Evaluation of Panel Data Models: Some Suggestions from Time Series

Two Deficits and Economic Growth: Case of CEE Countries in Transition

“…In this paper we applied the method of evaluating panel datasets developed by Granger and Huang [14]. This approach focuses on the forecasting properties of examined models rather than on significance tests (as in the case of the traditional approach).…”

“…If model (5) forecasts a change in GDP more accurately than model (6), one may claim that information on the past values of government surplus/deficit is indeed important [14]. In other words, G_BALANCE Granger causes economic growth in the countries listed in group I.…”

“…Following papers of Granger and Huang [14], Weinhold and Reis [37], Pér-ez-Moreno [31] and Gurgul and Lach [16], we used two forecast-based testing procedures to test for Granger causality in the discussed framework:…”

“…PROCEDURE I is not always as powerful as PROCEDURE II, however, it is robust to any covariance between and heteroscedasticity of the errors [14]. For the sake of the comprehensiveness of our research we additionally applied a traditional in-sample Granger causality procedure:…”

“…Barro [14] in his well-known paper demonstrated that economic agents rationally expect that a higher fiscal deficit will cause higher taxes in the future. The expected measure is increasing by the current savings.…”