The survey based monthly US ISM production index and Eurozone manufacturing PMI output index provide early information on industrial output growth before the release of the official industrial production index. I use the Carlson and Parkin probability method to construct monthly growth estimates from the qualitative responses of the US ISM production index and the Eurozone manufacturing PMI output index. I apply the method under different assumptions on the cross-sectional distribution of output growth using the uniform, logistic and Laplace distribution. I show that alternative distribution assumptions lead to very similar estimates. I also test the performance of the different growth estimates in an out of sample forecasting exercise of actual industrial production growth. All growth estimates beat a simple autoregressive model of output growth. Distribution assumptions again matter little most of the time except during the financial crisis when the estimates constructed using the Laplace distributional assumption perform the best. My findings are consistent with recent findings of Bottazzi and Sechi (2006) that the distribution of firm growth rates has a Laplace distribution.
Carlson-Parkin method Diffusion index forecasting ISM PMI purchasing managers’ surveys qualitative response data