Working Paper: CEPR ID: DP17521
Authors: Ippei Fujiwara; Kiminori Matsuyama
Abstract: It has been well-documented that the Phillips curve has flattened, making central bankers wary of the reduced effectiveness of monetary policy to achieve price stability. There has also been a growing concern about higher market concentration and the rising profit margins and markup rates. Are these two events merely coincidental or causally related? To address this issue, this paper extends the canonical New Keynesian model to introduce markup-rate changes caused by endogenous entry, by using Homothetic Single Aggregator (HSA), a class of flexible and tractable homothetic demand systems, which contain CES and Translog as special cases. Under Marshall’s second law of demand (i.e., the price elasticity of demand goes up with its price) and the third law of demand (i.e., the rate of increase in the price elasticity goes down with its price), a higher entry cost, which leads to market concentration, causes the flattening of the Phillips curve for two reasons. First, market concentration leads to a rise of the markup rate and a decline of the pass-through rate, which structurally reduces the slope of the Phillips curve. Second, an endogenous change in the number of firms generates an endogenous cost-push shock through strategic complementarity. Due to this endogeneity of the cost-push shock, a naive regression of the inflation rate on the real marginal cost leads to the negative omitted variables bias, which is amplified with more concentration. This weakens the estimated relationship between the inflation rate and the marginal cost. These two reasons, one structural and one observational, together can go a long way toward understanding the flattening of the Phillips curve.
Keywords: New Keynesian Phillips Curve; Market Concentration; Monopolistic Competition; Endogenous Entry; HSA; Variable Markups; Pass-through Rates; Supply Side Effects of Monetary Policy; Omitted Variable Bias; Cyclicality of Markups
JEL Codes: E31; E52
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| increase in market concentration (L19) | structural reduction in the slope of the Phillips curve (E31) |
| increase in market concentration (L19) | increase in the markup rate (E43) |
| increase in the markup rate (E43) | reduction in the slope of the Phillips curve (E31) |
| increase in market concentration (L19) | decrease in the pass-through rate (E43) |
| decrease in the pass-through rate (E43) | reduction in the slope of the Phillips curve (E31) |
| naive regression of inflation on real marginal cost (E31) | negative omitted variable bias (C20) |
| increase in market concentration (L19) | exacerbation of negative omitted variable bias (C21) |
| negative omitted variable bias (C20) | weakened estimated relationship between inflation and marginal cost (E31) |