Writers: Kristoffer Algerstam and Nils Charbonnel
Program: Bachelor program in Business studies
Företagsekonomiska institutionen
Uppsala University
Springterm 2020
Thesis Advisor: Adri De Ridder
Share repurchases and
abnormal returns
Sammanfattning
Studien har undersökt avvikelseavkastningar för aktier under perioden de haft ett aktivt återköpsprogram. Studien har även undersökt ifall intensiteten av aktieåterköp har påverkat företagens avvikelseavkastning. Studiens observationsperiod har varit mellan 2010 och 2019. Under observationsperioden har vi genom Jensen’s Alpha approach utfört en tvärsnittsstudie för att analysera perioderna som företagen har haft aktiva återköpsprogram. Resultaten visar att företag har haft positivt avvikande avkastningar mellan 1,8% till 6% under perioden de genomgått ett återköpsprogram. Resultaten angående intensiteten av återköp visar att den större delen av återköp har skett under den andra halvan av återköpsprogrammen, dock finner vi ingen statistiksignifikans att detta har påverkat avvikande avkastningarna under denna period.
Abstract
In this paper we examine abnormal returns during active repurchasing programs and if the intensity of repurchasing programs impacts the returns. Through the Jensen’s Alpha approach our findings show us that positive abnormal returns are experienced by repurchasing firms under our study period that ranges from 2010 to 2019. The results show us that during active repurchasing programs companies have showed positive average annual abnormal returns ranging from 1,8% to 6%. We also find that the intensity of share repurchases does not have a statistically significant effect on the given abnormal returns. However, our results indicate that the abnormal returns are higher when the repurchases occurred, rather than when they are authorized.
Table of contents
1 Introduction ... 2
1.1 Problematization ... 3
2 Theoretical review ... 4
2.1 Literature review... 4
2.2 Liquidity-driven price impact ... 5
2.3 Legal framework... 5
2.4 Hypothesis testing ... 6
2.4.1 Hypothesis 1... 6
2.4.2 Hypothesis 2... 6
3 Method ... 7
3.1 Sample and data collection ... 7
3.2 Model specification ... 8
3.2.1 Measurement of abnormal share-price performance ... 8
3.2.2 Calendar-time portfolio ... 9 3.2.3 Jensen’s alpha ... 10 3.4 Regression analysis ... 11 4 Results ... 14 4.1 Abnormal returns ... 14 4.2 Repurchase activity... 19
5 Analysis and Conclusions ... 22
1 Introduction
Since the 10th of March 2000, companies listed on the Swedish Stock Exchange have had the
right to buy their own shares. When a company repurchases its own shares, the bought shares contributes to the companies’ treasury stock. Fama and French (2001) define treasury stock as the cumulative stock in possession of the issuing company. While the company is holding treasury stock, the stock does not count as shares outstanding. The treasury can either be sold in the open market, used to employee stock ownership plans, continue to be held as treasury stock or be retired. Retirement of securities refers to the elimination of issuing companies’ treasury stock. Retirement of shares is irrevocable, meaning that the total number of shares diminishes.
Comment and Jarrell (1991) point out that positive abnormal returns have been observed after firms have announced that they are going to initiate repurchasing programs. Almeida, Fos and Kronlund (2015) highlight the fact that firms tend to repurchase their own shares when their stock price is undervalued. Almeida et al. (2015) also mention that repurchases are not only a tool of market signalling but can be used to drive financial measurement and ratios up. This can for instance, be done through increasing earnings-per-share (EPS) by reducing the number of outstanding shares. Almeida et al. (2015) mention that key ratios such as the EPS are an important tool for executives and that their compensation sometimes can depend on EPS or similar key performance indicators. As highlighted by Bens et al (2003), the growth of share repurchases is faster than the growth of dividends.
Almeida et al. (2015) further discusses that the real effects of share repurchases can be destructive to the firm due to the alternative use of the capital being spent. Such capital could otherwise be used to purchase new assets or be set aside towards marketing campaigns. Therefore, there seems to be a discrepancy between the motives behind repurchases. Jiang et al. (2013) also provides insight on how the repurchases can be a substitute to dividends. In their findings they state that dividends are negatively related to buyback premiums. This strengthens previously mentioned theories mentioning that companies tend to buyback their own stock when their price seems undervalued.
Previous studies concerning share repurchases primarily focused on the 2000-2010 period. These studies were concentrated on the market short-term reaction to the announcement of buyback programs. By investigating data from 2010 to 2019, while focusing on long term
abnormal returns we wish to contribute to the literature with new insights about the Swedish market reaction to repurchase programs.
By using the calendar-time portfolio approach, we measure abnormal returns for companies during share repurchase programs. With the Fama and French (1993) three-factor model and the Carhart (1997) four-factor regression, we observed positive average annual abnormal returns of 2-6%. When we examine long-run abnormal returns conditional to the intensity of the share repurchase program, we find no significant correlation between the value of repurchases and abnormal returns.
1.1 Problematization
Previous studies have primarily focused on the short-term effects before and following repurchases. These studies have primarily been concentrated on the short-term effects rather than the long-term effects in order to try to filter out other factors that potentially could impact the investigating variable in the long run. When investigating the long-term effect of repurchased shares, factors such as economic cycles and periods of political distress are difficult to account for to achieve a result focused on the investigating variable.
