Brad Setser

Follow the Money

Cross border flows, with a bit of macroeconomics

Print Print Cite Cite
Style: MLA APA Chicago Close


The New New China

by Brad Setser
January 20, 2006

I asked DOR — an active participant in the comments section of this blog — to do a guest post on the recent revisions to China's GDP numbers.   He has a real job, so his identity has to remain (somewhat) masked.   His post follows:

US$214 for everyone. That's the per capita increase in China's GDP under the adjustments announced in the second week of January. Its about the same price as a month-long ISDN connection in Shanghai, a night in a good hotel in Guangzhou or the rent on a 10 sq m room near the Oriental Plaza in Beijing. 

The revisions, the first since the early 1990s, were derived from a detailed, two-year economic survey. The results show an economy that was larger by Rmb 2,336.3 billion (US$281.9 billion) in 2004, or 17.1% (note that the Rmb appreciation since July 2005 has no effect on the revisions to 2004 data). According to the survey, the services sector is now thought to be 49.4% larger than before, manufacturing 2.3% and the primary agricultural, forestry and fishing 0.9% larger.

On the supply side, the revisions are mainly in domestic trade, catering, transport, storage, telecoms and real estate. Flip the national acccounts to the demand side, and one discovers an extra 22% (Rmb 1,087.9 billion). Fixed capital formation, by way of comparison, is now thought to be Rmb468.5 billion larger. Consumption thus rises from 36.3% to 37.8% while investments drops from 52.9% to 48.1%. Net trade also declined, from 9.1% to 6.3%, which leaves an unfortably larger residual (7.8%, up from 1.7%) in the new data. Inventory build-up, anyone?

Among the 18 provinces for which I have found revised data (more complete data would be most welcome), the average change among those showing larger economies is 11.1% while those reporting smaller economies shrank by an average 4.2 %. These 18 (of 31 sub-national entities, which for convenience will be referred to as provinces) represent 67.5% of the pre-adjustment provincial total or 80% of the pre-adjustment national GDP. The difference between the two figures is due to an inconsistency: the sum of the provinces' economies has exceeded the stated national GDP for many years, often by as much as 20%.

Guangdong Province, the largest sub-national economy, increased in size by Rmb282.5 billion (17.6%, equal to US$34 billion) whereas centrally located Hubei Province shrank by Rmb 67.8 billion (10.7%). The largest percent increase was in Beijing (up 41.7% or Rmb177.7 billion) while the smallest was Chongqing at 1% (Rmb2.7 billion).

Given the expected ease of documenting government activities, the adjustment to Beijing's data are puzzling. The municipal press release states that Rmb 154.4 billion (86.9% of the total) arises from revised service sector data.

Revised Ratios

The change in denominator means that every "/GDP" figure needs to be revised (it does not, however, improve the quality of any numerators). Hence, broad money (M2) is now 158.9% of GDP, not 186.1%; and narrow money (M1) drops from 70.3% to 60.0%. Central government revenues are now equal to 16.5% of the economy (down from 19.3%) and spending 17.7% (from 20.8%). The fiscal deficit didn't shrink much, just two-tenths of a percent, to 1.3%. Overall, the new figures suggest the Mainland and Hong Kong governments are roughly equal fiscal burdens to their respective economies.

On the current account, those who worried that 4.2% was too large in 2004 can rest assured it was only 3.6% (but, net trade was almost double that level, as noted above). Goods and services exports amounted to 35% of GDP (not 41%) and net direct foreign investment 2.8% (down from 3.2%).

Economists who tried to guess the size of the economy based on energy consumption need to revisit their spreadsheets. The country now generates Rmb11,534 worth of GDP for every ton of oil equivalent consumed, rather than Rmb9,848.

Stop making sense

Aside from incomplete data, there is also a problem with inconsistent pre-revision figures. According to Commerce Vice Minister Huang Hai, quoted in the state-controlled Beijing Review (May 13, 2005), China's final consumption as a share of GDP was 61.1%, 59.8%, 58.2%, 55.4% and 58.5% in 2000-04, respectively. United Nations figures suggest that final consumption didn't decline as quickly as Vice Minister Huang suggests, but neither set of numbers is even close to those cited by National Bureau of Statistics Director Li Deshui this month.

