How do we know that China is overinvesting?

By Michael Pettis

For years I have been arguing that the Achilles heel of the Chinese growth model is the unsustainable rise in debt that comes as a necessary consequence of capital misallocation fueled by bank lending. Capital misallocation, I argued, was the nearly inevitable consequence of high investment growth over many years in a system in which price signals are severely distorted and there is political incentive to maximize economic activity in the near term. If capital misallocation is funded by debt, the increase in debt is necessarily unsustainable.

These should have never been considered surprising revelations since the historical precedents for investment-led growth “miracles”, of which there are many, are pretty clear. Still, it was only in the past two or three years that the problem of wasted investment was widely acknowledged, although even here not universally. A number of China bulls that fought most strongly several years ago against the idea that China was misallocating capital on a grand scale are still fighting the good fight.

I mention this because of an article that came out in last Friday’s South China Morning Post about China’s investment in the electric car industry. The electric car industry was often Exhibit A in the argument that Chinese investment was in the aggregate rational and economically sensible. This industry is clearly the industry of the future, the China bulls argued, and China’s massive investment in the technology, which would allow the country to dominate one of the key sectors of the future, showed why it was mistaken to complain about capital misallocation. This kind of investment was actually very clever stuff.

I have to confess I was never comfortable with this argument, although criticizing investment in the electric car industry was always a little like punching a six-year-old child. It is really hard to do it without looking like a heel, and inevitably someone points out that since the electric car industry is clearly the industry of the future, and since China is getting an early lead, it follows that this is a very smart investment for China.

But remember, even if the first two points turn out to be true (and if it is so obvious, why doesn’t everyone else do it?), the third doesn’t obviously follow. The point is not whether or not electric cars will one day be an important business, and certainly not whether electric cars are a “good thing”. What matters as far as this debate is concerned is

1) Whether the total economic costs of investment are less than the total economic benefits, and

2) Whether there is a mismatch in the timing of costs and benefits.

The first point determines whether the investment is ever wealth enhancing, and the second determines whether or not in the medium term there is a worsening of the underlying imbalance. This means that in discussing whether or not capital is being misallocated it is not enough to assert that electric cars will one day be an important technology.

Electric cars?

My problem with Chinese investment in the electric car industry had nothing to do with my superior knowledge of the prospects for the industry (I have none). It had mostly to do with my basic instinct that risky high-technology ventures are not best funded and directed by companies, industries and policymakers who are historically weak in the technology sector, especially when they have no shareholder or budget constraints and have almost unlimited access to heavily-subsidized capital. This seemed to me a recipe for wasted investment.

This is why the SCMP article interested me:

Beijing is having a serious rethink over its ambitious electric vehicles policy. The central government has assembled the top experts and policymakers on electric and hybrid cars at a meeting in Wuzhen, in Zhejiang province, this week. One expert attending the meeting said the government was increasingly concerned about problems in the industry, including its cost-effectiveness, technological difficulties and the uncertain benefits to the environment, according to the 21st Century Business Herald.

It said Beijing was reconsidering its support for pure electric cars and may rethink how to spend the 100 billion yuan (HK$122.6 billion) fund set up to develop green vehicles – perhaps by shifting resources to hybrid cars.

Beijing is soon to release the final version of its long-awaited green car development plan, which has been under drafting by the Ministry of Industry and Information Technology since 2009. But there is now growing debate as to how to make best use of funding set aside to help develop the next generation of cars.

The central government announced its ambitious plan to develop a new generation of green vehicles – focusing on pure electric cars – two years ago. However, in July, Premier Wen Jiabao wrote a long article in the Communist Party’s mouthpiece Qiushi magazine in which he said he was confused by the latest developments and the future of electric cars. He also criticised the lack of co-ordination and planning of local authorities and warned against committing resources to premature technologies. “Whether electric cars will be a mature product, we don’t know,” he wrote.

A few days after this article appeared, BYD, the main maker of electric car in China saw its stock soar by 26% in a day. Here is what an article in South China Morning Post said was the reason:

Shares in the Warren Buffett-backed BYD soared 26 per cent to a three-month high yesterday on hopes that new policies to steer mainland car buyers towards electric vehicles would help turn around slumping sales at the Shenzhen-based manufacturer of petrol and electric cars.

