Press a Single Button, Produce All the World’s Food

This is an extreme example to illustrate how to think about personal income and how it relates to a worker’s productive output. Imagine that agriculture becomes so automated over the next century or so that the entire world’s output can be produced by pressing a single button. Yes, it’s silly to imagine that the button-pressing could not be automated as well, but bear with me. The entire world’s agricultural output can be done with a single person who has no particular skills.

Think for a moment about our single button-pressing laborer. How much does he earn? Should it be trillions of dollars, since he creates so much of the world’s total value? He creates all of a necessary ingredient to a functioning economy (and a living human species, for that matter). But that’s not the right answer, because he’s easily replaceable. He is not uniquely qualified. *Anyone* can push that button (for the sake of this example). So you only need to pay him enough to attract him to that job. He needs to be paid just as much as any other unskilled worker with a similarly difficult (which is to say similarly easy) job. He’ll earn roughly what an unskilled landscaping worker or dishwasher earns per unit of time worked, because the product of his labor, though massive, would be produced whether we had him or not.

Hopefully this simple example points out that you can’t just look at the value of the final product and claim that the worker produced it all. You have to talk about a person’s *marginal* productivity, which is the difference between what you’d get *with* vs *without* that person in the world. A person’s income depends strongly on this, the marginal product of labor. In no sense can you simply map a person’s labor onto a final product, determine the value of that product, and thus quantify the worker’s contribution. In most cases, some or most of that final product would be produced anyway by someone else. In other cases (say, a CEO of a company who creates a truly novel product, one that wouldn’t exist but for her/him), the final product would be substantially different without that person.

Now we’ll change the example a bit. Suppose again that there is a machine that produces all the agricultural output in the world, but it requires quite a lot of skill to use. And there’s no upper-limit to this skill; the very best “agriculture machine operator” in the world isn’t identical to the median or 80^th or 99^th or even 99.99^th percentile. Suppose there is variation all the way to the top. There are so many idiosyncrasies and complexities to running the agriculture machine that you absolutely need the very best, very most productive person in the world to do it. The relative value of the very best and the second best might be only a 1% difference in productivity, but that translates into tens of billions of dollars. Humanity should be willing to pay something comparable to make sure they attract the very best person to this job. If the very best machine operator can produce $100 billion more in output each year than the second best, then getting this person for a $50 billion annual salary is an absolute steal.

In the simple button-pushing case, where an unskilled worker can do the job and it’s nearly impossible to screw it up, humanity will be just as well off if the current button-pusher ceased to be. In the other example, if you don’t put the best guy in the world in that job, humanity really will be a great deal worse off for it.

I’m not seeking to morally justify income inequality with this post, although I personally think that income differences that are driven by true productivity differences are perfectly fair in a moral sense. My goal with this post is far less ambitious. I’m trying to get the reader to consider *why* these big differentials in income exist. What is missing from many rants against inequality is any *theory* about the determinants of income. You can’t possibly know that something (like a person’s income, for example) is too high or too low unless you have some kind of theory telling you what it should be. If someone actually had such a theory, and it told them that everyone (at least every identifiable class or sufficiently large grouping of people) earned basically what the theory said they should earn, it would be a lot harder to claim that inequality is a big problem. If someone’s theory tells them that, say, farmers are underpaid but bankers are overpaid, then we can open up that theory to scrutiny and see if it’s failing to capture something important. This can be an enlightening exercise. It needs to take place.

                                                                                                                    

World gdp = $75 trillion in nominal terms, $108 trillion in real terms (adjusting for purchasing power). https://en.wikipedia.org/wiki/List_of_countries_by_GDP_sector_composition

World Agriculture = 6.1% GDP https://en.wikipedia.org/wiki/Gross_world_product

World Agriculture = 4.5 Trillion in nominal terms, 6.5 Trillion in real terms.

But it’s kind of silly to talk about the “total” value of agriculture. The first 90% is probably necessary for humanity to exist at its current population. It's silly to talk about its "total value," which you might argue is infinite. The next 10% is worth quite a bit less than 1/9^th of the first 90%. This last 10% goes into making our food tastier (feeding cattle and making sugary snacks), not meeting our basic caloric needs. An additional 10% beyond what currently exists possibly has negative value if you account for health issues related to overeating or the costs associated with storing borderline-useless food and dumping spoiled food. I acknowledge I am playing fast and loose with the marginal value of food in my post; if the extra food produced by Mr. Super Producer in the 4^th paragraph isn’t very valuable, then his compensation will obviously be less than I imply above. Hopefully you can look past that and acknowledge the point about the marginal product of labor. You can always change the example to something that doesn’t have a steeply declining marginal value.

Another aside. The examples above describe a ridiculous amount of productivity coming from a vanishingly small proportion of the population. These scenarios are not absurdly far off, if you think in relative terms. Two hundred years ago almost everyone in the United States was a farmer; today almost nobody is. To a crude-but-relevant first approximation, the percent of the US population who were farmers has gone from 100% to 0%. The developing world is seeing dramatic economic growth, so presumably (I should say hopefully) they will follow a similar trajectory, so that what’s true of the US will eventually be true everywhere. The point here is that absurdly few people are now doing the work of what used to be absurdly many people.