Dividends, Earnings, and Stock Prices Author(s): M. J ...

[Pages:37]Dividends, Earnings, and Stock Prices Author(s): M. J. Gordon Source: The Review of Economics and Statistics, Vol. 41, No. 2, Part 1 (May, 1959), pp. 99-105 Published by: The MIT Press Stable URL: Accessed: 21/03/2010 12:05 Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at . JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at . Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission. JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@.

The MIT Press is collaborating with JSTOR to digitize, preserve and extend access to The Review of Economics and Statistics.



DIVIDENDS, EARNINGS, AND STOCK PRICES

M. J. Gordon*

T HE threepossiblehypotheseswithrespect attractivebuy at its currentprice.2 Grahamand

to what an investor pays for when he ac- Dodd go so far as to state that stock prices

quires a share of common stock are that he is should bear a specified relation to earnings and

buying (i) both the dividendsand the earnings, dividends,but they neitherpresent nor cite data

(2) the dividends,and (3) the earnings. It may to support the generalization.3 The distin-

be argued that most commonlyhe is buying the guished theoretical book on investment value

price at some future date, but if the future price by J. B. Williams contains several chapters de-

will be related to the expected dividends and/or voted to the application of the theory, but his

earnings on that date, we need not go beyond empiricalwork is in the tradition of the invest-

the three hypotheses stated. This paper will ment analyst's approach.4The only study along

critically evaluate the hypotheses by deriving the lines suggested here that is known to the

the relation among the variables that follows writer is a recent one on bank stocks by David

from each hypothesis and then testing the the- Durand.5

ories with cross-section sample data. That is, In contrastwith the dearthof publishedstud-

price, dividend, and earnings data for a sample ies the writer has encountered a number of

of corporationsas of a point in time will be used unpublished cross-section regressions of stock

to test the relation among the variables predict- prices on dividends, earnings, and sometimes

ed by each hypothesis.

other variables. In these the correlationswere

The variation in price among commonstocks high, but the values of the regressioncoefficients

is of considerable interest for the discovery of and their variation among samples (different

profitableinvestmentopportunities,forthe guid- industries or different years) made the eco-

ance of corporate financial policy, and for the nomic significanceof the results so questionable

understandingof the psychology of investment that the investigators were persuaded to aban-

behavior.' Although one would expect that this don their studies. There is reason to believe

interest would find expression in cross-section that the unsatisfactory nature of the findingsis

statistical studies, a search of the literature is due in large measure to the inadequacy of the

unrewarding.

theory employed in interpretingthe model, and

Cross-section studies of a sort are used ex- it is hoped that this paper will contribute to a

tensively by security analysts to arrive at buy more effective use of cross-section stock price

and sell recommendations.The values of certain studies by presenting what might be called the

attributes such as the dividend yield, growth in elementary theory of the variation in stock

sales, and managementability are obtained and prices with dividends and earnings.

comparedfor two or morestocks. Then, by some Before proceeding,it may be noted that there

weighting process, a conclusion is reached from have been some time series studies of the varia-

this information that a stock is or is not an tion in stock prices with dividends and other

* The researchfor this paperwas supportedby the Sloan variables. The focus of these studies has been

ResearchFund of the School of IndustrialManagementat the relation between the stock market and the

Massachusetts Institute of Technology. The author has benefited from the advice of Professors Edwin Kuh, Eli Shapiro,and GregoryChow. The computationswere done

business cycle6 and the discovery of profitable

2Illustrationsof this method of analysismay be found

in part at the M.I.T. ComputationCenter.

in texts on investmentanalysissuch as: Grahamand Dodd,

'Assumethat the hypothesisstockprice,P f (xi, X2,...), SecurityAnalysis,3rd ed. (New York, i95i); and Dowrie

is stated so that it can be tested, and it is found to do a good job of explainingthe variationin price amongstocks.

and'

Fuller,Investments(New York, I94I). GrahamandDodd, op. cit., 454 ff.

The model and its coefficientsthereby shed light on what investorsconsiderand the weight they give these variables

'The Theory of Investment Value (Cambridge,I938). 5Bank StockPricesand the BankCapitalProblem,Occa-

in buying commonstocks. This informationis valuableto sional Paper 54, National Bureau of Economic Research

corporationsinsofar as the prices of their stocks influence (New York, I957).

their financialplans. It is also true that a stock sellingat a

'J. Tinbergen,"The Dynamics of Share-PriceForma-

price above or below that predictedby the model deserves tion," this REVIEW,XXi (November I939), 153-60; and

specialconsiderationby investors.

