an investigation of financial ratios in predicting firms future performance an application of pr's...

Upload: kemalaristanti

Post on 21-Feb-2018

217 views

Category:

Documents


0 download

TRANSCRIPT

  • 7/24/2019 An Investigation of Financial Ratios in Predicting Firms Future Performance an Application of Pr's Methodology

    1/96

    INFORMATION TO USERS

    [I

    This manuscript has been reproduced from the microfilm master. UMI

    films the text directly from the original or copy submitted. Thus, some

    thesis and dissertation copies are in typewriter face, while others may be

    from any type o f computer printer.

    Th e quality of this reproduction is dependent upon the quality of the

    copy subm itted. Broken or indistinct print, colored or poor quality

    illustrations and photographs, print bleedthrough, substandard margins,

    and improper alignment can adversely afreet reproduction.

    In the unlikely event that the author did not send UMI a complete

    manuscript and there are missing pages, these wiil be noted. Also, if

    unauthorized copyright material had to be removed, a note will indicate

    the deletion.

    Oversize materials (e.g., maps, drawings, charts) are reproduced by

    sectioning the original, beginning at the upper left-hand comer and

    continuing from left to right in equal sections with small overlaps. Each

    original is also photographed in one exposure and is included in reduced

    form at the back o f the book.

    Photographs included in the original manuscript have been reproduced

    xerographically in this copy. Higher quality 6 x 9 black and white

    photographic prints are available for any photographs or illustrations

    appearing in this copy for an additional charge. Contact UMI directly to

    order.

    UMIA Bell & Howell Information Company

    300 North Zeeb Road, Ann Arbor MI 48106-1346 USA

    313/761-4700 800/521-0600

    eproduced with permission of the copyright owner. Further reproduction prohibited without permission.

  • 7/24/2019 An Investigation of Financial Ratios in Predicting Firms Future Performance an Application of Pr's Methodology

    2/96produced with permission of the copyright owner. Further reproduction prohibited without permission.

  • 7/24/2019 An Investigation of Financial Ratios in Predicting Firms Future Performance an Application of Pr's Methodology

    3/96

    AN INVESTIGATION OF FINANCIAL RATIOS IN PREDICTING FIRMSFUTURE PERFORMANCE: AN APPLICATION OF PRs METHODOLOGY

    A Dissertation

    Presented to

    the Faculty of the College of Business Administration

    University of Houston

    In Partial Fulfillment

    of the Requirements o f the Degree

    Doctor of Philosophy

    by

    Lijyun Angela Hwang

    April 18, 1997

    eproduced with permission of the copyright owner. Further reproduction prohibited without permission.

  • 7/24/2019 An Investigation of Financial Ratios in Predicting Firms Future Performance an Application of Pr's Methodology

    4/96

    UMI Number: 9817079

    UMI Microform 9817079Copyright 1998, by UMI Company. All rights reserved.

    This microform edition is protected against unauthorizedcopying under Title 17, United States Code.

    UMI300 North Zeeb Ro adAnn Arbor, MI 48103

    eproduced with permission of the copyright owner. Further reproduction prohibited withou t permission.

  • 7/24/2019 An Investigation of Financial Ratios in Predicting Firms Future Performance an Application of Pr's Methodology

    5/96

    AN INVESTIGATION OF FINANCIAL RATIOS IN PREDICTING FIRMS

    FUTURE PERFORMANCE: AN APPLICATION OF PRs METHODOLOGY

    APPROVED:

    C.S7 Ames Cheng, Associate Pr|, then the ratio is likely to be a profitability-oriented ratio.

    Gh hit (15)

    The Retums-Target Prediction Model:

    bhXhn (16)

    Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.

  • 7/24/2019 An Investigation of Financial Ratios in Predicting Firms Future Performance an Application of Pr's Methodology

    68/96

    If a a; 0 and \a\ * \b\,then the ratio is likely to be a risk-oriented ratio.

    I f a=*=0 and * \b\, then the ratio is likely to be a ratio with a mixture of both

    profitability and risk.

    Based on these relationships, the scheme of classifying financial ratios can be illustrated as

    follows:

    CLASSIFICATION OF FINANCIAL RATIOS BY COMPARING LOGISTIC

    COEFFICIENTS OF EARNtNGS-TARGET PREDICTION MODEL

    AND RETURNS-TARGET PREDICTION MODEL

    a sare the coefficient estimates form the eamings-target prediction model. 6 s are

    the coefficient estimates form the retums-target prediction model. The difference of"a 5 and bs is evaluated by taking the mean o f paired differences o f a and 6during each evaluation period.

    a ~0 a *0

    a ~ b Not useful Profitability

    a &\b Risk Mixed

    Table 5 reports the yearly coefficients obtained every three years from univariate

    logistic regressions of the respective models. For example, the coefficients for year 1980

    are obtained by pooling firm-year observations o f 1977-1979. Panels A and B report the

    yearly coefficients for the eamings-target prediction model and the retums-target

    prediction model during 1980-1986. Based on these yearly logistic regressions, an across-

    Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.

  • 7/24/2019 An Investigation of Financial Ratios in Predicting Firms Future Performance an Application of Pr's Methodology

    69/96

    R

    d

    d

    ith

    i

    i

    fth

    iht

    F

    th

    d

    ti

    hibit

    d

    ith

    t

    i

    i

    -and-hold exc ess returns are the dependent var iables I'mthe earnmgs-target prediction miHlel and the retums-target prediction model respect ively Independent variables are

    those identified by principal component analysis Coefticieu ts labeled for year 1980 are obtained by pooling firm-year observations from 1977-1979.

    Panel A: Earnings-Targct Prediction Model

    Net profit margin %A in Net profitmargin

    Debt-equityratio

    %A in Debt-equity ratio

    Inventory /totalassets

    %A in Inventory/total assets

    Quick ratio %A in Quickratio

    1980 1.399 -0.030 -0.113 0.390 -1.586* -1.010b 0.403* 0.448"

    1981 -2.703 -0.091 -0.062 0.855* -1.195b -1.572* 0.2 94 1 0.081

    1982 -4.043* -0 .1571 -0.092 0.67 lb -0.264 -1.294* 0.130 -0.117

    1983 -8.102* -0.250* -0.075 0.293 -0.401 -1.880* 0.107 0.185

    1984 -5.095* -0.172* -0.169 0.082 -1.302b -1.191* 0,173 0.293

    1985 -4.539* -0.141* -0.062 0.037 -1.182 -1.472* -0.013 0.332

    1986 -4.048* -0.094 0.041 0.478* -1.578* -1.000b -0.061 -0.143

    Meana -3.876k - A / i/ -0.076 0.401" -1.073* -1.345* 0.148 0.154

    Mean Sii c rror 1.609 0.063 0.082 0.241 0.566 0.411 0.117 0.246

    Wald Chi-square* 5.799 4.481 0.858 2.770 3.597 10.720 1.604 0.393

    Panel B: Returns-" 'arget Prediction IVodel

    1980 4.1 78 1, 0.007 -0.012 -0.151 -1.225b -1.152* -0.026 0.218

    1981 1.776 0.046 -0.022 -0.080 -0.128 -0.618 0.134 -0.058

    1982 -0.815 0.085 -0.1941 -0.103 0.873 -0.121 0.380* -0.246

    1983 4.372* -0.046 -0.081 -0.216 0.107 -0.210 0.004 -0.053

    1984 -0.199 -0.072 -0.045 -0.355 -0.457 -0.422 -0.049 0.435"

    1985 1.885 -0.034 -0.024 -0.368 -1.688* -0.732 -0.203 0.450

    1986 7.187* 0.125* -0.090 -0.112 -1.895* -0.062 -0.086 0.300

    Mean 2.626" 0.016 -0.067 -0.198 -0.630 -0.474 0.022 0.149

    Mean Std Error 1.583 0.060 0.082 0.255 0.562 0.401 0.115 0.243

    Wald Chi-square 2.754 0.070 0.672 0.598 1.260 1.400 0.037 0.378

    Panel C: Evaluation of Ratio Attribute

    Mean dilTerencer 1.360 0.074 0.021 0.203 0.162 0.871 0.043 -0.023

    Prob>|T| 0.302 0.039 0.516 0.098 0.539 0.006 0.619 0.707

    a not equal to 0s ** ** * * *+*

    Attribute Profitability Mix Not useful Mix Profitability Mix Not useful Nol useful

  • 7/24/2019 An Investigation of Financial Ratios in Predicting Firms Future Performance an Application of Pr's Methodology

    70/96

    f

    TAB1.H5 (CONTINUED)UNIVARIATE ANALYSIS OF LlXJISTIC COIiM-lCU-lN IS AND FINANCIAL ATTRIBUTES FOR RATIOS IDENTIFIED BY PRINCIPAL COMPONENT ANALYSIS

