Kandidatafhandling fra London School of Economics om alpha afkast til investorer
Kort version omtalt i pressen til dig der ikke ønsker at læse den fulde afhandling
Den fulde afhandling til dig der gerne vil forstå detaljerne
Alpha returns by Insiders - London School of Economics
MSc Finance (Full-time); Topics in Portfolio Management; Title: Alpha Returns by Insiders; Year: 2009-2010; Word count: 6499
“The copyright of this dissertation rests with the author and no quotation from it or information derived from it may be published without prior written consent of the author.”
Alpha returns by Insiders, This online version does not include the used dataset or datatables.
Abstract In this paper I investigate whether insiders are capable of earning an abnormal return in terms of a CAPM alpha return from investing in their own company stock. The conducted analysis was based on trades done by insiders in Danish blue chip companies in the period from April 2005 – April 2010. The analysis is the first attempt done to research this topic for the Danish market. The dataset used for this paper is a unique and new dataset that I created manually by reading through filling from the included companies to the Nasdaq OMX Nordic. My finding and conclusion is that insiders are capable of earning a substantial alpha return. This shows that the Danish market is not strong form efficient. This study shows that the private information reveled by the trades of insiders is valuable. Non inside investors in the market should therefore pay attention to trades conducted by insiders and use this information when evaluating a stocks potential future return.
Introduction The topic of legal trades performed by insiders has only been given very little attention in academic literature. Due to the insiders strong knowledge of the company it is likely that the insiders would be better capable of predicting future returns of the stock of the company for whom they are working than the average investor in the market. A legal framework has been established in most markets to protect other actors in the market place from insiders trading on private information. Nonetheless it is still possible that insiders trading in the market could have superior information on the future prospects of a company stock. Should this be the case other investors would be wise to keep strong attention to trades conducted by insiders. An investor should therefore be reluctant to invest in a company where an insider recently has been selling stocks, and be positively biased to invest in companies where an insider recently has been buying.
The research question that will be investigated in the paper will be whether insiders earn positive alpha returns in the market. This will be tested on a twelve and three month horizon. The following four hypothesizes will be tested: Hypothesis 1a: Following a buy by an insider the stock bought will outperform the market over the following twelve months. Hypothesis 1b: Following a buy by an insider the stock bought will outperform the market over the following three months. Hypothesis 2a: Following a sale by an insider the stock sold will underperform compared to the market over the following twelve months. Hypothesis 2b: Following a sale by an insider the stock sold will underperform compared to the market over the following three months.
The reason why this time horizon was chosen was to test whether an insider trade can be expected to revel valuable private information on the company’s fundamentals and hence the future performance of the stock.
The analysis has been based on stocks traded at Nasdaq OMX Nordic. The stocks included in the analysis have been stocks included in the Danish C20 index at April 1st 2010. The time period analyzed is from April 1st 2005-April 1st 2010. The underlying dataset used to create this analysis had previously not been created by others. I created the dataset used for this paper by manually reading through all filings to the Nasdaq OMX Nordic including information on insider trades or holdings of insiders.[1] In total 4007 filings were read. In the analysis 601 insider trades will be investigated on a risk adjusted basis. To do this 601 regressions have been conducted to determine alpha returns. The findings from the study show a clear pattern that abnormal alpha returns are earned by insiders buying and selling. The implication of this study is therefore to suggest that actors in the market should pay attention to the trades conducted by insiders, since these might reveal significant information about the future performance of the stock.
Literature review The topic of abnormal returns related to insider trading is largely uncovered for most markets. Until recent years very few papers have been written on the topic of abnormal returns or alpha returns from insider trading and related topics. Due to the low attention the topic has had, the results by following insiders are generally not discussed in academic or practical books of portfolio theory or asset management. Further very few papers have been written on the topic. However some sporadic attention has been dedicated to the topic in the last decades. As more managers’ compensation plans include warrants and stock options the level of attention has been slowly growing (Fest, 2008).
Older studies have shown that abnormal returns can be earned by insiders (Jaffe, 1974; Finnerty, 1976,1; Rozeff and Zaman, 1998; Lin and Howe, 1990). More recent papers have also shown a tendency that insiders can earn abnormal returns (Jiang and Zaman, 2009). The assumption by the papers is that insiders earn their abnormal returns by having better knowledge of their companies’ future prospective than does the average actor in the market. Ke et al., have investigated the topic even further and found that this abnormal return steamed from insiders being better at timing the market due to their superior knowledge (Ke et al., 2003). A study for the UK further analyzed in which periods insider trades contained most information, and found that the amount of ownership by directors had an influence on the size of abnormal returns. (Fidrmuc et al, 2006). Further it has also been proven that the market seems to be able to detect also illegal trading and incorporate this into stock prices (Meulbroek, L.K, 1992).
Other studies have been more skeptical about the abnormal returns earned by insiders. In a controversial paper on the Swedish market from 2009 Kallunki et al. did not find that abnormal returns were on average earned by Swedish insiders. In this paper it was concluded that insiders trade with many more motives than earning abnormal returns by exploiting private information. Other motives for insiders trading are found to be balancing objectives, behavioral biases and tax considerations (Kallunki, 2009). No academic papers have investigated abnormal returns by insiders in the Danish market.
Alpha returns by following trades of insiders have also in recent years found its way to the media outside the trading rooms and the academic world. Several news channels and news papers have explained how investors could have benefitted by following the trades of insiders (Song and Paul, 2003; Francis and Hays, 2002). In Canada a webpage under the name www.canadianinsider.com has been created with the purpose of surveying the trades of insiders (Johnston, 2005). Some media coverage has however been skeptical towards the potential of following trades of insiders (Adamson, 2001).
Insiders do clearly have access to information that other actors in the stock market do not posses. Assuming that this information could be used to earn higher returns by trading with uninformed investors this issue has been given a lot of attention by sovereign legislative bodies. Out of a total of 103 countries with stock markets, 87 countries have made regulations on insider trading (Bhattacharya and Daouk, 2000). It has continuously been discussed whether insiders should be allowed to trade in their own stock (Minenna, 2003, Meulbroek, 1992). Arguments for why insiders should or should not be allowed to trade in their own stock can be shortened to three main arguments for and against (Bainbridge, 1998). The arguments against insider trading have been that use of knowledge by insiders would be theft of knowledge from the firm(Georges, 1976), a market egalitarianism theory stating that all investors should trade based on equal information (Langevoort, 1987) and finally that trades by insiders with superior knowledge could harm the integrity of the market place (Bhattacharya and Daouk, 2000). Arguments defending insider trades have been that trades by insiders have no victims since it pushes prices to the direction of the preferential information(Herzel and Katz, 1987), that it is an important way to compensate managers (Manne, 1966) and that insider trading is helping to increase market efficiency (Finnerty, 1976,2). This was also found by Meulbroek, who stated that insider trading leads to private information being incorporated into stock prices and therefore results in more informative prices (Meulbroek, 1992).
Overall in the few papers that have been written on the topic it appears that some abnormal returns are earned by insiders. However since this topic is still in its infancy future studies with a different outcome could quickly change the conclusion. Abnormal returns by insiders have only been investigated at a higher academic level in the US and few other OECD markets. Much more knowledge is therefore to be discovered and clarified in the coming decades. Since it might appear to be the case that insiders earn abnormal returns this is a highly interesting topic that deserves more attention. The ambition with this paper will therefore be to fill out this gap for the Danish market and to create new knowledge on this relatively undiscovered topic.
Regulations on legal insider trading Protection of investors from the potential negative impact of insiders using superior information in the market has been incorporated differently across countries. In the United States the first rules regulating insider trading were created with the Securities Exchange Act from 1934 (Securities Exchange Act, 1934; Minenna, 2003). The rules have continuously been changed and were recently tightened in connection with the passing of the Sarbanes-Oxley Act of 2002 (Brochet, 2010). In the European Union insider trading is regulated by an ECC Directive from 1989 (ECC Directive, 1989). The regulation covers all trades in markets that are considered to be open to the public and operates regularly. Each member state is allowed to enforce prohibitions on trades related to securities that are allowed to be traded on a market in the member state (Minenna, 2003). Different countries have passed varying laws on the topic within the European Union. In the United Kingdom insider trading is regulated through the Criminal Justice Act (1993) and the Financial Service and Market Act (2000).
