the conclusion that is given is prone to not reflecting the real situation and possibly giving a wrong conclusion. Although the real bid -ask spreads are not used in this study, a variable is used that is expected to lead to a sample where the spread is negligible. After correcting for bid-ask spreads and transaction costs Atkins and Dyl (1990) found that the overreaction phenomenon does not necessarily means an opportunity for investors to gain abnormal returns and thus the efficient market hypothesis still holds.

The purpose of this study is to examine the short-term negative overreaction hypothesis in the American stock market correcting for risk, liquidity and transaction costs. I use the New York Stock Exchange (NYSE) after and during the financial crisis of 2008 and to examine if there is a possibility to gain higher-than-market returns from this. No research has used the latest data and the latest studies have overlooked the illiquidity of stocks and therefore the bid-ask spreads. This contributes to research that can be put in practice.

(6)

6

To answer this question, I use event study where data is collected from CRSP and an event study is conducted. Then the abnormal returns in the day following the day on which the security’s price decreased are being tested for significance. Furthermore, by including the financial crisis in the sample period, the possibility arises to see if there are more market inefficiencies during crisis periods than in normal market conditions. Malkiel (2003) criticized earlier studies for only having found results during crisis periods. Because this study investigates both periods together and separately it would become able to give a conclusion about this comment.

The remainder of this paper is organized as follows. First the relevant literature will be reviewed. Hereafter the data and methodology will be described in section 3 and 4. In section 5 the results are analyzed and discussed. In the last section, a summary of the results and conclusion is provided.

1. Literature Review

1.1 Overreaction Hypothesis

The overreaction hypothesis in itself assumes that there is a certain correction that is to be made after new information has become public that is ‘appropriate’. The overreaction phenomenon arises when price adjustment in securities is greater than this appropriate number. One of the first observations of overreaction in the securities market was by Keynes (1936). The overreaction theory is one of the many topics that would lead to the conclusion that the Efficient Market Hypothesis is, at least not completely, true and that it may be possible to predict the movement of a stock’s price.

The first research about the anomaly was in 1985 by De Bondt and Thaler. In their study, they concluded that the overreaction hypothesis exists. To come to this conclusion, they made two portfolios of stocks, one consisted of stocks which performed weakly in the last periods and one consisted of stocks which performed strongly in the last periods and tested whether the portfolio of the ‘losing stocks’ performed better than the ‘winner stocks’ and also compared the return of the portfolios with the market using CAPM. They argue that if stock prices overshoot when new information becomes public it would be possible to predict

(7)

7

the reversal effect just by looking at past prices. They proved that extreme movements in a stock’s price are followed by movements in the opposite directions.

In 1987 De Bondt and Thaler did a follow-up study about the overreaction hypothesis. In this research, they delivered additional evidence that supports the overreaction hypothesis. They examined the overreaction effect after putting in control variables for risk, company size and seasonality. This time they also factored in differences risk using CAPM. It became clear that the overreaction effect didn’t arise from a change in risk. Although they found that losing companies are often smaller companies they also concluded that company size is not an explanation for overreaction.

This is inconsistent with a research that was published in 1989. Paul Zarowin (1989) also researched the overreaction effect. The conclusion of his paper was that there was indeed evidence to support the theory that losing stocks outperform winning stocks, however, this couldn’t be caused by the overreaction effect. When firms comparable in size were compared it showed that there was no significant difference in performance between the two. This would mean that not the differences in returns by previous losers and winners was the result of firm size and the size effect.

A division in the researches about the overreaction phenomenon can be made by looking at the duration of the research period after the day that the overreaction occurs. To be more exact the division can be made by labeling research ‘long-term’ and ‘short-term’. In this study a study is defined as long-term if the observations in the research are exceeding 30 trading days after the overreaction happens. Often times long-term studies focus on monthly returns. Short term researches are defined as such when the research period is less than 30 trading days after the overreaction occurs. In short-term researches daily returns are predominantly used in the data. The researches of De Bondt and Thaler (1987) and Zarowin (1989) were both focused on the long-term and did not include any variables for illiquid stocks. This study is focused on the short-term because it is expected that the overreaction phenomenon will be more apparent in the short-term.