Paul G. Darling,"A SurrogativeMeasureof BusinessCon-

[99]

Ioo

THE REVIEW OF ECONOMICS AND STATISTICS

investment opportunities.7They have not been dend coefficientis consideredan estimate of the

concernedwith explainingthe variation in price rate of profit, we want to know whether the

among stocks, and it is questionable whether estimate is reasonableon groundsbroaderthan

such data can be effectively used for this pur- statistical significance. Good preferred stocks

pose. Auto-correlationin the time series would sold in these years at dividendyields of four to

impair the significance of the regression coeffi- five per cent, and companies acquired in merg-

cients for many of the variables. Possibly even ers were purchased for about five times their

more important, the use of time series assumes earnings before income taxes. Therefore, we

that the coefficientof a variableis constantover would expect the rate of profit on common

time but differentamong stocks. The exact op- stocks to fall between four and ten per cent and

posite is assumed in any attempt to explain the coefficient in question to fall between ten

preference among investment opportunities. and twenty-five. Further, we would expect a

The Sample

particularrank in the coefficients. Corporations in the chemical industry are considered to have

To test each of the theories, price, dividend, the advantages of, size, growth, and stability;

and earnings data were obtained for four in- foods represent an industry that is considered

dustries and two years, so that there are eight stable; steels represent an industry with large

samplesin all. The years chosenwere I95I and corporationswhich are consideredvulnerableto

I954, and the industries and number of cor- cyclical fluctuations; and machine tools repre-

porations for each industry are Chemicals, 32; sent an industry of comparatively small cor-

Foods, 52; Steel, 34; and MachineTools, 46. porationswhich are also vulnerableto the busi-

Including only those corporationswhich con- ness cycle. Accordingly, one might expect the

formed to a narrow definition of the industries rate of profit to vary among the industries in

mentioned did not provide samples of adequate the order just given. Further, I95I was a year

size. Therefore, certain fringe classifications of war profits with the outlook for the future

were included in each category. For instance, somewhat uncertain. By contrast, while there

Chemicals includes pharmaceutical manufac- was some talk of recession in I954, there was

turers,and Steel includesforgingmanufacturers little evidence that the high level of income ex-

and certain other fabricatorsof steel as well as tending back a number of years would fall

the basic steel producers. In general, while the sharply in the near future. Accordingly, one

corporations included in each sample can be might expect that the coefficients would differ

consideredto comeunderthe label, there is con- in a predictablemannerbetween the two years.

siderable variation among them in such attri-

butes as size, profitability,structureof the mar-

DividendsandEarnings

kets in which they buy and sell, and investor Given the task of explaining the variation in

status.8

price among common stocks, the investigator

The use of eight samples rather than one may observe that stockholdersare interested in

provides a more rigoroustest of the hypotheses. The industry and year selection of the data has the further advantage of allowing the use of a priori economic knowledge in evaluating the

both dividend and income per share and derive immediately from this observation the model:

P = ao + a,D + a2Y

(I)

regression statistics. For instance, if the divi- where P = the year-end price, D = the year's

fidence and Its Relation to Stock Prices," Journal of Finance, dividend, and Y = the year's income. The

x (December I955), 442-58. 'The outstanding example of this is The Value Line In-

vestment Survey. In addition, numerous articles in the Analysts Journal and the Journal of Finance analyze the change over time of price with other variables. A paper of some interest is D. Harkavy, "The Relation Between Retained Earnings and Common Stock Prices for Large, Listed

Corporations," Journal of Finance, viii (September I953),

equation may be considered of interest solely for the multiple correlation between the actual and predicted price, in which case no meaning can be given to the regression coefficients. Alternatively, the equation may be read to mean that the coefficientsa, and a2representthe value

I83-97. 'A list of the corporations and a description of how they

were selected may be obtained from the writer on request.

the market places on dividends and earnings respectively, a possible objective being the

DIVIDENDS, EARNINGS, AND STOCK PRICES

measurementof the relative importanceof the

The DividendHypothesis

two variables. However, a share of stock like The hypothesis that the investor buys the

any other asset is purchased for the expected dividendwhenhe aquiresa shareof stock seems

future income it provides. This income may be intuitively plausible because the dividend is

the dividendor it may be the earningsper share, literally the payment stream that he expects to

but it cannot be both. The model is therefore receive. In implementingthe hypothesis it must

conceptually weak.

