You want a bigger paycheck. Convince me.

There’s a common myth that standard economics predicts that people are paid an amount of money equal to the value of the things they produce. Actually, this isn’t true - in fact, the idea doesn’t even make sense.

In standard economics, your wage is equal to the marginal productivity of the company that hires you. That means that you get paid an amount equal to the amount that the company’s production increases when it hires you. Now, that may sound kind of like “you get paid what you produce,” but it’s not the same thing.

People generally don’t produce things individually. They produce things together, with the assistance of capital such as machines and buildings. So when a company hires you, its marginal productivity changes, because your presence affects the marginal productivity of everyone else at the company. In other words, in a competitive, classical economy, your wages are set not by your own effort and skill, but by how your effort and skill combine with the effort and skills of all the other people at the company, as well as all the company’s capital assets.

This brings me to a famous chart that has been making its way around the Web for a few years now - the divergence of wages from productivity. Here’s the chart:

It looks like until the early 1970s, wages and productivity grew together, like economics says they ought to, but after that they diverged. Suddenly, people are producing more, and not getting compensated for it.

What happened? Are markets failing to clear? Is the productivity or the wage numbers being misreported? Is classical economics just wrong?

Well, maybe. But actually, this interpretation of the chart is wrong. It’s a mistake. What the chart shows is average productivity, not marginal productivity.

Average productivity is just the total value of stuff produced, divided by the number of people producing it. It’s a very different thing from marginal productivity. No economic model says that people get paid based on average productivity. If they did, there would be no income left over for capital - no profits, rents or interest. We’d be living in a sort of a Marxist world, where labor is the only thing with any value.

But we’re not living in that kind of world. So average productivity and wages are perfectly free to diverge. In a Solow model - the most basic growth model that economists use, and one that all economics students still learn - the gap between wages and average productivity depends on the steady-state growth rate of the economy, on the rate of population growth, on the rate at which capital depreciates, and on a parameter called the “capital share of income,” which represents how much of national income is captured by capital owners.

Using this simple model, we see that a number of things have happened in recent years that could have caused the divergence between wages and average productivity. Population growth has slowed down. The long-term growth trend has slowed down. And, since around 1980 - and much more so since 2000 - the share of income going to labor has started to decrease:

This last fact - the increase in capital’s share of the national pie - is really the big mystery. Why has it happened? There are two basic competing theses: the rise of the robots and the great labor dump.

The rise of the robots is the idea that capital goods have gotten cheaper. This has made it profitable for companies to replace humans with machines. Economists Lukas Karabarbounis and Brent Neiman find some evidence for this.

That evidence is challenged by Michael Elsby et al., who find instead that a glut of labor from China, India, and other developing countries - Thomas Friedman’s global flattening - has been the biggest factor behind the fall in labor’s share of the pie. Outsourcing and global supply chains, along with the fall of communism, mean that suddenly, workers in rich countries are competing directly with workers everywhere, many of whom live in capital-poor countries. That will naturally make labor less scarce, and hence less valuable. I call this the “great labor dump” hypothesis. It’s a lot more optimistic than the “rise of the robots,” since the labor dump will eventually peter out, and labor will bounce back, along with many of the industries lost to foreign competition (unless better ones have replaced them)

Whatever is causing the shift, though, it’s wrong to think that Americans used to get paid a fair wage for what we produced, and are now getting cheated of our just desserts. In a globalized capitalist economy, you don’t get paid what you produce - in fact, you don’t produce anything without others to help. What you get paid is what you can convince other people to give you.

Noah Smith is an assistant professor of finance at Stony Brook University and a freelance writer for finance and business publications. From Bloomberg News.