A variable to exclude illiquid stocks is included. This is because illiquid stocks often mean that the bid-ask spreads are high and that the stocks are not available for trading at all for regular investors. Furthermore, the Fama-French Three Factor Model is used as a way to consider a security’s risk.

(8)

8

1.2 Evidence for Overreaction in the Short-Term

The topic of overreaction in the stock market has been researched multiple times and focus can be put on a lot of things. Because this paper’s focus lies on the short-term, it is interesting to look at what past researches found about the overreaction in the short-term.

Bremer and Sweeney (1991) looked at stocks from the Fortune 500 and characterized an overreaction as a decrease in stock price from -10% in a day or less. On the day that the decrease happened they marked it as an event and looked and the returns from the sequential days. From their research, it appeared that using the trigger strategy of -10% the cumulative return of the following 2 days was 2.2%. This may not necessarily mean an opportunity to attain higher-than-market returns, but it does mean that a certain correction in a share price is not efficient and that a share’s price is not always reflecting or quickly reflecting all the information. Bremer and Sweeney also tried other trigger values, like -7,5% and -15%. The results are unchanged, but they did find that larger values (in absolute value) will have a bigger rebound return. To show that this pattern was not exclusive to the American stock market, Bremer, Hiraki and Sweeney (1997) did a follow-up study on the Japanese stock market. In this study, they again found that there was some kind of overreaction, however ordinary investors were not able to profit from this. They concluded this because they argue that broker firms respond to large price declines not by trading for their own profit but providing their preferred customers with enough supply and liquidity of the stock. Next to Bremer and Sweeney (1991) and Bremer, Hiraki and Sweeney (1997) there are more studies that show that a reversal in stock price exists after a large decrease in share price. Including in these studies are Lehmann (1990), Atkins and Dyl (1990) and Park (1995).

More recently Alwathnani, Dubofsky, & Al-Zoubi (2017), Milian (2015) and Bai & Qin (2015) examined the overreaction effect when company earnings came out. The studies were performed in multiple markets and all found evidence of a certain overreaction when unexpected earnings result became public. Even after controlling for risk with the Fama-French Three Factor model the overreaction is still significant. These studies did not look into the question if there are abnormal returns to be gained by regular investors, something this study does try to find out.

The reason for the phenomenon and the question if a higher-than-market return is achievable is debatable. The following section will elaborate the different explanations for the overreaction phenomenon.

(9)

9

1.3 Explanations for the Overreaction Phenomenon

The explanations for the overreaction effect can be categorized in two different sub categories: one where the efficient market hypothesis holds and one where the EMH does not hold.

The first explanation lies in the field of psychology. It follows from irrational behavior in the decision-making process. De Bondt and Thaler (1985) argue that investors give more weight to recent information than to all the information that came before. When new negative information about a company becomes public investors would value this information more than they should, and the stock price would decline more than is actually efficient. This results in a reversion of the stock price to get to the equilibrium price level. As a result, the overreaction phenomenon occurs. Of course, this would conclude that the efficient market hypothesis does not hold. Adding to this explanation in more recent years is the research of Piccolo, Chaudhury, Souza and Vieira da Silva (2017), who expect that investors have overconfidence in events that are big or heavy but do not occur often. Because of this, investors overreact.

This explanation of the overreaction phenomenon is criticized by Fama (1998) and Malkiel (2003). Fama says that events that are used in psychological studies are selective and that the theory is not relevant to the data. Also, he asserts it would still only be able to explain a little part of the anomalies. Malkiel argues that the support for the overreaction hypothesis is not uniform across different studies. Especially studies in times where the economy is performing poorly are the biggest empirical results in favor of the overreaction hypothesis says Malkiel. In conclusion Fama and Malkiel argue that the overreaction hypothesis would be rejected if researched correctly and that the efficient market hypothesis still holds, despite researches that prove the phenomenon.

Piccolo & Chaudhury (2018) later researched the overreaction phenomenon and whether there would be a difference in the level of overreaction when investor sentiment differs. Using daily returns from the S&P500 and the investor sentiment index they concluded, like Malkiel (2003) stated, that anomalies are more visible and pronounced in economically bad times. This research will also look at if there is a difference between a period where the

(10)

10

economy is performing poorly, namely the financial crisis, and periods where the economy is performing normally.

