Today at work we had a lecture from Joseph Stiglitz, a well known Nobel-prize winning economist. Given his credentials I was hoping for insightful analysis and enlightening comment, but got neither. It was rather disappointing. Lots of time spent talking about stuff we already knew (global warming is bad kids!) and very little time spent on solutions. From this I conclude that the guys at the top are just as unsure of what to do as us guys at the bottom.
a modest proposal
There is probably something deeply wrong with this idea. If so be sure to point it out in the comments (on plan99). I don’t bite!
The current patent system has many problems, and I won’t enumerate them all because that’d be preaching to the choir. If you read this journal you probably work in software and have been burned by it before in some way. This proposal is designed to fix the following problems:
- Patents last for 20 years, a number chosen pretty much arbitrarily. For software this is way too long, for other areas it might actually be too short. When the patent system was created by Henry VIII I’m sure this seemed like a reasonable length of time – after all, the whole point was to stop trade secrets being passed down the through a family. When you are dealing with generations, 20 years probably seems like an instant.
- Patents grant a monopoly to the owner, and monopolies are bad.
- There is no connection between how much it cost to develop an idea and how much you can get for it. So there is an epidemic of patent trolls that simply buy up patents on obvious things then use them to blackmail big companies.
a new system
Let us assume that company A develops a new wonder-drug, and it costs $100 million to do so. This figure is wrong by an order of magnitude of course but it’ll make the maths easier. The value is known because A must provide an audit trail showing in a provable manner what the costs of the project were. Let’s oversimplify for now and assume that every project must pay for itself.
We have a variable II which is a percentage, let’s set it at 20% to start with. II stands for Innovation Incentive and is some arbitrary figure that can be adjusted upwards to increase the incentive to invent things, and downwards to decrease the incentive. I’ve tried to avoid arbitrary figures in this system but we need at least one.
20% of 100 million is 20 million, so we add these two numbers together and put it into a “pool”, which now stands at $120 million.
Anybody who is in the pool has a license to use the technology, anybody out of the pool does not. Unlike the existing patent system though the originators of the patent cannot control who is allowed into the pool – anybody can join as long as they contribute to the value of the pool.
The contributions work like this. When company B wants to join it must pay half of the original sum, or $60 million. For this price B gets the ability to compete with A. Perhaps it cannot make as much money as A can because it lacks first mover advantage, or maybe it can make more due to learning from As mistakes. It doesn’t matter.
The $60 million that B pays goes straight into As bank account. Now both company A and B are down $60 million, but both have equal access to the technology.
Let us assume company C wishes to join the pool. It must pay $40 million ($120m/3). The $40 million is split both ways between A and B, so they both receive $20 million each. Now C is down $40 million, as is B, as is A. They have effectively all paid $40 million towards the cost of developing the technology.
Company D wishes to join, and must pay $30 million, split 3 ways, so now A B and C receive $10 million each. They are now all down $30 million.
And so it continues on and on, with more and more companies joining the fray as the barrier to entry gets lower and lower. The drug that at first may have been affordable only by rich Westerners eventually become available as generics thanks to the African drug companies only having to contribute a small amount to get access to the technology. Effectively the cost of developing the technology has been spread out across many companies.
As this process continues, the amount that each company has “lost” will reduce to zero. Eventually company A will get back $120 million or some figure quite close to it …. it’ll never reach exactly $120 million of course, which is why we need the Innovation Incentive percentage at the start – to ensure A makes a profit on developing the technology. Company B will get close to $60 million etc.
Increasing the II will make the products based on the technology more expensive for the end user. However, it makes the rewards for creating the technology proportionally greater. Decreasing it will decrease costs but also make creating new patents less attractive.
the advantages
This represents an improvement over the existing system in several ways:
- There is no fixed length of time before an idea becomes “free”, because the idea is not really restricted. Anybody who can stump up the entry price can join in, and if the idea takes off they will eventually get their money back.
- It isn’t possible to patent obvious things and then extract billions from your competitors, because you have to prove that developing the idea cost you something. It’s not enough to merely have a “brain wave”, your patent must represent real, hard work.
- You can’t force competitors out of the market a la Eolas because you don’t have a monopoly.
- The cost of licensing the technology will eventually drop so close to zero that the pool can be abandoned entirely, leading the way for commoditisation (eg, generic drugs).
the disadvantages
Firstly, I’m not entirely sure my thinking is straight. I have a vague feeling that this system doesn’t actually work how I think it works, but am not smart enough to see why not.
Apart from that:
- How many inventors are pushed on by the promise of theoretically limitless rewards? Would a fixed II kill innovation?
- The cost of entering the pool early might not be outweighed by first mover advantage
- It’s not clear how you could incrementally evolve the current system towards it. I have a few ideas on this but it’ll have to wait for another post.
- The example figure of 20% I gave for the II might be way too low. Maybe 200 percent is more appropriate.
- Some companies research many ideas that cost money but never pan out, then rely on massive earnings from a golden goose to save them. In the fractional pool model, this becomes a lot harder because you cannot show that the costs incurred in one failed project contributed to the cost of the successful project.
- The need to audit research costs might be unworkable or prove offputting to some companies.
Clearly there are many unanswered questions here and many potential pitfalls, but given that the existing system is so deeply flawed surely anything must be better than it.