The rally came after the weekend publication of a directive signed by four government ministries encouraging 25 pilot cities, including major markets such as Beijing and Shanghai, to “actively study” exemptions for electric cars from license plate lotteries and auctions, as well as a host of other purchase restrictions.

The only way to make electric cars economically viable in China, in other words, is to put into place administrative measures that divert buyers, but as any economics student can tell you, these kinds of administrative measures simply shift resources from one sector of the economy to another without creating wealth. In fact because they force consumers to choose something that they otherwise wouldn’t, they actually reduce overall wealth.

It is hard for me to be very optimistic that the success of electric cars in China will increase the wealth of the Chinese people. It seems to me that investment in high-prestige areas like electric cars, solar panels, and so on for technologically backward countries with low worker productivity may be a little like investment in the space program or in the Olympics. They may have positive political and even psychological returns, but on a strictly economic basis they reduce overall wealth and exacerbate domestic imbalances.

In that case for all the importance the industry might have one day, debt levels in China must continue to rise unsustainably. The key to determining their economic impact of any investment must be the two points that I posit above.

Is capital is being wasted?

First, the additional wealth that it creates for the economy must exceed the debt servicing cost to the economy of the investment, and of course the debt servicing cost must be the “true” debt servicing cost, in which interest rates are implicitly raised to an economically justified rate, and not the heavily subsidized rate that simply transfers part of the cost to the household sector. If it does not, the domestic imbalances are exacerbated and debt rises unsustainably.

Second, even if over the long term the benefits justify the cost, if they do not do so over the short and medium term, then the imbalances are exacerbated for some period before they are resolved. In and of itself this shouldn’t matter if we are concerned about long-term growth prospects, but in the case of an economy with serious domestic imbalances and the risk that these imbalances may derail the economy, it will matter. To get to the long-term you have to go through the short and medium term, and disruptions in the short and medium term may eliminate the longer-term benefits.

Of course the question of whether or not China is misallocating capital can be endlessly debated because it is very hard to prove except in retrospect. I would argue that there are several reasons why we should believe that capital has been wasted on a large scale for many years. The first reason is simply historical precedents, something which unfortunately rarely enters into most economic analysis. No country in history that has had anywhere near the growth in investment as China has not had a serious problem in subsequent years, in which debt rose to crisis levels and growth ground to a stop.

The fact that China is so poor is often proposed as an argument as to why this cannot also be the case for China, but of course this is a nonsensical argument. Poorer countries with lower levels of worker productivity are less able, not more able, to absorb very high levels of investment. This may seem counterintuitive at first, but only if you believe that there is a single optimal level of investment for every country regardless of its specific conditions. If the purpose of investment is to save labor and labor cost, then it should be clear that the lower the level of worker productivity and the cheaper and more abundant the amount of labor, the less investment in capital stock is justified.

This is why when so many analysts compare the per capita capital stock of China with that of the US or Japan, and then announce that this proves China has a long ways to go before it runs out of investment opportunities, I am always surprised, and even a little skeptical about their economic backgrounds. This comparison simply does not make sense.

If it did, overinvestment crises would be largely limited to rich countries, not poor countries – something that is certainly not confirmed by history. Anyway I find bizarre the idea that the best comparison for China, one of the poorest countries in the world even if you accept the validity of GDP numbers and ignore the very low GDP share of household income, is the US or Japan, two of the richest and most technologically advanced countries in the world.

But I think there are more formal reasons to believe that China is misallocating capital. Common sense suggests that when there is massive investment with

• very little accountability,
• severely distorted prices,
• an incentive structure that concentrates the benefits of investment in specific jurisdictions and over a short time period while spreading the costs throughout the national banking system and over the debt repayment period (which can be decades),
• no or very limited budget constraints,
• factional and regional conflicts, and
• shifts in responsibility as the instigators of the investment are promoted (often because of the positive impact of their own investment initiatives),

it would be a rare system in history that did not tend towards substantial capital misallocation.

Certainly the evidence on SOE investment suggests that this is indeed what happened. A number of studies have suggested that if over the past decade you add up direct subsidies, the impact of monopoly pricing (which is of course simply a tax on households) and the interest rate subsidy, they total anywhere from six to ten times the aggregate profitability of the SOE sector. This means that unless the externalities associated with the SOEs are also at least six to ten times their aggregate profitability, they are actually value destroyers.

