Markup and price dynamics:linking micro to macro

We analyze the aggregate markup of a small-open economy, Belgium, using a firm-level dataset that includes all non-financial, private firms. The dataset covers the period 1980-2016 and merges the annual firm accounts over three periods when firms faced different reporting thresholds for the key variables we use. After harmonizing the data, we find that for the median firm the revenue share of service intermediates doubles, to some extent at the expense of in-house employment. As this general patterns holds true for the vast majority of firms and all sectors of the economy, we need to control for it in the calculation of our firm-level markup estimates. We document increasing markups in the overall economy throughout the first fifteen years of our sample, 1980-1995, and a continued rise in manufacturing until the early 2000s. In the remaining years, the aggregate markup, although cyclical, remained relatively stable. These patterns are driven by the dynamics in the sales-to-expenditure ratio, with only a small role for changes in the technology parameters. Two decompositions illustrate that the aggregate pattern masks systematic dynamics at the sector and firm level. We find that in periods where the aggregate markup rises—for the full economy or for one of the major sectors—it is almost entirely due to the within component, i.e. firm-level markup growth. In periods where the aggregate markup is stable, the average hides a strong process of reallocation. Firms or sectors with high markups increase their market share, which raises the aggregate markup, but this is dominated by a negative correlation between changes in market share and markups, which depresses the aggregate.