Profitability analysis of informative insiders and market efficiency

Identification of insiders whose transactions allow us to predict market movements of securities. Analysis of the predictive effectiveness of various approaches to identify potentially informative transactions at various levels of market liquidity.

Рубрика Банковское, биржевое дело и страхование
Вид курсовая работа
Язык английский
Дата добавления 07.12.2019
Размер файла 1023,5 K

Отправить свою хорошую работу в базу знаний просто. Используйте форму, расположенную ниже

Студенты, аспиранты, молодые ученые, использующие базу знаний в своей учебе и работе, будут вам очень благодарны.

Размещено на http://www.allbest.ru/

Анализ прибыльности информативных инсайдеров и эффективность рынка

Introduction

transaction market insider

According to Security Exchange Commission's legislation, insiders are obliged to disclose part of their trades. The trades to be disclosed are those that involve the securities of the company, in which insider is occupied. Such disclosures are expected to convey new information to the market and potentially may signal the future share price realization. However, not all insiders convey the valuable information in their trades given the fact that some of them may trade for the liquidity or portfolio reasons without possessing any knowledge. Only part of them may trade based on the privately known fundamental information. It is therefore expected that transactions disclosed by such insider may serve as a reliable predictor of future performance of corresponding stock. As the same time, insider who are involved in such transactions are considered to be informed. Therefore, if there is a consistent approach that allows determining, which insider disclosing his trade is indeed informed regarding the future, then it may serve as a basis for a strategy that is likely to yield the abnormal returns.

There are already studies, which examined this topic and proposed various ways to distinguish the insiders whose transactions have strong predictive power. In general, all these methods use the past performance or activity of insiders and rely on some pre-defined threshold to separate those, who are likely to grant the highest signal regarding future return by disclosing their transactions. For instance, if some insiders consistently outperformed the market in specific periods, this may serve as a signal that they are informed regarding the future stock price realization. One can notice that the logic, which authors of previous studies employed, is intuitive and based simply on past observations. The main contribution of this study is that it assesses the predictive power of transactions chosen and measured by available approaches under various market liquidity levels. The intuition explaining why the introduction of various liquidity scenarios may provide different results strongly matches with common sense. At the same time, this intuition is also based on the theoretical model, unlike all previous related studies.

The core intuition standing behind contribution proposed in the study is the following: market maker who sets the bid and ask prices faces uncertainty regarding the future share price realization. Given the higher uncertainty, market maker will set wider bid-ask spread to insure from unexpected price changes. Assuming that market maker is aware that within some companies more insiders trade on private information, however without understanding who these insiders are and what this information actually is, more uncertainty is faced by him. Namely, market maker understands that price of particular security is very likely to move. Acknowledging the fact of a likely move, but without having information on the direction, larger bid-ask spreads will be set to ensure more safety regarding the profits of market maker. Thus, it is possible that if there are some stocks involved in transactions of likely to be informed insiders and given that this is observed by market maker, then he may set higher bid-ask spread. This allows to hypothesize that there may be a potential link between the size of bid-ask spreads quoted and the average level of `informativeness' of insiders trading with the corresponding securities. In other words, if market is segmented by the different levels of stocks' liquidity then different levels of insiders' `informativeness' can be observed within each segment.

The goal of this study is to actually evaluate the already achieved results by testing whether these hold for any level of market liquidity. That is to test whether proposed attributes used to distinguish informed insiders indeed result in picking the transactions, which serve as reliable predictors of future performance on any segment.

The literature review part will introduce the theoretical model, which supports this intuition relating degree of `informativeness' with liquidity level. This model is developed by Fishman and Hagerty and is called “Mandatory disclosure of trades and market liquidity” (1995). Further part will discuss the foundations and predictions of this model in more details. Furthermore, previously mentioned empirical studies, which propose the methods of revealing informed insiders are also included into the literature review. Both explanation of methods that authors proposed as well as key results achieved will be discussed. The review part will be completed with explicit hypothesis statement.

The result of all considered empirical studies shows that one can earn abnormal returns by using the information inferred from those insiders who were regarded as informed according to the particular methodology proposed. In this study, the opportunity of earning abnormal returns will also be tested given different levels of liquidity. The results of this test will allow judging on market efficiency because in fact opportunities of earning abnormal returns based on publicly available information is evaluated.

To summarize, this paper will start from the theoretical set up that explains why liquidity level can be related to the level of insiders' `informativeness'. It will then advance to discussion of the existing studies devoted to distinguishing informed insiders on the market. After that, explicit hypothesis will be defined. The next section includes the empirical approach and its implementation. Finally, the results will be presented.

Contribution

As it was mentioned in the introduction, studies, which are related to the topic of determining informed insiders, propose methods based on past observation of performance and (or) behavior of these insiders. These attributes from the past are either average profitability or some seasonal patterns (when it is hypothesized that informed insiders do (or do not) trade in some particular periods). An important contribution of this research is that is allows assessing the applicability of these attributes under different liquidity levels. In other words, this study will evaluate whether some method previously discovered works not only on the whole universe of stocks, but also on each separate level of market liquidity. In case it is indeed observed that results depend on the level of liquidity, it will provide opportunities of revealing informed insiders more precisely. This will be achieved by excluding those insiders within particular liquidity segment of the market, for whom already discovered attributes do not work, although these attributes were proved to be significant on the whole sample in previous studies (that is without segmentation by liquidity). Consequently, it should result in a more precise group of informed insiders revealed. Involving this newly recognized group while picking stocks for trading portfolio may result in stronger results comparing to those achieved in previous studies.