be recognized that the stockholder is interested

The unfortunate consequence of this prag- in the entire sequence of dividend payments

matic approachto the measurementof the vari- that he may expect and not merely the current

ation in stock prices with dividendand earnings value. For the purposeof arrivingat an opera-

is illustrated by the data of Table i. The divi- tional model we may represent this infinite se-

dend coefficient for chemicals in I95I is nega- quence by two quantities, one the current divi-

tive and machine tools has the highest coeffi- dend and the other a measure of the expected

cient. Between I95I and I954 the chemicals growth in the dividend.

coefficient changes from approximately zero to Among the events which will lead to an in25. Many of the dividend coefficientsare ma- crease in a corporation'sdividend are: successterially below ten, and in I954 the highest co- ful tradingon its equity, an increasein its return

efficient is five times the lowest. The income on investment, and selling additional common

coefficients,with the exception of chemicals in stock when the rate of profitthe corporationcan

I95I, are extraordinarily low as measures of earn is above the rate at which its stock is sell-

the price the market is willing to pay for earn- ing. However, there is no doubt that the most

ings.

importantand predictable cause of growth in a

corporation'sdividendis retainedearnings. For

TABLE I.- MODEL I, REGRESSION OF PRICE ON DIVIDEND AND INCOME

those interested in a more rigorousargumentit has been shown that if a corporationis expected to earn a return r on investment and retain a

Sample

Constant term

I95I-Chemicals

-7.0

Foods

.J

Steels

5.5

Machinetools 2.4

I954- Chemicals -3.0

Coefficientand standarderrorof

D

Y

-.8

(5.2) 7.0

(I-5) 6.6

(i.8) I2.0

(I.2) 25.7

I6.7

(3.I) 5.5

(.9)

2.0

(.6) .8

(.5) .3

Multiple correlation

.93 .90 .86

fraction b of its income, the corporation'sdividend can be expected to grow at the rate br.9 If the investment or book value per share of common stock is B, then

br Y-D y

Y-D

(2)

Y

B

B(2

.90

Investors are interested in growth and not rate

.92 of growth, since a high rate of growth starting

Food

-4

(5.2)

(3.3)

with a low initial value will pay off in the heavi-

I0.4 (2.2)

5.6 (I.0)

.9I

ly discounted distant future, and it will not be

Steels

8.7

8.4

2.0

*94 as attractive as a lower rate of growth starting

Machine tools 6.3

(I -7) 5.5

(I-4)

(.8) 4.I

.89 from a higher initial value. Therefore, in a

(.6)

model where price and dividend are absolute

quantities, it is likely that retained earnings

per share without deflation by book value is a

Machine tools in I95I and chemicals in I954 bettermeasureof growththanthe rateof growth.

have incomecoefficientsthat are not significant- The previous discussion has provided the

ly different from zero, and three of the other economicrationalefor using the equation

coefficients are materially below five. Armed only with the theory just stated, it would be

P = ao + a, D + a2 (Y-D)

(3)

most difficult to infer from the data the existence of a logical structure in the pricing of common stocks.

'The argument is developed more fully in M. J. Gordon and Eli Shapiro, "Capital Equipment Analysis: The Required Rate of Profit," Management Science, DiI (October

I956), I02-IO.

I02

THE REVIEW OF ECONOMICS AND STATISTICS

to represent the hypothesis that the investor is a linear functionof the same variables,so that

buys the dividend when he acquires a share of they both yield the same correlationcoefficients.

stock. The reciprocalof the dividendcoefficient The earnings and retainedearnings coefficients

may be looked on as an estimate of the rate of are the same, since the change in earningsis the

profit the market requires on common stocks same as the change in retained earnings when

without growth, and the retained earnings co- the dividendis held constant. The differencein

efficient is the estimate of what the market is the dividend coefficientsis due to the fact that

willing to pay for growth.

in equation (i) the increasein dividendinvolves

Table 2 presents the eight sample estimates a correspondingreduction in retained earnings,

of the model's coefficients. The I95i dividend whereas in equation (3) retained earnings is

coefficients are considerably superior to those held constant.

of Model I under the criteria stated earlier for their absolute and relative values. Only the machinetools coefficientappears comparatively

TABLE 2. - MODEL II, REGRESSION OF PRICE ON DIVIDEND AND RETAINED EARNINGS

high. The I954 coefficientsvary among the industries as expected and they fall within the expected range. The spread in the coefficients is only one-half the range of those in Model I,

but it still seems quite large. In particular one

Sample

I 95 I-Chemicals Foods

Constant term -7.0 .1

Coefficientand standarderrorof

D

Y-D

I5.9

I6.7

(2.7)

(3 .)

