Digitized Honors Theses (2002-2017)

Date of Award

5-2004

Document Type

Undergraduate Thesis

Degree Name

BS

Faculty Mentor

Ross Dickens

Advisor(s)

Natalya Delcoure, Ph.D., Kenneth Hunsader, Ph.D.

Abstract

Using quarterly data, and a sample of 237 dividend paying companies selected from the S&P 500 index, this research tests the accuracy of stock prices predicted by Gordon's (1962) constant growth dividend discount model. For the model input (k), the methodology uses observed rates of return and Single Index Market Model (SilvfM) estimates. For (g), the process uses growth measures of dividends, Earnings Per Share (EPS) and Free Cash Flow per Share (FCF). Finally, for (D1), the study assumes next year's dividend to be known to investors. This research tests differences, correlations, and standard errors between predicted and observed prices to determine the accuracy of the model's price predictions. The study also details the variability of the model predictions, as well as the degree to which variability of model input determinants affects prediction accuracy. Results indicate the constant growth dividend discount model is seriously flawed, and not applicable to valuing shares of stock. No set of input variables consistently predicts prices accurately, and for the companies that best meet the model's assumptions, only six of twenty-four variable input sets predict prices not significantly different from observed prices. Finally, variability in the model's input variables explains little of the variability in predicted prices.

Download

Share

COinS