If you like this idea, please let me know by writing a short message in the comments. Alternatively digg it, blog it, discuss it, slashdot it etc. Thanks.
October 14, 2006 at 1:58 am |
That sounds like a neat idea for industries like medicine, where the initial investment is high. It completely falls on its face for the original use-case of patents — the invididual innovator.
If I were to develop some amazing invention in my spare time, investing a grand total of $100 in materials, this system would give virtually no chance to build a business around it. Under the current patent system, I might be awarded a patent, start a business and make millions because I’ve got the only game in town. When you open it up like this, I make $120 off people joining the pool, and have absolutely no chance at actually selling my invention because all the companies who have decided they want to participate have vast manufacturing and marketing resources, whereas I’m just someone starting out in a garage.
So, maybe this could be a good system for some industries. Perhaps the real problem is that the patent system is trying to cover so many different use-cases and industries with a single, one-size-fits-all (but not really) system
October 14, 2006 at 2:43 am |
Your biggest problem is “The cost of entering the pool early might not be outweighed by first mover advantage”. There’s a huge, huge free-rider issue; in your example, there’s no difference between dude number 2 to sign up and dude number 3 to sign up, except that dude number 2 is *ten million dollars* worse off. In that sort of situation, you sit tight and wait for the next guy to sign up first and then sign up ten seconds after him.
It also rather assumes that lots of companies *will* sign up to defray the costs, something that’s by no means guaranteed. Take AIDS medication, for example; you’re not going to get 20 Western drug companies all wanting to distribute the same drug, because there’s not enough market for it, and so the pool never gets shared out enough to allow African drug companies to play.
October 14, 2006 at 3:37 pm |
As Stuart Langridge already has said, the most obvious problem is the cost of entering the pool. Chances are that those 100 million bucks has been spent by company A over many years, but company B has to pay a half of this sum upfront.
October 14, 2006 at 7:31 pm |
Hmm. But dude number 2 is only 10 million dollars worse off temporarily – as competitors sign up, that money will eventually be returned (almost). I see the problem with people waiting 10 seconds though and getting off lighter on the initial payments – potentially some damping factor could counteract this tendency. Or maybe it’s a fatal flaw.
Davids point isn’t really valid I think, because the way the patent system works today makes it basically impossible for a small time inventor to enforce any patents they may have thanks to cross licensing etc. Look at Google – the quintessential garage-based inventors, but they didn’t get rich through patents. Virtually all search engines today use link analysis, and Larry & Sergey would have been insane to try and sue Microsoft over such a thing. It just doesn’t work in practice. Instead they got rich by first mover advantage – they had the best search engine at a time when other companies were ignoring the sector and grew quickly, using a few ideas from other companies (Overture) along the way.
Also the original point of the patent system wasn’t to protect the small inventor. It was to break the cycle of trade secrets being passed down through families – checking Wikipedia I see it was Henry VI not Henry VIII I was thinking of, but the patent on stained glass was the one I meant.
If other companies don’t sign up to defray the costs, then you are alone in the market and have effectively paid for the costs of the research yourself – the same as today if you hold a patent nobody wants or can license. So yeah, you aren’t guaranteed to make your money back … that only happens if lots of companies think it’s a good idea and want to enter the market. The point of the system is that you win either way – if nobody else believes in the idea you can go ahead and prove them wrong, but you bear the brunt of the cost yourself. If everybody thinks it’s a good idea, you can’t wield a monopoly on it, but you make your money back.
The problem with nobody wanting to go first I think is a bigger issue. It reminds me of auction theory …. maybe if I think about this some more I can come up with some clever solution like Vickrey did.
October 15, 2006 at 2:09 pm |
In my view, the main flaw is that A would never get paid by its competitors: patent and trade secret are used to avoid competition indeed, if not to kill it.
Moreover, IP licensing constitutes a large part of the profit nowadays, hence, it’s not just a matter to cover the cost, but rather to get a rent.
Fundamentaly there is also a natural hypocrisy : competition is good for the other, not for my buisness; excepted when I need to enter in a market.
October 15, 2006 at 2:13 pm |
Those are flaws from the perspective of the lucky company that owns tons of patents and has the ability to enforce them – ie, not most people. Obviously any other system than the current one will be worse for those people. That doesn’t mean we shouldn’t create one.
October 16, 2006 at 3:37 pm |
Not one, several indeed. As David says the actual system is a one-size-fit-all scheme, but the industry is too organic for that.
Now, I agree that the profit on IP is enjoyed by a minority. Beside that, competitors join together usually on field where they don’t compete, but the present idea leads to the contrary. It’s a good idea for a government which has something to license, frequencies for example, but would want to keep a high-level of competition. (though, a bandwitdh is not sharable indefinitely)
Moreover, the futur runs the buisness more than the present and the past. Of course, recovering his invest import, but the present idea don’t preserve the futur since the competition put it in task.
IMHO, the current system has to be severly reformed by having at least a set of rules depending the industrial field, the inventor (corporation or individu). But, maybe nowadays, the patent office should paid the inventor against a reduced duration of the patent. Afterwhat, the patent don’t fall in the public domain yet, the patent office resell his invest in the way that you expose. Hence the patent office will have a good incentive to make a good job and have a good expertize in prospection since it will be economically liable to have granted a false or dumb innovation. Briefly, the patent office should wet its shirt too.