The second explanation for the overreaction effect can be found in other anomalies of the market and in the methods that are being used to investigate the phenomenon. Chan (1988) finds in his research that the abnormal returns when using a contrarian strategy trying to profit from the overreaction effect are relatively small when control variables for risk are added. He therefore claims that the estimation of the returns is sensitive to the methods that are used. Chan used the Capital Asset Pricing model to come to this conclusion. In his conclusion the efficient market hypothesis would therefore hold. In this research also control variables for risk are included. The Fama-French Three factor model is used.

Zarowin (1989) argues that the overreaction hypothesis can be refuted by company size differences. In his study Zarowin finds that the stocks that performed poorly in previous times do outperform stocks that performed well in previous times, however, after pairing poor earning companies with the best earning companies there is no empirical result that the results in price performance differs. He therefore suggests that the size differences between companies is to blame for what is perceived by many as the overreaction phenomenon.

Countering these explanations, De Bondt and Thaler’s research pointed out that the overreaction hypothesis could not be attributed to either firm size or risk. Their argument to refute the size effect is that the market probably does not recognize well enough the potential of small firms’ higher growth potential. To refute the risk explanation, they included a risk variable in their research, namely CAPM. Even when using this risk-adjusted measure they still found that loser stocks outperformed winner stocks.

In the third explanation for the overreaction phenomenon the efficient market hypothesis also holds. According to the following studies the bid-ask spread explains the overreaction phenomenon. The conclusion in these explanations is that it is not possible for investors to profit from the overreaction hypothesis and achieve higher-than-market returns.

Atkins and Dyl (1990) find evidence of an overreaction effect when looked at average prices. They also looked at the bid ask spread and argue that the magnitude of the overreaction phenomenon is small when this spread and transaction costs are accounted for. Because of the big differences in the bid and ask prices and also counting for transaction costs

(11)

11

they conclude that ordinary investors are not able to profit from the overreaction hypothesis. Therefore, they conclude that both the overreaction hypothesis and the efficient market hypothesis hold. In my research the bid-ask spread is dealt with by only using stocks in the data where there is a big enough volume so that the spread is negligible and furthermore a market model is used as to account for risk. Something that Atkins and Dyl (1990) did not do and was criticized by earlier named Chan (1988).

Adding to the bid-ask spread explanations for overreaction in the market is the research of Cox and Peterson (1994). They find significant reversals in the stock price after a day where the stock price increases or decreases significantly. Like Zarowin, they found that small firms reverse more than big firms. They argue that big one-day price declines would lead to selling pressure and therefore enhance the chance that a closing transaction is at a bid price, which because of the bid-ask bounce would lead to a reversal the following day. To conclude, it may seem that a stock is overreacting, but investors could never profit from this because of the little liquidity. To account for liquidity in this research a stock is only included in the sample if it fulfills the condition that there were at least 1000000 pieces traded on the day that the stock declined.

Park (1995) also claimed that a large part of the overreaction phenomenon is caused by a systematically rapidly changing bid-ask spread, which would lead to higher bid-ask spreads. When the contrarian investment strategy is further investigated he finds that the average abnormal returns are not large enough to count for the loss that is made through the bid-ask spread and the transaction costs. He concludes that there is evidence of market overreaction, but it is not possible for investors to profit from this.

On the contrary is the research of Lehmann (1990). In his research he claims to find evidence that it is possible to profit from arbitrage using the predictive pattern following from the overreaction hypothesis. Even after controlling for bid-ask prices and transaction costs he claims to find evidence. Lehmann concludes “The results strongly suggest rejection of the

(12)

12

2. Data

The sample consist of companies who are selected by using a trigger strategy like Bremer and Sweeney (1991) and Cox and Peterson (1994) did in their research. Whenever a stock from the NYSE had a return of -10% or less an event was created for the date that the decrease occurred. Besides the daily return the stock also had to have a minimum daily trading volume of 1,000,000. This way illiquid stocks where the bid-ask spread is big compared to the regular traded stock are skipped. A too big bid-ask spread would block the opportunity to make a good return, even though stock price would increase. The adding of the volume trigger also has the advantage that it filters out some small stocks. Zarowin (1989) attributed the overreaction theory to differences in company sizes and that poor performing stocks were usually small stock. This variable would filter some smaller stocks that could lead to outliers in the data. The data consists of daily returns and trading volume of individual stocks from May 2007 to May 2016. This time span is chosen because contains a lot of interesting moments, for example the financial crisis of 2008 and the market’s recovery from that, and because it is the latest data that was readily available. All of the data used are from CRSP (Center for Research in Security Prices). In total, 2001 different companies that are traded on the New York Stock Exchange have been investigated. Altogether this led to 10763 events.

