To relax M&M assumptions of perfect capital market, tax factor should be firstly considered due to its significant influence on dividends policy as well as on the firm’s value. Generally speaking, tax brackets on dividends and capital gains differ to a large extent. Because investors only care the after tax return, high tax imposed on dividend leads to higher required rate of pre-tax return from investors. Thus the dividend is affected on the demand side. Consequently, managers tend to prefer higher retention ratio of earnings to maximize shareholders’ wealth and the firm’s value. The tax effect theory suggests that capital gains have tax advantages over dividend not only because the lower tax rates imposed on capital gains, but also because the tax on dividend are collected immediately in contrast to the tax on capital gains, which could be delayed until the date the capital gain is realized in cash. Therefore, favourable tax policy on capital gains makes the companies reluctant to pay high level dividends in order to lower the cost of capital and thereby increase the firm’s value. According to Brennan’s (1970) after tax version of capital asset pricing model, ceteris paribus, stocks with higher tax are less valued due to the significant tax reduction on dividend.
Black and Scholes (1974) examined Brennan’s model with empirical data and found that the coefficient of dividend influence was insignificant, which is to say, no obvious tax effect could be found. However, their research methods and particularly the definition of dividend yield was criticised by Litzenberger and Ramaswamy (1979).Litzenberger and Ramaswamy (1979) used monthly dividend yield definition to categorize stocks and found that the coefficient of dividend variable is positively significant, which supported the Brennan’s model. Miller and Scholes (1982) pointed out that Litzenberger and Ramaswamy’s conclusion was flawed since short term dividend yield was inappropriate to test the general impact of dividends tax on the stock prices. They also added that the positively significant dividend coefficient was caused by the information bias as they ignored the fact that announcement of dividend omissions might positively affect the dividend yield coefficient for non-dividend class of stocks. Nevertheless, after taking into account of the information bias, Litzenberger and Ramaswamy (1982) reached the results that still consistent with their previous studies. A more recent study conducted by Morgan and Thomas (1998) using the market data in UK suggested that within the 1973 imputation tax system the dividend gained tax treatment advantages over capital gains, thus it is reasonable to predict the negative relationship between dividend yield and risk adjusted returns. However, the relation between them was found to be positive and non-linear, which was inconsistent with Brennan’s model. Thus, the empirical research failed to give conceivable evidence to demonstrate the tax effect.
Location:Home > Academic Researches > Details
Liu Shangchao:Whether Low Dividends Increase Stock Value — theoretical examination on the tax effect hypothesis
2012-12-27 13:07