How Badly Flawed Is Chinese Economic Data? The Opening Bid Is $1 Trillion

by: Christopher Balding

Short Abstract:

Baseline Chinese economic data is unreliable. Conservatively, correcting for housing price inflation in the Chinese CPI data adds approximately 1% to annual consumer price inflation in China, reducing real GDP by more than $1 trillion.

Long Abstract:

Baseline Chinese economic data is unreliable. Taking published National Bureau of Statistics China data on the components of consumer price inflation, I attempt to reconcile the official data to third party data. Three problems are apparent in official NBSC data on inflation. First, the base data on housing price inflation is manipulated. According to the NBSC, urban private housing occupants enjoyed a total price increase of only 6% between 2000 and 2011. Second, while renters faced cumulative price increases in excess of 50% during the same period, the NBSC classifies most Chinese households as private housing occupants making them subject to the significantly lower inflation rate. Third, despite beginning in the year 2000 with nearly two-thirds of Chinese households in rural areas, the NBSC applies a straight 80/20 urban/rural private housing weighting throughout our time sample. This further skews the accuracy of the final data. To correct for these manipulative practices, I use third party and related NBSC data to better estimate the change in consumer prices in China between 2000 and 2011. I find that using conservative assumptions about price increases the annual CPI in China by approximately 1%. This reduces real Chinese GDP by 8-12% or more than $1 trillion in PPP terms.

Key Facts

  • According to the National Bureau of Statistics China (NBSC), the price of private housing in China rose by a total of 8.14% between 2000 and 2011 and only 5.99% in urban areas, where Chinese were moving to in large numbers.
  • According to the NBSC, the annual price of private housing in rural areas grew at 1.67%, more than three times faster than prices in urban areas at 0.53%.
  • According to the NBSC, the price increase of renting outpaced the change in the price of private housing by nearly 50%, making it significantly more advantageous to purchase an apartment between 2000 and 2011 in China.
  • According to the NBSC, only 12% of Chinese households are renters, skewing inflation data.
  • To calculate the total private housing price change, the NBSC utilizes a straight 80% weighting of the urban population and a 20% weighting of the rural population despite the fact that in 2000 nearly two-thirds of Chinese households were rural and only rose to 51% urban in 2011.
  • According to the NBSC, rural home values rose by 249% between 2000 and 2011 for a compounded annual growth rate of 8.65%.
  • According to the third party data, from the first quarter of 2000 to the first quarter of 2010, the nominal value of apartments in urban areas in China rose nearly threefold.
  • When using third party data and reconciling NBSC data in place of official housing inflation data, annual Chinese CPI increases by approximately 1% annually.
  • Incorporating this change in CPI into real GDP calculations reduces total real GDP by a mid-range estimate of 8-12% or $1 trillion in PPP.
  • Other significant discrepancies exist that will likely increase the size of the needed restatement of total real GDP.
  • According to NBSC data, the food component of the CPI in China was responsible for 99% of inflation between 2003 and 2011. This implies that the NBSC is claiming that the only prices to rise in China between 2003 and 2011 were food prices.


Since China opened up to the world in 1979, it has been an economic juggernaut, officially expanding by a compounded growth rate of 10% between 1980 and 2012. This rate actually accelerated after 1999, averaging 10.2% from 2000 to 2012. China has pulled hundreds of millions of its citizens out of extreme poverty becoming one of the world's largest economies.

However, there is strong evidence indicating that the rate of real Chinese GDP growth and ultimately total real GDP may be significantly overstated. This discrepancy stems primarily from significant and systematic irregularities in the inflation data maintained by China and the pass-through impact on a wide variety of other data such as real GDP and disposable income. If inflation data is not accurate, or is willfully fraudulent as appears to be the case, it will impact many other areas of economic and financial data leading to large disparities over time.

Though there are undoubtedly other irregularities within Chinese inflation data, the focus of this paper is the discrepancy around price changes in the cost of housing. The cost of housing is normally one of the biggest single line items in many national price baskets. Consequently, changes in its composition will have a disproportionate impact on the calculation of real GDP relative to other items. There is strong evidence that China has systematically manipulated its housing price data in order to lower official inflation, which, using conservative estimates, has lowered annual inflation by approximately 1%. Over this time frame, this would conservatively reduce real Chinese GDP by approximately 10%.