    Panel A: Earnings-Targct Prediction Model

    Sales/total assets %A in sales /

    total assets

    %A in LT debt %A in total

    assets

    Cash dividend as

    % of cash flows

    Sales to fixed

    assets

    Sales to

    inventory

    Sales to

    working capital

    1980 -0.610* 1.1261 -0.145" -1.692* 1.956* -0.04Sb 0.008 -0.0131981 -0.250* 0.474 -0.140" -1.256* 1.453* -0.013 0.015 0.001

    1982 0.173 0.891" -0.111 -0.936* 0.800" 0.051* 0.031* 0.013"

    1983 0.196 0.173 -0.111 -2.450* 0.798b 0.036* 0.020* 0.006

    1984 -0.069 1.119b -0.184* -2.181* 0.551 -0.006 0.017 -0.003

    1985 -0.090 -0.065 -0.288* -1.943* 0.697* -0.055* 0.000 -0.014b

    1986 -0.163 -0.161 -0.206b -0.829" 1.151* -0.046b 0.009 -0.017*

    Mean -0.116 0.508 -o. i6r -1.612 i . o s r -0.011 0.014 -0.004

    Mean Std Error 0.132 0.522 0.092 0.439 0.398 0.019 0.012 0.007

    Chi-square 0.774 0.948 3.395 13.481 7.061 0.340 1.436 0.259

    Panel B: Rctums-Target Prediction Model

    1980 -0.542c 1.332b -0.072 0.248 -1.604* -0.025 0.001 -0.006

    1981 -0.048 1.076b -0.074 -0.494 0.125 0.048' 0.010 -0.006

    1982 0.388* 0.783 -0.041 -0.739* 0.218 0.082* 0.014 -0.013"

    1983 0.469* 0.448 -0.119 -1.693' 0.724b&00o 0.035' 0.012"

    1984 0.410* 0.366 -0.077 -1.343* 0.147 0.016 0.038' 0.012*

    1985 0.245 0.276 -0.305* -1.795* 0.545 -0.031 0.043' 0.015b1986 0.200 0.427 -0.154 -0.581 0.686* -0.035 0.03 lb 0.004

    Mean 0.160 0.673 -0.120 -0 .W / 0.120 0.015 0.025* 0.003

    Mean Std Error 0.141 0.517 0.091 0.428 0.415 0.019 0.012 0.008

    Wald Chi-square 1.300 1.691 1.730 4.564 0.084 0.573 4.133 0.118

    Panel C: Evaluation of the Ratio Attr )UtCS

    Mean difference -0.107 -0.100 0.049 0.628 0.480 -0,005 -0.010 0.000

    Prob>|T| 0.195 0.560 0.028 0.012 0.022 0.627 0.236 0.971

    a not equal to b * *** *+*

    Attribute Not useful Not useful Mix Mix Mix Not useful Not useful Not useful

  • 7/24/2019 An Investigation of Financial Ratios in Predicting Firms Future Performance an Application of Pr's Methodology

    71/96

    S7

    T A U 1 ,1 - 5 ( C O N T I N I J I ' D )

    UNIVARIATH ANALYSIS Ql; LOGISTIC COITTICTI-NTS AND FINANCIAL ATTRIBlITHS FOR RATIOS IDENTIFIED BY PRINCIPAL COMPONENT ANAI.YS1S

    '' *c Denotes statistical significance at the 0.10.0 .05 . and 0.01 levels respectively, based on a two-tail t-test.

    ' Similar to Lev and Thiagarajan j 1993|, the yearly logistic regression is used to obtain an across-year mean, standard error to calculate Wald chi-square foreach financial ratio.

    cThe statistical meanings of Wald chi-square is similar to that of t-value in regression analysis. It indicates the statistical significance of a coefficient.

    r Mean difference is the mean of the differences between the absolute value of the coefficient for each variable obtained from the eamings-target prediction

    model and the corresponding coefficient from the retums-target prediction model during the same time periods. As a result, 7 paired observations for each

    financial ratio are included in the mean tests. As both of the parametric test (paired t lest) and the nonparamctric test (paired Wilcoxon singed rank test)

    produce similar results, only the paired t test is reported as Prob>|T|.

    KThis row summarizes across-year (mean) coefficients that are statistically significant in the earnings model as reported in Panel A. *. **, and *** denote the

    significance level at 0.10. 0.05. and 0.01 respectively.

  • 7/24/2019 An Investigation of Financial Ratios in Predicting Firms Future Performance an Application of Pr's Methodology

    72/96

    year mean, standard error are used to calculate Wald chi-square for each financial ratio to

    evaluate their statistical significance. This approach attempts to control for upward bias

    due to cross-sectional correlation in a pooled analysis (Bernard [1987], Lev and

    Thiagarajan [1993], and Cheng, Liu and Schaefer [1996]).

    Panel C of Table 5 reports the classification o f the financial attributes of the 16

    financial ratios based on the above framework. Net profit margin and inventory/total

    assets are classified as profitability-oriented ratios. Percentage change (%A) in net profit

    margin, %A in debt-equity ratio, %A in inventory/total assets, %A in long-term debt, %A

    in total assets, and cash dividend as a percent o f cash flows are classified as ratios with a

    mixture o f profitability and risk. Debt-equity ratio, quick ratio, %A in quick ratio,

    sales/total assets, %A in sales/total assets, sales to fixed assets, sales to inventory and sales

    to working capital are classified as non-useful ratios. None o f the ratios are classified as

    pure risk-oriented ratios.

    6.2.3 Prs Constructs

    6.2 .3.1 An Initial Evaluation Based on Multivariate Logistic Regressions

    The empirical models used to evaluate the constructs of Pr in the two multivariate

    logistic regressions are expressed as follows:

    The Eamings-Target Prediction model:

    AEiti v ' / 1 - = 2 aioch,, ( 15 )

    " h=\

    with permission of the copyright owner. Further reproduction prohibited without permission.

  • 7/24/2019 An Investigation of Financial Ratios in Predicting Firms Future Performance an Application of Pr's Methodology

    73/96

    5 3

    The Retums-Target Prediction model:

    m

    R~i = ^bkXh,t (16)>1=1

    where,xh(/ = 1 ,2 m) is the set o f mfinancial ratios for firm iat timeI,and ahand bh

    are the vectors o f the estimated coefficients for the financial ratios in the respective

    models. The definition o f the dependent variables are the same in both the multivariate

    models and the univariate models. Table 6 reports the yearly coefficients obtained every

    three years from multivariate logistic regressions of the respective models. The estimation

    of coefficients are similar to that of the univariate regressions except that only one

    financial ratio is included in the univariate models but all 16 financial rations are included

    in the multivariate models.

    A coefficient estimated in a logistic model indicates the effect of a unit change in

    an independent variable on the probability o f the occurrence o f the dependent variable.

    Given a significant statistical test, a positive sign o f a parameter estimate suggests that the

    probability o f one-year-ahead earnings changes (excess returns) increases with the value of

    the financial ratio. A comparison o f the coefficients in the univariate analysis in Table 5

    and multivariate analysis in Table 6 reveals consistent signs between the two analyses.

    This implies the collinearity is not a problem for the multivariate analysis. However, the

    significance level o f the coefficients is stronger in the univariate analysis than in the

    multivariate analysis. This is expected because univariate analysis only accounts for the

    effect of an individual independent variable. Ratios that are statist ically significant in the

    eamings-target prediction model include: %A in debt-equity ratio, %A in inventory/total

    Reproduced with permission of the copyright owner. Further reproduction prohibited withou t permission.

  • 7/24/2019 An Investigation of Financial Ratios in Predicting Firms Future Performance an Application of Pr's Methodology

    74/96

    R

    d

    d

    ith

    i

    i

    fth

    iht

    F

    th

    d

    ti

    hibit

    d

    ith

    t

    i

    i

    61)

    TABLE 6

    MULTIVARIATE ANALYSIS OF THE LOGISTIC COEFFICIENTS FOR RATIOS IDENTIFIED BY PRINCIPAL COMPONENT ANALYSIS

    Coefficients of logistic regressions arc estimated for every three .years. Signs of onc-ycar-ahcad earnings change without drift and signs of 12-month market-

    adjusted buy-and-hold excess returns are the dependent variables for the eamings-target and the returns-target prediction model respectively. Independent

    variables are those identified by principal component analysis. That is. coefficients labeled for year 1980 arc obtained by pooling firm-year observations from

    1977-1979. The coefficients are then fitted to the accounting data of year 1980 to obtain the eamings-target Pr (E-Pr) and retums-target Pr (R-Pr).

