>If jobs can be exported overseas to places with much lower wages (and lax/nonexistent environmental laws), then all a higher minimum wage (here) does is put more people out of a job.

As far as I know, the economic literature is pretty clear that this is not the case in the US. Even regional changes, where an employer can reasonably relocate outside the e.g. city boundary without much difficulty, don't show job losses.

A lot of manufacturing left the USA along time ago due so I’m not sure why the economics literature says it doesn’t matter? Yes, the service jobs that now dominate in the USA are largely unaffected, but that’s more due to survivor bias than any underlying economic mechanism.

Automation and self service (like in European cafeterias) is largely the response when services become too uneconomical to be done manually. Those elevator operators aren’t coming back to the USA.

But there's no evidence I'm aware of suggesting that minimum wages were the cause of deindustrialization in the US. I suspect most of those factory jobs were far above the minimum wage.

And people often wave the automation threat around - but why hasn't it happened yet? In concrete studies of minimum wage increases that actually occurred, we do not find job losses.

Minimum wage is, sure, not the cause of deindustrialization. Rather, it's what people tend to advocate as a cure for a large underclass stuck in poverty, but it's not addressing the core issue. We have a larger underclass stuck in poverty, because the kinds of jobs that historically allowed them a way out of it were sent to places with even poorer people, and the change that allowed that was a change in government policy (tariffs, etc.). So a minimum wage, like a blood transfusion, might help out for a bit, but it's not addressing the underlying issue, which was a (major, sustained) change in government policy.

Minimum wage in the USA has always been paltry, and increasingly irrelevant over time: very few work for it now. Labor costs rise because people at the low end get better options, not because of a gov mandated increase.

But this is just the government being conservative, actual significant increases that affect more than one percent of the population will have an affect.

A ton of minimum wage jobs simply cannot be exported or automated at this time.

Basically all jobs left. Your McDonalds Big Mac isnt getting grilled in the Phillipines.

There definitely are experiments with robotic cooks [0].

Not widespread yet, but neither were cars nor computers in their early days. The ratio of robots to humans in kitchens is likely to grow in the future. From its current very low value to something more meaningful.

[0] https://singularityhub.com/2020/11/06/flippy-the-fast-food-r...

Horn and Hardart closed in the 1970s because given the choice, people would rather not eat in automated restaurans. But, with wealth inequality as high as it is the "automated" restaurant may be making a comeback.

Just as income inequality is bringing a new wave of app-based delivery servants.

Or people will simply prepare their own food.

I can tell you absolutely that is my response. I'm not going to eat some automated thing that nobody's checking, or has poorly maintained for the least dollars.

It will be a long while.

Besides, nothing about that automation actually changes the funding labor problem.

I have always found it tough to sell a robot a good meal too.

There will be basic labors we need doing and a lot of people doing them for the foreseeable future.

Wouldn't an automation tax change the funding labor problem?

Assuming an already expensive automation makes any kind of economic sense, it could. But right now for a vast majority of basic labors, automation remains out of reach being impractical economically, if not technically.

And robots don't need a good movie, a night out, a good meal, or any of those things, so whatever taxes has to find their way to the people who do present real demand for those things.

And those people do have costs, unless we simply allow them to cease to exist and not show up for work anymore.

>vast majority of basic labors

We might need some qualifiers to understand exactly what you mean. Most automation is readily within reach if the job is rote and the economic differential is there. (FWIW, I used to build and program assembly line robuts).

>whatever taxes has to find their way to the people who do present real demand for those things.

I think this is the intent. Call it UBI, Freedom Dividend, or whatever, but the goal is to fund the societal base that has the actual need for whatever automation (or humans) will produce.

Those are not real answers in my view. How we value human labor is a real answer, and the fact is we need to start with the cost of existing and or showing up for work.

As for rote... look, automation is expensive to establish, adapt transport and maintain, and need infrastructure to support it all.

People, at their real cost, are a GREAT deal and that will be true for a very long time yet.

Go work through a few cases. I have experience in automation. It is the little things that cost a lot.

We are more likely to succeed with augmented labor, and already do. Better tools = labor multipliers.

Automation is expensive, but so is human labor. Once you cost in all the human costs, like paid time off, the variability in performance, administrative overhead, etc. it can be a lot cheaper to automate. Which is exactly why it's done. As for cases, I worked in automotive assembly line robots. Upfront costs are high, sure, but nothing compared to the legacy costs of the dozens of the unionized human labor they supplant. (I'm not against unionization btw, only bringing it up because it tends to - fairly IMO - sway the power to the worker in terms of higher wages). An automation tax would at least make up for some of the lost revenue that supports societal infrastructure. It would also sway the economics to extend the time horizon of when it usurps specific segments of the labor economy. Automation is never fully automated, so I would consider it all augmented. It just tends to shift the labor towards fewer and more skilled workers.

"Can be a lot cheaper"

Yep. I am not anti automation. Cars, for example. Great use case, as you say.

Scale there, appetite for capital investment there, ability to weather high maintenance, change costs all there.

There is an opposite end of the scale, and it is that end we are a very long away from.

Also, the likes of FANUC do have dark running facilities that use robots to build robots.

I do have experience too. That stuff is expensive, complex, and very rigid. Not applicable, at this time, to the vast majority of basic labors we depend on every day.

Finally, one can automate workers away, or automation can free them to do more things.

The latter is a lot more sustainable, but a lot less sexy, and not so well aligned with the dominant business justifications used to sell initial automation.

>automation can free them to do more things.

This was my latter point. Automation is often augmented, to use your words. This is more true the higher you get in skills. A Da Vinci robot does not replace a surgeon. But lower level skills are where the difficulty is at. A fork-lift driver may be automated away completely by a parts-picking robot. This is where unions are beneficial in my view. They can ensure at least a relatively soft landing with retraining, replacement, etc.

Yeah, had to respond on this one. I think we largely see it the same way, and did a little feeling around to center on the core concept.

Amazon did find a niche, the package movement, where everything aligns nicely in a fashion not too unlike the cars. IMHO, the big difference is more dynamic decisions are needed for the packages, where the cars require more complexity, but are very consistent otherwise. Software innovation got Amazon where they wanted to be, and it's a good, genuine innovation that did expand the automation space.

Similar levels of effort are necessary for all those basic labors. This is why I maintain we just won't be there for a considerable time, despite successes like Amazon out there. Scale, and dynamic locations, terrain and decisions are huge barriers to entry. So many cases... and mechanisms / processes that can address them, and or minimize them.

Costs on all that are off the charts, as is maintenance, simple acquisition of the machinery.

More people centric augmentation will happen sooner. Mostly, this will preserve people from wearing out and or increase their throughput more than anything else. It may be we extend capabilities in a more transparent way too. That's going on now, of course, but maybe some of that will scale, get cheap, someone finds another niche...

All good stuff really.

But, "the robots are coming" really does not apply for a solid majority of the basic labors out there we depend on. Of course, that doesn't mean we won't be automating. We will, and it's a good thing regardless of how we sort out the humans economically, which was your point, and one I agree with.

Cheers! Good discussion. Have a great week.

Exactly! And tons of those jobs are paying less than it costs laborers to work them too.

That is a raw subsidy, and it's US paying so big players get a bigger profit.

We should not pay a dime, and they make a smaller one, but they also earned it fair and square too, not mooching on the public dime.