Patents, Intellectual Property and the Rise of the Rent-Seeking Society with Alex Tabarrok | a16z

Transcript

uh very pleased to welcome Alex tabarak who's a professor at George Mason University um normally when I introduce economists like for a talk I often feel obliged to kind of mention this specialization like I'll say oh it's a they're a market designer or maybe if they're really General I'll say they're a micro Economist but Alex I have to say just he's an economist he's so broad he's worked on sort of Economics of crime Health economics including a lot of work on vaccine distribution more recently he's done work in crypto as well and then today he's going to be telling us about intellectual property it should be really interesting so Alex all yours great thanks very much Tim yeah no no crypto no blockchains today that would violate comparative advantage so I'm

going to talk about patents copyrights uh intellectual property more generally and what I call the rise of the ren seeking Society so if you take a look at the number of patents uh issued uh in uh in recent times you can see it was actually pretty flat through most of the 20th century and then you see this big increase beginning in the 1980s with this creation of the federal circuit court special patent Court in 1982 we'll come back to that and when you first look at this and you think oh well yeah sure I mean listen uh patents are going up and all of this Innovation and computers and you know it makes sense we live in a high-tech Society to see this big increase in patents and yet however

when you take a look at productivity a total factor of productivity what you see is that there was a massive decline in total Factor productivity this is what my colleague Tyler Cowan calls the great stagnation so you have this big increase until 1973 972 somewhere around there and then you have this decline in total Factor productivity briefly catching up a little bit in the 1990s we'll talk about that and then it you know it's just flat and in the 2000s 2020s uh it's just flat okay total Factor productivity is like what we care about the most it is basically how much output you're getting from the same inputs of Labor and

capital okay so uh basically you're looking at output minus you sort of take away uh what you can explain of output from your inputs of capital and labor and then as that changes over time okay you're thinking well that's technological progress right so uh you know the computer today has about the same amount of silicon uh as you know 30 years ago but it's much faster right it's much more capable that's total Factor productivity right so that's the idea and this is really what we care about because this determines our uh state of living standard of living okay so it's the

ability to get more from less which is really uh a determinant of our our standard of living there's an interesting way of putting this by the way and that is this is kind of an aside but if total Factor productivity had continued to increase on its post-world War II to you know 73 Trend okay if it continued to continue to increase on that uh same Trend we would be living in the world of 2092 rather than in the world of 2023 and I would much rather live in the world of 2092 with better Medical Care and their technology oh this can maybe we would

have flying cars okay so that's just kind of a way of thinking about this okay now uh so what I'm going to talk about what I want to begin talking about with patents is I first want to convince you I that patents are not necessary okay for at least some types of innovation and then that patents can actually reduce uh Innovation that's a possibility as well and I want to be clear I'm not I don't want to blame I'm not blaming patents for the decline in total Factor productivity that would be a much bigger much stronger claim but I do want to make these two claims okay pans are not necessary for at least some Innovations and in some cases they can

reduce innovation so beginning with the first claim uh the patents are not necessary it's actually the case that most Innovations okay are not patented so we think about you know economic growth every year maybe we're up two two and a half percent you know in the United States if we're doing well okay and that's like just millions and millions of improvements and Innovations and things getting better and so forth most of these Innovations are not patented at all uh and in fact there are many fields which are very Innovative which patents are not even allowed okay or not even used like fashion so every year we're in

Fashion District okay you walk down the streets and you see new designs and new cuts and new types of clothing new materials and so forth none of this is uh patented uh cooking if you watch Top Chef a guy you see new recipes you see Molecular Gastronomy right new ingredients uh sous-vide okay right new techniques this is not patented we didn't used to allow patents for business methods nobody really complained oh we don't have enough business methods okay uh we need to to get more business methods no one was really saying that but the court said okay now business methods can be patented even semiconductors

uh there was actually no IP for semiconductors until quite quite recently like 1970s okay and we still had plenty of semiconductors there's also an interesting historical a set of examples uh in the 19th century there was a sort of anti-patent movement a little bit similar to what we're seeing uh today and but it was quite strong and that a number of countries actually got rid got rid of their patent systems okay including Denmark the Netherlands Switzerland pretty major countries and so the question then becomes became uh did these countries become less innovative

