Online appendix to “Productivity and Potential Output Before, During, and After the Great Recession”

This appendix provides supporting materials and analysis. • Appendix A discusses the role of data revisions in overstating the strength of productivity growth early in the Great Recession. • Appendix B describes the data used in the paper. • Appendix C discusses why, in the context of the Basu, Fernald, Oulton, and Srinivasan (2003) model, one expects the industry slowdown to be most pronounced in IT-intensive industris. • Appendix D presents variants of the neoclassical growth model in order to derive the link between exogenous technology and endogenous labor-productivity growth. It also discusses the empirical fit of the model over the post-war period, which shows the gains from using the multi-sector model rather than a one-sector model, and to allowing for land growth in the model. Appendix A: Role of data revisions Real-time data obscured the slowdown in trend, and overstated productivity’s strength early in the recession. Figure A1 shows labor productivity by vintage. The dates correspond to the annual (or, in 2009 and 2013, comprehensive) NIPA revisions; these revisions incorporate additional source data for previous years. Almost every revision since 2005 has lowered the path of labor productivity, with most revision to output (the numerator). These revisions made the slowdown more apparent. Real-time data also overstated the strength of labor productivity growth early in the recession. Until the 2010 revision, productivity appeared to have risen sharply and steadily throughout the recession. The sizeable downward revisions suggest some of the challenges of doing analysis in real time