In Denmark insider trading is governed by the Law on Trading of Securities[2], that was latest updated in 2005 (Bekendtgørelse af lov om værdipapirhandel m.v., 2005). It is here stated that managing employees for companies that have issued securities to be traded on an exchange or have asked to be listed on an exchange have to notify the company of any trades in the company’s stock or affiliated securities. This has to be done at the latest one day after the trade. Leading employees are only allowed to trade in certain time periods and are only allowed to trade if all material information that could change the stock price economically significant has been revealed to the market. It is then required that the issuing company at the latest the following trade day notifies the supervisory body (Finanstilsynet). The supervisory body then immediately makes the information available to the public.
By a leading employee in the company is understood a member of the direction, a member of the board or a member of a supervisory body connected to the company. Other leading employees, who have access to confidential information on the company, and who have competence and authority to take decisions that have material importance for the company’s future are also included by the directive. Further close relatives to any of the mentioned above are also included in the directive. This includes but is not restricted to family members, companies owned by relatives or by others covered by the directive and people living in the same household.
Data and Summary Statistics Creation of the data set The dataset that this paper is based upon was not previously gathered for other studies when I started the project. As an initial starting point the data therefore had to be created. The companies included in this paper are the stocks listed on the Nasdaq OMX Nordic included in the C20 index. The index included the 20 most traded stocks in the Danish market. The time interval investigated is the five year period from April 1st 2005 until April 1st 2010.
No database or list exists of all insider trades in the Danish market. The data was therefore manually created by reading through 4007 fillings from the included companies, resulting in more than 1100 initial registered insider trades. Fillings including insider trades were then recorded. The trades registered were recorded in four categories: Insiders clean buys, Insiders clean sales, Insiders dirty buys, Insiders dirty sales. The first category, Insiders clean buys included all buys were insiders had been buying regular stocks at the exchange at the listed market price. The second category, Insiders clean sales included all sales were insiders sold company stock in the open market that had not been bought on special terms within the latest month. The third category Insiders, dirty buys included all the clean buys, but also purchases that were non regular open market purchases. This could be but is not excluded to conversion and utilization of options and warrants. Stocks bought as a part of the company’s compensation plan, stocks sold at a preferential price, stocks distributed to employees and gifts. The stocks in the fourth category, Insiders dirty sales included all clean sales, but also stocks bought at different terms than the market price and sold within the following month. This could be but is not excluded to converted and utilized options and warrants that were after a short time interval sold in the market. The reason why clean open market buys and sells were separated from dirty transactions was that I thought that insiders potentially could be expected to have different motives for e.g. an open market buy and the utilization of e.g. an expiring option. I therefore thought that it could be the case that a stronger signal could be included in stock purchases and sales done at market terms.
Criteria’s for accepting insider trades as data points Not all data points were accepted as valid for this paper. Data points were discarded for several reasons. Reasons for discarding were the following: Insiders had traded with relatives and the net position exchanged equaled out. Insiders had sold or bought stock from companies partly or totally owned by the insider. Distributions of warrants and options were not included. Buys and sales of drawing rights were not included. If an insider had been both buying and selling the net change in the position were recorded and several transactions were then combined into one. Transactions in which an insider and the insiders’ relatives had conducted many or complicated diverse directed trades were discarded, if it was not clear whether the insider was attempting to increase or decrease the exposure to the stock. When an insider bought stocks by converting options and quickly sold the stock within the following month this was recorded as a dirty sale. If the insider held onto the stock and then sold the stock at a later point in time this was recorded as a clean sale. Finally insider transactions performed in the last year of the analysis was not used since it would not be possible to track the performance in a full year following the transaction. After setting up these boundary conditions a total of 601 insider trades were accepted.
The stock data The data on the stocks were obtained from the Nasdaq OMX Nordic. Adjusted stock prices were used to take into account corporate actions. When working with the data, I observed some irregularities with the data of the stock DSV, where a stock split in May 2007 was not incorporated into the data. Following this I contacted the Nasdaq OMX Nordic that confirmed that there was a mistake in the data provided by the Nasdaq OMX Nordic. The discrepancy found was then corrected manually in the data set. One stock that filed insider trades in another country was discarded[3]. To compare the returns of the stocks the OMX Copenhagen Benchmark Gross Index was used. The index includes all the largest stocks and most liquid stocks in the Danish market. A gross index was used to take into account dividends. All companies used in the collected data are large liquid stocks included in the index. A potential other index that could be used would be the OMXC20 that is narrower and also includes all the stocks used for the analysis. However data on this index were only available as a PI Index not taking into account dividends. Therefore the Copenhagen Benchmark Gross index was more appropriate to use.
The stock data for the trades accepted are summarized in table 1 below[4].
*Three open market trades had an abnormally high value and are not representative for the general purchases and an adjusted value is therefore given in the table.
The total numbers of trades accepted were 601 trades. The total number of stocks traded was 4.7 million. The average number of stocks in a trade was 7,765. The total value of stocks traded in the time investigated was £352 million, with an average value pr trade of £ 586,266.
From the dataset it can be seen that that the most common transaction in the period was an open market buy. 290 data points where insiders bought stocks in the observed period of time were accepted. The corresponding number of sales were only 141. The average value of the buy transactions was £ 651,509, while the average value of the sales was £ 427,631. The value for buys is however misleading due to three very high open market buys. When adjusted for these three trades the average value of the open market buys drops by 64% to £ 232,271. From the data it can therefore be seen that buys are more common among insiders than sales[5]. However the average value of selling transactions is generally higher than the value of the buy transactions. When considering dirty buys it can be seen that this category includes 60 extra trades compared with clean buys. The dirty sales category includes 110 more trades than the clean sales. From the data it can therefore be seen that it has been more common that insiders sell their position rather than holding after using options and warrants. The value of the average buys including dirty trades has been £ 204,218, while the average sell has been worth £ 587,596. This indicates that managers are more likely to use options, warrants and other positions and hold the stock when the trade is of a smaller value, while when the value of options and warrants reaches a higher value insider are more prone to sell their positions. This could be an indication that insiders are not willing to hold a too large share of their wealth in the company for which they are working since they are already by their expected future income highly exposed to the future viability of the company. The difference in size of the positions bought and sold could also indicate that insiders wish to diversify their holdings if their utilization of options and warrants results in an overweight of the respective stock in the average insiders portfolio.
The average value of the trade is smaller than the value of trades from other markets. In the US the average value of insider sales was in the period from 1995-1999, a study 10 year older than this, determined to be above £ 662.000[6] ($ 1.031 million) (Aktas et al., 2008). The reason why insider trades in Denmark might be of a lower value than in other countries could potentially be thought to occur from a general lower level of compensation for top management compared with other countries (Business, 2006).
Methodology The model An important assumption behind the study is that insiders engage in open market sales and buys with the purpose of utilizing information to earn abnormal alpha returns. To see if insiders were able to earn a significant abnormal return a comprehensive scientific approach was followed.
When determining a way to measure the performance of insider trades it is important to differentiate between, what is noise and what is actual performance by the insider trading. When just looking at the return of the insiders trades without a measure of comparison the picture does not reflect the performance of the insider. The return could be a result of market movements that do not have anything to do with the insiders stock picking ability. Further many investors are risk averse and require a premium to hold more risky assets. To include this in the model market risk was adjusted for when determining whether a stock had delivered an abnormal return.
