Stock splits and abnormal returns in an overactive internet market segment.

Stretcher, Robert ; McLain, Michael ; Maheshwari, Sharad 等

INTRODUCTION

The significance of dividend policy to managerial finance continues to be a topic of contention. The work on the role of dividend policy is quite extensive. Stock splits are perhaps even more questionable in terms of the rationale and results of their use. A variety of arguments concerning stock splits have been pursued in past literature. Copeland (1979) provides six reasons for splitting a stock: maintenance of a price range for the firm's shares, reduction of odd-lot trading (since high stock price reduces divisibility), creation of an increase in trading volume, increased brokerage revenue, lowering of bid-ask price, and to encourage an increase the number of shareholders. Ikenberry, Rankine & Stice (1996) found that stock splits most often occur when there has been a substantial increase in the price of the stock, or when a stock trades at a high price. They also found that stock splits allow the investor an excess return during the period after the announcement. Additionally, they discovered that only short-term positive results were achieved when firms had low pre-split share prices.

Copeland (1979) defined liquidity as "changes in the proportional share volume traded and change in transaction costs as a percent of volume traded" and found that there was reduced liquidity following a stock split. He also determined that the announcement information about a stock split was disseminated within a two to three week time frame. Since the study was carried out prior to deregulation of brokerage commissions and the advent of the $8 trade, these results may vary based upon today's market. The information flow through cable television (e.g.: CNBC) and through various Internet sites facilitates dissemination of news about a stock split.

Other studies of stock splits yield further insight into the dynamics of market reactions. Brennan and Copeland (1988) determined that companies with stock splits had greater variance in returns on the announcement date of the stock split. The Beta of the stock would increase around the ex-date and on the day following. There was also a permanent increase in the stocks' average Betas after the ex-date. This followed the work of Ohlson and Penman (1985), who concluded that stock returns would increase immediately following the effective date of a stock split.

From a value viewpoint, it may be argued that a stock split does nothing more than change the denomination of the number of shares held, while the value per share changes such that the total value remains constant. From this perspective, stock splits would appear benign in terms of affecting any change in wealth. It is analogous to the idea that a five dollar bill is equivalent to five one dollar bills. The wealth is equal, no matter which denomination is held.

Other studies have pursued the possibility that, in an imperfect world where information is not heterogeneous to all market subgroups, there is the possibility of information content in any variety of managerial actions, including stock splits. Public announcements of stock splits may have the effect of drawing attention to the company's condition. This would be especially useful to management if the firm is undervalued, because a closer examination of the firm by an outsider may have the effect of a positive revaluation and, thus, higher bids. Penman (1983) asserts that, if stock splits signal manager's future value of the firm, the stock price should react upon the time of the announcement. Upon an announcement, therefore, investors should reassess the value of the firm.

There seems to be considerable belief that an optimal stock price range exists, although there is little empirical support for that belief (Lakonishok & Lev, p. 929, 1987). A stock split does not increase the shareholders proportional ownership of the firm, but only increases the number of shares outstanding. Since the number of shares rise upon a stock split, the question arises about why the price of the stock would increase. McNichols and Dravid (1990) state that splits realign the price of the stock prices to a preferred trading range. Splits increase the number of shares outstanding with the presumption that by increasing the number of shares it will result in increasing the number of shareholders, and thereby increasing the number of trades in the stock. A price range that allows for trading flexibility would be one that prevents the per share value from rising to levels that would rule out small-scale investors. It would also keep shares from appearing 'too cheap' by preventing share values from dropping below some value, determined by the perception of the markets. Angel (1997) found that share prices are relatively stable over an extended period of time. Interestingly, different countries maintain different average share prices. In the U.S., the average price per share on the New York Stock Exchange was relatively stable during the period of 1924 until 1994, even though the Standard and Poor's Index had a substantial increase in value during the same period.

Ikenberry, Rankine and Stice (1996) found that stock splits occur more frequently during a period of a rising bull stock market. This suggests that there exists some underlying reason for split frequency. This study explores the possibility of whether or not stock split announcements are strategically advisable in chaotic market conditions. Specifically, we examine stocks within an overactive segment of the stock market, internet stocks. We observe changes in value during the late 1990's, a period characterized by extreme price increases and considerable price volatility. In the flurry of market activity, we consider the question of whether stock split announcements are an effective way for firms to gain attention, with the objective of excess returns.