Furthermore, this irregularity in the calculation of inflation may help explain discrepancies in other macroeconomic variables such as growth in money supply, nominal, and real GDP growth in China. Taken over a longer time horizon, there are discrepancies in other aspects of Chinese data that may be explained by the inflation data.

This article is broken into three sections. First, I begin with an examination of Chinese price data focusing on the housing market. Second, I study broader macroeconomic variables including cross-country data from other emerging markets and developed countries for comparison. Third, using this data and plausible assumptions, I produce an estimate of real Chinese GDP.

Chinese Housing Price Data

The National Bureau of Statistics of China (NBSC) houses the primary national economic, financial, and social data. It is publicly available in Chinese and English with monthly, quarterly, annual, and census data, including a separate portal for data searches. Within their annual data collection, in most years' Section 9, they maintain a lot of data on prices for producers, consumers, and other areas of interest. In Table 3 under Section 9, they produce the consumer price index change with a lengthy list of major cost line items and their individual annual price changes.

At the bottom of Table 3 under Section 9, the NBSC presents the price change on "Renting" and "Private Housing" with numbers giving the price change from the previous year multiplied by 100. Consequently, the number 105 implies a 5% increase from the previous year, 95 means a drop of 5% from the previous year, and 100 is no change. The raw data for year 2000 to 2011 is available in Table 1.

Table 1

Click to enlarge

Source: National Bureau of Statistics of China

Even before moving to more detailed analysis, there are two items that cannot be overlooked even to casual observation. First, the price of private housing appears to grow quite slowly. According to raw NBSC data, from 2000 to 2011, private housing price growth was negative or flat in 4 out of 12 years. To anyone with even a rudimentary knowledge of China, this seems questionable at best. Second, the price of renting grows faster than the price of private housing. In 8 of the 12 years, the price of renting exceeds the price increase to private housing. Furthermore, renting is a much bigger winner than loser. For instance, in 2009, the price of renting increased 1.6% while private housing dropped 14.7% for a total differential of 16.3%. In 2002, renting increased in price 4.4% while private housing dropped 4.6% for a total differential of 9%. The obvious difference between rental and private housing prices implies that it is becoming more affordable to purchase a home in China than rent.

If we transform the raw data into indexed data with the base year 2000, the discrepancies between Chinese housing price data and common sense become increasingly stark. The indexed data with 2000 as the base year equal to 100 is presented below in Table 2.

Table 2

Click to enlarge

Source: National Bureau of Statistics of China and the author's calculation

There are numerous results that need mentioning. First, according to the NBSC, the price of private housing in China rose by a total of 8.14% between 2000 and 2011. More implausibly, the price of private housing rose by only 5.99% in urban areas, where Chinese were moving to in large numbers. Second, according to the NBSC, the annual price of private housing in rural areas grew at 1.67%, more than three times faster than prices in urban areas at 0.53%. In other words, it became more expensive in China to live in rural areas than in urban centers. Third and finally, according to the NBSC, the price increase of renting outpaced the change in the price of private housing by nearly 50%. Put another way, the price to rent ratio should have made it significantly more advantageous to purchase an apartment between 2000 and 2011 in China.

There are three further ways in which the NBSC manipulates inflation data, which are more subtle. First, according to the NBSC, only 12% of Chinese households are renters. As noted in Appendix 1, the NBSC counts only 11.95% of households as renters. When including all categories of private housing as defined by the NBSC, and excluding the undefined "Other" category, a total of 85.4% of Chinese households are counted as private housing occupants. This has a very big implication: even though the consumer price of renting increased a total of 53% from 2000 to 2011, this price applied to less than 15% of Chinese households according to the NBSC. The remaining 85% of households are counted as only seeing an 8% increase in their price of housing. In other words, even beyond the manipulation of basic price data, the NBSC is further slanting it by claiming an 85% weighting on the lowest price.

Second, the NBSC is skewing the distribution of population between rural and urban when calculating the total change to the private housing price index. The past decade in China has witnessed one of the largest migrations in human history from rural farms to urban apartment dwellings in China. In 2000, nearly two-thirds of the Chinese population was classified as rural. By 2011, the number classified as urban residents topped 51% according to the NBSC. However, when calculating the implied population distribution of the total private housing price index between rural and urban residents, it becomes obvious the NBSC has not accounted for actual population dynamics in China. These results can be seen in Table 3.