The model was created based on regressions. The daily return for the “stock i” over a given twelve month interval was regressed on the return of the Copenhagen Benchmark Index to determine if a positive alpha value had been earned during the following three or twelve month period. This was measured by the CAPM alpha value. The regression process was complicated by the different relevant time horizons for each insider trade. The trades performed by the insiders were done at different dates. A new data list for stock returns and index returns with the relevant following three or twelve months was therefore created separately for each regression. Following this it was possible to make the 601 regressions separately. The abnormal alpha return was then given by:
The model assumes that CAPM holds. The average abnormal alpha return over all the included data points was calculated by taking the average value for all data points. To test the significance of the findings the t-stat value was calculated. This was done by taking the average value of the alpha return and divide with the standard error of α[7].
To test if the results found were general across all sectors the analysis was also done on the separate industry groups. The same method was followed to see if abnormal returns were clustered in certain industries as for the full data sample.
When insiders trade the trade was typically filed to the exchange on the day of the trade or the day after. Trades were therefore generally available to the public within the first or second trading day. However in some instances the insiders trades were published later. The starting date used to determine the abnormal return in this study is the day that the insider traded. This was chosen because the price on the trading day is the price that the insider based the trading decision upon. Another approach could be to investigate abnormal returns from the day om which information about the trade was made public. This would then not show the alpha returns earned by insiders, but show the alpha returns that other actors in the market could have earned by replicating the trades of insiders, when information about the trade became public. Both approaches were tested and based on the given data set the difference in the results were minimal.
To test for clustering of alpha returns within industries the investigated stocks were grouped as can be seen in table 2. Analyses were then conducted on each industry separately.
Table 2
Results and discussion Full sample The result found for alpha returns can be seen in table 3.
Table 3
* Significance on 90% level. **Significance on 95% level. *** Significance on 99% level.
When looking at the “Clean Buys 1 year” it can be seen that insiders were able to earn substantial alpha returns. The average alpha return earned by insiders was 4.2% on a yearly basis. An open market buy by an insider must therefore be considered a strong indication that the stock will outperform the market in the following 12 months. The t-value found is very high. The result can therefore be proven to be significant on a 99% basis. The 99% confidence interval shows that alpha returns by following insiders are expected to be in the interval 1.2%-7.2%. This result is in line with returns found in other studies in other markets (Aktas et al., 2008). Hypothesis 1a can thereby be confirmed that following an insider buying the stock will outperform the market in the following twelve months, when looking at the “Clean Buys 1 year” category. The “clean buy three months” trades can be seen to give an expected positive alpha return of 2.8%. However t-stat value for this result is relatively low. Their therefore appears to be a trend towards a positive alpha, but hypothesis 1b cannot be proven statistically significant on a high confidence level.
When looking at the “Clean Sales 1 year” it can be seen that on average the stocks sold underperformed the market by 3.4% in the following twelve months. The result is also found to be significant on a 99% level, and the result found must therefore be considered to be very strong. When looking at the 99% confidence interval it can be seen that the stock is expected to underperform the market with between 0.2%-6.6%. The implication of this result is that stock owners in companies where insiders are selling should take insider sales as a strong signal that the stock will underperform the market in the following twelve months. Investors could therefore on average do well by reducing their position after an insider sale. Hypothesis 2a can therefore be shown to hold for “Clean Sales 1 year”. In the shorter three month period the result is also positive and statistically significant on a 95% confidence level. Hypothesis 2b can therefore also be confirmed.
When looking at the “Dirty Buys 1 year” the average alpha return can be found to be 4.1%. This is slightly below the value found for the clean buys. This is however nonetheless still a strong outperformance, indicating that even when insiders buy stocks at preferential terms and hold them an alpha return can be expected. On a 99% confidence level this amounts to an alpha return of 1.5%-6.7%. This therefore shows that hypothesis 1a does not only hold for Clean Buys, but also holds for buys on non market terms on a 1 year basis. On a three month basis the result is positive with an average alpha return of 2.6%. The result is however less certain and hypothesis 1b can only be proven significant on a 75% confidence level.
When considering the Dirty Sales the picture however changes. The “Dirty Sales 1 year” on average underperforms the market by 0.1%, which indicates that these stocks almost perform equivalently to the market. A sale by an insider when including non market terms trades is therefore not an as good indication as a clean sale. The result for the dirty sales is not significant on either a 75% or a 90% confidence level. For the three month period the result is slightly better but still not highly statistically significant. Neither hypothesis 2a nor 2b can therefore be confirmed to hold for “Dirty Sales”. The generally weak results found for dirty sales could potentially be explained by the fact that insiders conducting dirty sales might not be selling due to an expectation that the stock will underperform the market, but could be selling for other reasons such a balancing the insiders portfolio. This is in line with the suggestion given for insider trades in the investigation made by Kallunki et al. in the Swedish market (Kallunki et al., 2009). If an insider is compensated with options, warrants and other stock related instruments it is likely that the insider would potentially not have chosen to allocate a that large fraction of his wealth in the stock and would therefore decide to sell for diversification reasons even though the firm does not have bad future prospects.
In total there seems to be a strong case for insiders outperforming the market on a one year basis. When looking at the three month the same trends seems to be present that insiders do earn abnormal returns, but the results are less significant. This could be an indication that the trades of insiders are based on their long term fundamental expectation to the company’s performance.
Results separated into categories Table 4
*Significance on 90% level. **Significance on 95% level. *** Significance on 99% level.
In table 4 the data has here been used to calculate alpha returns return in various industries represented in the C20 index to see if some insiders are better at predicting future performance than others.
When looking at the financial stocks it can be seen that “Clean Buys 1 year” insiders do on average earn an abnormal return of 1.8% but this is not significant. When looking at the three month period it can however interestingly be seen that the alpha value is negative indicating that managers in the short run underperformed, but over performed on the 1 year basis. I tried to investigate this finding in the dataset and found a pattern that financial managers had conducted a large number of trades towards the end of the financial crisis, but some months before the markets reversed, which then resulted in a short run underperformance. A hypothesis could be that these managers might have been able to see the reverse of the financial stocks before the market, but where not capable of optimally timing their investments.
For the “Clean Sales 1 year” the results can be proven to be significant on a 95% confidence level and investors can expect to see the stock underperform the market by 0.2%-12.8%. For the “Clean Sales 3 moths” financial stocks also underperformed. This was found significant on a 90% confidence level. Investors could therefore expect to see a very large under performance in financial stocks after insiders have been selling. When investigating this finding further in the data I did some interesting observations. Before the financial crisis had strongly caught momentum substantial sales was done by the management within the banking sector. Investors could therefore have protected themselves from devastating looses on financial stocks if the transactions done by insiders had been followed.
When looking into the industrial sector a very large outperformance can be seen when insiders buys either Clean open market buys or Dirty Buys on both a three month and one year horizon. The average outperformance seen in the sample is at overwhelmingly 15.8% for the one year period and 19.9% on a three month period. Insiders therefore clearly seem to be able to predict future increases in stock prices. When considering the stock included in the industrial segment in the C20 index it can be seen that many of these companies are selling large projects e.g. windmills and plants for the mineral and cement industry. A part of the explanation why these insiders earned those high alpha returns could be that the time period between start of negotiations, to the signing of a final contract could be very long within those industries. Insiders could therefore be expected to have a better idea of the level of new contracts the company would win within the upcoming three months to one year than the consensus by analytics in the market.
In the pharmaceutical industry an interesting result was found under “Dirty Sales 1 year”. Here the stocks sold by insiders actually outperformed the market by 3%. When investigating the Dirty Sales conducted by insiders in the pharmaceutical industry it could be seen that a large share of these sales were connected to pharmaceutical companies that has stock options accounting for a very substantial part of the compensation plan. These sales could therefore very well be reflecting insiders wanting to decrease the relative share of their positions invested in the company. When looking at the three month period the over performance of Dirty Sales disappears and reverses to a negative average alpha of -2%. This could indicate that these portfolio allocations are timed with insider expectations of short run underperformance. When looking at the three months clean sales, insiders in the pharmaceutical industry performs better than the market, which again points towards these insiders in the pharmaceutical industry might be able to outperform the market in the short run when selling.
