Supriyadi—The Predictive Ability of Earnings vs Cash Flow Data to Predict Future Cash Flows
Gadjah Mada International Journal of Business
September 1999, Vol. 1, No. 2, pp. 113—132

THE PREDICTIVE ABILITY OF EARNINGS
VERSUS CASH FLOW DATA TO
PREDICT FUTURE CASH FLOWS:
A FIRM-SPECIFIC ANALYSIS*1
Supriyadi
This study evaluated the value-relevance of accounting information
(earnings and cash flows) in Indonesia to predict a firm’s future operating
cash flows. The predictive usefulness of earnings and cash flows in
association with future cash flows is of interest for three reasons. They
include providing empirical evidence on the relevant accounting information to assess a firm’s future cash flows, information about the behavior and
properties of Indonesian accounting information, and evidence of – or at
least providing a basis for evaluating–the validity of the Indonesian
Accounting Standards Committee (KPSAK) assertion on the usefulness of
accounting information to assess future cash flows.
The study evaluated three cash flow prediction models that employed
cash flow, earnings, and a combination of earnings-cash flow variables.
The models were applied on a firm-specific data set. The data used in this
study were semi-annual data for the 61 sample firms (manufacturing firms)
listed in the Jakarta Stock Exchange (JSX) spanning the years 1990-1997.
The results of this study supported the proposed hypothesis that cash flow
data provided better information to assess a firm’s future cash flows than
* This paper is part of my dissertation from the University of Kentucky, U.S.A. I would like to thank my
committee members, Professors Michael G. Tearney, James A. Knoblett, Mukhtar M. Ali, Steve J. Ott, and,
especially, Stuart B. Keller (Chair), for their thoughtful comments and suggestions. This paper has been
presented in the Simposium Nasional Akuntansi II (SNA II), September, 24-25, 1999 in Malang, Indonesia.
1
The term “cash flows” used in this study refers to cash flows from operations. This term is used
interchangeably with the terms of “cash flows from operations” and “operating cash flows.”

113

Gadjah Mada International Journal of Business, September 1999, Vol. 1, No. 2

earnings data. Since this study employed manufacturing firms only, future
research is necessary to evaluate the robustness of the results to other
populations of firms and/or by using an alternative deflator of earnings and
cash flows, such as consumer price index (CPI) or market value of the firms.
Further extensions of this study include additional refinements of the
prediction models on an industry-specific basis and disaggregating cash
flow variables into operating, investing, and financing components in
order to measure the value-relevance of the statement of cash flows.
Keywords: accounting information; cash flows; earnings; future cash flows

Introduction
In September 1994, the Indonesian
Accounting Standards Committee
(KPSAK) issued a new set of accounting
standards called “Pernyataan Standar
Akuntansi Keuangan (PSAK).” The new
set of accounting standards (PSAK) replaced the 1984 Indonesian accounting
standards called “Prinsip Akuntansi Indonesia (PAI) 1984.” These new standards
were wholly adopted from the International Accounting Standard (IAS) released
by the International Accounting Standard
Committee (IASC). In general, the purpose of issuing PSAK was to harmonize
Indonesian accounting standards with IAS
and to respond to world globalization since
it was believed by KPSAK that IAS was
acceptable worldwide [Indonesian Accountants’ Association (IAI) in PSAK
1995, Sambutan Ketua Umum].
The primary objective of accounting
information stated in the Framework for
the Preparation and Presentation of Financial Statements in Indonesia is to provide
useful information for assessing the
amounts, timing, and uncertainty of pro-

spective cash flows to the firm [para. 14,
and SFAC #1]. The FASB in SFAC #1
asserts that “the objectives of financial
reporting are not immutable—they are affected by the economic, legal, political,
and social environment in which financial
reporting takes place.”
Some studies that evaluate the effects
of economic and social environmental factors on accounting standards find evidence
supporting this assertion (Cooke and
Wallace 1990; Ndubizu 1992; Doupnik
and Salter 1995). They find that cultural
and economic differences result in differences in how investors and other financial
statement users value the same piece of
accounting information. In other words, it
is highly possible that the same reporting
objective could be achieved using different types of accounting information due to
different economic and environmental factors. Hence, empirical results found in the
U.S. that support the FASB’s assertion
that earnings provides better information
to assess future cash flows than cash flow
itself, may not apply to Indonesia.2 All of
these conditions lead to the following research questions:

2
Since Ball and Brown’s (1968) study in the evaluation of accounting income numbers, there have been
many studies that evaluate the usefulness or information content of earnings (Beaver and Dukes 1972; Beaver
et al. 1982; Hughes and Ricks 1987; Lev 1989; Kothari and Zimmerman 1995; among others). Under the CAPM,
these studies implicitly support the FASB assertion that earnings provide better information to assess future cash
flows than cash flow itself.

114

Supriyadi—The Predictive Ability of Earnings vs Cash Flow Data to Predict Future Cash Flows

What type of accounting information in Indonesia is most useful to
investors in assessing a firm’s future cash flows?
Does earnings provide more information than cash flows to predict
an Indonesian firm’s future cash
flows?
This study evaluated the value-relevance of accounting information (earnings and cash flows) in Indonesia to predict a firm’s future operating cash flows.
The predictive usefulness of accounting
information in association with future cash
flows is of interest for three reasons. They
include providing empirical evidence on
the relevant accounting information to assess a firm’s future cash flows, information about the behavior and properties of
Indonesian accounting information, and
evidence of – or at least providing a basis
for evaluating – the validity of the Indonesian Accounting Standards Committee
(KPSAK) assertion on the usefulness of
accounting information to assess future
cash flows.
The study evaluated three cash flow
prediction models that employed earnings
and cash flow variables extracted from
Indonesian companies’ financial statements. The models were applied on a firmspecific regression of the financial data
from selected Indonesian firms. The data
used in this study were semi-annual data
for the 61 sample firms (manufacturing
firms) listed in the Jakarta Stock Exchange
(JSX) spanning the years 1990-1996. Three
sets of accounting information that had
been widely used in the U.S. studies to
evaluate the value-relevance of accounting information to assess a firm’s future
cash flows (the FASB’s assertion) were
used in this study. They include earnings,
cash flows, and a combination of earnings

