A Nonparametric Approach to Evaluating Inflation-Targeting Regimes

We use a variety of nonparametric test statistics to evaluate the inflation- targeting regimes of Australia, Canada, New Zealand, Sweden and the UK. We argue that a sensible approach of evaluation must rely on a variety of methods, among them parametric and nonparametric econometric methods, for robustness and completeness. Our evaluation strategy is based on examining two possible policy implications of inflation targeting: First, a welfare implication and second, a real variability implication. The welfare implication involves evaluating a utility function, and tested by testing whether (1) the distributions of the levels and the growth rates of private consumption and leisure per capita remained unchanged under inflation targeting, i.e., first-order stochastic dominance; and (2) testing a linear combination of consumption and leisure per capita, where the parameter describing the utility of leisure or the relative preference of leisure is calibrated. Then we introduce nonparametric univariate and multivariate statistical methods to test whether the first and second moments of a variety of real variables, such as the real exchange rate depreciation rate, real GDP per capita growth rate in addition to private consumption per capita and leisure per capita growth rates, remained unchanged under inflation targeting, decreased or increased significantly. There seems to be some evidence of increased welfare under inflation-targeting regimes, but no concrete evidence is found that inflation targeting policy, in general, reduces real variability. Some cross country differences are also found.