Josh Hawley is a familiar type: the prep-school/Ivy League toff who grew up wealthy and has, for obviously self-interested reasons, appointed himself tribune of the plebs, champion of the working class, and zealous defender of literal McJobs—in this case, jobs working the drive-through window at fast-food restaurants.
I suppose that young Josh Hawley might have learned something about the drive-through window at the bank of which his father was the president, but here is the voice of experience from an actual graduate of Burger King University and a veteran of the drive-through window at University Avenue and Loop 289 in Lubbock, Texas:
Don’t save these jobs.
Really. Don’t.
(“Burger King University” is what they used to call the training videos you had to spend an afternoon watching if you worked at Burger King in the 1980s. These were pretty much what you’d imagine them being.)
Hawley is joined by Democratic Sen. Richard Blumenthal (now there’s a brace of them) in pushing a destructive, anti-market, anti-innovation, AI-regulation model that would base decisions in part on “effects on employment.” That sounds anodyne enough, but, in hearings with Microsoft President Brad Smith, Hawley made clear that he intends to see this construed in something very close to a literally Luddite fashion—he doesn’t want to see automation replace a single job, including (this was his example, and he dwelled on it) jobs running the drive-through windows at fast-food restaurants.
(The Luddites, if you don’t know, were early 19th-century English terrorists who destroyed machinery, most famously in textile mills, to prevent automating repetitive tasks done by low-wage workers.)
While it may not be entirely obvious to those of you who have had frustrating experiences with Siri, understanding speech is something at which artificial intelligence is increasingly good. (As usual, the most relevant question is: Compared to what? Compared to, say, a Starbucks barista.) Krishna Gupta, CEO of Presto, a company that makes AI-driven drive-through interfaces, told the Wall Street Journal in July: “In three years I don’t think there’s going to be any human taking an order in any drive-through in the U.S.”
I have my doubts about that prediction, but many human drive-through clerks will likely be replaced by AI. Why? The usual answer people give is, “Because the computer is cheaper.” But that isn’t quite right. If all that mattered were cheap labor, then Porsche would manufacture all of its 911s in Port-au-Prince rather than in Zuffenhausen. What matters isn’t simply the per-unit cost of labor (or of labor substitutes, such as automation) but the cost relative to the output, which is another way of saying productivity.
German autoworkers earn, on average, around $35 an hour in wages. Never mind that cheap Chinese labor everybody is worried about, that’s a good deal more than the $28 an hour the average unionized U.S. autoworker makes.
Why don’t the suits at Porsche go to where the labor is cheaper? What is it? German nationalism? Lack of ambition?
The truth is, German autoworkers are really, really good at what they do. They aren’t the only people in the world who can make a Porsche—all the 911s are made in Zuffenhausen, but Porsche makes Cayennes in Slovakia and in Malaysia, where the factory specializes in right-hand-drive models for the local market—but, as with their colleagues down the road building Mercedes-Maybach S-Class sedans in Sindelfingen, the German autoworkers who make Porsche 911s are extraordinarily skilled, experienced, competent, and proven. When I visited the Mercedes plant in Germany a few years ago, the workers were remarkably confident that what would keep their jobs safe in the long run wasn’t any kind of protectionist policy or trade barrier, but the fact that they are very, very good at what they do. Add to the skilled workforce Germany’s highly developed, trade-oriented economy, trade relationships within the European Union and beyond, competent (though by no means unimprovable) governance, relatively productive relationships between corporate management and labor unions, etc., and that $35 an hour starts to look pretty good. Most other countries don’t offer that kind of economic environment, the absence of which imposes enormous costs.
And that’s why Porsche doesn’t pay workers in Haiti 48 cents an hour to build 911s: it’s too expensive. They can’t afford those low wages.
Low-wage jobs are always disappearing. They are too expensive.
My father was occasionally employed picking cotton back in the 1940s and 1950s, and his parents were by that time old hands at it. Cotton-picking provides a very useful illustration for how automation actually works in the labor market. At one terrible point in American history, cotton-picking was work done by slaves. (Cotton, rice, and sugar were the main drivers of slavery in the United States.) When the United States finally abolished slavery, cotton-picking was done by very poor workers, in many cases former slaves who had little choice when it came to occupation. (Jim Crow was in part about social segregation, but much of that organized racial suppression was about reducing competition in the agricultural labor market.) Cotton-picking was unpleasant work—fibers had to be harvested by hand by workers dragging 100-pound bags of cotton behind them with fingers bloodied from working with the spiky plants. Between 1850 and 1950, there were nearly 2,000 patents filed for cotton-harvesting machines and parts, none of which enjoyed much success. The cotton gin had been around since the 18th century, but cotton-picking remained manual labor.
