Why Do Investors Forecast Macroeconomic Data: Effect of Perfect Forecasting on Portfolio Performance (2014)
Undergraduate: Alex Moehring
Faculty Advisor: Mike Aguilar
Department: Economics
This paper examines the portfolio response to scheduled macroeconomic news events using both daily and high frequency data. This is accomplished by comparing Sharpe ratios of portfolios formed using naive forecasting methods for expected return and volatility with those formed using ex ante knowledge of the release value as an additional term in the conditional mean and conditional variance equations. For tractability, the unconditional covariance is used to forecast the covariances required in Modern Portfolio Theory. The workhorse Autoregressive (AR(1)) and Generalized Autoregressive Conditional Heteroscedasticity (GARCH(1,1)) models are used to forecast expected return and variance respectively. The hypothetical portfolio is created at the end of the period before the release, and sold at the end of the period of the release. This paper finds that from the time period 2003-2012, there is little evidence to support the claim that knowledge of the release value improves portfolio performance. It is concluded that, although macroeconomic news events do affect the assets studied, there is no way to systematically include knowledge of the release values into your portfolio.