2020/09/23 by Brook T. Russell, Russell, Brook T., Whitney K. Huang +1
Economics, Econometrics and Finance · Agricultural and Biological Sciences · #Monetary Policy and Economic Impact #Agricultural Economics and Policy #Market Dynamics and Volatility
paper · pdf · doi:10.48550/arxiv.2009.11098
The block maxima approach is an important method in univariate extreme value\nanalysis. While assuming that block maxima are independent results in\nstraightforward analysis, the resulting inferences maybe invalid when a series\nof block maxima exhibits dependence. We propose a model, based on a first-order\nMarkov assumption, that incorporates dependence between successive block maxima\nthrough the use of a bivariate logistic dependence structure while maintaining\ngeneralized extreme value (GEV) marginal distributions. Modeling dependence in\nthis manner allows us to better estimate extreme quantiles when block maxima\nexhibit short-ranged dependence. We demonstrate via a simulation study that our\nfirst-order Markov GEV model performs well when successive block maxima are\ndependent, while still being reasonably robust when maxima are independent. We\napply our method to two polar annual minimum air temperature data sets that\nexhibit short-ranged dependence structures, and find that the proposed model\nyields modified estimates of high quantiles.\n