China's investment conundrum: IMF misses the point of policy?

There is huge debate around the sustainability of China's growth model.  In the bullish camp, you have those that point out that China's growth potential is still high given a high proportion of the population is yet to urbanise, while the capital stock is still low and technology has a long way to improve.

One popular way of showing this from a commodity perspective is in terms of consumption per capita vs GDP per capita, which is broadly following the development path of other more developed economies at a given level of GDP per capita and still has someway to go.

The bearish camp believe that growth rates expected for the next few years are unsustainable given the unbalanced nature of GDP.  Investment accounts for a huge proportion of output and is a sign of misallocation of resources.  Current levels of investment, let alone growth, are not sustainable and a painful adjustment will occur.

A recent paper from the IMF titled "Is China Over-Investing and Does it Matter?" tackles this debate and can be found by clicking here. In a nutshell, they tend towards the bearish camp, although we think their analysis perhaps over accentuates the degree of over-investment rates.

To derive their results, they use a panel of data from 36 emerging markets and estimate a model of investment/GDP using variables such as real GDP per capita, credit as a percent of GDP, age dependancy ratios and an "uncertainty" variable in which they use standard deviations of output.  The end result is summarised in the chart below, where actual investment ratios in China are way above estimates based on where other countries have been before.

One key issue with this approach for me is the impact of the "uncertainty" term, which is the standard deviation of 3-year rolling GDP.  The model suggests that if GDP falls relatively sharply, so should investment, which intuitively makes sense.

But does rolling loses in output really capture "uncertainty"?  Furthermore, does it make sense that investment must fall if policy makers can ensure otherwise?  In China's case, investment received a huge boost in order to offset the impact of uncertainty on other parts of the economy, rather than slide the same way. Prior to this policy driven boost, investment was somewhat higher than modelled, but this has been the case for the time period modelled.

This suggests the recent deviation, which is most apparent relative to the history of modelled outcomes, is a cyclical policy issue rather than a structural one. Why more recent data weren't included in the analysis is strange, as it would provide a better picture of whether the divergence is still as large as the 2009 data.

The last few years hasn't seen a fall in this ratio and investment is probably unsustainably too high. But I think this paper misses the point of why it was high in the first place. A policy induced surge in investment was in all likelihood a better alternative than a severe recession.  And at a more basic level, if you don't understand why it was high, it is difficult to understand how it will adjust and how far it has to go.

To be sure, there is some adjustment already occurring, with the stability of this ratio in the last few years suggesting other parts of GDP have grown enough to stop China's investment/GDP ratio rising further.

I think there has to be some kind of pain involved in reducing the level of investment, although I dont think markets or analysts are massively overly optimistic about potential growth rates as a consequence.  As the IMF paper points out, China avoided the cardinal sin of unhedged external borrowing, or any external borrowing which rules out a current account crisis which usually inflicts huge short term pain.  Perhaps a slower grind lower is more likely, although how messy it could end up ultimately depends on how much pressure banks will come under.