Productivity growth is carefully scrutinized by macroeconomists be-cause it plays key roles in understanding private savings behaviour,the sources of macroeconomic shocks, the evolution of internationalcompetitiveness and the solvency of public pension systems, amongother things. However, estimates of recent and expected productivitygrowth rates suﬀer from two potential problems: (i) recent estimatesof growth trends are imprecise, and (ii) recently published data oftenundergo important revisions.This paper documents the statistical (un)reliability of several meas-ures of aggregate productivity growth in the US by examining theextent to which they are revised over time. We also examine the ex-tent to which such revisions contribute to errors in forecasts of USproductivity growth.We ﬁnd that data revisions typically cause appreciable changes inpublished estimates of productivity growth rates across a range of dif-ferent productivity measures. Substantial revisions often occur yearsafter the initial data release, which we argue contributes signiﬁcantlyto the overall uncertainty policymakers face.This emphasizes the need for means of reducing the uncertaintyfacing policymakers and policies robust to uncertainty about currenteconomic conditions.
C22, J24, O47
productivity, real-time analysis, data revisions, Greenbookprojections