For instance, if abnormal returns have consistently been found subsequent to repurchase events, a simple hypothesis could be that the repurchases provide abnormal returns for various reasons. However, if abnormal returns are found in the long-run after repurchases have occurred the hypothesis that specifically the repurchases are the origin of the abnormal return is not as clear. The reasoning for originates in the number of factors that are investigated when choosing timeframe. In the short-term merely focusing on the repurchase event is logical, if there are no other events occurring during that time. In the long-term changes in fluctuations of sales, economic cycles or similar phenomenon might play a larger role. Therefore, we find it interesting to investigate long-term effects of share repurchases, but risk adjusted by calendar-time portfolios. By using calendar-time portfolios we look at
consecutive results sequent to repurchases between longer timeframes, rather than picking a specific start period and ending period for the observation.
2 Theoretical review
2.1 Literature review
The announcement of open market repurchase programs have extensively been documented as positive news by the stock market in the literature. For instance, De Ridder and Råsbrant (2014) findings suggest that companies performing share repurchases have experienced short- term abnormal returns of 1,94%.
In Ikenberry et al. (1995, 2000) various studies they have examined how the announcement of repurchases impacted the price of the stock, the long-run firm performance following open market repurchases and how companies use their repurchase authorization as a tool to
increase performance. Their results show that positive abnormal returns after repurchases can be found in the shorter timeframe, but also in different markets. Comment and Jarrel (1991) who examined the signalling theory in combination with repurchases, suggesting that the firms present, and use repurchase programs as a communication tool to the market when they deem themselves undervalued. Their results also showed abnormal returns following
repurchases, they concluded that their results were in line with the signalling theory, and the liquidity driven impact caused by increased demand.
Other studies regarding share repurchases have been focusing on the motives behind the repurchasing decision such as Almeida et al. (2015) who suggests that firms tend to repurchase shares to increase financial ratios depending on outstanding stock, such as earrings-per-share or similar metrics to beat estimated earnings set by analysts. However, the primary theory in literature to explain the positive effect of repurchase
announcements is the signalling theory. The signalling theory is widely acknowledged, many studies, such as Stephens and Weisbach (1998), reveal abnormal returns in link with share repurchase announcements. Comment and Jarrell (1991) explain that financial decisions convey information about the firm’s value. The signalling hypothesis states that the decision to repurchase shares indicates a current undervaluation of the share as earlier mentioned. Although Ikenberry and Vermaelen (1996) point out that open market repurchase programs are not hard commitments by firms to purchase shares, but solely gives the firm the choice to repurchase shares. However, as the authorization to repurchase shares is not a hard
commitment to actually repurchase shares the signalling theory could as well be an
inadequate hypothesis to explain the positive abnormal returns resulting from open market repurchase program announcements.
2.2 Liquidity-driven price impact
McNally et al. (2006) findings suggest that share repurchases itself provides price support. Which could be another plausible explanation for abnormal returns originating from repurchasing decisions. Meaning that the repurchases themselves provide liquidity, thus resulting in price support. Råsbrant (2013) refers to this as the liquidity driven price impact hypothesis. Which could be one plausible explanation for these positive abnormal returns after repurchasing decisions but could also be caused by chance and might be sample specific as argued by Kothari and Warner (1997) and Fama (1998).
Yook (2010) points out that companies that announced share repurchases had positive stock performance even when they did not repurchase any shares during the program. This contradicts the liquidity driven price impact hypothesis, as it is empirical evidence that abnormal returns will occur even when the firm does not influence the demand.
However, the primary weakness to the liquidity-driven price impact or “price support” is discussed by Ikenberry and Vermaelen (1996). As mentioned earlier, they point out that the repurchase announcement is merely an announcement of authorization to repurchase shares, and not to fully commit to repurchase shares suggesting that examinations following
announcement could provide inadequate information to create a direct connection to abnormal returns.
2.3 Legal framework
In order to better comprehend the mechanics behind share repurchases it is necessary to understand the regulations surrounding it. In Sweden, share repurchases are regulated by the Swedish Companies Act. One of the main regulations states that a firm can only hold a maximum of 10% of outstanding shares as treasury stock at any given time. Repurchases must be validated by at least two third of the shareholders and the authorization is only valid until the next annual general meeting. This means that buy back periods cannot last longer than the time span between two annual general meetings. Companies are also limited in the number of shares that they are allowed to buy each day. The daily purchase cannot be greater than 25% of the daily turnover of the share that day.
2.4 Hypothesis testing
2.4.1 Hypothesis 1If previous studies and theories concerning share repurchases holds, firms who announce repurchasing programs should show abnormal returns in the shorter timeframe. To investigate if repurchasing firms actually show abnormal returns in connection to repurchases and not only in connection to the announcement of authorization to pursue, our first hypothesis is stated as followed:
Firms show positive abnormal returns during active repurchase programs.
𝐻 : 𝐴𝑅 ≤0
𝐻 : 𝐴𝑅 >0
Where 𝐴𝑅 is the twelve months average abnormal return as measured by Jensen’s alpha.
2.4.2 Hypothesis 2
As previous studies also indicated, the higher demand caused by the share repurchase program can lead to a short-term price increase. This phenomenon is referred to as price support and suggests that the additional demand coming from repurchases moves the supply and demand equilibrium which in turn leads to higher prices.
