176 – Business Analytics and Financial Econometrics
Identification Strategies for Models of Innovation, R&D, and Productivity
Juana Sanchez
UCLA
Increasing productivity is a major priority for the US and other advanced nations. Because of this, the link between innovation input (R&D), innovation output (process, product and logistic innovation) and productivity is a relevant policy issue. This presentation is about the capability of currently available survey data such as the Business Research and Development and Innovation Survey (BRDIS) to establish a causal relationship that is useful for making predictions about the consequences of changing innovation inputs. The observational nature of BRDIS makes it necessary to recognize selection, simultaneity and, more generally, endogeneity before any links between R&D, innovation and productivity can be established. Lacking random assignment policy evaluations and a causal interpretation of the findings, a new identification strategy based on structural equation models is presented in this paper that tries to address all those problems.