---

Amid rising concerns about deflation in the Eurozone, it is very important to be able to model and estimate corresponding risks to the economy. This work attempts to determine if Bayesian vector auto-regression (BVAR) model produces accurate inflation forecasts and can be used for indicating deflation. It is argued that BVAR improves the forecasting performance of the larger-scale models. Point forecasts of the GDP deflator growth in the US and Japan show smaller estimation errors, while the historical values of this inflation measure appear to be in the same range with density forecasts. The ability of BVAR model to indicate deflation risks at the proper time could be proved useful for the Eurozone targeting this problem.

---

Liliya Maliauka, MSc. in Economics: A graduate from the University of Bonn. A Knowledge Analyst for the Financial Institutions Practice Area at The Boston Consulting Group GmbH, Frankfurt am Main.

---