(by Tracy Alloway) It's a question of more economic importance than you might think, given that it's still widely believed that China will lead the world economy into recovery. A low urbanisation ratio — the proportion of people living in urban areas — would suggest that it still has lots of growth to get through. A higher urbanisation ratio would suggest that much of the "development" in this developing country has already taken place.
Pivot Capital set off the debate with a research note that's garnered its fair share of media attention. Here's a summary of their thesis:
· China's expansion cycle surpassing historical precedents: It is widely believed that China is still in an early development phase and therefore in a position to expand capital spending for years to come. However, both in its duration and intensity, China's capital spending boom is now outstripping previous great transformation periods.
· Policy actions not sustainable into 2010. This year's burst in economic activity has been inflated by a front-loaded stimulus package and a surge in credit growth. Given their exceptional and forced nature we believe growth rates in government-driven lending and capital spending will collapse in 2010.
· Overcapacity and falling marginal returns on investment: Analysis of industrial capacity, urbanisation and infrastructure development shows that China's industrialisation and structural modernisation are largely complete. Combine this with falling returns on investment, and it becomes obvious that China's long-term investment needs are grossly overestimated.
In short, Pivot thinks China's consumption growth cannot replace the investment boom.
Here's a bit more detail on the urbanisation point:
Urbanisation is perhaps the most quoted fundamental trend that should drive China's growth for the next few decades. At a reported 45%, China's urbanisation rate is low not only compared to the rates of 70-80% for developed countries, but also to the world average of 50%. Since an increase of 10% in China's urbanisation rate means that roughly 130mn people would need to move into cities, this is obviously a very powerful argument in favour of continued high growth in Chinese construction activity. For example, if China was to reach the urbanisation rate of Russia (73%) in the next 40 years, it would need to add a city the size of New York every year.
As with many of the figures that come out of China, however, Pivot thinks there may be a problem with the official ratio:
However, as is very often the case with such comparisons, numbers are not what they seem. China's definition of an urban centre includes, amongst other things, population density of above 1,500 people per square kilometre. By that definition Western cities like Houston (2.2m people with density of 1,375/km2) or Brisbane (1.9 m people with density of 918/km2) could technically not be counted as cities. Back in China, a lot of the so-called "villages" and "townships" are in fact highly industrialised. Qiaotou, home to 64,000 people, produces 60% of the world's buttons and 70% of its zippers. Songxia with 110,000 people is the umbrella capital of the world: it produces 500mn umbrellas per year. Bordering on the edge of surreal is the story3 in the Wall Street Journal about the "village" of Shaliuhe at the outskirts of the city of Tangshan where a month before the Olympic Games, in order to reduce pollution, 26 inefficient cement factories were dynamited. Workers at the local Dafeng Steel Mill had to take an early vacation.
According to a recent OECD report, "…there are likely many more villages and undesignated towns in China that would in most OECD countries be formally designated as towns, and far more towns, both statutory and undesignated, that would be administratively defined as cities. The scale of China's urbanisation is therefore likely to be considerably understated by official definitions." As early as 1986 a UN study noted that changes to criteria used in China's or India's censuses could increase the world's urbanisation level by several percentage points.
Since urbanisation, construction and infrastructure projects have been a relatively easy way of boosting China's GDP, an under-estimated urbanisation rate would be no insignificant matter. Projections for future GDP would have to be dramatically revised (downwards), commodity demand would likely wane, real estate prices would lose a key area of support, and there would probably be bank losses and government losses, not to mention domestic political fallout.
Lucky for China, then, that the analysts at Standard Chartered are reviewing — and disputing — Pivot's claim in a piece of research out on Wednesday:
We went digging, and the main basis for the two claims [that China's definition of what consitutes an urban area is too low, and that "sprawling cities" are being under-accounted for since they don't have high enough population densities to be classified as cities] appears to be an OECD report published earlier this year ("Urban Trends and Policy in China‟, Lamia Kamal-Chaoui, Edward Leman, and Rufei Zhang, January 2009). This is a big, serious report which is based on independent statistical analysis of China's urbanisation patterns. We have been in contact with the authors.
