The ninth thing we're told is that even though manufacturing is becoming a smaller part of our economy, of all economies, it's still very important oh yes indeed it is! There's a certain sadness in watching the argument develop in this chapter in fact.

Chang is quite right on his facts: it isn't that manufacturing output has shrunk at all. In the UK that was rising until 2007 (and we don't know how much of the subsequent fall is recession related or structural) and it's still rising in the US. It's just been rising less than the growth of the rest of the economy: thus falling as a proportion but not absolutely. Manufacturing employment has been falling substantially: a small part of this is simple reclassifcation. The graphic designers who used to work at the factory were counted as manufacturing workers: now they take their cocaine in Soho lofts they're service workers. The majority of that workforce fall is because of rising productivity: we simply need fewer people to make ever more stuff as we become more efficient at using labour to make things.

Chang gets all of this right: then he makes something of an intellectual leap and it's sad to see the tumble into the chasm of illogicality.

Given all of the above he says that manufacturing is still important and we should work to increase the portion of the economy that is such. The argument being that as productivity is easier to increase in manufacturing than it is in services, as manufacturing becomes an ever smaller part of the economy then total productivity growth will fall. Well, yes, it will, undoubtedly (and there's rather a clue as to why productivity growth has been falling in rich societies in recent decades: because that manufacturing where productivity growth is faster has been becoming an ever smaller part of the economy).

But to then insist that we must have more manufacturing in order to improve productivity growth is most odd. For we don't desire productivity growth per se. It's nice to have, for sure, it means we can make more stuff with fewer inputs. But even with that we only actually want to make more stuff, become more efficient at making stuff, that we actually want. There's no point in becoming more efficient at making Simon Cowell for example, as we've all got a surfeit already. And so it is with things that are manufactured. We don't want to simply become more productive: we want to have more of the things that we want with the resources we've got available, not more things that are manufactured just because it improves average productivity.

It's also true that manufacturing (yes, output is rising, but as a portion) of the global economy is falling. So the advice that every country whould focus more on manufacturing is ridiculous. Manufacturing what for whom? If manufacturing is carriages and services are cars (bear with me) Chang's insistence here is like saying we should all be making more buggy whips. Sure, no one particularly wants them but we're getting ever so much better at making them that average productivity would rise as a result of doing so.

The point here being that "productivity" isn't some thing that we should reify. It is indeed the secret to rising living standards: as Paul Krugman has said productivity isn't everything but in the long run it's almost everything. What Chang's missed though is that the output is measured at market prices: if we overproduce manufactures simply because this will raise the productivity number then their market price will fall: and productivity won't in fact increase at that point, will it?

So even though his basic facts are right here his prescription still fails. For while we would like rising productivity and it's easier to raise productivity in manufacturing than servicves this does not then mean that we want to throw resources at manufacturing.

Another way of clarifying this point is that, as we said yesterday, the purpose of all production is consumption. Sure, it would be nice to be more efficient at production: but only of things that people want to consume. And as it happens, it appears that further units of services produce greater consumer surplus than further units of manufactures.

Chang also goes on to point out that developing countries must concentrate on manufactures as this is the only way to raise their general productivity, that productivity increase that by definition leads to becoming a developed country. And I'd agree that it's highly likely that the developing countries will go through their own industrial revolutions. But not for this particular reason:

"If you base your development largely upon services from early on, your long term productivity rate is going to be much slower than when you base it on manufacturing."

The confusion here is that yes, when you're at the technological frontier then improving manufacturing productivity is easier than service productivity. For when you're at that frontier there's an awful lot of head scratching and pondering about what to do next. When you're well behind that frontier then there's no particular reason to think that this is so: indeed, we might think that services are easier. Take, just as an example, retailing in India and the computer hardware industry in India. In which do we think it would be easier to improve productivity?

I'd argue in retailing, the service, rather than computing, the manufacture. To improve the productivity of retailing all we've got to allow (or, err, the Government of India has to allow) is WalMart and Tesco in to start building the standard retailing logistics chain. That's going to be far easier (and cheaper!) than trying to build silicon fab plants (and all the rest) in a country without reliable electricity supplies. Or we might argue that we could improve the various state bureaucracies by computerising them away from the current quill pen and parchment systems.

That it is more difficult for us rich world people to improve services productivity than manufacturing such is entirely true. That the same is true of those places mired in seventeenth century services productivity is not. And what's really interesting about this argument is that if services are 70% of the economy (as they are, UK and US alike) and current poor world services became as productive and efficient as ours, then we'd see those poor countries become vastly richer whatever they do about their manufacturing. Simply because they'll be getting more services for the same as the current input: that's what increasing productivity means you see?