Monday, November 19, 2012

Is the Chinese Construction Industry Too Large?

China Bears have been predicting a disastrous end to the experiment in "Capitalism with Chinese Characteristics" for a number of years now. Notwithstanding these gloomy prognostications, China has so far gone from strength to new strength with each passing year (with only relatively minor setbacks along the way). Thus it has become easy to dismiss the bears. Still, if history is any guide, they'll be right eventually and so it's worth trying to understand what might be the signs of that.

I believe that the best way to think about post-Deng-Xiaoping China is that the society is engaged in a crash program to build a modern developed economy, complete with freeways, high speed rail, large cities, excellent airports, world class universities and a modern electricity grid and energy generating infrastructure. The Chinese population is over 1.3 billion people, and so, when the whole project is complete, China will be the highest population developed country by a factor of about four.

So how far along are they in this project? I think the easiest way to keep track is to look at the fraction of the Chinese population that is urban. In developed economies, 95%+ of the population lives in cities, whereas, in traditional pre-industrial societies, that fraction tops out at about 20% (and can be as little as 3-5%, depending on the society). China's urbanization fraction (according to their census) looks like this:

Here I have included my own hand-crafted notional projection as the red dashed line.  You are at liberty to critique this but it seems to me one reasonably plausible continuation of the history to date.

This suggests that the massive nationwide construction project can keep going for another decade, perhaps two, before really slowing down. When you take this perspective, many of the major China-bear arguments seem less than compelling. For example, bears like to focus on the fact that there is a fair amount of housing space being kept empty: much held as retirement investments and not rented out. To some, this suggests an overheating of the construction market. But, looked at in the long term, this may be rational behavior; as the urbanization process continues, there will be ample demand for those apartments at some point and they may well prove decent retirement vehicles for investors who may lack other good options.

Other bear arguments also are difficult to make in a way that is completely convincing. For example, Michael Pettis argues that the hallmark of an excessive boom is debt being used to fund economically unproductive projects, and he compares China to Japan in the 1980s and 1990s. But this seems to me to be making a comparison to the wrong stage in the process: China is only just crossing the 50% urbanization mark now, whereas Japan reached 50% urbanization in the late 1940s. At that point, the end of the Japanese boom was decades off.  When the crash finally started in 1990, the Japanese population was 78% urbanized.  China is urbanizing faster than Japan did, but still it seems unlikely that it will reach that level before the mid to late 2020s.

One way to think about the question "Is China getting so overheated that trouble is inevitable?" is to look at the size of the construction industry in relation to the population. Whenever full development comes in China, it seems reasonable to assume that the country will then look approximately like other developed countries.  It will grow at the 2-3% typical of other developed economies (resources permitting) because its growth will then be constrained by the same rate of innovation and productivity improvement that constrains the growth of all the other developed countries today.  The recent accelerated 8-10% growth rates are a function of being able to rapidly deploy productivity enhancing technologies like freeways, electric grids, and high speed trains that have been invented and well worked out elsewhere.

Similarly, when fully developed, China is likely to turn over its building stock and other infrastructure at rates similar to other mature developed economies.  This means that in pursuing the very rapid development model that it is, China has some risk of creating a construction sector that is much larger than it will eventually need - the industry that is required to move the population into the cities at the breakneck pace currently being achieved may be much too large for the needs of a fully mature Chinese economy. If we could demonstrate that the Chinese construction industry was in this position, then we could say with some confidence that it was going to have to eventually shrink.

This would likely be a painful process. We've recently seen in Spain and Ireland (not to mention Florida and Nevada) what happens when a building boom goes past the point of what is really needed and then has to shrink. The adjustment tends to be brutal and to cause ripple effects across the entire economy and society. In democracies, the process almost invariably involves throwing out the incumbent government. In an autocratic society like China, a major construction depression could pose a risk to the entire political system since it doesn't have a ready-made way to accommodate sufficiently severe dissatisfaction amongst the citizenry.

Cross-country statistics that allow us to assess this issue are not over-abundant. So far, I have found two. The first is cement production (available from the USGS for a number of countries). Cement is the critical ingredient in concrete (being mixed with sand and gravel in roughly fixed proportions) and thus cement consumption is a rough proxy for overall construction activity. Because cement is heavy and doesn't store well (being damaged by moisture), it tends to be manufactured relatively close to where it's used and, while there is some trade, in large economies like the US, China, and Europe, most cement is domestically produced. Thus cement production is also a decent proxy for construction activity (if even a little rougher than consumption data would be, where I able to find any).

