Robust bootstrap inference for linear time-varying coefficient models : some Monte Carlo evidence

We propose two robust bootstrap-based simultaneous inference methods for time series models featuring time-varying coefficients and conduct an extensive simulation study to assess their performance. Our exploration covers a wide range of scenarios, encompassing serially correlated, heteroscedastic, endogenous, nonlinear, and nonstationary error processes. Additionally, we consider situations where the regressors exhibit unit roots, thus delving into a nonlinear cointegration framework. We find that the proposed moving block bootstrap and sieve wild bootstrap methods show superior, robust small sample performance, in terms of empirical coverage and length, compared to the sieve bootstrap introduced by Friedrich and Lin (2022) for stationary models. We then revisit two empirical studies: herding effects in the Chinese new energy market and consumption behaviors in the U.S. Our findings strongly support the presence of herding behaviors before 2016, aligning with earlier studies. However, we diverge from previous research by finding no substantial herding evidence between around 2018 and 2021. In the second example, we find a time-varying cointegrating relationship between consumption and income in the U.S.