and in a very nice paper by uh Petra Moser the economic historian uh she looks at these countries and finds in fact note they were quite Innovative these countries continue to be among the most Innovative countries uh in the world the direction of innovation changed a little bit the sort of types of Innovations which happened uh changed a little bit but overall these countries even without patent systems continue to be very innovative and I want to spend some time to talk about the most Innovative U.S company a company you're all familiar with but you might want to think different because the company which I'm actually thinking about is Walmart okay Walmart

now why is Walmart the most Innovative U.S company well first of all let's just point out how incredibly important Walmart is it is the world's largest private employer okay it's over 11 000 stores 27 countries uh the world's largest firm by revenues okay almost uh 600 611 billion in revenues last time I checked uh it saves consumers by one estimate from some years ago a quarter of a trillion dollars uh a year okay it does actually have low prices right and uh if you remember that productivity

boom of the 1990s which I pointed out in the first slide turns out that a huge fraction of that came from retailing and a huge fraction of that came from Walmart okay Walmart was simply much more productive than its rivals by some estimates 40 percent more productive in the retail sector than its Rivals and why why how do they do this how do they pull off this miracle how did they become the world's largest private employer the company with the biggest revenues of any other firm how did that happen was it some like Secret Sauce some technological Miracle some you know

something they had like no it's none of these things it was Sam Walton relentlessly improving productivity on multiple margins very simple stuff like cross-training of the staff in different departments use of barcodes so they use like some technology they used RFID chips okay but this is public technology this was not something which they had any special access to in fact uh to illustrate that you know IP and patents was not was not what made them super productive uh you can go and look and ask how many

patents does Walmart have the stock how many patents does Walmart have and when I checked this out it was like 60. okay now including one for a shoebox okay now to put this in context IBM uh uh files for like 30 patents a day okay so 60 the stock of moment it's nothing it's nothing so Walmart incredibly important incredibly productive not based upon IP okay so let's now turn to the costs some of the costs of intellectual property and we'll begin kind of with a static

example and probably for even the non-economists here probably seeing this diagram uh before and uh what the diagram says right is that when you give a patent that creates a monopoly okay which allows the firm to get some profit and the idea is that profit then allows firms to recoup their sunk costs okay they have some research and development costs this profit allows the firm to charge higher than market prices for some time period and earn these profits and recoup those sunk costs of research and development and the idea is that the cost of this is that since the price is above the

marginal cost we have some deadweight loss we have some transactions which fail to occur even though they would be socially beneficial because the price is higher than the marginal marginal cost it's a very standard uh story in fact I have told this story before it's actually not correct it's actually under underplace things uh and you can begin to see that first of all this story you might say is a little bit uncharitable because well you say oh well there's this deadweight loss but the counter argument to that would be well this product wouldn't exist if it weren't for the patent so how can you

say that there's this deadweight loss when the whole thing wouldn't even be there if it weren't for the patent okay and that actually tells you something it tells you that the cost of the patent system is not actually this deadweight loss it's the deadway loss on all the other products which are patented but where the patent was not required in order to produce the product in order to produce the innovation so it's all of these other deadweight losses which is actually part of the cost of the patent system that's the sort of static story which I think most of you maybe with the

exception of this Edition uh how probably heard before sorry Alex so the point there is you're saying the old story is roughly correct if indeed the product would not exist but for the protection of the patent yeah so it it's fine if you're looking at my point here is that when you're when you're trying to evaluate the patent system you have to look at not just the product which is created because of the patent but the monopolies which are created in

other products which would not necessarily be monopolies if we didn't have the patent system so patents were only issued to things that would not otherwise be produced sure yeah exactly if we could uh just pick and choose right that's right yeah so the the counter story which may or may not be correct is that well we provide temporary Monopoly but this Innovation might generate some technological spillovers that may undo this kind of deadway laws at some point sure and we can add I mean that that says yeah I mean that

uh there could be sure there could also be technological spillovers to the unpatented products as well right so you know yeah but then the question is whether these Innovations may take place or not and yeah so that's kind of like an accounting exercise yeah absolutely Dennis says okay all there all these deadweight losses from you know sort of giving Monopoly rights to all these different side products but if we're accounting for all the side products we should also be accounting for all the spillovers from the patented products that would not be produced otherwise sure but that I would but that I would say we should count for the spillovers

uh from the the monopolized product right I mean you get less you get fewer spillovers absolutely so that doesn't affect the uh that doesn't affect the okay the question is how does the monopolization of those you know sort of side products compare to the spillovers from the pad the products that are only produced if there are patents sure sure fair enough fair enough okay so that's kind of the static uh story um yeah you also get Spence destroys you also get expense distortions in this case right they're like because you have a a patent not something there's no need to Fair the innovated so like people's quality doesn't improve as fast as you

