Maybe our economic policy should go deeper than 101-level economics then! Because comparative advantage is a dynamic quantity which changes over time, and while some advantages (like geography) are fixed, others are built by investment.
Here's a video [1] which explains why, in 1955, manufacturing household goods was cheaper to make in the US than in China (and why, at the time, they thought this manufacturing dominance was the thing that backed the US position as a global superpower). It's not because Americans worked more cheaply than Chinese workers, it's because American factories had a well-developed tool-and-die expertise, which meant that when anyone in the world wanted to make something, they were well-advised to travel the US to get it made.
Econ 101's comparative advantage is true at an instantaneous point in time, which is a good start, but if perhaps it's just "knowing enough to be dangerous". Economic policymakers (and company leaders) would do well to think about comparative advantage as planning an optimal trajectory over time, which can mean sacrificing a short-term optimum in exchange for a long-term optimum, and if there even is a textbook solution for that, it's going to look less like a 101-level intersection of straight lines, and more like an iterative optimization over nonlinear differential equations.
> it's because American factories had a well-developed tool-and-die expertise, which meant that when anyone in the world wanted to make something, they were well-advised to travel the US to get it made.
Also because you couldn't offshore production to China or most other places even if you could provide all that due to various geopolitical, economic, social, institutional and other reasons.
mostly container shipping didn't exist, but things like tool and die products cost enough per kilogram that even air shipping is economical, to say nothing of integrated circuits
I think you're missing the point of what I'm saying. The US has steadily moved away from those past competencies because there was more profit to be made elsewhere.
And, sure, there are absolutely network effects with related goods and industries that have steepened that movement. If it was a win to change the allocation of resources when e.g. steelmaking was strong in the US, it's even more of a win after steelmaking withered.
> and if there even is a textbook solution for that,
It's not quite what you're saying, but the closest work I have read is 'Dynamic Optimization: The calculus of variations and optimal control in economics and management' by Kamien et al. It is all about estimating gradients and plotting trajectories in dynamical economic systems.
The feeling of misunderstanding is mutual! I agree that there was more profit to be made elsewhere. But I'm arguing that those profits were short-term profits which may well have come at long-term expense. If you follow the local gradient of profitability, you'll always find great short-term returns selling off your seed corn. Unlike what Econ 101 asserts about maximizing comparative advantage being the most profitable strategy, there is absolutely no guarantee that following a locally-optimal comparative-advantage strategy is globally optimal over a long-term window, where advantages are path-dependent.
Manufacturing is the core example of path-dependent advantages, because (unlike what any econ 101 textbook teaches), marginal costs decline with increasing production quantity in the manufacturing sector. This means the more you make, the better you are at making more things!
Fun fact: You should sell your seed corn, because the best hybrid seeds, crossed from especially made inbreds that you'd never want to use for yield, are so much better than the second generation crossing that you'll always lose money replanting.
There is never any guarantee that profits are long term or short term, or that your manufacturing specialization is going to remain useful, instead of being a dead end. Retaining specialization on, say, cathod tubes wasn't exactly profitable. See all the camera manufacturers that zigged when they should have zagged, and used their manufacturing strength to unprofitability. All of this is hidden by talking about 'manufacturing' in very large terms, but the real world doesn't work like that. Specifically, semiconductors were a very nice place to keep expertise in, and paid off. Internal combustion engines, and filaments for incandescent lighbulbs probably not.
Even in cases where we are looking at the same kind of manufacturing in multiple places, competitive advantages are lost. There are parts of Europe taht still have metallurgy and never attempted to divest, but lost comparative advantages because better technology came in at the wrong time in the capital depreciation curve: They invested heavily at the slightly wrong time, still had expensive labor, so they became far less competitive, at least for a while. Did they not pray enough to the manufacturing god? Did the Netherlands get lucky, or was sufficient dedication to manufacturing that led them to have ASML in their borders? Is the fact that Novo Nordisk found the most important pharmaceutical in the world a matter of Danish superior industrial policy, or did they just get lucky compared to the many other places with large investments in pharma that didn't get anywhere near that lucky?
The path dependence is not so predictable, and the path that makes you better today can lead you down a cliff. It's all gambles, and whoever claims they can predict what is the right one in the long run is being overconfident
ASML's success is partly (gross simplification) because it was the biggest local tech company (Philips) realising that they're too big to be effective, so they made a startup-esque new company which allowed them to be lean and engineering-focused. It's a good story of proper accounting allowing good company structures to persist inside of a bloated company.
Van den Brink gave a great interview some time ago, I'll see if I can translate it and post it here.
Pharma has plenty of examples of things like that. Discovery is made, nobody thinks it’s worth pursuing, the world and knowledge base changes, someone goes back and says “this deserves another look”.
A lot of the GLP-1 success is based on progress made in the diabetes space (incidentally so), the refinement of molecules and better understanding of how obesity could be treated.
>Manufacturing is the core example of path-dependent advantages, because (unlike what any econ 101 textbook teaches), marginal costs decline with increasing production quantity in the manufacturing sector. This means the more you make, the better you are at making more things!