To calculate the total private housing price change, the NBSC utilized essentially an automatic 80% weighting of the urban population and a 20% weighting of the rural population. In every year except 2000 and 2005, utilizing a strict 80% urban weighting yields a private housing price change within one- tenth of one percent of the actual NBSC number. In 2005, the difference is less than twenty-five basis points. Given the fact that nearly two-thirds of Chinese residents began 2000 as rural residents the NBSC decision to give an 80% weight of the total price index to urban residents strikes one as particularly egregious.

Third, the NBSC significantly under-weights the cost of housing in the total consumer price inflation basket. According to one report, from 2000 through 2010, the NBSC gave only a 13% weighting to housing in the CPI basket (click here for more details). To put this weighting in perspective, it gave a higher weighting to "education and cultural articles" and only a slightly lower weighting to "clothing." In 2011, the NBSC reweighted the housing portion to a 17.2% percent weighting to housing despite the fact that it grew significantly slower than all other components of the CPI. In other words, even though housing fell significantly relative to other items in the NBSC basket between 2000 and 2010, it was magically reweighted upward in 2011. The NBSC is implicitly saying its own statistics are unreliable.

This fraudulent compilation of price index data has an important implication: it drastically lowers the price impact of housing even more than the manipulated price data indicates. With 85% of residents classified as occupying private housing and an 80% weighting given to urban residents in the price index, this implies that according to the NBSC nearly 70% of Chinese residents enjoyed a cumulative increase in the price of housing of 6% between 2000 and 2011 (this number is obtained by taking the 85% of private housing dwellers, multiplying by the 80% weighting given to urban dwellers, then utilizing the accumulated total price change for urban private housing dwellers in Table 2). Put another way, the NBSC is claiming that 68% of Chinese households faced annual price increases of 0.53%. If we utilize a simple and more realistic assumption that the price of housing should represent 20-30% of a sample price basket, this would imply a contribution to the Chinese CPI of one-tenth of one percent annually. Any increase in the urban price of private housing would have a significant pass-through effect on other economic data.

Let's turn from price basket data to comparisons against other changes of price in housing and real estate beginning with changes in the value of rural home values. The NBSC rural home value price with index base year 2000 is presented below in Table 4.

Table 4: Rural Home Value Raw Price and Index

Click to enlarge

Source: National Bureau of Statistics of China and author's calculation

According to the NBSC, rural home values rose by 249% between 2000 and 2011 for a compounded annual growth rate of 8.65%. As there is no known third party research on the pricing of rural real estate and housing price data in China, the veracity of these statistics is virtually impossible to independently corroborate. However, even taking the NBSC rural house value as a baseline, it stretches the boundaries of plausibility to absurd lengths that rural home values would increase 9% annually while rural private housing prices increased only 1.7% as noted in Table 2. It is important to briefly note that according to the NBSC, 96% of rural households are classified as private housing occupants, hence this analysis will focus on the price of private housing ignoring the renting price change (this data is available in the back in Appendix 2).

There are a couple of additional points about this. First, the growth in rural private home values rose close to the official real GDP and well below nominal GDP numbers. It is at best implausible in a statistical series when nominal GDP grows more than 500% and rural home values increase 350% that the price of rural housing would grow just 19% (we will review in a later section the relationship to Chinese and cross country macroeconomic aggregates). Second, while a significant migration from rural to urban areas took place during this time sample, the rising rural home value indicates rising demand that would impact the price index. In other words, despite the migration from rural to urban areas, rising rural incomes prompted increases in home values. Third, while consumer price index measures will not move in perfect correlation with asset prices, the enormity of the disparity deserves skepticism. Fourth, while it is difficult to independently verify the rural home value data, it is extremely problematic that the NBSC presents two such closely related numbers that vary so enormously. Given the reduction in the consumer price measure, this appears to be more strategic data manipulation.

Turning to urban housing prices, the overall picture is strikingly similar though, with different data limitations. The NBSC did not publish a continual urban value throughout this time period as they did with the rural value. When they did publish value data, it gained such widespread Chinese skepticism, that the NBSC stopped publishing it in 2011. The official "70 Cities Index" showed an annual urban value increase of just over 4% annually from 2006 through 2010 (Wu et al. 2012). Even more unbelievably, the "70 Cities Index" claimed that the average annual real growth in Chinese boom towns like Shenzhen was a miniscule 0.04% while in Shanghai it was only 1% (Deng et al. 2012). As official estimates of private housing prices are implausible and widely discredited within China, we turn to third party data on the change in value of private housing.

