value (price stability. is incorporated into the study via the - PDF document

Attitudes Toward Characteristics of Common Stocks A Summary Presentation Haskel Benishay, Northwestern University* Introduction and Summary I. This study examines empirically the -The relation between determinants of the rate of return on the market value of common stocks for seven cross sections equity and a market of 111 common stocks for the years 1958- stock index which may 1964 in a multiple regression analysis. be thought of as the Theoretically, the common stock rate of "conformity" of the mar- return under investigation is the ratio ket value of the equity of the expected income associated with of a firm to the index a particular equity to the market value of the values of the of that equity. This rate is empiri- equities of the market cally represented by a ratio of a weight- as a whole. ed average of past annual earnings to current market value of the equity. All but the last two variables were employed in an earlier empirical The empirical rate of return is study of cross section data for the bypothesized to be a function of two The earlier study years 1954 -1957 [1]. groups of variables: a corrective group and a later discussion associated with and an explanatory one. (1) The correc- some of its controversial aspects [2,8] tive variables are employed to attenuate which this provided a benchmark from errors involved in the measurement of study was launched. A controversial re- the empirical representation of expected sult of the earlier study was a regres- income in the numerator of the rate of sion finding of a market preference for (2) The explanatory variables the variability of common stock equi- return. are presumed to represent factors which It was a result fortified by con- ties. exert a real influence on the relative sistency in the various regressions They are em- desirability of stocks. rather than a high level of significance ployed to provide an explanation of the in each one regression separately. In rate of return based on preferences and this study one of the major interests aversions of people in the market to was whether a preference for variability various attributes of common stocks. will persist in new data and with the The variables employed within the two inclusion in the regressions of addi- groups are: tional relevant variables. Another ma- jor interest, whetted in recent years by Corrective: -Trend(growth) in the mar- work on portfolio selection [4,10,11,15] ket value of equity. was the response of the market to the conformity) of equity conformity (non- -The pay -out ratio or the market value to the market index. ratio of dividends to earnings. To satisfy the first interest a measure of the skewness (third moment) Explanatory: -The stability of the in- of the equity distribution is added on- come (earnings) stream. to the explanatory variables. To sat- -The stability of equity isfy the second an additional dimension value (price stability. is incorporated into the study via the inclusion of a conformity measure rep- -The size of the firm. resented by the coefficient of the -The debt -equity ratio. linear relation between the firm's equity market value and the value of -The skewness of the dis- Standard and Poor's 425 stock index. tribution of equity values. * The author is a professor of managerial economics at Northwestern University. Len Wiltberger helped in the collection of data and William Melberg assisted in data Responsibility for errors is, of course, mine. processing. 318

H. Variables Employed Why should the independent vari- ables be expected to account for dif- ferences in rates of return on equity? Brief answers are provided below for each variable. Some additional discus- sion follows for equity variability, equity skewness and conformity to the movement of the market, taken as a group. The new and most noteworthy results A. The "Correctors." in this study, as told by multiple re- gressions, are: The Rate of Return: is empirically measured by a ratio of a weighted mean -Equities of non- conform- of the earnings of the company in the ing stocks sell at a nine years preceding and including the premium. Or, the market cross -section year to the arithmetic prefers stocks which do mean of the high and low of the market not conform to the move- values of company equity during the ment of the index. cross -section year. The weights em- ployed for the nine observations of the -Equities whose market weighted mean of earnings decline expo- values are variable sell nentially as the observations recede in- at a premium. Or, the to the past away from the cross -section market prefers variabili- year. The weights for the i`th year ty to stability of the back, where i =1 refers to the cross - market values of common section year, are: stock equities. -Equities whose market values are positively skewed sell at a premium. Or, the market prefers stocks whose market values are positively skewed. Thus for the cross -section year, the weight is The other results as told by our 9 regressions are: Z (.8)1 (.8) i=1 -The equities of large firms sell at a premium for the farthest year away, in the market. 9 -The equities of firms (.8)1 (.8) with larger debt -equity i=1 ratios sell at a dis- count. The weighted average of company -The equities of firms earnings represents, in theory, expect- whose earnings are more ed earnings (income) of the company. stable tend to sell at a Needless to say, the expected earnings premium (a mixed result). sought and the weighted mean of earnings employed are not the same. It is manda- The two corrective variables, aimed tory to emphasize that this creates a at attenuating the errors of measurement large sized and fundamental empirical of earnings, perform as expected: problem. If rates of return were meas- ured without error then the differences -The past growth in equity in the rates would be attributable is negatively associated wholly to characteristics of the company with the rate of return. of the type represented by the explana- tory variables below. But they are not. -The pay -out ratio is And hence the need for variables whose negatively associated function is to attenuate errors. with the rate of return. 319