PATENTS, R&D AND LAG EFFECTS: EVIDENCE FROM FLEXIBLE METHODS FOR COUNT PANEL DATA ON MANUFACTURING FIRMS

Hausman, Hall and Griliches (1984) and Hall, Griliches and Hausman (1986) investigated whether there was a lag in the patent-R&D relationship for the U.S. manufacturing sector using 1970¿s data. They found that there was little evidence of anything but contemporaneous movement of patents and R&D. We reexamine this important issue employing new longitudinal patent data at the firm level for the U.S. manufacturing sector from 1982 to 1992. To address unique features of the data, we estimate various distributed lag and dynamic multiplicative panel count data models. The paper also develops a new class of count panel data models based on series expansion of the distribution of individual effects. The empirical analyses show that, although results are somewhat sensitive to different estimation methods, the contemporaneous relationship between patenting and R&D expenditures continues to be rather strong, accounting for over 60% of the total R&D elasticity. Regarding the lag structure of the patents-R&D relationship, we do find a significant lag in all empirical specifications. Moreover, the estimated lag effects are higher than have previously been found, suggesting that the contribution of R&D history to current patenting has increased from the 1970¿s to the 1980¿s.