Размещено на http://www.allbest.ru/

Федеральное государственное автономное образовательное учреждение высшего профессионального образования

«Национальный исследовательский университет «Высшая школа экономики»

Международный институт экономики и финансов

Курсовая работа / Выпускная квалификационная работа - Бакалаврская работа

по направлению подготовки 38.03.01 «Банковское дело и финансы»

образовательная программа «Программа двух дипломов по экономике НИУ ВШЭ и Лондонского университета»

«Сигнальный эффект обратного выкупа акций в России» («Signaling effect of the «buyback» of shares in Russia»)

Кючюк Джесур

Рецензент В.И. Черняк

Научный руководитель А.Э. Улугова

Москва 2016



Table of Contents

Introduction

Chapter 1. Review of literature

1.1 General theoretical framework and methods of «buyback»

1.2 Reasons of «buyback»

Chapter 2. Methodology and Data

2.1 Hypotheses

2.2 Methodology

2.2.1 Event study

2.2.2 Regression analysis

2.3 Data for the research

Chapter 3. Empirical Conclusions

3.1 Event study conclusion

3.2 Regression analysis conclusions

Conclusion

References

Appendix 1



Introduction

In recent several years the role of Russian financial markets is growing constantly. The participants of Russian financial markets gaining the bigger role in the overall financial world perspective over years. Thus the optimal financial decisions by the firms are more important and crucial than before.

Managers of the firms are faced with lots of decision making problems in modern world, such as: payout policy, redistribution of excess capital etc. Moreover, the first factor taken under consideration by the companies management is shareholder wealth. Nevertheless, stated above factor can be harmed by the mispricing of the firm by the market or by the presence of the asymmetry of information problem. At the same time for such purposes repurchase of own shares of the company is the common way for several firms abroad.

The main aim of the following study is to outline and test the main reasons of managers to participate in «buyback» operations in Russia. What is more, this work is also concentrated on the effect on the stock price by the announcement of firms buying back their own shares.

To answer to stated questions following procedures should be intaken:

1. Deep analysis of the theoretical framework and the empirical studies of the «buyback».

2. The choice of the methodology of the study and the choice of the compatible with Russian realities.

3. The collection of needed data of stock prices and financial indicators for the firms participated in «buyback» in Russia.

4. Evaluation of the effect of the «buyback» announcement on the stock price and signaling effect.

5. Outlining and empirically studying the main reasons for repurchase of own shares on Russian financial markets.

6. Interpret the conclusions of the empirical research about the effect on the stock prices and determinants of the motivation of the «buyback»

Plenty of researches and empirical studies was conducted on the basis of «buyback» on leading financial markets such as U.S. and U.K. and only few of them concentrated on developing markets, but the «buyback» phenomena stays almost unstudied.

Talking about novelty of the following work several new methodologies can be outlined:

1. The study of the effect on stock price by «buyback» announcement conducting event study methodology in Russia.

2. The research of main determinants of repurchase of shares of firms traded in Russia applying regression analysis.

3. The study of the effect of takeover danger on motivation of firms for repurchase of shares on Russian financial markets using dummy variable.

4. Applying Logit regression model and studying marginal effects of variables.

In the first chapter of the work the main theoretical aspects and empirical studies will be concerned in order to evaluate the main reasons for «buyback» and effect of such event on several factors. Moreover, obtained information will be used as back-up arguments for the original research of the following work. Next step will be deep discussion of the methodology of the research and data collected in order to perform it. After that chapter 3 will outline the main conclusions obtained from the empirical and theoretical studies of the work.



Chapter 1. Review of literature

In this part I will outline the most common causes which answer the question «Why firms repurchase their own stocks» and «Which processes do they usually use do perform such operations». Considered facts will be backed up by theoretical framework and empirical evidence.

1.1 General theoretical framework and methods of «buyback»

To begin with considering U.S financial markets Grullon and Michaely (2002) in their «Dividends, Share Repurchases, and the Substitution Hypothesis» work mentioned that corporations tended to use dividends as payout mechanism over «buyback» of their shares. Nevertheless they based their work on the growing use of repurchasing stocks mechanism starting from 1980. Initially expenditures on «buyback» policies were 4,8% and grew constantly to 41% over the twenty years starting from 1980. Trying to explain such effect a lot of empirical research papers outlined several reasons why companies started to use repurchasing mechanism more often nowadays.

Continuing the general information about stock repurchasing processes I should clearly outline the ways how company did it. Theo Vermaelen in 2006 evaluated 4 main options which companies used for «buyback».

· Fixed-price tender offer.

Such mechanism usually implies that company makes an announcement in which basically investors face the offer to repurchase fixed amount of shares on fixed price.

· Dutch auction tender offer.

This process in contrast to fixed-price tender offers a variety of prices to investors expecting from them to choose their minimum price which they can accept. Dutch auction tender was used by Tribune in 2006 which represented the repurchase of 25% of outstanding shares according to The New York Times.