Chapter 1. Literature review

The study of Fishman and Hagerty, called “Mandatory disclosure of trades and market liquidity” (1995) plays an important role in theoretical set up. The authors compare the effect on insiders' trading profits under various scenarios depending on whether or not insiders disclose their trades. Authors consider three cases: there is no obligatory disclosure of trades at all, disclosure is mandatory and disclosure is voluntary. The main research goal of their paper does not totally coincide with the current one. However, authors partially address the link between the liquidity (measured by the bid-ask spread) as an intermediate step of their analysis. The case of mandatory scenario is the one, which is partially relevant in the context of this paper (where `informativeness' and market liquidity are linked). However, `no disclosure' scenario should be partially considered as well, because it serves as a basis when `disclosure' case results are predicted.

The model set up in the study considers two dates, during which trading is possible. Each trader is endowed with cash w and risky asset z. The risky asset is valued as и and in the future it takes either (high) or (low) values with equal probabilities. It is assumed that there are 2N outsiders and definitely do not possess any private information regarding the future realized stock value. The first half of outsiders trades at date 1 and second half - at date 2 (this assumption was introduced by authors in order to simplify the analysis). There is also one insider, who can be either informed (that is he knows future realization value of the stock) or uninformed (namely, possess same level of information as other market participants). The market believes that insider is informed with probability q. All traders have a concave utility function of wealth. There is also a market maker who establishes the bid-ask spreads at each period. Each of the participants: outsider, insider and market maker, has some strategy of behavior.

The modeling of trading activity employs all the components mentioned above in order to derive the equilibrium. The whole process is divided into 2 periods, which consist of:

Picture 1Source: Fishman, Hagerty, 1995

First of all, the scenario when insiders do not disclose their transactions should be considered. These results will be necessary in order to achieve the results under mandatory disclosure scenario.

At date 1, if the insider is informed he will purchase stock if anticipates an increase in its value and will sell otherwise. If the insider is uninformed, he will behave as well as outsiders by hedging the current positons. Namely, hedging implies selling the currently available number of stocks or purchasing the number of stock, which has previously been shorted. The split between outsiders who long and short is equal, in other words, outsiders will buy and sell with equal probabilities.

While establishing the bid and ask quotes, market maker considers his potential profit, which is expected to attain the minimum nonnegative value. The profit of market maker is considered separately for buy and sell transactions that is bid and ask quotes are established by solving different equations, although principles are similar. The equation determining profit consists of the value earned (lost) on the difference between realized value and prices quoted by market maker as well as the total number of traders participating on sell side in case of bid and buy side in case of ask. The number of traders on each side is conditional on the future stock value and is tied to the probability that insider is informed, in other words it is used to derive the expected number of traders given that future price is high or low:

Picture 2Source: Fishman, Hagerty, 1995

Equating this profit to 0 yields the equation determining the bid value. Relying on the same logic, the profit equation on the ask side is provided. Equating it to zero yields an ask quote.

At the period 2 market maker is capable to update the bid ask quotes based solely on order flow. Now, the trading strategy of informed insider at period 2 should be considered. In fact, it is the same as in previous period, so he will buy again if aware that realized value is high and vice versa. Uninformed insider and outsiders will wish to restore their stock positions, however they will do so only if the bid and ask updated by market maker, match with their own reservation bid and ask prices, at which they would be ready to buy. In other words, if ask quote is too high or bid quote is too low, outsider (and insider if uninformed) will not participate in the date 2 trading. Again, in order to find the value of bid and ask, the profit of market maker from each side of transactions should be considered. The relative value of bid and ask in fact determine whether outsiders (and uninformed insider) will engage in the date 2 transactions. Therefore, depending on their decision, the order flow will vary and thus yield bid and ask quotes separately for each scenario.

After attaining the results on bid-ask quotes in the `no disclosure' scenario, the next step is to derive the spread when disclosure is mandatory.

Insiders follow the same strategy as in `no disclosure' scenario in both dates 1 and 2. The market maker quotes similar bid and ask at date 1 too. The key difference now is that market maker updates his beliefs based not on the order flow, but on the disclosure provided by insider after date 1. Updated beliefs of market maker play the key role while establishing date 2 bid and ask quotes.

The final rule set by market maker depends on whether insider disclosed buy or sell transaction after date 1. If he disclosed buy, then market maker considers it to be equivalent to the information conveyed by N+1 total buys per period in case of absence of disclosure. This is because N outsiders will participate on the buy side regardless of scenario, while it is now known for sure that insider acquired the stock. If insider discloses sells, this is equivalent as 0 buys conveyed.

Authors prove that in fact there exists an equilibrium, in which all participants follow their defined strategies for both `no disclosure' and `disclosure' scenarios. The part of assessing impact on insiders' profit is rather irrelevant in the current context, so it is omitted in the literature review part.