Here’s the weird—and stupidly familiar—part: The politicians did not want those mechanical cotton-pickers to succeed. They wanted those workers to stay in the field—because they didn’t know what else to do with them. As Donald Holley wrote in The Second Great Emancipation: The Mechanical Cotton Picker, Black Migration, and How They Shaped the Modern South:
As early as 1932 the mechanical cotton picker evoked the worry that continued throughout the decade: The United States was in a major depression with thousands of people unemployed. The invention of a mechanical picker would have the effect of putting additional hundreds of thousands of people out of work. Additionally, the South was a poor and backward region; and its poorest people—those at the bottom of the agricultural ladder in the cotton fields, the lowly sharecroppers and tenant farmers—were the ones most vulnerable to the impact of cotton mechanization. Dispossessed of their pathetic livelihoods, these jobless masses would pour into Northern cities in a vain search for jobs.
The introduction of International Harvester’s Model H-10-H in 1942 changed all that. Work that had once taken many slaves or many very poor low-wage workers became work that could be done much more effectively by fewer people—by machines and machine-operators. Today, cotton-picking is a very high-tech, high-capital enterprise, with one man in a sophisticated harvester doing work that would have once employed hundreds. Many of those operators are, in fact, sharecroppers—they own their own machinery and harvest crops for big landowners in exchange for a share of the proceeds. The big landowners are saved the necessity of paying out hundreds of thousands of dollars (over a few seasons, more than $1 million) on machinery and learning how to use it, and the machinery owner gets to use his investment to make money off of land that he didn’t have to buy, pay taxes on, and assume the risk for. These modern entrepreneurial cotton-pickers make a lot of money in many cases—the successful ones can end up multimillionaire sharecroppers.
If Hawley believes that we would be better with more jobs and less automation, then I invite him to go find himself a nine-foot pick sack and show us what a Yale man is really made of.
As it turns out, eliminating 99.9 percent of the jobs in cotton-picking did not leave the nation plagued by ravening hordes of unemployed former farmworkers forming a restive, pre-revolutionary proletariat. The first half of the 1940s were not great economically (the GDP numbers look good on paper, but real life was grim as war, rationing, and New Deal central planning took their toll), but U.S. economic output more than doubled between 1947 and 1958. Between 1958 and 1969, U.S. economic output more than doubled again. Automating cotton-picking did not produce a gigantic mass of miserable unemployables “dispossessed of their pathetic livelihoods”—it helped to unleash the most wildly productive workforce the world had ever seen and contributed (one factor among very many) to the creation of the most affluent, most free, and most dynamic society the world had ever seen.
Hawley and Blumenthal—and all economic nationalists—are fundamentally pessimistic, believing the people they represent are too weak and too stupid to manage their own affairs without the benefit of political patronage. They see human beings as liabilities. That’s why the people who advocate interventionist industrial policy have historically been the same people who advocate population control—every person, in their view, is just another mouth to feed. (Not their own children, of course, who aren’t going to be losing jobs they never had taking drive-through orders at McDonalds—those are precious, unique gifts from God—it’s your kids who are the problem.) It’s the old Malthusian stuff, slightly updated: Too many people, not enough jobs.
If you view human beings as black holes into which resources must be shoveled, then a job—any job—looks like a solution to your problem. It does not matter to Hawley whether Worker X does anything that is economically useful (and doing work that could be more efficiently done by a computer is not economically useful). If Worker X gets a paycheck (and, preferably, health insurance and a pension), then somebody else has solved a problem that might otherwise fall into the lap of the welfare state. Another way of thinking about Hawley-style Luddism is as a complicated welfare program, one in which the payments are shifted from taxpayers and government agencies to businesses that are coerced into making decisions that run against their economic interests in order to satisfy the political interests of Hawley and his ilk. Economically, it’s the same thing as writing those workers a check, but doing it through opaque, back-door means that satisfy the No. 1 political need of the great nationalist-populist: It keeps events from exposing his political cowardice.
The more enlightened view understands that human beings are not liabilities but assets. Human effort is inherently valuable. And that doesn’t just mean the effort of geniuses and very clever people and people in high-status occupations—or people in paid formal work, for that matter. There’s a reason we don’t grow our own food. My time is valuable, and so I try to apply my working hours to the thing I do best. (Economics for English Majors: This is “comparative advantage.”) The people who grow my food, deliver my packages, refine my diesel, sew my boots, and put together the computer on which I am typing this all perform the very valuable service of doing things I am not good at—or doing things I might be pretty good at if I did them all day but am not as good at doing as doing what I do for a living—so that I can concentrate all my effort on doing what I do best. We free-market types like to talk about competition—and competition is important—but capitalism is profoundly cooperative: This marvelously productive worldwide economy is something we all do together. The value created by the guy picking avocados isn’t just in avocado production—it’s in the work of everybody else in the world who gets to do something else because Avocado Guy, and millions of others like him, have liberated the rest of us from the need to produce our own food, mill our own flours, manufacture our own tires, etc.