The OECD report does indeed claim that China's official urbanisation rate is "understated". The 1,500ppkm2 density requirement is part of the National Bureau of Statistics‟ 1999 definition of an urban area, along with a requirement that the area possess a "contiguity of urban construction". The OECD authors point out, however, that "contiguity‟ is not defined, that the 1,500/km2 requirement is "considerably higher than in many other countries", and that China's suburban sprawl is therefore not covered by this definition.
From here it gets a bit lengthy but bear with us.
Standard Chartered has recruited the services of a Chinese geography specialist:
So we come back to what is really the key issue at hand: what exactly is urban?
One of the easiest mistakes to make is to follow a government's administrative classification and define everyone living in a city or town as urban. But administrative classifications inevitably lag reality on the ground.
To get around this, statisticians turned to population densities. In 1950, for instance, the US Census Bureau defined as urban areas with more than 2,500 people living in a built-up area with a population density in excess of about 400ppkm2, even if such areas were outside formal cities. (It is interesting to note that the US does not have a single official definition of „urban‟; this is left up to each federal agency.) Since then, demographers worldwide have fine-tuned the definition by looking at the physical qualities of an area to determine whether it is urban or not.
China's demographers have also gotten cleverer. On this point, we have also been helped immensely by Kam Wing Chan, a professor of geography at the University of Washington. The system for defining what is urban has improved considerably in recent years:
• For the past decade, China's definition of urban has included people who live outside a city's administrative boundaries but who live in an area with a population density of more than 1,500ppkm2. So if a township has grown into a large area with that density, it is included.
• Before 2000, China's urban definition was based on hukou (household registration), so none of the 150mn rural migrants living in urban areas were counted. Since 2000, the census has defined migrants who have lived in an urban area for six months or more as urban residents. (This is definitely an improvement, but there is some dispute over how well the surveyors are able to find resident migrants, on which more below.)
• In 2006, the NBS rolled out a new definition for what constitutes an urban population. It now relies almost exclusively on the physical attributes of an area rather than its administrative identity or - importantly - its population density. So, as long as an area "looks" urban, its population is classified as such (with at least one important exception, which we discuss below). Operationally, this is tricky to roll out, as local government demographers have to follow a qualitative list of attributes rather than just calculate population density. But it is nonetheless a good refinement.
• The new definition uses smaller units of population (resident committees within cities and villages, rather than the larger towns and townships which were used in the 2005 survey and before). This is another important step towards fine-tuning the measurement. One big weakness of China's methodology until 2006 was that it was based on administrative regions rather than functional areas. So before then, if a township reached the 1,500ppkm2 cut-off point, all of its population was counted as urban, and vice versa. Now, the unit is the village, and this should result in more accurate data.
• One nice thing about the new definition is that it generates more or less the same number as the previous method. This is important, as it weakens the claim that China has already completed the urbanisation process. To see this, we just need to look at the numbers generated by the new definition. The NBS found that China had 594mn urbanites in 2007, compared to 562mn in 2005 (using the former definition). If one assumes a 2-3% annual increase in urban residents, the new number is basically in line with the previous number - so we can conclude that the previous definition using the 1,500ppkm2 standard was basically OK. (However, this is based on the assumption that the new definition has been rolled out everywhere without too many problems.)
Two weaknesses remain in the new survey method, however. First, the exception we mentioned above. The OECD report points out that those living in urban-looking areas who do not receive urban services are still being classified as rural. This is clearly troubling, though we do not know how big a difference it makes to the overall number.