and cash flows (Greenberg et al. 1986;
Finger 1994; Lorek and Willinger 1996;
among others).
The results of this study supported
the proposed hypothesis that cash flow
data provided better information to assess
a firm’s future cash flows than earnings
data. The result indicated that cash flows
significantly provided extra information
over and above earnings for 59 percent of
the sample firms while earnings was significantly found in 25 percent of the sample
firms. In terms of the predictive ability,
cash flow and earnings combined exhibited the lowest MAPE (Mean Absolute
Percentage Error) and significantly dominated the earnings model but not the cash
flow model. Furthermore, this result also
indicated that the effects of seasonal factors on future cash flows were more severe
than those of adjacent factors.
Since this study employed manufacturing firms only, future research is necessary to evaluate the robustness of the results to other populations of firms and/or
by using an alternative deflator of earnings
and cash flows, such as consumer price
index (CPI) or market value of the firms.
Further extensions of this study include
additional refinements of the prediction
models on an industry-specific basis and
disaggregating cash flow variables into
operating, investing, and financing components in order to measure the valuerelevance of the statement of cash flows.
The remaining part of this paper is
organized as follows. Section II describes
the prior studies on the predictive ability
of earnings and cash flows. Section III
discusses hypothesis development and
Section IV describes cash flow forecasting models. Sections V and VI discuss the
empirical results and hypothesis tests, respectively. Finally, Section VII concludes
the paper.
115

Gadjah Mada International Journal of Business, September 1999, Vol. 1, No. 2

Prior Studies
The FASB contends that accrual earnings information provides better information for prediction of future cash flows
than cash flow information itself (SFAC
#1, para. 44). Empirical studies in the
United States that test the FASB’s assertion fall into two categories, namely, studies that examine capital market effects of
accounting information and those that directly examine the ability of accounting
information to predict future cash flows
(or surrogates of future cash flows).3 However, these two categories always deal
with the predicted values. The former
employs the predicted values, as a proxy
for market expectations, in order to measure the unexpected values of an accounting variable under consideration, while
the latter uses the predicted values to evaluate the predictive ability of an accounting
variable(s).
The basic assumption underlying research in the information content of earnings and earnings forecasting literature is
that accounting earnings is a good surrogate of future cash flows. However, since
the accounting earnings incorporates accounting accruals and excludes investment activities, it is only under certain
(extreme) conditions that expected accounting earnings equal future cash flows
(Watts and Zimmerman 1986). Furthermore, management’s discretion in choosing accrual methods is also deemed to
decrease reliability of earnings as a surrogate of operating cash flows (Healy 1985;
Dechow et al. 1995; Guay et al 1996).
Nevertheless, studies in the information
content of cash flow data provide incon3

sistent evidence (Neill et al. 1991). Early
studies that employ simple cash flow measures (earnings plus depreciation) fail to
detect information content of cash flows,
while later studies that measure information content of operating cash flows find a
significant association between cash flows
and stock prices (Rayburn 1986; Bowen et
al. 1987; Wilson 1987; Livnat and Zarowin
1990; among others).
Few studies on the usefulness of accounting information have been conducted
in Indonesia. Most of these studies measure the information content of earnings in
association with stock prices (capital market research) on the Jakarta Stock Exchange (JSX). These studies provide inconsistent evidence regarding the value of
earnings in setting stock prices on the JSX
(Setiawati 1995; Husnan et al. 1996; Hanafi
1997). Three possible reasons may be put
forward for this inconclusive evidence,
namely, inappropriate research methodology, an inefficient market, and valueless
accounting information. As far as the research methodology, these studies may
fail to appropriately specify window
events, select variable measurement, or
select a proxy of the market.
Husnan (1992) indicates that the efficiency of the JSX was still in weak form,
but it has increased since 1990. Also, using 1991-1996 data, Sugiyanto (1998) finds
consistent results with Husnan’s (1992)
that the Indonesian stock market was not
efficient, at least in the semi-strong form.
This finding, together with the fact that the
JSX activities are still considered “thin
activities,” shows that the assumption of a
semi-strong efficient market is violated.

Based on the assumption that the financial statements’ users use the best data available, where best is
defined as most accurate, Brown (1993) states that the predictive ability and association studies are two sides
of the same coin. This means that examining the value-relevance of accounting information using the association
and predictive ability approaches should result in a similar conclusion (Brown 1993).

116

Supriyadi—The Predictive Ability of Earnings vs Cash Flow Data to Predict Future Cash Flows

Some possible reasons explain the
lack of value-relevance in accounting information that may have cause the inconsistent results found in these early studies.
For example, since the inflation rate in
Indonesia is relatively high, it could decrease the value of historical cost-based
earnings as a measure of a firm’s performance (PSAK does not have an accounting standard for inflation or changing
prices). Moreover, the lack of disclosures
required by the Indonesian accounting standards (Sutton 1997; Saudagaran and Diga
1997) should also reduce the quality of
accrual earnings.
Machfoedz (1994) measures the usefulness of accounting information in Indonesia based on 84 manufacturing firms
listed on the JSX for the years of 19891992. Even though using firms listed on
the JSX, his study does not directly measure the effect of accounting information
on stock prices. Machfoedz defines the
usefulness of accounting information in
terms of an association between financial
ratios (extracted from financial statements)
and future earnings changes. He asserts
that financial ratios significantly associate
with a year ahead of earnings changes, but
the association significantly decreases with
two year ahead of earnings changes.
Three things can be noted from
Machfoedz’s study. First, since there is
inconsistent evidence regarding the value
of earnings in the JSX, relating the financial ratios and earnings may not reflect the
way investors process accounting information. Second, the use of earnings
changes as a benchmark of information
usefulness assumes that information users
(investors) follow a naive random walk
model. Since there is no prior research in
user response to accounting information

in Indonesia, further study to validate this
assumption is needed. Finally, under the
EMH and CAPM theories, his study does
not provide evidence concerning the valuerelevance of accounting information in
asset valuation. In terms of the usefulness
of accounting information as stated in the
PSAK, his study does not directly provide
evidence on the ability of accounting information to assess a firm’s future cash
flows.
Parawiyati and Baridwan (1998)
evaluate the predictive ability of earnings
and cash flows to predict future earnings
and cash flows. They measure the predictive ability of earnings versus cash flows
based on the degree of association (using
linear regressions) between earnings (cash
flows) and future earnings (cash flows).
However, this study does not measure the
real forecast errors of their predictive model
used in their study.