3. Methodology

To analyze the overreaction effect an event study is used. At the time of the event 𝑡 = 0. Where 𝑡 = 1 is the first trading day after the event, 𝑡 = 2 is the second trading day after the event, etc. The event window is from 𝑡 = −10 to 𝑡 = 20. And the estimation period for the betas is 200 days before the event window.

(13)

13

To compute abnormal returns the Fama and French Three Factor Model is used. This asset pricing model considers differences in risk between the securities and also the movement of the general market. The model looks as follows:

𝑅* = 𝛼*+ 𝛽*./𝑅.− 𝑅01 + 𝛽*2𝑆𝑀𝐵 + 𝛽*6𝐻𝑀𝐿 + 𝑅0+ 𝜀*

Where 𝑅* is the daily return of stock 𝑖. 𝑅0 is the risk-free return and 𝑅. is the market return.

This way abnormal returns can be calculated. The abnormal return is the difference between a security’s actual return and predicted return for a certain day 𝑡 around the event.

𝐴𝑅*< = 𝑅* − (𝛽*./𝑅.− 𝑅01 + 𝛽*2𝑆𝑀𝐵 + 𝛽*6𝐻𝑀𝐿 + 𝑅0)

To calculate the total abnormal returns the cumulative abnormal returns (CAR) has to be calculated. This simply is the sum of all the abnormal returns.

𝐶𝐴𝑅<= @ 𝐴𝑅*<

When the CAR is calculated the cumulative average abnormal return (CAAR) can be calculated by dividing the CAR by the number of events.

𝐶𝐴𝐴𝑅< = 𝐶𝐴𝑅< 𝑁B The t-statistics are calculated using the following formulas

𝑡CD=

𝐴𝑅*<

𝑆CD 𝑎𝑛𝑑 𝑡HCD =

𝐶𝐴𝐴𝑅< 𝑆HCCDI

Where 𝑆CD is the standard deviation of the abnormal returns in the estimation window. And

𝑆HCCDI is the standard deviation of the cumulative average abnormal return.

4. Results and Analysis

4.1 Descriptive Statistics

In Table 1 the descriptive statistics are shown for the sample that is being used. In total the sample contains 10759 events that are collected from 2007 to 2016. The Bank of International published a paper in 2010 where the crisis on the financial markets started in 2008 and in the end of 2009 economic and financial prospects seemed to improve and there were green shoots again. According to this paper the crisis ended in the end of 2009 and therefore, in this study the crisis period is marked as the years 2008 and 2009. This is also in line with what can

(14)

14

be seen in the data. These two years have a relatively high number of events where a stock’s price decreased more than 10% in a day. 2008 had 3276 and 2009 had 2102 points where this happened. This counts for about 50% of the events that occurred in the whole sample period.

The post crisis period is from 2010 to 2016. This period has about 45% of the whole sample size. The year 2010 had a relatively low number of events and this is interpreted as a sign that the crisis period has ended.

The mean decline of a stock’s price on 𝑡 = 0 is 13.152% in the sample. This is very similar to what Cox & Peterson (1994) found in their research. For 8 out of the 10 years the cumulative abnormal returns from the 2 days following the initial price decline was negative. On average the cumulative abnormal returns after 2 days are -0.337% and after 20 days 3.162%. This is inconsistent with earlier research. Atkins & Dyl (1990), Bremer & Sweeney (1991) and Cox & Peterson (1991) all found evidence that in the 2 days following the drop there are positive abnormal returns. This study suggests that the overreaction is not corrected by the market after at least two days.

In half of the years of the sample there are significantly abnormal returns found when holding the stock for 20 days after the drop. It is interesting to see that the years from the crisis period show very significant abnormal returns in the 20 days following the stock price decline, but the years directly following the crisis period do not show significant abnormal returns. 2014 is the only year that shows negative abnormal returns in after the drop that are significant at 5%. 2015 and 2016 then again show positive abnormal returns.

