Pattern-Based Expectations in Macroeconomics

• Applicant: Professor Dr. Tobias Frank Rötheli
• Subject Area: Economic Theory
• Term: from 2012 to 2017
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 227064168

This project is concerned with the elicitation and modeling of expectations concerning macroeconomic variables like inflation and income. These expectations play a key role in many decisions that affect macroeconomic outcomes. Among the considerations where expectations play a role we focus on the pricing decisions on goods markets and financial markets and the consumption/saving decision. Through these channels expectations influence the course of the business cycle and inflation. Models of expectations formation are thus a critical ingredient in any macroeconomic analysis and influence economic policy making. Research on expectations over the past decades has been strongly influenced by the hypothesis of rational expectations. A substantial body of empirical research (much of it experimental in nature or based on survey methods) has questioned the validity of this hypothesis. The present project follows in the footsteps of this research and proposes a model of boundedly rational expectations. The project uses concepts from cognitive psychology and elicits expectations data under laboratory conditions. The theoretical foundation of this approach is the concept of pattern recognition. Building on insights from the field of financial economics we show subjects an array of patterns that is just sufficient to use the experimental data in the quantification of models of expectations. In the process of eliciting data we thus find expectations concerning inflation and income for one and five years into the future. In the various experimental treatments of the study we investigate differences between expectations regarding inflation and income, connections between short- and long-term expectations and the separate role of expert predictions and central banks¿ inflation targets on individuals¿ expectations. The experimental data is then used to quantify time-series models of expectations which are in turn applied to the study the dynamics of inflation expectations, the heterogeneity of expectations as well as expectations uncertainty. These models are then further applied in investigations of inflation dynamics, the course of interest rates, and consumption.

• DFG Programme: Research Grants