    Panel A: Eamings-Target Prediction ModelNet profit

    margin

    %A in Net

    profit margin

    Debt-equity

    ratio

    %A in Debt-

    equity ratio

    Inventory

    /total assets

    %A in Inventory /

    total assets

    Quick ratio %A in Quick

    ratio

    1980 3.143 0.019 0.049 1.859" -1.752b -0.607 0.255 0.642

    1981 0.304 -0.016 0.036 2.164' -1.169 -1.366' 0.343 0.119

    1982 0.475 -0.1411 -0.133 1.851" 0.205 -1.373" 0.114 -0.066

    1983 -2.862 -0.1611, -0.161 1 .2 0 2 " -0.597 -1.560" 0.086 -0.054

    1984 -4.653" -0.1511, -0.2921 1.075" -2.164" -0.999" 0.147 0.223

    1985 -4.658b -0.086 -0.223b 1.162" -1.143 -1.390' -0 . 1 0 2 0.386

    1986 -5.830" -0.073 -0.107 1.408" -2.216b -1.050* -0.135 0.146

    Mean** -2 . 0 1 2 -0.087 -0.119 1.531' -1.262 - l . l 9 f 0 . 1 0 1 0.199

    Mean Std Error 2.362 0.068 0.107 0.423 1.074 0.489 0.157 0.323

    Wald Chi-Square 0.725 1.661 1.230 13.104 1.382 5.954 0.415 0.380

    Mean Std. Estimate* -0.046 -0.050 -0.045 0.187 -0.070 -0,094 0.025 0.026

    Panel B: Retums-Target Prediction Model

    1980 5.829* -0.052 -0.039 0.530 -2.555* -1.528" -0 . 2 2 1 -0.172

    1981 5.332b 0.006 0.077 0.492 -0.841 -0.978b -0.029 -0.3281982 -0.045 0.088 -0.096 0.248 0.861 -0.424 0.271 -0.640

    1983 -1.675 -0 . 0 1 0 -0.098 0.298 1.300 0.071 0.016 -0.075

    1984 3.007 -0.077 -0.009 0.297 1.287 0 . 0 0 2 -O.I18 0.476

    1985 5.727" -0.038 -0.049 0.855" 0 . 0 2 0 -0.109 -0.356b 0.723

    1986 9.345" 0.074 -0 . 0 1 1 0.85 lh -1.476 0.374 -0.3901 0 .6 6 8 "

    Mean 3.931a -0 . 0 0 1 -0.032 0.510 -0 . 2 0 1 -0.370 -0.118 0.093

    Mean Std Error 2.271 0.065 0,103 0.394 1.057 0.473 0.152 0.313

    Wald Chi-Square 2.997 0 . 0 0 0 0.097 1.679 0.036 0.614 0.604 0.089

    Mean Std. Estimate 0.081 -0 . 0 0 0 -0 . 0 1 2 0.065 -0 . 0 1 2 -0.027 -0.037 0.015

  • 7/24/2019 An Investigation of Financial Ratios in Predicting Firms Future Performance an Application of Pr's Methodology

    75/96

    R

    d

    d

    ith

    i

    i

    fth

    iht

    F

    th

    d

    ti

    hibit

    d

    ith

    t

    i

    i

    < > l

    TABLE 6 (CONTINUED)

    MULTIVARIATE ANALYSIS OF THE LOGISTIC COEFFICIENTS FOR RATIOS IDENTIFIED BY PRINCIPAL COMPONENT ANALYSIS

    Panel A. Earnings-Targct Prediction Model

    Sales /

    total assets

    %A in sales /

    total assets

    %A in LT debt %A in total

    assets

    Cash dividend as

    % of cash flows

    Sales to fixed

    assets

    Sales to

    inventory

    Sales to working

    capital

    1980 -0 . 1 0 1 1.124" -0.2741 -2.296 1.292b -0,006 -0.024 -0.0071981 -0.074 1.057" -0.322 -1.759 0.737 0.009 -0 . 0 1 1 0.005

    1982 -0.059 1,553c -0.2481 - 1. 1 0 2 " 0.555 0.06 lb 0.029 0 . 0 1 1

    1983 -0.037 0.755 -0.025 -2.432 0.502 0.051" 0.014 0.005

    1984 -0.047 l.793c -0.132 -1.703 0.358 0 . 0 2 2 -0 . 0 0 0 -0.007

    1985 0.007 0.594 -0.400c -1.243h 0.292 -0.037 0.005 -0.028

    1986 -0.008 0.446 -0.578c 0.086 1.190b -0.008 0.013 -0.044

    Mean -0.046 1.046" -0.28? -1.493* 0,704 0.013 0.004 -0.009

    Mean Std Error 0.047 0.615 0.128 0.652 0.456 0.028 0.019 0 . 0 1 0

    Wald Chi-Square 1 . 2 0 0 2.894 4.906 5.245 2.382 0.217 0.038 0.819

    Mean Std. Estimate -0.048 0.063 -0 . 1 0 1 -0 . 1 1 1 0.056 0 . 0 2 2 0 . 0 1 1 -0.044

    Panel B: Retums-Target Prediction Model

    1980 -0.068 1.693 -0.114 -0.262 -2.125b 0.036 -0.036* -0.016

    1981 -0.023 1.071" -0.057 -0.891 -0 . 1 1 0 0.080 0.005 -0.009

    1982 0.130 0.642 0.043 -0.680 0.209 0.064b 0.037* -0 . 0 1 0

    1983 0.009 -0.184 -0.004 -1.546 0.715* 0.040 0.053 0.008

    1984 0.031 -0.078 0.032 -1.833 -0.030 0.017 0.054 0.0041985 0.041 -0.056 -0.261 -2.289 -0.017 0.003 0,044b -0.004

    1986 0.066 0.151 -0.247 -1.176 -0.045 0.017 0.014 -0 . 0 1 1

    Mean 0 . 0 1 0 0.463 -0.087 -1.23? -0 . 2 0 0 0.037 0.024 -0.006

    Mean Std Error 0 . 0 2 1 0.596 0.119 0.624 0.433 0.027 0 . 0 2 0 0 . 0 1 0

    Wald Chi-Square 0.226 0.603 0.533 3.950 0.215 1.804 1.557 0.326

    Mean Std. Estimate 0.026 0.027 -0.030 -0.094 -0 . 0 1 2 0.061 0.068 -0 . 0 2 0

    ' ' Denotes statistical significance at the 0.10, 0. 05, and 0 01 levels respectively, based on two-tail t-tesl

    1Simil ar to Lev and 1'hiagarajan 11993 1, the yearly logis tic regress ion is us ed to obtain an across-year mean, standard error to calculate Wa ld Chi-Square for each

    financial ratio

    Standardized coeificients are reported here in order to compute the relative influence of independent variables in the models

  • 7/24/2019 An Investigation of Financial Ratios in Predicting Firms Future Performance an Application of Pr's Methodology

    76/96

    62

    assets, %A in sales/total assets, %A in long-term debt, %A in total assets (see Panel A in

    Table 6). In the retums-target prediction model, only two financial ratios are statistically

    significant: net profit margin and %A in total assets (see Panel B in Table 6).

    The signs of these ratios in both models are consistent except for the net profit

    margin, which is operationalized as the ratio of earnings before extraordinary items to

    sales. In the earnings model, the sign o f the net profit margin is negative which indicates

    that an increase in the current earnings level will decrease the probability of an increase in

    the future earnings change. This observation is consistent with the earnings reversal

    behavior documented in previous studies (Freeman, Ohlson, and Penman [1982], Ou

    [1990], Elgers and Pfeiffer [1996]). In the returns model, however, the net profit margin's

    sign is positive which suggests that higher current earnings lead to higher probability of

    future excess returns.

    The other significant ratio in the retums-target prediction model is %A in total

    assets, and it has a negative sign. These observations o f a positive coefficient for net

    profit margin and a negative coeffcient for %A in total assets suggest that the probability

    of future excess returns for a firm will increase if operating profits can grow without tying

    up more investment in assets. These empirical results confirms the relationship between

    returns and accounting information specified in the finance free cash flow concept. Firm

    value equals the present value of discounted future free cash flow, and free cash flow is a

    eproduced with permission of the copyright owner. Further reproduction prohibited without permission.

  • 7/24/2019 An Investigation of Financial Ratios in Predicting Firms Future Performance an Application of Pr's Methodology

    77/96

    63

    positive function of net operating margin (profitability) and a negative function of growth

    in operating assets as suggested in Ohlson and Feltham [1995] (p. 695).25

    As reported in Panel A of Table 6, financial ratios that are significant in the

    eamings-target prediction model are (%A in debt-equity ratio, %A in inventory/total

    assets, %A in sales/total assets, %Ain long-term debt, %A in total assets. Based on Table

    5, they are all classified as ratios with a mixture of profitability and risk except for %A in

    sales/total assets, which is categorized as a non-useful ratio in the univariate logistic

    analysis. Although these ratios are classified as ratios with a mixture of profitability and

    risk, they are more profitability oriented because they are statistically significant in the

    eamings-target prediction model but not significant in the retums-target prediction model.