Hypothesis Development
Accrual earnings are based on two
important accounting principles, namely,
the revenue recognition and matching principles. The revenue recognition principle
requires a firm to recognize revenues when
it has performed all, or a substantial portion of, services that have to be rendered,
and cash receipts from the transaction are
reasonably certain. The matching principle requires a firm to recognize all expenses associated with revenues in the
same period in which the revenues have
been recognized. Since this accrual process is deemed to mitigate the timing and
matching problems inherent in cash flows,
it is believed that earnings more closely
represent a firm’s performance (Dechow
1995).4 Hence, studies in the usefulness of

4
Since cash flows are presumed to be the main focus of most investors, firm performance here shows the
ability (and the risk) of a firm to generate current and future cash flows.

117

Gadjah Mada International Journal of Business, September 1999, Vol. 1, No. 2

earnings (accrual based earnings) for investment decisions are based on the hypothesis that earnings is a good surrogate
for a firm’s future cash flows (Beaver
1968; Ball and Brown 1968; Easton 1985;
among others).
However, since management usually
has some discretion over the recognition
of accruals, accrual earnings may be a
noisy measure of the firm’s performance.
While the accounting conventions of objectivity and verifiability may limit the
flexibility of management’s discretion
(Watts and Zimmerman 1986), they also
may reduce the ability of earnings to reflect a firm’s performance. Hence, the
accrual process may result in a trade-off of
information quality between reliability and
relevance (SFAC #2 para 90; Ball 1989).
This condition leads to the possibility that
earnings may become less reliable as a
measure of a firm’s performance than cash
flow data (Dechow 1994). Dechow et al.
(1995) examine a sample of firms to evaluate earnings management. They find that
accruals that reverse in the subsequent
year reflect earning manipulations.
Moreover, Rayburn (1986), Wilson
(1986 and 1987), Bowen et al. (1987), and
Livnat and Zarowin (1990) also document
that cash flow data provide incremental
information content beyond earnings. They
measure the incremental information content based on the association between unexpected cash flows and stock returns.
These findings are consistent with the argument supporting the issuance of standards in cash flow statement (SFAS #95),
cash flow information adds information
contained in earnings (Wolk and Tearney
1997). Nevertheless, these studies do not
provide evidence of the relative ability of
earnings versus cash flows to assess a
firm’s future cash flows.

118

Sloan (1996) indicates that the persistence of earnings performance depends
on the magnitudes of the cash and accruals
components of earnings. The higher the
accrual component, the lower the persistence of earnings performance, while the
higher the cash flow component, the higher
the persistence of earnings performance.
Furthermore, the results of market-based
studies of the incremental information
content of earnings components indicate
that the market reacts differently to various earning components (Fairfield et al.
1996; Strong and Walker 1993; Ohlson
and Penman 1992).
The differences in cultural and economic factors that exist between the U.S.
and Indonesia may cause different values
to be assigned to the same piece of information. One example of economic factors
that affect the value of accounting information is the impact of inflation rate on
historical-based costing of accounting information. The inflation rate is a proxy of
an economic condition that has been widely
used in economic studies. Ndubizu (1992)
and Doupnik and Salter (1995) find significant evidence supporting the dampening effect of inflation rates on the value of
accounting information in various countries. The relatively high inflation rate
(double digit) in Indonesia during this
decade could decrease the benefits of a
historical cost-based system of accounting. Accounting information may become
less relevant under historical-based costing. High inflation rates will cause historical cost-based earnings to be overstated.
Consequently, its predictive value decreases (Wolk and Tearney 1997). Therefore, cash flow data that are free from
inflation effects should provide a better
indication of future cash flows than earnings.

Supriyadi—The Predictive Ability of Earnings vs Cash Flow Data to Predict Future Cash Flows

The differences in the accounting
standards between the U.S. and Indonesia
should also affect the usefulness of accounting information. The lack of a core
set of standards and comparability combined with differing interpretations of IAS
(PSAK) (Sutton 1997) may decrease the
usefulness of accounting information. For
instance, management’s discretion over
the accrual process may not be fully disclosed to the public through the financial
statements. Consequently, accrual earnings becomes a noisy measure of a firm’s
performance. Therefore, cash flow data
that are free from the effects of
management’s discretion over accruals
should provide a better indication of future
cash flows than earnings. This leads to the
following null hypothesis:
H01: Cash flow data do not outperform
earnings as the predictor of a firm’s
future cash flows.
To examine whether earnings provide incremental information to assess
future cash flows in the presence of cash
flow variables, Finger (1994) combines
earnings and cash flows as predictors of
future cash flows. She finds evidence that
earnings adds information to cash flows.
Although the ability of earnings to assess
future cash flows may be less relevant in
Indonesia, earnings may still add relevant
information. Therefore, combining earnings and cash flows in the forecasting
model may increase the power of the model.
This leads to the following null hypotheses:
H02: A combination of cash flow data and
earnings does not outperform either
earnings or cash flow data alone as

the predictor of a firm’s future cash
flows.
H03: When the forecasting model employs
earnings and cash flow data as the
explanatory variables, cash flow
(earnings) data do not provide extra
information over and above earnings (cash flow) data.

Cash Flow Forecasting Models
Testing the ability of earnings versus
cash flows to predict future cash flows (the
FASB’s contention) requires models that
employ either earnings or cash flows alone
as a predictor of cash flows (Bowen et al.
1986; Greenberg et al 1986; Murdoch and
Krause 1989, 1990; Finger 1994). Furthermore, to examine whether earnings
provide incremental information to assess
future cash flows in the presence of cash
flow variables, Finger (1994) combines
earnings and cash flows as predictors of
future cash flows.5 The models employed
in this study were developed based on
these earlier models. In summary, there
are three cash flow forecasting models
used in this study. These models are presented in Equation 1.
Equation 1: Cash Flow Forecasting
Models
CFOt = a + b1CFOt-1 + b2CFOt-2 +
b2D + et
CFOt = a + b1EAt-1 + b2EAt-2 + b2D +
et
CFOt = a + b1CFOt-1 + b2CFOt-2 +
g1EAt-1 + g2EAt-2 + b2D + et

5

The argument supporting the issuance of a standard requiring a Cash Flow Statement (SFAS #95) is that
cash flow information adds to the information contained in earnings. This means that cash flow information does
not eliminate the usefulness of earnings but it adds information to earnings. Therefore, Finger’s model (1994)
is consistent with the FASB assertion.

