An empirical analysis of the dynamic relation among investment, earnings and dividends

Richard A. DeFusco, Lee M. Dunham, John Geppert

Research output: Contribution to journalArticle

1 Citation (Scopus)

Abstract

Purpose – The purpose of this paper is to examine the dynamic relationships among investment, earnings and dividends for US firms. The sample period is 1950-2006. Design/methodology/approach – The authors use a firm-level vector auto-regression (VAR) framework to examine the firm-level dynamics among investment, earnings and dividends. The firm-level VAR yields Granger causality results, impulse response functions, and variance decompositions characterizing the dynamics of these three variables at the firm level. Findings – For the average firm in the sample, Miller and Modigliani dividend policy irrelevance is not supported, even in the long run; the shocks to dividends do have long-run consequences for investment and vice versa. Dividend changes are an ineffective signal of future earnings in both the short and long-term. The cost of an increased dividend is on average an immediate decrease of $3 in investment for every dollar increase in dividends and the effect is persistent up to six years after the increase in dividends. Research limitations/implications – The firm-level VAR used in the study requires that sample firms have long histories of investment, earnings and dividend data. The study addresses the interaction between dividends and investment and therefore necessitates examining dividend-paying firms. By the nature of the research question, the sample firms will not be representative in all respects to the universe of firms. The most striking difference between the sample and the universe of firms is firm size. As such, the study's conclusions are most applicable to larger, stable, dividend-paying firms. The study is also limited to dividend payout. Alternative payout policies, such as share repurchases, are not considered in this work. Practical implications – In theory, increases in dividends can signal higher future earnings; however, the evidence does not support this hypothesis. When capital markets are constrained or incomplete, increases in dividends come at a cost to investment. Firms should consider alternative methods of signaling future earnings that have less of an impact on investment. Investors should carefully evaluate the possible impact of an increase in dividends on investment and future earnings growth. Originality/value – This study is the first to examine the dynamics of earnings, dividends and investment at a firm level and over such a long sample period. By including the dynamics of earnings, the authors emphasize the potential opportunity costs that increasing dividends has on investment when capital markets are imperfect. The dynamic system also allows the authors to consider long-run effects as well as immediate responses to system shocks.

Original languageEnglish (US)
Pages (from-to)118-136
Number of pages19
JournalManagerial Finance
Volume40
Issue number2
DOIs
StatePublished - Jan 7 2014

Fingerprint

Dividends
Empirical analysis
Vector autoregression
Capital markets
Costs
Variance decomposition
Earnings growth
Interaction
Opportunity cost
Payout policy
Firm size
Investors
Dividend policy
Dividend payout
Share repurchases
Design methodology
Dynamic systems
Impulse response function
Granger causality

All Science Journal Classification (ASJC) codes

  • Finance
  • Strategy and Management

Cite this

An empirical analysis of the dynamic relation among investment, earnings and dividends. / DeFusco, Richard A.; Dunham, Lee M.; Geppert, John.

In: Managerial Finance, Vol. 40, No. 2, 07.01.2014, p. 118-136.

Research output: Contribution to journalArticle

DeFusco, Richard A. ; Dunham, Lee M. ; Geppert, John. / An empirical analysis of the dynamic relation among investment, earnings and dividends. In: Managerial Finance. 2014 ; Vol. 40, No. 2. pp. 118-136.
@article{6434eef07f2244339bb41f3ec91bc608,
title = "An empirical analysis of the dynamic relation among investment, earnings and dividends",
abstract = "Purpose – The purpose of this paper is to examine the dynamic relationships among investment, earnings and dividends for US firms. The sample period is 1950-2006. Design/methodology/approach – The authors use a firm-level vector auto-regression (VAR) framework to examine the firm-level dynamics among investment, earnings and dividends. The firm-level VAR yields Granger causality results, impulse response functions, and variance decompositions characterizing the dynamics of these three variables at the firm level. Findings – For the average firm in the sample, Miller and Modigliani dividend policy irrelevance is not supported, even in the long run; the shocks to dividends do have long-run consequences for investment and vice versa. Dividend changes are an ineffective signal of future earnings in both the short and long-term. The cost of an increased dividend is on average an immediate decrease of $3 in investment for every dollar increase in dividends and the effect is persistent up to six years after the increase in dividends. Research limitations/implications – The firm-level VAR used in the study requires that sample firms have long histories of investment, earnings and dividend data. The study addresses the interaction between dividends and investment and therefore necessitates examining dividend-paying firms. By the nature of the research question, the sample firms will not be representative in all respects to the universe of firms. The most striking difference between the sample and the universe of firms is firm size. As such, the study's conclusions are most applicable to larger, stable, dividend-paying firms. The study is also limited to dividend payout. Alternative payout policies, such as share repurchases, are not considered in this work. Practical implications – In theory, increases in dividends can signal higher future earnings; however, the evidence does not support this hypothesis. When capital markets are constrained or incomplete, increases in dividends come at a cost to investment. Firms should consider alternative methods of signaling future earnings that have less of an impact on investment. Investors should carefully evaluate the possible impact of an increase in dividends on investment and future earnings growth. Originality/value – This study is the first to examine the dynamics of earnings, dividends and investment at a firm level and over such a long sample period. By including the dynamics of earnings, the authors emphasize the potential opportunity costs that increasing dividends has on investment when capital markets are imperfect. The dynamic system also allows the authors to consider long-run effects as well as immediate responses to system shocks.",
author = "DeFusco, {Richard A.} and Dunham, {Lee M.} and John Geppert",
year = "2014",
month = "1",
day = "7",
doi = "10.1108/MF-04-2013-0090",
language = "English (US)",
volume = "40",
pages = "118--136",
journal = "Managerial Finance",
issn = "0307-4358",
publisher = "Emerald Group Publishing Ltd.",
number = "2",

}

TY - JOUR

T1 - An empirical analysis of the dynamic relation among investment, earnings and dividends

AU - DeFusco, Richard A.

