Title : Forecasting and Scenario Analysis of Korean Macro Economy Using a Large Non-linear Bayesian VAR Model

Author : KyuHo Kang(Korea University), Dowan Kim(BOK)

Forecasting Korea's key macro variables and analyzing their sensitivity to external shocks, this study performs a scenario analysis using a large vector autoregressive (VAR) model. Notably, we developed a new large Bayesian VAR model that considers the possibility of gradual structural changes in macro variables, These gradual structural changes are modeled as a stochastic trend, and we introduce a novel Bayesian algorithm that captures the existence of this trend while estimating it for each variable. Our sample data is from the first quarter of 2003 to the fourth quarter of 2022. We use twenty-seven domestic and foreign macro variables for the estimation. The primary findings are: first, we forecast Korea's economic growth for 2023 to lie within the mid-to-high 1% range. Second, a robust US economy will bolster the Korean domestic economy. However, a surge in raw materials and crude oil costs will adversely affect the Korean economy and its price stability.