3. Data and Research Methodology
3.4 Research Methodology
Many methods can be used to measure abnormal returns, such as cumulative abnormal returns, buy-and-hold abnormal returns, and calendar-time portfolios. In fact, Brav and
Gompers (1997) argued that the approach used to calculate cumulative and buy-and-hold abnormal returns with cross-sectional dependence in sample observations can lead to poorly specified test statistics in sampling situations. The calendar-time portfolio methods were found to eliminate the problem of cross-sectional dependence among sample firms because the returns of sample firms were aggregated into a single portfolio. The general approach to calendar-time portfolios was discussed by Fama (1998) and implemented in recent work by Loughran and Ritter (1995) and Brav and Gompers (1997).
By following Fama and French’s methodology (p.8-p.10, 1993) very closely, this study used six portfolios to examine common risk factors in returns related to size and book-to-market equity. The portfolios were formed on a monthly basis from a simple sorting of firms into two groups based on size and three groups based on the BE/ME. First, in June of each year t from1991 to 2005, all stocks (expect financial stocks) listed on the TSE were ranked by size (price times shares). The median size was used to split all stocks in the sample into two groups, namely, small firms or big firms (S or B). Second, the stocks were also assigned to three BE/ME groups based on breakpoints for the bottom 30 percent (low, or L), the middle 40 percent (medium, or M), and the top 30 percent (high, or H) of the ranked values of the BE/ME for all stocks listed on the TSE. The ranked values of the BE/ME for all stocks listed on the TSE were measured at the end of December in year t-1 from 1991 to 2005. Thus, six portfolios (S/L, S/M, S/H, B/L, B/M, and B/H) were finally constructed from the intersection of the two sizes and three BE/ME groups and monthly equal-weighted and value-weighted returns on six portfolios were calculated from July of year t to June of year t+1.
The Fama and French three-factor model (1993, 1996) was used in this paper to test the long-term abnormal stock returns and is specified in Equation (1):
pt
where is the average raw return for stocks in calendar month t (where a sample stock is
included if month t is within the 60-month period following its R&D increase), is the average of five major commercial banks’ 1-month deposit rates, is the value-weighted market index return, is the return on a portfolio of small stocks minus the return on a portfolio of large stocks, and is the return on a portfolio of stocks with high book-to-market ratios minus the return on a portfolio with low book-to-market ratios. Hence, the inference was based on the t-statistic derived from the time-series of the monthly calendar-time portfolio abnormal returns. All t-statistics were adjusted for heteroskedasticity using White’s (1980) method.
Previous studies have argued that although the calendar-time portfolio methods yield more robust test statistics in nonrandom samples, the calendar-time portfolio methods often yield misspecified test statistics in nonrandom samples. In addition, calendar-time abnormal returns do not precisely measure investor experiences. Value-weighted and equal-weighted calendar-time returns were both used in this paper. A debate revolves around the use of value-weighted versus equal-weighted calendar-time returns. Loughran and Ritter (2000) argued that equal weighting was better because it did not obscure the mispricing that was more likely to occur with smaller firms. On the other hand, Fama (1998) argued that value weighting was more appropriate because it more accurately gave the total wealth effects experienced by investor.
4. Empirical Results
Table 3 shows the results of Equation (1), which are the long-term abnormal returns for the full sample using the Fama and French three-factor model for the 1991 to 2005 period.
The alphas, which represent the abnormal return measure, were found to be significantly positive (0.11%) with an equal-weighted measure, but to be insignificant with a
value-weighted measure. The small abnormal returns for the full sample may have been due to the relatively small amounts of spending on R&D by firms in Taiwan. This result was similar to that of Eberhart, Maxwell, and Siddique (2004), who found alphas with equal-weighted and value-weighted measures of 0.69% and 0.43%, respectively, to be both significantly positive in the U.S. stock market from 1974 to 2001.
Because the firms with increases in R&D in Taiwan were concentrated in the electronics industry, the sample was separated into firms that were in the electronics industry and those that were in the non-electronics industry in order to explore their differences in this paper.
