Gonna just start referring to Russell Westbrook as a rent seeker whenever he does something hoggish.
Coincidentally recently read the part of Book IV in "Progress and Poverty" where George examines this issue. My sense is that these companies are good for much of their lifecycle but eventually need a little pruning.
Why does showing higher productivity in more-concentrated industries support the "superstar" explanation? If marginal productivity is *decreasing* in each input to production, and oligopoly distorts the economy by producing too little of a good, then we should expect more concentrated industries to have higher average productivity. I'm not sure what empirical strategy one could use to tease out these two effects since both imply a causal relationship from higher concentration to higher productivity.
Companies seem to go through a life cycle. The ones that get big almost always had some kind of advantage. Sometimes it was intellectual property like Microsoft or Apple. Sometimes it was a good throwing hand with the dynamite like Standard Oil. Sometimes it was industrial efficiency like MacDonald's or Ford.
The problem is that once a company gets big, it doesn't need to become more efficient. Why compete when you can buy your competitors? Why innovate when you can squash innovation? AT&T, Kodak and IBM were good examples of that. They had a productive edge, but at a certain point it made more sense from a business point of view to focus on market power.
One problem with companies in this state is that they make the economy overall less resilient and less efficient. By the 1960s, Detroit was in its doldrums after decades of innovation, but VW could come in and grab 10% of the market, and the Japanese could come in a bit later and grab even more. To be fair, Detroit is more innovative than it was, but a lot of this has been government pressure on safety, efficiency and so on.
The next stage of the cycle involves politics. The government can just let the company go broke. It can apply pressure to force greater efficiency either through regulation or breakups. Alternatively, it can step back and just prop up the business as it has done with fossil fuels and outfits like Boeing. At this point it's about policy.
There really does seem to be a sweet spot. Just like larger companies tend to be better employers than mom-and-pops but eventually start exploiting labor (via monopsony) and collecting rents instead of getting better at exploiting resources.
Sometimes big trees need pruning if even for their offspring to grow and it's fine. One company doesn't define a sector (or if it does, then that's bad by itself).
What on earth are you referring to with Standard Oil? They got big because they (specifically, John Rockefeller) was so maniacally focused on improving efficiency. During it’s lifespan, it vastly increased production whilst also making the price a fraction of what it was.
I agree. We should design the economic system to encourage more superstar behavior, and stop stagnating companies from poisoning the well for future superstars.
I don't know if there is data on this but it seems to me that the interconnectedness and complexity of modern software technology results in strong network effects for large tech companies. For example, it would require building and battle testing a mind boggling amount software systems to reproduce Amazon.
It would take decades of evolution and selection of millions of lines of code to reach the intricate compatibility across systems and the fit with real world complexity.
I'm pretty sure no one inside Amazon understands all of Amazon's systems and how they connect to each other. There's just too much complexity there for mere humans, even ones with deep inside knowledge, to be able to reproduce. If you want to create this types of system without reusing existing code, there is no choice but to slowly evolve the code base again from scratch. Computer code is the genetic material of modern businesses. This makes it very challenging for most companies, even very large ones to challenge an incumbent.
I once had the job of going around to every employee of a large financial institution and finding out what they did and how it fit into the larger process (I was documenting internal controls andncompliance with banking regulations).
It was almost comical how no one understood how their jobs fit into the overall process, even in management.
This was fairly complex stuff but nothing like what must be going on at Amazon.
The hard part of producing software is figuring out what it should do. Once that's out there, it is surprisingly easy to duplicate. Look at the recent Rosetta protein folding software that came out a few months after Alphafold. Software history is full of such examples e.g. 123 after VisiCalc, Gnu C and Linux after UNIX, Apple Maps after Google Maps, Lyft after Uber. They say that one can tell the pioneers in an industry by the arrows in their backs.
Amazon has some impressive software but also a good supply of mediocre software. They have reasonably good web interface software. (Granted, I'd love to be able to type in the title of a book or movie and have it pop up on the first page of results, but that's probably too much an algorithmic challenge.) They have good logistics code for order fulfillment. They seem to have straightened out some of the problems from 2020. Their warehouse robotics software, however, lags the industry, possibly because warehouse workers are cheap and union busting well understood.
AWS is impressive. Bezos insisted that it be designed as a set of services with straightforward APIs. Amazon has a lot of competition in this area now, also from big companies, but if you are a big outfit like Netflix, AWS can be a good fit.