119

Gadjah Mada International Journal of Business, September 1999, Vol. 1, No. 2

Where,
CFOt = cash flows from operations
CFOt-i = the lagged values of cash flows
from operations
EAt-i = the lagged values of earnings.
t
= time variable measured semi-annually
D
= dummy variable to proxy different accounting standards, D = 1
for 1995-1997 financial data and
D = 0, otherwise.
In this study, the lagged variables t-1
and t-2 (for semi-annual data) are used in
Models (1), (2), and (3) to capture adjacent
effects (t-1) and seasonal effects (t-2) of
the independent variables on the CFO.
In this study, all variables used in the
model will be extracted either from the
balance sheet and income statement, or the
statement of changes in the financial position or cash flows. Specifically, the variables are as follows (Neill et al. 1990;
Wilson 1986, 1987; Livnat and Zarowin
1990).
● Cash Flows from Operations = Earnings before extraordinary items ± Noncurrent Accruals ± Current Accruals
● Current Accruals = the net changes in
the working capital accounts except for
changes in cash, marketable securities,
and short-term debt.

● Noncurrent Accruals = Depreciation

and Amortization + Adjustment for
other noncurrent accruals.
● Earnings = Earnings before extraordinary items
The values of the variables used in
the models are scaled by the total assets of
the related firm at the beginning of the
period (semi-annual period). The purpose
of scaling the variables is to control for
heteroscedasticity which may arise due to
the level of economic activities that varies
over time. Scaling by total assets transforms measures of different time periods
into comparable measurements.

Empirical Results
Sample and Descriptive Statistics
The data used in this study were semiannual data for the 61 sample firms (manufacturing firms) listed in the JSX spanning
the years 1990-1996. Initially, there are 72
firms which had all the data needed. From
these firms, eleven firms were dropped as
the result of checking for outliers.
This study performed an analysis that
used the 1990-96 period data set that consisted of semi-annual data from the first
semi-annual reporting period of 1990 to

Table 1. Descriptive Statistics for Variables Used in the Prediction Models Distributions of Pooled Cross-sectional Data for 61 Firms (Semi-annual Data 19901996)
Variable

N

Mean

S.D.

CFO
EA

793 0.077 0.133
793 0.081 0.107

Minimum Maximum Median
-0.464
0.014

0.449
0.763

0.072
0.054

Lower
Upper
Quartile Quartile
0.016
0.043

0.132
0.078

Note:
Variable definitions: CFO = cash flows from operations; EA = net income before extraordinary
items.
The variables had each been deflated by total assets at the beginning of each period.
120

Supriyadi—The Predictive Ability of Earnings vs Cash Flow Data to Predict Future Cash Flows

Table 2. Pearson Correlations between Dependent Variable and Independent Variables Distribution of Correlations for 61 Firms (Semi-annual Data, 19901996)
CFOt

CFOt-1

CFOt-2

EA t-1

EA t-2

Mean
S.D.
Maximum
Minimum
Median
Lower Quartile
Upper Quartile

-0.17
0.34
0.74
-0.78
-0.19
-0.44
-0.01

0.21
0.47
0.83
-0.69
0.17
-0.23
0.61

0.08
0.32
0.80
-0.55
0.13
-0.17
0.31

0.08
0.29
0.74
-0.55
0.13
-0.20
0.29

the second semi-annual reporting period
of 1996 (12 observations for each firm).
The results of the regression (coefficient
parameters) on 12 observations were then
used to forecast cash flows from operations for the first semi-annual reporting
period of 1997.
Table 1 presents the descriptive statistics of data for variables used in the
models. The values of the variables have
each been deflated by the total assets at the
beginning of each period. Moreover, the
distribution of contemporaneous correlation between cash flows from operations
(the dependent variable for all forecasting
models) and the independent variables is
presented in Table 2.

Regression Results
Three regression models were used
to analyze the data. Cash flows from operations were the dependent variable for
all regression models. The first model was
called the cash flow model, while the second model was called the earnings model
and the third model was called the earnings-cash flow model.
Since the earnings-cash flow model
consists of cash flows and earnings as the

regressors, analyzing this model is expected to provide an answer on the valuerelevance of cash flows versus earnings in
predicting a firm’s future cash flows. Furthermore, analysis of this model can also
be used to test whether earnings (cash
flows) provide incremental predictive ability in the presence of cash flows (earnings). Hence, further evaluation of the
cash flow and earnings models is expected
to strengthen the results of analyzing the
earnings-cash flow model. Thus, in the
following parts, the results of regressing
the earnings-cash flow model are presented
first.
The Earnings-Cash Flow Model
(Model 3)
Table 3 summarizes the results of
estimating the earnings-cash flow model
based on firm-specific regressions. The
results provide an early indication that
cash flows from operations may outperform earnings as a predictor of a firm’s
future cash flows. The results indicate that
the model is robust for 12 (20%) and 22
(36%) firms at the 0.05 and 0.10. The
adjusted R2 ranges from the lowest of 0.45 to the highest of 0.74 with the mean of
121

Gadjah Mada International Journal of Business, September 1999, Vol. 1, No. 2

Table 3. Regression Results - Distribution of Adjusted R2 and Other Statistical
Measures for 61 Firms (Semi-annual Data, 1990-1996)
Model 3: CFOt = b0 + b1 CFO t-1 + b2 CFOt-2 + b3 EAt-1 + b4 EAt-2 + b4 D +e
Mean Adjusted R2
Maximum Adjusted R2
Minimum Adjusted R2
Median Adjusted R2
Lower Quartile Adjusted R2
Upper Quartile Adjusted R2

0.34
0.74
-0.45
0.40
0.12
0.43

Number of
Firms with bi •0

Intercp.

CFO t-1

at the 0.05 level
at the 0.10 level

14 (23%) 16 (26%)
20 (33%) 24 (39%)

25 (41%) 3 (5%)
8 (13%) 11 (18%)
34 (56%) 11 (18%) 13 (21%) 22 (36%)

0.34 and median of 0.40. For the coefficient parameters of cash flows from operations, lags 1 and 2 cash flows from operations are significant for 16 (26%) and 24
(39%) firms at the 0.05 level, and 25
(41%) and 34 (56%) firms at the 0.10
level. In comparison, lags 1 and 2 earnings
are significant for 3 (5%) and 11 (18%)
firms at the 0.05 level, and 8 (13%) and 13
(21%) firms at the 0.10 level. Finally, the
dummy variable of changes in accounting
standards is significant for 11 (18%) and
for 13 (21%) firms at the 0.05 and 0.10
levels.

0.05 and 0.10, lag 2 cash flows from operations is significant for 28 (46%) and 35
(57%) firms, while lag 1 cash flows from
operations is significant for 17 (28%) and
27 (44%) firms. The dummy variable is
significant for 28% and 33% of the firms at
the 0.05 and 0.10 level, respectively.

