HPI Forecast Validation Report
HPI Forecast Validation Report - February 2019
The CoreLogic HPI Forecast Validation Report analyzes the results of the HPI Forecast versus the actual HPI results over a 12-month period, comparing our forecasted HPI for November 2018 (made in November 2017), with the actual HPI Index results for November 2018.
This bi-annual report includes:
- Comparison of the National Forecasted HPI Versus the Actual HPI
- Forecasted Versus Actual Results for the Most Populous CBSAs
- Analysis of Overvalued and Undervalued Markets
- And much more
The CoreLogic HPI Forecast is a projection of home prices using the CoreLogic HPI and other economic variables. Values are derived from state-level forecasts by weighing indices according to the number of owner-occupied households for each state.
HPI Validation Forecast Methodology
Each month, CoreLogic publishes the CoreLogic Home Price Index. The HPI contains the current and historical index values going back to January 1976. There is a 5-week lag between the HPI release date and the most current index value. For example, the HPI release in early October 2018 contains index values through June 2018.
As new sales transactions are recorded and made available to the HPI, index values for previous months are restated to reflect the new sales data. Given the varying rates at which counties record sales transactions, all historical HPI values are subject to revision. Large restatements will, in general, be limited to the last 3 months of index values.
HPI Version 4.0 was first calculated in April 2016 and first published in June 2016. Each month, CoreLogic stores the published HPI for future back testing purposes.
HPI Forecast Data
Each month, CoreLogic publishes the CoreLogic HPI Price Forecasts (HPIF). The HPIF contains the predicted monthly HPI values for the ensuing 30 years. For example, the June 2016 HPIF contains a predicted HPI value for June 2016, July 2016, .... to April 2046.
HPIF Version 4.4 was first calculated in April 2016 and first published in June 2016. Each month, CoreLogic stores the HPIF for future back testing purposes.
Each month, CoreLogic conducts an ex-post back-test to determine how accurately the HPIF is predicting the HPI increase 12 months into the future. The analysis is done for Tier 11 (Single Family Combined) at the national and state level as well as for the largest 50 CBSAs.
The back-test analysis compares the actual HPI increase in the preceding 12 months to the HPIF predicted increase from 12 months earlier. For example, in August 2017, we compared the actual HPI increase between June 2016 and June 2017 with the increase forecast by the HPIF in June 2016. Because of the HPI revisions, we back test using HPI values after they have undergone 1 month of revision (i.e., we use index values from the prior release month.)