Forecasting macroeconomic data for an emerging market with a nonlinear DSGE model

Abstract

This paper considers the forecasting performance of a nonlinear dynamic stochastic general equilibrium (DSGE) model. The results are compared to a wide selection of competing models, which include a linear DSGE model and a variety of vector autoregressive (VAR) models. The parameters in the VAR models are estimated with classical and Bayesian techniques; where some of the Bayesian models are augmented with stochastic-variable-selection, time-varying parameters, endogenous structural breaks and various forms of prior-shrinkage (where the Minnesota prior is included as a special case). The structure of the DSGE models follow that of New-Keynesian varieties, which allow for nominal and real rigidities. The nonlinear DSGE model makes use of the second-order solution method of Schmitt-Grohé and Uribe (2004) and a particle filter is used to generate values for the unobserved variables. Most of the parameters in these models are estimated using maximum likelihood techniques. The models are applied to the macroeconomic data of South Africa, which is classified as an emerging market economy. The initial in-sample period of 1960Q1 to 1999Q4 is used to generate an eight-step ahead forecast. The models are then estimated recursively, by extending the in-sample period by a quarter, to generate successive forecasts over the out-of-sample period, 2000Q1 to 2011Q4. We find that the forecasting performance of the nonlinear DSGE model is almost always superior to that of the linear counterpart; particularly over longer forecasting horizons. The nonlinear DSGE model also outperforms the selection of VAR models in most cases.JEL Classifications: E0; C5; C11; C61; C63

Forecasting using a Nonlinear DSGE Model

“…In a few studies, small-scale nonlinear DSGE models are estimated (Pichler, 2008; Gust et al, 2012; Fernandez-Villaverde et al, 2010). Most of them use only three observed variables: output, the nominal interest rate and inflation (Amisano and Tristani, 2010; Pichler, 2008;Balcilar et al, 2014;Gust et al, 2012). A few studies use other observed variables: Doh (2011) uses additional data about the yield curve, while, Hall (2012) uses consumption instead of output.…”

DSGE model-based forecasting of modelled and nonmodelled inflation variables in South Africa

“…These studies develop more sophisticated frameworks that allow for various real and nominal rigidities in closed and small open economy NKDSGE models for South Africa and find that DSGEbased inflation forecasts tend to outperform those obtained from BVAR and VAR models. In addition, Balcilar et al (2013) show that by including non-linearities in the DSGE framework to account for structural changes in South Africa's economy, Non-linear DSGE-based forecasts outperform those from the linear counterpart, as well as VAR models. In the same breadth, other studies (e.g.…”

“…Overall, the preferred choice is the specification with λ = 1. On the other hand, RMSE statistics associated with the auxiliary 7 In addition to comparing the DSGE model's performance in forecasting core variables to that of the benchmark autoregressive AR(1) model, we also considered the forecast performance of the Vector Autoregressive (VAR), Bayesian VAR (BVAR) models as well as forecast combination based on the simple mean of the forecasts from the AR, VAR and BVAR, all estimated with one lag. On the whole (for h = 2, h = 4 and h = 12), we find that DSGE-based forecasts for the inflation variable are still associated with smallest forecast errors compared to the other benchmarks.…”

Do we need a global VAR model to forecast inflation and output in South Africa?

“…Gupta and Kabundi (2010;2011) investigate large-scale Factor Models (FMs) for forecast performance, while Ngoie and Zellner (2012) illustrate the forecasting power of a disaggregated Marshallian macroeconomic model. DSGE models used for forecasting include various closed-economy Dynamic Stochastic General Equilibrium (DSGE) models (Liu & Gupta, 2007;Liu, Gupta & Schaling, 2009;, small open economy New Keynesian DSGE models (Steinbach, Mathuloe & Smit, 2009;Alpanda, Kotzé & Woglom, 2011), small open economy New Keynesian DSGE-VAR model (Gupta & Steinbach, 2013) and a recent closedeconomy nonlinear DSGE model (Balcilar, Gupta & Kotzé, 2013).…”