might even they're like in the same products yeah yeah so uh so well now we're I want to sort of turn to uh the sort of a more Dynamic uh issues which arise uh with uh patents and here I want to contrast sort of a linear Innovation model which is the one which sort of first comes to mind with a more cumulative uh Innovation model okay so if you have a cumulative Innovation model where you're thinking about a product which you're building on top of other products okay or ideas which are building on top of other ideas well uh then it can be quickly become sort of obvious that if you have a patent here

which is raising the cost then you could raise the cost of building on top of uh other ideas and that can reduce this uh virtuous circle okay can reduce the Verge can cut the virtuous circle all right so let me give you a example uh I'm going to call this the tail of the mouse okay the tail of the mouse and uh it begins with the uncle Mouse which is the first patented animal okay uh this was in 1998 it was uh patented by Harvard and Dupont it's a new mouse and it was actually a set of techniques for

um changing the genes in uh mice lines okay and that's extremely useful because uh you then study these mice and can learn all kinds of things about disease and treatments and so forth an onco is cancer so the first ones that which were sort of inserted these insertion of the genes were cancer genes so they created this new uh Mouse and as you know if you build a better Mouse the world will beat a path to your door lots of people wanted to use this mouse right but DuPont uh they charged a lot of high fees charged really high fees to just use any of these ideas you want to use these

ideas to insert genes whether to create like you know an uncle rat or a onco a primate or whatever okay or insertion of other genes you have to pay DuPont in addition you have to report annually what are you doing with these ideas these techniques and to write an annual report to DuPont uh anyone using these ideas automatically automatically gave up a share of any of the downstream licensed Revenue and you are not allowed to share your new mice lines so if you created using these techniques a new type of mouse you're

not allowed to share that okay and that really violated all of these things but at that especially really violated the Norms of biological science at the time uh because they were much used to and much more used to an Open Access model okay uh they were much more used to the Jackson laboratory model which is a quite of an amazing thing like when you create a new mouse uh the way it works is you send some examples to the Jackson laboratory and they keep them for you okay they keep them and then if anybody else wants to study

these mice okay you can go to the Jackson laboratory and this is their web page and you can go find the mouse that you want so if you want to study like fat mice obese mice balding mice okay cancerous mice you know hairy mice you go to the Jackson laboratory and you find your mouse and then you can do your studies in a consistent line okay so with all the other papers which have been written on this so this was not very popular with the scientists and in fact they began sort of uh sharing their mice under the table okay however however all of this outrage okay led to an

agreement okay with the NIH to free the ankle Mouse to lift a bunch of these restrictions which is what happened in 1999 okay they lifted a bunch of these restrictions and then there's this very clever paper which I've been drawing upon by Fiona Murray and uh colleagues which uses this natural experiment to ask well what happened okay when these restrictions were lifted these ip-like restrictions what happened to the scientific discovery process okay when these restrictions were lifted okay and they're going to do that by looking

at scholarly citations because each new mouse line is associated with a mouse article okay the key article which defined this type of mouse right so there are new mouth lines created with the restricted techniques the DuPont techniques and there are mouth lines created using the non-restricted earlier techniques which were did not have all of this IP like uh uh restrictions so what Fiona Murray and colleagues do is they do a difference in difference

which is to say they look at citations to the restricted Mouse articles relative to the unrestricted ones before the restrictions were lifted and after the restrictions were lifted and what they find is that the citations to the mouse articles which were restricted go way up relative to the ones which are always unrestricted okay so there's a flowering of research using these new uh now unrestricted Mouse lines moreover you find new researchers entering the field okay you can see that

from the citations you find experimentation went up by as measured by new keywords okay and you find new ideas developed in a wide and diverse set of fields as measured by citations from new journals okay so what you really see is this is the idea that these restrictions really held back the willingness of scientists to use these techniques and when the restrictions were lifted you had a big increase in innovation moreover no evidence for a reduction in the number of new mouse lines after

the restrictions were lifted so that's I think an interesting illustration of how these IP like processes can restrict reduce innovation you know there's a very famous example of that point in crypto yeah which maybe you now right right yeah so I was told last night that uh Satoshi uh used the sort of less efficient less uh uh clean ebsca I guess that is uh a citation uh uh uh signature uh management system as opposed to the Snore because the Snore was patented and I think a lot of