As F.A. Hayek and Ludwig von Mises (among others) pointed out, you can’t plan that. The relevant information is distributed throughout society, every data point is contingent on potentially millions of other factors, and the lifespan of a relevant data point may only be seconds. I don’t know how many avocados I’m going to want to buy the next time I go to the grocery store—and Hawley and other would-be central planners sure as hell don’t know, either. We let markets—including the labor market—work because the actors in the marketplace collectively are in possession of the relevant knowledge that central planners cannot get their hands on. It is not that the free market is unplanned—it is exquisitely, complexly planned by billions of people, each acting on the problems closest to him, about which he knows the most and is best positioned to act. You cannot impose a rational plan on that–you can only impose rigidity and ignorance on it, pretending that that is a plan.
Some people will try to tell you that the middle classes have not seen any real economic progress since the 1970s or 1980s. This is pure baloney. (Unless you are getting it from the New York Times, in which case, it is mortadella.) The best economic data we have do not support that conclusion, but, if you don’t want to dig into the economics, then go to a library. Read a book written in the 1970s or the 1980s. Read a magazine. Look at pictures, look at the advertisements—trust the evidence of your own eyes. Some of you are old enough to remember what a typical house was like in 1975, what a grocery store was like in 1981, how life was actually lived. The Brady Bunch was not a documentary, but there is a reason people laughed at the show’s jokes about the high price of meat—a complaint coming from the family of a prosperous architect who lived in a large modern house with six children, a stay-at-home wife, and a live-in servant. There’s a reason people reading The Stand in 1978 nodded along with Stephen King’s observations about the price of butter. In 1950, groceries (“food at home,” in economists’ parlance) accounted for more than 16 percent of household spending; today, that figure is less than 5 percent. And before you say it, no, that hasn’t been more than offset by eating out: Food “away from home” went from 3.8 percent of household spending in 1950 to 4.7 in 2019, a modest increase, which is part of why overall food spending declined from 20 percent of household spending in 1950 to less than 10 percent today.
Of course we are far better off today than we were 30 or 40 years ago. The evidence is all around you.
That doesn’t mean things cannot—or should not—be even better. There is always room for improvement. If you think that improvement is going to come from letting Sens. Hawley and Blumenthal call the shots in the technology industry, then you are a fool.
And Furthermore …
Burger King wasn’t a bad place to work at all. It was far from the worst job I’ve ever had, and it certainly was a happier, healthier, and more positive workplace than was, say, The Atlantic. We were across the way from a cowboy bar, which meant that after last call, we were flooded with all the drunks who hadn’t … found something better to do … when the bar closed. They were a lively but mostly friendly bunch. One of them, barely able to keep on his feet but holding on to his sense of humor, looked over the menu and asked me, “What do you recommend?”
I thought about it for a second.
“Alcoholics Anonymous.”
I worked the drive-through, cleaned the flame broiler at night (a job that required me to wear something pretty close to a hazmat suit), cooked, everything. But, as it turned out, I spent much of the summer outside the restaurant: They were building some of those playscapes with the slides and the ball-pits and whatnot, and they asked me to help. It was a pretty good deal: If I remember correctly, I got to clock in for my usual hourly wage and then got paid an extra $10 an hour in cash (this may not have been strictly legal), and, at the end, they gave me a coupon book for something like 100 free meals, which were of little interest to me (the food was fine from a hygiene point of view, and there was very little of the fast-food horror stories you’re probably imagining, but I didn’t and don’t like Burger King food) but kept my friends in Whoppers for a couple of months.
The franchisee I worked for was rumored to own a Rolls Royce, though I never saw it. I’m glad he made his paper—it couldn’t have been a fun business. But Burger King did alright by me. And if that job hadn’t been there because of AI or automation, do you know what I would have done?
Something else.
Words About Words
Here is a really illuminating opening in a Slate article:
The Washington Post published a story on Monday that may harm Virginia Democrats’ chances of preventing a complete Republican state government takeover this November.
Based on a tip from a Republican operative to whom the Post granted anonymity, the paper reported that a Democratic candidate running for a House of Delegates seat in a competitive district has livestreamed videos of herself having sex with her husband on the online platform Chaturbate.
Let me ask you this: If there had been a story about, say, Donald Trump Jr. and Kimberly Guilfoyle doing sex acts for money on Chaturbate, do you think that the Slate lead would prioritize the following facts? 1. It might harm Republicans’ electoral hopes; 2. It came from a Democratic operative; 3. Said operative was granted anonymity by the paper. I mean, holy heck! Here we have a political candidate doing live porn to supplement her income, and Slate is only interested in the fact that this might help Republicans win in Virginia, and in pissing on the Washington Post for publishing the story without disclosing the name of the source who tipped them off.