AU - Dunham, Lee M.

AU - Geppert, John

PY - 2014/1/7

Y1 - 2014/1/7

N2 - Purpose – The purpose of this paper is to examine the dynamic relationships among investment, earnings and dividends for US firms. The sample period is 1950-2006. Design/methodology/approach – The authors use a firm-level vector auto-regression (VAR) framework to examine the firm-level dynamics among investment, earnings and dividends. The firm-level VAR yields Granger causality results, impulse response functions, and variance decompositions characterizing the dynamics of these three variables at the firm level. Findings – For the average firm in the sample, Miller and Modigliani dividend policy irrelevance is not supported, even in the long run; the shocks to dividends do have long-run consequences for investment and vice versa. Dividend changes are an ineffective signal of future earnings in both the short and long-term. The cost of an increased dividend is on average an immediate decrease of $3 in investment for every dollar increase in dividends and the effect is persistent up to six years after the increase in dividends. Research limitations/implications – The firm-level VAR used in the study requires that sample firms have long histories of investment, earnings and dividend data. The study addresses the interaction between dividends and investment and therefore necessitates examining dividend-paying firms. By the nature of the research question, the sample firms will not be representative in all respects to the universe of firms. The most striking difference between the sample and the universe of firms is firm size. As such, the study's conclusions are most applicable to larger, stable, dividend-paying firms. The study is also limited to dividend payout. Alternative payout policies, such as share repurchases, are not considered in this work. Practical implications – In theory, increases in dividends can signal higher future earnings; however, the evidence does not support this hypothesis. When capital markets are constrained or incomplete, increases in dividends come at a cost to investment. Firms should consider alternative methods of signaling future earnings that have less of an impact on investment. Investors should carefully evaluate the possible impact of an increase in dividends on investment and future earnings growth. Originality/value – This study is the first to examine the dynamics of earnings, dividends and investment at a firm level and over such a long sample period. By including the dynamics of earnings, the authors emphasize the potential opportunity costs that increasing dividends has on investment when capital markets are imperfect. The dynamic system also allows the authors to consider long-run effects as well as immediate responses to system shocks.

AB - Purpose – The purpose of this paper is to examine the dynamic relationships among investment, earnings and dividends for US firms. The sample period is 1950-2006. Design/methodology/approach – The authors use a firm-level vector auto-regression (VAR) framework to examine the firm-level dynamics among investment, earnings and dividends. The firm-level VAR yields Granger causality results, impulse response functions, and variance decompositions characterizing the dynamics of these three variables at the firm level. Findings – For the average firm in the sample, Miller and Modigliani dividend policy irrelevance is not supported, even in the long run; the shocks to dividends do have long-run consequences for investment and vice versa. Dividend changes are an ineffective signal of future earnings in both the short and long-term. The cost of an increased dividend is on average an immediate decrease of $3 in investment for every dollar increase in dividends and the effect is persistent up to six years after the increase in dividends. Research limitations/implications – The firm-level VAR used in the study requires that sample firms have long histories of investment, earnings and dividend data. The study addresses the interaction between dividends and investment and therefore necessitates examining dividend-paying firms. By the nature of the research question, the sample firms will not be representative in all respects to the universe of firms. The most striking difference between the sample and the universe of firms is firm size. As such, the study's conclusions are most applicable to larger, stable, dividend-paying firms. The study is also limited to dividend payout. Alternative payout policies, such as share repurchases, are not considered in this work. Practical implications – In theory, increases in dividends can signal higher future earnings; however, the evidence does not support this hypothesis. When capital markets are constrained or incomplete, increases in dividends come at a cost to investment. Firms should consider alternative methods of signaling future earnings that have less of an impact on investment. Investors should carefully evaluate the possible impact of an increase in dividends on investment and future earnings growth. Originality/value – This study is the first to examine the dynamics of earnings, dividends and investment at a firm level and over such a long sample period. By including the dynamics of earnings, the authors emphasize the potential opportunity costs that increasing dividends has on investment when capital markets are imperfect. The dynamic system also allows the authors to consider long-run effects as well as immediate responses to system shocks.

UR - http://www.scopus.com/inward/record.url?scp=84996551919&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84996551919&partnerID=8YFLogxK

U2 - 10.1108/MF-04-2013-0090

DO - 10.1108/MF-04-2013-0090

M3 - Article

VL - 40

SP - 118

EP - 136

JO - Managerial Finance

JF - Managerial Finance

SN - 0307-4358

IS - 2

ER -