Alternative output gap estimates for the Maltese economy

Abstract

This dissertation sets to identify reliable real-time estimates of the output gap for the Maltese economy that are relevant for policymaking, through the utilisation of a real-time database for the period between 2004 and 2017. The study examines and tests for various methods based on univariate and multivariate filters and Principal Component Analysis. The study also produces estimates from a suite of models, which is constructed by taking the mid-point of the full range of estimates identified. The estimates produced under these methods are then compared with the output gap estimates produced for Malta by the EU’s Commonly Agreed Methodology. The comparison is done on the basis of the stability of real-time estimates, their plausibility in key turning points, their ability to explain inflation and their complexity. Compared to previous studies and in line with a-priori expectations, the HP filter with a lambda of 100 and 6.25, the Kalman filter with drift and the Principal Component Analysis resulted in consistent estimates. While, the univariate Kalman filter with no drift and the multivariate Kalman filter with drift resulted in unfavourable results and were thus excluded from the suite of models. The results from testing indicate that in comparison to the output gap estimates produced by the EU’s Commonly Agreed Methodology, the estimates derived by the suite of models approach are stable, produce estimates that are more plausible, and are less complex to estimate, with the exception of the explanatory power of inflation. In line with the crucial role of structural budget balance measure within the EU’s fiscal policy framework and the measure’s reliance on output gap estimates, the suite of models framework developed within this study may aid policymakers in improving on the stability, plausibility and timeliness of real-time output gap estimates.