Finally and most importantly, the bid-ask spread under mandatory disclosure is set by the following equation:

Picture 3Source: Fishman, Hagerty, 1995

One can observe that the function is concave in q and is maximized at q=0.5. The intuition behind it is the following: given both very high and very low probability of particular insider to be informed, market maker can judge about future price with more confidence. While as it is obvious for the high-probability case, intuition is less obvious in case of lowest probability. Disclosure of insider who is very unlikely to be informed still provides a strong signal to the market maker for the following reasons: authors show that regardless of the fact whether insider is informed or not, his disclosure will impact subsequent bid-ask spreads. As a result, even if insider is uninformed, he has an opportunity to move the stock price in a desired direction and then reverse the trade at period 2 to earn the profits. If the market highly anticipates such strategy from uninformed insider, he can also anticipate more precisely the future price movement, which is affected by uninformed insider's reversal strategy. Therefore, market maker is quite confident in the future price, even understanding that uninformed insider provides disclosure. High level of confidence allows market maker to set the low bid-ask spread and enhance the liquidity of particular stock.

As a result, the plot for the bid-ask spread as a function of q will have the following shape:

Picture 4Source: Fishman, Hagerty, 1995

Now it is theoretically justified that the chance of insider to be informed is related to the quoted bid-ask spread on particular stock. Although link is established, one can see from the graph that this relation works in opposite directions. There are several arguments in favor of the relevance of the left tail on the graph. First of all, the previously mentioned intuition that the more insiders are trading on private information, the more unknown price movements will be faced by the market maker. As a consequence, the higher bid-ask spread will be quoted. Secondly, significant share of them trades for liquidity reasons, trying to optimize the composition of portfolio between cash and stock. Furthermore, some insiders engage in transaction by exercising options, which have been granted as a performance bonus, for instance. Considering all these facts, it can be assumed that real world probability of insider to be informed is rather low and is definitely below 50%.

The literature review part will now proceed to the observation of previous studies devoted to the search of methods of revealing informed insiders. Consequently, the best practices of these studies will be combined and afterwards the results of their application to be tested under various liquidity scenarios. There will be three studies, which are examined in this section.

The first study from Cohen, Malloy and Pomorski called “Decoding Inside Information” (The Journal of Finance, 2012) proposed that uninformative insiders are those, who trade for liquidity needs. In other words, they sell securities when need to receive cash for personal purposes, while either not having enough information about the future, or not expecting significant decreases in company's performance. Authors hypothesized that such `liquidity traders' follow the specific seasonal pattern. Namely, they trade during the same month every year. The possible explanation for such pattern is the fact that companies provide bonuses (for top managers and directors usually in the form of stocks) at some predefined period of time, like the end of fiscal quarter or year. As a result, some insiders will dispose such securities immediately to obtain cash. Authors assumed that liquidity traders are those who exercised transactions during the same month during the last 3 years. Consequently, they formed portfolios based on the transaction of all the other insiders (these are considered to be informed) and showed that portfolio strategy, which relied on such insiders indeed provided significantly positive abnormal returns.

It may be argued that such approach of distinguishing opportunistic insiders may seem as too simplified. It is very likely that most of the traders that follow seasonal patterns are indeed liquidity traders, however, there is no guarantee that part of liquidity traders have another or do not have seasonal pattern at all. Therefore, alternative approaches should be considered as well. The consequent studies classified insiders using another methodology based on their historical performance.

The next study, which deserves attention in this field was conducted by Usman Ali , David Hirshleifer (Journal of Financial Economics, 2017). Authors analyzed the transactions of insiders, which were conducted in the period preceding the quarterly earnings announcements (QEA), which public companies provide prior the issuance of official financial statement. In fact, the stock prices significantly move during these period and provide good profit opportunities if some private information is known. Ali et. al examined the transactions of individual insiders, which were conducted around the QEA period and correspondingly ranked these insiders into 5 quantiles according to their average profitability. The subsequent analysis included two major parts.

First of all, authors employed the linear regression approach where transaction disclosed by insider serves as a unit of observation. They regressed the future one-month return on the introduced dummy variable, which equals one, if the transaction has been disclosed by insider assigned to the most profitable (5th) quantile. Besides this indicator, authors also included common Fama-French factors, which played the role of control variables. It was demonstrated that this indicator serves as significant predictor of the future return. Secondly, authors have constructed portfolios, which follow the transactions of insiders for each of the five quantiles. To do this, in the end of each month authors included a particular stock into portfolio if it was traded by the insider assigned to one of the five quantiles during the corresponding period. Each portfolio was rebalanced at the end of each month. The portfolio based on top quantile with the most profitable insiders has resulted in the highest and significant alphas (based on Fama and French factors), thus showing that such approach works and allows distinguishing informed insiders, at least partly. This study, given its strong results, provides a methodic, which is reasonable to test under various liquidity scenarios.

Another study by Cline, Gokkaya, and Xi Liu, called “The Persistence of Opportunistic Insider Trading” (Financial Management, 2017) approached the process of revealing informed insider from the different perspective. On the one hand, they introduce less detailed segmentation of insiders by their profitability. In this study, the classification is binary and insiders are divided on the basis of whether their profits are above or below zero (that is there are 2 groups instead of 5). Thus, it is likely that such insiders will be outperformed by the top quantile insiders determined by the approach of Ali et. al. Also, while evaluating past performance, Cline et. al consider all previous trades of particular insiders, not only those, which occurred around the QEA period.

On the other hand, the advantage of the given study is that authors actually distinguished between different positions of insiders, using dummies for CEOs, CFOs, presidents and chairmen. The large shareholders were considered as well. Authors observed that trades of the shareholders are in fact not informative in the sense that coefficients for this type of insiders' transactions turned out to be insignificant for both buy and sell trades. Regarding the directors and officers' comparison, the evidence is rather mixed regarding the significance of each grade in the buy and sell scenarios separately. It cannot be concluded that either CEO or CFO are more efficient. This leaves the possibility for further investigation of whether insider's seniority affects his capability to predict future performance of the stock.

