This research primarily focuses on estimating macroeconomic models where households, firms, and policymakers have limited information and must learn about the state of the world from private and public signals. Evidence from forecasts suggests that economic agents have imperfect and often different information and beliefs about future macroeconomic and financial conditions, and a growing body of economic research has explored how information imperfections may affect the transmission of macroeconomic shocks. Furthermore, since at least Keynes, economists have debated the extent to which differences in beliefs or information determine asset prices, and the role of asset prices in aggregating and transmitting information. Understanding what economic agents learn from prices, especially financial prices, is important for understanding the interaction between Wall Street, monetary policy, and macroeconomic outcomes.
This researcher combines macroeconomic and financial data with forecasts to quantify the importance of imperfect information for business cycles and monetary policy, and to assess the information content of prices. One ongoing project focuses on how traders who have noisy, private signals about macroeconomic variables learn from the prices of different bonds about the state of the world. Another ongoing project demonstrates in calibrated examples how the information content of central bank policies is affected by the quality of central bank information. Future work will use private-sector interest rate forecasts to estimate the perceived and actual “shadow rate” – the effective rate of interest set by the Federal Reserve as a result of quantitative easing – in order to assess whether the extraordinary policy measures after the financial crisis were hampered by policymakers’ limited ability to communicate their intent.
A challenge in assessing economic models with realistic information frictions is computational. Finding solutions to these models involves solving relatively large systems of nonlinear equations, which is computationally burdensome. Estimation requires solving the systems hundreds of thousands of times to find parameters that best fit the data, and the use of panels of individual forecasts means that the data sets are relatively large (at least by the standards of macroeconomics), which increases the difficulty of estimation. Hence, MSI's resources greatly enhance the ability to carry out this research agenda.