I 2.5

5-5

Multiple correlation

.93 .90

might wonder at the high chemicals-I954 coefficient,the low steels-i95I and machinetoolsI954 values, and the strong inverse correlation

Steels

(I.I)

5.5

8.6

(I.5)

Machine tools 2.4

I2.8

(.9)

2.0

.86

(.6)

.8

.90

between the coefficientsand the constant terms.

(I.0)

(.5)

Turning now to the retained earnings coeffi- I954- Chemicals -3.0

30.0 (2.6)

.3

.92

(3-3)

cients, what would we expect of them? Since

Foods

-.4

I5.9

5.6

.9I

they representthe price the market is willing to pay for growth in the dividend, with retained earningsserving as an index of growth,the only

Steels

(I.5)

(I.0)

8.7

I0.4

2.0

.94

(I -4)

Machine tools 6.3

9.6

(.8)

4.I

.89

statement with respect to their values that fol-

(I.2)

(.6)

lows from the theory is that they should be

positive. It may be thought nonetheless that The dividend hypothesis provides a more

their values seem low, and the absence of sta- reasonable interpretationof equation (i) than

tistical significanceat the five per cent level for the interpretationgiven in the previous section.

two coefficients,machine tools-I95I and chem- If growth is valued highly, an increase in the

icals-Ig954,is particularlydisturbing. The really dividend with a correspondingreduction in re-

surprising result is the negative chemicals co- tained earningswill not increase the value of a

efficients for I954. On the other hand there is share as much as when a low value is placed on

some a prioricredibilityin the findings. Growth growth. There is some tendency for the a, co-

is most uncertain and it becomes quantitatively efficientsto vary among industries accordingly.

importantby comparisonwith the current divi- Another point to be noted is that the standard

dend in the distant future. Also, apart from the error of a, is below that for a,. This combined

I954 chemicals there is a roughcorrespondence with the higher values of the formercoefficients

between the rank of the coefficientsand notions means that the change in price with the divi-

as to the comparative stability of earnings dend can be predicted with much greater accu-

amongthe industries.

racy when retained earnings are held constant

The reader may have noted (i) the multiple than when the increase comes out of retained

correlationcoefficientsin Tables i and 2 are the earnings.

same for each industry year, (2) the earnings and retained earningscoefficients,a2and a2 are

The EarningsHypothesis

the same, and (3) the dividendcoefficienta, = a, The thirdhypothesis is that the investorbuys

+ a2. On the first point, in both equationsprice the income per share when he acquires a share

DIVIDENDS, EARNINGS, AND STOCK PRICES

I03

of stock. The rationale is that regardless of whether they are distributed to him the stock-

00

Po = I(I-b)Yte

ktdt

0

holder has an ownership right in the earnings per share. He receives the dividend in cash and

fC (I-b)Yoebrte-ktdt.

0

(4)

the retained earnings in a rise in the share's The price of the share is finite and the integravalue, and if he wants additional cash he can tion may be carriedout if k > br, in which case always sell a fraction of his equity. In short,

the corporateentity is a legal fiction that is not material with respect to his rights in the cor-

= (I-b) k -br

Y?

(5)

poration or the value he places on them.'0 One It may be noted that if k = r, equation (5) can argue further that the different tax treat- reduces to ment of dividends and capital gains creates a

stockholderpreferencefor retainedearnings. The hypothesis may be tested by referenceto

Po = i Y?'

