ABSTRACT
The literature on the subject of forecast evaluation is vast, and a profusion of tests exists to assess the forecast performance of statistical and other methodologies. This chapter notes that forecast evaluation extends beyond quantitative forecasting to cover judgmental forecasting. In this chapter we present conventional forecast accuracy tests. These tests are commonly used in real estate to assess the forecasting ability of models and should be seen as the starting but essential step in forecast evaluation. The chapter illustrates the key aspects of the forecast assessment exercise with detailed examples drawn on the office rent models the authors estimated with Hong Kong data and a model of all property prices (capital values) for the US. A comprehensive forecast evaluation requires a large dataset of forecast errors for in-depth testing. This may not be feasible in several markets where the authors have limited data. The chapter shows more practices to generate a sample of forecasts and forecast errors for evaluation.