· Repurchase of shares on the open market.

This is the most commonly used way of «buyback» in the modern financial world. According to Vermaelen 2006 90% of all repurchasing policies were performed through repurchase on the open market. Such method represents operations on open market in order to specify the amount of needed shares followed buy repurchasing of them.

· Repurchase of shares from large investors.

Usually this mechanism constitutes buy-back on discount or at premium from large investors or potential raiders.

1.2 Reasons of «buyback»

After answering «How» I will concentrate on more difficult and ambigious question «Why». First of all, continuing discussion about payout policies there is empirical evidence that «buyback» can be considered as substitute to dividends payout. Repurchasing stocks increases demand on companies shares hence increases the price of stock. Moreover, «buyback» of shares can be seen as signal that company is undervalued which can also increase the price of the stock. Nevertheless, the reason of «buyback» backed up by the fact that company is undervalued will be discussed later. Summing up, the payments to shareholders can be achieved through increase of the stock price after repurchase thus «buyback» can be сonsidered as alternative to dividends.

On the one hand, Brav, Graham, Harvey and Michaely (2005) in their «Payout policy in the 21^st century» research conducted survey of 384 companies which showed that dividends decisions are made at the same time with investment choices, but share repurchase usually performed after investment decisions. Moreover, according to such results from survey Brav concluded that choice between dividends and repurchase is question concerned about own preferences of the firm. Summing up the article, it can be stated that one of the causes for using «buyback» is that it is a good alternative to dividends and one of the reasons why it can be better depends on the priorities of the company.

On the other hand, choice between dividends payout policy and «buyback» operations can be explained due to difference in tax rates on dividends and personal capital gains. Since different countries have different taxing systems sometimes when capital gains has tax advantage over dividends companies have an incentive to redistribute their income using repurchase of stocks rather than dividends.

Tax rates on dividends and capital gains in U.S. in 1961-2011

SOURCE: HOLT, American Council for Capital Formation and LMCM analysis

To confirm such statement the study of Lie and Lie (2000) can be taken under consideration, in which they analyzed choice of payout policies over different tax rates. They empirically proved the very logical fact that managers of the companies are sensitive to tax faced by investors. In other words companies tends to repurchase shares when tax rates of capital gains are lower than tax rates on dividends.

Moreover, besides the tax effect study of Jagannathan, Stephens & Weisbach (2000) tried to understand what facts determine the choice between stock repurchase and dividend payout. They empirically proved that companies with less stable cash flows and less operating income share in overall income prefer to use repurchase operations over dividends policy. Other study analysed the choice of «buyback» and dividends from the signaling perspectives. Ofer & Thakor (1987) constructed the theoretical model which showed the signalling effects with respect to different payout policies. Even that the model was very difficult, they got the result that both dividend and repurchase policies are followed by signals and moreover that signals are not dominated by each other.

These studies explain why companies appeared to substitute dividends payout mechanisms with «buyback» of own shares. What is more, such studies (Fenn and Liang (1998) , Jagannathan, Stephens & Weisbach (2000), . Ofer & Thakor (1987) )confirmed that firms still use dividends payout policy even if tax rate on dividends payments are higher that tax rate on capital gains.

Continuing the discussion of reasons for repurchasing shares after «alternative to dividends» reason the most direct and straightforward reason is that managers of firms appeared to think that their companies are not priced fairly and that their true value is higher than the actual one. Going back to Brav, Graham, Harvey and Michaely (2005) survey the proved that the general reason of «buyback» of shares is the fact that market undervalue firms. In result managers can buy shares on discount to their fair value and create shareholder value since they will expect prices to go up soon. The idea of undervaluation can be also considered as signaling hypothesis. Since managers start to repurchase their own stocks investors can see that actions as a signal that future horizons of company development included in the stock price are not very precise. Going back again to question «Why companies repurchase their own shares» one can give an explanation due to undervaluation and signaling theories.

Arguing information above, we can rely on some empirical studies. Since we concluded that managers are willing to perform «buyback» only when prices are low and not fair, there are several researchers which proved this fact. Dittmar (2000) testing hypothesis of what determines the decision to buyback for firms in her «Why do firms repurchase stocks» study achieved the result that managers use «buyback» mechanism to take advantage of undervaluation of firm. This result was unexpected since very large corporations were included in sample and they were expected to be fairly priced. There are more empirical studies that arguing about «signaling effect» as the main reason for «buyback». For instance, Stephens and Wiesbach (1998) contribute to this theory with their result that firms who repurchased their stocks had negative returns on shares of their company. This study also «speaks for» signaling effect as the reason for «buyback». There are two more papers which mention signaling effect such as Lakonishok, Shleifer and Vishny (1994) and survey of Chief Financial Officers (Brav 2005). Lakonishok, Shleifer and Vishny (1994) stated that firms with low market - to - book ratios had positive and even outperforming market returns in periods after the announcement of «buyback». Latter survey of CFOs showed that the biggest incentive for repurchasing own shares is due to the fact that managers saw perspectives of buying a stock at undervalued price knowing that firm is mispriced