To compile improved pricing estimates outside the official data, I borrow heavily from the research carried out by the Department of Construction Management at Tsinghua University in Beijing and the Institute of Real Estate Studies at the National University of Singapore (Wu et al. 2010). In three papers, researchers lay out in great detail third party pricing values, compare them to official indicators, and improve price index methodologies. The figures, tables, and data are borrowed from their research. There are numerous pricing and value patterns, which are worth noting. First, the asset value of private housing has risen much faster than official data claims. Despite the NBSC data claiming that the price of private housing has risen by 8% since 2000, third party data indicates that underlying asset values have risen rapidly and especially in recent years. According to the Tsinghua University researchers, from the first quarter of 2000 to the first quarter of 2010, the nominal value of apartments in urban areas in China rose nearly threefold as presented below in Figure 1.

There are a couple of points worth noting about Figure 1. First, the urban home price value has risen much faster than official estimates. Second, the nominal value increase closely approximates the change in the rural housing price as given by the NBSC. Third, the real value is close to the nominal value because of the manipulated consumer price inflation data. If inflation data accurately reflects consumer prices, specifically in housing prices, the differences will become much larger. This furthers the argument that official Chinese data grossly understates the change since 2000 in housing prices and, for the purposes of this paper, the specific pass-through impact on price indexes and real GDP (please see Appendix 4 for additional information on the change in Chinese urban home values).

Next, I turn to the price of private urban housing relative to the price of renting. According to official NBSC CPI statistics presented in Table 2, the accumulated price change since 2000 between renting and private housing in urban areas was nearly ten times greater in 2011. In other words, it became significantly more advantageous to rent in urban areas than purchasing. However, third party data indicates that this is inaccurate as presented in Figure 2 (Wu 2010).

Figure 2: Price to Rent Ratio in Eight Major Chinese Cities, 2007(q1)-2010(q1)

While this data figure does not cover the entire time frame under question and only covers 8 major cities, it demonstrates quite clearly that private housing prices are significantly outpacing rental price increases as claimed by the NBSC. It is noteworthy that most cities entered the sample at elevated to significantly elevated levels and end the sample at highly elevated price to rent ratios. In other words, third party data simply does not support the NBSC private housing to rental market price data.

There are numerous problems revealed with the NBSC compilation of residential pricing data. First, the base price data is obviously manipulated. Second, the smallest price increases are seen in urban private housing owners which given the weights skews the results. Third, given the division between urban and rural residents it significantly overweights urban residents. Fourth, NBSC data significantly underweights in the total CPI basket the impact of housing. Fifth, the official data does not reflect independent third party data with regards to changes in the price of housing. While we do not expect the price change in consumer prices to be perfectly correlated with asset values, the sheer enormity of the discrepancies presented are a real cause for concern. Furthermore, while we do not expect third party data to perfectly mirror official data, the enormity of the discrepancy is noteworthy. Even the patterns revealed, such as the difference between private housing and renting, are so starkly different as to warrant investigation. Using this data as a baseline for estimating better prices, I will use the primary findings of this research to better estimate the impact on Chinese inflation and real GDP growth.

Chinese Home Prices and Real Estate Values in Macroeconomic Context

Despite the evidence that the NBSC has manipulated the change in housing price data to systematically lower inflation data, it is important to place this data within the larger picture of Chinese macroeconomic data since 2000. We begin with a comparison to official Chinese CPI data as presented below in Table 5.

Table 5: Private Housing Relative to Chinese CPI

Click to enlarge

Source: National Bureau of Statistics of China and author's calculation

According to the NBSC, consumer price inflation (CPI) increased at a rate of 2.43% between 2000 and 2011. This resulted in an official accumulated price level increase of 30.2%. When we divide the consumer change in the price of private housing by the CPI, the data implies that the real price of private housing fell in China by a total of 17% or at an average annual rate of 1.68%. In other words, according to the NBSC the real price of private housing dropped by almost 20%.