Number of Firms with Significant
F-values at the 0.05 Level

12 (20%)

Number of Firms with Significant
F-values at the 0.10 Level

22 (36%)

The Cash Flow Model (Model 1)
Table 4 summarizes the results of
estimating the cash flow model with lags 1
and 2 cash flows from operations and a
dummy as independent variables. The results of firm-specific regressions indicate
that 20 (33%) and 28 (46%) of the firms’
F values are significant at the 0.05 and
0.10. The mean adjusted R2 is 0.31 with
the highest adjusted R2 of 0.62, the lowest
of -0.16, and the median of 0.29. In terms
of significant coefficient parameters at
122

CFO t-2

EA t-1

EA t-2

D

The Earnings Model (Model 2)
Table 5 presents the results of estimating the earnings model with lags 1 and
2 earnings and a dummy as the independent variables. The results of firm-specific
regressions indicate that 8 (13%) and 11
(18%) of the firms exhibit significant overall regression at the 0.05 and 0.10. The
mean adjusted R2 is 0.12 with the highest
adjusted R2 of 0.55, the lowest of -0.33,
and the median of 0.03. In terms of significant coefficient parameters at 0.05 and
0.10, lag 2 earnings is significant for 10
(17%) and 19 (31%) firms, while lag 1
earnings is significant for 7 (12%) and 13
(21%) firms. The dummy variable is significant for 18% and 21% of the firms at
the 0.05 and 0.10 level, respectively.

Supriyadi—The Predictive Ability of Earnings vs Cash Flow Data to Predict Future Cash Flows

Table 4. Regression Results - Distribution of Adjusted R2 and Other Statistical
Measures for 61 Firms (Semi-annual Data, 1990-1996)
Model 1: CFOt = b0 + b1 CFO t-1 + b2 CFOt-2 + b3 D + e
Mean Adjusted R2
0.31
2
Maximum Adjusted R
0.62
Minimum Adjusted R2
-0.16
Median Adjusted R2
0.29
2
Lower Quartile Adjusted R 0.09
Upper Quartile Adjusted R2 0.37

Number of Firms with Significant
F-values at the 0.05 Level

20 (33%)

Number of Firms with Significant
F-values at the 0.10 Level

28 (46%)

Number of
Firms with bi • 0

Intercept

CFO t-1

CFO t-2

D

at the 0.05 level
at the 0.10 level

31 (51%)
36 (59%)

17 (28%)
27 (44%)

28 (46%)
35 (57%)

17 (28%)
20 (33%)

Table 5. Regression Results - Distribution of Adjusted R2 and Other Statistical
Measures for 61 Firms (Semi-annual Data, 1990-1996)
Model 2: CFOt = b0 + b1 EA t-1 + b2 EAt-2 + b3 D + e
Mean Adjusted R2
Maximum Adjusted R2
Minimum Adjusted R2
Median Adjusted R2
Lower Quartile Adjusted R2
Upper Quartile Adjusted R2

0.12
0.55
-0.33
0.03
-0.16
0.11

Number of Firms with Significant
F-values at the 0.05 Level

8 (13%)

Number of Firms with Significant
F-values at the 0.10 Level

11 (18%)

Number of
Firms with bI • 0

Intercept

EA t-1

EA t-2

D

at the 0.05 level
at the 0.10 level

10 (17%)
18 (30%)

7 (12%)
13 (21%)

10 (17%)
19 (31%)

11 (18%)
13 (21%)

Comparing the regression results of
the cash flow, earnings, and earnings-cash
flow models provides an early indication
that cash flows from operations outperform earnings as the predictor of future
cash flows. Furthermore, the marginal increase in adjusted R2 for the earnings-cash
flow model compared to the cash flow

model also indicates that earnings adds
little to the overall explanatory power. The
average adjusted R2 for the cash flow model
is 0.31, the earnings model is 0.12, and the
earnings-cash flow model is 0.34. The
dummy variable is only significant for
some firms. This indicates that the new
accounting standards may not have the
same effects for all sample firms.
123

Gadjah Mada International Journal of Business, September 1999, Vol. 1, No. 2

Table 6. Distribution of Absolute Percentage Errors (APE)d
Model 1

Model 2

Model 3

Mean (MAPE)
S.D.
Maximum
Minimum
Median
Lower Quartile
Upper Quartile

0.64
0.37
1.98
0.01
0.66
0.37
0.80

0.82
0.59
2.88
0.09
0.60
0.45
1.03

0.62
0.40
2.73
0.13
0.58
0.37
0.79

Normal Distribution
Test (W)

0.96**

0.83*

0.82*

d

APE is defined as the absolute value of (Actual CFO –Forecast CFO)/Actual CFO.
* significant at 0.01; ** significant at 0.05
Predictive Ability Results
One-step-ahead semi-annual cash
flow predictions are generated in an ex
ante fashion (out-of-sample forecast) for
the five cash flow prediction models. The
results of the regressions are used to predict cash flows from operations for the
first semi-annual reporting period of 1997.
In terms of predictive ability, one error
metric, absolute percentage error (APE) is
computed. APE is defined as the absolute
value of the difference between the actual
and forecast value divided by the actual
value of cash flows from operations. Predictive ability of one model compared to
other models (among models) is determined using mean absolute percentage
error (MAPE).
Table 6 summarizes the distribution
of absolute percentage error from three
regression models. Normal distribution
tests indicate that APEs resulting from
each of the three models are normally
distributed. The results of the tests show
that the probabilities of normal distribution for APEs are statistically significant
124

at 0.01 for the earnings and earnings-cash
flow models, and at 0.05 for the cash flow
model. Therefore, a t- or an F-test is employed to test the difference between two
means or among means of APEs.