The impact on ratios

This doesn’t prove that the same must be the case for infrastructure investment, real estate investment, and R&D initiatives, but the only relevant difference between SOEs and infrastructure, real estate and R&D investment, it seems to me, is that in the former case there is at least some way of accounting for results, whereas in the latter there isn’t. And since in the former case the results are not encouraging, what are the conditions that limit wasteful investment in the latter? There may be some, but I don’t know what they are.

But there are other ways to try to judge if there is likely to be massive capital misallocation. We can deduce whether or not capital is being misallocated by determining what impact large-scale misallocation might have on various ratios. We can then look to see if this is indeed happening.

I would argue that if capital is being misallocated on a substantial scale over many years at least three things should happen. First, GDP growth per unit of investment should decline sharply. Why? Because the costs of misallocation are pushed into the future as debt is serviced. As these costs accumulate, they should reduce future GDP growth, and so as we get to that future we would need rapidly rising investment to generate the same amount of growth.

This certainly seems to be the case in China, especially over the past three years, during which time investment growth rates surged while GDP growth rates stayed more or less constant. There are three complications with this argument, however.

• It is not just investment that matters, especially for China, where the external account matters a great deal, so the relationship between rising investment per unit of GDP growth will have been severely understated in the 2002-07 period, when the current account surplus surged, and severely overstated in the 2008-10 period, when it declined.
• It is natural for the growth impact of rising investment to decline even if capital is not being misallocated, since the marginal benefit of additional investment will tend to fall anyway. The fact that this is happening in China proves nothing. Only the fact that it is happening much more quickly than in other cases would suggest that capital has been misallocated.

• If capital has been misallocated for many years, then GDP growth numbers in recent years are overstated, which places us in a circular position. The declining impact of investment on growth is greater if we already know that investment has been misallocated.

The second ratio that would be affected by substantial misallocated capital is the debt level, and if capital is being misallocated debt should be rising at an unsustainable pace. Why? Because when liabilities rise faster than assets, by definition the increase in debt is unsustainable, and if debt-funded capital is being wasted, by definition the value of assets is rising more slowly than the value of debt.

Is debt in China rising at an unsustainable pace? We can’t prove it one way or the other, but I think most analysts would agree that debt is rising incredibly quickly. It may very well be that this is sustainable, but then it seems to me that this would imply an acceleration of growth in recent years, something we haven’t seen. If the value of total assets is rising at anywhere near the speed of debt, in other words, it would seem to me that we have reached a point in which the growth impact of investment is higher than it has ever been in the past thirty years. This is unlikely. Debt almost certainly is rising faster than GDP or any other measure of the capacity to repay.

Third, if capital is being wasted, and the resulting losses are not showing up in the accounting for the investment, then by definition the losses must have been absorbed by some other sector of the economy. In that case we should see their share of total wealth decline sharply as they are forced to cover the huge waste in investment.

Creating value

Is it possible to identify this sector in China and to judge whether or not this is happening? Of course it is. The sector that is responsible for absorbing the costs – transferred in the form of low interest rates, an undervalued currency, low wage growth relative to productivity growth, monopoly pricing, environmental degradation, and so on – is the household sector. Has their share of total wealth declined? Of course it has, especially in the past decade, when it has all but collapsed.

Now notice that with the exception of the studies suggesting that SOEs misallocate capital, this is all circumstantial evidence. It is nonetheless pretty powerful circumstantial evidence, and I think it strongly confirms what both common sense and historical precedents suggest. It doesn’t prove that capital is being misallocated in China, but it certainly suggests that anyone who isn’t at least seriously worried about this possibility is pretty credulous.

To go to the other extreme, I also worry about investment in social housing, which many China bulls insist thoroughly undermines the argument that China is wasting investment. I realize that once again I am going to make an argument that may at first seem very controversial, but I suspect that as in some of the other “controversial” arguments, within a year or two this will be more widely accepted. I say this because although it seems at first counterintuitive, the logic to me is quite compelling.

Is creating more social housing in China a bad idea? To suggest that it may be wasteful sounds even more of a heel’s argument than to suggest that investing in electric cars is wasteful. After all China certainly needs social housing, so how can it be wasteful?