(15)

15

In Figure 1 the cumulative abnormal returns are shown for stocks that showed a -10% or less return for the period surrounding the event. The large decline at 𝑡 = 0 is obviously reflecting the decrease of the stock price that caused the security to be in the sample. Figure 1 shows that in the 20-day period following the large decline the stock price seems to have abnormal returns. Furthermore, it is interesting to see that before the event the cumulative abnormal return floats around 0%, which indicates that holding the stocks before the event would not lead to abnormal returns. And that there are no suggestions that the abnormal returns that are shown after the event are caused by something else than the overreaction phenomenon.

(16)

16

Table 2 shows the mean daily returns and the cumulative abnormal returns from the whole sample. The cumulative abnormal returns and the t-statistics are calculated as how they are described in the section 3. The mean abnormal return is the mean of the abnormal return as calculated in section 3. The mean daily abnormal return is positive for 18 out of the 20 days after the price drop. Also, 13 out of the 20 days following the initial price drop are statistically significant from 0 using a two-tailed t-test at a 0.05 level. Figure 1 and Table 1 show that in this specific time window total abnormal return would be highest at day 19 after the event. The biggest mean abnormal return is found on the 3rd trading day after the drop. The 4th day

also shows high mean abnormal return with a high t-value. This suggests that the 3rd and 4th

days after the drop are very likely to have abnormal returns and thus that the overreaction phenomenon becomes visible.

(17)

17

In Table 3 and Table 4 the daily mean returns and the cumulative abnormal returns are given for respectably the crisis period and the post-crisis period. For both periods the third day shows positive abnormal returns at the 99% level. This result is also very apparent in the whole sample. The reason for this could be that the market does not react instantly to overreaction.

(18)
(19)

19

4.2 Overreaction Hypothesis

From the descriptive statistics it becomes clear that there is some kind of an overreaction that occurs during the initial price drop. Looking at Table 2, the third and fourth trading day after the drop shows that these days have very significant positive abnormal returns and as well the 20-day period following the event show very significant abnormal returns. Furthermore, the fact that the abnormal returns are not significantly different from 0 in the 10 days preceding the event indicates that there was an overreaction of investors to whatever news came out at the event date and not what happened in the 10-day period preceding the drop. The confirmation of the overreaction hypothesis is partly in line with what earlier research suggested. Bremer and Sweeney (1991), Lehmann (1990), Atkins and Dyl (1990) and Park (1995) all also confirmed the existence of the overreaction phenomenon. There is a difference between these studies and this study in the conclusion of the overreaction phenomenon however. The difference lies in the way that the overreaction becomes visible. In this study the overreaction does not become visible within the two days succeeding the drop. In researches like Bremer & Sweeney (1991) and Cox & Peterson (1994) the overreaction phenomenon can be found in the two days following the drop. Results from my research suggests that the overreaction is much more slowly corrected for than what was anticipated earlier.

To answer the question if regular investors have the opportunity to profit from the overreaction phenomenon transaction costs also have to be accounted for. At discount brokers, it is possible to trade US securities with a maximum of 0.5% of the trading value in commission. This means that the fees for buying and selling a security would lead to about 1% of the trading value. From Table 1 and Table 2 combined it becomes clear that investors can attain abnormal returns by buying on the second trading day after the initial drop and selling the 19th trading day after the drop. This would lead to abnormal returns of more than 3.1% in

a period of 17 trading days. Because the bid-ask spread is accounted for using volume, the abnormal returns are higher than 3% and the transaction costs are maximized to 1%, it can be concluded that there is an opportunity for regular investors to attain higher-than-market returns. This conclusion is also not entirely in line with what earlier studies suggested. Because the abnormal returns are adjusted for risk, using the Fama-French Three Factor model, the

(20)

20

idea that the overreaction hypothesis is non-existent when controlling for risk can be rejected. This is what De Bondt & Thaler found in their research in 1987, but Chan (1988) did not agree on this. The size difference as stated by Zarowin (1989) as the cause of the overreaction effect can also be rejected. This is because of the condition that has been put on the volume. Small stocks are therefore excluded from the sample. Furthermore, this research is in line with the research of Lehmann (1990), who controlled for the bid-ask spread and transaction costs to finally conclude that investors do have the opportunity to gain abnormal returns using a contrarian strategy. This study also adds in a market model and both come to the same conclusion. This is not the same what Atkins & Dyl (1990), Cox & Peterson (1991) and Park (1995) concluded from their research. They argue that investors do not have the possibility to gain abnormal returns.

