Authored by Michael Pettis, originally posted at China Financial Markets,
The IMF’s Il Houng Lee, Murtaza Syed, and Liu Xueyan have published a very interesting and widely noticed study called “Is China Over-Investing and Does it Matter?” In it they argue that there is strong evidence that China is overinvesting significantly. According to the abstract:
Now close to 50 percent of GDP, this paper assesses the appropriateness of China’s current investment levels. It finds that China’s capital-to-output ratio is within the range of other emerging markets, but its economic growth rates stand out, partly due to a surge in investment over the last decade. Moreover, its investment is significantly higher than suggested by cross-country panel estimation.
This deviation has been accumulating over the last decade, and at nearly 10 percent of GDP is now larger and more persistent than experienced by other Asian economies leading up to the Asian crisis. However, because its investment is predominantly financed by domestic savings, a crisis appears unlikely when assessed against dependency on external funding. But this does not mean that the cost is absent. Rather, it is distributed to other sectors of the economy through a hidden transfer of resources, estimated at an average of 4 percent of GDP per year.
The article is well worth reading because it makes a very strong case, perhaps a little late, for what many of us have been arguing for the past seven or eight years. China’s investment rate is so high, we have argued, that even ignoring the tremendous evidence of misallocated investment, unless we can confidently propose that Beijing has uncovered a secret formula that allows it (and the tens of thousands of minor government officials and SOE heads who can unleash investment without much oversight) to identify high quality investment in a way that no other country in history has been able, there is likely to be a systematic tendency to wasted investment.
Interestingly enough, while two of the authors of the study work for the IMF, Liu Xueyan, the third, is a Senior Fellow in the Institute of Economic Research at the National Development and Reform Commission (NDRC) of China. I don’t know how much we should read into this, but it is worth noting that both the World Bank report in March and this IMF study have involved input from important mainland think tanks.
This is noteworthy because both the World Bank report and this study have come out very strongly in the direction that the China “skeptics” have been arguing for many years. They identify the urgent need for adjustment and suggest – very delicately – how difficult it will be. I assume that this is all part of the tough debate that is taking place within policy-making circles over the need to implement the very difficult political reforms that will be a necessary part of the economic rebalancing, and I guess the reformers are eager to recruit the World Bank and the IMF to their points of view.
How much overinvestment?
One of the implications of the study is that households and SMEs have been forced to subsidize growth at a cost to them of well over 4% of GDP annually. My own back-of-the-envelope calculations suggest that the cost to households is actually 5-8% of GDP – perhaps because I also include the implicit subsidy to recapitalize the banks in the form of the excess spread between the lending and deposit rates – but certainly I agree with the IMF study that this has been a massive transfer to subsidize growth.
This subsidy also explains most of the collapse in the household share of GDP over the past twelve years. With household income only 50% of GDP, a transfer every year of 4% of GDP requires ferocious growth in household income for it just to keep pace with GDP, something it has never done until, possibly, this year.
The size of the transfer makes it very clear that without eliminating this subsidy – which basically means abandoning the growth model – it will be almost impossible to get the household and consumption shares of GDP to rise if China still hopes to maintain high GDP growth. The transfer of wealth from the household sector to maintain high levels of investment is simply too great, and this will be made all the more clear as the growth impact per unit of investment declines.
Another implication of the IMF study is that to get into line with other equivalent countries at this stage of its economic takeoff, China would have to reduce the investment share of GDP by at least ten percentage points and perhaps as much as twenty. Aside from pointing out that the sectors of the economy that have benefitted from such extraordinarily high investments are unlikely to celebrate such a finding, I have three comments. First, after many years in which China has invested far more than other countries at its stage of development, one could presumably argue that in order to get back to the “correct” ratio, investment should be lower than the peer group, not equal to the peer group. In that case investment has to drop by a lot more than ten percentage points.
After all if China’s deviation from the experience of other countries is meaningful, then after a few years of substantial deviation, it cannot be enough for China simply to return to the mean. It must come in lower than the mean for a few years so that on average the deviation is eliminated.
Second, even if China had kept investment at the “correct” level, as measured by the peer group, this would not imply that China has not overinvested. I haven’t been able to dig deeply into the comparison countries, but the study does list them, and a very quick glance suggests that many of these countries, after years of very high investment, themselves experienced deep crises or “lost decades”.
This implies to me that these countries themselves overinvested, and so even if Chinese investment levels were not much higher than that of the peer group (and it was mainly in the past decade that Chinese investment rose to much higher levels than that of the peer group, and not in the 1990s, exactly as we have been suggesting using more qualitative measures), this could nonetheless be worrying. China would still have a difficult adjustment for the same reasons that many if not most of the peer group countries also had difficult adjustments.
The average number driven by the peer group sample, in other words, is not in itself an “optimal” level of investment. It might already be too high. That Chinese investment levels have been so much higher than theirs is all the more worrying.
How much would growth have to slow?
My third point is more technical. If Chinese investment levels are much higher than optimal (assuming the peer group average is indeed optimal), of course the best solution for China is immediately to reduce investment until it reaches the right level. The longer investment rates are too high, the greater the impact of losses that have eventually to be amortized, and the worse off China is likely to be.
But it will be very hard for China to bring investment down as a share of GDP by ten full percentage points very quickly. Let us assume instead that China has five years to bring investment levels down to the “correct” level, and let us assume further that the “correct” level is indeed ten percentage points below where it is today. Both assumptions are, I think, dangerous because I am not convinced that an investment level of 40% of GDP is the “correct” level for China going forward (I think it must be much lower) and I don’t think China has five years to make the necessary adjustment without running a serious risk of a financial crisis.
But let us ignore both objections and give China five years to bring investment down to 40% of GDP from its current level of 50%. Chinese investment must grow at a much lower rate than GDP for this to happen. How much lower? The arithmetic is simple. It depends on what we assume GDP growth will be over the next five years, but investment has to grow by roughly 4.5 percentage points or more below the GDP growth rate for this condition to be met.
If Chinese GDP grows at 7%, in other words, Chinese investment must grow at 2.3%. If China grows at 5%, investment must grow at 0.4%. And if China grows at 3%, which is much closer to my ten-year view, investment growth must actually contract by 1.5%. Only in this way will investment drop by ten percentage points as a share of GDP in the next five years.
The conclusion should be obvious, but to many analysts, especially on the sell side, it probably needs nonetheless to be spelled out. Any meaningful rebalancing in China’s extraordinary rate of overinvestment is only consistent with a very sharp reduction in the growth rate of investment, and perhaps even a contraction in investment growth.
In fact I think over the next few years China will indeed undergo a sharp contraction in investment growth, but my point here is simply to suggest that even under the most optimistic of scenarios it will be very hard to keep investment growth high. Either Beijing moves quickly to bring investment growth down sharply, or overinvestment will contribute to further financial fragility leading, ultimately, to the point where credit cannot expand quickly enough and investment will collapse anyway.
This is just arithmetic. The extent of Chinese overinvestment – even if we assume that it has not already caused significant fragility in the banking system and enormous hidden losses yet to be amortized – requires a very sharp contraction just to get back to a “normal” which, in the past, was anyway associated with difficult economic adjustments. It is hard to imagine how such a sharp contraction in investment will itself not lead to a sharp drop in GDP growth, and the IMF paper recognizes this:
To the extent that elevated levels of investment during the post-crisis period in China were somehow abnormal and necessitated by the sharp external slowdown, the challenge now is how to return to a more “normal” level of investment without compromising growth and macroeconomic stability.