Ca' Zorzi, Michele Chudik, Alexander Dieppe, Alistair
Year of Publication:
ECB Working Paper No. 995
This paper examines two competing approaches for calculating current account benchmarks, i.e. the external sustainability approach á la Lane and Milesi-Ferretti (LM) versus the structural current accounts literature (SCA) based on panel econometric techniques. The aim is to gauge the medium term adjustment in current account positions that may be required in some central and eastern European countries. As regards the LM approach, we show how the outcome is especially sensitive to (i) the normative choice for external indebtedness and (ii) the decision to exclude the foreign direct investment subcomponent from the NFA aggregate. Turning our search to the SCA approach, we assess its sensitivity to model and parameter uncertainty by setting different selection criteria to choose amongst the over 8000 possible combinations of fundamentals. Furthermore, to test the robustness of our findings we combine all models, attaching to each a probability (Bayesian Averaging of Classical Estimates). We show both the LM and SCA methodologies are not immune from severe drawbacks and conceptual difficulties. Nevertheless pulling together the results of both approaches point to the countries that may need a current account adjustment over a medium term horizon.
Capital flows central and eastern Europe current account financial integration model combination Model uncertainty panel data