Working Paper: NBER ID: w19893
Authors: Emi Nakamura; J. N. Steinsson; Miao Liu
Abstract: China has experienced remarkably stable growth and inflation in recent years according to official statistics. We construct alternative estimates using detailed information on Chinese household purchasing patterns. As households become richer, a smaller fraction of total expenditures are spent on necessities such as grain and a larger fraction on luxuries such as eating out. We use systematic discrepancies between cross-sectional and time-series Engel curves to construct alternative estimates of Chinese growth and inflation. Our estimates suggest that official statistics present a smoothed version of reality. Official inflation rose in the 2000's, but our estimates indicate that true inflation was still higher and consumption growth was overstated over this period. In contrast, inflation was overstated and growth understated during the low-inflation 1990's. Similar patterns emerge from the data whether we base our estimates on major categories such as food or clothing as a fraction of total expenditures or subcategories such as grain as a fraction of food expenditures or garments as a fraction of clothing expenditures.
Keywords: Chinese growth; inflation; Engel curves
JEL Codes: D12; E21; E31; O11
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| Official inflation statistics in China have been systematically overstated during the late 1990s (E31) | True inflation was higher than reported (E31) |
| As households become wealthier (D19) | They allocate a smaller fraction of their expenditures to necessities and a larger fraction to luxuries (D12) |
| Downward shift of Engel curves from 1995 to 2000 (D12) | Measured expenditure growth is biased downward and inflation is biased upward (E31) |
| Upward shift of Engel curves from 2006 to 2008 (D12) | Official growth statistics were too high during this period and inflation measures were too low (E31) |
| Engel curve-based inflation estimates (E31) | Highly correlated with official statistics but show larger swings (E32) |
| Significant slowdown in urban consumption growth in 2007 and 2008 coinciding with a spike in inflation driven by supply shocks (F61) | Pervasive nature of Engel curve shifts supports mismeasurement in official statistics (F62) |