Yes figuring out what software should do is the hard part but in a massive platform like Amazon there is an incredible amount of subtle details and irreducible complexity embodied in the code that makes up "what software should do". I don't think you can compare that to technological leaps in core abstractions like UNIX and C or the idea of a spreadsheet.
Unix and C were great simplifications of core building blocks found in previously overengineered OS systems. Ken Thompson wrote Unix in assembly in three weeks (as an improvement over Multics, an overly complex, costly, multi year, multi organisation research OS of the time) by getting rid of cruft features, forgoing support for non power-of-two word length CPUs (which were common at the time), removing complex headers for creating files, allowing any program to process any files and making files a very plain thing that you can create in a single command without parameters. Then was added a simple layer of permissions on these files and everything including hardware and drivers were made to use the same simple files and permission interface through the /dev virtual file system so everything including hardware could be handled and composed and securely connected with simple tools like pipes and accessed through the same APIs as files. Once Ken Thompson simplified the core of operating systems to it's most simple composable, securable components, yes that idea spread and was copied by others. But that was an exercise that was more like solving a difficult equation than sorting out and optimizing the infinite amount of details in global retail markets.
Multics was a completely different system. It was trying to solve a hard problem, creating a secure computer system, and it was amazingly successful. It was one of the few computer systems approved to support multiple users at different DoD classification levels. Very few people were sweating attack surfaces in the 1970s. Multics was a test bed for exploring a lot of ideas that are extremely relevant today.
UNIX was about getting a useful subset of the system on a minicomputer, the PDP 9 (18 bit word) and later the PDP 11. UNIX is a good example of how one can rapidly develop software once one knows what it should do, and it wasn't the only system modeled on Multics. As for getting rid of "cruft", the software industry has been hard at work putting a lot of that "cruft" back in for the last nearly 50 years.
UNIX was popular because it was licensed to universities for free, and it was better in many ways than the DOS that DEC offered at the time. In the 1970s, AT&T was still fighting the antitrust forces that would dismantle it a decade later, so the company considered supporting it a good PR move but not directly relevant to its own business.
The only difference between large and small companies that are no longer under the founder is that the large ones is available capitol. Capitol which they use to bribe politicians and the court so that they are never actually held responsible for their malevolence. Self regulation is simply corruption at its height.
I have no business weighing in on a topic about which I am a mere fan boy. I had to look up Justice Brandeis, Neo-Brandeisian and Judge Learned Hand, for instance. And that was just in Noah’s very accessible column. In the replies I was even further out of my depth.
However, I am a fan boy with a huge stake in the debate (my quality and “quantity” of life) and one who tends towards the “…from Column A…Column B” with sliding scales that are based on the size/importance to the economy and, more importantly, humanity’s future.
Like I encouraged my addiction clients to do, I like to try and “think it through!” Do not just define the problem based on the obvious issue at hand (paying your bills). Set about solving the underlying problem: Alcohol and other drug addiction.
Like the addict “You’re concentrating on the wrong item!”
And…corporations might not even be the biggest threat to our future.
Individuals pursuing their own dreams with "billions and billions” of dollars might easily take humanity along with them, headlong into disastrous “Unintended Consequences.” Sound far-fetched? One *young* male was just looking for a way to meet hot chicks, but it snowballed into a platform with effects on humanity that even he could not begin to predict or control.
Amazonian serfs don’t care if Bezos gets to Mars ahead of Musk and Branson. If an alien civilization came along as the pyramids were being built and asked the slaves, “Would you rather take part in the construction of enduring monuments to the Pharaoh, quite possibly getting crushed in the process, or have small plot of land that would feed you and generations of your family?”
“Gee, I don’t know…can I think about it?”
But nobody ever asks the serfs, who are conned and forced by necessity into “voting” with their hard-earned money for individuals and politicians who bought their way to the top with the serf's hard-earned money.
I’m barely scratching the surface of the underlying problem here, but I’m having a cash flow problem, so I need to go to work to help support my great-grands…
Well, many (chiefly Schumpeter, but also Landes I think) argued that it was large companies that invested the most in R&D, thus driving long-term growth.
Imo it's due to an exploding labour supply due to the Soviet Block and China joining the world market at once, which made investing in new technology less profitable given all the cheap labour.
Gonna just start referring to Russell Westbrook as a rent seeker whenever he does something hoggish.
Coincidentally recently read the part of Book IV in "Progress and Poverty" where George examines this issue. My sense is that these companies are good for much of their lifecycle but eventually need a little pruning.