To further deepen our investigation on abnormal returns and share repurchases we have chosen to examine the intensity of repurchase programs and whether it have an impact on abnormal returns. To investigate the price support phenomenon our second hypothesis is formulated as followed:
Positive abnormal returns are driven by firms’ provided liquidity which results in increased price support.
3 Method
3.1 Sample and data collection
In our analysis we focus on companies forming the Stockholm Large Cap Index and have initiated share repurchasing programs between the annual general meetings of 2010 and 2019. Out of the sample firms’ financial companies and institutions have been excluded, due to their financial structural and regulation difference to non-financial companies. Companies that continuously have performed repurchases during the entire period are also excluded from the sample. Companies that have been under active repurchase programs during the entire period were excluded, in order to even out the distribution of data and to create as few outliers as possible. All data used in this study derives from the financial database Thomson Reuters Eikon and individual annual reports from firms’ in the sample.
The dependant variables that we used are the ratio of shares to treasury shares, the yearly fraction of repurchased shares, the market-to-book ratio, the market capitalisation and the average value of shares repurchased per day. Some of these dependant variables, such as the market-to-book ratio and the market capitalisation, were available on the Thomson Reuters Eikon platform. The others had to be calculated independently. To calculate the ratio of shares to treasury shares we had to find the number of outstanding at the end of each month as well as the quantity of treasury shares. In order to find the fraction of repurchased shares and the average value of daily shares repurchases we had to gather data concerning each share repurchase transactions, including transaction dates and the monetary value of every
transaction. This data gathering was the most time intensive as there was no platform that had accurate information concerning share repurchase transactions. We had to look at individual companies press releases and annual reports in order to find accurate information.
The data comprising our benchmark used to calculate our buy-and-hold abnormal returns consists of all publicly traded firms on the Stockholm Stock Exchange. These firms were collected using the Thomson Reuters Eikon code LSWSEALI. We had to screen out all the companies that had incomplete information on Thomson Reuters Eikon. Our study was confined to the ten-year period ranging from 2010 to 2019. This time period is interesting to study as it was a period of strong economic growth. The choice to study this time period was motivated by the lack of studies focusing on the repurchases during this decade, in addition to the fact that the sample of firms repurchasing shares was large enough.
Furthermore, the period has been a time of prosperity on the stock markets. This leads to companies having more freedom to undertake share repurchase programs. The previous ten-year period was the first-time companies had the opportunity to repurchase companies, this meant that the majority of companies were performing repurchases simultaneously. Many of the repurchase programs were primarily motivated by the fact that it was the first-time companies were allowed to do so. The number of repurchases in the period we studied was considerably lower but companies that performed repurchases had higher incentives to do so. The fact that these companies chose to repurchase shares instead of increasing dividends show that they find value in share repurchase programs. Therefore, we think that it is
interesting to study how the abnormal returns linked to these repurchases as they are signs of how the market perceives these repurchases.
3.2 Model specification
The independent variable in this study is the abnormal returns linked to share repurchase programs. We decided to utilise different tools to measure abnormal returns, the buy and hold abnormal returns and the Jensen’s Alpha method.
3.2.1 Measurement of abnormal share-price performance
In order to perform our analysis, we calculated the buy and hold abnormal returns (BHAR) for each company against the benchmark for both 6 months and 12 months after the annual general meeting whereas the repurchase program has been enacted. We used the following method to calculate the buy and hold abnormal return;
𝐵𝐻𝐴𝑅, = [ (1 + 𝑟, ) − 1] − [ (1 + 𝑟 , ) − 1]
Where BHAR is the buy and hold abnormal returns for the company i during the period t. 𝑟,
is the monthly raw return and 𝑟 , represent the return of the benchmark during the
period t. The benchmark consists of the equally weighted index of all the companies listed on the Stockholm Stock Exchange. Hence, a positive (negative) BHAR demonstrates that firms undergoing repurchase programs outperforms (underperforms) the benchmark. For the purpose of our regression, we calculated the mean buy-hand-hold abnormal return with the following formula:
𝐵𝐻𝐴𝑅 =1
𝑛 𝐵𝐻𝐴𝑅,
In order to be able to verify the liquidity driven price impact hypothesis we measured the fraction of repurchased shares for each firm. The repurchase fraction was calculated by dividing the total monetary value of repurchased shares, to their market capitalizations from the end of the year prior to the repurchasing program. Using these findings, we were able to divide the repurchasing firms in two groups. The first group represent firms that primarily repurchased shares during the first half of the program, and the second group represent firms that primarily repurchased shares in the second half of the program.
3.2.2 Calendar-time portfolio
We used the calendar-time portfolio method following Axelsson and Brissman (2011) methodology. This method is advocated by both Fama (1998) and Mitchell and Stafford (2000) arguing that it is better suited to measure long term abnormal returns than the buy-and-hold abnormal return method. The buy-and-buy-and-hold abnormal return method alternative to the calendar-time portfolio shows weakness due to its pseudo timing aspects. According to Schulz (2003) BHAR method automatically generates underperformance following clustering of issues experiencing common event. De Ridder (2008) argues that the BHAR method generates more extreme returns because of the compounding effect. Hence, the calendar-time approach was chosen for the study.
