Output measurement and technological shocks in business cycles

Is the importance of technology shocks in accounting for the fluctuations in output and hours sensitive to the measure of output? Using a vector autoregression, we empirically investigate the contribution of technological shocks to economic fluctuations when output is defined in consumption units, as commonly used in many macroeconomic models, and when output is tabulated according to the Divisia index in the U.S. National Income and Product Accounts (NIPAs). Based on a standard neoclassical growth framework that allows for a mapping of theory into any desired measure of output, we establish that the same restrictions may be used to identify technology shocks in a vector autoregression model regardless of the measure of output. However, our estimation reveals that while the combination of both investment-specific technology shocks and neutral technology shocks accounts for a large portion of the business cycle variability of hours and output, the choice of the measure of output via the use of the associated deflator greatly affects the amplification of the shocks in the responses and the variability of the responses in output and hours.