Recently in our SpicyIP (web) office, Sumathi pointed out an interesting article in the Economist to us. The article was pretty dramatic, to the extent that it essentially called patents nonsense. Indeed, it posted its fair share of examples of frivolous patents in support of its argument. However, in the course of our discussion over the article, Kruttika and I took decidedly different positions on how valid such a view was. We decided to do two contrast posts on the relationship between patents and innovation. Without further ado, here is my contribution.
Nowadays, there seems to be a larger and larger number of people saying that maybe patents and copyrights aren't the best way to go ahead with innovation and creation, the main concerns being that they restrict access to these innovations, as well as that this lack of access, albeit temporary, leads to a restriction on development of further innovations - i.e., slows down innovation overall.
But patents serve as incentives for innovation, don't they? The average person wouldn't work on an innovation unless he had a way of at least making back his investment, as well as probably making a little bit of profit off of it. So why is there this growing sentiment that more patents are a 'bad' thing? Arguably in copyrights, the incentive to create is not as commonly attributed to financial benefits, however it can largely be assumed to have the same rationale. Since this goes against what we've been told consistently, I decided to delve a little more into this topic to help readers understand why more and more people are starting to say this - presenting at least one normative argument for describing why patents and copyrights may not be the most efficient, nor the most equitable method of driving innovation. This is especially important in the current context of the argument over the Bayh Dole Bill which is being fiercely debated over. As we have pointed out in previous posts on the bill, several research institutions and universities are not very comfortable with the idea of the extravagant patenting that this Bill would have them do but haven't been able to voice their views.
Looking into the working of the patent system, the main driving force is clearly economic in nature. Patents require inventors to disclose their inventions, as well as give them a period of exclusivity over this information which they can use to recoup their investment as well as profit off their invention; and this profit motive would serve as an incentive to encourage the making and introduction of more inventions. Thus what this tells us is that patents are one way of inducing innovation/creation and promoting research and development. (For the purposes of this argument, I'm disregarding practical inefficiencies of this system - ie, in reality, the disclosure in a patent is, a large number of times, insufficient for actually developing the product being described; administration and litigation costs, failures in marketing etc.)
In the economics of information, knowledge or information is known as what is called a 'public good'. As noted by Paul Samuelson more than half a century ago, public goods have two main characteristics, and these two characteristics help in understanding the efficiency of patents as incentive for innovation.
Public goods are:
Non-rivalrous in nature - This means that my 'consumption' of it, does not extinguish your benefits from 'consuming' it. Information which I give you, will not take it away from me. This is opposed to a private good, which, if I consume, you cannot. Therefore another person can freely 'consume' or 'enjoy' the same 'product' if it is a public good. This means there is no marginal cost involved for each addition usage.
Non-excludable in nature - This means that it is difficult or impossible to prevent one person from using a good without stopping other users from the same. The typical example for this is a the light from a lighthouse. Either it is in use, or it is not. One cannot prevent only one ship in particular from using its light. This nature of informational goods can be thought to be the basis for IP law. Without external measures, once information is out there, it's out for everyone to use. IP law brings in a restriction on this, since it is essentially a state imposed right to exclude others from using the same information.
As was mentioned, there is no marginal cost of another person enjoying the same public good, therefore it is more efficient to distribute knowledge freely, than to charge for it since it involves factoring in an external cost which wasn't present nor required for the earlier equation. - Because now consumers will under utilize the knowledge, the marginal cost of which is 0, but are charged for a 'marginal cost' anyway, leading to a social loss. This is called Static Inefficiency and leads to what is called a Deadweight loss.
In the patent system, this ‘static inefficiency’ is deemed necessary, because it is viewed as the proxy for necessary incentive required to enhance innovation. Therefore the justification for this kind of static inefficiency, is a system which has dynamic efficiency – one in which the increased rate of innovation makes up enough social value so as to balance out the social loss in allowing for the exclusion rights. This is a long term effect as opposed to the short term effect that static inefficiency has.