The calendar-time portfolio method is to construct and examine portfolios for companies that experience a similar type of event during the same time period. The portfolios are composed of each company currently performing a share repurchase program. We defined a share repurchase program as the period between two annual general meetings where a company performed at least one share repurchase. We did not include companies that announced share repurchases programs but did not perform any buyback during the period.
Each firm enter the portfolio the day following the annual general and remains for twelve consecutive months. If a company begins a new repurchase program within the first twelve months, it stays within the portfolio for another twelve months. This way we constructed portfolios from April 2010 to Mars 2019. The number of firms in each portfolio varies with each calendar month, therefore, has the portfolios been rebalanced for each period. We constructed both an equally weighted and a value weighted portfolio for each calendar month.
The value weighted portfolios are weighted by market capitalization on the first day of that month. We used periods of 21 trading days in order to measure months, this way every month is of the same length throughout the year.
3.2.3 Jensen’s alpha
For each portfolio, raw returns are calculated for each event month. Using the Jensen’s alpha approach, we made a time series regression of our calendar time portfolios to find average abnormal returns for each event year studied. We made our first regressions using Fama and French (1993) three factor model (1) and completed them using Carhart’s (1997) momentum factor (2).
𝑟 , − 𝑟𝑓 = 𝛼 + 𝛽 (𝑟𝑚 − 𝑟𝑓 ) + 𝛽 (𝑆𝑀𝐵 ) + 𝛽 (𝐻𝑀𝐿 ) + 𝜀 , (1)
𝑟 , − 𝑟𝑓 = 𝛼 + 𝛽 (𝑟𝑚 − 𝑟𝑓 ) + 𝛽 (𝑆𝑀𝐵 ) + 𝛽 (𝐻𝑀𝐿 ) + 𝛽 (𝑊𝑀𝐿 ) + 𝜀 , (2)
Where 𝑟 , is the raw returns for a portfolio p during month t. 𝑟𝑓 is the one month T-bill for
month t,𝑟 , − 𝑟𝑓 is the excess return for a portfolio p during month t, 𝑟𝑚 − 𝑟𝑓 is the
market premium for the event month t, 𝑆𝑀𝐵 is the difference between the returns of value-weighted portfolios of large and small firms during month t and 𝐻𝑀𝐿 is the difference in returns of value-weighted portfolios of high and low book-to-market firms during month t. The WML factor is defined by the return on high momentum stocks minus the return on low momentum stocks.
The intercept, 𝛼 , indicate the average abnormal return for each month during the sample period. This result will indicate whether the portfolios outperformed the market during the repurchase programs. We performed these regressions for both our equally weighted portfolios and the value weighted. We looked at three different time periods in our analysis, first we focused on the entire period of the programs. Thereafter we divided our findings into the first and the second half of the repurchase programs, 1 to 6 months and 7 to 12 months respectively. The aim was to find whether the abnormal return differed during the two periods, as we have observed that repurchases occur primarily during the second half of the programs.
3.4 Regression analysis
In order to test our second hypothesis, we need to analyse the relationship between the abnormal returns and the different firms repurchase patterns. In order to study this
relationship, we performed a fixed effect regression analysis using buy and hold abnormal returns as the dependant variable. In our analysis, we fully exploit all daily repurchase trades, during each program for our sample firms and our sample period.
𝐵𝐻𝐴𝑅, = 𝛼 + 𝛽 𝑇𝑅𝐸𝑂𝑈𝑇, + 𝛽 𝑅𝐸𝑃𝐹𝑅𝐴𝐶, + 𝛽 𝐴𝐶𝑇𝐼𝑉𝐼𝑇𝑌, + 𝛽
𝑀
𝐵 , + 𝛽 𝑆𝐼𝑍𝐸, + 𝜀,
Where 𝐵𝐻𝐴𝑅, is the Buy-and-Hold abnormal return for company i during year t. 𝑇𝑅𝐸𝑂𝑈𝑇,
is the fraction of treasury shares to outstanding shares held by company i at the end of year t. 𝑅𝐸𝑃𝐹𝑅𝐴𝐶, is the fraction of the value of repurchased shares during year t to the market cap
of company i at the end of year t. 𝐴𝐶𝑇𝐼𝑉𝐼𝑇𝑌, is the average value of daily repurchases for
company i during year t.
, is the market to book ratio for company i at the end of year t.
