Assessing Accuracy: Hotel Horizons® Forecasts

August 5, 2015

Executive Summary

The staff at PKF Hospitality Research, A CBRE Company proudly presents this report documenting the accuracy of our Hotel Horizons® forecasts of financial performance for U.S. lodging markets. This assessment is developed from comparisons between forecast changes in hotel market performance and actual changes during the period 2010 through 2014. We also compare our econometric forecasts to forecasts using an intuitive scenario. The analysis covers short-term accuracy defined as rolling one-year forecasts versus actual results for RevPAR, ADR, occupancy, supply, and demand. Long-term accuracy over the five-year interval also is reported. We provide evidence of annual forecasting accuracy for all U.S. hotels, hotel chain-scales, across locations, and for 50 U.S. markets. Various statistical measures, such as mean absolute error and Theil’s U statistics, are relied on to assess accuracy.
The main findings from this self-assessment are as follows:
  1. The Hotel Horizons® approach of combining econometric analysis and expert judgment is shown to be consistently more accurate than an intuitive approach based on common sense methods such as using a previous year’s performance as a guide.
  2. Forecast accuracy declines around turning points in the hotel cycle and improves as time moves further from the turning point, which is during the up-phases and down-phases of the cycle.
  3. To some extent, the data we receive from other sources impacts the accuracy of our forecasts. This includes changes in historical data series, and the forecasting errors of data providers upon whom we rely.
  4. Our forecasts largely proved to err on the side of caution as 41 of the 50 MSAs we cover had actual performance that exceeded our forecast.
  5. Forecasts for the upper-priced chain scales tended to be more accurate over time than the lowerpriced chains.