Looking at how Apple said it would be impossible to get US chips, so much this. It needed a lot of investment to onshore chip production again. And we should onshore more high value manufacturing to keep the supply chain working in one place.
The EU has been better at keeping manufacturing competence, but I see a lot of these short term comparative advantage econ 101 ideas taking over in the EU as well.
The thing is there are also network effects, expertise building and marginal cost improvements to be built up in high value items and services.
The United States was able to build a tremendous economy by building up those systems while continuing to benefit from its older manufacturing base for decades.
The United States economy is far from perfect, but it hasn’t traded away a long-term asset for only short-term ones as you’re suggesting.
That last statement is absolutely true, but if you have a constrained domestic supply chain, high employee cost, and/or constrained margins on finished products, you're still going to have come out behind if you persist with domestic manufacturing rather than offshore. This is the calculus OEMs faced in the 1990s-2000s. The big bet that they all made is to assume relatively stable geopolitics, and that there wouldn't ultimately be a squeeze on the potential manufacturing constraints (labor, supply chain, capacity). Ultimately, it's proven to have been by far the smartest decision for high-vol / low-mix stuff: electronic components and consumer electronics (not to mention apparel and many industrial products).
Like I said in my previous comment, though, this doesn't mean the capability to build has left the US (or Europe). Just that the decision to continue investing in manufacturing things that aren't competitively profitable has been made and the capacity has been allocated to higher margin manufacturing (regulated industries, complex products, and products where customers are less price sensitive).
> The big bet that they all made is to assume relatively stable geopolitics
It's not just geopolitics. China required partnering with local companies and sharing IP. Even if they were geopolitically friendly, Western countries set up to build their own Chinese competition from scratch in exchange for lower labour costs, believing that either they could out-innovate China at design (even when Americans no longer understand how their own products get made), or that they'd be retired by the time it did matter.
> . Unlike what Econ 101 asserts about maximizing comparative advantage being the most profitable strategy, there is absolutely no guarantee that following a locally-optimal comparative-advantage strategy is globally optimal over a long-term window,
I think there's little doubt that our change in allocation of resources has been advantageous versus staying an economy focused on primary metals and relatively simple manufacturing. Do you really feel otherwise?
Of course, economic assessments and the behavior of markets generally assumes free choice by participants. So there's always:
1. Geopolitical risks: state leverage can turn a local absolute advantage in e.g. producing war materiel into other advantages.
2. Sure, we could back ourselves into a corner, ultimately, by not being able to provide a key part of the value chain by following that gradient. (Geopolitics can be related, too, in that states can gather together lots of small advantages and use them in coordinated ways against other states).
So our state, of course, needs to focus on countering those actions of other parties. And maintaining some diversity beyond what is economically optimal can add resilience.
(One point I make in class: our textbook pretty much says that price controls are always dumb... but that there are plenty of reasons that a country might desire to have a surplus of food or to not be dependent upon another country).
The track record of those who would seek to centrally plan and optimize for some future outcome instead of following that profit gradient has been very poor. Not to say that it's never worked: but generally following the profit gradient has yielded better outcomes.
> because of (unlike what any econ 101 textbook teaches), marginal costs decline with production quantity in the manufacturing sector.
Unlike what any econ 101 teaches? Talking about LRATC, returns to scale, etc, is a big part of my unit 3. If you're not referring to that and instead e.g. Wright's law, that too is mentioned.
I wasn't talking about "base metals and relatively simple manufacturing". Were you? When Tim Cook explained, in 2017, why the iphone had to be made in China, he explained that it's because China dominates advanced manufacturing, and has skill that cannot be replicated elsewhere.
The behavior of markets assumes free choices by participants that rewards the participants who make those choices. I do not dispute that the CEOs who were responsible for shipping supply chains to China were following their incentives, and it worked out well for them. I would argue that there are alterations to regulations on corporate governance which would increase long-term profitability of those corporations overall, but that the key people in the corporations aren't properly incentivized to pass them, nor are shareholders sufficiently informed or coordinated.
> Talking about LRATC, returns to scale, etc, is a big part of my unit 3
In your unit 3, do you draw LRATC curve as a parabola? Because that's the wrong shape for manufactured goods. Not only do average costs decrease, so do marginal costs, and this is monotonic over all but the shortest timescales. Wright's law is about half of the reason, yes.
> I wasn't talking about "base metals and relatively simple manufacturing". Were you?
A whole lot of the decline that we're talking about has been in those sectors. Microchips and aerospace grew; simple consumer goods and steel manufacturing fell through the floor.
> The behavior of markets assumes free choices by participants t...
Incentives can be, and often are misaligned. However, the context of our discussion is talking about large overall economic growth that has outpaced manufacturing growth, even though it is still positive. This isn't evidence of misaligned incentives.
> In your unit 3, do you draw LRATC curve as a parabola? Because that's the wrong shape for manufactured goods.
It's absolutely a bathtub.
It's steep-downward sloping, mostly flat for a loooooonnnggg time, and then upward sloping. Indeed, this understanding of the shape of LRATC originally comes from study of manufactured goods. At some point coordination gets hard and further increases in quantity require using resources that are not well suited for the task.