Table 6: Fixed effects regression on buy-and-hold abnormal returns (BHAR) during repurchase programs Alpha TREOUT REPFRAC ACTIVITY M/B SIZE Companies that
primarily
repurchased shares during the first half of the program BHAR 6 months -0,62* (-1,90) -0,75 (-0,27) 11,7 (1,37) -0,00 (-0,33) -0,02 (-1,50) 0,07* (2,00) BHAR 12 months -0,58 (-1,20) -1,93 (-0,48) 11,83 (0,94) 0,00 (0,15) -0,02 (-0,93) 0,07 (1,35) Companies that primarily repurchased shares during the second half of the program
BHAR 6 months 0,36 (1,15) -2,23 (-0,87) 1,64 (0,81) 0,00 (0,80) 0,01 (0,67) -0,04 (-1,34) BHAR 12 months 0,01 (0,03) 3,72 (1,46) 1,03 (0,51) 0,00 (0,33) 0,03 (1,34) -0,02 (-0,51) All BHAR 6 months 0,01 (0,03) -1,36 (-0,73) 2,05 (1,27) 0,00 (0,93) -0,01 (0,95) -0,00 (-0,03) BHAR 12 months -0,08 (-0,30) 1,79 (0,80) 1,71 (0,88) 0,00 (0,32) 0,01 (0,41) 0,00 (0,15)
Notes: This table presents the results from the following firm fixed effect regression analysis; 𝐵𝐻𝐴𝑅, = 𝛼 + 𝛽 𝑇𝑅𝐸𝑂𝑈𝑇, + 𝛽 𝑅𝐸𝑃𝐹𝑅𝐴𝐶, + 𝛽 𝐴𝐶𝑇𝐼𝑉𝐼𝑇𝑌, + 𝛽
𝑀
𝐵 , + 𝛽 𝑆𝐼𝑍𝐸, + 𝜀,
Where 𝐵𝐻𝐴𝑅, is the Buy-and-Hold abnormal return for company i during year t. 𝑇𝑅𝐸𝑂𝑈𝑇, is the fraction of treasury
shares to outstanding shares held by company i at the end of year t. 𝑅𝐸𝑃𝐹𝑅𝐴𝐶, is the fraction of the value of repurchased
shares during year t to the market cap of company i at the end of year t. 𝐴𝐶𝑇𝐼𝑉𝐼𝑇𝑌, is the average value of daily
repurchases for company i during year t.
, is the market to book ratio for company i at the end of year t. 𝑆𝐼𝑍𝐸, is the
natural logarithm of market value for company i at the end of year t
In Table 6, we present the results from the fixed effect regression. We found that the coefficient of the variable for the repurchase size, REPFRAC, is positive, although not significant. This indicates that the more shares firms’ have repurchased relative to their market capitalization the higher BHAR they have experienced under the observation period. We observe that this coefficient is considerably higher for companies that repurchased most shares in the first half of the program. This indicates that firms repurchasing during the first half of the program are more sensible to the repurchase fraction. This could be an indicator that companies that give evidence of repurchases early on perform significantly better.
We observe that the coefficient for the ACTIVITY variable is near zero in every regression performed. Which could indicate that the volume of daily share repurchases has a very insignificant impact on the abnormal returns. These results are not in accordance with the liquidity driven price impact hypothesis, hypothesis 2, that suggest increased liquidity provides increased price support.
The other independent variables all show little to no significance. Treasury shares to
outstanding shares show no statistical significance in any of the instances but does however, presents a pattern. Every coefficient of the TREOUT presents a negative correlation with BHAR for the first 6 months of the repurchase program, which indicates that firms holding a high amount of treasury shares have weaker performance when performing share repurchases. Market-to-book also shows statistical insignificance for at all levels with coefficient being close to zero. This could indicate that market-to-book does not have any effect on buy-and-hold abnormal returns in our model. We capture the size effect at a 10% significance level for the SIZE effect for companies that primarily repurchased shares in the first half of the
program. This can be interpreted as signalling that companies that have had a greater market capitalization has performed better than those with lesser market capitalization in the first group. It is however important to consider that all companies observed are large capitalization companies.
4 Results
4.1 Abnormal returns
Table 1 present the descriptive statistics for monthly raw returns of the equally weighted portfolios. The portfolios were constructed with the methodology explained in section 4.2. We observe that the average raw return for repurchasing companies is positive on average, meaning that the repurchasing companies tend to gain value during the repurchase programs. Table 2 describes the statistics for the monthly raw returns of the value weighted portfolios. The monthly raw return is slightly higher for the equally weighted portfolio, this indicate that a larger repurchasing companies performed slightly worse than smaller firms.
In order to identify abnormal returns, we used Jensen’s alpha methodology. Which consists of finding abnormal returns by performing regressions of excess returns. We made a regression using the Fama and French (1993) three factor regression model, using the excess returns as the dependent variable. Table 3 presents the results of the regressions, for the 12 months of the programs. We also made a regression focused on the first half as well as the second half of the program, for both the equally weighted and the value weighted portfolios. In order to further deepen our analysis, we also regressed our data using the Carhart (1997) four factor regression model, using the same methodology.
Using the Fama and French (1993) three factor regression model, we found monthly average abnormal return of 0,22% (t=0,44) for the total sample period when regressing the value weighted portfolios. The equally weighted portfolios average abnormal return amounted to 0,39% (t=1,17). This results in a 2,64% yearly abnormal return for the value weighted portfolios, compared to 4,68% for the equally weighted.
When adding Carhart (1997) momentum factor to the analysis our results change quite significantly. The value weighted portfolios average abnormal returns dropped to 0,15% (t=0,29) per month, 1,8% annually. While the equally weighted portfolios average abnormal returns increased to 0,50% (t=1,39), corresponding to 6% yearly abnormal returns.
The regressions show evidence of abnormal returns in the long term. This is evidence that the companies that have performed share repurchase programs have outperformed the general market during the time the program was active. This confirms our first hypothesis.