In the transport sector insiders can be seen to outperform the market when selling both “Clean Sales” and “Dirty Sales” on a one year horizon. When buying on a 1 year basis or in any three month category the insiders of transport companies are however not capable of producing an alpha return. In general insiders in transport companies are therefore less capable of predicting their stock returns than the average insider in this analysis. A potential explanation for this could be that the return of these stocks are highly affected of the overall state of the economy and less by company specific information compared with e.g. the industrial segment. The same could be said about the category utilities, where insiders are capable of outperforming when buying on a one year basis, but in no other categories.
Finally in the category others, insiders buying can also be seen to outperform the market. When investigating this data I found that the companies located in this category have the common denominator that they are very innovative companies. These companies could maybe be thought to have a good idea about the prospects for their new products which could be utilized in the market place.
In total there seems to be a case for following insiders within specific industries. However when separating the data into different categories, this results in a higher standard error due to the lower number of observations in each category compared with the full study. This makes it more difficult to prove statistical significance. Altogether when looking at the result for the entire test sample the message seems to be crystal clear, insiders do earn abnormal returns in the market on a twelve month horizon and on a lower confidence level there is also an indication that insiders outperform on a three month horizon. Other investors should therefore keep an eye on the trades of insiders.
Implication In total the results found are very strong and clearly shows that markets cannot be thought of as strong form efficient, since insiders with private information are able to earn abnormal returns. The implication of this study is therefore that an investor can on average obtain a positive alpha return by following the trades of insiders. An investor would therefore be wise to keep his eyes open, and takes into account the trades conducted by insiders. Following insiders could therefore be a way to increase the expected return without taking extra risk. Over time when more information is published about abnormal returns earned by insiders it is likely that more actors in the market would pay more attention to the insiders’ trades. This could result in stocks immediately increasing in value after a buy by an insider and in immediate falls following a sale. This could in the future potentially reduce the value of following insiders, but for now a consistent alpha return seems possible. Based on the findings from this study a potential trading strategy could be to set up a zero cost portfolio going long insider clean buys and short insider clean sales on a 1 year horizon.
Conclusion The returns earned by insiders trading in the market have generally been given very little attention in academic literature. A potential explanation for this could be that legislation has in many countries been passed to protect other investors from insiders using materially important information to trade in the market. A potential perception could therefore be that insiders would not be capable to earn abnormal returns in the market due to the legislation. By looking at this first study of insider trades in the Danish market it can be seen that insiders are capable of earning a substantial alpha return. The results are stronger when looking at a twelve month horizon compared with a three month horizon. This could be an indication that insiders trading decisions are based on their expectations of company’s long run fundamentals. The conclusion from this study is that investors should pay considerable attention to trades by insiders and use this information actively. This paper therefore proves that the Danish C20 index is not strong form efficient, and by following the trades of insiders investors can utilize the insiders’ private information to obtain a positive alpha return.
Limitations and suggestions to future studies:
Due to the time consuming nature of the manual data construction by reading filings of companies the paper was limited to stocks in the C20 Index including only blue chip companies. In future studies it could be interesting to include a larger number of stocks to increase the number of trades and thereby being able to make a more general conclusion covering the entire market place.
The risk adjusted returns used in this study incorporates market risk. A study in Hong Kong from 1993-1998 showed that a substantial part of the abnormal returns were earned in transactions by insiders in smaller companies (Cheuk et al., 2006). In future studies if data should be available at that point in time it could be interesting to investigate whether other risk factors could partly explain the abnormal returns earned by insiders.
Acknowledgements I would like to thank Dr. Michela Verardo from the Financial Markets Group at London School of Economics for her constructive comments and suggestions that helped to structure this study.
References • Adamson, D., 2001. When insider stock purchases mislead. CBS MarketWatch. June 11.
• Aktas et al., 2008. Legal insider trading and market efficiency. Journal of Banking & Finance 32, 1379-1392.
• Bainbridge, S., 1998. Insider trading: an overview, Encyclopedia of Law and Economics, Nber.
• Bekendtgørelse om lov om værdipapirhandel m.v., 2005. LBK nr 843 af 07/09/2005.
• Bhattacharya, U., Daouk, H., 2000. The World Price of Insider Trading, Working Paper.
• Brochet, F., 2010. Information Content of Insider Trades before and after the Sarbanes-Oxley Act. The Accounting Review vol. 85, No. 2, 419-446.
• Business, 2006. Topchefer tjener stadig for lidt. Berlingske Tidende, 23-06-2006. First entered 13th of April 2010. http://www.business.dk/diverse/topchefer-tjener-stadig-lidt
• Cheuk, M.-Y. et al., 2006. Insider trading in Hong Kong: Some stylized facts. Pacific-Basin Finance Journal 14, 73-90.
• ECC Directive n. 89/592 coordinating regulation on insider trading, 1989, European Union.
• Fest, G., 2008. Overseeing “Legal Insider Trading. Bank Technology News. February 2008.
• Fidrmuc et al., 2006. Insider Trading, New Releases and Ownership Concentration. The Journal of Finance, Vol. LXI, No. 6, 2931-2973.
• Finnerty, J., 1976,1. Insiders’ activity and inside information: a multivariate analysis. Journal of Quantitative Financial Analysis 22, 109-126.
• Finnerty, J., 1976,2. Insiders and Market Efficiency. Journal of Finance 31, 1141-1148.
• Francis, B., Hays, K., 2002. How to Profit from Legal Insider Trading. CNNFN August 2002. August 15, 05:00PM.
• Georges, S., 1976. L’utilisation en bourse d’information privilegees dans le droit des Etats-Units, Economica.
• Herzel, L., Katz, L., 1987. Insider Trading: Who Loses? Lloyds Bank Review.
• Jaffe, J., 1974. Special and insider trading. Journal of Business 47, 410-428.
• Jiang, X., Zaman, M., 2009. Aggregate insider trading: Contrarian beliefs or superior information. Journal of Banking & Finance 2009, doi:10.1016/j.jbankfin.2009.11.016
• Johnston, K.,2005. Website tracks legal insider trading. The Leader-Post. Section: Business & Agriculture.
• Kallunki, J. et all, 2009. Why do insiders trade? Evidence based on unique data on Swedish insiders. Journal of Accounting and Economics 48,37-53.
• Ke, B., Huddart, S., Petroni, K., 2003. What insiders know about future earnings and how they use it: Evidence from insider trades. Journal of Accounting and Economics 35, 315-346.
• Langevoort, L., 1987. Regulation, Enforcement and Prevention. Clark Boardman Company, New York.
• Lin, J., Howe, J., 1990. Insider trading in the OTC market. Journal of Finance 45, 1273-1284.
• Manne, H., 1966. Insider trading and the stock market. Free Press, New York
• Michaud, R., 1989. Financial Analyst Journal. Vol. 45, No 1. January-February, 31-42.
• Minenna, M., 2003. Insider trading, abnormal return and preferential information: Supervising through a probabilistic model. Journal of Banking and Fiance 27, 59-86.
• Meulbroek, L. K, 1992. An Empirical Analysis of Illegal Insider Trading. Journal of Finance, Nr. 5 Dec. 1992.
• Rozeff, M., Zaman, M., 1998. Overreaction and insider trading: evidence from growth and value portfolio. Journal of Finance 53, 701-716.
• Sarbanes-Oxley, 2002, section 403, USA.
• Securities Exchange Act of 1934, USA.
• Song, J., Paul, B., 2003. Cover Story: Insider watching. The Edge Singapore
[1] Filings from April 1st 2005- April 1st 2010 were read for the companies included in the analysis.
[2] Translated from Danish. Danish name: Bekendgørelse af lov om værdipapirhandel m.v..
[3] The stock discarded was the Nordea Bank. This company filed in Sweden.
[4] The conversion of DKK (Danish Kroner) to GPB (Great Bristish Pound) was based on the prevailing exchange rate at noon the 12th of April 2010.
[5] This would also be the case if all insider trades had been accepted as data points.
[6] With the exchange rate Pound/Dollars as of 13th of April 2010.