To back up above information by words of one of the greatest investor of all times Warren Buffet which he said at 2000 and outlined two key reasons why the firm should conduct `buy-back»:

«There is only one combination of facts that makes it advisable for a company to repurchase its shares: First, the company has available funds -- cash plus sensible borrowing capacity -- beyond the near-term needs of the business and, second, finds its stock selling in the market below its intrinsic value, conservatively calculated.» (Warren Buffet, 2000 )

Theories stated above outlined that «buyback» of shares mostly driven by motivation of taking advantage of undervaluing prices. Nevertheless, there are still several research papers which were concentrated on ex-post effect of repurchasing shares in order to signal.

Theo Vermaelen (1981) performed an event study in his «Common stock repurchases and market signalling: An empirical study» work to see if there any abnormal returns after stocks repurchase. He included several announcements about «buyback» and get rid of companies which had an announcements of other important information in close dates to repurchase announcements. He constructed expected returns using beta method described in Scholes and Williams in 1977. He used beta basis to calculate control portfolio and compared it to actual returns. So he used such notation fo AR (Abnormal Return):

According to rankings which he got Vermaelen divided securities to 10 equal portfolios. Calculating AR and CAR ( Cumulative Abnormal Return) he got that on time period plus three days; plus sixty dates CAR was insignificant. Summing up his study author concluded that there was no any signal effect due to «buyback».

Similar results gained Liano, Huang and Manakyan(2003) in their «Market Reaction to Open Market Stock Repurchases and Industry Affiliation» where about 1000 open market announcements in 1982 - 1997 period were taken as data. This sample was rearranged to 17 categories depending from which industry the firm is. The average proportion of repurchased shares in outstanding shares was 7,21% and average beta of companies performing «buyback» was 0,78. In contrast to Vermaelen(1981) study they used different methodology calculation abnormal return of shares:



Where is the daily return on CRSP value - weighted index on day t and is the return of stock i in industry j at time t.

In result they got following facts about signaling of «buyback»: in pre-announcement period ( -20;-3 ) the returns was negative and significant while at the announcement date ( -2;+2 ) returns of the stocks were positive and significant. Such conclusions as negative return before the announcement is consistent with studies which proved that companies with negative returns and underpriced shares tend to use «buyback» as a signal and positive returns at dates of announcements showed that repurchasing of stocks increase their value. Moreover according to their study there was no significant AR and CAR at ( +3;+60 ) which is consistent with Vermaelen (1981) results. Also such results can be a signal that markets were semi-strong efficient, since prices reflected the public information immediately ( CAR ( -2;+2) > 0 ) and agents could not exploit information for their own benefits.

To continue with there were several studies which analyzed the effect of volume of repurchase of shares on the strength of the signal. Liu & Ziebart( 1997) took sample of Australian companies and divided them to two subgroups with positive and negative CAR. So analyzing the effect of volume of «buyback» to strength of the signal they found strong and significant relationship between them. In other words more shares firm announce to repurchase the stronger the effect on signal to investors by managers.

What is more Padgett in 2007 got an opposite result of his study. Conducting his research he got insignificant coefficient for the volume on CAR in his regression model. In contrast to Liu & Ziebart(1997) Padgett showed that strength of the signal is not dependent on the announced volume of «buyback» of companies. Moreover, he got a positive relationship of the size of the company and the volume of «buyback». Reading both articles I came to conclusion that such opposite results may exist due to different samples of companies. Liu & Ziebart(1997) examined Australian companies while Padgett(2007) concentrated on UK companies. Talking about Padgett study it is worth discussing another result which he obtained: Signal to investors is not the main motivation for UK companies for repurchase operations, since other motives has dominating effect over signaling hypothesis.

All of the mentioned studies cover only short-run signaling effect and talking about long-run effect, Lakonishok & Vermalen and Ikenberry (1995) studied relationship between «buyback» and signaling effect on longer time horizons. Lakonishok & Vermalen and Ikenberry took sample of 1060 companies in 1989 - 1995 period and used Fama and French model (1992)

where M is market variable, SML is small capitalized stocks over low capitalized stocks snd HML is high market - to - book value over small mrket - to - book value). Conducting their study they got result which was significant monthly positive AR over three subsequent years. What is more, they got significant negative CAR before the announcement. This study is also consistent with the main idea that companies perform repurchasing when the have negative returns to signal to investors that their company is undervalued.

Other not that significant explanations for «buyback» are:

1. Change in capital structure.

Since companies tend to achieve optimal capital structure and optimal level of debt, Kivi 2006 got the conclusion that companies may issue new debt and use repurchasing mechanism to use additional attracted funds. Moreover, the survey of Dixon, Palmer, Stradling and Woodhead (2008) achieved the result that gaining optimal capital structure was the primary motivation for share repurchase.This theory unfortunately doesn't have enough empirical studies to rely on.

