Hacker News new | past | comments | ask | show | jobs | submit login
Patents and innovation in economic history [pdf] (gwern.net)
49 points by gwern on March 16, 2016 | hide | past | favorite | 21 comments

TL;DR: politicians and lazy academics typically consider patents to be the driver of new inventions. Turns out that research doesn't bear that out -- who knew?! In fact when patent laws are too strong or too broad, development slows down. Shocker!

(Actually it's good to have someone actually look into the data and challenge the dominant assumptions. And someone from the existing power structure, like the NBER in this case.)

That is a rather misleading and incomplete TL;DR. The paper actually references a number of other papers from the author (Petra Moser) and presents a number of salient conclusions. Here's one important point: Patents encourage innovation in industries where other methods of maintaining advantage, such as secrecy, are not available.

> If patents are important in some industries (such as manufacturing machinery) but not in others (such as scientific instruments or chemicals in the 19th century), changes in patent laws may influence the direction if not the level of technical change (Moser 2005). These patterns are borne out in exhibition data: Countries without patent laws contribute as many exhibits and prize winners as countries with patent law. But their innovations are disproportionately focused on industries in which secrecy is effective, so that inventors are less dependent on patents.

Moser 2005 also points out that as the efficacy of reverse engineering increased (today you can reverse engineer almost anything pretty cost-effectively), more exhibitions started to focus on industries with stronger patent protection.

This is a curious point, however: it almost suggests that the proportion of innovators in a given population is (was) fixed, and they would innovate where the promise of rewards is highest.

Another important point: Cheaper patents meant "democratization of invention". There were more exhibitions with patents from people in rural areas in the US, where patents were cheaper, compared to UK, where patents were more expensive and almost all inventors were based in London.

One problem with this paper is that it makes broad generalizations based on narrow data: The fact that plant patents didn't increase the number of new rose varieties registered gets the headline "Intellectual Property Rights for Living Organisms Have Not Encouraged Innovation". It draws conclusions about innovation based on exhibit data where, by the author's own admission (Moser 2005), exhibitions may have significantly discouraged the inclusion of exhibits that were easy to copy.

Also, it's important to look through the other cited studies, because they almost always present data that is different from other papers. Some data can be more compelling than others.

Nonetheless Moser provides a fascinating view into how innovation may be influenced by incentives and policies.

> "This is a curious point, however: it almost suggests that the proportion of innovators in a given population is (was) fixed, and they would innovate where the promise of rewards is highest."

True but I'll add that often for individuals to quit their jobs and focus on innovating requires more than just "the promise of rewards". Immediate funding often trumps the promise - and that funding comes from some kind of investor. Many investors view patents as: 1) a way to recoup at least some their investment dollars if the innovator can't monetize, 2) prevent competitors from simply copying the results of the R&D.

I find it very likely that many "politicians and lazy academics" actually already know it to be false that patents is a cost-effective driver of innovation. However, it is just so darn easy to count patent and claim something like "innovation is rising" or "patent count is falling, we have to hand out more patents so innovation can be seen to rise" etc.

Those "many" are still surely the minority. Most people, including academics and politicians, buy the patent propaganda.

Moser references James Bessen who wrote a highly-cited paper that has been largely debunked [1][2][3]. Specifically, the paper claimed patents cost the economy $29B per year.

> "Actually it's good to have someone actually look into the data and challenge the dominant assumptions"

Challenging the dominant assumption would be to challenge Bessen's number and the numerous articles that automatically assume that value to be true.

[1] http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2117421 [2] http://www.iam-media.com/blog/Detail.aspx?g=454c1adc-52c3-4c... [3] https://gametimeip.wordpress.com/2012/07/30/patent-scholars-...

I wonder if there have been any attempts to quantify the economic fruits of government-funded research, e.g. by ARPA (especially before being limited to narrowly defense-related projects) or NSF, or research directly undertaken by executive agencies like NASA, etc.