cryptographers have regretted that ever since the irony is that it hadn't expired the year Bitcoin launched so it actually would have been it would have been fine he missed the deadline but it was just widely perceived as being patented right right yeah so uh kind of uh sort of summing this up uh this part you know uh uh Newton famously said that uh if I have seen further than others it is only by standing on the shoulders of giants and the question then becomes would Newton have been able to see so far had to had he had to pay for the privilege of standing on these shoulders we can also see here that when you have

a lot of these patents and the IP like restrictions we can actually get the tragedy of the anti-comments okay so the tragedy of the commons is when you have a resource which doesn't have owners that's unowned and you have too many people using the resource and so it becomes overused and under maintained the tragedy the anti-commons is when you have too many people who can say no too many people who can say no to use the resource and so the resource becomes uh underused okay underused the idea here is if you're trying to build a product

and it depends upon sort of four other products and each of them demands you know 25 of the profits that's a problem if you're building a product and it depends upon 10 you know other products and each of them demands 20 of the resources that's a real problem okay a real set of transaction scoffs there to get agreement and in fact we saw this particularly in the early days of the cell phone industry uh this is a picture of all of the lawsuits okay in the cell phone industry who is suing uh who okay and uh obviously this was no way to create innovation

so uh what happened well we had all of this lawsuits and eventually what happened is you may recall this but Google Microsoft Samsung and so forth Apple they went on a patent buying spree okay and they brought up these billion dollar patent arsenals okay and the reason they did that was not because they needed the knowledge in this patent Arsenal they already had the knowledge Okay the reason they did this was because they wanted to have a plausible way of countersuing anyone who sued them

so they bought these patent arsenals these billion dollar patent arsenals so that someone sued them the control through their Arsenal and find plausibly or semi-plausibly some patent which they could then use to counter Sue okay so uh this led to what I call sort of the mutually assured destruction uh view of uh Innovation okay and to give it it's due it worked you know just like the nuclear mutually assured destruction has managed to avoid nuclear Wars but it still doesn't seem to me to be the ideal way of doing it and doesn't seem to me to be the ideal way of creating room for

innovation for one reason because it's clear that only the big could do this then only the really large companies can afford to buy these patent arsenals okay and in fact uh you see this uh in the data that there's been a inequality a greater Rising inequality in uh patents who owns them and who gets them okay so this is the share of patents held by the top one percent of patenting firms okay and that's increasing right there's a greater inequality they become more concentrated more concentrated in

the top the top one percent the number of patents are created by firms uh for the very first time creating a new patent okay that's been going down followed by more than 50 percent a fall by more than 50 percent yeah is there a con is there a counter argument on the big versus small that patents basically enable innovation of the small ones that the big could easily imitate and therefore enables the smaller participants to innovate in the first place so I mean there is this argument that it allows um it gives the so it yes there is that argument that it

gives a small firm something to sell at least okay um but if we're thinking about sort of modern innovation you're typically not building a product on a single patent okay it requires creation it requires using a whole bunch of patents and that the small firms can't do okay that's what the small firms can't do and as as we pointed out there there's fewer and fewer of these uh small firms with patents and they're flowing to the big firms uh this is also the big firms are buying the patents okay they are buying up the patents again a really large increase in

the share of the top one percent buyers and small buyers are in the decline there's also a very interesting paper by accident and my former student actually Nathan goldslag and they have this fantastic database on patents and all of the scientists all of the innovators which created the patents and they can actually track these uh scientists or these innovators over time and what they find first of all is that the innovators at incumbent firms are increasing okay and then they do this micro uh study where they look at what happens to

innovator earnings when the innovators uh are hired by an incumbent firm relative to very similar innovators who are you stay at an entrance okay and what they find is that when the innovator is hired at the incumbent firm their earnings go way up sort of makes sense but actually their Innovation goes down okay so you might ask well why are these incumbents hiring these innovators paying them a lot

and actually seeing their Innovation go down doesn't quite make sense but there is at least one story which not necessarily the only story but it's one story which uh makes sense of this data and that is uh the arrow replacement effect okay so the arrow replacement effect is argument by Ken Arrow who points out that the cost of innovation to a monopolist is higher than to a competitive firm because the monopolist has an opportunity cost the monopolist okay that Innovation May threaten the profits they already have