Table 3 also presents higher abnormal returns for both the value and equally weighted portfolio during the second half of the program in comparison with the first. The equally weighted portfolio shows abnormal returns of -0,01% for the first half of the program and 0,33% for the second half of the program. The abnormal returns for the value weighted portfolio show similar results of 0,05% for the first half and 0,71% for the second half of the program. These results are in accordance with the data presented in Table 5, Panel B, that show us that the average size of repurchases amounted to 0,34% in the first half, and 1,34% in the second half of the repurchase program. These results could highlight the flaw in the signalling theory. As mentioned by Ikenberry and Vermaelen (1996) share repurchase announcements are not firm commitments to repurchase shares, but merely the authorization to repurchase.
Table 1: Descriptive Statistics: Equally weighted, calendar time portfolio raw returns.
Year Mean Median STDEV Q1 Q3 Min Max
2010 1,27% 2,54% 4,82% -1,74% 4,75% -10,14% 7,25% 2011 -0,37% -1,87% 5,49% -5,37% 4,28% -6,36% 9,37% 2012 1,21% 1,84% 3,64% -2,44% 4,51% -4,08% 5,81% 2013 2,12% 2,17% 3,38% -1,36% 4,83% -2,78% 8,28% 2014 1,98% 1,57% 3,43% -0,34% 4,54% -4,60% 7,37% 2015 -0,30% -0,44% 3,83% -4,02% 3,56% -5,97% 5,25% 2016 0,85% 1,01% 1,57% -0,56% 2,20% -1,67% 3,30% 2017 -0,51% -1,70% 2,89% -2,41% 0,83% -3,75% 5,72% 2018 0,40% 2,05% 3,73% -2,40% 3,14% -7,88% 4,72% All 0,74% 0,84% 3,77% -2,04% 3,56% -10,14% 9,37
Notes: This table presents the descriptive statistics for the monthly raw returns of the equally weighted portfolios.
Table 2: Descriptive Statistics: Value weighted, calendar time portfolio raw returns.
Year Mean Median STDEV Q1 Q3 Min Max
2010 1,30% 1,73% 4,93% -3,13% 6,49% -7,60% 7,45% 2011 0,592% 0,00% 5,15% -3,40% 5,01% -8,65% 10,25% 2012 -0,03% 1,03% 5,66% -1,29% 3,76% -15,12% 6,44% 2013 0,78% 0,61% 3,61% -1,69% 4,03% -5,46% 6,39% 2014 1,88% 1,07% 4,42% -1,83% 6,38% -3,63% 9,10% 2015 -1,21% -2,26% 4,13% -4,42% 3,39% -7,06% 5,14% 2016 -0,28% -0,06% 2,89% -2,19% 2,02% -6,04% 4,44% 2017 1,25% -0,46% 6,15% -1,95% 3,75% -8,15% 17,44% 2018 1,71% 1,77% 10,12% -4,34% 8,86% -18,87% 18,20% All 0,67% 0,14% 5,46% -2,27% 4,09% -18,87% 18,20%
Table 3: Long-run performance of shares preforming share repurchase programs as estimated by the Fama and French three factor model using the calendar time portfolio approach.
Holding period months 1-12 Holding period months 1-6 Holding period months 7-12
Coefficient
t-stat
Coefficient
t-stat
Coefficient
t-stat
Panel A: Value Weighted
Intercept (αp) 0,22 0,44 -0,01 -0,02 0,33 0,45
𝑟𝑚 −𝑟𝑓 0,52*** 4,43 0,41** 2,38 0,65*** 3,82
𝑆𝑀𝐵𝑡 0,06 0,20 -0,13 -0,28 0,16 0,37
𝐻𝑀𝐿𝑡 -0,54** -2,21 -0,57 -1,44 -0,48 -1,50
𝑅2 0,16 0,10 0,23
Panel B: Equally weighted
Intercept (αp) 0,39 1,17 0,05 0,10 0,71 1,50
𝑟𝑚 −𝑟𝑓 0,41*** 5,21 0,34*** 2,89 0,46*** 4,14
𝑆𝑀𝐵𝑡 0,22 1,08 -0,02 -0,06 0,45 1,57
𝐻𝑀𝐿𝑡 -0,28* -1,70 -0,22 -0,85 -0,35* -1,67
𝑅2 0,21 0,16 0,28
Notes: This table presents the regression results for the Fama and French (1992) three factor regression model: 𝑟, − 𝑟𝑓 = 𝛼 , + 𝛽 (𝑟𝑚 − 𝑟𝑓 ) + 𝛽 (𝑆𝑀𝐵 ) + 𝛽 (𝐻𝑀𝐿 ) + 𝜀 ,
Panel A presents the results for value weighted returns, while the panel B presents the equally weighted portfolio returns. The t-statistics are in parentheses. ***, **, * shows significance levels of 1 percent, 5 percent and 10 percent respectively.
Table 4: Long-run performance of shares preforming share repurchase programs as estimated by the Carhart four factor model using the calendar time portfolio approach.
Holding period months 1-12 Holding period months 1-6 Holding period months 7-12
Coefficient
t-stat
Coefficient
t-stat
Coefficient
t-stat
Panel A: Value Weighted
Intercept (αp) 0,15 0,29 -0,37 -0,47 0,40 0,53 𝑟𝑚 −𝑟𝑓 0,52 4,42 0,40** 2,29 0,63*** 3,50 𝑆𝑀𝐵𝑡 0,06 0,19 -0,13 -0,28 0,17 0,38 𝐻𝑀𝐿𝑡 -0,49 -1,85 -0,34 -0,76 -0,53 -1,48 W𝑀𝐿 0,08 0,37 0,37 1,19 -0,09 -0,31 𝑅2 0,16 0,13 0,23
Panel B: Equally weighted
Intercept (αp) 0,50 1,39 0,04 0,06 0,87* 1,76 𝑟𝑚 −𝑟𝑓 0,40*** 5,05 0,34 2,85 0,42*** 3,61 𝑆𝑀𝐵𝑡 0,23 1,09 -0,02 -0,06 0,46 1,61 𝐻𝑀𝐿𝑡 -0,34* -1,91 -0,22 -0,72 -0,46* -1,96 W𝑀𝐿 -0,12 -0,87 0,01 0,07 -0,20 -1,04 𝑅2 0,22 0,16 0,29
Notes: This table presents the regression results for the Carhart (1997) four factor regression model: 𝑟, − 𝑟𝑓 = 𝛼 , + 𝛽 (𝑟𝑚 − 𝑟𝑓 ) + 𝛽 (𝑆𝑀𝐵 ) + 𝛽 (𝐻𝑀𝐿 ) + 𝛽 (𝑊𝑀𝐿 ) + 𝜀 ,
Panel A presents the results for value weighted returns, while the panel B presents the equally weighted portfolio returns. The t-statistics are in parentheses. ***, **, * shows significance levels of 1 percent, 5 percent and 10 percent respectively.
4.2 Repurchase activity
Table 5 Panel A present statistics concerning the value of repurchase programs listed by the year during which the program was initiated. The full sample consists of 55 programs, valid for one year, amounting to a total repurchased value of 11,9 billion SEK. The average value of a program is 216 million SEK, which represents 0,85% of the repurchasing’s firms’ market capitalization. The mean BHAR for the 12 months period of the programs amounted to 2,28%.
We compared our results to the findings of previous studies also focusing on the Stockholm Stock Exchange. De Ridder (2008) found that during the period of March 2000 to 2004, the average share repurchase programs amounted to 4,85% of the outstanding shares of the repurchasing firms. The significantly lower fraction of repurchased shares could indicate that repurchasing companies have less incentives to repurchase large portions of their shares. This could also depend on the fact that the 2000 to 2004 period was the first period during which repurchases were authorized on the SSE.
Panel B present the descriptive statistics for the ratio of repurchased shares. We separated repurchasing firms into two groups, depending on which half of the repurchase program most repurchases occurred. We observe that the companies that primarily repurchased shares during the first half of the program, bought a smaller number of shares. The average value of repurchases amounting to 0,34% of total repurchases, compared to 1,34% for the second group.
Table 5 / Panel A: Share Repurchase Programs Statistics by Calendar Year
Year n Total Value of
Repurchase Programs in SEK Program Size (%) Mean BHAR (%) t-statistic 2010 7 660,935,037 0,39 9,91% -3,69 2011 6 1,574,947,870 0,31 3,40% -0,54 2012 5 1,304,229,412 1,17 9,57% -3,53 2013 4 565,398,359 1,36 -0,33% 1,26 2014 4 1,558,354,692 1,40 10,94% -4,19 2015 5 1,645,490,054 0,72 -13,44% 7,60 2016 6 1,623,704,408 0,61 6,43% -2,01 2017 7 595,931,973 0,92 -5,57% 3,80 2018 11 2,397,996,352 1,04 0,27% 0,97 All 55 11,926,998,157 0,85 2,28% 1,14
Notes: Column one shows the year during which the repurchase programs started. The second column represents the number of share repurchase programs active per year. The total value of repurchase programs presents the combined monetary value of all repurchases each year in SEK. The third column presents the percentage of the repurchased value to the market
capitalization of the repurchasing firm. The fourth column shows the average buy-and-hold abnormal return for repurchasing firms during a repurchase program. The buy-and-hold abnormal return was calculated by subtracting product of the
repurchasing company’s return to the Stockholm Stock Exchange index returns for the 12 months period of the program.
Table 5 / Panel B: Descriptive Statistics of Share Repurchase Fraction per Program
Mean Median STDEV Q1 Q3 Min Max n
Companies that primarily repurchased shares during the first half of the
program
0,34% 0,11% 0,53% 0,06% 0,28% 0,00% 2,35% 27
Companies that primarily repurchased shares during the second half of the program
1,34% 0,48% 1,87% 0,12% 2,36% 0,00% 5,78% 28
All 0,85% 0,26% 1,46% 0,09% 0,64% 0,00% 5,78% 55
Notes: This table presents the descriptive statistics for the size of the share repurchase programs. This was calculated by dividing the value of repurchased shares to the market capitalization of the repurchasing firm. Repurchasing firms were grouped in two groups, depending on which half of the program most repurchases occurred.
5 Analysis and Conclusions
In this study we investigate whether companies experience abnormal returns during an active repurchase program and if positive abnormal returns can be explained by the liquidity provided by firms repurchasing activities.
