Tim Worstall writes:

The aim, point, and process of economic advance is to kill jobs- to get the task done with the use of less human labour...Economic advance is about using less labour to do any one task. This is why we obsess over productivity. It's not so that we can gain more from the same amount of labour either. It's so that some labour is freed up to go and do something else, which is the important matter.

This was true once, but it's more questionable now.

To see why, let's take a simple example*. Imagine there are two sectors in the economy, each employing 100 people. Sector A produces 100 units (productivity is one) whilst sector B produces 300, so total output is 400. Now imagine that sector A sees technical progress so that those 100 units can now be produced by just ten workers. The other 90 go to work in sector B.

In this example, output rises to 670 units. That's growth of 67.5%. And because those 90 workers are more productive in sector B than they were in sector A, their real wages are higher.

This demonstrates Tim's point. Labour in sector A is freed to do something more productive. That's progress.

This is not a contrived example. You can think of it as being a simple description of the shift from agriculture to industry in the 19^th century west or in China more recently. Workers stopped working with spades and hoes as agriculture became mechanized and moved into factories where they worked with more machinery and so were more productive.

Let's consider another example. Again, we start with 100 workers in sector A and 100 in sector B. The difference is that this time sector B is a low-productivity sector; it produces only 50 units. Now imagine that sector A becomes fully automated so its 100 workers must move into sector B. Total output now rises by only one-third. That's half the rate of my first example. This is despite the fact that technical progress in sector A is faster in this case than in the first.

Note that I'm conceding a lot to Tim here, as I'm assuming full employment.

And here's the thing. Just as my first example was reasonably realistic, so is the second. It's consistent with a point Acemoglu and Restrepo make - that any incipient downward pressure on wages due to workers being displaced by technology can be offset by employers having an incentive to create low-productivity jobs.

You can think of the displaced workers in this example as being those in good manufacturing jobs who have to take less productive work in shops or call centres.

Alternatively, because trade and technology are essentially the same things (as David Ricardo pointed out) you can think of this example as a stylized description of the impact on US manufacturing of trade liberalization. Yes, the economy is bigger in period 2 than in period 1. But those workers who moved from sector A to sector B saw their productivity halve - and by extension their real wages. This illustrates Dani Rodrik's point (pdf) - that even if globalization has net benefits, it can also have awkward local effects (pdf)**.

My point here is merely that technical progress and destroying jobs are not sufficient to achieve good economic growth, even where markets are functioning well. As Daniel Cohen says: "if a worker's individual productivity does not increase, growth is necessarily weak."

* I'm borrowing here from Daniel Cohen, who in turn borrowed from Alfred Sauvy.

** My example overstates the actual effects of globalization, which has been only one of many causes of increased inequality (pdf).