The answer, of course, depends on what we mean by wasteful. There is little question in my mind that lack of housing for the poor in China is a big problem, but we have to separate social, political and economic benefits. In order to understand the impact of a significant increase in investment in social housing on China’s rebalancing and debt levels, only the economic impact matters. I don’t doubt that there are huge social benefits to providing the poor with adequate housing, and there are also important political benefits to doing so, but are there economic benefits?

It depends. If the economic benefits of providing cheap housing exceed the economic costs, then a surge in social housing investment is good for Chinese rebalancing, good for Chinese wealth creation, and good for Chinese debt servicing ability. If not, it isn’t.

The economic costs are easy to understand. They are the cost of servicing the debt associated with the investment at whatever the appropriate interest rate would be. What is that rate? Remember that if the nominal interest rate is lower than the nominal GDP growth rate, net lenders (households depositors) lose relative to net borrowers, and the imbalances in the Chinese economy get worse.

This suggests that the correct debt servicing cost should be equal to the nominal GDP growth rate if we are concerned about rebalancing. Chinese nominal growth rates are around 14-15%, but I would argue that since GDP growth is probably overstated if I am right about wasted investment, perhaps we can adjust the appropriate interest downward to 11-12%.

If we are indifferent to domestic imbalances and are only interested in wealth creation, then the “correct” debt servicing cost could be lower, but it would have to be equal to inflation, plus the cost of liquidity, plus some fairly high risk premium to account for the volatility around expected outcomes in China’s economy. Of course we don’t know what inflation in China is. CPI inflation last year was 5-6%, but most analysts believe it understates real inflation. The GDP deflator is 9-10%.

Economic value of social housing

Where does that leave the appropriate debt servicing cost? I don’t really know, but probably still pretty close to 11-12% even if you don’t care about domestic rebalancing, and a hell of a lot higher than the 6-7% (or lower) at which local governments are likely to be able to borrow from the banks.

So much for the costs. What are the economic benefits of social housing? There are, I would argue, two different components. First, if social housing improves the productivity of labor, perhaps by giving better access to work and healthier living conditions, then this improvement in productivity is part of the economic value generated.

Second, families that receive social housing have had an increase in their wealth (by lowering their expected housing costs). This should boost their consumption and so create an additional source of demand that will drive economic growth.

If the combination of these two things – the annual increase in worker productivity and the additional annual spending that social housing releases into the economy – is greater than annual debt servicing cost, then the expected surge in social housing investment in the next few years will help rebalance the economy and will improve China’s ability to service its debt. If not, then it won’t.

I have ignored one other factor, and that is the income redistribution impact. To the extent that spending on social housing redistributes income from the average Chinese to the poorer Chinese, it should increase consumption somewhat, since the poor consume a larger share of income than the rich. But there is a flip side to this. The poorer the recipients of social housing are, the less they will contribute to economic growth and rebalancing because the lower their productivity and the less likely the consumption released by moving them into housing will justify the cost of the housing.

The point of all of this is not to say that social housing is a bad investment. I am arguing, instead, that when analysts say that China can keep growing healthily at elevated rates for many more years because policymakers are planning to invest in social housing, they may be completely wrong because they misunderstand the impact of the investment.

We shouldn’t assume that social housing expenditures are really a radically new form of economic growth that will replace the old form of growth driven by misallocated capital. It could very well be the same kind of growth, in which case we are still stuck with the problem of worsening domestic imbalances (and so a declining consumption share of GDP and more reliance on exports and investment), and unsustainable increases in debt. We have, in other words, resolved nothing as far as the underlying economic problem is concerned.

Analysts too easily argue that investment in social housing is China’s trump card that will guarantee sustainable annual growth of 8-9% or more for the next several years. It isn’t. First, there is almost no way social housing investment will be large enough to replace investment in manufacturing capacity, other real estate development, and infrastructure, and if there is a problem with the latter, it will continue. Second, investment in social housing may itself be economically wasteful. Certainly there is no reason simply to assume that, given the conditions that have encouraged capital misallocation for over a decade, calling the investment “social housing” will change anything.

This is an abbreviated version of the newsletter that went out two weeks ago. Academics, journalists, and government and NGO officials who want to subscribe to the newsletter should write to me at [email protected], stating your affiliation, please. Investors who want to buy a subscription should write to me, also at that address.