The Convergence of Price–cost Margins

Abstract

This paper provides estimates of sectoral price–cost margin (PCM) trends in thirteen OECD countries over the last three decades, once controlled for inflation and cycle effects. It concludes with the absence of a decrease in average PCM. More strikingly, it establishes a clear pattern of PCM convergence both across countries and sectors. This convergence means that high margins have shrunk and low margins grown. Better capital market efficiency might be a driving force in the PCM convergence. These results point to a need to search for factors countervailing the pro-competitive effect on markups.

Appendix

1.1 Impacts of inflation and cycles

Price changes and observed PCMs are estimated to be negatively linked: a decrease in inflationary pressures induces larger (observed) PCMs, as in Blanchard (1997). This is consistent with price stickiness forcing firms to cut their margins in the face of unfavourable cost developments. However, the variable DEFL is significant for only five of the 13 countries at the 5% level, which is likely to be due to the correlation with oil price variables over the period. When it is significant, it implies that a decrease of 10 points in the GDP deflator, not uncommon since 1980, leads to an increase in observed PCMs between 1 and 2 points. Moreover, the two oil price variables are jointly very significant. Oil price changes between 1980 and the end period entail, beyond the DEFL impact, an average increase of 0.008 in observed PCMs for all countries, ranging from −0.007 for the UK—the only negative point—to 0.040 for Japan, very dependent on oil. Overall, disinflation between 1980 and 2000 has triggered an average transitory increase in PCMs of 0.015 across countries.

At the sectoral level, the EMPCYC estimates confirm the counter-cyclicality of markups, supported empirically by Bils (1987) and Oliveira Martins and Scarpetta (2002) among others. On average per country, the effect of EMPCYC is counter-cyclical for ten countries, being significant for nine of them at the 90% confidence level. Overall, a cycle materialising in an increase of 1% above trend in sectoral employment induces a decrease of 0.0008 in PCMs. On the other hand, the estimated impact of the macroeconomic cycle, through the GAP variable clearly leans towards the pro-cyclicality of markups. This may be due to some externality in demand and is consistent with the observed pro-cyclicality of accounting profits. From the latter observation, scepticism about the counter-cyclicality of markups is implied in Christiano et al. (1997). On balance, these estimates may provide an explanation for why the debate concerning the cyclicality of markups remains unresolved. There may be a supply-driven counter-cyclical partial equilibrium effect dampened by a pro-cyclical general equilibrium one. On average across countries, an increase in the output gap of 1 point of GDP results in an average increase in PCMs of 0.002/0.003. Although the average sensitivity is three times larger than the EMPCYC one, employment at the sector level could fluctuate much more than at the country level.