This next chart shows cement production/capita in China and a selection of developed economies:

There is some cause for concern here. Mature developed economies (the US, France, Germany, Japan, Italy) seem to use 0.3-0.6 metric tons of cement per head of population per year. Spain, which had a major euro-driven construction boom in the early 2000s, reached a peak of 1.2 tons/capita/yr in 2005-2006, before entering an ongoing crash which has devastated the Spanish economy (with 25% unemployment for example). Korea (and at an earlier stage, Japan) reached higher levels of cement production during their booms but have been sliding since (I don't currently have access to stats for this prior to 1994).

Notwithstanding its huge population, China has now reached a level of cement production that is far in excess of the developed economies at around 1.5 tons/capita/year in 2011.  Growth in this indicator looks as follows:

If these kinds of growth rates continue, it seems that this indicator will increasingly soar to altitudes from which it's bound to crash. Presumably, the larger the construction industry gets, relative to its eventually required size, the worse the inevitable adjustment problems are going to be.

A less alarming picture emerges by looking at housing statistics. For a small number of countries, the UN maintains comparative statistics on the amount of housing square footage added each year. Looking at those (again on a per-capita basis) we get this

Thus the rate at which China is adding housing, specifically, is not yet out of line with developed countries, and indeed is still only around half the values that Korea has been achieving in recent years. This suggests that serious unsustainability in housing production is still a number of years off.

So how then are we to interpret the cement production statistics? One possibility is that Chinese buildings have tended to be more cement intensive than buildings in other economies - but it's not immediately obvious why this would be. Perhaps a more likely possibility is that the Chinese construction industry at present is weighted toward heavy construction involving a great deal of concrete: dams, freeways, canals, bridges, and the like. Perhaps it is this sector of the economy that is starting to reach beyond the size that will eventually be needed. Further research (and better data) is required to shed more light on this question.

Even if some aspects of the Chinese construction industry have now reached a size that cannot be sustained, that does not in itself tell us too much about the timing of the retrenchment. In general, it's often possible to keep a boom going for a number of years past the point of maximum wisdom; I'm fond of my corollaries to Stein's Law, which states that "If something cannot go on forever, it will stop."  Those are:
  • It will go on a lot longer than we think 
  • It will end badly when it does stop.
In China in particular, there is a considerable amount of useful infrastructure that could still be built, suggesting that perhaps the construction industry needn't shrink just yet. However, it's worth noting also that there has been a material slowdown in China's economy in 2011 and 2012 and it's possible there is an effective shift beginning to a lower rate of growth in heavy construction with policy-makers wanting to shift to developing other sectors of the economy.

Looking at the cement statistics, that would likely be a good idea.


A Quaker in a Strange Land said...

Well done, Stuart, and well explained.

A Quaker in a Strange Land said...

There is also a significant "Catch-22" developing. Europe is China's largest trading partner. The increase in China's Oil imports seems to be coming directly from Europe's former allotment of the World's exported Oil. Total oil imports/consumption in Spain, Italy, Greece, Portugal et al are off and it is pretty clear where that Oil is now going. I think a reasonable argument can be made about Europe's demand falling with its Oil imports (among other things). As China continues its development of productive infrastructure its internal demand will continue to rise - at the expense of its largest customer, Europe.

Ergo, at some point in the not to distant future either China's domestic demand will fill Europe's current role, or there will be a serious contraction in both China's manufacturing AND construction.

Stuart Staniford said...

One thing I started wondering about last night is whether the very accelerated rate of China's urbanization is going to run into a novel constraint which is the age of potential migrants. I would assume that in processes of urbanization generally, it's young people that move to the city: older peasants seem likely to lack the skills/education to take up a job in the city (and the one child policy will make it hard for their not very numerous grandkids to take up the slack and support them). Even in the very late stages of US urbanization you see this pattern where the average age of farmers is high and has been rising for decades. But a good half of China's urbanization is going to happen in only 40 years - half a lifetime. Thus I wonder if China might run out of young potential migrants some time before it runs out of rural population altogether. This wouldn't really have happened in the rather slower urbanization processes that other countries have followed.

Alexander Ac said...

Ok Stuart,

and what about this?

World Resources Institute identifies 1,200 coal plants in planning across 59 countries, with about three-quarters in China and India

Any advice for climate activists?


Stuart Staniford said...