The explanation for why the overreaction phenomenon exists is most logically the explanation that lies in the psychology field. Like De Bondt & Thaler (1985), Piccolo, Chaudhury, Souza and Vieira da Silva (2017) theorize it is expected that investors tend to overvalue recent information and therefore would act too extreme for what was appropriate.

4.3 Crisis versus Post-Crisis

In Table 1 the mean decline on 𝑡 = 0 and the cumulative abnormal returns on 𝑡 = 2 and 𝑡 = 20 are shown for both the crisis period (2008-2009) and the post-crisis period (2010-2016). The crisis period showed a mean decline of 12.502% and the post-crisis a slightly higher mean decline of 13.788%. This is interesting to notice because it would be expected that during a crisis period there would be more higher decliners than in a normal period. Apparently, the crisis period had a more events where the decline is closer to the trigger of -10%.

The interesting thing to notice then is that the crisis period has positive mean cumulative return after 2 days, but the post-crisis period has a negative return. The positive return is in line with earlier research like Bremer & Sweeney (1991), but the negative return is not. The discrepancy between the crisis period and the post-crisis period in this time frame could be because of investor sentiment. Like Piccolo & Chaudhury (2018) pointed out in their study, when investor sentiment is low the overreaction phenomenon is more pronounced. This could also mean that the initial price drop is relatively even higher than what should be, so this is corrected for faster.

(21)

21

The mean cumulative return after 20 days also differs a lot between the two periods. The crisis period shows abnormal returns of 4.846% and the post-crisis period shows abnormal returns of 1.307%. Both these abnormal returns are significant at 99%. It is interesting to see the big difference between the abnormal returns. This discrepancy is also in line with the earlier mentioned studies of Piccolo & Chaudhury (2018) and what Malkiel (2003) asserted in his study.

Because of the crisis period showing a larger and significant reversal of the price both after 2 days and after 19 days than the post crisis period. It can be concluded that the overreaction effect is more pronounced during crises.

(22)

22

5. Conclusion

Evidence is found of an overreaction effect when a stock decreases 10% or more. Investors seem to have the opportunity to gain abnormal positive returns when buying the security after the price drop and selling after 19 trading days. This conclusion is partly in line with earlier studies. Like earlier studies there is evidence found of an overreaction effect, however, unlike earlier studies the price reversal is not significant until the third trading day after the initial price drop.

Furthermore, there is a discrepancy between some earlier studies like Atkins and DyI (1990), who proved the existence of the overreaction phenomenon, but debunked the hypothesis that traders could gain abnormal returns for this. I also found evidence that the overreaction effect is larger in crisis times like the 2008-2009 financial crisis. This is in alignment with what Malkiel (2003) hypothesized in his study. He asserts that there is a chance that the market is less efficient during periods of crisis. Also, this is in line with the recent study of Piccolo and Chaudhury (2018) who found that the overreaction phenomenon is more pronounced when investor sentiment is low, like in a crisis.

The most logical explanation for the overreaction hypothesis would be the one that De Bondt and Thaler (1985) suggested in their study. This was the suggestion that investors tend to lay extra weight on recent information which would leave to irrational behavior and the overreaction effect.

For future research it would be interesting to also look at other crises and see if those periods also show larger and more significant abnormal returns. Also, in this study volume is used as a measure for liquidity and therefore the bid-ask spread. In future studies the real bid and ask prices could be used.

(23)

23

6. References

Alwathnani, A.M., Dubofsky, D.A. & Al-Zoubi, H.A. (2017). Under-or-overreaction: Market responses to announcement of earnings surprises. International Review of Financial Analysis,

52, 160-171.

Atkins, A.B. & Dyl, E.A. (1990). Price Reversals, Bid-Ask Spreads, and Market Efficiency. The

Journal of Financial and Quantitive Analysis, 25 (4), 535-547.

Bai, M. & Qin, Y. (2015). Short sales constraints and price adjustments to earnings announcements: Evidence from the Hong Kong market. International Review of Financial

Analysis, 42, 304-315.