In this study, this idea will be developed further by introducing an alternative way to classify insiders by their seniority. The classification will be based on the criteria filled in the Form 4, in which insiders report transactions. Within Form 4, insider can be classified as: director, officer, owner of more than 10% of stocks or other. This segmentation is simpler as it aggregates insiders into groups on higher level by contrasting directors and officers. It is possible that difference between capability to predict future performance varies stronger between director and officer than between CEO and CFO as it was compared in the study by Cline et. al.

As a result, insiders prescribed to quantile 5 according to the methodic of Ali et al. will additionally be classified further into directors and officers. Given that authors achieved no significant results in for the large shareholders, this group will be omitted in the analysis of this study. Hence, the two variables of interest would be dummy variables corresponding to whether transaction was performed by: 1) Director assigned to quantile 5 by past profitability; 2) Officer assigned to quantile 5 by profitability. These variables will be in the focus during regression analysis with future return as dependent variable.

To conclude the literature review section, the relevant hypotheses should be established. As it was discussed, it is reasonable to expect that results achieved by Ali and Hirshleifer as well as by Cline, Gokkaya and Liu may vary depending on the level of market liquidity. This is because predictive power of insiders' actions causes uncertainty in market makers' profits and motivates him to establish wider bid-ask spreads. Thus, the first hypothesis proposed is the following:

H1: Predictive power of transactions disclosed by insiders who were chosen according to the currently developed approaches will vary depending on the level of market liquidity

The predictive power of particular insider does not yet guarantee that following his transaction will provide profit opportunity for outsider monitoring corresponding disclosures. The next task is thus to understand whether portfolios, which consist of the stocks traded by quantile 5 directors and officers demonstrate varying performance depending on liquidity of these stocks. Formally, the second hypothesis is:

H2: Portfolios constructed on the basis of quantile 5 directors and officers' disclosures will exhibit varying performance depending on the liquidity of stocks included in these portfolios

It should be mentioned that achieving significant and positive abnormal returns can serve as an evidence against the efficient market hypothesis. The significant ability of insiders to predict future stock performance places doubt on the relevance of strong-form market efficiency, which implies that abnormal returns cannot be earned even when trading on private insider information.

Chapter 2. Empirical set up

After stating the hypotheses, the empirical set up in more details should be considered. This is logically divided into the 2 parts: data gathering and analysis (with description of how each hypothesis will be tested).

The data gathering is an operose process for the reasons that various pieces of data come from different sources as well as the fact that overall sample is rather large and requires time to obtain it. The environment used for all further gathering and processing is R Studio. The first and most important part of data is represented by transactions, which were conducted by the insiders between 2012 and 2017. According to the modern Security Exchange Commission's requirements, insider is obliged to report all his transactions in the special Form 4 and submit it into the Electronic Data Gathering, Analysis, and Retrieval (EDGAR). The typical form consists of 3 main parts and looks the following way:

Picture 8Source: sec.gov

The first part has the general information about insider and the company he is occupied in: names, address, ticker and nature of insider's ownership. One becomes an insider due to either being a director or a senior officer or beneficial owner of more than 10% of voting shares. It is important to distinguish between these types due to the fact that 10% owner may not have an access to the same information as its management. It will also be important for testing the hypotheses about insider's seniority, which is applicable to employees only.

The second part contains information about transactions being reported. Namely, this includes that title of security, date of transaction, amount and prices.

The third part is similar to the previous, however it is devoted to derivative securities involved. In this study we focus only on non-derivative securities, thus this part of form will be ignored hereafter.

Interestingly, the HTML format (see the previous picture) does not provide all the information supplied by insider. It is also difficult to gather data from this format. As a result, from the technical point of view, the XML format of representing data is more convenient and informative. In fact, the scripts used for this study involve processing the XML format. Unlike HTML, it looks less representative, however:

Picture 9Source: sec.gov

The data here is represented in the form of tree, where particular piece of information is enclosed within various tags. We can see, for instance, the tag <rptOwnerCik>, which provides the individual number of particular insider - Central index key (CIK). This information is not presented in the HTML-type form, while it is quite important: using the individual number of each insider will allow tracing all his previous actions and being sure that this is the one person.

Besides raw transactions, another important piece of data to be collected was related to stocks, especially prices and returns, which are further used as dependent variables in regressions as well as measures of constructed portfolios performance. The convenient library called `BatchGetSymbols' was used to download stock-related data. The source of information that this package involves is Yahoo.Finance, which is one of the most popular open platforms containing stock data. For each transaction, the future returns have been mapped for further analysis. In spite of access to the Bloomberg terminal, it is less preferred option for obtaining the stock data given the limit of 250 tickers per months. The total count of tickers involved in this study exceeds 5000.

Another supplementary data included the dates of earnings announcements for the period from 2011 to 2018, which was sourced from the “Fidelity investments service” using the scrapper written specially for this purpose, because no ready open source solution has been found.

The final piece of data included other stock and market related data, namely monthly factor data as well as capitalization, Book-to-market ratio and bid-ask spread. The factor data was soured from Fama and French web site, while as other number were downloaded using `Quantmod' package in R.

Given all the necessary data gathered involving the tools described above, the paper advances to the more detailed description of analytical part.