Alex: Weep?

Seriously, I think western climate activists will be most effective by focussing on making our own societies carbon neutral. We will have very limited ability to influence China and India if we aren't clearly setting a decent example (and particularly in the US, we've mostly been setting a terrible example so far). Developing countries are always going to tend to be late adopters. "Think globally but act locally" is probably very good advice here.

sunbeam said...

I have to wonder what effect automation is going to have on the Chinese economy.

Anonymous said...

Stuart -

I think you're dead on in your second half analysis that we're seeing a slightly skewed number on cement production due to the Chinese infrastructure build out that is accompanying the residential construction boom. I'm not sure where you could get the historical figures on international cement production, but it seems that it would make more sense to look at the production numbers per capita during periods of intense infrastructure construction (i.e. U.S. 1956 Interstate Highway Act, Spain's autopistas/autovias system from the 1960's, etc). Second, I wouldn't discount the nature of Chinese construction as being more cement heavy then other economies. Housing demand in the U.S. is heavily weighted towards single family detached homes (~60% of the market historically), meanwhile Chinese housing demand is driven by multi-unit apartment buildings where most of the urbanizing population can only afford minimal housing space (and developers have higher margins). The latter is concrete intensive while the former uses very little apart from the slab.

James said...

I have found Chovanec's multipart blog postings about China's housing situation, part 1 here:

- He is an economist
- He actually lives in China
- He promised five posts on the subject, has only posted three, and has mysteriously stalled. I wonder if he is rethinking his thesis.

Interesting nonetheless.

Anonymous said...

Re: China and Japan.

China's population is aging at a greater rate now than Japan's was in the 40's. So while from the stage-of-urbanization factor it might look like China has several decades of high construction growth to go, from a demographic standpoint, China right now is probably around where Japan was in the 80's or even 90's. So Michael Pettis is probably going to be more right than wrong. Elderly people aren't going to move en masse to urban areas as much as younger people are.

Don Johnson said...

Great post as usual. Unusually, as a planner living in China I may have something to add here.

The potential housing market is always the bedrock argument of the optimists - XX million new urban residents coming down the pike, you can't go wrong! But I think the argument of the bears is not that too much housing is being built, but that the wrong kind of housing (i.e. luxury flats) is being built and that a grotesquely large proportion of the Chinese economy, especially the banking system and local governmental revenues, depends on it continuing to get built and to rise in price beyond any possible market's ability to pay for it. So you are right that "there will be ample demand for those apartments at some point." The problem is that the road from here to there might be paved with the wreckage of developers, banks, local government finances, and everyone else who has invested in the housing market, which turns out to be practically everyone.

Regarding the efficiency of investment in infrastructure, it's essential to remember that China has not developed evenly. The coast, especially around Shanghai and other cities, is very close to developed-world levels of per capita GDP and in many cases now has better (because all brand new) infrastructure. Whereas the west, northeast and southwest are a decade or two behind, and this makes a big difference as to the ability of new infrastructure to raise productivity and boost the economy.

Finally, China is more concrete heavy than the US for both reasons you mention - they are at a more concrete-intensive stage of development (the high speed rail system alone uses mind-bending amounts of concrete) and Chinese residential construction is essentially all concrete, no wood or steel at all. Even in commercial structures, steel is much less common than in the US.

Sorry for the lengthy comment. My personal view is that an actual crash brought on by some combination of housing crisis and mis-spent investment is unlikely -
the central government has the financial resources and discipline to do whatever is necessary to avoid a mess. But a series of expensive bailouts that end up shaving several points off of growth for a long time is very plausible, and also quite bad.

James said...

Unknown - thank you very much for that lengthy comment. You sound very well informed to my ears.

Unknown said...

I read your recent post on peak oil, not clear we will quickly fall hard. however, even plateau may be hard for china to continue rapid growth. Exporters are exporting less, Europe and US have so far stepped aside to allow more for china, but it takes higher prices to persuade consumers to step aside. Not a problem per se for chinese, but this implies at best slow growth and maybe perpetual recession in the west...

Pettis point is that china must soon move to internal consumption, and that this change also implies much reduced investment along with sharply reduced growth. Corruption, pollution, reduced exports, ever higher cost of oil, capital flight, plus no ability to change government to vent resentment... consider that Japan did not have anything like this combination of problems until their housing/equity bubbles popped.

You make a good point that they may eventually need more housing, but IMO other problems will soon slash growth.