Of course, the quantity at which costs slope upwards may be at an impractically large quantity for any industry-- in which case that industry is likely to be a natural monopoly. And there are some recent arguments that coordination is easier thanks to information technology and that it is even harder to reach diseconomies of scale.
I endeavour to convince you that you are teaching your students a falsehood. Natural resource industries have bathtub-shaped average costs. Average costs fall strictly monotonically for manufacturing, and marginal costs either fall or remain constant. Constant marginal costs are what you get if you don't even bother to solve coordination problems, and just copy-and-paste your whole assembly line instead (except even then you can't help but gain economies of scale, if only from your tooling suppliers). The misconception that it's a bathtub does not come from the study of manufactured goods, it comes from thought experiments about manufactured goods done by people who never managed quote requests at a real factory. Empirical studies done on actual firms almost never show rising marginal costs at any quantity.
That this error has permeated introductory economics is a very, very big problem.
> Constant marginal costs are what you get if you don't even bother to solve coordination problems
You still have coordination problems on the supply and distribution side.
> Average costs fall strictly monotonically for manufacturing
This is an extraordinary claim that is easy to refute with simple thought experiments. e.g. You think that if I want 103% of the units that a set of equipment from ASML can deliver, that average costs will be lower than producing 100%? Or do you mean "strictly monotonically" in some other sense?
Being able to vary your capital in the long run doesn't mean that you can have 10.3 sets of photolithography apparatus.
> and just copy-and-paste your whole assembly line instead
If you copy and paste and have everything truly independent, without the need for any coordination of resources, what you effectively have is multiple firms. In practice, firms still need to allocate scarce resources among lines.
> The misconception that it's a bathtub does not come from the study of manufactured goods , it comes from thought experiments about manufactured goods
This is a falsehood. Bain conducted reams of real-world research on manufacturing, plant size, firm size, and returns to scale, and this informs today's idea of LRAC. Of course, this research is 70 years old, and recent data is more ambiguous. As I've said, some believe that information technology has changed everything.
I’ve been interested in Henry Charles Carey, the chief economic advisor to Abraham Lincoln. He wrote a book called the “Harmony of Interests” about the need for state policy + markets (to contrast with purely free markets). Lots of data and rigorous argumentation.
Apparently this was known as the “American School” of economics — and it dominated from the mid-late 1800s for over a century.
it seems that you both agree that it depends on the timescale; asml's next model of machine may be able to produce 10% more, or 91% less, and in either case that extra 3% of your demand will lower the average costs
> and marginal costs either fall or remain constant
Not true. If your factory can make N widgets per year, and you want to make N+1 widgets, the marginal cost of the N+1th widget is vastly greater than the Nth widget.
I think the parent comment was talking about building factories and amortizing their costs over your unit production, whereas you assume the factory is a fixed cost with fixed capacity and looking at the marginal cost to produce a unit, which is really rare in many real world situations.
For most goods, the factory doesn't run anywhere near N, and the fixed costs are 6 or more orders of magnitude higher than the marginal widget costs, so your business is well served just by finding any method to use that plant more effectivly. As an example, I was quoted $60,000 for a mold which would have produced parts at $0.005. (Very small plastic widgets.) At that ratio, any amount of scale will increase my profit, since the marginal costs, even if they increase by a factor of 10 or 100, are negligible to the cost of the tooling. (And the global market for this widget is measured in hundreds of units.) Any amount of reusing the mold is going to save me money. Sure we have problems if we need N+1 widgets in less than 1/N more time, but if we expected to need 2N widgets, we could reuse the tooling design at a second factory, and marginal costs actually do keep dropping.
But in real life, by the time you received orders for N/2 widgets, you were already breaking ground on your next factory. And if you get an order for 100N, you smile because now you can switch to a higher productivity manufacturing method, like stamping instead of machining; at 10000N you can invest in mass-producing the machines-that-make-the-machines. This is how we end up being able to make even complex products like cars so cheaply that we have more cars than people.
You know, part of the problem with the massive youth unemployment in China is that the Chinese, just like most people, don't want to work in blue collar advanced manufacturing. You might call jobs like consulting or investment banking as "useless jobs", but that kind of comfy white collar job is what everyone is sending their kids to university for.
Here's a video [1] which explains why, in 1955, manufacturing household goods was cheaper to make in the US than in China (and why, at the time, they thought this manufacturing dominance was the thing that backed the US position as a global superpower). It's not because Americans worked more cheaply than Chinese workers, it's because American factories had a well-developed tool-and-die expertise, which meant that when anyone in the world wanted to make something, they were well-advised to travel the US to get it made.
Econ 101's comparative advantage is true at an instantaneous point in time, which is a good start, but if perhaps it's just "knowing enough to be dangerous". Economic policymakers (and company leaders) would do well to think about comparative advantage as planning an optimal trajectory over time, which can mean sacrificing a short-term optimum in exchange for a long-term optimum, and if there even is a textbook solution for that, it's going to look less like a 101-level intersection of straight lines, and more like an iterative optimization over nonlinear differential equations.
[1]: https://www.youtube.com/watch?v=QU6nsfoNWDI