But there's also the issue of urban sprawl:The second weakness is more important, and also more controversial. It is sprawl. China's cities and towns are sprawling outwards, absorbing rural towns and villages. Some of these areas become built up, but many do not. Farmers continue to farm on parcels of land as factories and residential apartment blocks pop up around them. How to define, let alone govern, such areas raises no end of complicated questions. Some of these residents behave like urbanites, commuting to work in their new cars, while next door, farmers eke out a living on a strip of farmland.In simple terms, Mr. Leman at the OECD basically sees much of this sprawl as urban. He has spent years compiling demographic data and running computer-based analysis of China's sprawl. His estimate — using a 2,000ppkm2 density ratio (even higher than the old official one) — is that China's real urban population ratio is roughly 50%. That implies some 60mn more people living urban lives than the official numbers allow. He believes that his data is better since it does not rely on reporting by administrative units, but rather on much more fine-tuned analysis. He also thinks there is still much more urbanisation to come — by 2020-25, the real ratio will be in the 60-65% range, an increase of 15-20mn urbanites, he predicts.
But others point out reasons to believe that China‟s real urbanisation ratio is lower, rather than higher, than the official number. Again, the mess that China‟s sprawl has created is the reason for this confusion. The 2000 census found that one in five urban workers (21%) were actually farmers. In other words, even within areas with population densities in excess of 1,500ppkm2 (based on the old definition), there were still large rural enclaves. So in theory, one should take these people out of the urban number.
But one should not take out all of that 21%, since even in modern cities, there are allotments and horticultural areas on the outskirts. Prof. Chan argues that one should delete about 10% from the urban number, or some 20-30mn people (including dependents).
And then there is the issue of migrants. Most analysts have traditionally assumed that the census surveyors generally miss migrants. They do not visit temporary construction sites or unregistered factories, or they visit when migrants are temporarily away at home - since some of the surveys take place at peak harvest times . . . But Prof. Chan believes that the 2000 census was different and that for technical reasons, the number was an overestimate (basically, the surveyors did not find enough migrants so they adjusted up their numbers, but by too much). So, he argues that we probably need to remove another 20mn, leaving us with 40mn fewer urbanites than the official number - a ratio of some 42%.
In sum then, the Standard Chartered analysts see the urbanisation ratio at somewhere close to the reported rate of 45 per cent:Mr. Leman and Prof. Chan have differing views, partly of how to classify all of that not-really-rural and not-really-urban sprawl. But the key takeaway from this debate is that China's urbanisation ratio is most likely in the 42-50% range.
So 45 per cent according to Standard Chartered or 60 per cent according to Pivot.
It's a relatively small percentage difference, but one with very big consequences.
China tiene 1300 millones de habitantes, de los cuales aprox 800 millones son rurales. De estos hay aprox 200 millones que se van a trasladar a trabajar como obreros a las urbes en los próximos 20 años. Es a estos a los que hay que buscar casa, a obreros (mano de obra muy barata) que tendrá que vivir en las ciudades, dígnamente.
Urbanizar China significa no construir casas sólo, sino también, construir calles, aceras, alumbrado, saneamiento, jardines, carreteras, autopistas, cercanías, alta velocidad (han comprado el Bombardier canadiense que alcanza los 370 km/h), puertos, aeropuertos, hospitales, colegios, polideportivos, centros culturales...
Como dice el post, se necesitaría construir una ciudad como Nueva York cada año durante los próximos 20 años.
Y sabemos que hay una cierta inflación inmobiliaria en el barrio de Salamanca homónimo chino. Pero no en Fuenlabrada, ni Leganes.
Y esto es muy importante para el futuro crecimiento. Estas conclusiones son las que te hacen confiar en que a largo plazo, las diferencias con otros países desarrollados serán muy grandes.
Y gracias a que somos inversores a largo plazo, los vaivenes que se producen en el corto nos joroban, pero no nos quitan el sueño. Es lo que tiene la bolsa y hay que acostumbrarse a ellos.
A MARC FABER, del newsletter “Gloom, Bloom and Doom Report” le han hecho una entrevista que te resumo:
