Dividend policy and the impact on the share price

Dividend policy is one of the most controversial aspects of the company, which has always caused a lot of controversy among financiers and scientists. Do firms pay dividends, and if so, what share of their profits? What will be the reaction of investors to the announcement of the dividend? Dividend policy is an important attribute of the company, as well as an important element of the financial strategy of the company.

The relevance of dividend policy is not only a huge amount of theoretical and empirical research. It is also observed due to the high demand of Russian firms, investment banks and consulting companies to develop a dividend policy. At the same time, despite the large amount of research in the field of dividend policy, it is necessary to note the fact that almost all of them were based on data for the developed financial markets of the USA, the UK, Germany, Japan, while the emerging market of Russia has not been paid in this regard, a lot of attention. As a result of radical differences between the Russian financial market and the markets of developed countries, it is impossible to draw conclusions about what classical and modern theories of dividend policy will have the right to life in relation to the domestic market, and which do not.

In this case, the aim of my work is to study the impact of dividend payments on the value of Russian companies by using one of econometric methods - the method of least squares.

If any dependence is to take place, the second target will be the analysis of the factors that can influence the value of dividend payments. As an example, I chose major Russian company in the real sector of the economy, whose shares are traded on the Russian stock exchange - LUKOIL.

To build the model and its further validation was taken major Russian public company whose shares are traded on the RTS stock exchange: LUKOIL.

The study analyzed the influence of such factors as company size (market capitalization), investment opportunities (the ratio of market value to its book value), the rating companies (dummy variables), the amount paid dividend (the ratio of the amount of the dividend to the profits of the company) to change the prices of the stocks before and after the announcement of dividend payment.

(PC = a1 + a2*DP + a3* INVEST + a4 * SIZE + a5 * CR + £t

}

{                             E (^ t ) = 0

(                         о (ct) = const

where

PC – price change,

DP –dividend payout,

INVEST –market-to-book value,

SIZE – market capitalization,

CR – credit rating;

£_t - the disturbance term [1].

I used OLS – method to check the model and got such results [2]:

Table 2 Regression statistics

Regression Statistics
Multiple R	0,92
R Square	0,85
Adjusted R Square	0,75
Standard Error	2,79
Observations	11

Table 3 Variance analysis

	df	SS	MS	F	Significance F
Regression	4	262,46	65,61	8,42	0,01
Residual	6	46,78	7,80
Total	10	309,24

Table 4

	Coeffici ents	Standard Error	t Stat	P-value	Lower 95%	Upper 95%	Lower 95,0%	Upper 95,0%
Intercept	14,87	5,63	2,64	0,04	1,10	28,64	1,10	28,64
DP	-95,35	27,92	-3,42	0,01	163,66	-27,04	-163,66	-27,04
INVEST	0,0008	0,0002	4,26	0,01	0,0003	0,001	0,0003	0,001
SIZE	-0,0009	0,0002	-3,80	0,01	-0,001	0,0003	-0,001	-0,0003
CR	-9,59	2,08	-4,62	0,00	-14,67	-4,51	-14,67	-4,51

The specification of our model with calculated parameters are presented below:

Y t = 14,87 – 95,35X 1t + 0,008X 2t - 0,0009X 3t – 9,59X 4t + ɛ t

(5,63)  (27,92)   (0,0002)  (0,0002)     (2,08)  (2,79)

F = 8,42; F crit. = 4,53

R² = 0,85; t crit. = 2,45

According to this model, increase in market-to book value by one million dollars, leads to increase in current value of price change by 0,0008 dollars. And increase in market capitalization of the company by one million dollars, leads to decrease in current value of price change by 0,0009 dollars.

Regression analysis allowed us to draw the following conclusions: on the current value of price change influence the investment opportunities of the company, capitalization, and credit rating. These results were confirmed signaling and Agency motives dividend policy. In addition, some specific result has led to claims that the theory satisfies the preferences of the investor also has the right to life in the Russian context.

R² is high (85%). It means that varies in X explains 85% of varies in Y. F crit. is less than F, therefore R² is not random and quality of specification of econometric model.

E (ɛ t ) = 0 is the first Gauss-Markov assumption that the error ɛ has an expected value of zero given value of the explanatory variable. This means that on average the errors balance out. This is not a restrictive assumption since we can always use a 0 so that this equation holds.

When checking the significance of the coefficients with the help of t-test, I found out that all coefficients passed it, because |t| ≥ t crit . In this case we shouldn’t exclude coefficients from the model.

The calculated level of significance 0,01<0,05 (table 3) confirms the R2 significance.

The Goldfield-Quandt test also confirms that we can use OLS – method to check the model.

Random disturbance are considered to be homoscedastic if both of the following inequalities are valid:

GQ ≤ F crit.

1/GQ ≤ F crit.

In our example both inequalities are valid, so the assumption about homoscedasticity of random disturbance is adequate.

In order to check the third assumption of the Gauss-Markov theorem about the absence of autocorrelation between adjacent random residuals in the model, I used Durbin-Watson test.

nn

DW = X ( e t - e t - 1 ) 2 / £ e;

= 2,96

= 2                     t = 1

In the model d L = 0,59 and d U = 1,93, so there is no information about autocorrelation of the model’s residuals, so we can use least square technique to estimate the model.

And at the end, we should define confidence interval and show that this model is adequate. For this, we should estimate the lower and upper boundaries ˆ for year, using the following formula: 99,5% boundary = Yt - tcrit. st-error, where tcrit. is calculated as it has been shown in part “t-test” and standard error = 2,79, Yˆ – predicted value of Yt. Then we should compare the empirical data for each data with the resulted interval boundaries.

Low level = -3,42

Upper level = 10,24

Empirical = 3,41

Our empirical for 3,41 of 2013 data lies between upper and lower boundaries predicted by our model.

Each of the above-mentioned theoretical studies have confirmed or denied the viability of one of the models for those in the financial and corporate systems, the study of which it was based. Such countries, as a rule, were the United States or Britain. But those conclusions, which were made for them in the advanced economies, could not find his evidence to the emerging market of the Russian Federation. Here was the main goal of my work: on the basis of empirical studies to test the validity of existing theories for Russian companies.