Bank for International Settlements. (2010). The international financial crisis and policy

challenges in Asia and the Pacific, 52. Monetary and Economic Department.

Bremer, M. & Sweeney, R. J. (1991). The Reversal of Large Stock-Price Decreases. The Journal

of Finance, 46 (2), 747-754.

Bremer, M., Hiraki, T. & Sweeney, R. J. (1997). Predictable Patterns after Large Stock Price Changes on the Tokyo Stock Exchange. Journal of Financial and Quantitative Analysis, 32 (3), 345-365.

Chan, K. C. (1988). On the Contrarian Investment Strategy. The Journal of Business, 61 (2), 147-163.

Cox, D.R. & Peterson, D.R. (1994). Stocks Returns Following Large One-Day Declines: Evidence on Short-Term Reversals and Longer-Term Performance. The Journal of Finance, 49 (1), 255-267.

(24)

24

De Bondt, W. F. M. & Thaler, R. H. (1985). Does the Stock Market Overreact? The Journal of

Finance, 40 (3), 793-805.

De Bondt, W.F.M. & Thaler, R.H. (1987). Further Evidence on Investor Overreaction and Stock Market Seasonality. The Journal of Finance, 42 (3), 557-581.

Fama, E.F. (1998). Market Efficiency, Long-Term Returns, and Behavioral Finance. The Journal

of Financial Economics, 3 (1), 283-306

Keynes, J.M. (1936). The General Theory of Employment, Interest and Money. London: Harcourt Brace Jovanovich.

Lehmann, B.N. (1990). Fads, Martingales, and Market Efficiency. The Quarterly Journal of

Economics, 105 (1), 1-28.

Lo, A.W. & MacKinlay, A. C. (1990). When Are Contrarian Profits Due to Stock Market Overreaction? The Review of Financial Studies, 3 (2), 175-205.

Malkiel, B.G. & Fama, E.F. (1970). Efficient Capital Markets: A Review of Theory and Empirical Work. The Journal of Finance, 25 (2), 383-417.

Malkiel, B.G. (1981). A Random Walk Down Wall Street. New York: Norton.

Malkiel, B.G. (2003). The Efficient Market Hypothesis and Its Critics. The Journal of Economic

Perspectives, 17 (1), 59-82.

Milian, J.A. (2015). Unsophisticated arbitrageurs and market efficiency: Overreacting to a history of underreaction. Journal of Accounting Research, 52, 175-220.

Park, J. (1995). A Market Microstructure Explanation for Predictable Variations in Stock Returns Following Large Price Changes. The Journal of Finance and Quantitive Analysis, 30 (2), 241-256.

(25)

25

Piccolo, P., Chaudhury, M., Souza, A., Vieira da Silva, W. (2017). Stock overreaction to extreme market events. The North American Journal of Economics and Finance, 41, 97-111.

Piccolo, P., Chaudhury, M. (2018). Overreaction to extreme market events and investor sentiment. Applied Economics Letter, 25 (2), 115-118.

Zarowin, P. (1989). Does the Stock Market Overreact to Corporate Earnings Information? The

(26)

26

Referenties

GERELATEERDE DOCUMENTEN

The results in the first 16-year test period however, do not support the overreaction hypothesis since the abnormal return of the winner portfolio is larger

Stakeholder involvement in the delivery of education programmes has been embraced by many successful schools. In order to benefit from stakeholders and partners alike, schools

Volterra series are used to analyze the amplitude of the non-linear contributions and their spatial distribution dynamically.. Due to transport there is a delay between the

Expert-assessment identified task support, perceived persuasiveness, unobtru- siveness, credibility, perceived effort and perceived effectiveness (as defined in the Persuasive

Hence, model 3 is run for both variables separately (this is shown in appendix 1; model 3.1 and model 3.2). higher intellectual or executive) and the other job

To assess the level of downward communications, we analyzed the number of followers from regional departments of the 5 national police twitter accounts?. As could be expected, these

Several techniques which have been used to increase the performance of the metal oxide semiconductor field effect transistor (MOSFET) are also applied to the FinFET; such as

cognitive screening instrument with a strong theoretical foundation, tested in a relatively large population of ALS patients, healthy control participants, ALS-FTD - and FTD