Files in this item



application/pdf8026480.pdf (7MB)Restricted to U of Illinois
(no description provided)PDF


Title:Behavioral Models of Dividend Policy and Their Implications to Financial Management
Author(s):Djarraya, Mohamed
Department / Program:Finance
Degree Granting Institution:University of Illinois at Urbana-Champaign
Subject(s):Business Administration, General
Abstract:This thesis examines the intertemporal change of dividend policy. Since the partial adjustment model and adaptive expectations version of the informational content of dividends are two different behavioral justifications for the same empirical equation, an approach for determining which rationale is appropriate to explain the dividend behavior is suggested. The confusion about making the proper interpretation of the coefficients associated with the empirical equation is avoided by testing a generalized model including the conceptual elements of the two behavioral justifications.
Before positing a solution to this problem, this thesis provides a review of the literature related to the behavioral models of dividend policy and the determinants of the payout ratio. The relationship between the dynamic aspects of dividend, investment, and financing behavior is fully discussed. The impact of dividend behavior upon the value of the firm and the informational content of dividends are also discussed.
The generalized model consists of a set of non-linear parameters which are estimated and discussed for each of the eight companies included in the sample. The econometric properties of several possible estimators of the model are described. It is argued that the ordinary least squares estimation method is inconsistent and does not enable us to solve for the parameters of the generalized model. A maximum likelihood estimation method using Marquardt's non-linear regression technique is presented under different specifications of the disturbance term.
Seasonalities in dividends and earnings data are investigated. The results reveal that seasonality does exist in the historical data in 81.25 percent of the companies for earnings per share and in 46.25 percent of the companies for dividends per share. For a large majority of firms, the failure to use the generalized model to explain the dividend behavior results in a misspecification which has serious consequences for the validity of either the partial adjustment or adaptive expectations models. When testing for auto-correlation, it is found that the auto-correlation parameter is on average negative, but not significantly different from zero.
Although the results support the generalized model for a large majority of firms, this model is not found to be suitable for every firm in the sample. Different specifications of dividend behavior are derived from the generalized model depending on the value and significance of the estimated parameters.
The overall implications of the results of this study to financial management are explored. Observable payout ratio and theoretical target payout ratio are compared for each individual company and for different groups of companies reflecting different dividend behavior. It is found that most companies pay more dividends than theory suggests. This is explained by the fact that most companies included in the sample are well established, and have easy access to capital markets. The generalized model is shown to be of great interest to financial manager because it helps him identify the appropriate dividend behavioral model of any company and thus supplies new information related to his dividend policy decision making process.
Issue Date:1980
Description:190 p.
Thesis (Ph.D.)--University of Illinois at Urbana-Champaign, 1980.
Other Identifier(s):(UMI)AAI8026480
Date Available in IDEALS:2014-12-14
Date Deposited:1980

This item appears in the following Collection(s)

Item Statistics