Changes in Cash: Persistence and Pricing Implications

• DOI: http://doi.org/10.1111/1475-679X.12050
• Author: PHILIP B. SHANE,JEFF ZEYUN CHEN
• Published date: 01 June 2014
• Date: 01 June 2014

DOI: 10.1111/1475-679X.12050

Journal of Accounting Research

Vol. 52 No. 3 June 2014

Printed in U.S.A.

Changes in Cash: Persistence and

Pricing Implications

JEFF ZEYUN CHEN

∗AND PHILIP B. SHANE

†

Received 4 August 2012; accepted 9 February 2014

ABSTRACT

This paper decomposes the cash component of earnings and analyzes persis-

tence characteristics and pricing implications of various subcomponents, with

particular attention to changes in cash. Changes in underlying fundamentals

might dictate changes in cash to new optimal levels. Alternatively, subopti-

mal changes in cash might result from agency costs allowing managers’ ac-

tions to diverge from the best interests of shareholders. We predict and find

that both suboptimal increases and decreases in cash bode poorly for future

earnings. In fact, we find that suboptimal increases (decreases) in cash have

less (greater) persistence than any of the earnings components we study, in-

cluding accruals and net distributions to both shareholders and debt holders.

Market efficiency tests indicate that the market severely punishes firms with

suboptimal decreases in cash, but we find no evidence to support the hubris

hypothesis that the market overreacts to the earnings implications of unwar-

ranted increases in cash.

1. Introduction

This paper decomposes the cash component of earnings and analyzes per-

sistence characteristics and pricing implications of various subcomponents,

∗University of Colorado at Boulder; †College of William & Mary.

Accepted by Philip Berger. We appreciate helpful comments by an anonymous referee,

John Hassell, Derek Oler (discussant), Anne Wyatt, and workshop participants at the 2011

American Accounting Association annual meeting, Baruch College, Boston University,College

of William and Mary, Washington and Lee University, and the Universities of Auckland, Col-

orado, Connecticut, Houston, Indiana at Indianapolis, Memphis, New South Wales, Queens-

land, and Virginia. Any errors are our own.

599

Copyright C, University of Chicago on behalf of the Accounting Research Center,2014

600 J.Z.CHEN AND P.B.SHANE

with particular attention to changes in cash.1We test hypotheses that these

persistence characteristics and pricing implications depend on the nature

of the change in cash. Changes in cash might occur in response to changes

in fundamental characteristics of the firm’s operating environment, such as

changing working capital requirements, more or less growth opportunities,

shifts requiring larger or smaller investment in tangible assets, and chang-

ing operating cash flow variability. We refer to these as “normal” changes in

cash, and we estimate normal changes with reference to a prediction model

developed in the finance literature (Bates, Kahle, and Stulz [BKS; 2009]).

Alternatively, suboptimal changes in cash might result from agency costs

allowing self-interested managers to act in ways that diverge from the best

interests of shareholders. Agency-related changes in cash include subopti-

mal cash build-up and suboptimal cash reduction. Unwarranted cash build-

up may lead to management wasting funds on negative net present value

(NPV) projects, pecuniary benefits, empire building, or exorbitant salaries

(Jensen [1986]). Suboptimal cash reduction creates the danger of having

insufficient cash to meet debt obligations, potentially constrains future in-

vestment in positive NPV projects and weakens the firm’s competitiveness in

product markets (Fresard [2010], Tang [2010], Oler and Picconi [2014]).

We refer to suboptimal changes in cash as “abnormal” changes, and we

estimate these abnormal changes as the residual from the model used to

predict normal changes.2

This paper builds on Dechow, Richardson, and Sloan [DRS; 2008]. DRS

disaggregate earnings into accruals and free cash flow. DRS then disaggre-

gate free cash flow into changes in cash, net distributions to debt holders,

and net distributions to stockholders. We further disaggregate the changes

in cash component into normal and abnormal changes as described above.

Since abnormal changes move the cash balance away from optimal levels,

we expect these abnormal changes to bode poorly for future earnings. A

positive component of current earnings that bodes poorly for future earn-

ings lacks persistence and, therefore, we expect abnormal increases in cash

to have relatively low persistence. A negative component of current earn-

ings that bodes poorly for future earnings has relatively high persistence,

so we expect abnormal decreases in cash to have relatively high persistence.

Our evidence is consistent with these expectations. In fact, we find that ab-

normal decreases (increases) in cash have greater (lower) persistence than

1Our definition of cash includes all financial assets.

2As described in BKS [2009], firms make precautionary adjustments to their cash balances

to maintain normal amounts of cash reserves, where normal changes are defined as those

predicted by BKS’s model of fundamental economic forces requiring such changes. BKS treat

abnormal changes in cash (the residual of their model) as reflective of agency problems be-

tween managers and shareholders or between shareholders and debt holders or, more gener-

ally, managers’ suboptimal cash policies.

CHANGES IN CASH 601

any of the earnings components we study, including accruals and net distri-

butions to both shareholders and debt holders.3

DRS find that, on average, the change in cash component of current earn-

ings has relatively low persistence (similar to the accruals component).

However, DRS do not investigate differences in the persistence of positive

and negative changes in cash. Our evidence suggests that lower earnings

follows both positive and negative changes in cash and, therefore, only pos-

itive changes in cash have low persistence, which drives the, on average,

lower persistence of total changes in cash discovered by DRS.

In evaluating market efficiency with respect to persistence characteristics

of earnings components, we find evidence similar to DRS that the market

overreacts to the persistence of accruals and changes in cash. DRS (p. 558)

attribute their finding of similar market overreaction and lower persistence

ofaccrualsandchanges incashto“...hubrisconcerning future investment

opportunities. If managers and investors are overoptimistic about the in-

vestment opportunities of certain firms, these firms invest more capital and

have less sustainable profitability.”

We argue that the hubris interpretation implies that positive abnormal

changes in cash lack persistence and that the market overreacts to the per-

sistence of this component of free cash flow. However, after we decompose

changes in cash into normal and abnormal parts, we fail to reject the null

hypothesis that the market rationally prices positive abnormal changes in

cash. Instead, market overreaction to the persistence of normal and neg-

ative abnormal changes in cash drives the DRS evidence that the market

overreacts, on average, to the persistence of total changes in cash. This re-

sult is contrary to an agency-driven hubris interpretation of market overre-

action to changes in cash, but is consistent with several finance studies that

find no evidence of market inefficiency with respect to the valuation im-

plications of excess cash holdings (e.g., Dittmar and Mahrt-Smith [2007],

Harford, Mansi, and Maxwell [2008], Fr´

esard and Salva [2010]).

Our evidence suggests that the market severely punishes firms that allow

cash balances to fall below optimal levels. With respect to the persistence

of fundamentals-driven changes in cash, we find evidence of market over-

reaction, perhaps due to insufficient adjustment for diminishing returns to

new investment or overestimation of negative income implications of fun-

damentals leading to cash reduction. The significant market overreaction

to the persistence of normal and abnormal decreases in cash might explain

3Persistence refers to the implications of current earnings for future earnings. By mod-

eling earnings as accruals plus free cash flow disaggregated into its change in cash and net

distribution to investor components, and regressing next year’s earnings on the current year’s

earnings components, we isolate the unique implications of current year accruals, changes

in cash, and net distributions to investors for next year’s earnings. In other words, the coef-

ficient on the change in cash variable represents the persistence of changes in cash, holding

constant accruals and net distributions to investors. Throughout the paper, we refer to such

a persistence coefficient as representing the persistence of changes in cash, and we estimate

persistence for both normal and abnormal changes.