Determinants of tax evasion in Greece: Econometric analysis of co-integration and causality, variance decomposition and impulse response analysis

Abstract

The purpose of this paper is tο examine the causality relationships and the degree of interdependence, between the level of tax evasion in Greece and a set of deterministic factors, using annual data for the period 1995 - 2018. The research methodology employed includes testing for stationarity with the Augmented Dickey-Fuller (ADF) test, cointegration test according to Engle-Granger approach, estimation Error Correction Models (ECM) to investigate the short-run and long-run relationships, variance decomposition and impulse response analysis. The results indicate a significant interdependence, which is an important tool for pursuing a targeted and effective policy to fight tax evasion. More specifically, the survey showed that the level of tax rates, the level of unemployment, the Rule of Laws index, the level of GDP, the level of non-performing loans, the government efficiency, the corruption perception index and the level of final consumption expenditure, affect the size of tax evasion in Greece, significantly. In addition, the results of variance decomposition and impulse response analysis, support the above findings, providing a quantitative representation of the causality relationships between the factors under investigation and tax evasion.