Hypothesis Tests
Hypotheses 1 and 2 were tested by
comparing MAPEs resulting from all three
models. Hypothesis 3 is tested by evaluating the earnings-cash flow model to determine whether subsets of coefficient parameters of cash flows and of earnings are
significantly different from zero. In terms
of variable contribution into the forecasting model, Hypothesis 3 evaluates the
usefulness of cash flow data versus earnings when they are used together in the
forecasting model. Therefore, since Hypotheses 1 and 2 test the predictive ability
of either cash flow or earnings data alone,
it is expected that testing Hypotheses 1, 2,
and 3 will result in a similar conclusion.
Specifically, since testing Hypothesis 3
indirectly can also test Hypotheses 1 and

Supriyadi—The Predictive Ability of Earnings vs Cash Flow Data to Predict Future Cash Flows

2, further tests of Hypotheses 1 and 2 is
purported to strengthen the results of testing Hypothesis 3. Hence, in the following,
the result of testing Hypothesis 3 is presented first.
An F test (“partial” F test) is employed to measure whether subset of coefficients of cash flows (earnings) in the
model is equal to zero. An F-statistic to test
Hypothesis 3 is calculated based on the
residual sum of squares resulting from the
full model (the original earnings-cash flow
model) and the reduced model. The reduced model is constructed based on the
earnings-cash flow model with the variables stated in the null hypothesis removed from the model. Specifically, testing the contribution of cash flows (earnings) is performed by comparing the earnings-cash flow model and the earnings

model (the cash flow model) to determine
whether the earnings-cash flow model is
significantly better than the earnings model
(the cash flow model).
The distribution of F-values to test
Hypothesis 3 (for both firm-specific and
pooled cross-sectional regressions) is presented in Table 7. These measures are used
to test Hypothesis 3, whether cash flows
(earnings) provide extra information over
and above earnings (cash flows). Since
there are two cash flow and earnings variables used in Model 3 (lags 1 and 2 for each
variable), a test of Hypothesis 3 is performed by evaluating the contribution of
lags 1 and 2 cash flows or earnings simultaneously.
The results of the firm-specific regressions indicate that cash flows are superior to earnings. For the period under

Table 7. Test of Hypothesis 3 - Distribution of F-value for Testing H0 as Stated Below
Model 3: CFOt = b0 + b1 CFO t-1 + b2 CFOt-2 + b3 EAt-1 + b4 EAt-2 + b5 D + e
The reduced models of Model 3 for H0 of
b1 = b2 = 0: CFOt = b0 + b3 EAt-1 + b4 EAt-2 + b5 D + e (Model 2)
b3 = b4 = 0: CFOt = b0 + b1 CFO t-1 + b2 CFOt-2 + b5 D + e (Model 1)

Mean F-valuec
Maximum F-value
Minimum F-value
Median F-value
Lower Quartile F-value
Upper Quartile F-value
Number of Firms with Significant F at 0.05
Number of Firms with Significant F at 0.10
c

b1 = b2 = 0

b3 = b4 = 0

3.98
15.83
0.22
3.46
0.85
5.06
20 (33%)
36 (59%)

1.64
6.45
0.01
0.66
0.37
2.97
5 (8%)
15 (25%)

F-value is calculated using the following formula:

F = ((SSER – SSEF)/(K – L + 1))/(SSEF/(n – K – 1)), where SSER = sum of squared errors from the
reduced model; SSEF = sum of squared errors from the full model;
K = the number of the independent variables used in the full model, L = the number of the
independent variables used in the reduced model (including the intercept), and
n = the number of observations.
125

Gadjah Mada International Journal of Business, September 1999, Vol. 1, No. 2

examination, 20 (33%) and 36 (59%) firms
reject the null hypothesis that cash flows
do not provide extra information over and
above earnings at the 0.05 and 0.10 levels,
respectively. In comparison, only 5 (8%)
and 15 (25%) firms reject the null hypothesis that earnings does not provide extra
information over and above cash flows at
the 0.05 and 0.10 levels, respectively.
Hypothesis 1 was also constructed to
address the first specific research question. This hypothesis tests the predictive
ability of cash flow versus earnings data

when each is employed alone in the model.
Hypothesis 2 is developed to measure the
relative predictive abilities of cash flow
and earnings data when they are used
together. The difference between Hypotheses 2 and 3 is that while Hypothesis 3
evaluates the contribution of each cash
flow versus earnings, Hypothesis 2 examines the predictive ability of both cash
flows and earnings. Hence, it was expected that the results of testing Hypotheses 1 and 2 would be consistent with the
results of Hypothesis 3 tests.

Table 8. Tests of Hypotheses 1 and 2 Comparison of Mean Absolute Percentage
Errors (MAPE) of CFO Predictions on June 30, 1997
APE on June 30, 1997
MAPE
S.D.
Maximum
Minimum
Median
Lower Quartile
Upper Quartile

Model 1

Model 2

Model 3

0.64
0.37
1.98
0.01
0.66
0.37
0.80

0.82
0.59
2.88
0.09
0.60
0.45
1.03

0.62
0.40
2.73
0.13
0.58
0.37
0.79

6/30/97
F-value

Lower APEh

Hypothesis Tests of MAPE Differences

t-value
1. Models 1 and 2
2. Models 1 and 3
3. Models 2 and 3
4. Models 1, 2, and 3
h

2.36**
0.23
2.34**

36 (59%)
31 (51%)
36 (59%)
3.53**

Comparison of the magnitude of APE between two models.
The numbers appearing in this column show the number of firms (percentage) from the lower
MAPE model which have lower APE.
For example, in the comparison of Models 1 and 2, Model 1 has a lower MAPE. Hence, 36 (59%)
in the row of Models 1 and 2 means that 36 firms (59%) out of the sample firms (from a total of
61 firms) has lower APE under the prediction of Model 1 than that of Model 2.
** significant at 0.05

126

Supriyadi—The Predictive Ability of Earnings vs Cash Flow Data to Predict Future Cash Flows

Hypothesis 1 is tested by comparing
MAPEs from the cash flow and earnings
models, while Hypothesis 2 is tested by
comparing MAPEs from the cash flow,
earnings, and earnings-cash flow models.
A t-test is employed to statistically measure MAPE differences resulting from two
models and an F-test is used to statistically
measure MAPE differences resulting from
more than two models. Table 8 presents
the results of MAPE comparisons including the statistical tests for Hypotheses 1
and 2.
The result of testing Hypothesis 1
that the MAPE of the cash flow model
(0.64) is significantly different from that
of the earnings model(0.82) at the 0.05
level (t-value = 2.36). Consistently, looking at the APE for each firm, the cash flow
model outperforms (has lower APE than)
the earnings model in 36 (59%) firms,
while the earnings model outperforms (has
lower APE than) the cash flow model in 25
(41%) firms. The result supports the expectation as stated in Hypothesis 1, cash
flow data provide better information to
forecast future cash flows than earnings.
Hypothesis 2 is tested by comparing
MAPEs from all three models. Pairwise ttest comparisons indicate that the cash
flow model (0.64) is not significantly different from that of the earnings-cash flow
model (0.62), while the earnings model
(0.82) is significantly different from that
of the earnings-cash flow model (0.62) at
the 0.05 level (t-value = 2.34). Furthermore, 36 (59%) firms have a lower APE as
measured by the earnings-cash flow model
than when only cash flow is included. On
the other side, 25 (41%) firms have a lower
APE under the earnings model than the
earnings-cash flow model. As expected,
the results of comparing the three models
simultaneously also support the expectation embodied in Hypothesis 2. The

MAPEs from these three models are significantly different at 0.05 (F-value = 3.53)
at which the earnings model has the highest MAPE. This result indicates that earnings adds little to the ability of cash flows
to predict future cash flows.