When we compare the change in the CPI component of private housing against other macroeconomic indicators, we again witness a similar divergence. Since 2000, real GDP, nominal GDP, and the money supply have all officially expanded enormously. While China has undoubtedly grown significantly in real terms, what has been little understood about this growth has been the accompanying expansion of the money base. Real GDP increased just under three fold between 2000 and 2011, but money grew more than twice as fast, increasing more than six fold over the same time frame. When we demonstrate the growth in these key macroeconomic variables, with the growth in the official price of private housing in China below in Figure 1, the discrepancy becomes increasingly obvious.

From 2000 to 2011, according to official statistics, M2 money grew more than six fold, nominal GDP nearly fivefold, and real GDP three fold, but the price of private housing by only 8%. If we focus on nothing other than the growth of official real GDP, it seems implausible that the price index of private housing would be falling so significantly relative to output. If we include the nearly five-fold increase in nominal GDP and six-fold growth in M2 money into the analysis, it again strains credibility that the price of private housing would remain nearly unchanged for eleven years.

The Chinese Growth Story in Cross Country Perspective

The sources and discrepancies of the Chinese growth story gain much more focus in a cross-country perspective. Comparing the major macroeconomic variables, there are some stark contrasts that further demonstrate the argument that systematic under reporting of inflation has significantly boosted Chinese real GDP growth. We begin with an examination of the increase in numerous measures of the money supply. Despite the reputation of China as managing prudent monetary policy, it has increased money by one of the largest amounts among emerging economies since 2000. The first figure is the growth of M1 among major emerging markets and select developed countries, presented below in Figure 2.

According to Figure 1 for major emerging market economies and the Euro Area, China experienced the third highest growth in M1 money (please see Appendix 3 for a figure, which includes India for 2006 to 2012 data). While M1 is a narrower form of money, given the difficulties in cross-country money growth comparisons, it helps provide a basis for comparing money growth in countries. As M1 comprises only the narrowest definitions of money, such as currency in circulation and demand deposits, it omits a significant volume of potential money expansion. Specifically in China, given the fixed currency which funneled the growth in money through banks and ultimately to state owned enterprises, the M1 definition of money may overlook significant growth in broader money. The growth of M2 money is similar, as can be seen in Figure 3.

Figure 3: M2 Growth of Major Economies from 2000-2011

According to IMF data, Chinese M2 growth was again the third highest among major emerging market economies. Chinese M2 growth outpaced almost every major emerging market except countries with rapid inflation. However, rapid growth in the money supply does not necessarily portend problems if there is growth in the economy. In fact, despite the Milton Friedman declaration that "inflation is always and everywhere a monetary phenomenon," most central banks have adopted strategies unrelated to explicit money supply growth targets to maintain stable prices.

One problem relying on other measures of monetary policy is their reliance of flawed or manipulated data. Let's take a simple variation of the well-known Taylor Rule, which has been rewritten as the Mankiw Rule. The Mankiw rule is written as the following:

While the discount is public and well known, both inflation and unemployment are subject to political manipulation. I have presented below the Mankiw Rule for China plotted with the discount rate, consumer price inflation, and estimated core inflation in Figure 4 (China does not breakout a specific "core inflation" number. However, given what we believe is the official CI weighting, I have estimated core CPI based upon this base).

When using official data and the Mankiw Rule, China appears to have run a tight monetary policy between 2003 and 2011. However, it would be mistaken to believe that China ran tight monetary policy since 2003 based upon fraudulent data. As this paper is primarily about inflation data, here I will focus on the problem with unemployment data. Since 2002, it has fluctuated constantly between 4% and 4.3%. Despite the enormous up economic changes, both to China and the global economy, it should induce serious skepticism that a rapidly growing export dependent economy enduring a global financial crisis never sees its unemployment rate move. The quarterly data is so rigid that years pass without the quarterly unemployment rate fluctuating even one-tenth of one percentage point. Even believing the general trend, it seems implausible that Chinese unemployment has been so rigid so long. In fact, between 2000 and 2012, China had the lowest standard deviation of 28 major economies, which is presented as a figure in Appendix 5. To provide some perspective, the Chinese standard deviation of unemployment is a third less than Japan over the same period. The absolute rigidity should provide real concern about the quality of data being furnished by official Chinese agencies.