Testing the Hypothesis 1 involves estimating the regression models under various liquidity segments of the markets and afterwards comparing the impact of previously described dummy variables. Each transaction disclosed by insider serves as a unit of observation. Depending on the profitability quantile as well as seniority insider, the predictive power of each transaction may vary. The goal is to analyze at which extent (if actually significant) one particular attribute will influence future return depending on the level of market liquidity. Thus, the universe of stocks will be divided into subsamples according to their liquidity level, which is proxied by the bid-ask spreads. By definition, higher bid-ask spread of the stock implies its lower liquidity. The liquidity segments are determined by distributing the stocks into terciles by the size of the average bid-ask spread quoted. This results in three subsamples. In addition, the whole sample will be considered as well. Finally, the control variables, such as Book-to-market, Size and Market risk premium are included as well.

To summarize, the model will be estimated separately for buy and sell transactions, like in the study of Cline et. al to see whether insiders `informativeness' depends also on the direction of the transaction. Each company's stock, which is involved in the transaction, is included into one of three liquidity groups: high, middle and low. The bid-ask spread is used as proxy for liquidity measure. The models will also be run on the whole sample, without distinction on liquidity to see the impact of informed insider on the whole market. As a result, it makes 8 regressions in total. It is expected that coefficients for quantile 5 (top 20%) insiders of various seniority (directors and officers) will vary depending on the stocks' liquidity within particular sample (Hypothesis 1).

Additional manipulations are required to sort insiders into the quantiles according to their profitability. In order to that, methodic similar to Ali, et. al (2017) will be employed and will involve several steps:

1) Obtain a sample of insiders trading not earlier than 21 trading days (around 1 calendar month) before the quarterly earnings announcement (QEA). The transactions occurring in the last 2 days preceding QEA are excluded from observations to avoid the excess noise, similarly to Ali et. al.

2) For each insider within this sample obtain the list of transactions, in which he participated and the following measure is calculated:

Where represents return on particular transaction n days after it and represent weighted average market return during the same period. The necessity to subtract the market return for the same period is explained by the fact that we are interested particularly in abnormal returns earned by insider, that whether he actually over-performs the market due to the private information.

3) Given that this study tends to focus on insiders rather than on their transactions, there should also be a method of consolidating all profits earned by insider during the period into one number. The method used in this study is slightly different from the one proposed by Ali et. al (2017). The average profit of each insider trading in the pre-QEA period is found as weighted sum of returns from buy transactions less weighted sum of sell transactions, where weights are calculated as absolute dollar value of each transaction divided by the total dollar volume of transactions conducted by particular insider:

Where T is the total number of transactions in which insider engaged during the pre-QEA window. is the weight of each transaction. The value is calculated as the number of stocks multiplied by the transaction price per share. In cases when securities were disposed, the consequent return was multiplied by -1. As a result, unlike in the study by Ali, et. al, in which arithmetic mean of returns on transactions was used, the given measure provides more weight to the transactions implying larger value.

For each insider, who engaged in pre-QEA transaction, such metrics are calculated on the basis of previous 3 years. Therefore, given a sample from 2012 it is possible to sort insiders by quantiles since 2015. The result of this part is a table representing the list of insiders' CIKs (unique number) and their corresponding profit quantiles in the perspective from 2015 to 2017.

Finally, there results are mapped to original list of all transactions reported by all insiders. Each transaction was performed by particular insider and therefore, it is possible to assign quantiles determined to particular transaction. There are transactions performed by insiders who did not trade during the QEA. For such transactions, no quantile is provided.

Turning to the testing of Hypothesis 2, in order to assess the profitability of the strategy based on picking same stocks as quantile 5 insiders, the portfolios have to somehow be constructed. The approach for constructing portfolios is the following. First of all, for each year (2015-2017), the list of insiders and quantiles is obtained, as well as corresponding tickers for the companies in which they are occupied. As a result, each year from 2015 to 2017, within the whole time frame, transactions are observed on the monthly basis. The rule for including the stock in portfolio was the following: if during the specific month, the transaction was made by insider who was ranked with particular profit quantile, such transaction is included into either buy or sell portfolio of corresponding profit quantile depending on whether one acquired or disposed the stock. In case there are several transactions performed by insiders within same quantile, the values are summed and if the direction of these transactions varies, both are included separately into corresponding buy or sell portfolios. Such portfolios are created separately for each quantile and also distinguish by the grade of insider within the company. Namely, the buy and sell portfolios are constructed for 1) directors, 2) officers and 3) both directors and officers, totaling 5 (quantiles) * 2 (buy or sell) * 3 (seniority groups) = 30 portfolios. At the same time, 10% owners are neglected, because from the intuitive point of view they are expected to have less or no access to company's private information at all and as a result demonstrated no significant results in previous studies. For each portfolio, stock positions are rebalanced each month, according to the transactions, which occurred during corresponding month. Given 3 years sample and the fact that each type of portfolio is reconstructed every month, the procedure is repeated 36 times. By the end of each period, returns of each position are obtained and weighted monthly return of portfolio is calculated. The weights are calculated according to the share within total value of transactions for particular stock, where value is measured as number of stocks per transaction multiplied by transaction price per share. The study by Ali et. al has shown higher performance of such value-based portfolios comparing to equal-weighted case. For this reason, value-based weighting approach will be employed in this study.

