Inefficiency in Macroeconomic News Forecasts: Effects on Asset Prices and Asset Allocation Rules

Abstract

This paper tests the efficiency of macroeconomic forecasts, contributing to the existing literature using a rolling-event approach. We construct a monthly economic surprises index, aggregating several macroeconomic news surprises for the nine largest economic areas (G9), which we further analyze the impact on stock, bonds and foreign exchange markets using monthly data. Consequently we extend both research branches mostly focused on efficiency analysis and event studies in macroeconomic news impact. Consistently with the slow adjustment of analysts to news, our results suggest the existence of persistent unexpected economic surprises, presenting a strong autocorrelation for the aggregated G9 economic areas and, individually for USA, Eurozone and Japan. Business cycle decomposition shows that this is more intense in recession phases. Moreover, we provide evidence of a significant relation between economic news surprises and the returns of bond and stock markets. At last, a comparative study of investment decisions and asset allocation rules is also provided, concluding that past economic surprises can be used to predict future returns, providing stronger hit-ratios and higher returns than buy-and-hold and auto-regressive based strategies.