Capital market efficiency of firms financing research and development

Private industrial research and development (R&D) as percentage of sales is significantly reflected in the beta coefficients of Sharpe's "diagonal" capital asset pricing model (CAPM) for 343 firms with an issue of common stock traded on either the New York or American Stock Exchange from the first of 1979 through 1988 and that financed R&D in each year from 1978 through 1988. Yet the correlation is less than {dollar}mid{dollar}0.20{dollar}mid{dollar} between nine measures of R&D and four measures of market-beta. The measures of firm-level and industry-level R&C partly explain the cross-section of expected excess real total gross returns for one to three out of nine decile portfolios formed monthly on the basis of the respective R&D measure. The lowest decile portfolio consisting of non-R&D firms is not priced for any of the R&D measures or for industry research group code ("group"). Decile portfolios formed on R&D expense are priced more frequently for firms reporting proprietary R&D expense only than for firms reporting R&D expense with engineering expense and contract R&D.; Market value of equity ("size") is one of a set of circular variables that is shown to provide a spurious explanation of return due to implicit autoregression or embedded identity. The circular set includes share price, dividends per share, number of shares outstanding, and any other variable that entails one or more of these variables. Size and lexical order of firm name ("name"), respectively, serve as upper and lower benchmark variables.; The multivariate CAPM consisting of the market-beta for a proxy of the stock market in combination with group and separately with size measured by the natural logarithm of market value of equity (LnMKEQ) as risk factors significantly prices group and size when decile portfolios are formed monthly on firm R&D expense. Market-beta is not priced jointly with either group or size in these tests.; The univariate Sharpe's CAPM with seemingly unrelated regression (SUR) prices three out of 25 portfolios formed monthly on R&D expense and then group and five out of 25 portfolios formed monthly on group and then R&D expense. With CAPM and SUR, five out of 25 portfolios formed monthly on R&D expense and then size are priced, and one out of 25 portfolios formed monthly on size and then R&D expense is priced. All risk premiums or prices are negative.; The data generally show that a rational market in equilibrium efficiently prices firm differences based on nine measures of R&D for the large majority of stocks. Nevertheless, there is evidence of persistent anomalous pricing due to R&D activity for our sample assuming a pure first-order autoregressive stochastic return-generative process with evidence of no martingales with zero drift. This finding is robust across group, sales, market data lead times, market proxies, disaggregation levels of firm R&D data, and method of estimation. Portfolio pricing excluding the January seasonal is either augmented or unaffected rather than attenuated. Group could be considered as a proxy for an industry effect.