216 A.. Mauri and M. P. Michaels As can be seen in Table 1, core strategies on DISCUSSION &D and advertising are principally influenced by industry-level factors. In particular for the 5- Our results support the complementarity between year period, 62 percent of the variation in R&d resource-based and industrial organizations per strategy is caused by industry factors, which is spectives. The results from core strategies support about double the size of the firm effect. The the strong influence of industry-level drivers on esults for the longer period are of similar magni- R&D and advertising investments, whereas the de. For advertising strategy, the influence of results for performance confirm the strong effect industry factors is larger for both periods(68- of firm-level drivers 69 percent ), which is also larger in magnitude The findings suggest that firms competing than the firm effect(24-25 percent). In addition, the same industry tend to develop homogeneous for both of the core strategy variables the error competitive strategies for investing in technology percent of the total variance. This shows that consistent with institutional theory. Manage so term is of a small magnitude, between 6 and 15 and marketing resources. The results are both industry and firm factors consistently explain to reduce the strategic gaps with relevant com- the variation of these resource strategies across petitors to gain legitimacy in the eyes of insti- companies, and over time. Companies in the same tutional investors and other important stakeholders industry present a homogeneous pattern in their (Meyer and Rowan, 1977; DiMaggio and Powell, R&D and advertising investments to develop core 1983). Under causal ambiguity, firms choose to resources. The results for ROA are similar to imitate the observable aspects of core strategies those presented by Rumelt(1991), though our For instance, the uncertainty inherent in the ample is very differently chosen, and the periods relation between the investments in specialized are longer. Using a 3-year period, Rumelt found resources and competitive advantage leads other that 46.4 per cent of variation was derived from firms to imitate the behavior of the more success firm factors and 8.3 percent from industry factors. ful ones(Alchain, 1950 The results in Table 1 confirm the predominant These findings also support the resource-based effects of firm-specific factors on performance. It view. Barriers to imitation depend on the degree interesting to note that there is a clear trend of observability of a resource( Godfrey and Hill toward reduction in firm effects as the time period 1995 ). Firms dedicate enormous attention to the of the study is increased. The magnitude of firm study of competitive moves on core resources effects for a 5-year period is 37 percent, whereas because of their potential impact on performance for the 15-year period it is only 25 percent. 2.3 However, easily implementable strategies such as the allocation of funds for R&D and advertising ted by selecting companies competing in more and Jacobson, 1992). The existence of conver precise industry segments, rathet sample was selected b observe variables confirms the competitive value usiness files data base. this data base collects of the difficult-to-observe resources as collis companies identified in the original sample were screened for and montgomery (1995)recently observed: sus the competition topic in financial analysts'reports. Industry tainable competitive advantage can be gained only by leveraging and combining competitively dis- lected in the original sample competing in the same segment. tinctive resources that exist at lower levels of total of 103 single-business companies in 35 segments aggregation ere identified following this procedure. The results using The results are also consistent with the exist- in Table 1. Due to the more hor ults sh nature of the ence of unique resource endowments for firms in le levels of vanance effects increased from almost 62 percent (as reported in R&D and advertising expenditures with rates of 10 and 10/3. modified performance variable. This try effects. This additi by modifyng the asset base of ROa by capitalizing the annual anonymous reviewer o 1998 John wiley Sons, Ltd. Srar.Mgmt.J.vol.19.211-219(1998