TIIP Newsletter

Introduction
Patent law requires that new inventions be non-obvious in order to receive patent protection. Because patents confer an exclusive property right, it makes sense to grant this right only to inventions that are sufficiently different from ones already patented and that represent a significant improvement over existing knowledge. Virtually all patent systems include a non-obviousness standard, also known as the “inventive step” requirement. This rule attempts to apply a uniform test for novelty to all patent applications.

But should the same standard apply to all industries? And should international patent harmonization treaties require uniform standards worldwide? The need for uniform policy is one of the central arguments for extending patent protection to software and business processes, but should patent policy really be “one-size-fits-all”?

This paper makes a strong case that the answer to this question is “no.” Hunt develops a simple model that relates patentability standards to industry characteristics.

Model
Hunt’s model considers sequential innovation, where each innovation builds on previous ones. This differs from other models of sequential innovation because it also considers the effect of patent policy on industry concentration and the ability of innovative firms to enter an industry.
The non-obviousness standard affects firms in two conflicting ways. A higher standard makes it more difficult for a new firm to obtain patent protection for an innovation—and more difficult to enter an industry where competitors hold existing patents. On the other hand, an invention that does meet the higher standard will be more profitable: it will be more difficult for other firms to develop new inventions that may challenge the patent-holder’s position in the market, and the patent-holder can expect to receive profits for a longer period before its invention becomes obsolete.

Given such conflicting incentives for innovators, policymakers face a tradeoff. If the non-obviousness standard is too weak, firms may have insufficient incentive to develop initial inventions because they fear rapid obsolescence. If the non-obviousness standard is too strong, firms will not develop sequential improvements because these may prove too difficult to patent.

Results
Hunt shows that socially beneficial innovation is greatest with an intermediate patentability standard that provides reasonable incentives for both initial inventors and sequential innovators. This result is similar to results of some economic growth models (see Aghion et al, 1999) that find innovation to be greatest when the market is competitive, yet not so competitive to eliminate all expectations of profit. But here this pattern is related to intellectual property policy.

The paper also finds that this optimum point changes across industries. Specifically, in industries that are inherently more innovative, the optimal inventive step is larger.. Innovative industries tend to have faster follow-on improvements, and their innovations tend to be more sequential, with many generations of improvements. Hunt’s result suggests that patent policy should be tailored to encourage dynamic innovation in these industries by allowing patents only on the most significant innovations. This policy encourages firm entry, but it leaves a larger share of incremental innovations unprotected.

This finding captures what seems to be a basic feature of many innovative industries. Industries such as software and semiconductors do indeed exhibit a rapid rate of sequential innovation and these industries have had low barriers to firm entry in the past. The model implies that these industries deserve a different standard of intellectual property protection, one that permits innovators to enter the market easily in the future as well.

Hunt’s model is, of course, a highly stylized one based on several strong assumptions. For instance, he assumes that sequential innovations completely replace previous inventions. Nevertheless, the simplicity of his model does not invalidate his main conclusions. If a “one-size-fits-all” policy is not optimal even in this simple model, then it is hardly likely to be optimal in a more complicated setting, where industry differences may be even greater.

Reference
Aghion, Philippe, Nicholas Bloom, Richard Blundell, Rachel Griffith, and Peter Howitt. (2002). “Competition and Innovation: An Inverted U Relationship,” Working Paper.



	Patentability, Innovation and Industry Structure Federal Reserve Bank of Philadelphia Working Paper 01-13 by Robert M. Hunt, Federal Reserve Bank of Philadelphia FULL TEXT
			-- Summary by James Bessen
			Introduction Patent law requires that new inventions be non-obvious in order to receive patent protection. Because patents confer an exclusive property right, it makes sense to grant this right only to inventions that are sufficiently different from ones already patented and that represent a significant improvement over existing knowledge. Virtually all patent systems include a non-obviousness standard, also known as the “inventive step” requirement. This rule attempts to apply a uniform test for novelty to all patent applications. But should the same standard apply to all industries? And should international patent harmonization treaties require uniform standards worldwide? The need for uniform policy is one of the central arguments for extending patent protection to software and business processes, but should patent policy really be “one-size-fits-all”? This paper makes a strong case that the answer to this question is “no.” Hunt develops a simple model that relates patentability standards to industry characteristics. Model Hunt’s model considers sequential innovation, where each innovation builds on previous ones. This differs from other models of sequential innovation because it also considers the effect of patent policy on industry concentration and the ability of innovative firms to enter an industry. The non-obviousness standard affects firms in two conflicting ways. A higher standard makes it more difficult for a new firm to obtain patent protection for an innovation—and more difficult to enter an industry where competitors hold existing patents. On the other hand, an invention that does meet the higher standard will be more profitable: it will be more difficult for other firms to develop new inventions that may challenge the patent-holder’s position in the market, and the patent-holder can expect to receive profits for a longer period before its invention becomes obsolete. Given such conflicting incentives for innovators, policymakers face a tradeoff. If the non-obviousness standard is too weak, firms may have insufficient incentive to develop initial inventions because they fear rapid obsolescence. If the non-obviousness standard is too strong, firms will not develop sequential improvements because these may prove too difficult to patent. Results Hunt shows that socially beneficial innovation is greatest with an intermediate patentability standard that provides reasonable incentives for both initial inventors and sequential innovators. This result is similar to results of some economic growth models (see Aghion et al, 1999) that find innovation to be greatest when the market is competitive, yet not so competitive to eliminate all expectations of profit. But here this pattern is related to intellectual property policy. The paper also finds that this optimum point changes across industries. Specifically, in industries that are inherently more innovative, the optimal inventive step is larger.. Innovative industries tend to have faster follow-on improvements, and their innovations tend to be more sequential, with many generations of improvements. Hunt’s result suggests that patent policy should be tailored to encourage dynamic innovation in these industries by allowing patents only on the most significant innovations. This policy encourages firm entry, but it leaves a larger share of incremental innovations unprotected. This finding captures what seems to be a basic feature of many innovative industries. Industries such as software and semiconductors do indeed exhibit a rapid rate of sequential innovation and these industries have had low barriers to firm entry in the past. The model implies that these industries deserve a different standard of intellectual property protection, one that permits innovators to enter the market easily in the future as well. Hunt’s model is, of course, a highly stylized one based on several strong assumptions. For instance, he assumes that sequential innovations completely replace previous inventions. Nevertheless, the simplicity of his model does not invalidate his main conclusions. If a “one-size-fits-all” policy is not optimal even in this simple model, then it is hardly likely to be optimal in a more complicated setting, where industry differences may be even greater. Reference Aghion, Philippe, Nicholas Bloom, Richard Blundell, Rachel Griffith, and Peter Howitt. (2002). “Competition and Innovation: An Inverted U Relationship,” Working Paper. © 2002. Verbatim copying and distribution of this entire article for noncommerical use are permitted provided this notice is preserved.