How Much Does Industry Matter, Really? 23 Table 1. Screened Compustat Business-segment Data (A)By year assets Median assets N (Smil) profit profit 982 5200 l1.310.9 1989 5030948 98 2855 11,4 1990 50291028 598 11.9 11.6 1991 51141040 198 5195660 10.310.3 1992 5232105l 90 10.0 5396 531977 10.0 9.8 1994 57331161 9,1 9.0 5112865 (B)By sector First digit nted represented segment Avg. profit of SIc Brief description of sector (mil) per segmer 0,1,2 Agriculture, Mining 11.2 4068 165 2006 Wholesale Retail Trade Lodging Entertainment Services 279 10.3 "Average ratio in percent of operating income to identifiable assets rather than real economic differences. This prob- lem is related to the general question of appropri ate industry definition Table 2. COv results developed from estimates of Equation 5a EMPIRICAL RESULTS Percent of In this section, we nt our estimates of al variance Equation 5 through uation 4 through anova methods. Year(oy) Industry(o) Table 2 shows the COV estimation of Equation 5. Corporate parent(op 18.68 4.33 Results are expressed as a percent of or, the total Segment specific(ot 31.71 form our Cov decomposition from oR to o?, we Moderr.-industry covariance(2CaB)-551 variance in business-segment profits. To trans- Corp 51.60 use an estimate of p equal to 0.3777. This esti-Error mate is obtained from the nested ANOVA analy- Total(o2) sis and is described in detail below plained by the model. The error, which equals 48.40 percent of the total variance, arises because business-segment profits are sub