Prospective Earnings Per Share

University of Cambridge, Judge Institute of Management Working Paper No. 06/2004

65 Pages Posted: 21 Feb 2005

See all articles by Graham Bates

Graham Bates

University of Cambridge - Judge Business School

M. A. H. Dempster

University of Cambridge - Centre for Financial Research; Cambridge Systems Associates Limited; University of Cambridge - Judge Business School

Hok Go

University of Cambridge - Judge Business School

Yee Sook Yong

University of Cambridge - Judge Business School

Date Written: 2004

Abstract

This report considers the relation between pro-forma and forecast consensus earnings per share (EPS) figures in terms of six measures identified, qualitatively, as good indicators for quality of earnings. These are, return on capital employed (RoCE), productive asset reinvestment ratio (PARR), cash realisation (CR), tax rate (TR), Standard and Poors (S&P) equity rating and S&P debt rating as identified by Merrill Lynch. The choices are thought to capture aspects of and differentiate between long term strategies and short term, non-sustainable earnings through financial engineering.

Analyses are run on ten years of data from 1992 to 2001 for 131 S&P500 companies also provided by Merrill Lynch, using consensus EPS forecasts one year ahead. This sample is smaller than 500 because the index is subject to change, with companies being added and removed based on their market share or other factors. S&P debt ratings were not available over the entire sample length for all companies thus cutting the two ratings from analyses. However, correlations between the two S&P ratings and other indicator measures are found to be high so their removal is not a significant problem. The ten years are split into three phases based on market sentiment:

1992-1995 Bull market, emphasis on sustainable growth 1996-1999 Bull market, emphasis on high earnings per share (bubble) 2000-2001 Bear market, refocussing on sustainability.

Given the radical change in conditions after 2000, the sample is enlarged for 2000 to 2001 and investigated separately. This increases sample size to 366, but does not change results significantly or give any further insights.

The ability of the four indicator measures to connect EPS predictions to released figures appears weak in analyses. Adding lagged information and market proxies improves the situation, but unfortunately not sufficiently for linear regression to be used confidently for out-of-sample prediction. Given a sample of ten observations in time, non-linear regressions were not carried out. The indicator measures are not thought to be effective indicators of companies' future performance.

Keywords: earnings per share, financial engineering

JEL Classification: G10

Suggested Citation

Bates, R. Graham and Dempster, M. A. H. and Go, Hok and Yong, Yee Sook, Prospective Earnings Per Share (2004). University of Cambridge, Judge Institute of Management Working Paper No. 06/2004, Available at SSRN: https://ssrn.com/abstract=670084 or http://dx.doi.org/10.2139/ssrn.670084

R. Graham Bates (Contact Author)

University of Cambridge - Judge Business School ( email )

Trumpington Street
Cambridge, CB2 1AG
United Kingdom
01223 339700 (Phone)
01223 339701 (Fax)

HOME PAGE: www.jims.cam.ac.uk

M. A. H. Dempster

University of Cambridge - Centre for Financial Research ( email )

Centre for Mathematical Sciences
Wilberforce Road
Cambridge, CB3 0WA
United Kingdom

Cambridge Systems Associates Limited ( email )

5-7 Portugal Place
Cambridge, CB5 8AF
United Kingdom

University of Cambridge - Judge Business School ( email )

Trumpington Street
Cambridge, CB2 1AG
United Kingdom

Hok Go

University of Cambridge - Judge Business School ( email )

Trumpington Street
Cambridge, CB2 1AG
United Kingdom

HOME PAGE: www.jims.cam.ac.uk

Yee Sook Yong

University of Cambridge - Judge Business School ( email )

Trumpington Street
Cambridge, CB2 1AG
United Kingdom

HOME PAGE: www.jims.cam.ac.uk

Do you have negative results from your research you’d like to share?

Paper statistics

Downloads
500
Abstract Views
3,314
Rank
104,005
PlumX Metrics