can this this explain the very high rate of innovation across both Israel's and Japan's industrial base (please put Gaza/Palestine aside for a moment) - does this sort of not too much not too little monopoly theory holds up in Israel ....Lots of small firms - strong social safety net - high rate of income inequality. But japan has giant companies lots of regulation little social safety net.. curious
I learned in my aerodynamics classes that if you assume no viscosity (friction), airplane wings produce no lift. The sharp trailing edge produces the friction necessary to sustain the pressure differential between the top and bottom surfaces. A barrier to entry, if you will.
This article is fine. But there's something of a human element missing: humans like me, who love inventing stuff and simply do it as a matter of course, whenever they're allowed to.
It's never been something any of my employers were interested in. They always had specific software they wanted done, specific capabilities and features. They paid me to do those and, consequently, I had little time to do the broad inventions that I was actually enthusiastic about, the foundational things that a wide variety of other software could have built on top of. I've built a lot of free shit in my career, such as the MilliKeys keyboard for PalmPilot, the Enhanced C# preprocessor, and Loyc trees. I could've built so much more. For example, in 2021 I made a really serialization system for C# called SyncLib in 2019, then rebuilt it from scratch (but much bigger and better) in 2021 after my old employer wouldn't let me have the source code of the original. So I asked my new employer for a 4-day work week and 20% reduced pay so I could work on it, and he agreed ― until the cofounder left and I had to do all the supposedly important work that makes the money (it's not actually important and the company need not exist, but I expect to make my first million this way). To this day SyncLib is sitting on a branch on GitHub, waiting for the day I have time between childcare, doomscrolling, and regular work to put the finishing touches on it and give it to the world for free.
There's probably no empirical study you can find on the economic effects of "open engineering", i.e. funding independent inventors and innovators for their foundational engineering work, much as we fund basic science. If no one funds it, who's gonna study it? There are companies on the one hand, who want to build specific things, and there's "basic science" on the other. In both environments there are mavericks ― people at companies who push for general-purpose open-source products that benefit the whole ecosystem instead of just their employer, and other people at universities, sacrificing time they're "supposed" to spend writing papers and doing administrative work, to produce a software package, reusable engineering document, or a polished web site with demos that attract people to build upon their work.
For physical goods, innovation probably happens "enough" through the regular VC and small business pipeline. But for software and other informational engineering work, the whole value proposition comes from the work product being available to use free. If you insist on charging one dollar for open source, most of its value evaporates instantly. If you want more innovation in this part of the economy, I think it's always been as simple as throwing a financial bone to the people who want to do the work. If we can fund basic science, why not basic engineering?
Today's AI is great. I use it every day. But I think it's ultimately going to destroy the world as we know it. So many people, championing the race to AGI, claim we need AGI because there aren't enough humans available to do the necessary innovation. I call bullshit. BULL. SHIT.
The innovation that Aghion et al address is almost entirely micro-innovation because that is what directed R&D is really good for. Macro-innovation is almost always from curiosity-driven research, the kind that scarcely any companies engage in, or can afford to. Mokyr knows this and I suspect that Aghion does too but so far at least it has been very resistant to any sort of endogenous modeling.
Trump and his lackey's are of course doing all that they can to strangle curiosity-driven research. They won't live to pay the price but their (and our) children and grandchildren will have reason to curse their memory.
I don’t get all the love for Aghion and Howitt; I think these kinds of uber macro models are all that is wrong about the profession. The issue with their paper isn’t so much that there are assumptions, because assumptions are unavoidable, it’s that they assume everything away and all of their assumptions are empirically incorrect.
Just to name a few. The assumption of frictionless labor reallocation fails in the face of data on long unemployment spells and regional dislocation following technological shocks (Autor, Dorn & Hanson 2013). The model’s memoryless Poisson innovation process ignores the strong path-dependence and cumulative knowledge effects observed in patent and R&D studies (Jaffe et al. 1993; Hall et al. 2001), while its claim that new technologies instantly render old ones obsolete is contradicted by evidence of long technological coexistence (Bresnahan & Trajtenberg 1995). Nor do innovators reliably enjoy temporary monopolies—real markets feature imitation, licensing, and network effects that quickly erode rents. Perfect foresight and exogenous R&D productivity bear little resemblance to the uncertainty, strategic learning, and institutional variation documented in innovation research (Nelson & Winter 1982; Bloom et al. 2019).
And we haven’t even gotten to the representative agent, the most reviled of all macro models, for which we have an entire sub-discipline called behavioral economics that picks apart every assumption. People don’t behave in this way, not by themselves, and not when aggregated up.
Assumptions are supposed to abstract away the superfluous details. But in the case of DSGE modeling, really most macro modeling, they exist to get a closed-form solution, not to actually show us anything. The real insight comes from Schumpeter; everything else can be thrown in the waste basket.
Given the prospect of exponential economic and technological growth, one can always arbitrarily discount the present in favor of a potentially exponentially better future. However, America fell into fascism first during this particular technological epoch, so I would hold off on popping the champagne about how American creative destruction and monopoly power is good in the long term at least until we figure out if we are going to have free elections in the future.
It is true that European companies are not behemoths like US ones but that is only because Europe and the USA have open capital flows and practically free trade. Europe has excellent Mathematics and Physics programs which basically underlie all of modern technology and could theoretically support a robust tech sector, but because America is willing to bear the costs of discounting the present wellbeing of its population, in the form of higher drug prices etc, European capital flows to the USA while Europe itself is content with exogenous technological growth flowing to it from a paternalistic ally. Of course given that the US is now descending into a techno-fascist state, Europe has some tough decisions ahead and the old model is no longer tenable.
"I would hold off on popping the champagne about how American creative destruction and monopoly power is good"
I am not an economist, but Elon's turn to the Dark Side has certainly soured me on the idea that tech bro visionaries amassing untold wealth are a net positive for society. See also: Andreessen, Benioff, various billionaires paying tribute to Trump and kissing his backside. Elon is tied with Vance for "most disappointing Dark Side turn by a person I used to like/admire." Trump doesn't count, because I never admired him in the first place.
Exactly, you can justify any amount of arbitrary cruelty and death in the short term if you convince yourself that in the long term you are saving humanity by spreading your sperm across the whole galaxy via exponential growth.
This was an excellent piece. One of my favorites in a long time. Count me in for more theory and policy analysis and less factional fighting (even as the latter is more engaging in some sense, and probably does bigger numbers)!
The thesis here is very compelling. I firmly believe that in a perfectly competitive market there would be far less innovation. But I'm curious about what *kind* of monopoly best suits increased innovation. Bell Labs was one of the most important contributors to American economic development in the 20th century.* Bell had a virtually complete monopoly over most telecommunications for most of the century (i.e., not just a "little bit" of monopoly). But it was also highly regulated by the government. In terms of just furthering the cause of innovation, that really seemed to hit the sweet spot. Is that kind of corporate/government cooperation something that should be considered again?
* The book "The Idea Factory" on Bell Labs is fantastic. It makes it clear that Bell Lab's importance went far beyond inventing the transistor and discovering the Big Bang. Its work, especially in the 1920s, was crucial for inventing and propagating all the things that made a nation-wide telecommunications network feasible (e.g., how to propagate calls over long distances without the signal degrading, the dial tone, how to enter phone numbers and on and on -- a million things we take for granted).
<rant>I'm just coming here to complain about Substack for a second. Please PLEASE make the footnotes linked in the email, not just on substack.com. I subscribe to what, 5-10 newsletters, some of whom (JVL's Triad) contain dozens of footnotes.
I read Matt's post in it's entirety, then either have to scroll back up and find the superscript number then back down to read the footnote and associate the two or have to click the link to the article on substack.com, scroll down to the footnotes, click the footnote number, that takes me back up to where it's noted and I can hover over to read the footnote. </rant>
A tech giant funded the climate tech research lab where I did my PhD. The school itself, a leader in tech, was founded by a wealthy monopolist. Much of the tech that drives the modern world came out of Bell Labs, Google, and other quasi-monopolies that could fund globally competitive research programs with the spare change in their couch cushions.
I’m old enough to remember the 1970’s and 80’s and how foreign competition disrupted “healthy” US markets in cars and electronics. There was certainly competition among the big US automakers at the time, for example, but there was still sclerosis, poor quality products, and other negative aspects that allowed for more competition.
One thing not mentioned in the piece are cartels, which also negatively impact actual competition.
Good article and a good argument for subscribing! Thank you!
can this this explain the very high rate of innovation across both Israel's and Japan's industrial base (please put Gaza/Palestine aside for a moment) - does this sort of not too much not too little monopoly theory holds up in Israel ....Lots of small firms - strong social safety net - high rate of income inequality. But japan has giant companies lots of regulation little social safety net.. curious
I learned in my aerodynamics classes that if you assume no viscosity (friction), airplane wings produce no lift. The sharp trailing edge produces the friction necessary to sustain the pressure differential between the top and bottom surfaces. A barrier to entry, if you will.
This article is fine. But there's something of a human element missing: humans like me, who love inventing stuff and simply do it as a matter of course, whenever they're allowed to.
It's never been something any of my employers were interested in. They always had specific software they wanted done, specific capabilities and features. They paid me to do those and, consequently, I had little time to do the broad inventions that I was actually enthusiastic about, the foundational things that a wide variety of other software could have built on top of. I've built a lot of free shit in my career, such as the MilliKeys keyboard for PalmPilot, the Enhanced C# preprocessor, and Loyc trees. I could've built so much more. For example, in 2021 I made a really serialization system for C# called SyncLib in 2019, then rebuilt it from scratch (but much bigger and better) in 2021 after my old employer wouldn't let me have the source code of the original. So I asked my new employer for a 4-day work week and 20% reduced pay so I could work on it, and he agreed ― until the cofounder left and I had to do all the supposedly important work that makes the money (it's not actually important and the company need not exist, but I expect to make my first million this way). To this day SyncLib is sitting on a branch on GitHub, waiting for the day I have time between childcare, doomscrolling, and regular work to put the finishing touches on it and give it to the world for free.
There's probably no empirical study you can find on the economic effects of "open engineering", i.e. funding independent inventors and innovators for their foundational engineering work, much as we fund basic science. If no one funds it, who's gonna study it? There are companies on the one hand, who want to build specific things, and there's "basic science" on the other. In both environments there are mavericks ― people at companies who push for general-purpose open-source products that benefit the whole ecosystem instead of just their employer, and other people at universities, sacrificing time they're "supposed" to spend writing papers and doing administrative work, to produce a software package, reusable engineering document, or a polished web site with demos that attract people to build upon their work.
For physical goods, innovation probably happens "enough" through the regular VC and small business pipeline. But for software and other informational engineering work, the whole value proposition comes from the work product being available to use free. If you insist on charging one dollar for open source, most of its value evaporates instantly. If you want more innovation in this part of the economy, I think it's always been as simple as throwing a financial bone to the people who want to do the work. If we can fund basic science, why not basic engineering?
Today's AI is great. I use it every day. But I think it's ultimately going to destroy the world as we know it. So many people, championing the race to AGI, claim we need AGI because there aren't enough humans available to do the necessary innovation. I call bullshit. BULL. SHIT.
Dumb question: are Americans more than 15% wealthier, making the 15% price increase irrelevant?
The innovation that Aghion et al address is almost entirely micro-innovation because that is what directed R&D is really good for. Macro-innovation is almost always from curiosity-driven research, the kind that scarcely any companies engage in, or can afford to. Mokyr knows this and I suspect that Aghion does too but so far at least it has been very resistant to any sort of endogenous modeling.
Trump and his lackey's are of course doing all that they can to strangle curiosity-driven research. They won't live to pay the price but their (and our) children and grandchildren will have reason to curse their memory.
I don’t get all the love for Aghion and Howitt; I think these kinds of uber macro models are all that is wrong about the profession. The issue with their paper isn’t so much that there are assumptions, because assumptions are unavoidable, it’s that they assume everything away and all of their assumptions are empirically incorrect.
Just to name a few. The assumption of frictionless labor reallocation fails in the face of data on long unemployment spells and regional dislocation following technological shocks (Autor, Dorn & Hanson 2013). The model’s memoryless Poisson innovation process ignores the strong path-dependence and cumulative knowledge effects observed in patent and R&D studies (Jaffe et al. 1993; Hall et al. 2001), while its claim that new technologies instantly render old ones obsolete is contradicted by evidence of long technological coexistence (Bresnahan & Trajtenberg 1995). Nor do innovators reliably enjoy temporary monopolies—real markets feature imitation, licensing, and network effects that quickly erode rents. Perfect foresight and exogenous R&D productivity bear little resemblance to the uncertainty, strategic learning, and institutional variation documented in innovation research (Nelson & Winter 1982; Bloom et al. 2019).
And we haven’t even gotten to the representative agent, the most reviled of all macro models, for which we have an entire sub-discipline called behavioral economics that picks apart every assumption. People don’t behave in this way, not by themselves, and not when aggregated up.
Assumptions are supposed to abstract away the superfluous details. But in the case of DSGE modeling, really most macro modeling, they exist to get a closed-form solution, not to actually show us anything. The real insight comes from Schumpeter; everything else can be thrown in the waste basket.
A little bit of monopoly, as a treat.
Here's an example of a near monopoly that was very innovative, ATT and Bell Labs say up to deregulation (late 70's early 80's).
Given the prospect of exponential economic and technological growth, one can always arbitrarily discount the present in favor of a potentially exponentially better future. However, America fell into fascism first during this particular technological epoch, so I would hold off on popping the champagne about how American creative destruction and monopoly power is good in the long term at least until we figure out if we are going to have free elections in the future.
It is true that European companies are not behemoths like US ones but that is only because Europe and the USA have open capital flows and practically free trade. Europe has excellent Mathematics and Physics programs which basically underlie all of modern technology and could theoretically support a robust tech sector, but because America is willing to bear the costs of discounting the present wellbeing of its population, in the form of higher drug prices etc, European capital flows to the USA while Europe itself is content with exogenous technological growth flowing to it from a paternalistic ally. Of course given that the US is now descending into a techno-fascist state, Europe has some tough decisions ahead and the old model is no longer tenable.
"I would hold off on popping the champagne about how American creative destruction and monopoly power is good"
I am not an economist, but Elon's turn to the Dark Side has certainly soured me on the idea that tech bro visionaries amassing untold wealth are a net positive for society. See also: Andreessen, Benioff, various billionaires paying tribute to Trump and kissing his backside. Elon is tied with Vance for "most disappointing Dark Side turn by a person I used to like/admire." Trump doesn't count, because I never admired him in the first place.
Exactly, you can justify any amount of arbitrary cruelty and death in the short term if you convince yourself that in the long term you are saving humanity by spreading your sperm across the whole galaxy via exponential growth.
This was an excellent piece. One of my favorites in a long time. Count me in for more theory and policy analysis and less factional fighting (even as the latter is more engaging in some sense, and probably does bigger numbers)!
Loved this write-up!
The thesis here is very compelling. I firmly believe that in a perfectly competitive market there would be far less innovation. But I'm curious about what *kind* of monopoly best suits increased innovation. Bell Labs was one of the most important contributors to American economic development in the 20th century.* Bell had a virtually complete monopoly over most telecommunications for most of the century (i.e., not just a "little bit" of monopoly). But it was also highly regulated by the government. In terms of just furthering the cause of innovation, that really seemed to hit the sweet spot. Is that kind of corporate/government cooperation something that should be considered again?
* The book "The Idea Factory" on Bell Labs is fantastic. It makes it clear that Bell Lab's importance went far beyond inventing the transistor and discovering the Big Bang. Its work, especially in the 1920s, was crucial for inventing and propagating all the things that made a nation-wide telecommunications network feasible (e.g., how to propagate calls over long distances without the signal degrading, the dial tone, how to enter phone numbers and on and on -- a million things we take for granted).
Enthusiastic second for The Idea Factory! Fantastic book.
<rant>I'm just coming here to complain about Substack for a second. Please PLEASE make the footnotes linked in the email, not just on substack.com. I subscribe to what, 5-10 newsletters, some of whom (JVL's Triad) contain dozens of footnotes.
I read Matt's post in it's entirety, then either have to scroll back up and find the superscript number then back down to read the footnote and associate the two or have to click the link to the article on substack.com, scroll down to the footnotes, click the footnote number, that takes me back up to where it's noted and I can hover over to read the footnote. </rant>
A tech giant funded the climate tech research lab where I did my PhD. The school itself, a leader in tech, was founded by a wealthy monopolist. Much of the tech that drives the modern world came out of Bell Labs, Google, and other quasi-monopolies that could fund globally competitive research programs with the spare change in their couch cushions.
I’m old enough to remember the 1970’s and 80’s and how foreign competition disrupted “healthy” US markets in cars and electronics. There was certainly competition among the big US automakers at the time, for example, but there was still sclerosis, poor quality products, and other negative aspects that allowed for more competition.
One thing not mentioned in the piece are cartels, which also negatively impact actual competition.