Our results confirm our first hypothesis that firms show positive abnormal returns during active repurchase programs. Through the Jensen’s Alpha approach, we find abnormal returns in all instances observed. In our Fama and French (1993) three factor model we find that the average annualized abnormal return for our value weighted portfolio amounting up to 2,64% and respectively 4,68% for the equally weighted portfolio. When we applied the Carhart (1997) four-factor model we found positive average annual abnormal returns of 1,8% for the value weighted portfolio and 6% for the equally weighted portfolio. The results differ from the three-factor model but are still in line with our first hypothesis that firms show abnormal returns during active repurchase programs.
In our findings we also detect that most of the abnormal return are experienced in the second half of the repurchasing program. This finding goes hand in hand with our other set of data that shows us that the value of repurchased shares was significantly higher in the second half of the program. These findings could potentially be interpreted as the flaw Ikenberry and Vermaelen (1996) highlighted on why the signalling theory may not be an adequate explanation for abnormal returns following repurchase announcements. Ikenberry and Vermaelen (1996) argued that the authorization for share repurchases merely gives the alternative to repurchase and are not a hard commitment to perform repurchases. However, our findings show us higher abnormal return when actual repurchases occur, rather than when the authorization is announced.
To test our second hypothesis, we investigated whether intensity of repurchases affects positive abnormal returns. To test this, we created a variable in which we summarized the value of shares repurchased and divided it by the number of days on which the shares were repurchased.
Our results regarding price support matches the theory and recent studies in the area that suggest that abnormal returns can be found from the additional demand and liquidity provided by repurchases. Although, our results indicate that the extra liquidity and demand coming from the price support phenomenon positively correlates with positive abnormal returns.
We cannot fully confirm that our second hypothesis is correct since our results are statistically insignificant, but only conclude that there may be a positive correlation between price support theory and positive abnormal returns. But even though our results are insignificant in a statistical way, they are in pair with similar studies and point towards the direction that price support and positive abnormal returns positively correlates. Therefore we conclude that our second hypothesis is not fully falsified, nor confirmed but suggests that further investigation could provide additional support to confirm the price support theory.
References
Almeida, H., Fos, V. & Kronlund, M. (2015) The real effects of share repurchases. Journal of Financial Economics, 119, p. 168-185.
Axelsson, L. & Brissman, P. (2011). Share repurchase announcements and abnormal returns for Swedish listed real estate companies. (master upp.) Kungliga Tekniska Högskolan.
https://www.diva-portal.org/smash/get/diva2:497867/FULLTEXT01.pdf
Bens, D., Nagar, V., Skinner, J. & Wong, F. (2003). Employee stock options, EPS dilution, and stock repurchases, Journal of Accounting and Economics, 36, p. 51-90.
Carhart, M. (1997). On persistence in mutual fund performance. Journal of Finance. 52, p. 57- 72.
Cheng, Y., Harford, J. & Zhang, T. (2015). Bonus Driven Repurchases. Journal of Financial and Quantitative Analysis. 50, 3, p. 447-475.
Comment, R. and Jarrell, G. (1991), The Relative Signalling Power of Dutch-auction and Fixed-Price Self-Tender Offers and Open-Market Share Repurchases, Journal of Finance, 46, 4, p. 1243-1271.
De Ridder, A. & Råsbrant, J. (2014). Share Repurchases: Does Frequency Matter? Studies in Economics and Finance, 31, p. 88-105.
De Ridder, A. (2008). Share Repurchases and Firm Behavior. International Journal of Theoretical and Applied Science, 12, 5, p. 605-631.
Fama, E. & French, K. (1993). Common Risk Factors in the Returns on Stocks and Bonds. Journal of Financial Economics, 33, p. 3-56.
Fama, E., (1998), Market efficiency, long-term returns, and behavioral finance, Journal of Financial Economics, 49, p. 283-306.
Fama, E. & French, K. (2001). Disappearing dividends: changing firm characteristics or lower propensity to pay? Journal of Financial Economics, 60, p. 3-43.
Ikenberry, D. & Vermaelen, T. (1996). The Option to Repurchase Stock. Financial Management, 25, 4, p. 9-24.
Ikenberry, D., Lakonishok, J. & Vermalaen, T. (1995). Market underreaction to open market share repurchases. Journal of Financial Economics, 39, p. 181-208
Ikenberry, D., Lakonishok, J. & Vermalaen, T. (2000). Stock Repurchases in Canada: Performance and Strategic Trading. The Journal of Finance, 55, p. 2373-2397.
Kothari, S.P. & Warner, J.B. (1997). Measuring long-horizon security price performance. Journal of Financial Economics, 43, p. 301-339.
McNally, W., Smith, B. & Barnes, T. (2006). The Price Impacts of Open Market Repurchase Trades. Journal of Business Finance & Accounting, 33, p. 735-752.
Råsbrant, J. (2013). The Price Impact of Open Market Share Repurchases. Working paper. Uppsala University. Downloaded from:
https://www.researchgate.net/publication/228315600
Schultz, P. (2003). Pseudo Market Timing and the Long-Run Underperformance of IPOs. The Journal of Finance, 58, p. 483-517.
Stephens, C. P., & Weisbach, M. S. (1998). Actual Share Reacquisitions in Open-Market Repurchase Programs. The Journal of Finance, 53, 1, p. 313-333.
Yook, K. C. (2010). Long-run stock performance following stock repurchases. The Quarterly Review of Economics and Finance, 50, 3, p. 323-331.
