Greg Clark's incredibly confused argument

Greg Clark has a very strange, even incoherent, argument in today's Washington Post.

Essentially, it boils down to, we're going to have to raise taxes because machines are going to replace all the labor that people do and thus we'll have to pay for them somehow. We raise taxes so that those who have been replaced by the machines will be able to eat.

What makes this near incoherent is that Clark himself is an economic historian. He knows that this argument has been going on since the invention of industrial machines themselves: we even have a name for those who espouse it, Luddites.

This has happened in agriculture, as machines have replaced hand labor and, no, people did not then starve in the fields. They went off and manned the factories. Then we increasingly automated the factories and industrial employment has been falling for about 50 years now. Did people then starve? No, they went off and occupied the service sector. Now Clark knows this, he's a historian of exactly these movements of labor between sectors over time. So why on earth is he suggesting that this time around it will be different?

Apparently machines are about to take over jobs that we previously thought that only humans could do and thus the low skilled worker will have no possible job to go to. There's a number of different problems with this idea: the first being that sure, those who came off the land a hundred years ago were unskilled at anything other than being potato farmers: but one of the definitions of a human being is a being that can acquire skills. In fact, we're rather known for it as a species.

The second failure of the argument is to assume that there's some limited number of jobs that need doing. This simply isn't so, it's regarded as one of the basic fallacies that needs to be beaten out of every economics student. We've unlimited desires and limited resources with which to satisfy them. Freeing up resources such as labor from satisfying one desire does not make that labor redundant, it simply means that another, different, desire can be satisfied.

But the real howler comes here:

And as machines expand their domain, basic wages could easily fall so low that families cannot support themselves without public assistance.

It's difficult to think of what mechanism Clark might be proposing here. So, OK, the machines start creating all of the things that we desire. Goods and services are indeed manufactured by these machines. Does this mean that real wages decline? No, certainly not. It would mean that we've got more goods and services for less human labor: that is usually taken as a sign that goods and services are becoming cheaper.

Yes, automation makes things cheaper. So if things are becoming cheaper, why would real wages be declining? Real wages, how many goods and services you can buy with labor, will be rising, not falling.

Clark's Op/Ed is a very strange one to come from an economist. Human beings cannot learn new skills, there's a limited number of things to be done and falling real prices for goods and services means a fall in real wages.

That's three economic fallacies in one piece: can anyone point to an Op/Ed (from an economist) with more than that many?