A note on Romer's openness-inflation relation: the responsiveness of AS and AD to economic openness and monetary policy

Abstract

Temple (2002) empirically challenges Romer's (1993) negative openness-inflation relation on empirical grounds. This article links economic openness to the slopes of aggregate supply (AS) and aggregate demand (AD) to explain why the openness-inflation relation can be ambiguous. Starting with a widely used assumption initiated by Romer (1993) that more open economies face greater output inflation tradeoffs, we demonstrate that greater output-inflation tradeoffs in more open economies (reflected in the steeper AS) induce policymakers to adopt more aggressive optimal monetary policy (reflected in the flatter AD). Empirical results from 15 developed countries’ data support our theoretical explanation on the recent empirical failure in finding the negative openness-inflation relation.

Notes

¹ The main objective of this article is to offer an explanation on the empirical failure of the openness-inflation relation even when Romer's (1993) assumption holds – AS is steeper in more open economies. Therefore, for the ease of analysis, we take Romer's hypothesis as an assumption such that we can directly investigate the relation between openness and both AD and AS empirically. Alternatively, one can model general equilibrium framework to allow the slope of AS to be endogenously determined. We refer interested readers to CGG (2001, 2002) as a good candidate for a general equilibrium model on the optimal monetary policy in relation to openness.

² Leeper and Roush (2003) empirically identify that as long as money demand is not interest inelastic, the money stock and interest rate would jointly transmit monetary policy in the US economy. Thus, ignoring the money stock in the empirical analysis would underestimate the measured effects of monetary policy on output and inflation levels.

³ Since different countries could have different versions of Taylor rule, we make the distinction between two versions possible in our empirical analyses.

⁴ These 19 countries are Australia, Austria, Belgium, Canada, Denmark, Finland, France, Germany, Italy, Japan, Netherlands, New Zealand, Norway, Portugal, Spain, Sweden, Switzerland, the United Kingdom and the United States.

⁵ Note that significant lags of (π _{i , t}  − π _{i, t −1}) are included in the regression to eliminate the serial correlation, if any, in the residuals.

⁶ These four countries are Belgium, Finland, Portugal and Switzerland.