View source for Hall Ziedonis (2001) - The Patent Paradox Revisited

{{Article
|Has page=Hall Ziedonis (2001) - The Patent Paradox Revisited
|Has bibtex key=
|Has article title=The Patent Paradox Revisited
|Has author=Hall Ziedonis
|Has year=2001
|In journal=
|In volume=
|In number=
|Has pages=
|Has publisher=
}}
*This page is referenced in the [[Patent Thicket Literature Review]]
*This page is listed on the [[PTLR Core Papers]] page

==Reference==

*Hall, B.H. and Ziedonis, R.H. (2001), "The patent paradox revisited: an empirical study of patenting in the US semiconductor industry, 1979-1995", RAND Journal of Economics, pp.101--128

 @article{hall2001patent,
   title={The patent paradox revisited: an empirical study of patenting in the US semiconductor industry, 1979-1995},
   author={Hall, B.H. and Ziedonis, R.H.},
   journal={RAND Journal of Economics},
   pages={101--128},
   year={2001},
   abstract={We examine the patenting behavior of firms in an industry characterized by rapid technological change and cumulative innovation. Recent survey evidence suggests that semiconductor firms do not rely heavily on patents to appropriate returns to R&D. Yet the propensity of semiconductor firms to patent has risen dramatically since the mid- 1980s. We explore this apparent paradox by conducting interviews with industry representatives and analyzing the patenting behavior of 95 U.S. semiconductor firms during 1979-1995. The results suggest that the 1980s strengthening of U.S. patent rights spawned "patent portfolio races" among capital-intensive firms, but it also facilitated entry by specialized design firms.},
   discipline={Econ},
   research_type={Theory, Empirical},
   industry={Semiconductor},
   thicket_stance={},
   thicket_stance_extract={},
   thicket_def={},
   thicket_def_extract={},  
   tags={},
   filename={Hall Ziedonis (2001) - The Patent Paradox Revisited.pdf}
 }

==File(s)==

*[[Media:Hall Ziedonis (2001) - The Patent Paradox Revisited.pdf|Download the PDF]]
*[[:Image:Hall Ziedonis (2001) - The Patent Paradox Revisited.pdf|Repository record]]

==Abstract==

We examine the patenting behavior of firms in an industry characterized by rapid technological change and cumulative innovation. Recent survey evidence suggests that semiconductor firms do not rely heavily on patents to appropriate returns to R&D. Yet the propensity of semiconductor firms to patent has risen dramatically since the mid- 1980s. We explore this apparent paradox by conducting interviews with industry representatives and analyzing the patenting behavior of 95 U.S. semiconductor firms during 1979-1995. The results suggest that the 1980s strengthening of U.S. patent rights spawned "patent portfolio races" among capital-intensive firms, but it also facilitated entry by specialized design firms.

==Review==

===Measures of thicket===

Capital intensity is used as a measure of whether firms engage in defensive patenting as would be expected in a patent thicket.

===Sample===

Quantitative data is analyzed for:
*95 U.S. publicly-owned semiconductor firms (SIC=3674) from 1979-1995:
**Financial data on each firm is drawn from Compustat, and includes the universe of wholly-owned public U.S. semiconductor firms between 1975 and 1996;
**The narrow SIC definition ensures that the firm's R&D expenditures relate to the technology area.
**Business directories and industry sources provided firm founding dates and identification of firm as a manufacturer or a design firm (20% of observations).
*17,228 patents in the patent portfolios of the 95 firms analyzed:
**Drawn from the NBER-Case Western patent database of 2,000 manufacturing firms (updated to reflect recent data and firm name changes);
**All patents applied for between 1979-1995.
The paper also presents results of field interviews for 7 semiconductor firms (4 manufacturers and 3 design firms) between April 1998 and November 1998.

