Time-Varying Spillover of US Trade War on the Growth of Emerging Economies

In the wake of an unprecedented increase in the trade policy-related uncertainty of the US since 2017, we analyze the ability of a newspaper-based trade policy uncertainty index (TPU) of the US in predicting the growth rate of emerging market economies. The newspaper-based trade policy uncertainty index was developed by Caldara et al. (2020), while GDP data was obtained from the Global Economic Database maintained by the Federal Reserve Bank of Dallas, in the form of an aggregate value for all emerging economies. We analyze the data in a bi-variate VAR(2) setup, with the number of lags identified using Akaike Information Criterion, using the novel multivariate time-varying causality framework developed by Rossi and Wang (2019) on an effective sample of 1984:Q3 to 2019:Q3. Starting with the standard constant parameter Granger causality test, we find no evidence of TPU to the growth of GDP. We then proceed with the time-varying parameter causality test of Wang and Rossi (2019), where we find overwhelming evidence of the role of trade uncertainty in impacting the growth of emerging markets in a statistically significant manner, with the effect being on the rise since the Great Recession. Our results are robust to the usage of an alternative econometric methodology, metric of trade uncertainty, and also over an out-of-sample forecasting exercise. While the time-varying predictive analysis is the focus of our paper, causality tests are silent about the sign of the impact of TPU on the GDP growth of the emerging markets. Given this, we estimate a time-varying parameter VAR model with stochastic volatility (TVP-VAR-SV). We find that the impact is sharp and negative. Our results imply that policymakers in emerging countries must be aware of the possible threat of an upcoming recession in the wake of heightened trade policy uncertainty in the US, and hence, must be ready to undertake necessary (expansionary) policies to prevent a downturn in their respective domestic economies. To draw optimal policy responses, authorities would need to utilize time-varying econometric models, since the effect of uncertainty related to trade policies of the US are indeed non-constant over time.