Effects of macroeconomic forecast disagreement on asset prices and risk premia

• Applicant: Professor Dr. Matthias Neuenkirch
• Subject Area: Economic Theory
• Term: since 2021
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 493001050

Expectations play an important role in macroeconomics and asset pricing. However, expectations regarding macroeconomic parameters – even those of professional forecasters – are characterized by a high degree of heterogeneity. Previous studies yield mixed results concerning the impact of forecast disagreement on asset prices; depending on the study, these can be positive or negative. Thus far, the results are inconclusive on whether the market reaction is mispricing due to market frictions (Miller hypothesis) or can be interpreted as a risk premium resulting from an efficient mechanism. Our project analyses this question with a new data set and a new econometric methodology. In addition to the Survey of Professional Forecasters (SPF) of the Federal Reserve Bank of Philadelphia, forecasts by Consensus Economics are utilized in the context of asset pricing applications. This new data set enables us to obtain measures of forecast disagreement on a monthly basis for several economies and hence expands the data previously used in the literature (only quarterly data for the United States in the SPF). In particular, this data set allows for a more precise determination of the dynamics of risk premia at a higher frequency and an examination of cross-country differences.We develop two additional perspectives to the discussion of whether there is mispricing or an efficient risk premium. First, we empirically investigate disagreement factors on an international level (instead of for the US only). If the Miller hypothesis holds, we expect that the influence of forecast disagreement on asset prices is particularly pronounced in markets where short selling is uncommon or even completely prohibited. For that purpose, we exploit cross-country differences in short-selling regulations and practices. Second, we explore the possibility that disagreement is not only a priced risk, but also influences the level of other risk premia via a time-varying transmission channel. Such a conjecture is supported by modern asset pricing models with recursive preferences and heterogeneous agents. Therefore, we estimate a dynamic multi-factor model where a shock in the disagreement of macroeconomic forecasts is considered (i) as an independent risk factor and (ii) as a determinant of the risk premium of other asset pricing factors. This allows us to trace back previous negative results of the risk hypothesis to an empirical misspecification of the factor structure and could explain the conflicting results of previous studies.

• DFG Programme: Research Grants
• Co-Investigators: Professor Dr. Christian Bauer; Dr. Dennis Umlandt