===Results===

Field interviews with semiconductor firms indicated that:
*stronger patent rights ushered in by the CAFC reshaped incentives to obtain patents.
*Large manufacturers had ramped up their patenting activity and viewed patents as "bargaining chips".
**May be a response to 1986 Kodak v Polaroid case that created chances of injunctions and thereby risks of patent infringement.
**Texas Instruments asserted rights over a number of patents in 1985-1986, which increased licensing revenues
**Exacerbating these effects, costs of production were higher and product life lower in this period.
*Smaller design firms viewed patents as a means to secure property rights over their innovations, and exclude competitors.

Empirical analysis:

*In base-line model, larger, capital-intensive firms appear to patent defensively.
:''"Capital intensity seems to have an important effect on the propensity to patent, one that is slightly larger and more significant than the influence of R&D spending. This is consistent with our first (strategic response) hypothesis."''

*Comparing the later pro-patent regime with earlier periods:
**The elasticity of patenting to capital intensity becomes larger and statistically significant going from an insignificant -0.030 and 0.128 before and during the regime change to 0.574 after, and "catches up" with the high patenting propensity of Texas Instruments.
*The elasticity of patenting to R&D weakens going from 0.457 and 0.530 before and during the regime change to an insignificant 0.41 after.
:''"During the era of strong patent rights, these firms' patenting behavior became less, not more, responsive to their investments in R&D. Consistent with the view that capital-intensive firms altered their behavior under the new pro-patent regime by patenting more aggressively, the coefficient of capital intensity changed from small and insignificant in periods 1 and 2 to positive and highly significant in period 3."''

*Chi-square tests for whether larger capital-intensive firms vs. smaller design firms are the same is rejected.
*The elasticity of patenting to capital intensity of larger capital-intensive firms is 0.64 compared to 0.30 for smaller design firms.
*The elasticity of patenting to capital intensity for smaller design firms entering after 1982 is a large, significant 1.60.
:''"We view this as strong confirmation of our interviewees' statements that patent rights are required to secure venture capital and other financing for entry as a specialized semiconductor design firm."''

*Citations to semiconductor patents fell as a ratio of manufacturing patents.
*Citations to other patents by Semiconductor patents rose.
:''"[T]here is weak evidence at best for the hypothesis suggested by our interviewees, that the quality of the average patent in this industry has fallen."''

*In aggregate, patenting propensity after 1982 increased, especially for new entrants.
:''"The patent propensity of the firms in our sample increased at a steady rate of 10% per year after about 1986, somewhat later than the 1982 introduction of the Court of Appeals for the Federal Circuit, for a total increase of approximately 100% by 1995."''

===Social Welfare Consequences===

:''"[S]tronger patent rights may have facilitated entry by specialized firms and contributed to vertical disintegration in this industry (Merges, 1997; Arora and Fosfuri, 1998). But these positive effects coincide with a process whereby firms amass vast patent portfolios simply as "bargaining chips," leading to "patent portfolio races."''

:''"It also leaves a number of important questions unresolved. What are the overall social welfare consequences associated with the type of "patent portfolio races" that we observe in this industry?"''

===Dependent Variable and Model===

*The dependent variables is the number of patents granted to a firm in a given year.
**Models are separately estimated for periods "before" 1979-1983, "during" 1984-1988 and "after" 1989-1993 establishment of more pro-patent CAFC regime, and is restricted to 34 firms publicly traded before 1984.
**Although Lagrange Multiplier tests favored negative binomial, a Poisson-based model with robust standard errors is used in the estimation, because robust standard errors should remain consistent and true distribution may still be Poisson.

*Independent control variables are:
**log of R&D spending (per employee when a control for employees is included in the specification);
**firm size as measured by the log of the number of employees;
**time dummies for each observation year;
**capital expenditures (log of property, plant and equipment per employee in 1992 $), as a measure of definsive patenting by larger firms;
**An indicator of market entrants (founded after 1982 the year of CAFC);
**Firm type (manufacturing or design firm);
**A Texas Instruments indicator in one specification;
**Firm age.