METHODOLOGY

A sample of 360 internet companies within the Worden Telechart 2000 database was compiled. Among these companies, 122 carried out stock splits within the period under consideration, from July 1, 1998 to March 30, 2000. Among these stock splits, 75 splits occurred in public markets. Announcement dates for these splits were acquired from the "Stock Splits and Stock Dividends" database from e-analytics.com, an internet site maintained by Equity Analytics, Ltd. For nine of the 75 splits, no announcement dates were available. This left 66 observations for the statistical analysis.

For each split, daily returns were calculated from 15 days prior to the split to 15 days after the split. Using the AMEX internet index (^IIX) as our comparison base, excess return for each day (daily return for the stock minus daily return for the index) formed the observations for the dataset.

The data were organized according to an announcement date, which represents day zero for all stock splits within the data set. For each day prior to and after the announcement date, summary statistics were calculated. The summary results appear in table 1.

INTERPRETATION OF RESULTS

As the summary in table 1 indicates, only the excess returns on day zero are significant and positive. The significant and positive result may be interpreted as a same-day positive market reaction to the stock split announcement. Interestingly, the results suggest that the days immediately preceding and immediately following the announcement entail no significant excess return, positive or negative. This represents a departure from results of studies done on general market data in not-so-chaotic time periods, which indicate at least a minimal level of significance on days close to the announcement date (Ikenberry, Rankine & Stice 1996).

There are many possible reasons for our results. In this volatile environment (i.e. CV's from 247% to 99,563%), especially where prices are generally increasing at a high rate, investors may be less concerned with the relatively small possibility of gain around a stock split announcement than they are with other aspects of the same market segment. During this time period, for example, passive investors realized exceptionally high returns simply by holding a diversified portfolio of 'tech stocks.' As one of our public policy officials described it, 'irrational exuberance' was the market emotion of preference. It may also indicate that signaling may be more difficult in a chaotic, noisy environment.

Interestingly, there is a somewhat significant result for day -3, prior to the split announcement. This may be due to the considerable degree of variance and relative variance that characterizes the entire data set. It could be argued that an information 'leak' three days prior to announcement could spark movement in excess returns, but given the negative sign on the coefficient, this would be counter-rational.

To develop the argument about the significance of the differences in excess returns between the days, two ANOVAs were conducted. In table 2, we test for the null hypothesis that there is no difference between the means of the different splits:

The P-value from the table indicates that the split averages are not significantly different from zero. This result lends credence to our conclusion about the differences in days. It also suggests that there are no significant 'offsetting' effects among the averages of table 1, where one significantly positive result might otherwise negate another significantly negative result on the same day.

In table 3, we test for the null hypothesis that the difference between the means among the different days is zero:

We reject the null hypothesis of zero difference between days. This is consistent with the result from table 1.

As a supplemental test, we omit the day 1 average and test for differences among the remaining days' averages:

The P-value indicates an acceptance of the null hypothesis, implying that all other days but zero have about the same average.

An interesting statistical result can also be observed from day -9 to day -2. On these days, not only did the averages all have negative signs, they also involved a total cumulative average of about 5.5%, a negative movement greater than the average positive movement on day zero. This significant negative run suggests that day zero returns may actually be a recovery of sorts.

IMPLICATIONS FOR INVESTMENT STRATEGY

The statistical summary suggests that, with an investment strategy, it would be difficult to derive benefit. In order for an investor to take advantage of the same-day excess return, prior knowledge of the split announcement would have to be available.

It is plausible that an investor capable of quickly executing trading orders could trade within the announcement day. Without intra-day data, though, determining the potential benefit is beyond this study. It does appear, however, that there is a lesser promise of excess returns (from reacting to stock split announcements) in this chaotic environment for the internet segment than may be possible in the general market.

REFERENCES

Agrawal, A.; J. F. Jaffe; and G. N. Mandelker (1992). "The Post-Merger Performance of Acquiring Firms in Acquisitions: A Reexamination of an Anomaly." Journal of Finance, 47 (Sept), 1605-1621.

Angel, J. J. (1994). "Tick Size, Share Prices, and Stock splits." Unpubl. Manuscript FINC 1377-11-694, Georgetown Univ.

Anshuman, V. R., and A. Kalay. (1993) "A Positive Theory of Stock Splits." Unpubl. Manuscript, Boston College.