[7] The standard errors obtained through the regressions are not corrected for clustering. However the vast majority of trades in the analysis take place at days with only one trade and this is therefore not likely to have any significant effect on the findings.
“The copyright of this dissertation rests with the author and no quotation from it or information derived from it may be published without prior written consent of the author.”
Alpha returns by Insiders, This online version does not include the used dataset or datatables.
Abstract In this paper I investigate whether insiders are capable of earning an abnormal return in terms of a CAPM alpha return from investing in their own company stock. The conducted analysis was based on trades done by insiders in Danish blue chip companies in the period from April 2005 – April 2010. The analysis is the first attempt done to research this topic for the Danish market. The dataset used for this paper is a unique and new dataset that I created manually by reading through filling from the included companies to the Nasdaq OMX Nordic. My finding and conclusion is that insiders are capable of earning a substantial alpha return. This shows that the Danish market is not strong form efficient. This study shows that the private information reveled by the trades of insiders is valuable. Non inside investors in the market should therefore pay attention to trades conducted by insiders and use this information when evaluating a stocks potential future return.
Introduction The topic of legal trades performed by insiders has only been given very little attention in academic literature. Due to the insiders strong knowledge of the company it is likely that the insiders would be better capable of predicting future returns of the stock of the company for whom they are working than the average investor in the market. A legal framework has been established in most markets to protect other actors in the market place from insiders trading on private information. Nonetheless it is still possible that insiders trading in the market could have superior information on the future prospects of a company stock. Should this be the case other investors would be wise to keep strong attention to trades conducted by insiders. An investor should therefore be reluctant to invest in a company where an insider recently has been selling stocks, and be positively biased to invest in companies where an insider recently has been buying.
The research question that will be investigated in the paper will be whether insiders earn positive alpha returns in the market. This will be tested on a twelve and three month horizon. The following four hypothesizes will be tested: Hypothesis 1a: Following a buy by an insider the stock bought will outperform the market over the following twelve months. Hypothesis 1b: Following a buy by an insider the stock bought will outperform the market over the following three months. Hypothesis 2a: Following a sale by an insider the stock sold will underperform compared to the market over the following twelve months. Hypothesis 2b: Following a sale by an insider the stock sold will underperform compared to the market over the following three months.
The reason why this time horizon was chosen was to test whether an insider trade can be expected to revel valuable private information on the company’s fundamentals and hence the future performance of the stock.
The analysis has been based on stocks traded at Nasdaq OMX Nordic. The stocks included in the analysis have been stocks included in the Danish C20 index at April 1st 2010. The time period analyzed is from April 1st 2005-April 1st 2010. The underlying dataset used to create this analysis had previously not been created by others. I created the dataset used for this paper by manually reading through all filings to the Nasdaq OMX Nordic including information on insider trades or holdings of insiders.[1] In total 4007 filings were read. In the analysis 601 insider trades will be investigated on a risk adjusted basis. To do this 601 regressions have been conducted to determine alpha returns. The findings from the study show a clear pattern that abnormal alpha returns are earned by insiders buying and selling. The implication of this study is therefore to suggest that actors in the market should pay attention to the trades conducted by insiders, since these might reveal significant information about the future performance of the stock.
Literature review The topic of abnormal returns related to insider trading is largely uncovered for most markets. Until recent years very few papers have been written on the topic of abnormal returns or alpha returns from insider trading and related topics. Due to the low attention the topic has had, the results by following insiders are generally not discussed in academic or practical books of portfolio theory or asset management. Further very few papers have been written on the topic. However some sporadic attention has been dedicated to the topic in the last decades. As more managers’ compensation plans include warrants and stock options the level of attention has been slowly growing (Fest, 2008).
Older studies have shown that abnormal returns can be earned by insiders (Jaffe, 1974; Finnerty, 1976,1; Rozeff and Zaman, 1998; Lin and Howe, 1990). More recent papers have also shown a tendency that insiders can earn abnormal returns (Jiang and Zaman, 2009). The assumption by the papers is that insiders earn their abnormal returns by having better knowledge of their companies’ future prospective than does the average actor in the market. Ke et al., have investigated the topic even further and found that this abnormal return steamed from insiders being better at timing the market due to their superior knowledge (Ke et al., 2003). A study for the UK further analyzed in which periods insider trades contained most information, and found that the amount of ownership by directors had an influence on the size of abnormal returns. (Fidrmuc et al, 2006). Further it has also been proven that the market seems to be able to detect also illegal trading and incorporate this into stock prices (Meulbroek, L.K, 1992).
Other studies have been more skeptical about the abnormal returns earned by insiders. In a controversial paper on the Swedish market from 2009 Kallunki et al. did not find that abnormal returns were on average earned by Swedish insiders. In this paper it was concluded that insiders trade with many more motives than earning abnormal returns by exploiting private information. Other motives for insiders trading are found to be balancing objectives, behavioral biases and tax considerations (Kallunki, 2009). No academic papers have investigated abnormal returns by insiders in the Danish market.
Alpha returns by following trades of insiders have also in recent years found its way to the media outside the trading rooms and the academic world. Several news channels and news papers have explained how investors could have benefitted by following the trades of insiders (Song and Paul, 2003; Francis and Hays, 2002). In Canada a webpage under the name www.canadianinsider.com has been created with the purpose of surveying the trades of insiders (Johnston, 2005). Some media coverage has however been skeptical towards the potential of following trades of insiders (Adamson, 2001).
Insiders do clearly have access to information that other actors in the stock market do not posses. Assuming that this information could be used to earn higher returns by trading with uninformed investors this issue has been given a lot of attention by sovereign legislative bodies. Out of a total of 103 countries with stock markets, 87 countries have made regulations on insider trading (Bhattacharya and Daouk, 2000). It has continuously been discussed whether insiders should be allowed to trade in their own stock (Minenna, 2003, Meulbroek, 1992). Arguments for why insiders should or should not be allowed to trade in their own stock can be shortened to three main arguments for and against (Bainbridge, 1998). The arguments against insider trading have been that use of knowledge by insiders would be theft of knowledge from the firm(Georges, 1976), a market egalitarianism theory stating that all investors should trade based on equal information (Langevoort, 1987) and finally that trades by insiders with superior knowledge could harm the integrity of the market place (Bhattacharya and Daouk, 2000). Arguments defending insider trades have been that trades by insiders have no victims since it pushes prices to the direction of the preferential information(Herzel and Katz, 1987), that it is an important way to compensate managers (Manne, 1966) and that insider trading is helping to increase market efficiency (Finnerty, 1976,2). This was also found by Meulbroek, who stated that insider trading leads to private information being incorporated into stock prices and therefore results in more informative prices (Meulbroek, 1992).
Overall in the few papers that have been written on the topic it appears that some abnormal returns are earned by insiders. However since this topic is still in its infancy future studies with a different outcome could quickly change the conclusion. Abnormal returns by insiders have only been investigated at a higher academic level in the US and few other OECD markets. Much more knowledge is therefore to be discovered and clarified in the coming decades. Since it might appear to be the case that insiders earn abnormal returns this is a highly interesting topic that deserves more attention. The ambition with this paper will therefore be to fill out this gap for the Danish market and to create new knowledge on this relatively undiscovered topic.
Regulations on legal insider trading Protection of investors from the potential negative impact of insiders using superior information in the market has been incorporated differently across countries. In the United States the first rules regulating insider trading were created with the Securities Exchange Act from 1934 (Securities Exchange Act, 1934; Minenna, 2003). The rules have continuously been changed and were recently tightened in connection with the passing of the Sarbanes-Oxley Act of 2002 (Brochet, 2010). In the European Union insider trading is regulated by an ECC Directive from 1989 (ECC Directive, 1989). The regulation covers all trades in markets that are considered to be open to the public and operates regularly. Each member state is allowed to enforce prohibitions on trades related to securities that are allowed to be traded on a market in the member state (Minenna, 2003). Different countries have passed varying laws on the topic within the European Union. In the United Kingdom insider trading is regulated through the Criminal Justice Act (1993) and the Financial Service and Market Act (2000).