2. Defense against Merges and Acquisitions.

Often companies buy their own shares back when they are afraid of being acquired by another firm, so the takeover deterrence can be also explanatory reason for motivation to repurchase stocks.

The comprehensive work in that shield is «The share repurchase and takeover deterrence» by Bagwell (1991). The main outcome of which was if the supply of firms share is upward-sloping there will be evidence that more shares company «buyback» instead of paying dividends the more it difficult for raider to acquire the firm. What is more, Billett and Xue (2007) modelled the takeover probability as latent variable and got the result consistent with Bagwell (1991), which stated that activity of firms in «buyback» increases when they are faced with bigger probability of takeover.

3. Employee stock options hypothesis.

Since firms give their employees stock options as compensation or for other reasons, managers are concerned about dividend payout policy. Managers granted with stock options will be motivated to participate in «buyback» because dividends paid decrease the value of their options. The studis of Fenn and Liang (1997) and Hurtt, Kreuze and Langsarn (2008) are consistent with stated hypothesis.

4. Free cashflow hypothesis.

Padgett and Wang (2007) in their study concluded the results, that in U.K. firms main motivation to buy their own share back is not the undervaluation hypothesis, but the free cashflow hypothesis. They interpreted the result that the primary reason for «buyback» was to redistribute excess capital of the firm using such operations. Woods and Brigham (1966) also concentrated on such hypothesis.

Above information about empirical studies and theoretical frameworks gives the wider perspective about the theme of this study and deeper look inside to the subject in order to analyze signaling effect of buy-back of shares. Moreover these studies will help in further work to study Russian financial market on «buyback» research basis.



Chapter 2. Methodology and Data

2.1 Hypotheses

In this part two methodologies of my research will be examined and discussed. Working on the theme two main problems were stated for research:

1. Buyback operations have significant effect on the price of the stock in subsequent periods on Russian financial market.

Basically if the operation of the repurchase of own stocks by firms will influence the stock price in later time, so it will be consistent with studies of Vermaelen and Liano, Huang and Manakyan who performed event study on the «buyback» dates.

2. Evaluate the determinants of motivation of firms for repurchasing stocks on Russian financial markets.

Plenty of studies were performed on leading financial markets such as U.S. or U.K. to evaluate the main determinants of «buyback», but just few of similar researchers were conducted on Russian financial markets(Kinyakin (2012)). One of the main aims of this study was to stylize research in order to determine the main factors which influenced the decision of the firms to participate in repurchase of their own shares.

During my research following hypotheses were stated.

1. CAR is significantly positive for firms after «buyback».

Afresh according to studies of Vermaelen and Liano, Huang and Manakyan (2003) who stated that «buyback» of shares have several effects such as:

· Signal for investors about undervaluation of the company.

· Increase in demand for stocks of the company

Both of the effects appeared to increase the price of the stock in following future, so direct task was to look at such effects on Russian financial markets and determine if the price of the stock have significantly positive CAR afterwards. This research will show if signaling effect works in proper way in Russia.

2. Undervaluation of the company increases the ptobability of «buyback».

Dittmar (2000) and Stephens and Wiesbach (1998) studied the undervaluation hypothesis. This hypothesis implies that main motivation for repurchasing own securities is the fact that market doesn't price the company fairly and due to the fact that managars have better information than market and market participants so they will show the undervaluation using «buyback». Stephens and Wiesbach (1998) and Dittmar(2000) got the results that undervaluation hypothesis was consistent with the real world. In my research I will expect positive significant relationship between `buy-back» of shares and factors which can show that the company is mispriced.

3. The smaller the company the lower probability of the repurchase.The problem of asymmetry of information.

The smaller the company the less company is monitors. This fact may increase the asymmetry of information between managers and investors. In general the bigger the asymmetry of information the less is signaling effect, so managers don't have enough motivation to repurchase its own stocks. In the study this hypothesis will be consistent if there will be negative dependence of the «buyback» and proxy for the size of the company.

4. The dividends payout policy negatively affects the probability of «buyback».

According to different tax systems and preferences of managers, payout policy for shareholders may be different for companies. First of all, in different countries there are different tax rates on dividends and capital gains, thus according to the fact above managers can choose to repurchase shares as alternative to dividends since tax on dividend may be higher than the tax on capital gains. Such relationship is consistent with study of Lie and Lie (2000). Moreover Brav, Graham, Harvey and Michaely(2005) got the result that the choice among payout policies is the question of preferences of the firm. Talking about my research it will be examined if firms that repurchased stocks paid low divided.

5. «buyback» positively depends on excess capital of the firm.

According to Fenn and Liang (1998) the free capital in firm appeared to spent on not effective investing decisions. This fact can tell about excess capital of firm as the reason for buying back own shares. Moreover, it can also help to avoid or minimize agency costs for company. In given study the significant positive relationship between repurchasement of shares and factors which will show free capital of the company will be expected.