Asquith, P.; P. Healy; and K. Palepu. (1989) "Earnings and Stock Splits." The Accounting Review, 64 (July), 387-403.

Baker, H. K., and P. L. Gallagher (1980). Management's View of Stock Splits." Financial Management, 9 (Summer), 73-77.

Baker, H. K., and A. Phillips (1994). "Why Companies Issue Stock Dividends." Financial Practice and Education, 3 (Fall), 29-27.

Baker, H. Kent, and Patricia L. Gallagher (1980). "Management's View of Stock Splits." Financial Management, Summer, 73-77.

Banz, R. W., and W. J. Breen (1986). "Sample Dependent Results Using Accounting and Market Data: Some Evidence." Journal of Finance, 41 (Sept.), 779-794.

Barber, B. M., and J. D. Lyon (1996). "Detecting Long-Run Abnormal Stock Returns: The Empirical Power and Specifications of Test Statistics." Unpubl. Manuscript, Univ. of California at Davis.

Brennan, M. J., and T. E. Copeland (1988). "Stock Splits, Stock Prices, and Transaction Costs." Journal of Financial Economics, 22 (Oct.), 83-101.

Brennen, J. J. and P. J. Hughes (1991). "Stock Prices and the Supply of Information." Journal of Finance, 46 (Dec), 1665-1691.

Conroy, R.; R. Harris; and B. Benet (1990). "The Effects of Stock Splits on Bid-Ask Spreads." Journal of Finance, 65 (Sept.), 1285-1295.

Copeland, T. "Liquidity Changes following Stock Splits (1979)." Journal of Finance, 34 (March), 115-141.

Copeland, Thomas E (1979). "Liquidity Changes Following Stock Splits." Journal of Finance, March, 115-141.

Cusatis, P. J.; J. A. Miles; and J. R. Woolridge (1993). "Restructuring through Spinoffs: The Stock Market Evidence." Journal of Financial Economics, 33 (June), 293-311.

Desai, H., and P. C. Jain (1995). "Dividend Signaling and Common Stock Returns following Stock Splits and Reverse Splits." Unpubl. Manuscript, Tulane Univ.

Dharan, B. G., and D. Ikenberry (1995). "The Long-Run Negative Drift of Post-Listing Stock Returns." Journal of Finance, 50 (Dec), 1547-1574.

Fama, E.; L. Fisher; M. Jensen; and R. Roll (1969). "The Adjustment of Stock Prices to New Information." International Economic Review, 10 (Feb), 1-21.

Grinblatt, M. S.; R. W. Masulis; and S. Titman (1984). "The Valuation Effects of Stock Splits and Stock Dividends." Journal of Financial Economics, 13 (Dec), 461-490.

Ikenberry, D., and J. Lakonishok (1993). "Corporate Governance through the Proxy Contest: Evidence and Implications." Journal of Business, 66 (July), 405-435.

Koski, J (1995). "Measurement Effects and the Variance of Returns after Stock Splits and Stock Dividends." Unpubl. Manuscript, Univ. of Wshington .

Lakonishok, J., and B. Lev (1987). "Stock Splits and Stock Dividends: Why, Who, and When." Journal of Finance, 62 (Sept), 913-932.

Lakonishok, J., A. Shleifer; and R. W. Vishny (1994). "Contrarian Investing, Extrapolation and Risk." Journal of Finance, 49 (Dec), 1541-1578. Lamoureux, C., and P. Poon (1987). "The Market Reaction to Stock Splits." Journal of Finance, 62 (Dec), 1347-1370.

Loughran, T., and J. R. Ritter (1995). "The New Issues Puzzle." Journal of Finance, 50 (March), 23-51.

Maloney, M., and J. Mulherin (1992). "The Effects of Splitting on the Ex: A Microstructure Reconciliation." Financial Management, 21 (Winter), 44-59.

McNichols, M., and A. Dravid (1990). "Stock Dividends, Stock Splits, and Signaling." Journal of Finance, 45 (July), 857-879.

McNichols, Maureen, and Ajay Dravid (1990). "Stock Dividends, Stock Splits, and Signaling." Journal of Finance, July, 857-879.

Michaely, R.; R. H. Thaler; and K. L. Womack (1995). "Price Reactions to Dividend Initiations and Omissions: Overreaction of Drift?" Journal of Finance, 50 (June), 573-608.

