alexham said in #5500 23h ago:
It is fashionable to attack financial capitalism. But I think some of this impulse is misplaced. On a financial capitalist vs. industrial policy axis, I lean centre-financial capitalist.
The reason naive financial capitalism is superior to naive industrial policy is because naive financial capitalism has darwinian selection amongst capital allocators / investors. Investors that generate high returns while minimizing risk, compound capital, partially due to retained earnings and partially due to new capital inflows. And investors that lose money, lose their capital. Furthermore, the ability to generate returns is non-stationary. As market regimes shift, old-fashioned investors gradually get replaced by newer entrants that start to outcompete them.
If you look at the biotech industry for example, the richest investors are those who are genuinely skilled in figuring out if a drug's gonna work or not. After a generation of selection we have capital markets that can (relatively) accurately price the probability of success of clinical trials. But we can't take this for granted. Not every capital market can actually allocate hundreds of mil to clinical stage companies efficiently. And it's not the only thing you need either. Here the bottleneck is our ability to generate small n proof of concept data cheaply and fastly, rather than an inability to interpret and invest based on this kind of data. In China it's the other way around. They don't really have the commercial capital allocation motion, but they are rock stars at generating POC data, which is why so many of their assets end up out-licensed to Western pharma. They generate the asset, we pay for the expensive part.
Naive industrial policy however, is not subject to this kind of selection policy. In fact, on the contrary there is an aversion to Darwinian selection. We want to help create jobs, not destroy inefficient industries or companies. There's a tendency to pursue everything-bagel outcomes and a tendency to king-make politically connected champions.
All successful industrial policy implementations inject some sort of cybernetic feedback mechanism to avoid this failure mode. Chinese industrial policy for example lets different states back their respective champions, and then those companies compete for the national market. Israel invested as a LP into VC funds at no more than 30-40 cents per dollar raised from other investors. Japanese and Koreans subsidized firms that could win in export markets.
We, on the other hand, just pay more and more to ever more inefficient shipyards and protect inefficient manufacturers through tariffs in industries such as steel making and automotive manufacturing. (I don't advocate pulling the plug on them, getting the Japanese build their car factories in the south and bring best practices and competition was great industrial policy!)
Of course, financial capitalism selects for the ability to make money which is often misaligned with civilizational and national security goals we may have. If we have a dysfunctional permitting system, other red tape and crazy unions, why not outsource the factory instead? That's easier than lobbying to reduce the cost of doing business, and you get to pay lower salaries as a bonus. And money made in China is as valuable as money made here. Ditto for returns generated via rent extraction vs value creation.
My goal in this note is thus not to advocate for pure libertarian economics necessarily, but to emphasize that if we want efficient capital allocation, we need cybernetic control built into the capital allocation process. This is inherent to finance (stronger in public markets vs private) but not industrial policy. Alignment of social and private interests, on the other hand, is more natural to industrial policy.
The reason naive financial capitalism is superior to naive industrial policy is because naive financial capitalism has darwinian selection amongst capital allocators / investors. Investors that generate high returns while minimizing risk, compound capital, partially due to retained earnings and partially due to new capital inflows. And investors that lose money, lose their capital. Furthermore, the ability to generate returns is non-stationary. As market regimes shift, old-fashioned investors gradually get replaced by newer entrants that start to outcompete them.
If you look at the biotech industry for example, the richest investors are those who are genuinely skilled in figuring out if a drug's gonna work or not. After a generation of selection we have capital markets that can (relatively) accurately price the probability of success of clinical trials. But we can't take this for granted. Not every capital market can actually allocate hundreds of mil to clinical stage companies efficiently. And it's not the only thing you need either. Here the bottleneck is our ability to generate small n proof of concept data cheaply and fastly, rather than an inability to interpret and invest based on this kind of data. In China it's the other way around. They don't really have the commercial capital allocation motion, but they are rock stars at generating POC data, which is why so many of their assets end up out-licensed to Western pharma. They generate the asset, we pay for the expensive part.
Naive industrial policy however, is not subject to this kind of selection policy. In fact, on the contrary there is an aversion to Darwinian selection. We want to help create jobs, not destroy inefficient industries or companies. There's a tendency to pursue everything-bagel outcomes and a tendency to king-make politically connected champions.
All successful industrial policy implementations inject some sort of cybernetic feedback mechanism to avoid this failure mode. Chinese industrial policy for example lets different states back their respective champions, and then those companies compete for the national market. Israel invested as a LP into VC funds at no more than 30-40 cents per dollar raised from other investors. Japanese and Koreans subsidized firms that could win in export markets.
We, on the other hand, just pay more and more to ever more inefficient shipyards and protect inefficient manufacturers through tariffs in industries such as steel making and automotive manufacturing. (I don't advocate pulling the plug on them, getting the Japanese build their car factories in the south and bring best practices and competition was great industrial policy!)
Of course, financial capitalism selects for the ability to make money which is often misaligned with civilizational and national security goals we may have. If we have a dysfunctional permitting system, other red tape and crazy unions, why not outsource the factory instead? That's easier than lobbying to reduce the cost of doing business, and you get to pay lower salaries as a bonus. And money made in China is as valuable as money made here. Ditto for returns generated via rent extraction vs value creation.
My goal in this note is thus not to advocate for pure libertarian economics necessarily, but to emphasize that if we want efficient capital allocation, we need cybernetic control built into the capital allocation process. This is inherent to finance (stronger in public markets vs private) but not industrial policy. Alignment of social and private interests, on the other hand, is more natural to industrial policy.
It is fashionable to