This paper attempts to answer the question of whether the principle of parsimony can be applied to interval forecasting for multivariate series. Toward that end, this paper proposes the block bootstrap prediction intervals based on parsimonious first‐order vector autoregression. The new intervals generalize standard bootstrap prediction intervals by allowing for serially correlated prediction errors. The unexplained serial correlation is accounted for by the generalized multivariate block bootstrap, which resamples two‐dimensional arrays of residuals. Different methods of block bootstraps are compared. A Monte Carlo experiment shows that, in most cases, the new intervals from a parsimonious model outperform the standard bootstrap intervals from a complex model. The proposed block bootstrap prediction intervals are illustrated using financial data for interest rates and exchange rates.

中文翻译：

---

简约一阶向量自回归的块自举预测间隔

本文试图回答是否可以将简约原则应用于多元序列的区间预测的问题。为此，本文提出了基于简约一阶向量自回归的块自举预测间隔。新间隔通过允许与序列相关的预测误差来概括标准的引导程序预测间隔。无法解释的序列相关性是由广义多元块自举引起的，该自举对二维残差数组重新采样。比较了块引导程序的不同方法。蒙特卡洛实验表明，在大多数情况下，简约模型的新间隔优于复杂模型的标准自举间隔。