Different Views of China's Final Consumption Expenditure
(% of GDP)
Year Huang Hai UN pre-revision post-revision
2000 61.1% 61.0%    
2001 59.8% 60.6%    
2002 58.2% 59.7%    
2003 55.4% 57.5%    
2004 58.5% N.A. 36.3% 37.8%


Sources: Huang Hai, Vice Minister of Commerce, quoted in Beijing Review (May 13, 2005;  Author's own calculations based on United Nations data; and pre-revision and post-revision figures from Li Deshui, National Bureau of Statistics Director, quoted in People's Daily Online (December 21, 2005)

China carries out economic surveys on various parts of the economy at regular intervals, and has laid out a timetable for future refinements. Because the new data is still maddeningly incomplete, and full figures are probably still 6-12 months away, readers will need to very carefully read economic analysis of the PRC economy. "New" insights based on old data are going to be common for a while.


  • Posted by Navin

    DOR Excellent work.

    If the % of consumption in GDP decreasing from 00 to 04, what is making for the total growth ? Inflows/ FDI ?

  • Posted by Joseph Wang

    Question: What does this do to capital productivity numbers?

    I was at a conference on Chinese Law and emerging markets, and I noticed that the bankers and economists were all very optimistic on China whereas the people that were most negative where the management academics. The reason for this is that the management academics focus on the capital productivity numbers which have been going down in the 1990’s.

    I’m curious what the new GDP numbers do to the capitial efficiency numbers.

  • Posted by Gcs

    for what its worth

    these numbers tell me nothing


    these numbers are a reflection of collectors proirities

    if net discoveries
    are nearly all services
    which makes sense …i guess

    how explain provincial declines ???

    are the provincial decliners
    seeing their industrial va jiggered
    away to other stages in a process
    or what ????

  • Posted by DOR

    Thanks, Navin
    For the most part, I don’t know the more complete answers to your (and other people’s) questions. Off the top of my head (and in the absence of any more complete data), I’d point to the rapid rise in capital investment and net trade as the reason for the total economy to rise faster.

    Joseph Wang,

    I wish I had read Carsten A. Holz’s FEER article (“Why China’s New GDP Data Matters”, available to subscribers at before writing this piece, as he does a credible job of examining the alternative conclusions that might be drawn from the available data. Among them is that depreciation or profits may have been enormously larger than previously thought. If China has been replacing capital equipment at an unprecedented pace, might capital productivity be dramatically increased?


    The national GDP figures may now be very close to the sum of the provincial figures (we don’t know yet), which may mean that the provincial data collectors are now thought to be more accurate or honest. Li Deshui says is that the overly optimistic (OK, falsified) manufacturing data that some provinces produced has now been reduced. In some provinces, that reduction would be more than the increase in services sector data.

  • Posted by Gcs

    dor thanx for clearing that up

    the gross changes
    in industrial
    production are signifiganteven if they basically net out

    and if i/m right the costal share
    of industrial production
    is actually greater now

    as to the measures one extracts for retun to capital

    this is really a sportsmans game
    at this point
    again i see tea leaf readers
    finding what pattern
    they are most looking for

  • Posted by bsetser

    DOR — first, thanks. second, I had not realized that the investment number had been revised up as well, so consumption remains under 40% of GDP (in the official numbers):

    “Consumption thus rises from 36.3% to 37.8% while investments drops from 52.9% to 48.1%. Net trade also declined, from 9.1% to 6.3%, which leaves an unfortably larger residual (7.8%, up from 1.7%) in the new data. Inventory build-up, anyone? ”

    Given that investment growth exceeded consumption growth in 2005, what would your estimate be for investment/ GDP in 05? presumably, it is close to 50%, and that implies saving of 55% plus given the current account (savings) surplus.


  • Posted by gabor

    Chinese consumption/GDP is a very important gauge to watch as it directly influences global inflation -> liquidity -> sustainability of the Asset Economy.
    Once Chinese start to increase their consumption faster than their capacity, global deflation will turn into inflation and the Asset Economy will collapse.

  • Posted by DOR


    Thanks for the opportunity to tap your audience.