[Graph taken from the wiki page on deadweight loss]
However, there is a strong argument that the patent system currently does not provide this either. There is a fundamental problem that the marginal private returns do not correspond to the to the marginal social returns.  This is because of the way that the patent system rewards innovators. When an innovator is granted exclusion rights, it is as a ‘reward’ for having come out with that innovation, since that innovation adds some value to the society at large. However, as is widely accepted, that had that particular innovator not come out with that knowledge at that point in time, another one would have at a later point in time. While, admittedly, it may usually be impossible to know when that later point of time is, the reward that the first innovator is given is therefore supposed to be the reward from bringing it out at that point in time; ‘X’ many years before another innovator would have done the same. From a social point of view, the innovators contribution is making this knowledge available “earlier, and only the extent to which that knowledge occurred earlier than it otherwise would have is what ought to be rewarded; economic efficiency requires that people’s compensation be related to their marginal social returns.”
However, the patent system gives the innovator the entire value of the contribution, and not just the value for bringing it out that much earlier. This worked for a long time, because the patent protection period of 20 years was less than or roughly equal to the marginal social returns. It took a longer time to develop and innovate. (This is a broad generalization of course) However, now, especially in certain fields such as software, with the rate of innovation so quick, the marginal private returns, proxied by the reward of 20 years for an innovation which very likely would've been innovated by someone else much before that 20 year period, well exceeds the marginal social contribution. With faster rates of innovation, the likelihood of the knowledge being discovered or invented by another innovator increases exponentially – therefore quite often it ends up exceeding the marginal social contribution. Therefore, especially with sequential innovation, there are growing instances of imbalance between the marginal private returns and the marginal social gains. Is the answer for this - shortening patent periods? I would think that's too simplistic an answer which doesn't take into account the intricate details of the trade-off between static inefficiencies and dynamic efficiencies.
In a sense, this is understood within the system as well. In the case of Scientific Principles and Theorems, (which are not allowed to be patented) the ratio between potential private benefits and the social value that the scientific principle has is so obviously skewed, that it would make no sense to reward the private individual for this contribution. Amongst other things, it could block/slow down whole streams of research. This would also explain why all patent systems recognize that scientific principles cannot be patented, though the reasoning may usually be more intuitive than corroborative.
As the article points out, as does much evidence, the pharmaceutical industry is one wherein patents have played an important role as being incentives for innovation. This is mainly due to the high investment costs in a product, as well as the relatively low reverse engineering costs for the same. However, all that this shows is that pharmaceuticals require a mechanism to promote innovation and that the firms that dominate under the current system are dependent on the tools which brought them to dominance. 
Patents may not only be an inefficient method of innovation, but may also actively impede the rate of innovation.
Sticking with examples from the poster child of strong IPRs, the pharma sector: it is often seen that it may be more beneficial for the innovator to invest less in new research, but rather invest in drugs, which are therapeutically similar to existing drugs (already proven to have a big market) with a different active ingredient (commonly known as a ‘me-too’ drugs). Thus, the social value addition that the drug makes is very small, since it is for the therapeutically same reason. At the same time the private benefit for the innovator is very large, since there is a trend towards oligopoly within therapeutic class levels, ensuring high rates of profits for the innovator.
The patent system also is capable of distorting research directions. This is because innovators will naturally tend their research towards patents which grant them more profits. This is especially significant in pharmaceuticals, because it leads to a huge social loss, in the form of lack of research towards diseases prevalent in developing countries (with ‘weaker’ markets), rare diseases and a bias towards maintenance drugs.
While I am not suggesting that the patent system be removed completely, for I don't know what effects that will have, nor am I currently suggesting an alternate, I do believe that in order to proceed with policy making in the system of innovation, it makes sense to critique and analyze the existing system, with all its problems, so as to reformulate and refine as necessary.
For those interested in more of the various critiques of the patent system with respect to innovation, Bessen and Meurer have some very insightful articles over at www.ResearchOnInnovation.org
To the readers who have made it till here, my apologies for the length of the post! And now over to Kruttika!
 Joseph Stiglitz, "Economic Foundations of Intellectual Property Rights", 57 Duke L.J. 1693 (2008); available here.
 Kapczynski, Chaifetz, Katz & Benkler, "Addressing Global Health Inequities: An Open Licensing Approach for University Innovations" 20 Berkeley Tech. L.J. 1031 (2005); available here.