Next, I focus on the increase in nominal GDP. While not widely focused on, nominal GDP is an important place to start when considering the Chinese growth story since 2000. As has already been noted both in Chinese specific and cross-country data, the money supply grew rapidly. Consequently, we would expect nominal GDP to grow quite rapidly. We examine cross-country nominal GDP growth below in Figure 5.

Despite having one of the highest increases in M2 growth in major emerging market economies, China had lower nominal GDP growth than Indonesia and Argentina. This means that to achieve the claimed Chinese level of real GDP growth, it would need to maintain throughout this period extremely low inflation. China would only grow faster than Indonesia, Argentina, and even India in real terms if its annual inflation was extremely low. We present inflation data below in Figure 6.

According to official statistics, between 2000 and 2011, China had one of the lowest accumulated inflation indexes of any major economy. Officially, total Chinese inflation was less than the United States and only slightly more than the UK and Canada. Given the enormous discrepancies between key macroeconomic variables, the Chinese claims to such extraordinarily low inflation appear tenuous. The low growth in inflation cascades directly into Chinese claims of other key variables, specifically real GDP growth. As Chinese nominal GDP growth was average for major emerging market economies, their real GDP growth depends enormously on low inflation. I present real GDP growth data for major emerging market economies below in Figure 7.

Figure 7 shows official data claiming that real Chinese GDP growth has grown significantly faster than other major emerging market economies. The growth in real Chinese GDP is nearly completely dependent upon the claims of emerging market growth, but developed country inflation rates. If Chinese inflation rates increase even marginally, then the Chinese growth story is substantially different between 2000 and 2011.

Estimating Chinese Inflation and Real GDP Growth Between 2000 and 2011

Chinese inflation data suffers from obvious manipulation. To correct this systematic bias, I begin to estimate, using a variety of reasonable assumptions based upon independent verifiable data, improved values of Chinese inflation. It is important to cover a couple of major general data and methodological issues. First, the NBSC throughout most of the time period in question assigned only a 13% weighting to private housing or rental prices, which by a variety of other measures seems extremely low. According to the NBSC, the price basket spent more on "education and cultural articles" and only slightly less on clothing (please click here for more information). The weighting for housing was only increased to 17.2% in 2011. To correct for this, we include a NBSC weighting, a 20% weighting, and a 30% weighting. Second, rather than just adding in the new housing inflation figure, it is also necessary to subtract out the previous number so as to avoid double counting. Third, rather than using straight line and obviously false population weighting data, I weight the pricing data based upon the proportion of rural and urban population and their respective categorization between private housing and renting. This provides us increased accuracy in the inflation data. These are the general assumptions made, but they are also realistic and conservative.

To accomplish our objective of improved estimations of Chinese inflation data, we need to make a number of assumptions in each of the three price scenarios. In Inflation Scenario 1, I make one primary assumption that the price index value of private housing will rise by no less than the price change to renting. If the official price of private housing rises faster that number is used; if the official price of renting is higher, that number is used. As previously noted in third party research and data, private housing prices increased faster than rental prices. This assumption significantly raises the price of private housing, but only to slightly more than the rental price. This happens because in most years, the rental price grew faster than the price of private housing in official NBSC statistics. In Inflation Scenario 2, I make the same assumption as in Inflation Scenario 1, but never allow the annual price change of private housing to drop beneath 4%. This changes the pricing data for five years for urban residents and eight for rural residents. It does not however, change the data enormously. In Inflation Scenario 1, the accumulated price difference between private housing and renting was 91% to 75%. In Scenario #2, the difference is only slightly larger at 99% to 75%. In other words, this is a small and conservative assumption that does not deviate significantly from other data. In Inflation Scenario 3, rather than basing the new numbers on existing NBSC inflation data, I use asset price value with assumptions about the underlying pass through to prices. For the urban private housing price I use Deng et al. (2012) and their annual urban price change from 2004 to 2011 dividing by two as asset prices rise faster than the price index, with all other assumptions from Price Inflation Scenario 2 for years 2000 to 2003. Urban rental values are obtained by dividing the price change from urban private housing in half. Rural private housing price is taken from NBSC data on rural home value with the same discounting factor applied to the price index. Rural rental prices are obtained in the same manner as urban rental prices, dividing the change in rural private housing prices by half. Using the outlined methodology and the IMF local currency GDP deflator, we add in the new inflation data, given the different assumptions with results shown below in Table 6.

