A0477
Title: Forecast combination puzzle in the HAR model
Authors: Andrey Vasnev - University of Sydney (Australia) [presenting]
Adam Clements - Queensland University of Technology (Australia)
Abstract: Given its simplicity and consistent empirical performance, the Heterogeneous Autoregressive (HAR) model of Corsi has become the benchmark model for predicting realized volatility. Many modifications and extensions to the original model have been proposed that often only provide incremental forecast improvements. A step back is taken, and the HAR model is viewed as a forecast combination that combines three predictors: previous day realization (or random walk forecast), previous week average, and previous month average. When apply- ing the Ordinary Least Squares (OLS) to combine the predictors, the HAR model uses optimal weights that are known to be problematic in the forecast combination literature. An average of simpler HAR-style models and a simple average forecast often outperforms the optimal combination in many empirical applications. The performance of these simple combination forecasts is investigated for the realized volatility of the Dow Jones Industrial Average equity index and a sample of individual constituent stocks, as well as across a range of other as- sets, commodities, exchange rates and a range of global equity market indices. In all cases, dramatic improvements in forecast accuracy across all horizons and different time periods are found. This is the first time the forecast combination puzzle has been identified in this context.