okay so the cost of innovation to a firm which already has these uh profits is quite High so they are less likely to want to innovate one way of doing that is to higher up a bunch of innovators and then put them on some other project which doesn't threaten your Monopoly rents so that's one story which makes sense of this is this sort of equivalent to like buying companies buying patents and then shelving them yeah yeah exactly that's like a that that would be the very direct way of doing it and uh what this paper shows is that you can do it surreptitiously okay by hiring the innovators at least I'd say as I say one

explanation for this the these patterns then can slow the diffusion can slow the diffusion of uh new ideas uh reducing growth and the way they do this is by protecting the leaders from the followers and I call this the Tiger Woods effect okay uh the story here is Imagine That Tiger Woods and myself are in a golf game okay well neither of us is going to put in much effort Tiger Woods because he doesn't have to because he's going to win for sure and me because I know I'm going to lose for sure so I don't put in much effort either okay and a patent can create a Tiger Woods effect where the leader

is insulated from the followers making it more difficult to catch up which was reducing the diffusion of ideas and this reduction in diffusion of ideas is something which shows up a lot okay it shows up a lot of data in the United States and in Europe as well so this is labor productivity um sort of the value added per worker between the frontier firms okay and the other firms in the industry this is for Europe and what you can see is that there's a a bigger Gap okay a widening gap between the productivity of

the top firms and the productivity of the average firm okay and one possible explanation again for this is that these Technologies these ideas are not diffusing as rapidly as they have in the past okay that's patents I want to say a little bit about copyright uh as well it's kind of some fun kind of fun as well copyright a little bit of History copyright terms uh began in 1790 at 14 years okay so your copyright would last uh 14 years now who wrote the first U.S Copyright Act uh historians have looked at this okay and you all know this the name of this

person this person's book it's possible it might even be around here somewhere so any guesses does anybody know first Copyright Act was written by Noah Webster okay author of Webster's Dictionary okay still being sold uh today right uh and this guy uh is Noah Webster who was Webster's cousin nobody you've probably gotten heard this name before this is Daniel Webster okay got a well-known famous Senator uh and Noah

and Daniel Webster were the key players also in the 1831 uh extension of copyright from 14 to 28 years all right so a lot of rents you can going on and remarkably Daniel Webster in the Senate was joined by this guy Oliver Ellsworth in the house there's no reason you would know this guy's name Oliver Ellsworth but who was he he was Daniel Webster's son-in-law okay so he was the guy also standing to gain by the extension of copyright law once his father-in-law passed away all right so right from the very beginning we see these sort of rent Seekers try to extend

uh copyright law now it actually stayed after the 1831 expansion from 14 to 28 years the the length uh stage did not increase again for some period of time though the law got much more broad so it started to apply to many more uh things uh including like news and photographs and the technology changed so the log got more abroad but it didn't lengthen very much until in 1976 it was lengthened to life of the author plus 50 years and then it was expanded in 1998 to life of the author plus 70 years

okay and this is one of my favorite stories because this expansion it's a life of the author plus 70th it applied to already existing works okay so there's clearly there's no incentive argument right you can say well we want to extend copyright so that we will get more you know authors writing books but this extension applied to stuff which already existed right so there's no possible incentive argument for this expansion obviously however it is the past authors particularly the successful ones who most have the ability to lobby for and who most want uh an extension

and this expansion in 1998 was opposed in a Friends of the Court brief by George akerloff Kenneth Arrow Ronald Coast Milton Friedman and James Buchanan to name just the Nobel prize-winning Economist who opposed this expansion okay a bunch of others as well nevertheless despite this intellectual Firepower never before seen okay they lost and they lost to a mouse okay they lost to Disney right Disney was the main uh lobbyist uh pushing for this in order to keep the uh

in order to keep copyright on Mickey Mouse uh you can see some of the costs of expanding this copyright this is a neat little paper um as I mentioned it applied to past existing works but if a work had gone into the public domain right then it stayed in the public domain right and the cut point was like 1924 okay so uh prior so 1923 there was no extension 1924 there is an extension okay and this is the number of imprint editions like that you could go to Amazon today and find okay so what you see here is that there's a a big decrease

at this cut point with the extension of the copyright law okay extension the copyright term you get this big decrease in the number of new editions of these works that you could find you know on Amazon and you know prices of course are going to be lower here and higher uh here right and that's just one kind of direct cost yeah curious why that is the case because I'm assuming even the later books are now free of copyright because it's been 70 years no no well no 1925. well um so so it would be life of the author plus 70 years

