Bayesian Estimation of a Small Open Economy New Keynesian DSGE-VAR Model: The Impact of Inflation Targeting in the Case of Vietnam after the Global Financial Crisis

Abstract

This thesis evaluates whether Vietnam’s monetary policy following the Global Financial Crisis can best be described as an inflation-targeting (IT) regime augmented by exchange rate smoothing. Using nine quarterly observables over 2011Q1-2025Q1, including key Vietnamese and U.S. macroeconomic variables, I estimate a Bayesian small open economy New Keynesian (SOE-NK) DSGE-VAR model to account for the Vietnamese economy's unique characteristics: its high trade openness, shallow FX markets, and incomplete exchange rate pass-through. I synthesize extant literature on BDSGE-VAR modeling and Vietnamese monetary policy with the aim of contributing a Vietnam-specific application of this innovative methodology to the field. My results support three conclusions. Chapter 4's posterior output confirms that Vietnam is indeed an IT regime, supported by evidence that the inflation coefficient satisfies the Taylor rule. Chapter 5 estimates the marginal likelihood of two policy rules, illustrating that the data strongly prefer the baseline specification, which allows for an exchange rate term, over the strict-IT (or pure-float) counterfactual. Chapter 6 describes the recursive out-of-sample forecast experiments using data from 2015Q1–2019Q4 and demonstrates that the effects of exchange rate smoothing are horizon-dependent. Specifically, although the baseline model tends to project higher output at the one-step-ahead horizon and lower output at medium horizons compared to the strict-IT counterfactual, it delivers stronger medium-horizon forecast accuracy for the Vietnamese output proxy variable. Taken together, my findings suggest that Vietnam’s post-GFC monetary framework is best understood not as strict IT, but as an "IT-style" regime in which exchange rate smoothing remains systematically crucial to the SBV's policy considerations.