As somebody who does not live and work in the US, it seems plausible to me that the H1B system helps prevent other countries from obtaining similar talent hubs as Sillicon Valley. A lot of the talent is in the US, which attracts more companies, which attracts more talent, which means it's easier to go to the US and work there and start companies there than to do it anywhere else.
There’s also the easier financing, but yes - I’m sure the outcome is dominated by the network effect of having denser talent. That’s one of the reasons Silicon Valley is so hard to replicate elsewhere.
> phrased differently, the goal is to help industry, not hurt workers. hurting some workers is an acceptable cost, not the goal.
The phrase "help industry" has many dimensions. The simplest of course is that by increasing labor supply and suppressing wages it increases profit margins, rewarding shareholders.
Another important function is that by having more workers overall in the US, it increases the productivity of the domestic industry itself, due to increased competition for jobs driving up the productivity of the average worker. This in turn makes the industry more competitive vs its equivalents in other countries.
The average worker (whether permanent resident or temporary/H1B) who doesn't have significant investments likely doesn't receive much of those productivity gains, since they mostly go to capital owners.
Long term, it boosts returns to capital while capping returns to labor, the same trend noted by Thomas Piketty some years back.
I dont think there long term impacts are so clear or cynical. the question is less about productivity, but network effect, number of jobs, and quality of jobs.
> I dont think there long term impacts are so clear or cynical
The economic impacts I described are looking backwards, not forward, and the data is pretty clear that long term returns on capital swamp the returns on labor (especially since the 1970s). STEM workers have been somewhat insulated from that due to the industries they work in growing in the past few decades faster than the labor supply. It's anyone's guess whether or not either trend will continue into the future.
> the question is less about productivity, but network effect, number of jobs, and quality of jobs.
I'd argue productivity and returns to capital are almost everything when it comes to what informs immigration policy from an economic lens. "Network effect" is a mechanism, not an outcome, and outcome metrics like "quality of job" or even "quality of life afforded by a job" are not a concern of such policies. On average, they might improve, or they might get worse, but productivity and returns on capital will always go up, whether they require workers or not.
I understand that you are trying to make a point about return on capital, but I dont understand how you are connecting it to the question of H1-B visas and if local benefits to industrial expertise outweigh the downward pressure from labor competition.
> I dont understand how you are connecting it to the question of H1-B visas and if local benefits to industrial expertise outweigh the downward pressure from labor competition.
Because what you are calling "local benefits to industrial expertise" is ultimately realized in the form of returns on capital.
Whether these benefits outweigh the costs is an open question.
When the tech industry's growth was very talent constrained as it was in the last few decades, arguably opening labor competition had the effect of increasing overall growth (mainly through new production invention). The list of immigrant technologists who have created new technologies and products - and jobs as a result - could probably fill an encyclopedia.
It's unknown whether that type of growth - the kind that creates more and better jobs - will continue, especially given recent developments in AI.
If the benefits going forward are largely going to be based on massive increases in labor efficiency, then it's not as clear that the benefits (mostly to capital) outweigh the costs (mostly to labor). Most business models in AI are predicated on replacing people, who are expensive, not making more or better goods. Sure, we'll get some neat robots along the way that actually make stuff, but that will likely be a small fraction of the money to be made.
Or perhaps we are at the dawn of a new era of technology which will make more and better jobs. We'll see.
OK, so you were changing the topic to something else you wanted to talk about. That was not clear to me. I thought you were making a rebuttal to what I was saying.
Correct, if you look at my comment I was unpacking what the phrase "to help industry" could mean, not rebutting your comment.
It's relevant to the original context because what helps industry (in terms of immigration regulation) might or might not help workers in that industry.
> Existing studies that show an increase in capital’s share of income miss the growing role of depreciation in short-lived capital, in items such as software, says MIT’s Matthew Rognlie in “Deciphering the Fall and Rise in the Net Capital Share.”
Subtracting depreciation isn't a fair comparison. The example uses software as a short-lived asset. Has the monetary value of Google's search algorithms depreciated? They've been upgraded with routine investment, but the scale of the returns on their upkeep vastly outweighs the capital investment, otherwise Google wouldn't be so profitable.
Software of the internally-developed sort isn't even depreciable [1], so it's not clear how its value for these purposes would be determined (short of assuming it represents a percentage of the business's value).
Also, from the paper linked in your article:
> Once all compensation of employees at the sawmill is
subtracted, the remainder is its gross capital income. Some of this capital
income will be paid to lenders in the form of interest, some will be paid
to the government in taxes on profits, and the rest may be retained on the
balance sheet of the sawmill or distributed as dividends to shareholders.
Gross capital income is thus a very broad concept, encompassing funds that
are ultimately paid out to many different recipients—it is unaffected, for
instance, by the split in financing between debt and equity.3
GROSS VERSUS NET: CONCEPTS An alternative to gross value-added is net
value-added, which subtracts depreciation. This can be divided into labor
and net capital income, the latter being gross capital income minus depreciation.
Everything which I have emphasized above are examples of returns to capital. Excluding them from consideration in this presentation is ignoring how a large amount of returns are channeled to owners of capital.
Debt-holders gain from interest and shareholders are enriched via dividends and share buybacks that never appear on the article's net income derived graph.
Of course, when you willfully ignore those huge tranches of returns, then housing looks like a major factor, because it is the common asset class that has been on a largely unchecked inflationary track.
Finally, your article from 2015 argues that the overall trend will reverse and labor's share of GDP will start increasing. Here's what has actually happened since then:
The brief spike in 2020 was due to pandemic era redistribution policies like the child tax credit, among others. Since those have been repealed, labor's share has continued its prior trend downwards.
The interesting thing about the return to labor vs capital line of argument is that generally speaking, capital doesnt consume the types of product that labor is interested in.
