China's GDP data reliability has been in question for years. Over the last couple of decades questions regularly arise on whether or not to believe official growth numbers out of Beijing. Recently the subject has gained traction in the news as stock market volatility elevated worries of China's economic prospects.
Even in the official data there is evidence that underlying growth is below the stated 7% numbers for Q1 & Q2. As I have mentioned in a previous blog, the froth from a financial market surge pushed China's growth numbers to 7% so far this year. Underling growth readings from the official data point to a GDP growth more likely around 6.5% so far this year. See my posting 7% GDP growth? Probably not. for more details.
Certainly if you look at other EM country GDP numbers over time, China's data looks suspiciously smooth. But while Beijing can be accused of smoothing the data, or suffering the same difficulties that other EM countries have measuring the entire economy during a rapid transition, recent charges that China's numbers are completely fabricated to hide a hard landing are less credible. Smoothing, calculation problems, distortions from some local leaders fluffing the data, and the occasional transient froth from specific sectors means the GDP number are far from total reliability, but they can still be in the right ballpark.
Few studies have been done on the reliability of China's GDP data. The most important that I know of was conducted by the Fed Reserve Bank of SF in an economic letter from March 25th 2013. After using alternative data from inside and outside of China they concluded that the GDP numbers were not unreliable stating that "These alternative domestic and foreign sources provide no evidence that China’s economic growth was slower than official data indicate.", and that "...these models suggest that Chinese growth has been in the ballpark of what official data have reported." The conclusion ruled out policymaker complete fabrication.
But in the end, the reality is that China's overall GDP number is of little use to most folks anyway, and a more nuanced analysis than in the past is required to know what is going on in individual sectors and the economy as a whole.
Much was made of future Premier Li Keqiang's WikiLeaks memo stating his distrust of official GDP numbers back in 2007 when he was running the Liaoning province. He instead favored using rail freight data, electricity output, and bank loans to gauge the economy. Certainly in his province of Liaoning, a significant supplier of ore and heavy industry, a decade ago those three numbers were good indicators. But in today's China where services account for over 49% of economic output and industry around 42%, those three numbers are completely unrepresentative of the broad economy (see my posting The "Li Keqiang" Index: Why is that still a thing?). The 2010 leaked Li Keqiang memo spawned some of the first alternative GDP calculations. A version of the "Li Keqiang" index is on the chart to the right.
There have been a wide variety of alternative GDP measures released over recent years. Some have very explicit ingredients, like the "Li Keqiang" index, which uses freight, electricity output, and bank loan data. But most alternative measures from research firms are black boxes, which is a concern. China's economic data is rooted in its communist planning past, with its main statistics bureau highly reliant on reporting numbers for heavy industries and the physical output of tangible things. As a result, most of China's economic data has to do with industry, with very little data available for crunching service sector statistics. My worry is that most alternative measures use the easy-to-get industrial data, therefore missing roughly half of economic output or more in the calculations.
Official data is 7%, around 6.5% if you strip out the froth from the financial sector early in the year. Bloomberg news service has its own alternative measure for growth, and its GDP tracker is running 6.6% as of July 2015. Capital Economics has the number running somewhere around 5%. Lombard Street Research pegged growth at 3.8% in Q2, indicating a hard landing. There are a number of alternative GDP measures out there to choose from, which means nowadays both soft landing and hard landing forecasters seem to have evidence on their side.
For my own monthly GDP measure, I calculate the numbers for industry and service GDP sectors independently using a wide variety of economic data and principal components analysis. Then, I combine the measures based on respective weightings in overall GDP. My indicators point to July growth running at 6.5% overall, with 5.6% from the industrial sector, and 7.8% from services.
Unreliable GDP numbers do not equate to hard landing
One implication of the rise of skepticism in China's numbers is that hard landing doomsayers and collapsists can now say that they have been proven right. Pundits, like Gordon Chang who has been warning for 15 years that a China collapse is imminent, are using data skepticism to bail out their own bad predictions. With GDP skepticism on their side, hard landing doomsayers and collapsists have now begun to dominate much of the China conversation.
There are plenty of indications that do not support the dramatic hard landing GDP numbers indicated by the more pessimistic alternative measures like Lombard's. Number crunching some data from outside of China, such as trade data from Australia and Brazil, does not confirm the dramatic declines by the more pessimistic alternative GDP measures (see my chart on the right, a generic unscaled combination of various data). In fact, on a volume basis, shipments from key exporting countries do not confirm a hard lading scenario. Ausrtalian iron ore shipments on a volume basis, Japan and Korea exports to China, and Brazilian shipments, all point to weaker demand for sure, but not a dramatic collapse of growth. The private Caixin manufacturing PMI is running at 47.3, indicating contraction in manufacturing, but far from the PMI numbers in the 30s in 2008 that would indicate China is in economic crisis territory. In the past, the general rule was 8% growth was the threshold for supporting employment, and anything below that would lead to unrest and massive unemployment. If China is growing well below half of the threshold widely accepted only 4-5 years ago we would expect to see problems from massive joblessness. Pollution is still the number one cause of social unrest, and massive unemployment has not been seen. Growth is certainly slowing, probably more than official numbers indicate. But, many other information sources do not confirm the hard landing indicated by some alternative GDP measure providers.
Stop looking at overall GDP for indications of economic activity
The overall GDP number might give us an indication of whether Beijing needs more or less stimulus, and frequently has an effect on risk assets around the world. But, because of restructuring and a highly uneven slowdown, China's 7% growth number is less relevant than it once was. For years Beijing has told us two things: They are going to restructure to quality growth (services and consumption playing a bigger role), and policymakers want to move growth to a lower sustainable trend rate. The only way to achieve those two goals is by letting industry and fixed investment drop, which is what we have seen, while maintaining rapid growth in services and consumption. Heavy industry has dropped like a bag of hammers, and services and consumption have remained strong, making economic growth highly uneven. A commodity exporter gets little use from the overall GDP in such a bifurcated growth scenario, as does an online retailer for example. Online retail sales are growing at almost 50%, all the while steel production is down almost 5% from last year.
In short, Beijing is probably guilty of smoothing data, and/or releasing flawed data as it struggles with the difficulties of measuring a rapidly changing vast economy. There is a small possibility that data is fabricated, as China's leadership will go to great lengths to maintain control over their perceived governing ability. But, a nefarious attempt to completely misrepresent economic growth in order to mask a hard landing is less probable as other information sources do not confirm that finding. If we accept the San Francisco Fed's 2013 analysis, even if we assume China's GDP numbers are flawed and not completely reliable, they are at least in the right ballpark.
But, GDP is not a useful number for China research anyway. A more nuanced and complex analysis is required to measure what is happening in the Chinese economy going forward.