In Denmark insider trading is governed by the Law on Trading of Securities[2], that was latest updated in 2005 (Bekendtgørelse af lov om værdipapirhandel m.v., 2005). It is here stated that managing employees for companies that have issued securities to be traded on an exchange or have asked to be listed on an exchange have to notify the company of any trades in the company’s stock or affiliated securities. This has to be done at the latest one day after the trade. Leading employees are only allowed to trade in certain time periods and are only allowed to trade if all material information that could change the stock price economically significant has been revealed to the market. It is then required that the issuing company at the latest the following trade day notifies the supervisory body (Finanstilsynet). The supervisory body then immediately makes the information available to the public.
By a leading employee in the company is understood a member of the direction, a member of the board or a member of a supervisory body connected to the company. Other leading employees, who have access to confidential information on the company, and who have competence and authority to take decisions that have material importance for the company’s future are also included by the directive. Further close relatives to any of the mentioned above are also included in the directive. This includes but is not restricted to family members, companies owned by relatives or by others covered by the directive and people living in the same household.
Data and Summary Statistics Creation of the data set The dataset that this paper is based upon was not previously gathered for other studies when I started the project. As an initial starting point the data therefore had to be created. The companies included in this paper are the stocks listed on the Nasdaq OMX Nordic included in the C20 index. The index included the 20 most traded stocks in the Danish market. The time interval investigated is the five year period from April 1st 2005 until April 1st 2010.
No database or list exists of all insider trades in the Danish market. The data was therefore manually created by reading through 4007 fillings from the included companies, resulting in more than 1100 initial registered insider trades. Fillings including insider trades were then recorded. The trades registered were recorded in four categories: Insiders clean buys, Insiders clean sales, Insiders dirty buys, Insiders dirty sales. The first category, Insiders clean buys included all buys were insiders had been buying regular stocks at the exchange at the listed market price. The second category, Insiders clean sales included all sales were insiders sold company stock in the open market that had not been bought on special terms within the latest month. The third category Insiders, dirty buys included all the clean buys, but also purchases that were non regular open market purchases. This could be but is not excluded to conversion and utilization of options and warrants. Stocks bought as a part of the company’s compensation plan, stocks sold at a preferential price, stocks distributed to employees and gifts. The stocks in the fourth category, Insiders dirty sales included all clean sales, but also stocks bought at different terms than the market price and sold within the following month. This could be but is not excluded to converted and utilized options and warrants that were after a short time interval sold in the market. The reason why clean open market buys and sells were separated from dirty transactions was that I thought that insiders potentially could be expected to have different motives for e.g. an open market buy and the utilization of e.g. an expiring option. I therefore thought that it could be the case that a stronger signal could be included in stock purchases and sales done at market terms.
Criteria’s for accepting insider trades as data points Not all data points were accepted as valid for this paper. Data points were discarded for several reasons. Reasons for discarding were the following: Insiders had traded with relatives and the net position exchanged equaled out. Insiders had sold or bought stock from companies partly or totally owned by the insider. Distributions of warrants and options were not included. Buys and sales of drawing rights were not included. If an insider had been both buying and selling the net change in the position were recorded and several transactions were then combined into one. Transactions in which an insider and the insiders’ relatives had conducted many or complicated diverse directed trades were discarded, if it was not clear whether the insider was attempting to increase or decrease the exposure to the stock. When an insider bought stocks by converting options and quickly sold the stock within the following month this was recorded as a dirty sale. If the insider held onto the stock and then sold the stock at a later point in time this was recorded as a clean sale. Finally insider transactions performed in the last year of the analysis was not used since it would not be possible to track the performance in a full year following the transaction. After setting up these boundary conditions a total of 601 insider trades were accepted.
The stock data The data on the stocks were obtained from the Nasdaq OMX Nordic. Adjusted stock prices were used to take into account corporate actions. When working with the data, I observed some irregularities with the data of the stock DSV, where a stock split in May 2007 was not incorporated into the data. Following this I contacted the Nasdaq OMX Nordic that confirmed that there was a mistake in the data provided by the Nasdaq OMX Nordic. The discrepancy found was then corrected manually in the data set. One stock that filed insider trades in another country was discarded[3]. To compare the returns of the stocks the OMX Copenhagen Benchmark Gross Index was used. The index includes all the largest stocks and most liquid stocks in the Danish market. A gross index was used to take into account dividends. All companies used in the collected data are large liquid stocks included in the index. A potential other index that could be used would be the OMXC20 that is narrower and also includes all the stocks used for the analysis. However data on this index were only available as a PI Index not taking into account dividends. Therefore the Copenhagen Benchmark Gross index was more appropriate to use.
The stock data for the trades accepted are summarized in table 1 below[4].
*Three open market trades had an abnormally high value and are not representative for the general purchases and an adjusted value is therefore given in the table.
The total numbers of trades accepted were 601 trades. The total number of stocks traded was 4.7 million. The average number of stocks in a trade was 7,765. The total value of stocks traded in the time investigated was £352 million, with an average value pr trade of £ 586,266.
From the dataset it can be seen that that the most common transaction in the period was an open market buy. 290 data points where insiders bought stocks in the observed period of time were accepted. The corresponding number of sales were only 141. The average value of the buy transactions was £ 651,509, while the average value of the sales was £ 427,631. The value for buys is however misleading due to three very high open market buys. When adjusted for these three trades the average value of the open market buys drops by 64% to £ 232,271. From the data it can therefore be seen that buys are more common among insiders than sales[5]. However the average value of selling transactions is generally higher than the value of the buy transactions. When considering dirty buys it can be seen that this category includes 60 extra trades compared with clean buys. The dirty sales category includes 110 more trades than the clean sales. From the data it can therefore be seen that it has been more common that insiders sell their position rather than holding after using options and warrants. The value of the average buys including dirty trades has been £ 204,218, while the average sell has been worth £ 587,596. This indicates that managers are more likely to use options, warrants and other positions and hold the stock when the trade is of a smaller value, while when the value of options and warrants reaches a higher value insider are more prone to sell their positions. This could be an indication that insiders are not willing to hold a too large share of their wealth in the company for which they are working since they are already by their expected future income highly exposed to the future viability of the company. The difference in size of the positions bought and sold could also indicate that insiders wish to diversify their holdings if their utilization of options and warrants results in an overweight of the respective stock in the average insiders portfolio.
The average value of the trade is smaller than the value of trades from other markets. In the US the average value of insider sales was in the period from 1995-1999, a study 10 year older than this, determined to be above £ 662.000[6] ($ 1.031 million) (Aktas et al., 2008). The reason why insider trades in Denmark might be of a lower value than in other countries could potentially be thought to occur from a general lower level of compensation for top management compared with other countries (Business, 2006).
Methodology The model An important assumption behind the study is that insiders engage in open market sales and buys with the purpose of utilizing information to earn abnormal alpha returns. To see if insiders were able to earn a significant abnormal return a comprehensive scientific approach was followed.
When determining a way to measure the performance of insider trades it is important to differentiate between, what is noise and what is actual performance by the insider trading. When just looking at the return of the insiders trades without a measure of comparison the picture does not reflect the performance of the insider. The return could be a result of market movements that do not have anything to do with the insiders stock picking ability. Further many investors are risk averse and require a premium to hold more risky assets. To include this in the model market risk was adjusted for when determining whether a stock had delivered an abnormal return.
The model was created based on regressions. The daily return for the “stock i” over a given twelve month interval was regressed on the return of the Copenhagen Benchmark Index to determine if a positive alpha value had been earned during the following three or twelve month period. This was measured by the CAPM alpha value. The regression process was complicated by the different relevant time horizons for each insider trade. The trades performed by the insiders were done at different dates. A new data list for stock returns and index returns with the relevant following three or twelve months was therefore created separately for each regression. Following this it was possible to make the 601 regressions separately. The abnormal alpha return was then given by:
The model assumes that CAPM holds. The average abnormal alpha return over all the included data points was calculated by taking the average value for all data points. To test the significance of the findings the t-stat value was calculated. This was done by taking the average value of the alpha return and divide with the standard error of α[7].