(6)

the data of Table 2. If the investor is indiffer- but this is not relevant to the question at issue.

ent to the fraction of earnings distributed, the For the earnings hypothesis to be valid, it is

dividend and retained earnings coefficients of necessary that k be independentof b. That is,

Model II should be the same. However, with the rate of profit requiredby the marketshould

the exception of chemicals-I95 I the difference be independent of the fraction of income re-

between the coefficients is statistically signifi- tained.

cant. Durand's bank stock study presents the We could reasonas follows. A necessarycon-

same picture on this question."

dition for the price of a stock to be finite is

Since the proposition that the rate of profit k > br. This condition is most easily satisfied

at which a common stock sells is independent if k is an increasing function of br, and if this

of the dividend rate has some intuitive merit, a is true we would also expect that k will vary

theoreticalexplanationof the statistical findings with b. Other things equal, the rate of profit

presentedabove is of interest. The firstpoint to required on a common stock will vary for a

be noted is that the dividend hypothesis is cor- corporation and among corporations inversely

rect regardlessof whether the earningshypoth- with the dividendrate.

esis is correct.The only point at issue is whether An argument with considerably more the-

the dividend hypothesis is unnecessary. Can oretical content can be derived from the two

one study the pricing of common stocks and related questions without considering the fraction of incomepaid in dividends? It is therefore possible to investigate the problem by using a more rigorous formulation of the dividend hy-

following assumptions, both of which appear reasonable. (i) The rate at which a future payment is discounted increases with its uncertainty; and (2) the uncertaintyof a futurepayment increases with the time in the future at which it will be received. It follows that the rate of

pothesis to establish the condition for the valid- profit at which a stream of expected payments

ity of the earningshypothesis.

is discountedis really an average of rates, each

Let k be the rate of profit at which a stock is weightedby the size of the payment. The larger

selling, Yt the income expected in year t, b the the distant payments relative to the near pay-

fraction of income the corporationis expected ments, the higher the average rate that equates

to retain, and r the rate of profit it is expected the streamof paymentswith the price, the latter

to earn on investment. The corporation'sdivi- obtainedby discountingeach future paymentat

dend is expected to grow at the rate br, and the its appropriate rate. The relative size of the

price of the stock at t = o is:

distant payments will of course vary with the

'LThis appears to be a widely held point of view in the economics literature. See for example Lutz and Lutz, The Theory of Investment of the Firm (Princeton, I95I). The question is nowhere considered explicitly, but it is implicit

in the material treated on pages 155 f. I Durand, op. cit., Io-i I.

rate of growth. Therefore, given the current earnings, the rate of profit required on a share increases with the fraction of income retained. The same reasoningprovidesan explanationfor the tendency of interest rates on bonds to in-

I04

THE REVIEW OF ECONOMICS AND STATISTICS

crease, other things being the same, with the retainedearningsmaybe expectedto varyamong

maturityof the bond.

corporationswith the confidencein the dividend

stream. This would suggest that the price of a

Refinementsin theModel

share varies with other variables such as the

Equation(3) is an extremelysimple and crude size of the corporation,the relation of debt to

expression of the dividend hypothesis, and in- equity, and the stability of its earning record.

sofar as the values of the coefficientsare suspect, Insofar as the values of these variables vary

it may be due to limitations of the model. In among industries, failure to include them intro-

this section we shall discuss the more important duces variation and error in the dividend and

limitations,suggest how they may be dealt with, retainedearningscoefficients.

and then present data for a model that attempts 4. In the presentmodel the variationin price

to overcomesome of these limitations.

with growth in the dividend is estimated by

i. Correlation between the variables and using an index of growth, retained earnings, as

variation in the coefficientsamong industries is the independentvariable. A model in which it is

due in part to the scale factor. The problem possible to use the rate of growth itself might

may be stated as follows. Assume a sample of yield better results. More important,the defini-

n corporationsfor all of which the dividend is tion of the rate of growth has considerablethe-

the same, the price differs among the shares, oretical merit to date nothing superior has

and the average of the prices is higher than the been proposed but there are empirical prob-

dividend. There is no correlationbetween divi- lems involved in using it. Variationin account-

dend and price. However, if n numbers are ing practice among firms makes the use of book

selected at random and the price and dividend value as a measure of return on investment

of each share is multipliedby one of these num- questionable. Also, the instability of corporate

bers, correlation between the variables will be retained earnings and the possibility that they

created. Further, if each of the n randomnum- vary over time differentlyamongindustriesmay

bers is first multiplied by a constant greater make the use of past values to predict the future

than one, the correlationand the regressionco- an heroic assumption. This is particularly true

efficient will be larger the larger the value of if investors give considerableweight, rationally

this constant. The presence of so-called high- or otherwise, to other variables in predicting

pricedand low-pricedstocks in a samplereflects future earnings. in some part this scale factor. It is possible that Table 3 presents the regression statistics for by deflatingthe data, say by book value, and/or the followingmodel

usinglogs we will moderatethe influenceof scale

P=po+/

d+ 32 (d-d)

on the coefficients.

