Who gets the gains from innovation?

In this paper we merge individual income data, firm-level data, patenting data, and IQ data in Finland over the period 1988–2012 to analyze the returns to invention for inventors and their coworkers or stakeholders within the same firm. We find that: (i) inventors collect only 8 percent of the total private return from invention; (ii) entrepreneurs get over 44 percent of the total gains; (iii) bluecollar workers get about 26 percent of the gains and the rest goes to white-collar workers. Moreover, entrepreneurs start with significant negative returns prior to the patent application, but their returns subsequently become highly positive.

That is by Philippe Aghion, Ufuk Akcigit, Ari Hyytinen, andd Otto Toivanen in the new AER.  Via Ben Southwood.


Consumers get 100% of the gains

Yes, the consumer is at the end of distribution. 70% of economy.

Well, Linus Torvalds got about 0% from his invention of Linux/Android.

To use a Finnish inventor who has probably had a greater impact on the world around us than any other living Finn.

I bet he is able to make much better money in consulting fees and conference speaking fees than if he had never invented Linux. Not all returns need to come from patent licencing. in fact this is a relatively new model of entrepreneurship - create something that people can use for free to get it widely adopted, then you get the necessary status to profit in other ways. This model is really only possible now that there are literally billions of people able to use your product.

'I bet he is able to make much better money in consulting fees and conference speaking fees than if he had never invented Linux.'


'create something that people can use for free to get it widely adopted, then you get the necessary status to profit in other ways'

Stallman would suggest (OK, explain in detail) that is not the point of the GPL, a point that Torvalds recognized when creating an OS that can be shared and expanded without requiring money or permission (as long as one follows the copyright terms embedded in the GPL, of course)

'This model is really only possible'

You do know why Torvalds created Linux, right? 'In early 1991 he purchased an IBM-compatible personal computer with a 33MHz Intel 386 processor and a huge 4MB of memory. This processor greatly appealed to him because it represented a tremendous improvement over earlier Intel chips. As intrigued as he was with the hardware, however, Torvalds was disappointed with the MS-DOS operating system that came with it. That operating system had not advanced sufficiently to even begin to take advantage of the vastly improved capabilities of the 386 chip, and he thus strongly preferred the much more powerful and stable UNIX operating system that he had become accustomed to using on the university's computers.

Consequently, Torvalds attempted to obtain a version of UNIX for his new computer. Fortunately (for the world), he could not find even a basic system for less than US$5,000. He also considered MINIX, a small clone of UNIX that was created by operating systems expert Andrew Tanenbaum in the Netherlands to teach UNIX to university students. However, although much more powerful than MS-DOS and designed to run on Intel x86 processors, MINIX still had some serious disadvantages. They included the facts that not all of the source code was made public, it lacked some of the features and performance of UNIX and there was a not-insignificant (although cheaper than for many other operating systems) licensing fee. ' http://www.linfo.org/linus.html

To add another person as an example, though not Finnish, there is a certain Tim Berners-Lee.

Lots of the current world's greatest inventors were utterly unconcerned about profit - instead, they were trying to solve a problem.

Really good post, thank you.

+1. I didn't know the genesis story.

Lots of the current world's greatest inventors were utterly unconcerned about profit - instead, they were trying to solve a problem.

I am not sure that granting patents nets out positive.

People are free to give their share away. It was his choice.

There is compensation and there is compensation. For instance I bet Torvalds has all the groupies he could want these days. Now that may or may not appeal to him - and he did marry one of his students. But also there is the adulation he gets at conferences in America.

However he certainly gets to be a pain in the backside to large numbers of people for a variety of reasons in ways that might not be tolerated at a real company. From which he does seem to take some pleasure.

Well, Linus Torvalds got about 0% from his invention of Linux/Android.

He was not to know it at the time he started Linux but that is not quite true. He was given part of the RedHat IPO and his share options were worth millions. Not a lot of millions but not nothing either. I don't know how much he still has.

The last estimate I saw for Linus Torvalds' net worth was $150M, and an approximate annual income of $10M.

Good for him - any idea how much comes from git?

What was Linus Torvalds' "invention"? Linux was just a copy of a proprietary product; distributing it free via the open source model was the sole reason for its popularity.

Don't get me wrong, Torvalds is very talented, and developing Linux was a huge contribution to the world. But it isn't an "invention" that he failed to profit from. And if he had maintained proprietary control of it, most likely no one would ever have heard of it.

'Linux was just a copy of a proprietary product'

No, it was an independently developed OS. Absolutely no copying was involved, as SCO spent millions proving.