Table 4 displays the long-term abnormal returns for the sub-samples of electronics and non-electronics industry firms. In panel A, the alpha for the electronics industry firms was significantly positive (0.6%) using equal-weighted returns but insignificant using value-weighted returns. In panel B, the alphas for the non-electronics industry firms were significantly negative (-0.54% and -0.4%, respectively) using equal-weighted and value-weighted returns. This result is the same as that of Chan, Martin, and Kensinger (1990) who stated that high-technology firms that announce increases in R&D spending obtain positive abnormal returns but that low-technology firms have negative abnormal returns.
This result is also similar to that of Eberhart, Maxwell, and Siddique (2004) who observed that the performance of high-tech firms’ R&D investment in the U.S. is better than that of low-tech firms. From the statistics, it can be inferred that the firms in the electronics industry that expend more on R&D will benefit, whereas it will not be advantageous for firms in the non-electronics industry to engage in R&D activity. This might be because firms in non-electronics industry usually spend money on marketing and advertising to boost their sales, and their economies of scale in R&D are too small for them to benefit from R&D.
Similarly, the sample is also separated into firms with high leverage and those with low leverage. The long-term abnormal returns for the sub-samples of high-leverage and low-leverage firms are presented in Table 5. In panels A and B, the alphas for both the
high-leverage and low-leverage firms are significantly positive (0.87% and 0.6%) when using equal-weighted returns but insignificant when using value-weighted returns. This result reveals that the high-leverage firms with R&D increases have higher long-term abnormal returns. Such firms may believe their R&D investments will result in abnormal profits, and so they are willing to take on more debt. As for shareholders, based on the benefits of debt monitoring, they will tend to believe that debt as opposed to managerial discretion will reduce agency cost. For managers, debt will tend to limit them from making free use of cash flow, and thus managers will work hard and make good investment decisions because high leverage may lead to bankruptcy (e.g., Jensen, 1966; Grossman and Hart, 1982; Stulz, 1990). In R&D activity, it seems good for firms to raise their debt level appropriately.
From the above, it can be seen that the firms in the electronics industry have higher abnormal returns than those in the non-electronics industry. Then, by adding the factor of leverage, we are able to explore whether electronics firms with high leverage obtain higher abnormal returns. The long-term abnormal stock returns for the different leverage levels of firms with increases in R&D in the electronics industry are presented in Table 6. With equal-weighted returns, the alpha (0.91%) for firms with high-leverage levels in the electronics industry is significantly positively higher than the alpha (-0.75%) for the firms with low-leverage levels. The reason for this may be that most firms in the electronics industry should be innovating as a result of engaging in R&D, and so the firms with higher leverage will be those that obtain more money to support their beneficial R&D projects which will raise their firms’ value. In brief, relatively little spending on R&D may result in less profitability from that R&D.
When value-weighted returns are used, the abnormal return estimates are insignificant across all categories of the Fama and French three-factor model except for firms in the non-electronics industry. Loughran and Ritter (2000) noted that the Fama and French three-factor model using value-weighted returns tends to underestimate abnormal returns
when the event is regarded as a managerial choice involving cash flows (such as equity issues) rather than routine events (i.e., quarterly earnings announcements). In addition, value-weighted portfolios can also include some periods in which a single firm accounts for a large proportion of the portfolio, resulting in a high variance of returns because this firm’s unique risk is not diversified away, and thereby leading to low t-statistics.
5. Conclusions
In this paper, we have empirically examined the effects of leverage on long-term abnormal returns following R&D increases in order to observe the performance of R&D investment in Taiwan. Our sample consisted of 3,390 observations for 645 firms listed on the Taiwan Stock Exchange (TSE) from September 1991 through June 2000 with increases in R&D.
The abnormal returns were measured with calendar-time returns when the ratio of R&D to total assets increased above 5%. Because R&D increases are based on accounting data, and not on formal announcements of increases in R&D, the abnormal returns were not measured using event-time returns. In this paper, the potential benefit of such increases in R&D is regarded as intangible information and the long-term abnormal returns that follow the firms’ R&D increases over a five-year period are used to test the efficient markets hypothesis (EMH). The results show that the abnormal returns are ambiguous in terms of providing evidence that the Taiwan market is efficient because the abnormal returns in the sub-samples are not all significant and the abnormal returns are too small. The small abnormal returns may result from the smallness of the scale of R&D conducted in Taiwan.
