Working Paper: NBER ID: w2676
Authors: Nancy L. Rose; Paul L. Joskow
Abstract: This paper investigates the effect of firm size and ownership structure on technology adoption decisions, using data on the electric utility industry. We argue that traditional models of technology diffusion are subject to sample selectivity biases that may overstate the effect of firm size on adoption probabilities. By extending conventional hazard rate models to use information on both adoption and non-adoption decisions, we differentiate between firms' opportunities for adoption and their underlying adoption propensities. The results suggest that large firms and investor-owned electric utilities are likely to adopt new technologies earlier than their smaller and publicly-owned counterparts. Moreover, the selection biases from conventional statistical models can lead one to overstate size effects by a factor of two and to understate ownership structure and factor cost effects by two to four times.
Keywords: technology diffusion; firm size; ownership structure; electric utilities
JEL Codes: O33; L94
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| Larger firms (L25) | Adopt new technologies earlier (O33) |
| Investor-owned utilities (L97) | Adopt new technologies earlier (O33) |
| Failing to account for higher building probabilities among larger firms (D21) | Overstatement of size effects on adoption propensities (C92) |
| Ownership structure (G32) | Adoption probabilities (C25) |