6: When do companies repurchase shares? Evidence on long-run performance after repurchase announcements.
Academic studies traditionally only examined short-term returns because of a strong prior belief in semi-strong efficient markets. Questioning the efficient market hypothesis used to be as popular in academic circles as questioning the existence of God in bible class. Moreover, dealing with long-term returns raises a number of methodological and statistical issues. In particular, long-term abnormal returns are very sensitive to the model of "normal" returns. So the results can always be criticized because any test of market efficiency is a joint test of efficiency and a model of market equilibrium. There is always a possibility that the excess returns observed are explained by an omitted "repurchase risk factor", something which would comfort proponents of market efficiency.
6.1 US evidence
Lakonishok and Vermaelen  examine long term price behavior after 258 fixed price repurchase tender offers announced between 1962 to 1986. They find that buying shares 1 month after the expiration and holding them for 2 years generates excess returns of 8%. Excess returns are calculated by comparing the returns of the buyback company with the returns of a portfolio of control firms with similar size and beta. As they report a cumulative abnormal return from the month before the announcement until the month after expiration of approximately 14%, if the market had been efficient, stock prices should have risen to 22%, not 14%. Note that this level corresponds to the repurchase premium observed in the sample. So when companies are offering to repurchase their shares at a price, that price is a good estimate of the fair value of the stock. However, the market seems to under-react and is too skeptical about the often stated managerial belief that the shares are undervalued or a good investment. On closer inspection, the abnormal returns seem to be confined to small firms, something which makes sense as it is more likely that smaller firms could be undervalued. Peyer and Vermaelen  find qualitatively similar results using data from 1986 to 2001 and now using the three-factor Fama-French  model as a benchmark.
Similar evidence of underreaction is reported by Ikenberry, Lakonishok and Vermaelen  who examine long-term excess returns after 1289 open market repurchase programs announced between 1980 and 1990. They test a strategy of buying companies that announce open market share buyback authorizations, and holding the shares for 4 years (48 months). In order to compute abnormal returns, Ikenberry et al. employ 4 benchmarks: a value-weighted stock market index (i.e. the S&P 500), an equally weighted market index, a portfolio with the same size and a portfolio with the same size and book-to-market ratio. Depending on the method used to compute normal returns, long-term cumulative average abnormal returns vary between 6% and 12%. Further analysis shows that most of the excess returns are generated by value firms (low market-to-book), rather than growth (high market-to-book) firms. This is consistent with the hypothesis that when value firms announce a repurchase it is more likely that the motivation for the buyback is driven by a desire to take advantage of an undervalued stock price. Weston and Siu  point out that, the higher the market-to-book ratio, the higher the increase in book return on equity resulting from a repurchase. The fact that high-market-to book firms have an incentive to repurchase stock to manipulate return on equity, may also explain why these repurchases are not followed by long-term excess returns.
Mitchell and Stafford  criticize Ikenberry et al.  because all the events are equally weighted, instead of value weighted. They show the abnormal returns disappear when they value-weight the events. While value-weighting gives a better idea whether the anomaly is an economically important event, the method is biased against finding abnormal returns : small firms, which are, per definition, less followed by analysts, are more likely to be undervalued than large firms. If one would build a portfolio to exploit the anomaly, the weighting should be based on the inverse of market capitalization, the opposite of Mitchell and Stafford's suggestion. Fama  points out that, the weighting should be determined by the research questions we try to answer. As in corporate finance the relevant research question is "what can I expect when a company announces a repurchase" equally weighting seems appropriate. Mitchell and Stafford  also criticize the buy-and-hold abnormal return methodology of Ikenberry et al.  and argue that it is biased. Schwert  shows that even if anomalies existed in a period in which they are identified, the activities of practitioners who implement strategies to take advantage of anomalous behavior can cause anomalies to disappear. Consistent with this argument, he documents that several anomalies have disappeared after the papers that highlighted them were published. Perhaps these repurchase anomalies are no longer valid today?
Peyer and Vermaelen  address these concerns by re-examining the Ikenberry, Lakonishok and Vermaelen  anomaly on a fresh 11-year data-set, employing Ibbotsons' RATS methodology. Based on a sample of 3465 observations, they confirm the results by Ikenberry et al. . Moreover, they find that the abnormal price behavior of the stock in the 6 months prior to the repurchase announcement is a better predictor of future excess returns than book-to-market or size: the sample of the 20% most beaten down stocks outperform the Fama-French  benchmark by approximately 45% in the 4 years after the repurchase. This is shown in Figure 6.1 which shows the abnormal price behavior during the four years following the repurchase announcement for 5 quintiles formed on the basis of the abnormal price behavior during the 6 months preceding the announcement. Moreover, consistent with Ikenberry et al.  the cumulative excess returns only become really economically significant after 2 years.