so uh at the time of the uh at the time of the study these are all still copied these are all because at the time the study not at the time of the 1998 yeah so so these are these are all uh uh this this is the dividing line for the uh 1998 La act okay and so these are still under copyright and these are not yeah now that's got a direct cost but there's actually I think a much more important sort of indirect costs and that is to think about the commons uh the public domain stuff is the public domain as actually being uh the

opportunity of the commons I call it the opportunity of the commons and I think about the intellectual Commons as a Fountainhead okay a Fountain Head that doesn't run dry but is a source of ideas to build upon to remix to combine to develop to change to alter and so forth so you think about Disney and he is drawing from the brothers Grim okay and reworking these stories uh from the brothers grip with Cinderella right and then you get a different version Ella and the prince a kind of a news a new telling right a remixing right and here we have you know the Hansel and Gretel

and then we have like a modern version of Hansel and Gretel okay right uh uh we got these but but you you think about Romeo and Juliet and you can have you know Romeo and Juliet as a musical you know in South Beach right that kind of thing this develop taking from The Commons and redeveloping and remixing it and so forth now it's true that not necessarily you know this is all of the remixes are not necessarily good uh you might have noticed uh heard perhaps recently that this year uh Winnie the Pooh okay went out of copyright okay and you thus we have

Winnie the Pooh blood and honey okay do not take your children to this movie okay under no circumstances should you do that but if we think about this comment it's kind of remarkable um so you know John Locke right said that it's okay to draw from the commons if enough it is good is left for others right now this intellectual Commons is what I call Super lockian okay because when the IP law is well designed everyone who draws from The Commons eventually restocks it okay when their

own revisions are made are coming to the public domain okay so IP law is super locking in and that those who draw from it also restore it also add to it and yet the problem here is that not everyone does this so people sometimes say that uh people who draw from The Commons are Free Riders but actually the Free Riders are the people who do not restock the commons so Disney Drew from The Commons The Brothers Grimm and so forth and the old folk tales Disney Drew from that but has not restocked the commons because they've kept their

own works in the private domain by extension of copyright law and you can see this kind of Destruction of the commons and this goes to your point had the pre-1976 rules been in place okay all of the following works would be available in the Commons today for anybody to rework mix combine alter or expand wins and socials history of the English-speaking peoples Minority Report Philip K dick James Bond okay James Bond would today be in the public domain okay we would have James Bond like we have

um Sherlock Holmes or Robin Hood okay we don't have that we don't have that in fact under current copyright law none of these Works will enter the public domain until 2052. so what do we do about this okay what kind of potential policy reforms are there well patent law is interesting because to a large extent patent law is actually judge created law not so much legislators matter but it's also judge created law so all this stuff I was talking about the extension of patents to business methods and so forth um all created by judges right and what the

judges can create the judges can also take back and in fact that has started to happen you know I mentioned there was this creation of this federal circuit court in 1982 special court for patents which was became very very Pro patent an expanding patent in all of this ways but more recently the Supreme Court okay has started to smack him down okay uh has started to say no and the Supreme Court has been saying no no you can't you can't patent uh genes you can't patent this you can't patent they've been pulling back in a wide variety of areas okay uh they've been

you can interpret patents in a lot of different ways you can interpret them narrowly or broadly okay judges can do that right uh ramping up the novelty non-obvious requirements uh ramping up the enablement uh requirement so there are a lot of things which the courts have begun to do particularly the Supreme Court to kind of pull patents back which I think has generally been uh positive there's also a question of like more fundamental reforms which I've written about it seems very peculiar actually why does every innovation get a 20-year patent

okay remember they the argument for patents is when your Innovation costs are much higher than your imitation costs okay so like Pharmaceuticals are the classic example your first pill costs a billion dollars your second pill costs 50 cents right so that's a classic case where you kind of you want patents but where when you're in Innovation costs are much closer to your imitation costs like software for example then there's much less of an argument for patents so why 20 years for everything okay uh how about creating a patent system which is more responsive to your Innovation

versus your imitation costs you could either literally require people to uh show some receipts for their r d costs okay or if that's too complicated uh let's make some divisions by type of Industry okay and maybe we should have like a three-year patent 10-year patent a 20-year patent okay if you want a 20-year patent you better show you know I got really high r d costs something of that that nature or as I say you could do it by broad class that's one big possible reform uh what is yeah you calibrate those numbers yeah so I think if you had so first of all we