Chapter 3. Results

The obtained data can be summarized in the following tables:

Table 1Source: calculations of the author

The table represents the statistics on the number of insider transactions, which occurred from 2015 to 2017. “Directors” and “Officers” rows demonstrate the number of transactions, in which corresponding type of insider was involved. “Quantile 5” is a total number of transactions, in which top 20% profitable insiders were involved.

2015

2016

2017

Buys

Sells

Buys

Sells

Buys

Sells

Total transactions

3845

17108

3397

15709

2459

18395

% of total per year

18,3%

81,7%

17,8%

82,2%

11,8%

88,2%

Directors

2357

6944

1634

11916

1757

7744

Officers

1708

13063

1893

18745

903

13940

Quantile 5

115

961

77

999

64

1124

Table 2 Source: calculations of the author

The table represents summary statistics on the firms, securities of which have been involved in the insider transactions. Size shows the market capitalization of the firm in $ Millions. BM is the book-to-market ratio based on the total value of equity reported in balance sheets.

N

Mean

25th Percentile

Median

75th Percentile

Size ($M)

60931

1132,04

2.6

11.7

904

BM

60931

0,4559

0,2279

0,5071

0,8536

Bid-Ask

60931

0,1044

0,01

0,02

0,07

# of unique companies

1901

According to the previously described procedure, 8 various regressions have been constructed. The results are summarized in the table:

Table 3Source: calculations of the author

Dependent variable: future 1-month return

Total Buy

Total Sell

Low Liq Buy

Low Liq Sell

Mid Liq Buy

Mid Liq Sell

High Liq Buy

High Liq Sell

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

B/M

-0.004

-0.003

-0.003

0.172***

0.033

-0.023

-0.082***

-0.005**

(0.005)

(0.002)

(0.003)

(0.026)

(0.063)

(0.026)

(0.030)

(0.002)

Size

-4.140

0.330**

4.344*

1.454***

-13.655

0.053

-6.888*

-1.290*

(2.650)

(0.134)

(2.578)

(0.297)

(13.276)

(0.150)

(3.790)

(0.694)

MRP

0.090***

0.094***

0.101***

0.091***

0.106***

0.097***

0.064***

0.093***

(0.007)

(0.001)

(0.008)

(0.003)

(0.015)

(0.002)

(0.009)

(0.003)

Director

-0.655***

0.476***

-1.358***

0.583***

-0.070

0.446***

-0.187

0.348***

(0.159)

(0.031)

(0.190)

(0.060)

(0.355)

(0.048)

(0.209)

(0.056)

Officer

0.231

0.580***

0.305*

0.742***

-0.552*

0.539***

1.155***

0.417***

(0.149)

(0.038)

(0.180)

(0.072)

(0.326)

(0.058)

(0.204)

(0.069)

10% Owner

1.196***

1.404***

0.487***

1.698***

2.241***

1.625***

0.609***

0.740***

(0.150)

(0.045)

(0.186)

(0.087)

(0.322)

(0.068)

(0.214)

(0.086)

Other

0.812***

-0.281***

2.115***

0.895***

0.026

-0.154

0.202

-1.425***

(0.253)

(0.091)

(0.297)

(0.165)

(0.573)

(0.153)

(0.338)

(0.157)

Quantile 5 Director

-1.362

0.107

5.607***

-0.700***

3.401*

-0.584***

2.750

0.059

(0.987)

(0.124)

(1.071)

(0.215)

(2.000)

(0.187)

(2.433)

(0.264)

Quantile 5 Officer

1.368

-0.391***

2.658

-0.893***

1.638

-0.597***

4.501

-0.713

(1.732)

(0.102)

(2.297)

(0.154)

(3.089)

(0.176)

(3.838)

(0.814)

Quantile 5 10% Owner

-0.863

-0.128

0.483

-0.020

-5.930

-0.801***

-2.420

-0.014

(0.603)

(0.141)

(0.504)

(0.231)

(9.189)

(0.233)

(2.431)

(0.316)

Quantile 5 Other

-3.086***

0.060

-2.122*

-1.414***

-2.732*

0.320

-0.196

2.150**

(0.841)

(0.363)

(1.149)

(0.431)

(1.432)

(1.466)

(3.448)

(0.907)

Constant

-5.204***

-6.909***

-5.099***

-7.305***

-5.993***

-6.856***

-4.547***

-6.554***

(0.218)

(0.047)

(0.256)

(0.090)

(0.475)

(0.074)

(0.299)

(0.087)

Observations

9,710

51,221

3,973

14,463

3,694

19,633

2,043

17,125

R2

0.030

0.097

0.077

0.117

0.030

0.115

0.045

0.080

Adjusted R2

0.029

0.097

0.074

0.116

0.027

0.115

0.040

0.080

Residual Std. Error

6.852

3.396

5.093

3.375

9.183

3.260

4.199

3.531

F Statistic

27.520***

502.662***

30.030***

173.953***

10.358***

232.564***

8.720***

135.557***

Note:

*p**p***p<0.01

The obtained results demonstrate the applicability of proposed hypothesis 1. Indeed, the dummy variables of core interest demonstrate the varying predictive power.