    On the demand side of the ledger, assume Final Consumption – including Government Consumption Expenditure (GVT: 12.0% of GDP in the old 2004 data, now unknown) and Private Consumption Expenditure (PCE: previously 43.3% of GDP, also now unknown) – has not been a major driver of growth, if it rose at all in 2005. Changes to National Account’s Gross Fixed Capital Formation (CAP) is very hard to see in the monthly or quarterly data produced in the past year (and, the numbers were based on the old methodologies). Let’s also remember that the official statistical discrepancy (Stat) ranged from -2.6% to -3.6% in recent (pre-revision) years, and that inventories maxed out at +0.4% of GDP (unrevised) in 2002.

    Assume CAP rose 1.9 percentage points (to 50%), and net trade from 6.3% to 10.0% (a lowball guess), AND the Stat remains at least 2% negative. That leaves 36.0% for Final Consumption. I don’t (can’t) know if that is correct, but it adds up.


    You may be right about PCE/GDP shifts affecting the asset economy. However, that really doesn’t keep the people in China awake at night. Not on their top 10 list of concerns.

  • Posted by HZ

    I think it is Keynes’ idea that benefits to the present outweigh risks to the future. For one thing no one has a crystal ball; for another in the long term we are all dead. Even in those dark years Keynes had faith in the adaptability of human to eventually sort things out.
    Demands for current sacrifices for future benefits are hardly ever justified. So if the Chinese overinvestment keeps people employed, while American overconsumption is what consumers want, so be it. If the trend can’t continue, it will reverse one day. But with people being adaptable, the ending will not be nearly as bad as bears would fear. If things do go terribly awry, it won’t be for economical reasons, anyway.

  • Posted by Garhane

    This is great detail but it would be most helpful to the general reader to see a note as to how this
    all sums up for a view of what China has become. How about a dozen or so summary measures, a note
    as to which can be credited at all, and a range of confidence or the lack of it.
    I have been reading accounts about China for several years and there seems to be a general view
    that “it”cannot continue while it (pick your measure from trade or other deficits to monetary topics)
    keeps on doing so. Either the numbers we get are so wrong none of them can be relied on at all, or
    nobody can claim to have done their sums correctly, or the method of financial analysis being used
    is not correct. Knowing at least a 60% answer (meaning something just a little better than chance)
    wojuld be very helpful.

  • Posted by Elaine Supkis

    Very funny stuff here.

    Guess what? The REAL information is who is fighting hardest for their jobs, their rights? Americans or Chinese?

    Answer: the Chinese, hands down. They have forced the government to respond, to change tax laws by removing taxes on farmlands, change mining laws (look who has been having mining deaths THIS month!) and in general, has the rulers running like crazy.

    Ameican workers seem to be a collection of wet noodles going about, heads down, as factories close and wages of 80% of the workers is below the rate of inflation and they are so hopeless and helpless and…MICE, NOT MEN!

    So, whatever is going on, the Chinese are up and up and we are down and down and going down badly. Ptui.

  • Posted by bsetser

    Garhane and HZ — one of Rudy Dornbusch’s better lines was something to the effect that “unsustainable trends last far longer than you think they can, and then unwind with more force that you thought possible”. I am not convinced that this will unwind quite as smoothly as HZ suggests.

    and HZ, how would you manage the strains created in those parts of the US that still make things as China moves up the value-added chain? Unfortunately, those areas tend to be in “flat-land” so rising home values won’t bailout out auto (parts) workers who lose their job … And the US doesn’t really have a mechanism for redistributing the capital gains in Organge County (or hedge fund managers profits) to a hard hit part of Michigan … even if economically, China can keep on adding to its reserves, I am not sure the political consensus that supports the status quo will hold.

    all the more so if home equity extraction slows globally and us consumption growth slows, making everything a bit more of a slog.

  • Posted by HZ

    I don’t see an easy solution. A softening of US consumer demands is probably the pre-requisite to our trade balance re-adjustment. Will that lead to a global recession? I hope people are smart and adaptive enough to re-adjust.
    There is no lack of jobs in US. The evidence is in the amount of legal and illegal immigrant labor we import each year. Agriculture/hospitality/construction/nursing/programming, you name it. It is very hard to get sympathy from Chinese or other developing economies where labor faces much more intense competetive pressure.
    On the other hand, competition in manufacturing will be there whether China is there or not: Detroit with its legacy cost, faces threat mostly from Japanese im-plants; in textile without China there are still India/Pakistan/Vietnam and countless other countries; in electronics there is always Taiwan/Korea/Southeast Asia.
    The most serious problems about China are political not trade: its political reform and internal stability, conflict across the Taiwan Strait or with Japan, how it will use its newly acquired military and soft power in the future, to name a few. I hope trade issues do not get blown out of proportion and lead US/China down on a path of conflicts.