6. Repurchase of stocks is more active when company face the takeover probability.

Main outcome of the study of Bagwell (1991) was that the more shares were repurchased buy the company the harder it was to acquire it for the raider. In fact, that may lead to determine the «buyback» of shares as the defensive strategy for company. The information above clears the fact that deterrence of takeover may be considered as the reason for firms to participate in «buyback». In following study the positive relationship between the willingness to acquire the company and the repurchasing of shares buy the target firm.

2.2 Methodology

In this section of the study information about methodology of the research will be provided. This work consists of two empirical studies which will be the best for testing hypothesis which were stated above. First one is event study to determine the first hypothesis if CAR is significantly positive and regression analysis with help of which the determinants of «buyback» on Russian financial markets will be evaluated.

2.2.1 Event study

MacKinlay in 1997 specified the method to use to see the effect of specific event on the price of the share. According to his work with the help of available financial data the method of event study helps to achieve the measurement of specific event on some factors of the firm, such as price of the stock. MacKinlay (1997)stated that this method may be very useful since having rationality of the market the influence of the effect will be reflected in security value. This study can be constructed using the information about stock prices in relatively short period. Even knowing the fact that first event study work was published by James Dolley in 1993, MacKinlay(1997) was the first who «put everything in one hand».

According to his study there are several steps which should be done in order to achieve needed results. Nevertheless there are no unique algorithm for event study, but he outlined general flow of analysis:

Step 1.

The starting point in order to do event study is outlying the specific event and all information needed connected with this specific date. In my work the specified event is the announcement of «buyback» of shares by the firm. According to the methodology of the event study it is also crucially important to get daily information about stock prices at the dates of the announcement.

Step 2.

The next step is to determine the dates of the specified event which will be called «event window» in following study.

According to event window one can measure the short-run effect and long-run effect. Since the very long event windows, for instance two years after the announcement, may include serious biases and provide inconsistent results. Specifying it more correctly such biases may be other announcements of the firm, such as announcement of dividends payout or announcement of planes acquisition of the firm. Given that I will more concentrate on short-run effects of the event on stock prices in the work. Nevertheless studying short-run effects have its own disadvantages.

To analyse short-run effects on the share price by «buyback» announcement two event windows will be considered:

· Three days before the announcement and four days after: (-3 +3)

· Three days before the announcement and forty days after: (-3 +40)

Since Russian market is considered as developing it may tell that information is reflected not immediately in stock price, so two event windows will be considered.

Moreover, it is important to include dates before the announcement. This may be explained due to information asymmetry problem. Since management of the company knows more than investors it will result in insider effect.

Step 3.

To continue the procedure the step of calculating daily returns of the stock should be undertaken. The algorithm is very simple:

Where is the price at current period; is the price of the stock in previous period. To clarify following mechanism, if return is positive that means that price of the stock went up and if it is negative that means that price decreased. Nevertheless, if return equals to zero, that means that price stayed unchanged.

Step 4.

The next stage is to find normal return. According to MacKinley(1997) description it could be done two ways:

· The constant mean return model.

· The market model.

Latter implies a stable linear relationship between the stock return and market return. To get needed result I should calculate the expected return according to market portfolio. In my case the expected return is calculated using the market index MICEX (Moscow Interbank Currency Exchange) . This will be done by finding stock prices and micex index at the same dates and running ordinary least squares regression with normal return as dependent variable. In my work that was calculated using add-in functions in excel.

Step 5. The following task is to calculate daily abnormal return. The abnormal return can be defined as the difference between actual return and expected return and can be written as:

Where is the abnormal return of stock i at period t; is the actual return of the security and can be defined as normal return and calculated using following equation:

Where are coefficients of the regression and is the return on MICEX index.

In other words if is higher than zero that means that firm outperformed the market at time t and vice versa.

Step 6.

To continue the analysis CAR should be calculated for each event window. CAR can be defined as cumulative abnormal return and after getting all daily abnormal returns it can be calculated as the sum of all abnormal returns for specified company in specified event window:

Where (-t t) are the borders of the event window, for instance in my study one of the specification will be (-3 +40).

Step 7.

The final step is to test significance of the CAR which was get using previously described operations. This can be done in several ways but I chose the way of constructing a new sample of CAR for each company with given event window to test hypothesis:

Where ACAR is the average CAR calculated from collected sample in following way:

Where N is the number of observations and is cumulative abnormal return for a given company in a given event window. Testing such hypothesis can be achieved by implying ordinary t-test. T-statistics is calculated with standard formula for t-test:

Where ACAR is average abnormal return, is the standard deviation of sample of CAR of companies and finally N is the number of the observations. The calculation of standard deviation was performed using add-in formulas in Microsoft Excel.

The final result is compared with critical values taken from student distribution tables and test performed on 1%; 5% and 10% level of significance.