Conclusion
In general, empirical analyses performed in this study provided results supporting the proposed hypothesis that cash
flow data provided better information to
assess future cash flows than earnings
data. The results of testing Hypothesis 3
on a firm-specific level indicated that cash
flows significantly provided extra information over and above earnings for 59
percent of the sample firms while earnings
was significantly found in 25 percent of
the sample firms. The results of testing
Hypotheses 1 and 2 provided supporting
evidence on the results of testing Hypothesis 3.
Furthermore, lag 2 cash flows significantly outperformed lag 2 earnings,
while lag 1 cash flows slightly dominated
lag 1 earnings in three regression models.
This result indicated that the effects of
seasonal factors on future cash flows were
more severe than those of adjacent factors.
In terms of the predictive ability, cash flow
and earnings combined exhibited the lowest MAPE and significantly dominated the
earnings model but not the cash flow model.
This study contributes to the literature of the usefulness of accounting information study in three folds. First, this
study examines the relevant variables that
can be used by investors in forming expectations of a firm’s future cash flows. A
good expectation model is very important
to a growing interest in capital-marketbased research in Indonesia (the JSX),
since an accurate expectation model will
127

Gadjah Mada International Journal of Business, September 1999, Vol. 1, No. 2

yield a stronger information content study.
Second, this study provides information
about the behavior of feasible cash flow
expectation models and the properties of
Indonesian accounting information that
may be useful for future research. Finally,
this also provides evidence of - or at least
it provides a basis for evaluating - the
validity of the KPSAK assertion regarding
the ability of accounting information to
assess future cash flows.
The use of the univariate and multivariate models (Models 1, 2, and 3) based
on a firm-specific may provide alternative
evidence on the properties of accounting
information. However, the limited timeseries data used may constrain the
generalizability of the results. Nevertheless, at least it is expected that this study
will provide a general framework in using
accounting information to forecast future
cash flows so that when more data are
available, the models used in the study will
be reliable.
Since the statement of cash flows is
not readily available for the period before
1995, some cash flows from operations
are calculated using information from the
balance sheet and income statement. This
condition may affect the validity of the
results. However, the use of the same
procedures to compute cash flows from
operations stated in PSAK #2 (Indonesian
Accounting Standard on the Cash Flow
Statements) in this study may minimize
this problem.

128

This study employs manufacturing
firms only as sample firms. The sampling
criteria may affect the generalizability of
the results. Therefore, future research is
necessary to evaluate the robustness of the
results to other populations of firms (nonmanufacturing firms). Furthermore, additional refinements of the prediction models on an industry-specific basis may further enhance the predictive power of the
cash flow forecasting models.
The robustness of the results can also
be evaluated by using an alternative deflator of earnings and cash flows as suggested by Finger (1994). This provides
opportunity to extend this study by employing other deflators of variables used in
the model. Such alternative deflators include consumer price index (CPI) and
market value of the firms. Further extensions include disaggregating cash flow
variables into operating, investing, and
financing components in order to measure
the predictive ability of disclosures prescribed by PSAK # 2. Disaggregation of
current accruals into their components can
also be used to evaluate various information extracted from the accrual process.
Finally, when data become available,
analysis of considerably longer time-series databases may improve the predictive
ability of the models. Also, this analysis
can be used to further evaluate the timeseries properties of semi-annual cash flow
data and the impact of structural changes
on the time-series properties.

Supriyadi—The Predictive Ability of Earnings vs Cash Flow Data to Predict Future Cash Flows

References
Ashton, R. 1974. The predictive-ability criterion and user prediction models. The
Accounting Review (October): 719-32.
Ball, R., 1992. The earnings-price anomaly. Journal of Accounting and Economics, 15:
319-45.
_____ and P. Brown. 1968. An empirical evaluation of accounting income numbers.
Journal of Accounting Research (Autumn): 159-77.
Bathke, Jr. A. W., and K. S. Lorek. 1984. The relationship between time-series models and
the security market’s expectation of quarterly earnings. The Accounting Review:
163-76.
Beaver, W. H., and P. A. Griffin, and W.R. Landsman. 1982. The incremental information
content of replacement cost earnings. Journal of Accounting and Economics (July):
15-39.
Bernard, V. L., and T. L. Stober. 1989. predictive ability as a criterion for the evaluation
of accounting data. The Accounting Review (October): 624-52.
______, and J. Thomas. 1990. Evidence that stock prices do not fully reflect the
implication of current earnings for future earnings. Journal of Accounting and
Economics (December): 305-340.
Bowen, R. M., D. Burgstahler, and L. A. Daley. 1986. Evidence on the relationships
between earnings and various measures of cash flows. The Accounting Review
(October): 713-25.
_____, D. Burgstahler, and L. A. Daley. 1987. The incremental information content of
accrual versus cash flows. The Accounting Review (October): 723-47.
Brown, L. D. 1993. Earnings forecasting research: Its implication for capital market
research. International Journal of Forecasting 9: 295-320.
Brown, L. D., and M. S. Rozeff. 1979. Univariate time-series models of quarterly
accounting earnings per share: A proposed model. Journal of Accounting Research:
179-89.
Chaney, P. K., D. C. Jeter, and C. M. Lewis. 1996. The use of Accruals in Income
smoothing: A Permanent Earnings Hypothesis. Working Paper. Vanderbilt University.
Cooke, T. E., and R. S. O. Wallace. 1990. Financial disclosure regulation and its
environment: A review and further analysis. Journal of Accounting and Public
Policy (9): 79-110.
Dechow, P. M. 1994. Accounting earnings and cash flows as measures of firm performance: The role of accounting accruals. Journal of Accounting and Economics: 342.

