Dividend Regressions: January 2008

---

Variables used in the regression

1. Dividend Yield = Dividends per share in most recent year/ Current Stock Price
2. Dividend Payout Ratio = Dividends / Net Income
3. Insider Holdings = Shares held by insiders/ Primary number of shares outstanding
4. Standard deviation = 3 year standard deviation in stock returns (or ln(stock prices). You can get this from the complete dataset for US companies under updated data.
5. ROE = Return on equity in most recent time period (Net Income/ Book value of Equity)
6. Expected Growth in EPS over next 5 years = Consensus analyst estimate (or your own) of expected growth in EPS . If you don't have an analyst estimate, use your own estimate of expected growth.
7. Market Debt to Capital = Debt/ (Debt + Market Value of Equity): If you have market value for debt, use it. If not, use book value of debt and market value of equity.

Dividend Yield Regression

Model Summary
Model	R	R Square	Adjusted R Square	Std. Error of the Estimate
1	.569^a	.324	.322	1.502
a. Predictors: (Constant), 3-yr Standard Deviation (Stock Price), Insider Holdings, Expected Growth in EPS: next 5 years

Coefficients^a,b
Model		Unstandardized Coefficients		Standardized Coefficients	t	Sig.
		B	Std. Error	Beta
1	(Constant)	.039	.001		37.378	.000
	Expected Growth in EPS: next 5 years	-.093	.006	-.422	-1.623E1	.000
	Insider Holdings	-.010	.004	-.063	-2.616	.009
	3-yr Standard Deviation (Stock Price)	-.039	.004	-.244	-9.385	.000
a. Dependent Variable: Dividend Yield
b. Weighted Least Squares Regression - Weighted by Market Cap

• How do I use this regression?

Assume that you want to estimate the dividend yield for a firm with the following characteristics:

Insider holdings = 9% of outstanding stock

3-year standard deviation = 45%

Expected Growth in EPS over next 5 years = 12%

• Predicted Value

Expected Dividend yield = 0.039 - 0.093 (.12) - 0.010 (.09) - 0.039 (.45) = .00939 = 0.94%

If your predicted value is less than zero, your predicted dividend yield is zero.

Dividend Payout Regression

Model Summary
Model	R	R Square	Adjusted R Square	Std. Error of the Estimate
1	.368^a	.135	.133	27.28093
a. Predictors: (Constant), ROE, Expected Growth in EPS: next 5 years, 3-yr Standard Deviation (Stock Price)

Coefficients^a,b
Model		Unstandardized Coefficients		Standardized Coefficients	t	Sig.
		B	Std. Error	Beta
1	(Constant)	.683	.025		27.410	.000
	Expected Growth in EPS: next 5 years	-.313	.121	-.077	-2.595	.010
	3-yr Standard Deviation (Stock Price)	-1.070	.099	-.322	-1.085E1	.000
	ROE	-.185	.060	-.088	-3.062	.002
a. Dependent Variable: Dividend Payout
b. Weighted Least Squares Regression - Weighted by Market Cap

• How do I use this regression?

ROE = 20%

3-year standard deviation = 45%

Expected Growth in EPS over next 5 years = 12%

• Predicted Value

Expected Dividend payout ratio= 0.683 - 0.313 (.12)- 1.070 (.45) - 0.185 (.20) = 0.1269 or 12.69%

If your predicted value is less than zero, your predicted dividend payout ratio is zero.