Table 6: Estimated Chinese IMF GDP Deflator Change From Improved Housing Price Data

Click to enlarge

Source: National Bureau of Statistics of China, International Monetary Fund, and author's calculation

There are a number of points to make about the GDP deflator adjustments. First, the range is not insignificant. Beginning with the most conservative of assumptions, where private housing prices only keep pace with official rental prices and a 13% weighting for housing, both assumptions we know to be inaccurate, real GDP should be reduced by more than 4%. Given a more realistic assumption with regards to the change in prices and the amount spent by Chinese residents on housing, this implies closer to an 8-12% reduction in real GDP. Translating this new deflation into PPP international dollars, this would reduce real GDP by more than $1 trillion, if we use a mid-range estimate.

There are a couple of concerning points when placing this analysis in context. First, it is disturbing that a statistical body would so obviously manipulate and produce blatantly fraudulent data. Though improved third party data exists and has been used, the claim that the housing component of CPI grew by less than 10% between 2000 and 2011 is nothing less than comical. Second, given the relative ease with which obvious statistical manipulation was found, it is quite likely that less obvious fraud is present in the CPI data. In fact, even a simple perusal of NBSC data reveals even more discrepancies. For instance, the NBSC from 2000 to 2010 assigned a 34% weighting to the food component of price inflation. In 2011, food was reweighted and assigned a 31.8% weighting within the CPI basket. Using solely NBSC data on the price increase of food in China, it becomes quickly apparent that this would require unrivaled intellectual gymnastics. Using NBSC data on the consumer price inflation and the food component, I present the estimated food share of the Chinese CPI below in Figure 9.

According to NBSC data, the food component of the CPI in China was responsible for 99% of inflation between 2003 and 2011. Thinking of this another way, this implies that the NBSC is claiming that the only prices to rise in China between 2003 and 2011 were food prices. To provide one more example, the NBSC claims that food received a 34% weighting throughout most of the decade, the data does not support this. In Table 7, I present the data using the NBSC weighting and the declared annual CPI.

Table 7 - Component Pricing Compared to the NBSC Claimed CPI

Click to enlarge

Source: National Bureau of Statistics and the author's calculation.

The component weighting of the CPI data reveals more inconsistent results. There is an annual discrepancy of .35%. Even when weighting the data according to the NBSC using their data, the numbers fail to reconcile. This simply furthers the discrepancies and problems underlying Chinese statistical data.


China has grown rapidly since its opening in 1979 to become one of the world's largest economies. However, its economic data is flawed at best and it appears to be politically manipulated. The problematic data stems from two primary problems. First, the base data on housing prices is grossly manipulated. Second, it systematically weights it to the lower price categories despite those groups being under-represented in the population. If we adjust the CPI data based upon conservative, reasonable, and independent data or assumptions about the changes in the price of housing in China, it lowers real GDP by a mid-range estimate of 8-12% or more than $1 trillion PPP USD. Given what we obviously locate, it seems likely that much larger revisions to Chinese real GDP and other economic data is needed to produce more reliable statistics.

Comparison Between the Hedonic Price Index and Two Official House Price Indices (2006 Q1-2010 Q4)

Click to enlarge

Source: Wu et al. 2012 Table 4 p. 43

Note: Wu et al. methodologically prefer the Hedonic Model. The Hedonic Model indicates approximately 150% increase between 2004 and the end of 2009. Even more conservative and traditional measures like the weighted or unweighted average indicate a doubling of prices during this time.

Note: While the above figure is not a specific index, it does indicate that the real average annual growth in most cities was above 10% with only one city averaging less than 5%. Considering this is the real price and not the nominal price, it implies that a very small number of major cities witnessed average annual nominal price appreciation of less than 10% between 2006 and 2010.

Appendix 5-Figure of Major Economy Standard Deviation of the Unemployment Rate of Major Economies

Click to enlarge

Source: International Monetary Fund World Economic Outlook and author's calculation

Note: The countries in the figure are Argentina, Brazil, China, Denmark, France, Germany, Hong Kong, Iceland, India, Indonesia, Ireland, Italy, Japan, Korea, Mexico, Netherlands, New Zealand, Nigeria, Norway, Pakistan, Russia, South Africa, Spain, Sweden, Switzerland, Turkey, the United Kingdom, and the United States.