Use of Deflators in Business Surveys: An Analysis Based on Italian Micro Data
Leandro DAurizio, Raffaele Tartaglia-Polcini,
This article aims at evaluating the deflation method used in the Bank of Italys business surveys. Here the deflators of investments and revenues for each business are individually collected. Currently, deflation for computing changes of monetary values at constant prices is based on sector averages of individual deflators. The option to deflate individual changes with the corresponding individual deflator is also explored. Efficiency gains appear to be linked to the existing correlation between nominal change and the deflator chosen and to the extent of measurement error. Some theoretical results are exposed, linking the deflation method used to the classical Laspeyres and Paasche formulae. Several estimators of real rates of change of revenues and investments, based on individual and average deflation, are then selected and tested in a 6-year simulation study. In the absence of error, the individual deflation of revenues significantly decreases the MSE of the estimators. When unbiased and symmetrical measurement errors in deflation are accounted for, regression-based average deflators and individual deflators of revenue rates of change perform better than the other deflation techniques in terms of MSE of the revenue real growth rate.
Deflation, rates of change, price indices, mean square error, simulation study, measurement error, reliability theory