Summary for non-specialists(167 kB)  

This study uses vector auto-regression (VAR) models and a panel fiscal reaction function (FRF) to simulate debt ratios for fifteen EU Member States according to four regimes which are the product of the type of errors (normal or bootstrapped) with the assumption on the structural primary balance (unchanged or determined by a panel FRF). This methodology should be used to make probabilistic assessments on the debt ratio rather than for providing point estimates. Results suggest that debt ratio paths are not normally distributed being positively skewed, and; primary balances show "fiscal fatigue" and partial mean reversion to historical trends. Debt sustainability scenarios should also be run using a FRF or some equivalent "mean reversion" hypothesis.