First of all, within the low liquidity segment, Directors assigned to quantile 5 profitability (that is the top 20% of the most profitable directors, who traded during the earnings announcement period) exhibit the highly significant (P-value is approximately 0) results among both buy and sell transactions. Furthermore, it can be seen that among low liquid companies, the fact of transaction being performed by quantile 5 directors adds 5.61% to the future return, 1-month since transaction date. The effect of sell transaction by director, although still highly significant, demonstrates relatively low effect with only -0.7 to future 1-month return. Regarding the officers, they demonstrate significance only within sell-side transactions on the low liquidity market. The size of the effect not large either, -0.89% only. The fact that directors have both coefficients significant, especially buy (given its rather large size of 5.61%) allows to hypothesize that directors may possess more private information than officers may. One of the possible explanations is that officers tend to belong to the too high level of company hierarchy, which excludes them from some day-to-day, still valuable and private knowledge about the facts and events being fundamental enough to drive the subsequent stock performance. The significance of the coefficients discussed can serve as an evidence against the strong form of market efficiency as it is clearly shown that insiders who are very likely to be informed indeed tend to profit on the private information. It is questionable however, whether using this statistical predictions will provide opportunities of earning abnormal returns on the market. First of all, it is not guaranteed that market will allow earning on this information once it becomes disclosed. In case market is still semi-strong efficient, the new information will be immediately incorporated into the price and therefore will remove any profit opportunities. Secondly, even if the market is not semi-strong efficient, the presence of opportunity to buy or sell stock before the whole market reacts still does not guarantee the profits. This is due to the fact that securities on this market are low liquid and therefore imply high bid-ask spreads. As a result, the profit from realized value of stock may be compensated by the wide bid-ask spread quoted by the market maker. One can also notice that as liquidity of the sample increases the coefficients of main interest lose their significance. To summarize on the results of low liquid market, it can be said that transactions of insiders selected by the methodology of Ali et al who are presented in the low-liquidity sample have significant predictive power of the future return. Although the possibility of successfully following such insiders and earn abnormal returns is not clear yet.

Turning to the medium liquidity market, the similar pattern is observed, although at lower scale in terms of both coefficients' absolute size as well as their significance. As a result, the fact some stock was involved within the transaction of quantile 5 director, now yields only additional 3.4% to future expected 1-month return, which is 2.2 p.p. lower. In addition, the coefficient is now significant at only 5% significance level. Regarding the sell transaction performed by quantile 5 director, the expected incremental effect on return is now around -0.58%, which is slightly lower in absolute terms. Still the coefficient demonstrates high significance. In case of officers, pattern is again similar: coefficient is insignificant in case of buying transaction and highly significant for sells. Interestingly, its value is very close to the one corresponding to quantile 5 director, implying that expected effect from this type of transaction will be similar regardless the seniority (officer or director) of the transacting.

In case of the highest by liquidity segment of the market, neither director nor officer demonstrate the significance of corresponding coefficients. This implies that those insiders who are expected to be the most informative according to the methodology of Ali et.al do not possess any predictive power within their disclosures.

The preliminary conclusion is that previously described observations are consistent with the Hypothesis 1. The segment that stands out the most is the one with most liquid stocks. Within this segment, the transactions picked by the method combining Ali's and Cline's show no predictive power at all. The intuition provided in the beginning of this study is also consistent with this result. If transactions of specific insiders do not exhibit significant predictive power, then insiders who perform these transactions are likely to be uninformed. In this case, market maker is not confused by the trading, which occurs on private insider information and has no incentives to quote high bid-ask spread, which would otherwise compensate for potential unexpected price changes.

Finally, the models for the whole market should be considered. In case of buy transactions, neither of insiders considered previously provide the significant signal. Regarding the sample of sell transactions, only officers demonstrate the significant impact of their transactions on future return. Still, the absolute value of that effect is lower than in case of both low and medium liquid markets. Overall, this mixed evidence obtained on the whole market relative to low and medium liquidity segments, demonstrates that indeed dividing the universe of stocks in subsamples according to the level of their liquidity allows obtaining additional insights on “informativeness” of insiders' transactions. Therefore, the regression analysis did not reveal any evidence allowing to reject the Hypothesis 1.

The first part of empirical analysis demonstrated that distinguishing samples by liquidity indeed provides various results. As a next step, it would be interesting to test whether performance of portfolios constructed by relying on the actions of quantile 5 insiders will also yield different results. This part of empirical analysis is devoted to testing the significance of abnormal returns, which are earned from portfolios that follow the transactions of insiders who traded the securities from the least as well as most liquid part of the market. The portfolios constructed involve all quantiles separately as well as distinction between directors and officers. As a result, 30 various portfolios have been constructed within each liquidity segment distinguishing by the quantile assigned to insider as well as seniority of these insiders within the company. The two methods are used to evaluate the abnormal returns. In the first one, returns of each portfolio are regressed on excess market returns solely. Formally the equation is:

The second method involved Fama French 3 factors:

This methodic is similar to the one employed in study by Ali et. al. The results are summarized in the followin Table X. High-liquidity portfolio alphas regressed on market excess returns. For the convenience, only alphas are reported from the models estimated, while as sensitivity to factors is not in the scope of current analysis.

Table 4.Source: calculations of the author

Low-liquidity portfolio alphas regressed on market excess returns. For the convenience, only alphas are reported from the models estimated, while as sensitivity to factors is not in the scope of current analysis.

...

