Measuring the NAIRU – An Structural VAR Approach.docVIP

Measuring the NAIRU – An Structural VAR Approach Hongmei Zhao Abstract In this paper, we investigate the NAIRU in a framework where inflation and the unemployment rate can respond to each other. The NAIRU is defined as the component of the actual unemployment rate that is uncorrelated to inflation in the long run. By this definition, the NAIRU and core inflation can be estimated simultaneously. Our estimation results show that the NAIRU falls dramatically at the end of 1990s and the long run vertical Phillips Curve shifts back from 6.8 per cent before 1997 to 4 per cent afterwards. Introduction At the end of 1990s, low inflation and low unemployment arising in the U.S. focused people’s attention to the time-varying NAIRU (Non Accelerating Inflation Rate of Unemployment) again. According to the Phillips Curve theory, inflation and unemployment cannot move towards the same direction. For this unusual economic phenomenon at the end of 1990s, a popular explanation is that the NAIRU has fallen with the lapse of time so that unemployment can stay at low without the danger of raising inflation (symposium on the NAIRU, JPE, 1997). Did the NAIRU really fall? This is the purpose of our paper – measure the NAIRU from a new point of view. If NAIRU does change as time goes on, we hope our method can give a more precise estimation to the NAIRU, thereby providing a more reliable reference index for the central banks to set the inflation target. The definition of the NAIRU originates from the Phillips Curve theory. The meaning of NAIRU itself is the unemployment rate when inflation tends to be stable. Traditionally, the NAIRU is estimated from Phillips equation according to its definition (Gordon, 1997; Staiger, Stock, Watson, 1997). But looking back the literature of NAIRU estimation, a perplexing problem is the existence of too wide standard error bands. To some extent, this implies the uncertainty of estimating the NAIRU. Laubach (1997) may shed some lights on the reason for