What Chang wants us to understand is that because we used to have protectionism and we still had economic growth and development then therefore we should have protectionism in those places where we want to have economic growth. In other words the poor countries should throw up trade barriers so that all the rich world megacorps cannot supply the people of those countries. Thus will industry develop and in the long term, wealth will be created.
There are a few problems with this argument. One of the most glaring is that he takes historical levels of tariffs as evidence of levels of historical protectionism. Which is an absurdity: until well into the 20th century transport costs were more important than whatever tariff levels were as a barrier to trade. Just as an example, it is true that US tariffs near doubled post Civil War. But actual trade barriers fell as transport prices (essentially, the ocean going steam ship) fell by more than that doubling of the tariffs. His historical evidence of tariff barriers is thus highly suspect. The reason that most countries developed their own industries is precisely because non tariff barriers, those high transport costs, were more important.
Another problem is that, as he actually points out but doesn't make the connection with, all of his examples who developed behind such tariff barriers and with infant industry protection etc simply were not democracies in any modern sense. Even the countries that developed behind them in the 20th century like Taiwan (or his native Korea) were not. Semi-fascist military dictatorships would be a more useful description of the political systems actually. And don't forget what the sort of planning that he's advocating means: not just that government should encourage certain industries but also that local people must be actively prevented from wasting their energies in things which are not part of the plan. It's extraordinarily difficult to think of a way in which a free and liberal democracy could do such things. Force some companies to enter ship building, yes, perhaps that could be done with carrots and not with sticks, but how would one, in any semblance of a liberal society, prevent someone from setting up to build ships if that's how they desired to waste their money? This is the sort of thing that did actually happen in those planned economies too.
Even if we grant him his thesis, that such planned and directed industry, protected by trade barriers, did lead to industrial development, I can't actually see how anything like it could be done in anything close to a free society. Indeed, I'd even be willing to consider the idea that the reason this "worked" in certain societies (like parts of East Asia) and did not work at all in others (parts of Latin American and Africa) was precisely that those two latter sets of societies were not authoritarian enough to allow it to work. People had enough freedom to be able to ignore the plan.
One further very important point from Chang's own argument. He does insist that only those countries that have got to the technological leading edge benefit from free trade. His argument is absolutely not that the rich countries of today, those on that leading edge, would benefit from restrictions on trade: quite the contrary. His argument, such as it is, applies only to developing, not developed, nations. So don't allow anyone to start using his arguments, faulty even as they are, to propose that the UK or the US, EU, should retreat behind tariff barriers. That's not what even he is saying.
We might also mention that historical evidence of restricted trade areas is interesting in an historical sense: but it's not really of any relevance today. This is becasue of the sheer scale of modern industry. Perhaps, maybe, it made sense for the US to build a steel industry behind barriers. There were a number of companies in it and between them they created a market, however protected it was. These days, even the EU isn't a large enough market, all 500 million of us, to produce, say, a viable computer industry. The idea that Tanzania (just as an example) should have tariff barriers in order to encourage an indigenous computer industry is therefore ridiculous. Or a car industry: it costs $1 billion just to plan out a major new car platform these days, let alone tool up to manufacture it.
The scale of modern industry is simply such that anyone trying to recreate any substantial part of it behind tariff barriers is just going to be making shoddy goods, very expensively, for no very good reason. You might, just about, get away with a little bit of restriction with the billion and more in China and or India. But the idea that Somalia will, with the appropriate planning and protection, ever have a viable steel, car, chemicals or comuter industry is simply nonsense. It might well end up producing firms in an interesting niche or other: but the creation of an entire industry for such a small number of people just isn't ever going to happen.
And there's one final overarching reason why this autarkic route to development is undesirable: it's immoral. Building up infant industries behind tariff barriers is very much a case of jam tomorrow, not jam today. The idea is to deliberately remove from the inhabitants of the country concerned the ability to consume the delights of the current world. So as to enrich those who own the industry within those tariff barriers. That populace is subjected to decades of worse consumption goods than they could have had. Even if this does, in the end, lead to development we've still impoverished the people in favour of the capitalists of that society. Not that I think it does lead to such development: but even if it did that's what is being urged.
Which rather brings us back to why I don't think this will work in a democracy, or in anything even vaguely approaching a free and liberal society. Yes, sure, economic growth is important but not at the cost of deliberately impoverishing this generation. And that's what infant industry protection does and not only do the voters appear not to be willing to sit still for that (and thus it only, if at all, succeeding under authoritarian regimes) I very seriously doubt that it's moral for us to go around insisting that they should.