Integrating the empirical tests of the natural rate hypothesis: A meta-regression analysis

A meta-analysis of thirty-four restriction tests from nine studies of the natural rate of unemployment hypothesis (NRU) finds the statistical trace of a false empirical hypothesis. A theme of bias and misspecification among those studies that tend to be more supportive of NRU emerges. When combined with a separate meta-analysis of NRU's falsifying hypothesis, unemployment 'hysteresis' (Stanley 2004a), the natural rate hypothesis may be regarded as empirically 'falsified' (Popper 1959). Monte Carlo simulations validate the meta-regression methods used here to integrate different restriction tests and to identify their limitations.