2.2.2 Regression analysis

The main aim of this section is to determine the main reasons for «buyback» in Russian financial markets. The research is proceeded with Logit Binary choice model in eViews 8 software. The significance of the coefficients is also tested with the help of Views. In order to understand what influences the decision of managers to buy their shares back the dependent variable BUYBACK was specified which takes value 1 if «buyback» occurred and 0 if companies didn't repurchase their securities back.

To begin with, the theoretical framework of logit model should be discussed. Logistic regression or simply Logit was introduced by David Cox in 1958 and imply that dependent variable is categorical. Originally logit model was made to estimate the probability of event (response of the dependent variable) in response to specified factors. Talking about given study the event probability of which will be estimated later is the presence of «buyback». The logit methodology is based on logistic function which can specified as:

Where:

Where Z is the dependent variable and are explanatory variables.

To understand it more clearly the logistic function can be explained by the following graphical image:

Logit function graph. Source: wiki.analytica

Collecting data about repurchase of own shares by firms in total sample of 45 effective dates of «buyback» announcements was achieved. Moreover at least 3 observations of each company were taken, one observation of the year before buyback with the value «1» of BUYBACK variable and different years which were not followed by repurchase of shares by company. In this case BUYBACK variable took 0 value.

Due to given study several variables to explain probability of «buyback» were added according to following hypothesis:

1. Undervaluation hypothesis.

This hypothesis states that decisions of firm to buy their shares back are mostly driven by the fact that market undervalue the company and investors may signal about that by this mechanism. As the proxy for undervaluation of the company the natural logarithm was chosen of price over earnings ratio in year prior to year of repurchase of the stocks. This ratio is usually used for valuing the company and low values of P/E index can signal about mispricing of the company by the market.

There are several ways in calculating this ratio but in general P/E index is calculated using following formula:

This ratio is often use to determine if company is priced correctly and includes the market perception of the risk of the given company. Collecting sample of 45 announcements of «buy-backs» such descriptive statistics for LNPE variable were gotten:



VARIABLE: LNPE
Average	StandardDeviation	Minimum	Maximum	Range
2.36	0.77	0	4.49	4.49

Due to the undervaluation hypothesis significant negative dependence of the BUYBACK variable and LNPE variable should be expected.

2. «buyback» as alternative to dividends hypothesis.

According to this hypothesis firms can construct payout policy to their shareholders using repurchase of share mechanism. Since buying back will increase demand for company stocks it will be followed by increase in price of the stocks in theory. Shareholders instead of dividends payment will get increased capital gains. The choice between dividends and «buyback» is determined by preferences of the firm and the difference in tax systems. As the measure of dividends payments DIV_PAYOUT variable was taken which shows how much dividends paid in cash were maid as the proportion of the net income of the company. Similar variable wes taken by Dittmar (2000) in her «Why do firms repurchase stocks study». The DIV_PAYOUT variable were calculated using the information of cash dividends paid over net income in year prior to the repurchase operations and was calculated by the formula :

And has descriptive statistics:

VARIABLE: DIV.PAYOUT
Average	Standart Deviation	Minimum	Maximum	Range
0.15	0.99	-9.00	4.41	13.41

3. Asymmetry information hypothesis.

This hypothesis implies that lower size firms have less incentive to repurchase stocks because they are less monitored and will have lower signaling effect of the «buyback».

The LNREV variable was taken as proxy for the size f the firm. This variable is the natural logarithm of the collected information of the revenues for the company. From the sample of firms following descriptive statistics for LNREV variable were achieved:

VARIABLE: LNREV
Average	Standart Deviation	Minimum	Maximum	Range
18.06	2.37	10.86	22.41	11.54

4. Excess capital hypothesis.

This hypothesis states that «buyback» is a good way to use excess capital of the company. As the proxy for that CASHFLOW variable was taken. Similar methodology was used afresh by Dittmar (2000). CASHFLOW variable is the variable which can approximately show how much free funds does company have. To be consistent with the hypothesis significant positive dependence of the BUYBACK with CASHFLOW variable should be expected. This variable is calculated as following:

From collected sample of the observations of total assets and earnings before interest and tax including depreciation following descriptive statistics of CASHFLOW variable arised:

VARIABLE: CASHFLOW
Average	Standart Deviation	Minimum	Maximum	Range
0.12	0.20	-1.45	0.42	1.87

5. Takeover deterrence hypothesis.

There are few studies which specifies the «buyback» as the defensive strategy against acquisition, since more share company repurchase harder it gets for acquirer to takeover the company Bagwell (1991). In order to measure the effect of the takeover on the decision to repurchase for the firm dummy variable TAKEOVER was included, which took value of «1» if one and half year before the «buyback» there were public announcements or rumors of raiders which inform about willingness to acquire the «buy-backing» firm. And this variable took «0» value if there were no such information. Going back to my sample good example is Dorogobuzh JSC, which made repurchase of 27 millions of their own stocks in 2013 and Akron Group announced their willingness to buy the control package of shares in Dorogobuzh JSC in the same year. In this case TAKEOVER variable took value of «1».

