Essays in Macroeconomics

Abstract

My dissertation consists of three independent chapters analyzing parameter estimation and structural change in applied macroeconomics. A first theme linking these papers is structural change, especially as it relates to the monetary policy transmission mechanism through the Phillips curve. A second theme is an assessment of small-sample statistical inference for impulse response functions after estimating macroeconomic models. Two of my chapters provide simulation studies of statistical coverage of standard test statistics after estimating impulse response functions in both atheoretical (local projection) and highly structural (dynamic stochastic general equilibrium) models. The first chapter of my dissertation, ``Using Survey Expectations to Estimate the New Keynesian Phillips Curve,'' provides new estimates of the parameters in the New Keynesian Phillips Curve, exploiting survey based expectations data provided by the Survey of Professional Forecasters and the Michigan Survey of Consumers. I find that the use of survey expectations in US data improves the fit of the textbook Phillips Curve model to the data and provides economically sensible estimates of its coefficients. The estimated model provides stable parameter estimates until the Great Recession, after which inflation becomes less dependent on marginal cost. Household and professional forecasts each contribute to the forward-looking component of inflation expectations, with household forecasts given more weight. The second chapter of my dissertation, ``Estimating Structural Breaks in Impulse Response Functions via the Local Projection Estimator,'' proposes an estimator for parameter instability in impulse response functions that are estimated by local projections. I use the estimator to investigate the presence of parameter instability in the Romer--Romer monetary policy shocks. I find evidence of a structural break in the impulse response coefficients in the late 1970s. In the early period, there is strong evidence that monetary policy shocks have real effects. There is little evidence that monetary policy shocks have real effects in the later period. Tax and oil price shocks exhibit little change in their effects on output throughout the postwar period. The third chapter of my dissertation, ``Standard Errors for Impulse Response Functions of Estimated DSGE Models,'' provides a method for constructing appropriate asymptotic standard errors for impulse responses of estimated dynamic stochastic general equilibrium models. The method requires only the matrices characterizing the model solution, the derivatives of those matrices with respect to the underlying structural parameters, and the covariance estimate of the structural parameters themselves. I provide simulation evidence on the small-sample properties of these standard errors.