Can firm micro data match macro trends? Comparing MultiProd and STAN

Better understanding about the drivers of aggregate productivity and wage inequality requires data that offer a representative picture of the underlying firm-level heterogeneity but are, at the same time, able to reproduce patterns observed in aggregate data. The OECD MultiProd project aims to gener...

Full description

Saved in:
Bibliographic Details
Published inOECD Science, Technology and Industry Working Papers Vol. 2019; no. 2; pp. 1 - 20
Main Authors Bajgar, Matej, Berlingieri, Giuseppe, Calligaris, Sara, Criscuolo, Chiara
Format Paper
LanguageEnglish
Published Paris OECD Publishing 14.06.2019
Organisation for Economic Cooperation and Development (OECD)
SeriesOECD Science, Technology and Industry Working Papers
Subjects
Aggregates
Datasets
Decomposition
Economic models
Economics
Employment
Heterogeneity
Macroeconomics
Population
Productivity
Regression analysis
Retailing industry
Science
Trends
Value added
Variables
Wages & salaries
Online AccessGet full text

Cover

More Information
Summary:Better understanding about the drivers of aggregate productivity and wage inequality requires data that offer a representative picture of the underlying firm-level heterogeneity but are, at the same time, able to reproduce patterns observed in aggregate data. The OECD MultiProd project aims to generate such data by collaborating with a network of national experts who apply a harmonised statistical code to representative business microdata across a large number of countries. This paper compares the project’s output to the OECD STAN database to test to what extent MultiProd data can be taken as reflecting the aggregate economies in question, and if they are able to reproduce patterns observed in aggregate data across years, industries and countries. The results suggest that (1) MultiProd captures a major part of gross output, value added and employment in most of the countries covered; and (2) MultiProd reproduces aggregate patterns relatively well, with median correlations over time, across industries and across countries between 0.75.
ISSN:1815-1965
DOI:10.1787/2b0ac915-en