Why are growth declines sharp and sudden?

…these econometric studies imply that the decline in growth rates should be gradual, since convergence to high incomes is gradual.  Actual experience, however, rarely displays such a smooth adjustment.  More commonly there is a sharp drop to a new, lower growth rate, as noted above.  In Japan the drop was probably precipitated by the OPEC oil price increases, given Japan’s heavy dependence on imported energy, and by the concurrent worldwide stagflation, particularly in Japan’s major export markets.  In Korea, enormous chaebol-initiated investment projects artificially supported growth for a time but also laid the foundations for the financial crisis that precipitated the sharp decline in growth.

This is one of the most neglected questions in macroeconomics and the theory of economic growth, and these days our understanding of the world, and our fiscal future, may well turn on this matter.  We still don’t have a good sense of why growth fall-offs are so sharp and why they seem so hard to reverse.

In any case that excerpt is from Barry Eichengreen, Dwight H. Perkins, and Khanho Shin, From Miracle to Maturity: The Growth of the Korean Economy.

More generally, I loved reading this book, perhaps all the more because I had just returned from Korea when it arrived.  It is clearly written and full of useful information on virtually every page.  In my opinion every economist should study the Korean growth miracle, as a) it is in some ways mankind’s most impressive and least precedent-backed growth miracle, and b) it overturns or at least challenges many preconceptions about economic growth.  It is perhaps the case where government policy has been most effective in spurring economic growth.  This is now one of the go-to books on that topic.


What does 19th century experience look like?

When governments stopped handing out resources, the growth declined sharply, the demand for government action forced wars to acquire resources for government to hand out.

Louisiana Purchase was to ensure lots of land available for growth.
War of 1812 was driven by demand for land expansion, stalled by New England resistance to taxes to fund war.
Andrew Jackson wanted more land; he became a politician to overturn Jefferson's impediments to land redistribution.
Background to Civil War was government limits on growth of plantation culture.
Intercontinental railroad signed by Lincoln was to spur growth.
War with Mexico was to increase land for government to redistribute to promote growth.
Free Silver movement was demand that government restore growth, after government sponsored mining claims created surplus of silver (under laws still in effect currently in dispute as people want to mine the National Parks).

As taxes were needed to fund these government actions, and opposition to taxes was historic and deep, the growth decline must have been sharp to spur the tax and spend. Of course, taxes were cheaper than buying the land from the people who claimed ownership in free market transactions.

When anyone talks of going back to America's root I ask if they want to go to war with Canada or Mexico again to get more land for government to redistribute.

So...what does the experience of the 19th century look like?

Isn't the Korean growth miracle pretty darn similar to the Japanese growth miracle that preceded it by 20 years or so?

Not in how it started and not in how it ended, so the simplest answer is "no, it wasn't all that similar".

Uh, OK, so you don't see any parallels between the way that Japan achieved fantastic economic growth by exporting progressively more complex manufactured goods to the west and the way that Korea achieved fantastic economic growth by exporting progressively more complex manufactured goods to the west?

It seems to that the question at hand isn't "how it started" or "how it ended", but rather, how did they achieve the growth that did take place? In that respect, they seem rather similar.

Well, yes, the broad export-focused model is similar. But then every East Asian state participated in the chain of steadily moving up the supply chain and letting some other East Asian state take on the downstream low-wage roles, including Singapore and Malaysia and Thailand and Indonesia, to varying degrees of success. Some were enthusiastic about FDI, some were not, some engaged in aggressive industrial policy, some did not, and political circumstances were all different.

Sure, which is why I said "similar" and not "exactly the same in every detail". It just seems odd to describe Korea as the "least-precedented" economic miracle when you have the erstwhile colonial master Japan just a few hundred miles away providing a clear precedent. I visited South Korea in 1985 and everyone there (ordinary South Koreans, not economists or anyone like that) was acutley aware that their country was modeling Japan's export-led manufacturing success. They all expressed optimism that they would be able to beat Japan at her own game.

Oh, that's true. Still, South Korea was a bombed-out agricultural wasteland with substantial corruption and initially weak government, whereas even post-Allied-firebombing Japan had substantial industry and one of the richer countries in the world.

A hypothesis:

All four of the East Asian tiger states, and for that matter their fellow southeast-Asian Newly Industrialized Country states, inflicted varying amounts of brutality and suppression on their population during their highest growth periods.

If the winners from rapid technological change don't have to compensate the losers, and the winner's gains show up in GDP and the loser's losses don't decrease it, then GDP is liable to grow fast, in particular once the winner's incentives regarding politics are considered. If and when the losers suddenly assert themselves, growth suddenly slows. Economic incentives smooth over time, but politics cannot perform intertemporal stability tradeoffs and so it is that which prevents smooth convergence; the financial crises only play a role in that they permanently damage the ability of the establishment to keep pushing losers aside. This would explain all the HPAEs save Japan, which is complicated.

If this is true, what does it say about the Soviet Union in the 1930s and 1950s (when they did achieve remarkable growth rates)?

Well, murdering losers from change is plausibly a necessary but not a sufficient condition. Taiwan doesn't seem to be significantly better off from the great purges. Neither Indonesia. Singapore got away with public but indefinite jailing, instead of mass killings, secret police, and disappearances. You can murder people to no end: just ask Pol Pot.

I don't think there's anyone, however Hayekian, who would dispute that Stalinist economic policy is really excellent at factor mobilization. After all, the alternative to centralized coercion is rarely glorious market paradise, but instead having many local manor-lords instead; one needs to dismantle centuries of self-sustaining feudal social contract.

The Soviets didn't export primarily. Having to sell your products on an open market forces discipline, which naturally weakens the authoritarian or totalitarian hold on reality.

And since we are talking about killing people, maybe it helps if communists were killed.

nitpick: the Soviets exported a lot, it was just to fellow Warsaw Pact states rather than via open competition with existing Western companies

Or it's just catchup growth and humans are good at copying. Some cultures seem particularly good at it.

I sometimes wonder if China's capitalist pragmatism is just a lack of pride.


Good point. The Singapore miracle, it seems to me, is impossible without an extant West.

"We still don’t have a good sense of why growth fall-offs are so sharp and why they seem so hard to reverse."

If people are discounting to net present value then small changes in predicted long run effects can lead to big changes in current behavior.

It seems to me that most economic models are essentially "first order/linear" and my mathematical sense is that "first order/linear" models never exhibit the type of behavior you refer to. You need to move to second order/nonlinear models to get the type of unstable/asymmetric behavior you describe.

Given the amount of electronic data out there, if you could get your hands on it (including, for example, Walmart POS data and VISA transaction data), it should be possible to analyze the evolution of the boom and the recession virtually hour by hour.

"Assuuuuuuuuume we have 40 million Koreans."

and a can opener. :-)

The sigmoid curve is a classic shape of adaption from one steady state to another. How would we expect a change of institutional conditions to effect the economy- over time and within limits. Shifting policy in Eastern Eutope should be similar. The supernormal productivity growth of the last 2 decades since the end of the Comecon will end abruptly as income levels near the levels implied by the institutional possibilities.

Slow growth and fast falls could be rooted in network effects. While we need to build relationships before we can exploit them for good, they will be already in place to facilitate the spread of systemic stress and precipitate lukewarm situations.

There is a growing interested in the math and science of networks of networks. I expect it to be one of the hottest fields in the near future.

Comments for this post are closed