Why does showing higher productivity in more-concentrated industries support the "superstar" explanation? If marginal productivity is *decreasing* in each input to production, and oligopoly distorts the economy by producing too little of a good, then we should expect more concentrated industries to have higher average productivity. I'm not sure what empirical strategy one could use to tease out these two effects since both imply a causal relationship from higher concentration to higher productivity.
Well, you could look at the firm level, as Gutierrez and Philippon do!
Companies seem to go through a life cycle. The ones that get big almost always had some kind of advantage. Sometimes it was intellectual property like Microsoft or Apple. Sometimes it was a good throwing hand with the dynamite like Standard Oil. Sometimes it was industrial efficiency like MacDonald's or Ford.
The problem is that once a company gets big, it doesn't need to become more efficient. Why compete when you can buy your competitors? Why innovate when you can squash innovation? AT&T, Kodak and IBM were good examples of that. They had a productive edge, but at a certain point it made more sense from a business point of view to focus on market power.
One problem with companies in this state is that they make the economy overall less resilient and less efficient. By the 1960s, Detroit was in its doldrums after decades of innovation, but VW could come in and grab 10% of the market, and the Japanese could come in a bit later and grab even more. To be fair, Detroit is more innovative than it was, but a lot of this has been government pressure on safety, efficiency and so on.
The next stage of the cycle involves politics. The government can just let the company go broke. It can apply pressure to force greater efficiency either through regulation or breakups. Alternatively, it can step back and just prop up the business as it has done with fossil fuels and outfits like Boeing. At this point it's about policy.
There really does seem to be a sweet spot. Just like larger companies tend to be better employers than mom-and-pops but eventually start exploiting labor (via monopsony) and collecting rents instead of getting better at exploiting resources.
Sometimes big trees need pruning if even for their offspring to grow and it's fine. One company doesn't define a sector (or if it does, then that's bad by itself).
What on earth are you referring to with Standard Oil? They got big because they (specifically, John Rockefeller) was so maniacally focused on improving efficiency. During it’s lifespan, it vastly increased production whilst also making the price a fraction of what it was.
Yeah, but being better with few sticks of dynamite helped provide an edge.
I agree. We should design the economic system to encourage more superstar behavior, and stop stagnating companies from poisoning the well for future superstars.
I don't know if there is data on this but it seems to me that the interconnectedness and complexity of modern software technology results in strong network effects for large tech companies. For example, it would require building and battle testing a mind boggling amount software systems to reproduce Amazon.
It would take decades of evolution and selection of millions of lines of code to reach the intricate compatibility across systems and the fit with real world complexity.
I'm pretty sure no one inside Amazon understands all of Amazon's systems and how they connect to each other. There's just too much complexity there for mere humans, even ones with deep inside knowledge, to be able to reproduce. If you want to create this types of system without reusing existing code, there is no choice but to slowly evolve the code base again from scratch. Computer code is the genetic material of modern businesses. This makes it very challenging for most companies, even very large ones to challenge an incumbent.
I once had the job of going around to every employee of a large financial institution and finding out what they did and how it fit into the larger process (I was documenting internal controls andncompliance with banking regulations).
It was almost comical how no one understood how their jobs fit into the overall process, even in management.
This was fairly complex stuff but nothing like what must be going on at Amazon.
The hard part of producing software is figuring out what it should do. Once that's out there, it is surprisingly easy to duplicate. Look at the recent Rosetta protein folding software that came out a few months after Alphafold. Software history is full of such examples e.g. 123 after VisiCalc, Gnu C and Linux after UNIX, Apple Maps after Google Maps, Lyft after Uber. They say that one can tell the pioneers in an industry by the arrows in their backs.
Amazon has some impressive software but also a good supply of mediocre software. They have reasonably good web interface software. (Granted, I'd love to be able to type in the title of a book or movie and have it pop up on the first page of results, but that's probably too much an algorithmic challenge.) They have good logistics code for order fulfillment. They seem to have straightened out some of the problems from 2020. Their warehouse robotics software, however, lags the industry, possibly because warehouse workers are cheap and union busting well understood.
AWS is impressive. Bezos insisted that it be designed as a set of services with straightforward APIs. Amazon has a lot of competition in this area now, also from big companies, but if you are a big outfit like Netflix, AWS can be a good fit.
