mitigatedchaos replied to your post: mitigatedchaos replied to your post: …
(also admittedly I am really tired right now)
thinking in equilibrium is hard and I’m not that great at it, either, but the additional insight I want to deliver is that technological innovation and capital accumulation responds to prices
Google invested in the driverless cars because it can create profits in the captured markets. That’s a Schumpeterian process: capital and labor for technological innovation is scarce, so rational firms allocate capital and labor where the captured market can deliver the greatest profits.
By necessity, that doesn’t happen “all at once”, as you suggested earlier. There are only so many software and intelligence engineers to go around. Google had to pay them dearly:
For the past year, Google’s car project has been a talent sieve, thanks to leadership changes, strategy doubts, new startup dreams and rivals luring self-driving technology experts. Another force pushing people out? Money. A lot of it.
[….]
A large multiplier was applied to the compensation packages in late 2015, resulting in multi-million dollar payments in some cases, according to the people familiar with the situation. One member of the team had a multiplier of 16 applied to bonuses and equity amassed over four years, one of the people said. They asked not to be identified talking about private matters.
[…]
It’s unclear how much the payouts cost Alphabet, however, Chief Financial Officer Ruth Porat talked about it during an earnings conference call with analysts in early 2016.
Operating expenses in the fourth quarter of 2015 rose 14 percent to $6.6 billion, “primarily driven by R&D expense, particularly affected by expenses resulting from project milestones in Other Bets established several years ago,” Porat said, according to a transcript. The CFO wasn’t specific, but one of the people familiar with the situation said the comments referred to the car project compensation.
Google is the most powerful firm in the world, but Google does not have limitless resources, and those resources are prudent enough to command their marginal product. Google already had an endowment here: detailed, comprehensive, machine-readable routes for every navigable roadway in the world; alongside efficient photograph recognition software.
Google, and competing firms, can make a coordinated push where the technology is right, but there are real resource constraints that limit the ‘revolutionary’ impacts of their technology. Technology doesn’t just develop itself, especially capital- and labor-intensive technologies like this one. Not yet, anyways. And not anytime soon.
Google isn’t making surgeonless-surgeries or builderless-buildings or teacherless-teaching, because there aren’t the profits in those markets to justify the costs of the technology and capital investments needed to undercut market incumbents – Google doesn’t have detailed, comprehensive, machine-readable databases of surgeries, building plans, or problem sets which it can process at nominal cost – which suggests that dramatic innovations in those markets aren’t just around the corner, either.
It takes work to do these things. It’s not going to happen all at once unless these are perfect general purpose technologies, with trivial adaptation and marketization costs. I don’t think that’s what they are.
The other constraint is that Schumpeterian innovation profits depend on market prices and incomes. Acemoglu makes them straightforwardly reflect factor prices: falling relative wages induce labor-augmenting demand, falling relative rents induce capital-augmenting demand.
But if you’re working from a model where capital elastically substitutes for labor, I think the equilibrium conditions look different. If your technology is so impressive that it materially reduces labor incomes by substituting for labor, then it will reduce the profits from the goods and services your technology produces.
Less demand for driverless cars means less people can effectively demand drives, right? That really cuts into your bottom line, doesn’t it? Why would you invest in capital-augmenting technology if the relative returns to that capital-augmenting technology are so low? Why would you develop a driverless car if no one could afford to drive it?
Even if you have dramatically different understandings of what the constraints are, I think you have to work out the equilibrium conditions, the individual, marginal choices they emerge from. Why would individuals and firms keep doing the things that get you to that outcome?
I’m not actually entirely averse to thinking of all production as a product of labor. It’s in search of the equilibrium in a closed system with labor that I devised a heuristical model I am adapting for the OTV game concept, because I wanted to figure out prices for a setting in the transhuman space future.
Of course, not all goods respond nice and linearly to labor (especially WRT time), e.g. land vs waitressing, so it doesn’t necessarily hold that, if differences in productivity between workers are sufficient, the price for someone’s labor cannot fall below the minimum in terms of absolute resources that they need to survive. Indeed, this already happens for sufficiently low-productivity workers, workers during famines, etc.
Anyhow,
- If my technology is so impressive that it reduces labor incomes, that doesn’t matter unless I’m the majority employer within the system. I can pull the money from someone else hiring these people, at least for a while.
- If I’m the CEO of most companies, I don’t get paid based on the conditions in ten years, I get paid on a much shorter horizon.
- And unless I did, it’s a problem for the Commons, not me.
- And it doesn’t matter if it’s incredibly profitable - it matters whether my competitors will pursue it. If my competitors make less profit per transaction but undercut my prices by 10%, soon I will be making no profit per transaction. (Although it’s true that at some point the correct decision becomes to exit the industry.)
But maybe I’m missing something here, since my own econ education didn’t go all that far.
