What can we learn from the history of gasoline crack spreads?: Long memory, structural breaks and modeling implications

摘要: Abstract In this paper, we investigate the long-range auto-correlations of crack spreads using a nonparametric method, named detrended moving average (MF-DMA). We find that display multiscaling behaviors and are dominated by anti-persistence (mean-reversion) in long-term. Moreover, multifractal, indicating various small large fluctuations different scaling behaviors. Using technique rolling windows, some extreme events can drive degree multifractality (complexity) to rise up. other words, these have negative impacts on market efficiency. However, effects not alike. also detect spread volatilities strong persistent behavior multifractality. Finally, discuss modeling implications findings auto-correlated patterns. Our results indicate ARFIMA-GARCH models capture major dynamics fluctuations. For fluctuations, they misspecified. Interestingly, do imply ARFIMA model which takes long memory into account outperform random walk sense out-of-sample prediction. The reason may be complexity exploited paper causes low predictability model.