Have we been mismeasuring Chinese inflation?

This is from a reputable source, although it is not an established consensus conclusion:

That is the tricky task that Emi Nakamura, Jón Steinsson and Miao Liu of Columbia University set themselves in a recent study. They start with an economic law first observed by a 19th-century statistician, Ernst Engel: richer households spend a smaller share of their income on food. Thus as a household becomes richer over time, its spending pattern should match that of households who were equally rich a year or two before.

But in China, they discovered something different. They compared urban households in 2006 with households that were, according to the official figures, equally rich in 2008. They discovered that the later households were devoting 3-4% more of their budgets to food. Perhaps they were not quite as rich as their 2006 counterparts, after all.

That would suggest the Chinese economy tanked more during the financial crisis than is usually believed.  This part I believe for sure:

Moreover, it turns out that China’s official figures do not always understate inflation. From 1996 to 2006, they actually exaggerated it in every year but one, according to the same method. As a result, urban consumption was growing even faster in this period than the official statistics conveyed. China’s policymakers had more to boast about than they knew.

The inflation figures calculated by the three economists are also remarkably well correlated with the official numbers. They rise and fall in unison. It is just that the unofficial figures rise faster and fall further. The trio conjecture that two competing biases are at work. First, new goods are often of higher quality than the ones they replace, but their price is the same. That would explain why China overstated inflation before 2007. More subtly, statisticians sometimes fail to grasp that new goods are merely upgrades of existing ones. So they invent new categories; that biases inflation towards zero. As a consequence, China’s official figures “present a smoothed version of reality,” the authors write.

You can read more here.  You will find the pdf of the paper here.