    6.2.3.2 A Further Look at Prs Constructs by Exam ining Trading

    Profitabilities of the Earnings-Target Prediction Model and the Returns-Target

    Prediction Model

    The interpretation o f Prs constructs cannot be complete without looking into the

    trading profitability for the two models. In HL, financial ratios that are significantly

    associated with future earnings increases (excess returns) are used to form eamings-target

    Pr (E-Pr) and retums-target P r (R-Pr), respectively. They found that the trading

    ' ' The free cash flow concept states that:Free cash flow,

    = Operating earnings, + Operating assets,./ - Operating assets,

    = Operating earnings, - (Operating assets, - Operating assets,./)

    = Operating earnings, - (A in Operating assets,)

    = [(Operating earnings, / Operating assets,./) - (A in Operating assets, / Operating assets,./)) * Operating assets,./= (Operating margin, - growth in operating assets,) * Operating assets,./

    eproduced with permission of the copyright owner. Further reproduction prohibited without permission.

  • 7/24/2019 An Investigation of Financial Ratios in Predicting Firms Future Performance an Application of Pr's Methodology

    78/96

    64

    profitability based on R-Pr dominates eamings-target Pr. In this study, principal

    component analysis is used to obtain a set o f financial ratios that only depends on the

    interrelationship among a large number o f financial ratios without considering their

    association with the dependent variables to form Pr.

    Trading profitability is measured by monthly rebalanced market-adjusted returns

    gained from a hedge portfolio over various holding periods as in Greig [1992], Hedge

    portfolios for the eamings-target prediction model and retums-target prediction model are

    formed based on two schemes. The first scheme is to form a hedge portfolio that take a

    long (short) position if Pr is greater (less) than 0.6 (0.4). The second scheme involves

    trading strategy by dichotomizing the whole sample. That is, trading strategy is based on

    taking a long position for Prs that is greater than 0.5 and taking a short position for Pr

    less than and equal to 0.5. While Table 7 provides the small prints for the by-year

    trading profitability to the long position, short position and hedge portfolio for various

    holding periods of 3, 6, 9, 12, 18, 24, 36, 48, 60 and 72 months. Table 8 highlights the

    mean returns for the average of the by-year returns reported in Table 7. To get a better

    picture of the trading profitability trends, the trading profitability over various holding

    periods are plo tted side-by-side. Finally, the trading profitability patterns for the various

    trading strategies, on a common scale, are comprehensively presented in Figure 3.

    Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.

  • 7/24/2019 An Investigation of Financial Ratios in Predicting Firms Future Performance an Application of Pr's Methodology

    79/96

    6

    I Mill- 7

    H Y - Y K A R T R A D I N G I 'R O I I I A H I I I T Y T O M il- : 11 I. IK .I -, P O R T F O L I O S O V I -R H O L D I N G P F R I O D S O T .3 . 6 , 9 . 1 2 . 1 8 . 2 4 , 3 6 . 4 8 . 6 0 A N D 7 2 M O N T H S

    T h e m o n t h l y r e b a l a n c e d m a r k e t a d j u s t e d r e t u r n s t o t h e h e d g e p n r t o l i o s a r c c a l c u l a t e d b y t a k i n g l o n g p o s i t i o n f o r f i rm s w i t h P r > 0 .6 ( P r - 0 . 5 ) a n d s h o r t f o r f ir m s w i t h P r- 0 4

    ( P r - - 0 . 5 ) f o r t h e c a i n i n g s - t a r g c t m o d e l a n d t h e r c t o r n s - t a r g c t m o d e l . P o r t f o l io s a r e fo r m e d t h r e e m o n t h s a f t e r t h e D e c e m b e r y e a r e n d f o r w h i c h t h e a c c o u n t i n g i n f o r m a t i o n is

    n s s m u c d t o b e p u b l i c l y a v a i l a b l e . T h e r e t u r n s a r e c a lc u l a t e d u s i n g e q u a t i o n s ( 1 0 ) . ( 1 1 ) . a n d ( 1 2 ) i n t h e t e x t . M e a n r e t u r n s a r e t h e ' a v e r a g e o f t h e b y - y e a r m e a n s .

    lannel A: Kamings-'Parget Model with Pr 0 6 for long posilon (n -1017) and lr 0.4 (n 472) lor short position

    No. of Observations Holding Period 3 Holding Period 6 Holding Period-9 Holding Period-12 Holding Periods-18

    Year Ixmg Short t-ong Short 1ledge 1AMlg Short 1ledge long Short Hedge Ixtng Shon Hedge 1aitig Short Hedge

    80 341 17 0.0457 -0.0086 0.0543 0.0146 -0.0876 0 1022 0.0207 -0.0744 0.0952 0.0371 -0.1527 0.1898 0 0218 -0.2638 0 285681 208 33 0 0100 -0.0229 0.0328 -0 0048 0.0071 -0.0119 0.0370 0.1177 -0.0807 0.0989 0.1783 -0.0794 0 1606 0.1739 -0.013382 56 127 0.0916 0.0389 0.0526 0.0808 0.0388 0.0420 0.0807 0.0551 0.0257 0 1063 0.0383 0.0680 00910 0.0124 0078683 50 107 0.0210 0 0121 0.0089 0.0094 -0.0089 -0.0006 -0.0797 -0.0332 -0.0465 -0.0597 -0.0143 0.0454 -0 0823 -0.0174 -0 064884 94 76 -0.0563 -0.0647 0.0083 -0.0281 -0 0532 0.0251 -0.0488 -0.0865 0.0377 -0.0215 -0.0766 0.0551 -0.0470 -0.1261 0.079083 161 42 -0.0113 0.0072 -0 0 184 -0.0123 -0.0419 0.0295 0.0132 -0.0229 0.0097 0.0283 0.0075 0 0208 0.0993 0.0246 0.123986 107 70 00199 -0.0308 0.0507 0.0534 0.0230 0.0765 0.0241 -0.0419 0.0659 0.1002 0.0717 0.0284 0 1008 0.1010 -0 0002

    Mean Returns 0 0172 -0.0098 0.0270 0.0134 -0.0241 0.0375 00030 -0. 0123 0. 0153 0.0414 0.0075 0.0339 00492 -0.0207 00698

    Pannel H: Famings-Targct Model with Pr 0.5 forlong posilon (n -1814) and P r 0.5 (n 1100) for short position

    80 444 39 0.0479 0.0545 -0.0066 0.0047 -0.0294 0.0341 0.0076 -0.0015 0.0091 0.0275 0.0413 0.0689 0.0195 -0.0954 0.114881 362 105 -0.0002 -0.0215 0.0213 -0.0140 -0.0033 0.0108 0.0293 0.0595 -0.0302 0.0927 0.1490 -0.0563 0.1605 0.1754 -0.015082 130 298 0 0616 0.0416 0.0199 0.0630 0.0507 0.0123 0. 0784 0.0680 0.0104 0 1054 0.060S 0.0448 0.0904 0.0330 0.057483 170 247 0.0237 0.0037 0.0200 0 0081 -0.0115 0.0196 -0.0248 -0.0196 -0.0052 -0.0075 -0.0040 -0.0035 -0.0202 -00153 -0.0049

    84 226 168 -0.0281 -0.0404 0.0122 0.0003 -0.0282 0.0286 -0.0065 -0.0429 0.0363 0.0214 0.0277 0.0491 0.0124 0.0676 0.080083 271 101 -0.0105 0.0220 -0.0325 0.0246 -0.0093 -0.0153 -0.0237 0.0121 -0.0116 0.0145 0.0230 0.0084 0.0534 0.0207 0.032786 211 142 0.0086 -0.0155 0 0241 00502 0.0037 0.0466 0.0259 0.0022 0.0238 0 1048 0.0924 0.0124 0.1102 0.0893 0.0209

    Mean Returns 0.0147 00064 0.0084 0.0125 -0 .0039 0.0164 0 0123 0 007 7 0.0047 0.0513 0.0360 0.0153 0.0609 0.0200 0.0409

    Pannel C: Retums-Targc t Model with 1Pr 0.6 for long positon (n 386) and Pr 0.4 (n; 487) forslion position