  • Posted by HZ

    ” And the US doesn’t really have a mechanism for
    redistributing the capital gains in Organge County (or hedge fund managers
    profits) to a hard hit part of Michigan ”

    There is a thing called “capital gain tax”. But seriously this is hardly something that we should expect our trade partners to solve for us. Maybe we need more taxes, more welfare and more relaxed attitudes towards creative types living on the dole.

  • Posted by DOR


    “What China has become” ? Hmm, a market-based economy under an increasingly less rigid (but still authoritarian) Leninist state? That is what it has become, but it isn’t anything really new to say so.

    China’s leadership has a very good record of identifying problems early on, working out the most optimal (least worst) solutions, and successfully implementing those solutions. There have been mistakes, but not as many as in other economies faced with similar challenges. Hence, there is good reason for confidence that China will be at least as successful at rebalancing its (positive) economic misalignments as the US will be at rebalancing its (negative) misalignments.

    Make no mistake about it: the numbers are wrong, every single one of them. They are not US or OECD standard, and have to be used much more cautiously than the numbers we are used to. However, they also are the only numbers we have, so unless we are going to abandon quantitative analysis, we are going to just have to keep on trying as best we can to make sense of them.

    Elaine Supkis,

    I knew the day would come when we would agree on something!

    When Americans are leaving their homes and traveling hundreds of miles to live in military-style dormitories with no social safety net, just to earn enough to send home to their families, then – and only then – I might consider it a serious question as to who’s fighting hardest for their jobs.

    Oh, and I can guarantee you that more Chinese miners die in any given month than do American miners. But, CNN doesn’t give 24 hour coverage to mere Chinese miners.


    I wouldn’t make too much of China’s political milieu. The country is the most stable it has been in centuries. Political reform is proceeding, albeit not at the pace Western observers would wish.

    Taiwan is a minor distraction, a sideshow, the Chinese equivalent of the Terri Schiavo deathwatch. China’s military power is purely land-based and domestic. They have no successful experience projecting power across their own land borders – let alone across water – since sometime back in the early Qing Dynasty. The nuclear weapons cannot be used, for the same reasons that the US and USSR couldn’t use them in the 1970s, and there isn’t an amphibious force worth considering.


  • Posted by Anonymous


    Thanks. Back when I was working at the IMF in the “wild east” of the FSU, we encountered a statistical/methodological anomaly in the way in which the various state statistical agencies compiled their price level series. The anomaly, known as the “Sauerbeck problem”, led to an overestimation of inflation and hence price levels. I wonder if the same problem exists in the Chinese data (they would presumably have received “technical assistance” from Gosplan in the 1950s when putting their numbers together). All of this wouldn’t really matter in the good old days of price controls, of course. But as liberalization increases and prices are freed, the Sauerbeck problem would tend to distort the breakdown between real GDP and prices. Perhaps the latest revisions to GDP reflect a correction for this anomaly?

  • Posted by Joseph Wang

    The reason I’m interested in the capital productivity numbers is that I was at a recent conference on Chinese Banking, and I noticed that the finance people were very optimistic on the Chinese economy whereas the management people were very pessimistic. This seemed to be due to the fact that they were looking at different numbers.

    The finance people were looking at GDP numbers and the NPL numbers and they look good. Another factor is that the finance people seemed very impressed by the Chinese bureaucrats that they had to deal with.

    The management people were negative, because the numbers that they were looking at were capital efficiency numbers and these have steadily declined over the 1990’s. However, I’m wondering if these numbers are a fluke. One problem is that the conventional wisdom in the 1990’s was that the Chinese government was fudging numbers. If it turns out that the Chinese government isn’t deliberately fudging numbers (which appears to be the case) this makes figuring out what is really going on much harder.

    (It was also interesting to contrast the views of these people with the typical view you get from journalists. One issue is that people who have had dealings with the banking regulators seem to come away with a much more favorable impression of the government than people that deal with local officials.)