already require for like the investment tax credit we already require people to you know uh show their receipts for investment so I don't think it's uh uh impossible to have the same thing for our r d uh second if you're going to have like a 3 a 10 and 20 year you really only need sort of crude classifications and you could make it uh yeah you could make it partly yeah uh uh uh you could create it with a participation constraint right so if you want a three year patent you get it very quickly you know you want a 10 not so bad you want a 20 then we require kind of more from you so I think it on that kind of level you could do it

pretty pretty reasonably thanks yeah I'm gonna say we already have two kinds of patents at least try with design and utility patterns correct so correct so yeah so I'm talking exactly so so that's right uh so I'm talking about um changing what we currently think about of the design patents yeah exactly um yeah a final type of intellectual property is actually intellect property right uh which is you may be familiar some of you may have had to sign some of these uh like non-competes okay so a non-compete agreement is when a employee agrees that

uh if the terminate their employment you know either they leave or they're fired either one okay then they agree they're not going to go work for a competitor for maybe two or three years okay that's a non-compete agreement and that's the type of intellect intellectual property it's like when you can't when ideas are in heads so it's a way of controlling the heads right if you can't keep the ideas with a trade secret or patent you try and keep the ideas by controlling the heads right um and these are non-compete agreements Biden has uh is arguing we should get rid of them okay which I think is a plausible our argument

um and they have gone up by a lot as well so 18 percent astounding really of the U.S Workforce is governed by a non-compete seems very very high like the base case for them is like if you have like a trade secret or you have a customer list okay you can sort of understand why you might want to non-compete but 18 of the US Workforce seems very high and that includes fifteen percent of workers without a college degree right and it's quite plausible and there is some evidence for this that these non-competes also reduce the diffusion of ideas

okay so one way in which ideas diffuse is by people moving from firm to firm and that in fact is part of what made California's Silicon Valley so successful is that California does not enforce non-competes so when workers move from firm to firm the take their ideas with them and you get this diffusion of ideas I think here there may also be a kind of prisoner's dilemma of Ip which applies to non-competes but applies to IP more generally and that is any firm of course is going to want to say yeah we want non-competes okay but when a firm comes to hire workers

it then finds it can't hire them because other firms have got non-competes so what's good for the firm may not be good for the industry all right you need to look on both sides like every firm is going to want non-competes but that makes their hiring more difficult as well so uh perhaps getting rid of them might be a good idea okay all right yes so let me summarize or let me conclude and that is that a lot of what IP is by no means all but a lot of what you think about it it's all about sort

of uh dividing the pie right okay and the problem with dividing the pie is that when a lot of resources go into dividing the pie you spent a lot of money on the lawyers and so forth and all of these laws about who gets what share okay a lot of resources go into dividing the pie then the pie actually gets smaller now IP is actually in the U.S Constitution right it was put into the U.S Constitution why to promote the progress of Science and the useful Arts and it makes sense that over time sort of Ip has gotten a lot stronger because the only people who pay

attention are the beneficiaries okay the rent Seekers the firms Disney's and so forth who want stronger and stronger IP laws right but I think it's incumbent upon the public to every now and then take a re-look to revisit this question and ask whether IP is still continuing to promote the progress of Science and the useful arts or whether we need to take a second look thanks very much I wonder if you could uh comment on all the kind of debate around regulation of large language models and how that sort of fits into what you've said today

right yeah so uh so there's a whole bunch of different questions uh one of which sort of goes well beyond IP right is the Doomer scenarios right the people who are worried about uh artificial intelligence um taking over the world and you know uh creating becoming more intelligent than us and you know sort of doing away with us which is actually gotten a lot of much more play in Silicon Valley uh than than you might imagine it's quite kind of surprising actually and my view of that is that it's all so far it's all

Theory it's all rationalization there's no empirical no empirical evidence for that as well so I want to wait some time before we see you know whether the AIS are taking over um in terms of now IP I guess the big issue is you know people are complaining right that uh uh mid-journey as learned about all of the art of the world and is then using that to create you know more art and my view is look that's exactly what artists have been doing for centuries that's what artists do okay so you know Michelangelo doesn't arise in a vacuum

you know he looks at what the artists around him are doing and he modifies he changes he does something a little bit different he builds upon those foundations and mid-journey is no different than Michelangelo so uh I see absolutely no reason to Grant uh IP to looking at artwork or to looking at these books or Internet materials which the large language models have been trained upon that's no different than saying you know like if you read a book then the author gets a share of what you then use to produce you know using you the knowledge

you got from the book like that's crazy so um yeah I I the large language models they're not databases okay they're not databases so they are learning from past materials whether it's photographs or artworks or text and then they're producing new original texts which have never before been seen in the history of the world right so you can go and ask you know chat gbt you know create a rap about Fred Flintstone and uh shakes in issue or take it to create a Shakespearean sonnet about Fred Flintstone and it will do it okay and

that's totally original it's creative it's using material from other sources but in just the same way that every artist has always done so I don't see any reason for regulating on those grounds Joe I guess to follow up on Tim's like there's this question because you can't I believe like you can't copyright the output of a machine um there has to be like human creativity so do you have an opinion on like what what level of involvement the human needs to have had in the creative process when it becomes like if you just say Fred Flintstone rap and push a button have you sufficiently contributed to the creative process for it to right be copyrightable so forget

about who influenced you but if you want to copyright the output right that's a good question and I don't have a uh yeah I don't have a well thought out answer on that um so the way it is now like mid-journey has some copyrights but they give very a very broad license and I think so I think that's something which we're just going to have to figure out um it's a pretty new pretty new technology uh so yeah I don't I don't have a good answer on that it's a good question I had a totally different question too

you didn't mention trademark very much um yeah in your opinion is trademark working the way it is I've heard this fear that Disney and some of the others when there are some of their early stuff is going out that they're going to sort of shift to asserting trademark rights instead of copyright on characters like Mickey Mouse and where the copyright because trademark never expires right fall back to trade right so in general I am much less worried about trademark um because that still leaves open a huge amount of competition trademark is really about um making sure that the consumer is getting what they think they're getting right is this really uh Yves Saint

Laurent bag and so forth right so it's really I see trademark as much more consumer friendly in that it's not creating a monopoly so much as helping consumers to figure out what is real and what is what is fake um so I'm I'm less worried about trademark however I think you raise a good point in that if Disney loses copyright which eventually we don't know when presumably eventually they will maybe they would try and Resort it in other ways um still I think if there's no uh if the consumers are not being fooled

then there isn't a case for a trademark so you know it's pretty obvious that uh A.W Mill and the author of Winnie the Pooh was not the author of blood and honey okay in the movie I showed you earlier right and so uh if if it was obvious if it's obvious that this new version of uh you know Mickey Mouse or this new version of James Bond you know if it's not uh if it's distinguishable then I don't see any reason for the trademark should intervene there yeah you should be able to use it with the trademark yeah so this graph about the decline of productivity yeah um is there like any empirical evidence it's tied to

the rise of this rent-seeking behavior and if so like what prompts the rise of the Strand seeking Behavior because in my mind laws typically follow like some other General Trend in in economic behavior so what is this why is there is a shift towards like more than seeking Behavior right yeah so it's a big debate about what caused the Great stagnation um you know one theory is that uh is a in fact they're they're actually a fewer ideas okay and that it's a scientific uh reduction which you know I mean if you think about the number of

fundamental new technologies uh in the post-world War II era or in coming you know around that time it's a lot of them right um and like the jet aircraft for example right uh but no improvement in the real almost no improvement in jet aircraft you know since the 1955 something like that um and is that more about the technology itself or is it about regulation like Banning SuperSonics uh a little bit hard to say I think there is some evidence that part of it is due to these forces which I've been talking about it's very difficult to correlate the this General decline with these individual forces but

if you if you look at that diffusion for example so in Industries where there's this greater diffusion you actually see slower aggregate productivity growth as well so uh on a industry level there is a connection between the slower diffusion of ideas and slower aggregate productivity growth but again I wouldn't it would be too much to ascribe all of this to describe all of that decrease in productivity to these forces as to what it's creating this I mean I think rent seeking has always been around um but part of it is on a big level is maybe regulation part of it if more risk

aversion and so forth um uh and and maybe part of it is uh that is becoming harder to come up with new ideas right there's actually some evidence for that in we all know Moore's Law right uh which is holding pretty well okay even today but if you look at the number of Engineers the number of uh you know electrical engineers sort of needed to to keep Moore's Law going that has been going up and up on up exponentially so in order to keep Moore's Law at the same Pace we've had to put more and more resources into this so uh so it's a combination of these things and it's hard to say which is the

most important okay thank you very much [Applause]