Saturday 1 November 2014

Regression to the Mean: Chinese Economic Growth in Context

The latest study making it's rounds about Chinese economic growth is this paper by Larry Summers and Lant Pritchett: Asiaphoria Meet Regression to the Mean.

I have to admit to being underwhelmed by this paper. Much smarter people then I seem to be quite impressed like Tyler Cowen here. My reaction is fairly simple: Of course Chinese growth is unsustainably high over some undefined time horizon but what in this paper wouldn't have applied in 1990, 1995, 2000, 2005, and 2010?

Perhaps the author's intent is to be obvious and non-specific (this paper still has plenty of useful info in this context), but why has this become the new 'it' study on China? (mood-affiliation?)

The paper makes three basic claims:

1. Growth rates are typically persistent and low.
2. Developing world growth rates are highly volatile and regress back to global means aggressively.
3. The Asian giants (China and India) should include a rapid regression to the mean as a realistic possibility.

None of which I really disagree with. But in the conclusion there is this:
India and even more so China are into essentially historically unprecedented episodes of growth. China’s super-rapid growth has already lasted three times longer than a typical episode and is the longest ever. The ends of episodes tend to see full regression to the mean, abruptly.  
It is impossible to argue that either China or India have the kinds of “quality institutions” that have been associated with the steady dynamic of growth in the currently high productivity countries. The risks of “sudden stops” are much higher with weak institutions and organizations for policy implementation. China and India have very different modalities of this risk, but both have tricky paths to continued prosperity. 

Again, it's not that I disagree with any of this. In fact, I would say that these are almost self evident truths. But my issue is that this paper doesn't wrestle with whether China's "historically unprecedented" episode of growth does in fact mean that 'this time is different'. I believe that the author's would agree that progression towards the "quality institutions" found in successfully transitioned countries promotes growth, and that neither China nor India posses sufficiently developed institutions, would they not then conclude that this represents an opportunity to avoid said regression?

It seems to me that, while this paper describes the rate of change of GDP (and by proxy development) it doesn't give the level of GDP (and of course by proxy the level of development).

And this isn't me talking about how wise and efficient autocracy can be. To me, the greatest contribution to these unparalleled growth rates is just how effectively the CCP pre-Deng destroyed the economy. Talking about levels seems paramount. And I think GDP per capita adds that context to the discussion.

When you think of the most relevant cultural/economic performance parallels to China, you probably think Japan, Taiwan, and Korea. China, of course, is unique in it's sheer size. But what is lost on people is it's uniqueness in it's absolute poverty. China was one of the poorest countries in the world when this growth phase started and has merely (and miraculously) risen to near middle income territory. But where were the rest of these countries in per capita terms when they "regressed to the mean"?

Japan '70: $15,162 (regressed to: 3.4% growth)
Korea '91: $9,591 (regressed to: 4.42% growth)
Taiwan '94: $11,932 (regressed to 3.48%)

How about the Per Capita GDP at the start of these phases?

Japan '59: $7,752
Korea '62: $1,123
Taiwan '62: $2,263

By Contrast corresponding Per Capita Chinese GDP:

China '77: $178
China '91: $495
China '13: $3,583

Per capita GDP Data World Bank WDI 2013 Data Set. All numbers reported in constant 2005 USD. Growth numbers from S&P paper.

So here's an assumption that I'll make and not provide adequate data to cover: moving away from centralized autocratic control of a country/economy produces favourable economic results with diminishing marginal returns. Think of wealth enhancing policy as a continuum; at 0 you have North Korea and at 10 you have, well an argument, but lets just say above 7 you have a large group of market based economies with some mixture of liberal democratic institutions. This of course excludes countries abnormally dependent on exporting natural resources (I'm looking at you Norway).

Now in an unscientific approach that assumes my under-supported assumption, what would best proxy a countries location along that continuum? In my opinion it's the per capita wealth or GDP measurement. In a manner as simple as Summers and Pritchett (S&P) arguing that countries regress back to the GDP growth mean. Therefore, China starting at say a 0 or 1 on the 'wealth enhancing policy continuum' might be at a 3 or 4 now. Countries that successfully regressed to the mean maybe did so around 6 or 7. Of course this isn't rigorous. But intuitively I think it's not a bad framework to use.

I'll try to come back to this at a later date and try to flush things out at a later date in more detail because I think it's often overlooked, particularly in China and India's case because the aggregate size of their economies are so much larger then any other country (at their stage in development has ever been). These countries, on per capita terms, are very, VERY poor (India in '13 had per capita GDP of $1,165). If per capita is any sort of proxy of how much "low hanging fruit" is available to propel growth, then the slow move along the development continuum results in the possibility of sustained abnormal growth levels (which China has already demonstrated). Judging by the many headlines you ready about the need for/lack of/inability to produce reform in China, I have to assume it's fairly well accepted that the long hanging fruit is out there and fairly obvious.

I think China is doing a good job of sustainably harvesting that low hanging fruit. And their ability to do so sustainably in an unprecedented manner over the past three decades should give experts a bit more pause when making recommendations with any confidence. Everyone agrees on the direction of reform. The speed and the end result (just because the CPC wouldn't want to hit a 10 on the democracy index, it doesn't meant they don't want to go from a 3 to a 5) are a debatable, as the CPC is in rarified economic growth air, shouldn't we consider that they aren't idiots and might have some insight?

This isn't saying that China WILL continue growing rapidly. As S&P allude to in their paper, developing countries growth rates are extremely variable. I interpret this as: though there are a much greater quantity of growth promoting possibilities, there is a certain symmetry in growth destabilizing risks. This of course comes from a guy who is endlessly fascinated and impressed with the abnormally amazing economic performance of China (and to a lesser extent India) described in this paper by S & P. While the current regime inherited a low base to grow from, they've been able to avoid that highly variable growth rate and the expected regression to the mean.

Let's hope that L & S can repeat this same paper in another 5 years time.

No comments:

Post a Comment