Sigh. I didn't mean he stole the source code. Why would you even think that?

To call Linux a copy is precisely the same as calling Windows a copy of the Macintosh operating system. Which Apple did (more or less), by the way - https://en.wikipedia.org/wiki/Apple_Computer,_Inc._v._Microsoft_Corp.

And just one of Linux's more interesting features, due to how Torvald's licensed his work, was how it could be easily adapted to run on a wide range of processors, while still allowing the user to work with Linux transparently.

Basically, Linux transcended the (generally artificial) constraints of essentially all previous OSes, with a world wide community of programmers that contributed freely, without requiring permission to try out new ideas. This was not anarchy, of course, but in comparison to how a company like IBM, Apple, or DEC developed an OS, this process was revolutionary. And before it proved itself in practice, derided as being absolutely impossible at creating a functional OS.

'Why would you even think that?'

Because of a broad PR effort to convince people precisely that, 15 years ago. (A similar effort continues with Java right now, by the way, even if the PR has generally faded from public view - https://en.wikipedia.org/wiki/Oracle_America,_Inc._v._Google,_Inc.

Torvald essentially reverse-engineered Unix in creating Linux without utilizing any of its source code. This is not a "copy", so prior is right on that score.

However, I would point out that BSD Unix preceded Linux by some years, and was also an effort to have a free version of Unix widely available.
It took a while but Net 2 (a complete version of Unix with no substantial AT&T source code) of BSD was released in 1991. Legal arguments surrounding BSD vs AT&T Unix were one of the reasons that Linux eventually gained share over BSD; other reasons included differences in the terms of the license agreement and the more widely-spread contributions to Linux as opposed to the more tightly controlled BSD development. So, "transcended the (generally artificial) constraints of essentially all previous OSes" is a bit of overstatement.

This is where a discussion of the GPL would be relevant, but this is not the place for it. Selecting the GPL was what allowed Torvalds to ensure that everyone could share in the development of Linux.

And that Torvalds alone was able to transcend the fragmentation that has always plagued Unix versions/derivatives/clones is pretty clear - whether that would be covered in the aspect of transcending constraints or being an overstatement is a matter of opinion.

Isn't that result what it should be? The inventor in my experience is not the one doing all the grind - e.g. sales and marketing - for the product, or at least not the one doing it effectively. As an inventor myself, I've seen how hard it is to market ideas. On their own, they are dime a dozen as they say.
I understand they only use narrowly defined technical innovation (patents) because that is what you can measure. I would like to see a study about what's the deal in the arts. Pop music on Youtube is a very innovative activity sector, but you usually need good production, polishing and finishing so innovations is def a team effort there also. I suspect the singers/headliners get more than 8% even though they are usually neither the authors of the tune nor of the text these days (some exceptions like Ed Sheeran notwithstanding).


Though I don't know Ed Sheeran.

If sales and marketing are not rewarded well it is probably because anyone can do it. There is no special talent attached to the role that needs a great deal of financial reward. If your sales guy quits, you hire another one.

Genius on the other hand is rare and worth paying. Although it seems not very much.

For someone to be highly reassessed, it's not just impact that is needed but measurable impact.

A marketing genius may produce a lot of value, but it's intently hard to measure.

Sales is easy to measure, and pay is often based on performance. In fields where sales talent makes a big difference, to salespeople are often quite highly compensated.

The closer to the transaction you are, the bigger your cut.

Inventors get 8% of the "total private return", entrepreneurs get 44% of the "total gain", blue-collar workers get about 26% of the "gains", and the "rest" goes to white-collar workers. Is this just sloppy writing or are the authors making distinctions between the terms? The paper is gated. The abstract does identify the four categories as "inventors and their coworkers or stakeholders within the same firm", so, for example, by the term "inventors" I assume that they are the actual inventors who work for the firms that develop, produce, and market the inventions as opposed to freelance inventors or investors who purchase patents from the actual inventors. Of course, more often than not inventors who work for firms more often than not give up the rights to their inventions as part of the employment relationship. I know an individual who owns the patent for something everyone who uses a car also uses, and every time one uses it the individual collects a royalty. He has collected lots of royalties. Lots. He was not the inventor, but an investor in the patent. His enormous returns would not be taken into account in this study. This isn't a criticism of the paper only an attempt to clarify the meaning of terms and the limits of the findings in the paper.

WOW !!! I am surprised that the inventor gets 8%. I would have guessed much less than that. I know I didn't get 8% average and where I did get a large return it was from utilizing the innovations in my own business (part of the entrepreneur share).

Comments for this post are closed