The overall average intensity of industrial R&D investment in Taiwan falls behind that of the developed countries. Enterprises are the entities that primarily engage in R&D in Taiwan, but Taiwan’s industries are dominated by small and medium-sized enterprises, which generally lack the resources needed for R&D. Therefore, the Taiwan government has
established a series of financial subsidy policies for enterprises to share their R&D risk.
Moreover, Taiwan needs to change its innovation model to encourage more frontier technical innovation or services, and the government should thus concentrate its resources on several farsighted innovation areas that have the potential for development in the future.
It is clear that the R&D investments in Taiwan are mainly concentrated on hi-tech product manufacturing. Our results suggest that the long-term performances of R&D investments in Taiwan’s electronics industry are better than those in Taiwan’s non-electronics industry. This may in turn suggest that firms in Taiwan’s non-electronics industry would not benefit so much from investment in R&D and that their profit would be increased by keeping costs down.
This paper provides evidence that leverage levels significantly affect the abnormal returns of R&D increases. The high-leverage firms with R&D increases have higher abnormal returns than the low-leverage firms. These results correspond to the debt monitoring hypothesis. In particular, in the electronics industry, the high-leverage firms have positive abnormal returns, while the low-leverage firms, by contrast, have negative abnormal returns. One possible reason for this is that firms take on debt to take advantage of economies of scale in good R&D investments, and the other is that debt reduces managerial discretion. To sum up, the agency cost of capital structure model indicates that investors can observe the leverage level of firms when predicting the market’s reaction to the quality of R&D.
References
Aboody, D., and B. Lev (2000), Information asymmetry R&D and insider gain, Journal of
Finance, 55, 2747-2766.
Brav, A., and P. Gompers (1997), Myth or reality? The long-term underperformance of initial public offerings: Evidence from venture and nonventure capital-backed companies,
Journal of Finance, 52, 1791-1821.
Chan, S. H., J. Martin, and J. Kensinger (1990), Corporate research and development expenditures and share value, Journal of Financial Economics, 26,255-276.
Chan, L. K. C., J. Lakonishok, and T. Sougiannis (2001), The stock market valuation of research and development expenditures, Journal of Finance, 56, 2431-2456.
Chang, P. L., and H.Y. Shih (2004), The innovation systems of Taiwan and China: A comparative analysis, Technovation, 24, 529-539.
Eberhart, A.C., and R. S. Akhtar (2002), The long-term performance of corporate bonds (and stocks) following seasoned equity offerings, Review of Financial Studies, 15, 1385-1406.
Eberhart, A. C., F. M. William, and R.S. Akhtar (2004), An examination of long-term abnormal stock returns and operating performance following R&D increases, Journal of
Finance, 59, 623-650.
Fama, E. F., and K. R. French (1992a), The cross-section of expected returns on stocks and bonds, Journal of Finance, 47, 427-465.
Fama, E. F., and K. R. French (1992b), The economic fundamentals of sizes and book-to-market equity, Working paper, Graduate School of Business, University of Chicago.
Fama, E. F., and K. R. French (1993), Common risk factors in the returns on stocks and bonds,
Journal of Financial Economics, 33, 3-56.
Fama, E.F., and K. R. French. (1996), Multifactor explanations of asset pricing anomalies,
Journal of Finance, 51, 55-85.
Fama, E. F., and K. R. French (1998), Market efficiency, long-term returns, and behavioral finance, Journal of Financial Economics, 49, 283-306.
Grossman, S. J., and O. D. Hart (1982), Corporate financial structure and managerial incentives, the economics of information and uncertainty, edited by John J. McCall.
University of Chicago Press.
Ho, Y. K., M. Tjahjapranata, and C. M. Yap (2006), Size, leverage, concentration, and R&D investment in generating growth opportunities, Journal of Business, 79, 851-876.
Lang, L. H. P., E. Ofenk, and R. M. Stulz (1996), Leverage, investment, and firm growth,
Journal of Financial Economics, 40, 3-29.
Lev, B., and T. Sougiannis (1996), The capitalization, amortization, and value relevance of R&D, Journal of Accounting and Economics, 21, 107-138.
Loughran, T., and J. Ritter (1995), The new issues puzzle, Journal of Finance, 50, 23-51.
Loughran, T., and J. Ritter (2000), Uniformly least powerful tests of market efficiency,
Journal of Financial Economics, 55, 361-389.
Lyon, J., B. Barber, and C. L. Tsai (1999), Improved methods for tests of long-run stock abnormal returns, Journal of Finance, 54, 165-201.
Jensen, M.C., and W.H. Meckling (1976), Theory of the firm: Managerial behavior, agency costs and ownership structure, Journal of Financial Economics, 3, 305-360.
Jensen, M.C. (1986), Agency costs of free cash flow, corporate finance, and takeovers,
American Economic Review, 76, 323-329.
Mathews, J. A. (2001), The origins and dynamics of Taiwan’s R&D consortia, Research
Policy, 1315, 1-20.
Ming, X., and Z. Chu (2004), The explanatory power of R&D for the cross-section of stock returns: Japan 1985-2000, Pacific-Basin Finance Journal, 12, 245-269.
Modigliani, F., and M. Miller (1958), The cost of capital, corporation finance and the theory
of investment, American Economic Review, 48, 261-297.
Modigliani, F., and M. Miller (1963), Corporation income taxes and the cost of capital: A correction, American Economic Review, 53, 261-297.
Myer, S. C., and N. S. Majluf (1984), Corporate financing and investment decisions when firms have information that investors do not have, Journal of Financial Economics, 13, 187-221.
Reinganum, M. R. (1981), A new empirical perspective on the CAPM, Journal of Financial
and Quantitative Analysis, 16, 439-462.
Stulz, R. M. (1990), Managerial discretion and optimal financing policies, Journal of
Financial Economics, 26, 3-27.
Szewczyk, S., G.. Tsetsekos, and Z. Zantut (1996), The valuation of corporate R&D expenditures: Evidence from investment opportunities and free cash flow, Financial
Management, 25, 105-110.
White, H. (1980), A heteroskedasticity-consistent covariance matrix estimator and a direct test for heteroskedasticity, Econometrica, 48, 817-838.
Zantout, Z. Z. (1997), A test of the debt-monitoring hypothesis: The case of corporate R&D expenditures, Financial Review, 32, 21-48.
Table 1 Descriptive Statistics
Table 1 shows descriptive statistics of the variables based on 3,390 observations of R&D increases for 646 firms from 1991 through 2000. R&D intensity, size and average monthly return are measured as of the beginning of the sample firm’s R&D increase month.
Mean Median Std. Deviation
R&D intensity measure :
R&D/total assets 0.0151 0.0095 0.0179
Size (NT$ million) 18,736.40 5,695 60,992.92
Market value equity/Book value equity (ME/BE)
2.6003 2.19 2.3781
Average monthly return (%) 1.5175 0 13.6862
Table 2 Distribution of Firms with R&D Increases in Industries
Table 2 shows the distribution of all firms in the sample using the two-digit industry code based on the categories of the Taiwan Stock Exchange Corp. from 1991 through 2000.
Industry Code Industry Number of firms with increases in
R&D/ Total Assets
1100 Cement 12
1200 Food 148
1300 Plastics 184
1400 Textiles 222
1500 Electrical Machinery 316
1600 Appliance Cable 110
1700 Chemical 268
1800 Glass Ceramics 85
1900 Paper and Pulp 97
2000 Iron and Steel 112
2100 Rubber 109
2200 Automobiles 58
2300 Electronics 1,437
2500 Construction 45
2600 Transportation 0
2700 Tourism 0
2900 Wholesale and Retail 0
9900 Others 187
Total 3,390
Table 3 Long-term Abnormal Stock Returns of Firms with R&D Increases - Full Sample
Long-term abnormal stock returns are provided for the full sample of 3,390 observations with R&D increases from 1991 to 2005 using the Fama and French (1993, 1996) three-factor model:
pt
for stocks in calendar month t (where a sample stock is included if month t is within the 60-month period following its R&D increase), is the average of five major commercial banks’ 1-month deposit rates, is the value-weighted market index return, is the return on a portfolio of small stocks minus the return on a portfolio of large stocks, and is the return on a portfolio of stocks with high book-to-market ratios minus the return on a portfolio with low book-to-market ratios. The intercept (α) in the above equation is the abnormal return measure.
Rpt
Equal-weighted return 0.0011 1.095 0.9176 -0.7129
(3.94)*** (14.5)*** (3.12) *** (-0.46)
Value-weighted return 0.0012 0.3260 -2.3048 -1.2413
(1.14) (7.01)*** (-17.86)*** (-2.53) ***
The symbols*, ** and *** denote statistical significance at the 10%, 5% and 1% levels, respectively.
Table 4 Long-term Abnormal Stock Returns of Firms with R&D Increases – Sub-sample
Long-term abnormal stock returns for the electronics sample of 1,437 observations and the non-electronics sample of 2,543 observations with R&D increases from 1991 to 2005 are provided using the Fama and French (1993, 1996) three-factor model:
pt
for stocks in calendar month t (where a sample stock is included if month t is within the 60-month period following its R&D increase), is the average of five major commercial banks’ 1-month deposit rates, is the value-weighted market index return, is the return on a portfolio of small stocks minus the return on a portfolio of large stocks, and is the return on a portfolio of stocks with high book-to-market ratios minus the return on a portfolio with low book-to-market ratios. The intercept (α) in the above equation is the abnormal return measure.
Rpt
Equal-weighted return 0.0060 0.7189 -0.5963 2.3640
(1.63)* (11.24)*** (-2.97)*** (1.04)
Value-weighted return 0.0035 0.4734 -1.7947 5.1099
(1.07) (4.62) *** (-7.07)*** (2.55)**
Panel B: Non-Electronics Sample
Intercept( α ) b s h
Equal-weighted return -0.0054 0.3598 -1.9972 -3.0847
(-1.73)** (6.90)*** (-9.71)*** (-1.23)
Value-weighted return -0.0040 0.1182 -3.2225 -2.9335 (-4.41)*** (7.23)*** (-49.95)*** (-3.92)***
The symbols*, ** and *** denote statistical significance at the 10%, 5% and 1% levels, respectively.
Table 5 Long-term Abnormal Stock Returns on the Leverage Levels of Firms with R&D Increases – Sub-sample
Long-term abnormal stock returns for the high-leverage sample of 1,696 observations and low-leverage sample of 1,694 observations with R&D increases from 1991 to 2005 are provided using the Fama and French (1993, 1996) three-factor model:
pt
for stocks in calendar month t (where a sample stock is included if month t is within the 60-month period following its R&D increase), is the average of five major commercial banks’ 1-month deposit rates, is the value-weighted market index return, is the return on a portfolio of small stocks minus the return on a portfolio of large stocks, and is the return on a portfolio of stocks with high book-to-market ratios minus the return on a portfolio with low book-to-market ratios. The intercept (α) in the above equation is the abnormal return measure.
Rpt
Equal-weighted return 0.0087 0.9898 0.3840 -0.4938
(2.65)*** (8.96)*** (1.08) (-0.38)
Value-weighted return 0.0010 0.3475 -2.2312 -0.0501
(0.83) (6.69)*** (-14.73)*** (-0.06)
Panel B: Low-leverage Sample
Intercept( α ) b s h
Equal-weighted return 0.006 0.7747 -0.0071 -3.6339
(2.61)*** (12.55)*** (-0.30) (-2.97)***
Value-weighted return -0.004 0.2221 -2.7391 -2.7374
(-0.54) (8.36)*** (-31.41)*** (-4.28)***
The symbols*, ** and *** denote statistical significance at the 10%, 5% and 1% levels, respectively.
Table 6 The Long-term Abnormal Stock Returns on the Leverage Levels of Firms with R&D Increases in the Electronics Industry
Long-term abnormal stock returns for a high-leverage sample of 718 observations and low-leverage sample of 719 observations with R&D increases in the electronics industry from 1991 to 2005 by using the Fama and French (1993, 1996) three-factor model:
pt
for stocks in calendar month t (where a sample stock is included if month t is within the 60-month period following its R&D increase), is the average of five major commercial banks’ 1-month deposit rates, is the value-weighted market index return, is the return on a portfolio of small stocks minus the return on a portfolio of large stocks, and is the return on a portfolio
for stocks in calendar month t (where a sample stock is included if month t is within the 60-month period following its R&D increase), is the average of five major commercial banks’ 1-month deposit rates, is the value-weighted market index return, is the return on a portfolio of small stocks minus the return on a portfolio of large stocks, and is the return on a portfolio