Abstract: We develop an estimation method that allows us to identify the slope of the Phillips curve using a bottom-up approach, as opposed to trying to estimate it using aggregate data. We construct a novel dataset at the firm-product-level, which enables us to observe prices, real output, and estimates of marginal costs for the entire manufacturing sector of Belgium over a thirty-year period at a quarterly frequency. By employing this detailed high-frequency data, we estimate structural forward-looking pricing equations that align with the theoretical micro-foundation of the New Keynesian Phillips curve. In particular we are able to identify the primitive parameters that govern the slope, including the degrees of price rigidity and of strategic complementarities in price setting. Our results indicate a significant sensitivity of inflation to changes in economic activity, where the latter is measured by real marginal cost as the underlying theory suggests. We estimate a slope coefficient for the (marginal cost based) Phillips curve of approximately five percent, which is nearly one order of magnitude larger than the most recent estimates in the literature that rely on either regional or aggregate data, and on the output gap or unemployment as the real activity indicator.
If you would like to be added to the distribution list or for further details regarding this seminar, please contact Erica Loh at el2378@nyu.edu