Joint convergence of sample autocovariance matrices when \(p/n\to 0\) with application (Q2284381)

This paper is about Limiting Spectral Distribution (LSD). The main result of this article is that, under some assumptions, the trace of any polynomial in \(\left\lbrace \hat{\Gamma}_{u}, \hat{\Gamma}_{u}^{*}, u \geq 0\right\rbrace\), where \(\hat{\Gamma}_{u}\) is the \(u\)-th order sample autocovariance matrix, is asymptotically normal. In addition, the results are illustrated through several examples (13 in total) in the context of time series, focusing on the inference (hypotheses tests and estimation of the unknown order) of high-dimensional MA processes.