6. Control variables.

Talking about control variables I chose the ratio which reflect structure of liabilities of the company. The ratio which was chosen is D/E ratio which is commonly specified as leverage ratio. LEVER variable is calculated as total debt over equity ratio and can be rewritten in mathematical terms as:

and has following descriptive statistics:

VARIABLE: LEVER
Average	Standard Deviation	Minimum	Maximum	Range
95.24	206.62	-252.14	1914.99	2167.13

Finally the regression equation can be rewritten as follows:



After the regression analysis the regression should be tested on multicollinearity in order to specify if it can be trusted from the statistical perspective:

The matrix of correlations among variables will be constructed and tested with critical value of 35%.

buyback share price fictitious

2.3 Data for the research

The collected data contains information about all «buy-backs» in Russia in 2001-2016 years. Since Russian financial sphere started developing process not a lot of years ago the data on such operations was very limited. Overall 166 announcements of repurchase of Russian companies were collected for the sample for analysis. Since my first-order task is to examine «buyback» on Russian financial markets Russian firms trading on foreign financial platforms were excluded from the sample. Moreover, such companies as Uralkali made announcements about repurchase more than 3 times a year, so repeated announcements firms were also excluded from the sample if it was less than 3 months between dates when information was available. Unfortunately, the financial data as Price over earnings ratio or net earnings before interest and tax including depreciation (EBITDA) of many companies was very limited or even absent, so I ended up with 45 effective dated of announcements.

As I mentioned in review of literature section there are several ways to perform repurchase of the stocks and different ways have it own specification and effects. In fact the operations on open market are the most commonly used way to make «buyback», so to make the research more one of the kind information only about open market repurchase were chosen for the sample.

All information about prices of the stocks, micex index, some financial data as total assets, net income, total debt, common equity, EBITDA, P/E ratio, cash dividend paid and information about «buy-backs» and dates of it were found with the help of Bloomberg informational system and Thomson Reuters Datastream service.



Chapter 3. Empirical Conclusions

In this part of the study empirical conclusions of the research will be shown. In this work two methods of analyse were conducted. First of all, event study was performed to see if CAR connected with «buyback» of shares is significantly positive. Afterwards, the regression analysis was used to outline determinants of motivation for managers to participate in «buyback» on Russian financial markets.

3.1 Event study conclusion

In order to perform event study, final sample of effective dates of announcement contained 97 observations. Afresh, tow event windows were taken to study short-run effects of the announcements of repurchase of shares of the company:

· Three days before the announcements and three days afterwards: (-3 +3)

· Three days before the announcement and forty days after the announcement (-3 +40 )

The simple t-test of significance was performed to study the first hypothesis stated in the methodology section. To begin with, analysis of the will be discussed.

Average cumulative abnormal return (-3 +3).

Source: Graph by the author.

From the graph it can be seen that evidently average cumulative abnormal return is positive three days before the announcement and considering the next three days it also stays positive. Due to the graphical analysis there is clear presence of insider information since CAR is becoming positive even before the announcement. It can be also seen on company example, in this case Oil Company Lukoil who published their «buyback» announcement at 16^th of April 2003:

COMPANY: LUKIOL
Days	Stock Price	Return	AR	CAR
-3	451.7	4.1%	3.1%	3.1%
-2	470	1.6%	0.6%	3.6%
-1	477.5	0.3%	0.8%	4.4%
0	479.1	0.9%	-0.2%	4.3%
1	483.3	1.8%	-0.4%	3.9%
2	491.9	3.9%	2.1%	6%
3	510.9	1%	0.8%	6.8%

Calculated Indicators for Lukoil JSC in (-3 +3) event window.

Source: Calculated by the author

On such example significant temp of growth of the price can be seen, which is consistent with the hypothesis that repurchase of share announcement tends to increase the stock price of the company.

Moving on to the statistical analysis sample of 97 was collected and descriptive statistics for that sample is as follows:

CAR(-3 +3)
Average	Standard Deviation	Min	Max	Range
0.01	0.05	-12.1%	15.0%	27.1%

The next step was to calculate t-statistics and compare it with critical values from students distribution. To test the stated hypothesis three confidence levels were taken : 10%; 5%; 1%.

The following values were taken from the table : 1.2; 1.67; 2.4 and t-statistics value calculated is 2, continuing the study and comparing it to critical values the null hypothesis that equals to zero is rejected at 10%; 5% level of confidence. Given that the results can be interpreted that there is statistical evidence that is significantly positive given event as announcement of «buyback» by the company. Speaking in general terms, the results of the event study is consistent with the first hypothesis stated in Methodology section, which basically means that the operation of the repurchase of the shares increases the price of the stock and stock outperform the market three days before and three days after the announcement.

The next step is to concentrate on the testing of the hypothesis based on sample.

Average cumulative abnormal return (-3 +40)

Source: Graph by the author.

