The influence of change in dispersion in analysts' forecasts on the earnings response coefficient.
Item
-
Title
-
The influence of change in dispersion in analysts' forecasts on the earnings response coefficient.
-
Identifier
-
AAI9807919
-
identifier
-
9807919
-
Creator
-
Cohn, Gordon M.
-
Contributor
-
Adviser: Harry Davis
-
Date
-
1997
-
Language
-
English
-
Publisher
-
City University of New York.
-
Subject
-
Business Administration, Accounting | Economics, Finance
-
Abstract
-
The dissertation examines how change in dispersion in analysts' forecasts ({dollar}\Delta D{dollar}) is an earnings response coefficient (ERC) determinant. Two competing hypotheses are examined. The first hypothesis predicts that {dollar}\Delta D{dollar} is a positive determinant of ERC and the second hypothesis predicts that {dollar}\Delta D{dollar} is a negative determinant of ERC.;The first hypothesis suggests that {dollar}\Delta D{dollar} can be used to measure the importance of the earnings surprise UE. As {dollar}\Delta D{dollar} becomes larger, the accompanying UE is expected to be more important to investors.;The second hypothesis is based on Kim and Verrechia's (1991) (KV) model. KV claim that there is a stronger stock price response to earnings announcements whose information is more precise. Previous research demonstrates that dispersion in analysts' forecasts is an inverse measure of earnings information precision. Based on KV and the previous research, {dollar}\Delta D{dollar} is hypothesized to be a negative ERC component.;A linear regression model tests the hypotheses. Cumulative abnormal return (CAR) is regressed onto unexpected earnings (UE), {dollar}\Delta D{dollar}, 'UE {dollar}\times{dollar} {dollar}\Delta D{dollar}' and control variables. UE is the forecast error.;Two sample populations are examined. One looks at {dollar}\Delta D{dollar} around the annual earnings announcements and the other around quarterly announcements.;The dissertation's results are statistically significant and consistent in both samples. A negative association is found between CAR and both {dollar}\Delta D{dollar} and UE. The relationship between CAR and 'UE {dollar}\times{dollar} {dollar}\Delta D{dollar}' is positive. The last finding confirms Hypothesis I.;Five contributions are made in this research. One, a new ERC determinant is developed. Two, {dollar}\Delta D{dollar} is shown to measure the importance of the earnings surprise. Three, evidence is brought for the hypothesis that {dollar}\Delta D{dollar} is a measure of information content. Four, this is the first study to demonstrate a negative association between the CAR and {dollar}\Delta D{dollar} around an earnings announcement. This result gives support to those researchers who claim that dispersion in analysts' forecasts is a measure of risk. Five, this dissertation uses the I/B/E/S (Institutional Brokers' Estimate System) detail rather than summary tapes to calculate {dollar}\Delta D{dollar}. As a result, more statistically significant statistics are obtained.
-
Type
-
dissertation
-
Source
-
PQT Legacy CUNY.xlsx
-
degree
-
Ph.D.
