Working Paper: NBER ID: w3935
Authors: Donald Siegel
Abstract: Extending a methodology developed by Lichtenberg and Griliches (1989), we examine the extent of measurement error in two independent indicators of price change: the producer price index (PPI) and the U.S. Census Bureau's unit value relative (UVR). Estimation of factor analytic models is improved by the availability of more accurate and comprehensive proxies for price and quality change within industries and a more complete specification of the econometric model. We find that the UVR is a "noisier measure of price change than the PPI and that the PPI adjusts for approximately 57% of product quality change.
Keywords: measurement error; price change; producer price index; unit value relative
JEL Codes: No JEL codes provided
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| Improper measurement of quality change (L15) | Distortion in estimates of real output (C51) |
| PPI adjusts for quality change (C43) | Measurement of productivity (O47) |
| PPI is a more reliable indicator of price change (E30) | Measurement of productivity (O47) |
| UVR is noisier measure of price change (R19) | Misestimation in productivity growth (O47) |
| Quality adjustments (C43) | Price indices (C43) |
| PPI shows stable relationship with true price changes (E31) | Measurement of productivity (O47) |
| UVR's noisiness (R19) | Significant misestimation in productivity growth (O47) |