From the graph the effects is rather ambigious but the more precise results may be obtained using statistical analysis. The same sample size was collected containing 97 observations of each company . From constructed sample descriptive statistics was achieved :



CAR(-3 +40)
Average	Standard Deviation	Min	Max	Range
0.24%	0.01	-0.94%	12.796%	13.74%

On the basis of study the same critical values were taken in order to perform test. Moreover, t-statics observed from the sample is as follows: 1.7 . Carrying on the study and comparing t -statistics with critical values the conclusion that there is strong statistical evidence that is significantly positive can be made. This is due to the fact that null hypothesis is rejected at two levels of confidence : 5% and 10%. Interpreting the result, one can have enough arguments to state that «buyback» of the shares positively influence the stock performance, in other words there is evidence that share prices appears to increase in its price and gain abnormal return due to the repurchase of the shares even at plus forty days basis. Moreover, the interpreted result is consistent with the first hypothesis stated in given research.

3.2 Regression analysis conclusions

After conducting event study methodology, the regression analysis was made in order to evaluate main reasons for managers to participate in «buyback» of shares and test stated hypothesis. After working on the model, collecting data for variables and finally running regression using special satistical software eViews 8, following regression results were obtained:

Variable	Coefficient	Std. Error	z-Statistic	Prob.
CASHFLOW	2.3522	1.1834	1.9875	0.0469
DIV_PAYOUT	0.4945	0.4221	1.1715	0.2414
LEVER	-0.0043	0.0027	-1.5549	0.12
LNPE	0.3249	0.2811	1.1556	0.2478
LNREV	0.0091	0.1021	0.0898	0.9284
TAKEOVER	3.3759	1.3356	2.5275	0.0115
C	-1.8352	2.0429	-0.8983	0.369
	16%

What is more, according to logit estimation model, marginal effects of variables should be calculated. The following table of marginal effects was calculated using additional special statistics software STATA:

logit: Changes in Probabilities for BUYBACK
Variable	MargEfct
LNPE	7.2%
LEVER	-0.1%
LNREV	0.2%
TAKEOVER	75.0%
DIV_PAYOUT	11.0%
CASHFLOW	52.2%

Testing multicollinearity of the model the following correlations of variables was constructed with the help of eViews8:

	Buyback	Cashflow	Div_Payout	Leverage	Lnpe	Lnrev	Takeover
BUYBACK	1	0.105589	0.142215	-0.141974	0.123224	0.025816	0.313152
CASHFLOW	0.105589	1	0.110702	-0.040586	-0.049694	0.278079	-0.202869
DIV_PAYOUT	0.142215	0.110702	1	0.032141	0.120567	0.092147	0.079486
LEVERAGE	-0.141974	-0.040586	0.032141	1	-0.137023	0.01123	-0.072615
LNPE	0.123224	-0.049694	0.120567	-0.137023	1	-0.031288	-0.006916
LNREV	0.025816	0.278079	0.092147	0.01123	-0.031288	1	-0.147505
TAKEOVER	0.313152	-0.202869	0.079486	-0.072615	-0.006916	-0.147505	1

From the table one can see that the highest correlation between the coefficients is 27% between CASHFLOW and LNREV, but this is lover than the critical 30%, so problem of multicollinearity was not evident in the constructed model.

Variable CASHFLOW has positive coefficient as expected, which assumes that more excess capital the firm has on its balance sheet the more motivated managers are to participate in «buyback» operations. Moreover, the coefficient is significant at 5% and 10 %level of confidence, which in consistent with the 5^th hypothesis stated in the study that excess capital is the reason for the repurchase. Interpreting the result, the conclusion is that in Russia firms appeared to buy their own shares back if the have excess free funds in order to redistribute them and minimize agency costs. This result is also consistent with findings of Dittmar (2000). Finally, looking at marginal effects there is statistical evidence that a unit increase in CASHFLOW variable increases the probability of «buyback» by about 52%.

Positive and significant coefficient at dummy TAKEOVER is consistent with the expectations of the study. The coefficient is significant at 5% and 10% level of significant. This outcome is consistent with the last hypothesis in given research and with the results obtained by Bagwell (1991). The facts above mean that if the company is at the danger of the acquisition this can play as the reason for company to decide to repurchase their own stocks. Moreover, due to the marginal effects analysis if the raider announced its willingness to acquire the company or there were rumours about it, the probability of «buyback» by the target increases by 75 percent.

Coefficients at LNPE, LNREV and DIV_PAYOUT are insignificant which rejects 2^nd 3^rd and 4^th hypothesis of this study. Such outcomes can be interpreted as Russian financial markets have its own specific which explain that firms are not motivated to buy their own shares back if their company is undervalued by the market in order to signal for investors. Moreover, Companies does not use repurchase of stocks as alternative for dividends in Russia. This can be consistent with the fact that the tax rate in Russia on dividends paid appeared to be lower than on capital gain in recent years according to the information collected. Finally, explaining insignificant coefficient at LNREV asymmetry of information can not be taken as the reason for «buyback» for Russian companies.

...