    80 58 114 0.0522 0.0467 0.0055 -0.0177 0.0152 -0.0329 -0.0208 -0.0007 -0.0201 0.0220 0.0293 -0.0513 00830 0.0196 0.102581 21 74 0.0053 -0.0269 0.0322 0.0028 -0.0524 0.0551 0.1149 -0.0279 0.1428 0.2581 0.0658 0.1922 0.2382 0.1277 0.110582 100 41 0. 0967 0.0553 0.0414 0.0767 0.0631 0.0136 00673 0.1082 -0.0409 0.0568 0.1911 0.1343 0.0571 0.0560 0.001283 65 38 0.0355 -0.0020 0.0374 0.0139 -0.0049 0.0188 -0 0259 -0.0479 0. 0220 0.0073 0.0415 0.0342 0 0133 -0.0535 0066884 47 48 -00222 -0.0641 0.0418 -0.0169 -0.1061 0.0892 -0.0359 -0 1364 0.100S -0 0151 0.1278 0.1127 -0.0335 0.1806 0 147183 55 72 0.0190 -0.0154 0.0344 0.0207 -0.0643 0.0850 0.0204 -0.0787 0.0991 0.0763 -0.0563 0.1326 0.0848 -0 0784 0.163286 40 100 0.0100 0.0234 0.0334 0. 0383 0.0020 0.0363 0.0272 -0.0520 0.0792 0.0643 0.0381 00262 0.0772 0.0706 00066

    Mean Returns 0 028) -0.0043 00323 0 0168 -0 0211 0.0379 0 0210 -00336 0.0546 0.0587 0.0141 0.0446 0 0506 -0.0055 0 0561

    Pannel 1): Eamings-Target Model with Pr 0.5 for long positon (n 1381 ) and Pr 0 5 (n 15.73) lor short position

    80 219 264 0.0483 0.0487 -0.0004 -0.0112 0 0131 -0 0243 -0 0089 0.0205 -0.0294 -0 0051 0.0452 -0 0502 0.0287 0.0435 -0 072281 154 313 0.0044 -0.0097 0.0141 0.0000 -0.0172 0.0172 0.0710 0.0194 0.0516 0.1398 0.0888 00510 0.1880 0.1524 (10356

    82 281 147 00539 0.0357 0.0183 0 0515 00598 -0 0083 00585 0.0958 -0.0374 0 0541 0.1130 -0.0589 0 0498 0.0501 0.000383 233 184 0 0252 -0.0045 0.0297 0 0058 -0.0151 00208 -0 0141 -0 0310 0.0169 00047 -0.0180 0 02 26 -0.0032 0.0347 0 031584 177 217 -00227 -0.0419 0.0192 0 0031 -0.0239 0 027 0 -0 0030 -0.0374 0.0343 0 0213 -0.0164 0.0376 0.0135 0.0504 0.06.3983 175 197 0,0014 -0.0046 0 0061 0 0014 -0.0397 0 0411 0 0103 0.0476 0.0579 00563 0.0177 0.0740 0 1031 0 0060 0 109186 142 211 0 0116 -0 0098 00214 00372 00271 0.0101 0 0485 -0.0054 0.0539 0 0907 0.1054 00147 0 0X44 0 1128 -01)281

    Mean Returns (10174 0 0020 0 0155 00125 0 0006 0 0120 0 0232 0 002) 0 0211 0 0517 0.0429 0 0088 0 0581 0 0382 0 0199

  • 7/24/2019 An Investigation of Financial Ratios in Predicting Firms Future Performance an Application of Pr's Methodology

    80/96

    iAm.i:7

    CDCO

    z

    v :

    &

    >

    k.* C C r- - ^f'. C >C (N

    v r- o -r cdr-* no

    I *-ca o y ;

    fl P* r*. f (**r- oonor- noo r- oo rs r- 5n CS

    ?9999

    r*. n **. 'C^1+-. = 5 9 9 0 0 9

    rs Onoooor-

    rs

    9 9 9 9 9

    P- r*. r*-.

    \o o rs o gor*oe o 00wo cn N - O^m~ rs

    c o 9 9 9 9 d

    X CCN c "Tc r f* - *t^ * >0 ***.**r

    2 c r s r s - r O' =- k9 9 9 9 9 9Ofl

    *2 rr, rs*r%rs SPrs

    g 0 0 0 p

    or>*nO' rs,. r vOp>nfS C\ e on 0

    d

    Qfi f

    d p d

    rs rT0000,, r*~, 00rs 0000rs 00chr- f*mri o -r N OO

    d p * p

    : 2 S ; o m

    w eS;S.;;5|!? 5 5 S C2I ; 5 S! Ori -= === 2 .-S =====5

    g ^ | o : 5 5 5 5 s 9 o ^ c 5 5 S = s o = i = = = ? T ^ ? = :i = 0' :r :? = = :? =

    ~ ao _

    _ . moooo nonor- rs d '

    Cnrs rr-y vn

    >s r> rs rs -* rs rs '

    r* r- r, c- y r s r s

    o OO-T 00 rs NTNr* .i - r r -. p~*w-i r.

    r O*r a

  • 7/24/2019 An Investigation of Financial Ratios in Predicting Firms Future Performance an Application of Pr's Methodology

    81/96

    of t he

    co

    pyright owner

    Furt her

    reproduct ion

    prohibit e

    dwit hout permission

    IAHI I S

    TRADING PROITIAI51I 11 Y I( ) ITU-: Ill- IKII l0 6 (Pr>() 5) and short for firms with

    Pr0.6 for long position (n= 1017) and Pr0.5 ft

    Months Long Short Hedge

    .3 0.0147 0.0064 0.0084

    6 0.0125 -0.0039 0.0164

    9 0.0123 0.0077 0.0047

    12 0.0513 0.0360 0.0153

    15 0.0593 0.0374 0.0219

    18 0.0609 0 . 0 2 0 0 0.0409

    24 0.0929 0.0457 0.0472

    .36 0.0950 0.0297 0.0653

    48 0.0864 -0.0152 0.1016

    54 0.0598 -0.0.388 00986

    60 0.0817 -0.0317 0 11.34

    72 0.0834 -00365 0.1198

    0 2500

    0.2000

    0 1500

    01000

    0.0500

    0.0000

    -0 0500

    -01000

    -0 1500

    -0.2000

    L I

    11AMlg I Short Hedge

    0.1200

    0 1000

    0.0800

    0.0600

    0.0400

    0.0200

    00000

    00 200

    -II 0-10011ong Slum Iledge

  • 7/24/2019 An Investigation of Financial Ratios in Predicting Firms Future Performance an Application of Pr's Methodology

    82/96

    TABLE X(CONTINUED)

    TRADING PROFITABILITY TO THE HEDGE PORTFOLIOS OVER HOLDING PERIODS OF 3.6 . 9. 12, 15, IX. 24. 36. 4X. 54.60 AND 72 MONTHS

    Panel C: Rcturns-Targcl Model with Pr> ( ) . 6 for long position (iv=386) and Pr0.5 for long position (n -1381) and Pring I Short I ledge

  • 7/24/2019 An Investigation of Financial Ratios in Predicting Firms Future Performance an Application of Pr's Methodology

    83/96

    O)CD

    zr.

    sfscis:

    2

    'TCd

    J /5

    i?;ifid

    cei -2O 2C_ =r r l O

    rti 80O feG UJ t/1

    x i s=rilu oX itH * Y.~ so o'OS .= ~o s r-^ 3 V> u

    a "Ca s o< J -s

    o "=~ --'w' */",* S _

    _ 3 X

    < = a-"C iG tn BP

    2 S'w o 5G -a

    iR

    v

    G ,XO ^EH LU

    oc

    sc

    o

    rs

    *r.

    r.

    (V

    i

    ur.r

    *i

    r * l

    t

    Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.

  • 7/24/2019 An Investigation of Financial Ratios in Predicting Firms Future Performance an Application of Pr's Methodology

    84/96

    In contrast to HL [1992], this study finds that the trading profitability of Retums-Target

    Pr (hereafter R-Pr) only dominates that o f Eamings-Target Pr (hereafter E-Pr) up to a 9-

    months holding period (see Panels A & B in Table 8 and Figure 3). After that, the

    trading profitability based on E-Pr continues to grow monotonically. On the other hand,

    the trading profitability based on R-Pr remains constant and eventually declines (see

    Panels C & D in Table 8 and Figure 3). Specifically, the excess returns observed from E-

    Pr are monotonic increasing with the holding period (from 2% to 24%) and persistent

    beyond 72 months. This suggests that E-Pr might only capture firms profitability

    prospects and fails to incorporate risk factors. As a result, the trading profitability

    generated by E-Pr is probably due to some omitted risk factors.

