The effects of inflation and openness on inequality across alternative monetary regimes

Abstract

This article provides the first empirical evidence by employing an endogenous switching regression model to investigate whether the adoption of inflation targeting (IT) alters the relationships between inflation and inequality as well as between openness and inequality. Using a large panel of countries over the 1980–2010 period, we find that policies aiming at stabilizing prices and opening to trade will benefit more from the reduction of income inequality, and the inequality-reducing effects are more pronounced in the IT countries. As a by-product, we find that the inverted-U relationship between inequality and development, that is, the Kuznets’ hypothesis, holds only in the Non-IT countries, not in the IT economies. The findings are unlikely to be driven by the presence of outliers, the use of estimators and the assumptions of errors distribution.

Keywords:

• inflation targeting
• income inequality
• inflation
• trade openness

JEL Classification:

• C24
• E4
• E5
• O1

Notes

¹ There exists a large literature evaluating the average treatment effects of IT on key macroeconomic variables per se, such as inflation, inflation variability, etc.

² For instance, while Lin and Ye (2007) show that IT exerts no effect on either inflation or inflation variability in industrial countries, Lin and Ye (2009) find that IT has large and significant impact in reducing both inflation and inflation variability in developing counterparts. In addition, Huang et al. (2013, forthcoming) show that income inequality is significantly larger in IT countries across the full sample. However, they also find that the prior harmful effect (more unequal income distribution) of IT is mainly driven by the developing countries subsample. In contrast, IT exerts no discernible effect on income distribution in the industrial countries subsample.

³ Please see Hasebe (2013) for more technical details.

⁴ In the following empirical analyses, further inclusion of additional controlling variables may decrease the number of countries, depending on data availability.

⁵ The full estimation results are omitted for brevity, but are available upon request.

⁶ We consider alternative combinations of Gaussian, Farlie-Gumbel-Morgenstern, Ali-Mikhail-Haq, Plackett, Frank, Clayton, Gumbel and Joe. Please see Table 1 of Hasebe (2013) for a detailed list of these copula functions.