Least squares estimators for unit root processes with locally stationary disturbance (Q764805)

Summary: A random walk is used as a model expressing equitableness and effectiveness of various financial phenomena. The random walk is included in unit root processes which form a class of nonstationary processes. Due to its nonstationarity, the least squares estimator (LSE) of the random walk does not satisfy asymptotic normality. However, it is well-known that the sequence of partial sum processes of a random walk weakly converges to a standard Brownian motion. This result is a functional central limit theorem (FCLT). We can derive the limiting distribution of the LSE of the unit root process from the FCLT result. The FCLT result has been extended to unit root processes with locally stationary process (LSP) innovations. This model includes two different types of nonstationarity. Since the LSP innovation has a time-varying spectral structure, it is suitable for describing empirical financial time series data. We derive the limiting distributions of the LSE of the unit root, near unit root and general integrated processes with LSP innovations. The testing problem between unit roots and near unit roots is also discussed. Furthermore, we suggest two kinds of extensions for the LSE, which include various well-known estimators as special cases.