    The R-Pr based trading profitability curve demonstrates a reverse V curve with a

    peak at the holding period o f 9 months (5.6%),26 which signifies that the market slowly

    adjusts to the fundamental value. In fact, the excess returns for R-Pr are not gre ate r than

    zero when one considers the transactions cost, which ranges between 4.3% and 9.5%

    (Holthausen and Larcker [1992]). The implication of the trading profitability observed

    from R-Pr suggests that the market is efficient with respect to the use of accounting

    information.

    6Although the highest trading profitability of 9.3% is shown at 36-month, it is not significantly higher tlian tlwtof 9-month at the p-value of 0.21 (one-tail test).

    eproduced with permission of the copyright owner. Further reproduction prohibited without permission.

  • 7/24/2019 An Investigation of Financial Ratios in Predicting Firms Future Performance an Application of Pr's Methodology

    85/96

    6.2.3 .3 Benchmarking With Trading Profitability Patterns of Two Famou

    Risk Proxies: Earnings-to-price Ratio and Book-to-Market Ratio

    The conjecture that E-Pr proxies for some omitted risk factor is further evaluated

    by comparing it with two benchmarks, namely trading strategies based on eamings-to-

    price ratio (E/P) and book to market ratio (B/M) respectively. These two ratios are well

    documented in studies o f market anomalies as proxies for omitted risk factors (Lev

    [1978], Fama and French [1992], Bernard and Wahlen [1995] and others). E/P is the

    earnings per share excluding extraordinary items (COMPUSTAT annual data item #58)

    scaled by the price as o f December fiscal year-end. B/M is the book value of common

    equity (COMPUSTAT annual data item #60) scaled by the market value of common

    equity (product o f COMPUSTAT annual data items #25 and #199) as o f December fiscal

    year-end. The monthly rebalanced market adjusted abnormal returns for each trading

    strategy are equally-weighted pooled cross-sectional and time-series observation during

    1980-1986. That is, once all observations have been assigned to a portfolio, the mean for

    the portfolio is based on an equally-weighted average o f all observations in that

    portfolio.27 Subsequently, cumulative abnormal returns for the long and short position are

    calculated for various holding periods. As the E-Pr based trading strategy involves taking

    long positions fo r 1017 firm-year observations with E-Pr>0 .6 and taking short positions

    472 firm-year observations with E-Pr

  • 7/24/2019 An Investigation of Financial Ratios in Predicting Firms Future Performance an Application of Pr's Methodology

    86/96

    72

    (B/M) based portfolios is determined by taking long positions for 1017 firm-year

    observation with highest E/P (B/M) and taking short positions for 472 firm-year

    observation with lowest E/P (B/M).

    Imposing the same frequency of stocks in long and short position as in the

    eamings-target prediction model, both E/P and B/M models demonstrate similar but

    stronger trading profitability patterns than does E-P r (see Panel D in Table 9). Moreover,

    stocks with long position have significantly higher E/P and B/M than those with short

    position for the eamings-target model but not for the retums-target model. The result is

    consistent with Ou and Penman [1989b] in which they find similar properties between E-

    Pr and E/P and conclude that ...information in prices that leads (future) earnings is

    contained in financial statements [p. I l l ] ,

    6.2.3,4 Evaluation o f Firm Size Effects

    In light of continuous trading profitability based on the Pr strategy, Greig [1992]

    examines the firm size effect on the cross-sectional relation between Pr and subsequent

    stock returns. He finds that the positive associations between Pr and subsequent stock

    returns disappear after allowing for cross-sectional differences in CAPM beta and

    implementing tighter controls of firm size than those in OP [1989a], The hypothesized

    effect of E-Pr on trading profitability should be positive. When Greig regresses the E-Pr

    and firm size, defined as the log of the market value o f equity, on the mean adjusted buy-

    and-hold return for the twelve-months (Table 6, p. 432 in Greig [1992]), the significance

    eproduced with permission of the copyright owner. Further reproduction prohibited without permission.

  • 7/24/2019 An Investigation of Financial Ratios in Predicting Firms Future Performance an Application of Pr's Methodology

    87/96

    R

    d

    d

    ith

    i

    i

    fth

    iht

    F

    th

    d

    ti

    hibit

    d

    ith

    t

    i

    TABLE 9. COMPARISON OP TRENDS OF T R A D I N G PROFITABILITY TO THE HEDGE PORTFOLIOS BASED ON

    EARNINGS-TA RGET Pr. EARNINGS-TO-PRICE RATIO (E/P) RATIO, AND BOOK-TO-MARKET (B/M) RATIO

    The monthly rebalance d market adjusted returns associated with the trading strategies based on carnings-bascd Pr (E-Pr). carnings-pricc ratio, and

    bo ok -to -m ark ct ra ti o ar c eq ua lly -w eig ht ed PO OL ED cr os s-s ec tio na l an d lim c- sc ric s ob se rva tio ns ov er the tim e pe rio d 1980 -198 6. Th e tr ad in g s tra teg y

    by g oi ng lo ng for fir ms wi th E -P r >0 .6 an d sho rt for fir ms wit h E-P r- 0 4 re su lts in 1017 fir m s an d 472 fir ms res pec tiv ely . Th e tr ad in g str at egy for the

    E/P (B/M) based portfolios is determine d by taking long for 1017 firm-vcar observations with highest E/P (B/M) and taking short for 472 firm-year_______________________________________________________ observat ions with lowest E /P (B/M ).

    N u m b e r

    P a n e l A .

    M o n t h s

    o f o b s e r v a t i o n s i s 1 0 1 7 f o r l o n g p o s i t i o n a n d 4 7 2 f o r s h o r t p o s i t io n .

    T r a d i n g P r o f i ta b i l it y B a s e d o n P a m i n g s - t o - P r ic e R a t io ( P . /P )

    L o n g S h o r t 1 l e d g e

    .16

    9

    121 5

    1 8

    2 4

    3 6

    4 8

    5 4

    6 0

    7 2

    P a n e l B :

    M o n t h s

    0 0 1 9 5

    0 . 0 1 5 2

    0 . 0 2 7 9

    0 . 0 7 8 9

    0 . 0 9 0 5

    0.10000 . 1 5 5 8

    0 . 1 9 7 2

    0 . 1 9 1 2

    0 . 1 7 3 7

    0 . 1 8 8 7

    0 . 2 0 5 1

    0 . 0 0 9 1

    - 0 . 0 0 3 8

    - 0 . 0 3 2 3

    0 0 0 4 4

    -0.0017- 0 . 0 1 2 8

    - 0 . 0 0 6 8

    - 0 . 0 3 5 4

    - 0 . 0 4 9 3

    - 0 0 7 4 5

    - 0 . 0 4 0 3

    - 0 . 0 5 3 0

    0 . 0 1 0 4

    0 0 1 9 0

    0 . 0 6 0 2

    0 . 0 7 4 5

    0 . 0 9 2 3

    0 . 1 1 2 9

    0 . 1 6 2 6

    0 2 3 2 6

    0 . 2 4 0 5

    0 . 2 4 8 2

    0 . 2 2 9 0

    0 . 2 5 8 1

    0 . 3 0 0 0

    0 . 2 5 0 0

    02000

    0 1 5 0 0

    01000

    0 . 0 5 0 0

    0.0000

    - 0 0 5 0 0

    -0 1000 tUing Short -Hedge

    T r a d i n g P r o f it a b i li ty B a s e d o n B o o k - l o - M a r k e l R a t i o ( B / M )

    L o n g S h o r t I l e d g e

    3

    6

    9

    1215

    1 8

    2 4

    3 6

    4 8

    5 4

    6(17 2

    0 . 0 2 1 6

    0 . 0 1 7 4

    0 . 0 2 6 5

    0 . 0 8 5 5

    0 . 0 9 5 1

    0 . 0 9 3 1

    0 . 1 5 4 3

    0 . 1 8 3 5

    0 . 1 7 7 4

    0 . 1 5 5 4

    ( I 1 7 8 4

    0 I 8 I < >

    0 . 0 1 4 0

    - 0 . 0 0 9 3

    - 0.0120

    - 0 . 0 0 4 7

    - 0 . 0 0 1 8

    - 0 . 0 2 6 9

    - 0 . 0 2 6 1

    - 0 . 0 5 1 3

    - 0 . 0 7 9 2

    - 0 . 1 0 8 0

    - 0 0 9 2 8

    - 0 1 0 4 2

    0 . 0 0 7 7

    0 . 0 2 6 7

    0 . 0 3 8 5

    0 . 0 9 0 2

    0 . 0 9 6 9

    0 . 120(1

    0 . 1 8 0 4

    0 . 2 3 4 8

    0 2 5 6 7

    0 2 6 3 4

    0 . 2 7 1 1

    0 2 8 5 7

    0 3 0 0 0

    0 . 2 5 0 0

    02000

    0 1 5 0 0

    0.1000

    0 0 5 0 0

    00000

    - 0 0 5 0 0

    -01000

    - 0 1 5 0 0

    I

    I "HI! I I S h o rt -Hedge

  • 7/24/2019 An Investigation of Financial Ratios in Predicting Firms Future Performance an Application of Pr's Methodology

    88/96

    f

    th

    iht

    F

    th

    d

    ti

    hibit

    d

    ith

    t

    i

    i

    7

    TABLE 9 (CONTINUED): COM PARISON OF TRENDS OF TRADING PROFITABILITY TO THE HEDGE PORTFOLIOS BASED ON

    Panel C Trading Profitability for Earnings-!ai'gel PTf!7

    Months Long Short 1ledge0 2500

    3 0.0181 -0 0 0 1 0 0 0214

    6 0.0091 -0 009(i 0 0189 0 2 0 0 0

    9 0.0116 -00092 0 0207 0150012 0.0491 0 01.18 0 0151

    15 0.0612 0 0 ) 0 1 005110 . 1 0 0 0

    18 0.0629 -0.0058 0.0687

    24 0.1181 - 0 0 0 1 2 0.1193 0.0500

    .16 0.1440 -0.0172 0 16120 0 0 0 0

    48 0.1.114 -00511 0.182654 0.1007 -00715 0.1722 -0 050060 0 . 1 2 1 0 - 0 0621 0 1810

    72 0.1.190 -00784 0 2174 I Ling I ledge

    Panel D: Comparison of Trading Profitability to the Hedge Portfolios Based on Earnings-Targcl Pr (E-Pr), Earnings-to-Pricc Ratio (E/P), and Book-to-

    Markcl Ratio (B/M)

    0 3000 nB/M

    E/ P0 2500-

    E-Pr0 2000

    0 1500-

    01000-

    0 0500

    000006 1 5 24 363 9 12 1 8 5 44 8 6 0 7 2

  • 7/24/2019 An Investigation of Financial Ratios in Predicting Firms Future Performance an Application of Pr's Methodology

    89/96

    75

    of E-Pr disappears and becomes negative. The control variable, firm size, is negative and

    statistically significant. Moreover, when he forms a trading strategy merely based on size

    quintiles, the results show that the size strategy outperfo rms OPs Pr measure in predicting

    abnormal returns by going long in the smallest firms and short in the largest firms.

    This study complements Greigs study [1992] by further examining whether E-Pr

    or R-Pr proxies for size effect after controlling for the E/P and B/M effects. Table 10

    reports the results o f regressing the market-adjusted buy-and-hold returns for the holding

    periods o f 12, 24 and 36 month subsequent to the computa tion o f Pr. Similar to Greig, E-

    Pr is not significant in explaining market-adjusted buy-and-hold returns in the presence of

    firm size (measured as log o f market value) with or w ithout controlling for E/P, B/M (see

    Panel A in Table 10). However, R-Pr remains significant despite the inclusion of firm size.

    E/P and/or B/M (see Panel B in Table 10). One noticeable observation is that the

    significance of firm size in both models disappears as the holding periods prolong. On the

    other hand, the significance o f E/P and B/M remains robust.

    Reproduced with permission of the copyright owner. Further reproduction prohibited withou t permission.

  • 7/24/2019 An Investigation of Financial Ratios in Predicting Firms Future Performance an Application of Pr's Methodology

    90/96

    76

    TABLE 10

    COMPARISON OF CONDITIONAL COEFFICIENTS (T STATISTICS) OF EARNINGS-TARGET

    PR (E-PR) AND RETURNS-TARGET PR (R-PR)

    The sample comprises 2.914 December year-end industr ial firms from 1980-1986. AR,t is the market-

    adjusted buy-and-hold return for the holding periods of 12 (24,36) months subsequent to thecomputation of E-P rlt (R-Prlt) for firm / at time t. E-P rIt (R-Pr,t) is the estimated probability o f an

    earnings (returns) increase in year r. Size,t is the log of market value of equity as of December fiscal

    year-end. E/Plt is earnings per share excluding extraordinary items (Compustat annual data item #58)

    scaled by the price (Compustat annual data item #199) as o f December fiscal year-end. B/M,t is the

    book value o f common equity' (Compustat annual data item #60) scaled by the market value o f common

    equity' (product of Compustat annual data items #25 and #199) as of December fiscal vear-end.

    Panel A: AR;t = Yo + Yi E-Pr;, + Yi E/P;, + 7 3 B/M ;t + Y4 Size;,

    Ym Yi Y2 Y3 Y-v Adj-R2

    Holding Period = 12 months

    0.024

    (0.692)

    0.042

    (0.857)-0.092

    (-2.567)b0.0019

    -0 . 0 2 2

    (-0.611)

    0.151

    (0.889)

    0.165

    (3.08)c

    0.027

    (2.524)b

    -0.008

    (-2.089)b0.0073

    Holding Period = 24 months

    0.058

    ( 1 .0 2 2 )

    0.046

    (0.565)

    -0.017

    (-2.842)'

    0.0023

    -0.046

    (-0.758)

    0.058

    (0.696)

    0.398

    (4.497)c

    0.06

    (3.406)'

    -0.014

    (-2.322)b

    0.014

    Holding Period = 36 months

    0.054

    ( 0 622)

    -0 . 0 2 1

    (-0.169)-0.009

    (-0.981)

    -0.0004

    -0.099

    (-1.074)

    -0.017

    (-0.131)

    0.53

    (3.808)c

    0.091

    (3.402)'

    -0.004

    (-0.423)

    0.0092

    Panel B: ARIf = y0+ y xR-Pr;t + 7 2 E/P;, + 7 3 B/M;, + 7 4 Size;,

    Y" Yi Y: Y3 Y4 Adj-R2

    Holding Period = 12 months

    -0.024

    (-0.65)

    0.152

    (2.299)b

    -0 . 0 1

    (-2.784)'

    0.0035

    -0.072

    (-1 743)-

    0.151

    (2.269)b

    0.148

    (2.793 )c

    0.03

    (2.882)'

    -0.008

    (-2.181)b

    0.0088

    Holding Period = 24 months

    -0.013

    (-2.215)

    0.208

    (1.889)

    -0.018

    (-3.014)'

    0.0034

    -0.113

    (-1 642)

    0 . 2

    (1.820)

    0.376

    (4.293 )c

    0.064

    (3.71)'

    -0.015

    (-2.393 )b

    0.015

    Holding Period = 36 months

    - 0 . 1 1 1

    (-1.174)

    0.331

    (1.979)b

    -0 . 0 1

    (-1.15)

    0 . 0 0 1

    -0.263

    (-2.498)b

    0.324

    (1.935)

    0.511

    (3.71 l)c

    0.095

    (3.576)'

    -0.005

    (-0.535)

    0.0106

    l c denotes statistical significance levels at 0.10. 0.05 . and 0.01 respectively.

    Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.

  • 7/24/2019 An Investigation of Financial Ratios in Predicting Firms Future Performance an Application of Pr's Methodology

    91/96

    Chapter 7

    CONCLUSIONS

    Ou and Penman [1989a] form a trading strategy based on the predicted probability

    of future earnings increase estimated by a set of financial ratios that can predict one-year-

    ahead earnings change. In contrast , Holthausen and Larcker [1992] extract information

    from financial ratios that can directly predict one-year-ahead excess returns to estimate the

    probability o f future excess returns, which is then used to form a trading strategy. Both

    trading strategies are able to generate substantial abnormal returns. However, the

    persistence of abnormal returns resulted from using these P r methods has puzzled

    researchers concerning the constructs o f Pr.

    Based on a typical firm valuation model, returns can be expressed as a function of

    the change o f earnings magnitude and a discount factor that reflects risk. If the change of

    earnings magnitude captures mainly profitability, then, the construct of eamings-target Pr

    (hereafter E-Pr) should be mainly related to profitability and retums-target Pr (hereafter

    R-Pr) should be affected by both profitability and risk. Hence, comparing the relationships

    between a common set o f financial ratios that compose Pr and the two prediction model

    allows one to characterize the attributes o f the financial ratios and to evaluate the relative

    importance of profitability and risk in predicting future unexpected earnings and future

    excess returns.

    77

    Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.

  • 7/24/2019 An Investigation of Financial Ratios in Predicting Firms Future Performance an Application of Pr's Methodology

    92/96

    78

    This study uses principal component analysis to identify a set of essentially

    uncorrelated financial ratios that a ttempts to capture distinct aspects o f firm

    characteristics. Because the set of financial ratios only depends on the interrelationship

    among a large number o f financial ratios to form Pr without considering the association of

    the ratios and the pred iction targets (i.e. the sign o f one-year ahead earnings change /

    excess return), they can be applied in different methodologies. Using principal component

    analysis, this study identifies a set of 16 essentially uncorrelated financial ratios from 59

    financial ratios that were originally investigated by OP and HL. Based on a simple firm

    valuation model, these ratios are subsequently classified into one of three categories:

    profitability-oriented ratios, risk-oriented ratios and ratios with a mixture of risk and

    profitability.

    Because the targets that the financial ratios aim to predict are different for the

    eamings-target prediction model and the retum s-target prediction model, investigating the

    patterns of abnormal returns or trading profitability generated by eamings-target Pr (E-

    Pr) and retums-target Pr (R-Pr) can provide insights regarding their respective economic

    meanings. As expected, this study finds that the construct o f E-Pr different from that o f

    R-Pr. Similar to eamings-to-price and book-to-market ratios, E-Pr proxies for some

    omitted risk factors. Yet, R-P r supports the premise that operating profitability and

    growth in operating assets are the two most important pieces of accounting information to

    evaluate firm value and tha t the market is efficient in using accounting information. This

    study complements Greigs study [1992] by further examines whether Pr is merely a

    Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.

  • 7/24/2019 An Investigation of Financial Ratios in Predicting Firms Future Performance an Application of Pr's Methodology

    93/96

    79

    manifestation o f the firm size effect. E-Pr is not significant in explaining market-adjusted

    adjusted buy-and-hold returns in the presence of firm size (measured as log of market

    value), eamings-to-price ratio, and book-to-market ratio. However, R-Pr remains

    significant despite the inclusion o f firm size, eamings-to-price ratio, and book-to-market

    ratio.

    The understanding o f Pr s constru cts provides an explanation for the economic

    meanings o f Pr and helps us to explain the market anomaly as reported in previous studies

    Furthermore, it provides evidence as to the relative importance o f profitability or risk in

    predicting future unexpected earnings and future excess returns.

    Several limitations in this studyare: First, this study attempts to capture all relevant

    aspects of firm characteristics from a relatively complete set of financial ratios, but there is

    no guarantee that the set of ratios used in this study is complete . Principal component

    analysis is a variable-sensitive statistical method; different financial patterns may be

    generated as different financial ratios are fitted in the model. Second, the criteria for

    selecting a financial ratio as a surrogate for each identified factor is somewhat subjective

    The appropriateness of the ratios selected for proxying the factors and the potential

    correlation among these ratios are also limitations to this study. Finally, this study is

    limited to manufacturing firms with December 31 fiscal year-end. Therefore, the results

    may not be generalized to other types o f firms and firms without a December 31 fiscal

    year-end.

    eproduced with permission of the copyright owner. Further reproduction prohibited withou t permission.

  • 7/24/2019 An Investigation of Financial Ratios in Predicting Firms Future Performance an Application of Pr's Methodology

    94/96

    BIBLIOGRAPHY

    Ball, R., 1992, The eamings-price anomaly, Journal o f Accounting & Economics 15, 319-345.

    Banz, R., 1981, The relationship between return and market value o f common stocks.

    Journal o f Financial Economics 9, 3-18.

    Bernard, V.L., 1987, Cross-sectional dependence and problems in inference in market-

    based accounting research, Journal of Accounting Research 25, 1-48.

    Bernard, V.L., 1993, Earnings qualify, book value, and the state o f financial statement

    analysis, in Earnings Quality, edited by S. A. Butler, The Center for Economic and

    Management Research, Norman, Oklahoma.

    Bernard, V.L., J. Thomas, and J. Wahlen, 1995, Accounting-based stock price anomalies:

    Separating market inefficiencies from research design flaws. Working Paper.

    Bowerman, B.L. and R.T. O Connell, 1990, Linear Statistical Models, Boston. MA:

    PWS-KENT Publishing Company.

    Chen, K.H. and T.A. Shimerda, 1981, An empirical analysis of useful financial ratios.

    Financial Management 10, 51-60.

    Cheng, C.S.A., S.H. Liu and T.F. Schaefer, 1996, Earnings permanence and the

    incremental information content of cash flows from operations. Journal ofAccounting Research 34, 173-181.

    Courtis, J.K., 1978, Modeling a financial ratios categorical framework, Journal of

    Business Finance & Accounting 5, 371-386.

    Devine, K. and L. Seaton, 1989, A comparison of quarterly and annual financial ratio

    stability: A factor analysis, Working Paper, Univ. of Nebraska-Lincoln.

    Dunteman, G.H., 1989, Principal Component Analysis, Beverly Hills: Sage Publications.

    80

    Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.

  • 7/24/2019 An Investigation of Financial Ratios in Predicting Firms Future Performance an Application of Pr's Methodology

    95/96

    Elgers, P.T., and R.J. Pfeiffer, Jr., 1996, Pr-Estimates for predicting and evaluating

    earnings changes: A Re-assessment, Working Paper, Univ. o f Massachusetts atAmherst.

    Fama, E.F., and K.R. French, 1992, The cross-section of expected stock returns, Journalo f Finance 47, 427-465.

    Feltham. G.A. and J.A. Ohlson, 1995, Valuation and clean surplus accounting for

    operating and financial activities. Contemporary Accounting Research 11, 689-731.

    Freeman, R., J. Ohlson, and S. Penman, 1982, Book rate of retu rn and prediction of

    earnings changes: An empirical investigation, Journal o f Accounting Research 20,639-653.

    Gombola, M.J., and J.E. Ketz, 1983, Financial ratio patterns in retail and manufacturing

    organizations, Financial Management 12, 45-56.

    Greene, W.H., 1993, Econometric Analysis, New York, NY: Macmillian Publishing

    Company.

    Greig, Anthony, 1992, Fundamental analysis and subsequent stock returns. Journal of

    Accounting & Economics 15, 413-442.

    Hair, J.F., R.E. Anderson and R.L. Tatham, 1987, Multivariate Data Analysis, New York.

    NY: Macmillian Publishing Company.

    Holthausen, R.W. and D.F. Larcker, 1992, The prediction o f stock returns using financial

    statement information, Journal o f Accounting & Economics 15, 374-411.

    Kanto, A.J. and T. Martikainen, A test on a priorifinancial characteristic of the firm,European Journal o f Operational Research 57, 13-23.

    Larcker, D.F., 1989, Discussion of accounting measurement, price-eamings ratios, and the

    information content o f security prices, Journal of Accounting Research 27, 145-

    152.

    Laurent, 1979, Improving the efficiency and effectiveness o f financial ratio analysis.

    Journal o f Business Finance & Accounting 6, 401-411.

    Lev, B. and R. Thiagarajan, 1993, Fundamental information analysis, Journal of

    Accounting Research 31,190-215 .

    with permission of the copyright owner. Further reproduction prohibited without permission.

  • 7/24/2019 An Investigation of Financial Ratios in Predicting Firms Future Performance an Application of Pr's Methodology

    96/96

    82

    Libby, R., 1975, Accounting ratios and the prediction o f failure: Some behavioral

    evidence, Journal of Accounting Research 13, 150-161.

    Liao. T.F., 1994, Interpreting Probability Models: Logistic. Proit. and Other Generalized

    Linear Models, Thousand Oaks, CA: Sage.

    Menard, S.W., 1995, Applied Logistic Regression Analysis, Thousand Oaks, CA: Sage.

    Ou, J.A. and S. H. Penman, 1989a, Financial statement analysis and the prediction o f

    stock returns, Journal of Accounting & Economics 11, 295-329.

    Ou, J.A. and S. H. Penman, 1989b, Accounting measurement, price-eamings ratio, and

    the information content o f security prices, Journal of Accounting Research 27

    (Supplement), 111-144.

    Ou, J.A., 1990, The information content ofnoneam ings accounting numbers as earnings

    predictors, Journal o f Accounting Research 28, 144-163.

    Penman, S . H., 1992, Financial statement information and the pricing o f earnings changes.

    The Accounting Review 67, 563-577.

    Pinches, G.E., K.A. Mingo, and J.K. Caruthers, 1973, The stability o f financial patterns in

    industrial organizations. Journal of Finance 28, 389-396.

    Pinches, G.E., A.A. Eubank, K.A. Mingo, and J.K. Caruthers, 1975, The hierarchical

    classification o f financial ratios, Journal o f Business Research 3, 295-3 10

    Seber, G.A.F., 1984, Multivariate Observations, New York, NY: John Wiley & Sons

    Stevens, D.L., 1973, Financial characteristics o f merged firms: A multivariate analysis,

    Journal o f Financial and Quantitative Analysis, 149-158.

    Stober , T.L., 1992, Summary financial statement measures and analysts forecasts of

    earnings, Journal of Accounting & Economics 15, 347-372.