Working Paper: NBER ID: w11746
Authors: Ricardo Reis
Abstract: This paper shows that conventional measures of cost-of-living inflation, based on static models of consumption, suffer from two problems. The first is an intertemporal substitution bias, as these measures neglect the ability of consumers to borrow and lend in response to price changes. The second problem is the omission of intertemporal prices, which capture relevant relative prices for a consumer who lives for many periods. The paper proposes a dynamic price index (DPI) that solves these problems. Theoretically, it shows that the DPI is forward-looking, responds by more to persistent shocks, includes assets prices, and distinguishes between durable and non-durable goods' prices. A constructed DPI for the United States from 1970 to 2008 differs markedly from the CPI, it is close to serially uncorrelated, it is mostly driven by the prices of houses and bonds, and is twice as high as the CPI in 2008.
Keywords: Dynamic Price Index; Inflation Measurement; Intertemporal Substitution
JEL Codes: E31; C43; J26; D91
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| conventional static measures of inflation (E31) | intertemporal substitution bias (D15) |
| intertemporal substitution bias (D15) | consumers adjust consumption patterns over time (D15) |
| DPI (Y10) | more accurate measure of inflation (E31) |
| DPI (Y10) | less serially correlated than static measures (C22) |
| DPI (Y10) | influenced significantly by prices of durable goods (D12) |
| higher asset prices (G19) | increase effective price of future consumption (D15) |
| increase effective price of future consumption (D15) | necessitate adjustments in wealth to maintain welfare (D69) |
| DPI (Y10) | higher dynamic inflation rate compared to CPI (E31) |
| DPI and CPI (C43) | low correlation (0.3421) (C10) |