Forecasting Housing Starts and Mortgage Rates using Predictive Modeling Method

Unlike home housing starts, mortgage rates in the US fell significantly during this past summer. In this paper, statistical and computational analyses of the forecasting of the Housing Starts and Mortgage rates were performed. As evident in our observations, the lower Mortgage rates, the shorter the return periods. In other words, Mortgage rates are more likely to reoccur; however it was observed that the duration of the return periods was increased as the independent variable increases for higher values. For the analysis, we used extreme value theory (EVT) type I to quantify tail behavior to assess extreme events within a probability distribution. Analyzing the minimum and maximum values of the sample was able to determine the probability for extreme events. It was analyzed for authenticity compared to previously observed events. Multiple fitting with the last decades of data were checked to see which model fits the data adequately.