To test if the results found were general across all sectors the analysis was also done on the separate industry groups. The same method was followed to see if abnormal returns were clustered in certain industries as for the full data sample.
When insiders trade the trade was typically filed to the exchange on the day of the trade or the day after. Trades were therefore generally available to the public within the first or second trading day. However in some instances the insiders trades were published later. The starting date used to determine the abnormal return in this study is the day that the insider traded. This was chosen because the price on the trading day is the price that the insider based the trading decision upon. Another approach could be to investigate abnormal returns from the day om which information about the trade was made public. This would then not show the alpha returns earned by insiders, but show the alpha returns that other actors in the market could have earned by replicating the trades of insiders, when information about the trade became public. Both approaches were tested and based on the given data set the difference in the results were minimal.
To test for clustering of alpha returns within industries the investigated stocks were grouped as can be seen in table 2. Analyses were then conducted on each industry separately.
Table 2
Results and discussion Full sample The result found for alpha returns can be seen in table 3.
Table 3
* Significance on 90% level. **Significance on 95% level. *** Significance on 99% level.
When looking at the “Clean Buys 1 year” it can be seen that insiders were able to earn substantial alpha returns. The average alpha return earned by insiders was 4.2% on a yearly basis. An open market buy by an insider must therefore be considered a strong indication that the stock will outperform the market in the following 12 months. The t-value found is very high. The result can therefore be proven to be significant on a 99% basis. The 99% confidence interval shows that alpha returns by following insiders are expected to be in the interval 1.2%-7.2%. This result is in line with returns found in other studies in other markets (Aktas et al., 2008). Hypothesis 1a can thereby be confirmed that following an insider buying the stock will outperform the market in the following twelve months, when looking at the “Clean Buys 1 year” category. The “clean buy three months” trades can be seen to give an expected positive alpha return of 2.8%. However t-stat value for this result is relatively low. Their therefore appears to be a trend towards a positive alpha, but hypothesis 1b cannot be proven statistically significant on a high confidence level.
When looking at the “Clean Sales 1 year” it can be seen that on average the stocks sold underperformed the market by 3.4% in the following twelve months. The result is also found to be significant on a 99% level, and the result found must therefore be considered to be very strong. When looking at the 99% confidence interval it can be seen that the stock is expected to underperform the market with between 0.2%-6.6%. The implication of this result is that stock owners in companies where insiders are selling should take insider sales as a strong signal that the stock will underperform the market in the following twelve months. Investors could therefore on average do well by reducing their position after an insider sale. Hypothesis 2a can therefore be shown to hold for “Clean Sales 1 year”. In the shorter three month period the result is also positive and statistically significant on a 95% confidence level. Hypothesis 2b can therefore also be confirmed.
When looking at the “Dirty Buys 1 year” the average alpha return can be found to be 4.1%. This is slightly below the value found for the clean buys. This is however nonetheless still a strong outperformance, indicating that even when insiders buy stocks at preferential terms and hold them an alpha return can be expected. On a 99% confidence level this amounts to an alpha return of 1.5%-6.7%. This therefore shows that hypothesis 1a does not only hold for Clean Buys, but also holds for buys on non market terms on a 1 year basis. On a three month basis the result is positive with an average alpha return of 2.6%. The result is however less certain and hypothesis 1b can only be proven significant on a 75% confidence level.
When considering the Dirty Sales the picture however changes. The “Dirty Sales 1 year” on average underperforms the market by 0.1%, which indicates that these stocks almost perform equivalently to the market. A sale by an insider when including non market terms trades is therefore not an as good indication as a clean sale. The result for the dirty sales is not significant on either a 75% or a 90% confidence level. For the three month period the result is slightly better but still not highly statistically significant. Neither hypothesis 2a nor 2b can therefore be confirmed to hold for “Dirty Sales”. The generally weak results found for dirty sales could potentially be explained by the fact that insiders conducting dirty sales might not be selling due to an expectation that the stock will underperform the market, but could be selling for other reasons such a balancing the insiders portfolio. This is in line with the suggestion given for insider trades in the investigation made by Kallunki et al. in the Swedish market (Kallunki et al., 2009). If an insider is compensated with options, warrants and other stock related instruments it is likely that the insider would potentially not have chosen to allocate a that large fraction of his wealth in the stock and would therefore decide to sell for diversification reasons even though the firm does not have bad future prospects.
In total there seems to be a strong case for insiders outperforming the market on a one year basis. When looking at the three month the same trends seems to be present that insiders do earn abnormal returns, but the results are less significant. This could be an indication that the trades of insiders are based on their long term fundamental expectation to the company’s performance.
Results separated into categories Table 4
*Significance on 90% level. **Significance on 95% level. *** Significance on 99% level.
In table 4 the data has here been used to calculate alpha returns return in various industries represented in the C20 index to see if some insiders are better at predicting future performance than others.
When looking at the financial stocks it can be seen that “Clean Buys 1 year” insiders do on average earn an abnormal return of 1.8% but this is not significant. When looking at the three month period it can however interestingly be seen that the alpha value is negative indicating that managers in the short run underperformed, but over performed on the 1 year basis. I tried to investigate this finding in the dataset and found a pattern that financial managers had conducted a large number of trades towards the end of the financial crisis, but some months before the markets reversed, which then resulted in a short run underperformance. A hypothesis could be that these managers might have been able to see the reverse of the financial stocks before the market, but where not capable of optimally timing their investments.
For the “Clean Sales 1 year” the results can be proven to be significant on a 95% confidence level and investors can expect to see the stock underperform the market by 0.2%-12.8%. For the “Clean Sales 3 moths” financial stocks also underperformed. This was found significant on a 90% confidence level. Investors could therefore expect to see a very large under performance in financial stocks after insiders have been selling. When investigating this finding further in the data I did some interesting observations. Before the financial crisis had strongly caught momentum substantial sales was done by the management within the banking sector. Investors could therefore have protected themselves from devastating looses on financial stocks if the transactions done by insiders had been followed.
When looking into the industrial sector a very large outperformance can be seen when insiders buys either Clean open market buys or Dirty Buys on both a three month and one year horizon. The average outperformance seen in the sample is at overwhelmingly 15.8% for the one year period and 19.9% on a three month period. Insiders therefore clearly seem to be able to predict future increases in stock prices. When considering the stock included in the industrial segment in the C20 index it can be seen that many of these companies are selling large projects e.g. windmills and plants for the mineral and cement industry. A part of the explanation why these insiders earned those high alpha returns could be that the time period between start of negotiations, to the signing of a final contract could be very long within those industries. Insiders could therefore be expected to have a better idea of the level of new contracts the company would win within the upcoming three months to one year than the consensus by analytics in the market.
In the pharmaceutical industry an interesting result was found under “Dirty Sales 1 year”. Here the stocks sold by insiders actually outperformed the market by 3%. When investigating the Dirty Sales conducted by insiders in the pharmaceutical industry it could be seen that a large share of these sales were connected to pharmaceutical companies that has stock options accounting for a very substantial part of the compensation plan. These sales could therefore very well be reflecting insiders wanting to decrease the relative share of their positions invested in the company. When looking at the three month period the over performance of Dirty Sales disappears and reverses to a negative average alpha of -2%. This could indicate that these portfolio allocations are timed with insider expectations of short run underperformance. When looking at the three months clean sales, insiders in the pharmaceutical industry performs better than the market, which again points towards these insiders in the pharmaceutical industry might be able to outperform the market in the short run when selling.