    I tend to side with the finance people. One can get too buried with numbers, and it’s always a good idea to just look outside the window and see if the numbers make any sense. The thing about economic numbers is that they are pretty easy to verify (i.e. if you want to get a sense of unemployment and inflation, just take a walk on the street).

    Things like capital efficiency are harder to get a handle on, but the notion that the Chinese economy is less efficient with capital in 2000 than in 1990 just “smells” incorrect.

  • Posted by Elaine Supkis

    Good grief.

    American miners die, so do Chinese. But the difference is, our mining commissioners are corrupt like the Chinese BUT they are NOT being PUNISHED!

    Far from it! Bush is rewarding them! So miners in America just have to go hang, eh? I don’t see them marching and yelling.

    Crying, yes.

    30,000 Ford workers will be thrown off the cliff this next two years, minimum. What are they doing? Eh? Are they, like the Remington workers who just found out their factory is going for a hike to Asia, going to go to work, heads down, until dismissed to die of drink or work three jobs while living in a foreclosed home?


    So who is rioting? Who is marching? Who is stopping the machinery? Ain’t Americans. They are now sheep.

  • Posted by DOR

    Sauerbeck problem: “a discrepancy between measured inflation and the targeted inflation path that was consistent with the programmed degree of monetary restraint” . . . even if I understood what it meant, I’m not sure it would apply.

    China hasn’t been on what I would think of as a FSU-style stabilization process, nor a sharp deceleration of money supply growth. Did I miss something?

    Joseph Wang,
    My first thought was that the financial people are anticipating big whacking fees while the managers are expecting big whacking headaches.

    Looking outside the window to see if the numbers make sense . . . one of the most efficient (and least utilized) tools in the economic toolbox.

    In the early 1990s, I was asked to go to Ho Chi Minh City to give a talk on the Vietnamese economy, about which I knew almost nothing. I went down a day early, and walked the streets for about 10 miles. What I saw was Taiwan in the late 1970s: markets bulging with great vegetables and cheap electronics, lots of people on bikes or motorcycles, dozens of people standing around a shop watching TV and open sewers. You gotta smell the economy to understand it (and, you gotta remember that it is the richest corner of the country).

    The other thing I notices, and began to track in various places, was shoes. Once you leave the farm, the first thing you want is something on your feet (flip flops). Then, maybe you get married and buy a “nice” pair of shoes.

    Of course, those become the everyday shoes for a long time (even for construction workers), but eventually you get rich enough to buy purpose-made shoes: shoes for walking, for working, for going out at night, etc. It is a very nice measure (beats PPP any day!).

  • Posted by Joseph Wang

    DOR: The self-interest part might be part of it, but it didn’t seem to be the whole story since most of the finance people were academics and government regulators as were the management people.

    Also, the people that were from the banks seemed pretty level headed (which was a good thing). One of the speakers was from a Western firm that had gotten control of a broken Chinese bank, and I got the same feeling that I would from someone who had just bought a old, badly broken house which they are in the process of renovating.

    I also got the sense that Zhou Xiaochuan (the current head of the PBC) was one of those people with massive historical impact, that no one has ever heard of.

    One quick sanity check: It is the case, is it not, that PRC nationals can’t easily buy shares of PRC companies listed on HK or NYC boards because of currency restrictions? I was talking to a lawyer about this, and when that factoid came out of my mouth it seemed really bizarre, but I think that is the case.

  • Posted by DOR

    Joseph Wang,
    Yes, you’re right on PRC nationals currency exchange controls: no NYSE or HKEx script for them.

    We get a flood of PRC-origin money here in HK, and while it is all perfectly legal once it is here, getting it here is strictly illegal.

    Bags of Rmb to buy apartmetns. Literally, bags of it.

  • Posted by jaichind

    The level of investments on the PRC is overestimated which in turn makes the capital efficiency look bad when in fact it has been getting better over the last decade. Land costs and investment sectore inflation are underestimated both which artifically pushes up the investment ratio to GDP. The incremental capital-output ratio on the PRC lower than the Asian Tigers although the PRC economy is less capital intenstive, a fact that will change in the future. The rate of return on asets and equity last couple of years are several times higher than the late 90s.