+/33g+34

(g-g).

(7)

2. The independentvariablesin equation(3)

are the currentvalues of dividendsand retained In this equation:

earnings. These quantities are of interest, how-

P = year-endpricedividedby bookvalue,

ever, only because they represent the latest

d = averagedividendfor the priorfive years

available information for the prediction of future dividends. Insofar as these current values departfromaveragesover somepriorperiod for extraordinaryreasons,investmentanalysts maintain that the changes should be discounted to arrive at what might be considered normal values. This suggests that some combinationof

dividedby bookvalue, d = currentyear's dividenddividedby book

value, g = averageretainedearningsforthepriorfive

yearsdividedby bookvalue, g = currentyear'sretainedearningsdividedby

bookvalue.

currentvalues and averages over a prior period The deflationby book value was undertaken

for dividends and retained earningswould pro- to eliminate the scale effect discussed previous-

vide a superior explanation of the variation in ly.'2 The objective was only partially accom-

price among shares.

plished, since correlation exists between the

3. The value the marketplaces on a dividend 'The use of deflatedvariablesin regressionanalysisis a expectation derived from past dividends and debatable question. See David Durand, op. cit., 56; and

DIVIDENDS, EARNINGS, AND STOCK PRICES

I05

TABLE 3.

Sample I95I - Chemicals

Foods Steels Machine tools I954-Chemicals Foods Steels Machine tools

REGRESSION OF PRICE ON DIVIDEND, RETAINED EARNINGS, CHANGE IN DIVIDEND, CHANGE IN RETAINED EARNINGS, ALL DEFLATED BY BOOK VALUE

Constant term -.23 .04 .I5 .I2 .54 -.03 .I8 .05

_ d

I2.42 (2.63) I4.04 (I .04)

9.88 (I.05) I2.62 (I.I7) I7.38 (2.92) I5.5I

(I.04) 9.69 (.99)

II.65 (I.I6)

Coefficientand standard errorof

_

d-d

g

9.79

(5.98) 8.o6

(2.49) 6.38

(I.87) 5.93

(2.75) I2.7I

(8.93) 8.74

(2.82) 3.85

(I.I3) 6.o6

(I.74)

I8.74 (5.96)

3.I6

(I.3 9) I.45

(I.09) .I2

(.99) .I2

(6.39) 5.15

(I.66) 2.02 (.68) 3.70

(I.I2)

g-g

I4.36 (5.6o)

4.57 (I .58)

.4I (i.o6)

I.II (.80) 3.44 (4.78) 5.96

(I.67) 2.85 (.67) I.92

(I.04)

Multiple correlation

.8o .90 .88 .9I .79 .92 *9I .87

deflatedand undeflatedvariables. For instance, the use of both the average value and the de-

correlationbetweenP andp for the eightsamples parture from average appears to have done

ranged from zero to .65 and was more than .4 some good. The range of the dividend coeffi-

for six of the samples.

cient has been reduced by comparison with

The use of d and (d- d) assumes that the Table 2, and the change in dividend coefficient investor values a stock on the basis of the aver- is interesting. All but the chemicals coefficients

age dividend during the prior five years and the are significant at the five per cent level, and

amount by which the currentvalue differsfrom they all are less than the d coefficients. There-

this average. The same reasoning applies to g fore, as expected, a rise in the dividend is dis-

and (g- g), which by the way should be inter- counted until the average has risen to the new

preted as deflated retained earnings and not as level.

growth rates in the context of this model. The The growth coefficients,however, are disap-

coefficients ,8 may be interpreted as follows: pointing. First, the values for g are if anything

1/3= /2 (or /3 = 4) implies that the investors poorer than the values for Y- D in Table 2. ignore the average dividend for the prior five Second, three of the eight coefficients are not

years and consider only the current dividend; statistically significantat the five per cent level.

2 = o implies that the current dividend is ig- Third, for some of the samples 34 3, which nored; ,81> A32impliesthat investorsadjustto means that investors are either indifferent to a change in the dividend with a lag,'3 i.e., the past performance or prefer a share for which

elasticity of expectations is less than one. The retainedearningshas increasedto one for which

opposite is true if 83, ................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download