Yes figuring out what software should do is the hard part but in a massive platform like Amazon there is an incredible amount of subtle details and irreducible complexity embodied in the code that makes up "what software should do". I don't think you can compare that to technological leaps in core abstractions like UNIX and C or the idea of a spreadsheet.
Unix and C were great simplifications of core building blocks found in previously overengineered OS systems. Ken Thompson wrote Unix in assembly in three weeks (as an improvement over Multics, an overly complex, costly, multi year, multi organisation research OS of the time) by getting rid of cruft features, forgoing support for non power-of-two word length CPUs (which were common at the time), removing complex headers for creating files, allowing any program to process any files and making files a very plain thing that you can create in a single command without parameters. Then was added a simple layer of permissions on these files and everything including hardware and drivers were made to use the same simple files and permission interface through the /dev virtual file system so everything including hardware could be handled and composed and securely connected with simple tools like pipes and accessed through the same APIs as files. Once Ken Thompson simplified the core of operating systems to it's most simple composable, securable components, yes that idea spread and was copied by others. But that was an exercise that was more like solving a difficult equation than sorting out and optimizing the infinite amount of details in global retail markets.
Multics was a completely different system. It was trying to solve a hard problem, creating a secure computer system, and it was amazingly successful. It was one of the few computer systems approved to support multiple users at different DoD classification levels. Very few people were sweating attack surfaces in the 1970s. Multics was a test bed for exploring a lot of ideas that are extremely relevant today.
UNIX was about getting a useful subset of the system on a minicomputer, the PDP 9 (18 bit word) and later the PDP 11. UNIX is a good example of how one can rapidly develop software once one knows what it should do, and it wasn't the only system modeled on Multics. As for getting rid of "cruft", the software industry has been hard at work putting a lot of that "cruft" back in for the last nearly 50 years.
UNIX was popular because it was licensed to universities for free, and it was better in many ways than the DOS that DEC offered at the time. In the 1970s, AT&T was still fighting the antitrust forces that would dismantle it a decade later, so the company considered supporting it a good PR move but not directly relevant to its own business.
Probably why Amazon makes so damn much from AWS.
The only difference between large and small companies that are no longer under the founder is that the large ones is available capitol. Capitol which they use to bribe politicians and the court so that they are never actually held responsible for their malevolence. Self regulation is simply corruption at its height.
This discussion has been bothering me for days.
I have no business weighing in on a topic about which I am a mere fan boy. I had to look up Justice Brandeis, Neo-Brandeisian and Judge Learned Hand, for instance. And that was just in Noah’s very accessible column. In the replies I was even further out of my depth.
However, I am a fan boy with a huge stake in the debate (my quality and “quantity” of life) and one who tends towards the “…from Column A…Column B” with sliding scales that are based on the size/importance to the economy and, more importantly, humanity’s future.
Like I encouraged my addiction clients to do, I like to try and “think it through!” Do not just define the problem based on the obvious issue at hand (paying your bills). Set about solving the underlying problem: Alcohol and other drug addiction.
Like the addict “You’re concentrating on the wrong item!”
And…corporations might not even be the biggest threat to our future.
Individuals pursuing their own dreams with "billions and billions” of dollars might easily take humanity along with them, headlong into disastrous “Unintended Consequences.” Sound far-fetched? One *young* male was just looking for a way to meet hot chicks, but it snowballed into a platform with effects on humanity that even he could not begin to predict or control.
Amazonian serfs don’t care if Bezos gets to Mars ahead of Musk and Branson. If an alien civilization came along as the pyramids were being built and asked the slaves, “Would you rather take part in the construction of enduring monuments to the Pharaoh, quite possibly getting crushed in the process, or have small plot of land that would feed you and generations of your family?”
“Gee, I don’t know…can I think about it?”
But nobody ever asks the serfs, who are conned and forced by necessity into “voting” with their hard-earned money for individuals and politicians who bought their way to the top with the serf's hard-earned money.
I’m barely scratching the surface of the underlying problem here, but I’m having a cash flow problem, so I need to go to work to help support my great-grands…
Dylan: “Money doesn’t talk, it swears.”
Productivity has no constituency these days
This was all known 45 years ago. Michael Porter.
It's a basic of market economics that the highest market share players make vastly higher cash flow. Experience curve
Well, many (chiefly Schumpeter, but also Landes I think) argued that it was large companies that invested the most in R&D, thus driving long-term growth.
👍
Imo it's due to an exploding labour supply due to the Soviet Block and China joining the world market at once, which made investing in new technology less profitable given all the cheap labour.