Fuller and Kerr (1981) proposed a technique to determine the CAPM beta of a firm that requires neither market data nor relies upon a specific model of accounting data. In their method, the analyst selects a proxy for the firm based on his or her analysis of accounting data, industry factors, and other information the analyst deems relevant. They report that this procedure produces ‘‘pure-play proxies’’ that are very close to the target firms in CAPM beta. However, this technique has not been utilized extensively in cost of capital estimation. This is probably due to two significant shortcomings. The first is the lack of pure-play firms to serve as proxies. Again referring to Jack Gunn in Financial Management (see Ang, 1989): ‘‘The toughest part was identifying comparable firms and single product firms who competed with the products we offered’’ (p. 19). Another problem is the inherently subjective nature of this approach; a different analyst may select a different proxy for the same division.