Director

Buy

Director

Sell

Officer

Buy

Officer

Sell

Director and Officer Buy

Director and Officer Sell

Quantile 1

-0.01

0.012

0.011

-0.091

0.043

-0.02

(0.082)

(0.093)

(0.079)

(0.041)

(0.054)

(0.030)

Quantile 2

0.034

-0.11

0.192

-0.125

0.542

-0.08


Подобные документы

  • The grey market is an over-the-counter market where dealers may execute orders for preferred customers as well as provide support for a new issue before it is actually issued. Sometimes, "dark markets” are referred to as a third type of grey market.

    презентация [2,1 M], добавлен 18.12.2013

  • Commercial banks as the main segment market economy. Principles and functions of commercial banks. Legal framework of commercial operation banks. The term "banking risks". Analysis of risks and methods of their regulation. Methods of risk management.

    дипломная работа [95,2 K], добавлен 19.01.2014

  • Main segments of the financial market: investment, loan, stock, insurance, foreign exchange markets. Top 10 currency traders of overall volume. Internationalization of the national currency. The ratio of US Dollar and Euro against ruble in 2009-2012.

    доклад [115,0 K], добавлен 14.12.2013

  • Theoretical basis of long-term loans: concept, types. Characteristics of the branch of Sberbank of Russia. Terms and conditions of lending to households in Sberbank of Russia. Financing of investment projects. Risk - the main problem in the credit market.

    реферат [28,0 K], добавлен 17.09.2013

  • The behavior of traders on financial markets. Rules used by traders to determine their trading policies. A computer model of the stock exchange. The basic idea and key definitions. A program realization of that model. Current and expected results.

    реферат [36,7 K], добавлен 14.02.2016

  • A bank: nature of activity, main business-processes and organizational structure, the market place and history. Definitions of the project and project management, the project life cycle. Management of development projects in a bank, the expected results.

    реферат [20,6 K], добавлен 14.02.2016

  • History of introduction of a modern banking system to the Muslim countries, features of their development and functioning in today's market economy. Perspectives of future development of Islamic banking in the world and in the Republic of Kazakhstan.

    курсовая работа [1,3 M], добавлен 19.04.2012

  • Meaning of currency operations and order in relation to the currency of legal entities and individuals. Currency regulation and currency control. Term (forward) operations in foreign currencies. Foreign exchange transactions. The transaction "swap".

    курсовая работа [35,0 K], добавлен 22.12.2011

  • Description of exchange stocks as financial point-of-sale platforms. Description of point-of-sale algorithm of broker trade at the financial market. Parameters of price gaps on financial auctions and optimization of currency point-of-sale algorithms.

    контрольная работа [1011,9 K], добавлен 14.02.2016

  • Development banking, increasing the degree of integration of the banking sector of Ukraine in the international financial community, empowerment of modern financial markets, increasing range of banking products. The management mechanism of bank liquidity.

    реферат [17,2 K], добавлен 26.05.2013

  • The history of the development of Internet banking in Kazakhstan and abroad. Analysis of the problems faced by banks in the development of this technology. Description of statistical of its use and the dynamics of change. Security practices for users.

    презентация [1,3 M], добавлен 24.05.2016

  • Business plans are an important test of clarity of thinking and clarity of the business. Reasons for writing a business plan. Market trends and the market niche for product. Business concept, market analysis. Company organization, financial plan.

    реферат [59,4 K], добавлен 15.09.2012

  • The stock market and economic growth: theoretical and analytical questions. Analysis of the mechanism of the financial market on the efficient allocation of resources in the economy and to define the specific role of stock market prices in the process.

    дипломная работа [5,3 M], добавлен 07.07.2013

  • Current situation and market condition in which is Indes, offered services and company problems. Segmentation of the market and an industry condition. The analysis of possibilities and threats of firm, action for advancement of business processes.

    курсовая работа [21,5 K], добавлен 08.01.2012

  • Types and functions exchange. Conjuncture of exchange market in theory. The concept of the exchange. Types of Exchanges and Exchange operations. The concept of market conditions, goals, and methods of analysis. Stages of market research product markets.

    курсовая работа [43,3 K], добавлен 08.02.2014

  • Natural gas market overview: volume, value, segmentation. Supply and demand Factors of natural gas. Internal rivalry & competitors' overview. Outlook of the EU's energy demand from 2007 to 2030. Drivers of supplier power in the EU natural gas market.

    курсовая работа [2,0 M], добавлен 10.11.2013

  • Ability of the company to reveal and consider further action of competitive forces and their dynamics. Analysis of environment and the target market. Functional divisions and different levels in which еhe external information gets into the organization.

    статья [10,7 K], добавлен 23.09.2011

  • Law of demand and law of Supply. Elasticity of supply and demand. Models of market and its impact on productivity. Kinds of market competition, methods of regulation of market. Indirect method of market regulation, tax, the governmental price control.

    реферат [8,7 K], добавлен 25.11.2009

  • Executive summary. Progect objectives. Keys to success. Progect opportunity. The analysis. Market segmentation. Competitors and competitive advantages. Target market segment strategy. Market trends and growth. The proposition. The business model.

    бизнес-план [2,0 M], добавлен 20.09.2008

  • Mergers and acquisitions: definitions, history and types of the deals. Previous studies of post-merger performance and announcement returns and Russian M&A market. Analysis of factors driving abnormal announcement returns and the effect of 2014 events.

    дипломная работа [7,0 M], добавлен 02.11.2015

Работы в архивах красиво оформлены согласно требованиям ВУЗов и содержат рисунки, диаграммы, формулы и т.д.
PPT, PPTX и PDF-файлы представлены только в архивах.
Рекомендуем скачать работу.