[FIGURE 6.1 OMITTED]
Finally, market under-reaction is also observed in private repurchases . Moreover, the under-reaction occurs, regardless of whether the firm buys back the shares at a premium or at a discount. Even greenmail transactions where the initial announcement return is negative, are followed by positive long-term excess returns which largely compensate for the premiums paid to potential hostile bidders. This shows that one should make a clear distinction between two questions: (1) why do managers repurchase stock and (2) when do they repurchase stock. Managers repurchase shares for a variety of reasons (especially in private transactions), but regardless of the reason, they do it when the stock is cheap.
Other evidence consistent with market timing are studies which report that insiders tend to buy stock when the company repurchases shares [17, 87, 102]. A final piece of evidence that may convince the reader is the performance of the KBC Equity Buyback America fund, an open-ended mutual fund sold by the Belgian bank KBC that invests in stocks (1) that announce share buybacks and (2) where the buyback motivation seems to be "undervaluation". The fund was started in July 1998, and I was responsible for the portfolio selection of the fund until February 2004. Table 6.1 shows the performance of all US mutual funds sold in Belgium on February 13, 2004. Regardless of the look-back period (1 year, 3 year or 5 years) the KBC buyback fund had the highest buy-and-hold return, a result difficult to explain by "luck". According to Standard and Poor's, (1) the Sharpe ratio of the fund over the 3-year look-back period was the highest of all 71 comparable funds sold in Belgium. Over the 5-year look-back period the fund's Sharpe ratio was the second largest of the 41 surviving funds.
6.2 International evidence
Subsequent research has found economically and statistically significant long run abnormal returns in Canada , the UK , and Hong Kong . Because in Canada firms have to disclose each month the number of shares repurchased, Ikenberry et al.  are able to directly test the market timing hypothesis. They find that companies who announce a repurchase, but don't complete it experience significant excess returns in the year after the open market authorization announcement, but zero excess returns in the following two years. However firms that complete the repurchase in the year after the authorization announcement experience zero excess returns during that year, but positive excess returns during the following years. Such a pattern is perfectly consistent with a market timing story and clearly inconsistent with theories (such as the agency cost or tax hypothesis) that assume that share prices should only go up if the repurchase takes place. Zhang  uses the same methodology as Ikenberry, Lakonishok and Vermaelen  using data from Hong Kong and find similar results: long-term excess returns are significantly positive for value stocks, not for growth stocks.
The most surprising finding is the fact that long-term excess returns in continental Europe are significantly negative . Lasfer explains the difference in results by the lack of minority shareholder protection in continental Europe (note that a repurchase strengthens the control of the remaining stockholders) and lower information asymmetries in European firms. However, because continental European share repurchase activity is such a recent phenomenon, researchers probably will have to wait a few more years before more definitive long-term event studies can be done.
6.3 The bottom line
When managers are asked "why do you repurchase stock?", the most popular answer is "because we are undervalued". The evidence of long-term event-studies in the 3 countries where share repurchases are the most popular (US, Canada and the UK) is consistent with the hypothesis that, at least on average, managers are able to take advantage of an undervalued stock price. This evidence, in combination with the negative long-term returns observed after equity issues  supports the hypothesis that managers have timing ability: companies repurchase stock when they are undervalued and issue shares when they are overvalued. Information asymmetries create opportunities to take advantage of uninformed investors, and benefit the existing long-term stockholders.
Table 6.1 Performance of KBC buyback fund over different investment horizons, on 13 February 2004 1y 3y 5y KBC Equity Fund 67.92 10.03 78.21 Buyback America Fidelity Funds 54.14 -0.13 57.98 American Growth Fund MLIIF US Opportunities 55.09 -4.98 57.46 Fund Class A (USD) MLIIF US Basic Value 45.95 5.15 36.35 Fund Class A (USD) Threadneedle Investment Funds ICVC American Select Growth 37.97 -21.91 19.58 MLIIF Basic Value Fund Class A (EUR) 23.68 -26.56 18.74 Threadneedle Investment Funds ICVC AmericanGrowth 35.76 -20.12 11.61 Parvest USA C 46.17 -13.90 11.35 Aberdeen Global American Growth Fund A 34.41 -22.24 9.36 Parvest USA D 43.74 -17.63 4.06 Aberdeen Global American Growth Fund B 33.00 -24.54 3.96 Credit Agricole Funds USA Private 32.81 -17.57 -2.80 Credit Agricole Funds USA Classic 32.68 -17.72 -3.02 JP Morgan Fleming Investment Funds JPMF US Select Equity Fund X 38.99 -10.59 -4.04 Source: <www.destandaard.be>
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Share Repurchases|
|Publication:||Foundations and Trends in Finance|
|Date:||May 1, 2005|
|Previous Article:||5: Why do stock prices increase around buyback announcements?|
|Next Article:||7: Dividends versus share repurchase.|