129

Gadjah Mada International Journal of Business, September 1999, Vol. 1, No. 2

Dechow, P. M., R. G. Sloan, and A. P. Sweeney. 1995. Detecting earnings management.
The Accounting Review 70: 193-225.
Doupnik, T. S., and S. B. Salter. 1995. External environment, culture, and accounting
practice: A preliminary test of a general model of international accounting development. The International Journal of Accounting (30): 189-207.
Easton, P. 1985. Accounting Earnings and Security Valuation: Empirical Evidence of the
Fundamental Links. Journal of Accounting Research (Supplement): 54-77.
Fairfield, P. M., R. J. Sweeney, and T. L. Yohn. 1996. Accounting classification and the
predictive content of earnings. The Accounting Review (July): 337-55.
Financial Accounting Standards Board. 1978. Statement of Financial Accounting Concepts No. 1: Objectives of Financial Reporting by Business Enterprises. Stamford,
CT: FASB.
Finger, C. A. 1994. The ability of earnings to predict future earnings and cash flow.
Journal of Accounting Research (Autumn): 210-23.
Foster, G. 1977. Quarterly accounting data: Time-series properties and predictive ability
results. The Accounting Review (January): 1-21.
Greenberg, R. R., G. L. Johson, and K. Ramesh. 1986. Earnings versus cash flow as a
predictor of future cash flow Mmasures. Journal of Accounting, Auditing, and
Finance: 266-77.
Guay, W. R., S. P. Kothari, and R. L. Watts. 1996. A Market-Based Evaluation of
Discretionary-Accrual Models. Working Paper. University of Rochester, N.Y.
Hanafi, M. 1997. Informasi laporan keuangan: Studi kasus pada emiten BEJ. KELOLA
Gadjah Mada University journal of Business (September) 16: 74-87.
Healy, P. M. 1985. The effect of bonus schemes on accounting decisions. The Journal of
Accounting and Economics (7):85-107.
Husnan, S. 1992. Effisiensi pasar modal Indonesia. Jurnal Ekonomi Indonesia (April): 2534.
_____, M. Hanafi, and A. Wibowo. 1996. Dampak pengumuman laporan keuangan
terhadap kegiatan perdagangan saham dan variabilitas tingkat keuntungan. Kelola
Gadjah Mada University Business Review 11 (May). Yogyakarta: Master of Management, G.M.U.
Ikatan Akuntan Indonesia (IAI). 1995. Pernyataan Standar Akuntansi Keuangan. Buku
1 dan 2. Jakarta, Indonesia.
Lev, B. 1989. On the usefulness of earnings and earnings research: Lessons and directions
from two decades of empirical research. Journal of Accounting Research (Supplement): 153-210.
Livnat, J., and P. Zarowin. 1990. The incremental information content of cash flow
components. Journal of Accounting and Economics (May): 25-46.
130

Supriyadi—The Predictive Ability of Earnings vs Cash Flow Data to Predict Future Cash Flows

Lorek, K. S. 1979. Predicting annual net earnings with quarterly earnings time-series
models. Journal of Accounting Research: 190-204.
_____, T. F. Schaefer, and G.L. Willinger. 1993. Time-series properties and predictive
ability of funds flow variables. The Accounting Review (January): 151-63.
_____, and G. L. Willinger. 1996. A multivariate time-series prediction model for cashflow data. The Accounting Review (January): 81-101.
Machfoedz, M. 1994. The Usefulness of Financial Ratios in Indonesia (Unpublished
Dissertation). University of Kentucky.
Murdoch, B., and P. Krause. 1989. An empirical investigation of the predictive power of
accrual and cash flow data in forecasting operating cash flow. Akron Business and
Economics Review (Fall): 100-13.
_____, 1990. Further evidence on the comparative ability of accounting data to predict
operating cash flows. The Mid-Atlantic Journal of Business (Winter): 1-13.
Neill, J. D. et al. 1991. The usefulness of cash flow data: A review and synthesis. Journal
of Accounting Literature (Fall): 117-50.
Ndubizu, G. A. 1992. Accounting disclosure methods and economic development: A
criterion for globalizing capital markets. The International Journal of Accounting
(27): 151-63.
Ohlson, J. 1990. A synthesis of security valuation theory and the role of dividends, cash
flows, and earnings. Contemporary Accounting Research (Spring): 648-76.
Pariwayati, and Z. Baridwan. 1998. Kemampuan laba dan arus kas dalam memprediksi
laba dan arus kas perusahaan go publik di Indonesia. Jurnal Riset Akuntansi
Indonesia. Vol. 1 No. 1. Januari: 1-11.
Rayburn, J. 1986. The association of operating cash Flow and accruals with security
returns. Journal of Accounting Research (Supplement): 112-33.
Saudagaran, S. M., and J. G. Diga. 1997. Financial reporting in emerging capital markets:
Characteristics and policy issues. Accounting Horizons (June): 41-64.
Setiawati, L. (1995). Kandungan Nilai Informasi Pengumuman Laba: Studi Kasus pada
Bursa Efek Jakarta. (Unpublished Thesis S1). Yogyakarta: Faculty of Economics,
G.M.U.
Sloan, R. G. 1996. Do stock prices fully reflect information in accruals and cash flows
about future earnings? The Accounting Review (July): 289-316.
Sugiyanto, C. 1998. Price linkages in selected Indonesian financial market. Kelola Gadjah
Mada University Business Review 17 (January): 50-60.
Supriyadi. 1998. The Association between Accounting Information and Future Cash
Flows: An Indonesian Case Study (Unpublished Dissertation). University Of Kentucky.

131

Gadjah Mada International Journal of Business, September 1999, Vol. 1, No. 2

Sutton, M. H. 1997. Financial reporting in U.S. capital markets: International dimensions.
Accounting Horizons (June): 96-102.
Watts, R. L., and J. L. Zimmerman. 1986. Positive Accounting Theory. New Jersey:
Prentice-Hall, Inc.
Watts, R. L., and R. W. Leftwich. 1977. The time-series of annual accounting earnings.
Journal of Accounting Research: 253-71.
Wilson, G. P. 1986. The relative information content of accruals and cash flows:
Combined evidence at the earnings announcement and annual report release date.
Journal of Accounting Research (Supplement): 165-200.
_____, 1987. The incremental information content of funds components of earnings after
controlling for earnings. The Accounting Review (April): 293-322.
Wolk, H. I., and M. G. Tearney. 1997. Accounting Theory: A Conceptual and Institutional
Approach. 4 ed. Cincinnati: South Western College Publishing Co.

132