Ohlson, J., and S. Penman (1985). "Volatility Increases Subsequent to Stock Splits: An Empirical Aberration." Journal of Financial Economics, 14 (June), 251-266.

Ritter, J. R (1991). "The Long-Run Performance of Initial Public Offerings." Journal of Finance, 46 (March), 1-27.

Spiess, D. K., and J. Affleck-Graves (1995). "Underperformance in Long-Run Stock Returns following Seasoned Equity Offerings." Journal of Financial Economics, 38, 243-267.

Szewczyk, S. H. and G. P. Tsetsekos (1995). "Why Stocks Split? Evidence from Institutional Ownership." Unpubl. Manuscript, Drexel Univ.

Womack, K (1996). "Do Brokerage Analysts' Recommendations Have Investment Value?" Journal of Finance, 51 (March), 137-167.

Robert Stretcher, Hampton University

Michael McLain, Hampton University

Sharad Maheshwari, Hampton University

Tracey Hayes, Credit Suisse First Boston

TABLE 1

 Day Average Sample SD CV Z value p-value

 -15 -0.11% 6.74% -6000% -0.1323 0.8948
 -14 -0.12% 6.28% -5270% -0.1506 0.8803
 -13 1.14% 5.80% 510% 1.5555 0.1198
 -12 -0.24% 6.44% -2710% -0.2929 0.7696
 -11 0.39% 6.65% 1695% 0.4684 0.6395
 -10 0.15% 6.10% 4039% 0.1981 0.8430
 -9 -0.08% 4.97% -5897% -0.1357 0.8921
 -8 -0.20% 6.84% -3454% -0.2316 0.8169
 -7 -0.13% 5.78% -4587% -0.1744 0.8615
 -6 -0.93% 4.74% -510% -1.5684 0.1168
 -5 -1.27% 5.83% -458% -1.7475 0.0805
 -4 -0.10% 6.27% -6413% -0.1248 0.9007
 -3 -1.61% 4.88% -303% -2.6367 0.0084
 -2 -1.19% 6.69% -563% -1.4220 0.1550
 -1 1.20% 11.01% 915% 0.8745 0.3819
 0 4.81% 11.90% 247% 3.2326 0.0012
 1 -0.84% 6.51% -776% -1.0303 0.3029
 2 -0.33% 6.78% -2074% -0.3858 0.6997
 3 1.01% 7.37% 731% 1.0944 0.2738
 4 1.52% 9.58% 629% 1.2712 0.2037
 5 -0.16% 6.98% -4471% -0.1789 0.8580
 6 0.45% 7.57% 1687% 0.4743 0.6353
 7 0.01% 6.43% 99564% 0.0080 0.9936
 8 1.06% 7.44% 703% 1.1374 0.2554
 9 -0.97% 5.36% -554% -1.4445 0.1486
 10 1.60% 7.24% 453% 1.7664 0.0773
 11 -0.37% 5.24% -1423% -0.5621 0.5740
 12 0.78% 7.96% 1025% 0.7803 0.4352
 13 -0.20% 5.94% -2920% -0.2740 0.7841
 14 1.16% 6.78% 583% 1.3727 0.1698
 15 -0.22% 6.71% -3117% -0.2567 0.7974

TABLE 2: ANOVA Between Splits
Ho: Avg (obs1) = Avg (obs2) = ... Avg (obs66)

Source of Variation SS df MS

Between Groups 0.276082 63 0.004382
Within Groups 9.482346 1915 0.004952
Total 9.758428 1978

Source of Variation F P-value F crit

Between Groups 0.885015 0.7273 1.317316
Within Groups
Total

TABLE 3: ANOVA Between Days
Ho: Avg (day -15) = Avg (day -14) = ... = Avg (day 15)

Source of Variation SS df MS

Between Groups 0.272426 30 0.009081
Within Groups 9.486002 1948 0.00487
Total 9.758428 1978

Source of Variation F P-value F crit

Between Groups 1.864803 0.003074 1.465002
Within Groups
Total

TABLE 4: ANOVA Between Days (omit day zero)
Ho: Avg(day-15) = Avg(day-14) = ... = Avg(day-1) =
Avg(day+1) = ... = Avg(day15)

Source of Variation SS df MS

Between Groups 0.131999 29 0.004552
Within Groups 8.593848 1885 0.004559
Total 8.725847 1914

Source of Variation F P-value F crit

Between Groups 0.998385 0.468284 1.473534
Within Groups
Total