In the transport sector insiders can be seen to outperform the market when selling both “Clean Sales” and “Dirty Sales” on a one year horizon. When buying on a 1 year basis or in any three month category the insiders of transport companies are however not capable of producing an alpha return. In general insiders in transport companies are therefore less capable of predicting their stock returns than the average insider in this analysis. A potential explanation for this could be that the return of these stocks are highly affected of the overall state of the economy and less by company specific information compared with e.g. the industrial segment. The same could be said about the category utilities, where insiders are capable of outperforming when buying on a one year basis, but in no other categories.
Finally in the category others, insiders buying can also be seen to outperform the market. When investigating this data I found that the companies located in this category have the common denominator that they are very innovative companies. These companies could maybe be thought to have a good idea about the prospects for their new products which could be utilized in the market place.
In total there seems to be a case for following insiders within specific industries. However when separating the data into different categories, this results in a higher standard error due to the lower number of observations in each category compared with the full study. This makes it more difficult to prove statistical significance. Altogether when looking at the result for the entire test sample the message seems to be crystal clear, insiders do earn abnormal returns in the market on a twelve month horizon and on a lower confidence level there is also an indication that insiders outperform on a three month horizon. Other investors should therefore keep an eye on the trades of insiders.
Implication In total the results found are very strong and clearly shows that markets cannot be thought of as strong form efficient, since insiders with private information are able to earn abnormal returns. The implication of this study is therefore that an investor can on average obtain a positive alpha return by following the trades of insiders. An investor would therefore be wise to keep his eyes open, and takes into account the trades conducted by insiders. Following insiders could therefore be a way to increase the expected return without taking extra risk. Over time when more information is published about abnormal returns earned by insiders it is likely that more actors in the market would pay more attention to the insiders’ trades. This could result in stocks immediately increasing in value after a buy by an insider and in immediate falls following a sale. This could in the future potentially reduce the value of following insiders, but for now a consistent alpha return seems possible. Based on the findings from this study a potential trading strategy could be to set up a zero cost portfolio going long insider clean buys and short insider clean sales on a 1 year horizon.
Conclusion The returns earned by insiders trading in the market have generally been given very little attention in academic literature. A potential explanation for this could be that legislation has in many countries been passed to protect other investors from insiders using materially important information to trade in the market. A potential perception could therefore be that insiders would not be capable to earn abnormal returns in the market due to the legislation. By looking at this first study of insider trades in the Danish market it can be seen that insiders are capable of earning a substantial alpha return. The results are stronger when looking at a twelve month horizon compared with a three month horizon. This could be an indication that insiders trading decisions are based on their expectations of company’s long run fundamentals. The conclusion from this study is that investors should pay considerable attention to trades by insiders and use this information actively. This paper therefore proves that the Danish C20 index is not strong form efficient, and by following the trades of insiders investors can utilize the insiders’ private information to obtain a positive alpha return.
Limitations and suggestions to future studies:
Due to the time consuming nature of the manual data construction by reading filings of companies the paper was limited to stocks in the C20 Index including only blue chip companies. In future studies it could be interesting to include a larger number of stocks to increase the number of trades and thereby being able to make a more general conclusion covering the entire market place.
The risk adjusted returns used in this study incorporates market risk. A study in Hong Kong from 1993-1998 showed that a substantial part of the abnormal returns were earned in transactions by insiders in smaller companies (Cheuk et al., 2006). In future studies if data should be available at that point in time it could be interesting to investigate whether other risk factors could partly explain the abnormal returns earned by insiders.
Acknowledgements I would like to thank Dr. Michela Verardo from the Financial Markets Group at London School of Economics for her constructive comments and suggestions that helped to structure this study.
References • Adamson, D., 2001. When insider stock purchases mislead. CBS MarketWatch. June 11.
• Aktas et al., 2008. Legal insider trading and market efficiency. Journal of Banking & Finance 32, 1379-1392.
• Bainbridge, S., 1998. Insider trading: an overview, Encyclopedia of Law and Economics, Nber.
• Bekendtgørelse om lov om værdipapirhandel m.v., 2005. LBK nr 843 af 07/09/2005.
• Bhattacharya, U., Daouk, H., 2000. The World Price of Insider Trading, Working Paper.
• Brochet, F., 2010. Information Content of Insider Trades before and after the Sarbanes-Oxley Act. The Accounting Review vol. 85, No. 2, 419-446.
• Business, 2006. Topchefer tjener stadig for lidt. Berlingske Tidende, 23-06-2006. First entered 13th of April 2010. http://www.business.dk/diverse/topchefer-tjener-stadig-lidt
• Cheuk, M.-Y. et al., 2006. Insider trading in Hong Kong: Some stylized facts. Pacific-Basin Finance Journal 14, 73-90.
• ECC Directive n. 89/592 coordinating regulation on insider trading, 1989, European Union.
• Fest, G., 2008. Overseeing “Legal Insider Trading. Bank Technology News. February 2008.
• Fidrmuc et al., 2006. Insider Trading, New Releases and Ownership Concentration. The Journal of Finance, Vol. LXI, No. 6, 2931-2973.
• Finnerty, J., 1976,1. Insiders’ activity and inside information: a multivariate analysis. Journal of Quantitative Financial Analysis 22, 109-126.
• Finnerty, J., 1976,2. Insiders and Market Efficiency. Journal of Finance 31, 1141-1148.
• Francis, B., Hays, K., 2002. How to Profit from Legal Insider Trading. CNNFN August 2002. August 15, 05:00PM.
• Georges, S., 1976. L’utilisation en bourse d’information privilegees dans le droit des Etats-Units, Economica.
• Herzel, L., Katz, L., 1987. Insider Trading: Who Loses? Lloyds Bank Review.
• Jaffe, J., 1974. Special and insider trading. Journal of Business 47, 410-428.
• Jiang, X., Zaman, M., 2009. Aggregate insider trading: Contrarian beliefs or superior information. Journal of Banking & Finance 2009, doi:10.1016/j.jbankfin.2009.11.016
• Johnston, K.,2005. Website tracks legal insider trading. The Leader-Post. Section: Business & Agriculture.
• Kallunki, J. et all, 2009. Why do insiders trade? Evidence based on unique data on Swedish insiders. Journal of Accounting and Economics 48,37-53.
• Ke, B., Huddart, S., Petroni, K., 2003. What insiders know about future earnings and how they use it: Evidence from insider trades. Journal of Accounting and Economics 35, 315-346.
• Langevoort, L., 1987. Regulation, Enforcement and Prevention. Clark Boardman Company, New York.
• Lin, J., Howe, J., 1990. Insider trading in the OTC market. Journal of Finance 45, 1273-1284.
• Manne, H., 1966. Insider trading and the stock market. Free Press, New York
• Michaud, R., 1989. Financial Analyst Journal. Vol. 45, No 1. January-February, 31-42.
• Minenna, M., 2003. Insider trading, abnormal return and preferential information: Supervising through a probabilistic model. Journal of Banking and Fiance 27, 59-86.
• Meulbroek, L. K, 1992. An Empirical Analysis of Illegal Insider Trading. Journal of Finance, Nr. 5 Dec. 1992.
• Rozeff, M., Zaman, M., 1998. Overreaction and insider trading: evidence from growth and value portfolio. Journal of Finance 53, 701-716.
• Sarbanes-Oxley, 2002, section 403, USA.
• Securities Exchange Act of 1934, USA.
• Song, J., Paul, B., 2003. Cover Story: Insider watching. The Edge Singapore
[1] Filings from April 1st 2005- April 1st 2010 were read for the companies included in the analysis.
[2] Translated from Danish. Danish name: Bekendgørelse af lov om værdipapirhandel m.v..
[3] The stock discarded was the Nordea Bank. This company filed in Sweden.
[4] The conversion of DKK (Danish Kroner) to GPB (Great Bristish Pound) was based on the prevailing exchange rate at noon the 12th of April 2010.
[5] This would also be the case if all insider trades had been accepted as data points.
[6] With the exchange rate Pound/Dollars as of 13th of April 2010.
[7] The standard errors obtained through the regressions are not corrected for clustering. However the vast majority of trades in the analysis take place at days with only one trade and this is therefore not likely to have any significant effect on the findings.