When productivity goes up, that doesnt mean workers are making 10X as many houses or hamburgers, which capitalist are eating.
For me, this begs the questions of what exactly is being produced when we say worker productivity has increased, and where is it going? If it is "stuff" being produced, surely it should be evident somewhere, like massive exports hoarded stockpiles. Alternatively, the productivity is an illusion because there is a corresponding inefficiency or deadweight loss, like paying some service workers to create problems and paying others to fix them.
In my experience close to 100% of productivity increases accrue exclusively to shareholders.
When my companies have produced more output from the same inputs (or the same output from less inputs in the case of mass layoffs), we return the cash to shareholders by way of a stock buyback or special dividend the following quarter.
Maybe in some companies they instead give workers raises or outsize holiday bonuses, but I’ve never seen this.
What about the output goods produced? if you make 200% more hamburgers, the shareholders arent eating them. If every US company has doubled production, where is all this stuff piling up?
> For me, this begs the questions of what exactly is being produced when we say worker productivity has increased, and where is it going?
Power.
Political power: policies written to benefit the highest bidder.
Financial power: more leverage in being able to dictate terms of borrowing by workers - and being able to force the government to borrow from capitalists instead of levying taxes on them.
Physical power: Being able to buy/influence law enforcement (themselves a type of worker) to protect the capitalist's interests over those of other workers.
Increasing worker productivity manifests as greater returns that predominately go to capital owners, not workers. That concentrated wealth in the hands of capital owners is wielded as power in the political, economic, and physical realms.
The "product" that the increased productivity buys is control over policy at whatever level of government, not more washing machines or tires.
Returns of what? I feel like this argument is leaving out words and skipping logical steps. Hamburgers are a product, cars are a product, cleaning services are a product. Wealth is not a product.
If you have a company and worker productivity goes up 200%, where does the product go? Wealth created selling that product may go to the owner, and carry power with it, but that doesn't answer the fundamental question. Where is the product?
> If you have a company and worker productivity goes up 200%, where does the product go?
In a mature industry, there is no new product, because all else equal, demand doesn't change. The company makes the same amount of product, but with fewer workers (aka layoffs).
Even in an industry serving growing demand, increase in worker productivity is not the cause of increase demand for product produced by that industry. Any growing enterprise knows it's first more important to focus on demand than increasing productivity, usually by hiring workers at the lowest cost possible. Otherwise, your competitor will serve your customers needs before you do. Premature optimization is a waste.
What increases demand for products is technological innovation plus a need/desire for more personal convenience, comfort, and time, coupled with the funds to purchase those in the hands of a growing population. Why have most companies have staked their future profits on the developing world's demand growth? Because the developing world has the desire for all of the above plus a growing population.
The question of where the new product goes has nothing to do with the question of worker productivity unless the workers have the funds to purchase those products. The product goes where the purchasing power is.
Capital's share of the return, however, goes into assets and as I described earlier, power. It doesn't go into purchasing any increase in product created.
>The question of where the new product goes has nothing to do with the question of worker productivity unless the workers have the funds to purchase those products. The product goes where the purchasing power is.
That is my exact question, who is purchasing the goods? we have high employment and have supposedly high productivity. We dont have massive national export surplus. You say capital isn't purchasing the goods, so what gives?
Where is the black hole that is consuming all of the goods, if the workers dont get them, the rich dont get them, and they aren't exported.
> Where is the black hole that is consuming all of the goods, if the workers dont get them, the rich dont get them, and they aren't exported.
Take new cars as an example. We are producing fewer of them [1], they are larger and more expensive, and they are mostly being sold to wealthier people. So yes in this case, capital owners (people more likely to have more wealth) are the ones purchasing the product.
Also, for a while we have been shifting towards a services based economy, so for a lot of this production growth, you won't see physical products. For example, you can't see the software IDE subscription I signed up for yesterday.
We also don't have a national export surplus because we import so many goods that are not worthwhile to manufacture here, while we export a ton of services, petroleum, and other raw extracted materials, all industries that scale with technology/capital/machinery and not labor.
This still seems to negate the point you made earlier that there are huge productivity gains and 100% of them have gone to shareholders. It doesn't seem realistic that they are using all of the new services. I feel like I'm repeating myself, so I think this is the last post.
> This still seems to negate the point you made earlier that there are huge productivity gains and 100% of them have gone to shareholders.
I didn't say 100%, I said most (re-read my comments upthread). Please don't misrepresent my words. I choose them carefully.
Greater productivity does not automatically equal a commensurate increase in products/services delivered, which seems to be the flawed assumption you are unable to get past.
Here is a concrete scenario to illustrate this.
A company makes 1M units of a product at a cost of $1/unit, and sells them for $1.50/unit. Profit/unit is $.50.
Productivity doubles, so the same million units can now be produced for $.50/unit. When sold for $1.50, profit is now $1/unit.
The $.50/unit increase in profit goes mostly to shareholders.
There are no new products, no new services.
In reality, demand varies over time, so product output varies with that, but the gains in profit mostly have gone to shareholders.
The only time they ever go to labor is when labor is in short supply or when labor organizes to demand a larger share.
> If harm was the goal, something like a STEM worker tax or cutting R&D tax incentives would be easier
These would affect all STEM workers equivalently. The H1-B program, whatever one thinks of its merits, hurts domestic STEM workers and helps immigrant STEM workers.
Perhaps the result is that the overall opportunities are greater because the larger talent pool results in more companies being formed. That depends a lot on how mature the industry is, and whether technological trends like generative AI will replace large swaths or STEM workers altogether.
Phrased differently, the goal is to help industry, not hurt workers. Hurting some workers is an acceptable cost, not the goal.
One idea is that having a thriving industrial ecosystem helps those same workers more than the downward pressure.