Estimating Nairu For The Romanian Economy

The paper tries to estimate the natural rate of unemployment in the case of Romania, concentrating on implementing an econometric schedule in order to estimate the value of the NAIRU. One of the major problems in literature is that of identification, and the solution could be in finding certain instrumental variable to be correlated with unemployment, but non-correlated with supply shocks. Practically, to identify such valid instruments is very difficult and, consequently, only few studies are reporting concluding results. However, there are many economists that questioned the assumption of a constant NAIRU. A growing literature seeks to estimate the path of a time-varying NAIRU. This literature is based on the idea that movements in U* are long-term shifts in the unemployment-inflation relationship, while the shock v includes short-run fluctuations. Some authors estimate U* by positing a stochastic process for U* (such as a random walk) and a stochastic process for v (such as white noise) and then using a statistical procedure that separates Phillips-curve shifts into these two kinds of shocks. We use an approach that is simpler but yields similar results. Firstly, we are supposing that it is given the value of parameter a, which expresses the slope of the unemployment-inflation trade-off. Then, we rearrange terms to obtain the following equation U*+(v/a)=U+(Dp/a). Its right-hand side can be computed from the statistical data, generating in this way an estimate of U*+(v/a), which in fact measures the shifts in the Phillips curve. Within this sum, U* represents the longer-term trends and v/a is proportional to the short-term shocks. Therefore, it is natural to try to extract U* from U*+(v/a) using a standard approach to estimate the trend in a series. As a rule, in literature it is used the Hodrick-Prescott filter. The HP filter is a generalization of a linear time trend that allows the slope of the trend to change gradually over time. Formally, the HP filter minimizes the sum of squared deviations between the trend and the actual series, with a penalty for curvature l that keeps the trend smooth. In the case in which there is no penalty, the filter will yield the original series; but when the penalty is very high, it will yield a linear time trend. To implement this procedure, we must choose two parameters: the Phillips curve slope, a, and the smoothing parameter, l, respectively (this makes that the trend, U*, be smoothed and not with large oscillations, by replacing the ordinary procedure of fitting every movement in U*+(v/a)). The choice of this parameter is largely arbitrarily. In fact, the HP filter is equivalent to an interpolation method. Therefore, given a time series, it is natural to consider as candidate every other method permitting to estimate a smooth trend. In our exercise on the Romanian economy during the transition period, we used three different function-filters: regress, loess, and ksmooth. On the basis of simulations, one may see the unfavorable impact of a positive difference between the effective unemployment rate and NAIRU on inflation dynamics (Dp). It is demonstrated the points in area (2D space) DU-Dp are distributed in the sectors II and IV (denoted in trigonometric sense) over the right line transcending the origin of the coordination axes. Eventual differences (the evading from the two mentioned sectors) can be attributed to the short run supply shocks.