But that is the Slate view of the world, which is also the view of much of the rest of the Democratic-aligned media: The news doesn’t really matter as much as whether telling the story might help Republicans.
It’s a broader and more general version of the “Republicans pounce!” phenomenon. It doesn’t matter what the story is, how depraved or embarrassing or foolish some Democrat may have been, the real story is how “Republicans pounce!” on the episode.
Ain’t gonna be no “Democrats pounce!” headlines as the latest Lauren Boebert story makes its way through the cycle.
Eric Adams Is Handing Republicans a Flamethrower for 2024. Again.
The problem with creating the impression of a nonexistent problem is that it becomes impossible to solve.
This is all to say: How best to address the migration issue is an important consideration. But Adams is not only bungling the response, he’s actively choosing to politicize it in a way that helps Republicans.
…
Meanwhile, national Democrats who invited the wolf into the henhouse can only hope that Adams has not yet again created a winning blueprint for Republicans to overperform in this coming election cycle.
Does anybody think that crime or immigration are in fact nonexistent problems in New York City? Nobody in New York, I would think. Exaggerated? Sure. Nonexistent? That isn’t journalism—that is propaganda.
Given the subject matter, as a journalist, did you have ethical concerns? Other journalists I’ve interviewed have struggled with the question of when you should or shouldn’t humanize leaders of violent groups like the Proud Boys. Did that cross your mind at all?
The question is asked of a journalist who has interviewed and profiled Proud Boy leader Enrique Tarrio. “Humanize” is, of course, an entirely empty word: Neither Enrique Tarrio nor the Proud Boys nor Adolf Hitler nor Pol Pot needs to be “humanized”—they are human, their crimes are human crimes, their failings are human failings. That’s what humans do, at least part of it. People often wrote of the danger of “normalizing” Donald Trump, as though Donald Trump weren’t as normal and all-American as diabetic amputations and school shootings. The complaint, put into more honest English, is this: “If you give readers a more complete and interesting view of a person, then that person will become less useful to us as a hate totem to employ for the purposes of community-building ritual-denunciation ceremonies.” But nobody writes that. And nobody will usually go so far as to say that the Washington Post should refuse to report the news when it is bad for Democrats, though sometimes they get close to it.
Other Wordiness …
Drive-through, not drive-thru.
Unexpected News …
Finally, I have something in common with Lauren Boebert. She seems like a fun date, but I still think my theater-ejection story is better.
Elsewhere
If Americans don’t seem grateful for the amazing, wonderful, outstanding state of the economy, it is because the economy ain’t so amazing, wonderful, and outstanding for most of them at the moment. More in the New York Post.
You can buy my most recent book, Big White Ghetto, here.
You can buy my other books here.
You can see my New York Post columns here.
You can check out my Competitive Enterprise Institute series on jobs and working, “How the World Works,” here. I interview a different guest each month: Veronique de Rugy is up now, and soon to come are Bryan Caplan, George Weigel, Klon Kitchen, and others. If you have an idea for a guest—or if you think you’d make a good guest—write to me at kevin@thedispatch.com.
In Closing
Eventually, the money runs out. Wall Street Journal:
The West Virginia University Board of Governors voted Friday to cut 28 academic programs and eliminate 143 faculty jobs—despite opposition from students, alums and faculty—in favor of what they say is a necessary step toward financial viability.
The public flagship university will eliminate a Ph.D. in mathematics, master’s in public administration, most foreign-language instruction, graduate degrees in higher-education administration, and ceramics and sculpture degrees. Faculty cuts will include positions in its schools of law, mathematical and data science and public health, and in its chemistry and plant and soil-sciences programs.
…
“We can’t keep every program. We can’t do everything that we’ve been doing, because we’ve lived beyond our means,” President E. Gordon Gee said at a faculty senate meeting earlier this week.
A Wall Street Journal analysis of financial records found that after adjusting for inflation, spending at the university increased by 38% between 2002 and 2022, or 29% on a per-student basis.
Protest all you like. Complain all you like. Call it racist or sexist or monocisheteronormative or anything else. In the end, the math is the math—in West Virginia and down the road in Washington.
Please note that we at The Dispatch hold ourselves, our work, and our commenters to a higher standard than other places on the internet. We welcome comments that foster genuine debate or discussion—including comments critical of us or our work—but responses that include ad hominem attacks on fellow Dispatch members or are intended to stoke fear and anger may be moderated.
You are currently using a limited time guest pass and do not have access to commenting. Consider subscribing to join the conversation.