The paper applies quantile regression technique, specifically, quantile vector autoregression to stochastic DSA and the generation of debt fan charts. Stochastic approach to DSA typically uses standard ordinary least squares (OLS) vector autoregression (VAR) and fan charts to depict the upside and downside risks surrounding public debt projections due to uncertain macroeconomic conditions. These VAR models rely on constant coefficients and random variables that are independent and identically distributed. However, empirical evidence suggests that macroeconomic variables are characterized by nonlinearities and asymmetries that linear regression models, such as OLS VAR, cannot capture. Several papers attempt to show how such nonlinearities can be accounted for by using quantile regression methods. Quantile regression allows for varying parameters for each quantile and facilitates the analysis of asymmetric dynamics. It is also a natural environment for stress testing exercises by estimating the reaction of the endogenous variable to tail shocks or future quantile realizations.

JEL codes: H63, C31