Government-sponsored basic research and scholarship often comes under fire from asshole lawmakers claiming to have the interests of “taxpayers” in mind (cf. e.g. https://en.wikipedia.org/wiki/Golden_Fleece_Award, or more recently https://www.congress.gov/bill/114th-congress/house-bill/3293...), but as far as I can tell, such research has been one of the primary drivers of “innovation” and economic growth in the US over the past 70+ years.

Thats the fallacy of the seen: You don't know what would have been discovered otherwise. Non-government funded discoveries include salk and Sabin polio vaccines and the chemiosmotic effect.

I personally notice that a lot of government funded research goes to PIs of questionable intelligence that have merely mastered grantsmanship... How many brilliant young scientists have had their careers cut short by being discouraged by the granting system instead of going on to make discoveries in science?

The grant process is terrible. Let's fix it.

Or are you implying that for some unspecified reason that would be impossible?

It's impossible. Responsible government has to have a mechanism for accountability, and this is antithetical to basic science.

By that logic the only way to fund basic science is for independently wealthy scientists to fund their own research, since otherwise they would be accountable to someone.

I expect the problematic accountability would be accountability for results, but why would that be necessary for government funded research? Negative results are progress, not failure. Now you know something doesn't work when you didn't before.

On the other hand, if there was accountability for e.g. the research being reproducible, why would that be antithetical to science?

Because protecting taxpayer money which is collected nonconsensually incurs a higher standard of accountability. Accountability is important because a scientist can claim basic science to engage in irresponsible dalliances, and cover by claiming "basic science". It happens all the time as a part of Grantsmanship

How is the problem of irresponsible scientists different for the government than private funders? I can assure you that private funders want the highest standard of accountability available.

Private funders can be assumed to have taken that risk when choosing to spend their money. Public funders do not. Like it or not, for example, congress has continued to appropriate millions for "holistic" and "integrative" medicine research. When for the most part it was discovered to be bunk, they fired the chief administrator and redoubled efforts under the justification that "not enough had been spent to achieve a positive result"

Discovered in what otherwise?

It's not like public research funding comes at the expense of private research. Public research funding allows there to be a higher total number of scientists doing research than there would be, all else equal, without it.

And there are clearly things patents don't provide the right incentives for. Anything that benefits all of humanity is very difficult to recover the value of as a patent because engaging in licensing transactions with every single person everywhere is prohibitively expensive. And then the license fees artificially reduce adoption of the invention until the patent expires due to supply and demand (price goes up, demand goes down), making it even harder to recover the full value of the invention. Which would imply that patents necessarily result in under-investment in research.

> "it's not like public research funding comes at the expense of private research"

Yes, it does. It sucks productive capacity of the scientists, and also has a system that's destructive to the creative potential of people actually doing it and trying to "make it", ultimately incompetent people are promoted, and that causes a death spiral as competent people are disinventivised from continuing to participate because the smart and observant ones quickly realize labor is exploited by their lessers, and they get out. Ironically, increasing public funding to the sciences makes this worse by rewarding the existing superstructure of leadership (and creating more bureaucracy to administer). The bureaucracy is even worse: who decides to become a science administrator instead of a PI? Only the people who really couldn't hack it. (With apologies to the two exceptions I know personally)

> "Anything that benefits all of humanity is very difficult to recover the value of as a patent because engaging in licensing transactions"

"Anything that benefits all of humanity" should be recognized as inherently valuable to fund and therefore would be useful to look at reducing the costs of "engaging in licensing transactions". In other words, society may benefit by making it easier (or less costly) for the researcher or innovator to be rewarded - thereby increasing the incentive to create something to benefit humanity.

What are you actually proposing? Some abstract means of reducing licensing transaction costs? How?

Salk was a university researcher, and his polio vaccine had its research and trials funded by public donations to a tax-exempt non-profit organization which was founded by FDR, the US president.

I’m not quite sure what you’re getting at. Are you suggesting that scrapping government funded research in favor of letting tax-exempt charities handle it instead would lead to just as much innovation?

This seems not to reference one of the strongest and most thorough recent reviews of this whole topic, "The Case Against Patents": https://www.aeaweb.org/articles.php?doi=10.1257/jep.27.1.3

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact