EU’s CAM and Potential GDP Estimates. Some Inconsistent Results in the Case of Romania

The paper quantifies the differences in the estimated potential GDP due to corrections to the used data series. The corrections are necessary, since the capital series used by the DG ECFIN, in the absence of a coherent series, was constructed based on an approximation of the initial stock, using the perpetual inventory method. Recently, the INS has finished recreating the capital series, and the differences in values are quite significant. Secondly, the CAM uses a value of 0.65 for the elasticity of labor, selected based on the average share of wages in the EU15 in the 1960 period. The average value for Romania computed for the 1995-2018 period from the statistical series is 0.45, significantly lower than the above-mentioned value. We used the CAM for computing an alternative potential GDP essentially the GAP interface provided, in order to make sure that the differences were not induced by differences in the estimation process (Kalman filter, Hodrick Prescott filter), but by the differences in the data. The estimated potential GDP using the new data proved to be remarkable similar to the DG ECFIN, but the output gap captured some differences. The authors’ computed gap is smaller in absolute value than that of the EU’s. Although the sign of the gap is mostly similar, the DG ECFIN gap is positive, while the authors’ gap is negative (although small) at the end of the interval (2017-2019). Since the sign of the GAP is most important in assessing the fiscal stance and performances of the EU countries, determining the European Commission recommendations on the economic policies that a country should adopt, according also to the matrix of fiscal requirements, the results raise some questions regarding the accuracy of the DG ECFIN estimations of the potential GDP.