Essays on macroeconomic dynamics and the econometrics of expectiles

This is a collection of studies considering models that behave differently in different scenarios. In four essays, we apply this approach to (1) macroeconomic dynamics and (2) expectile regression, which is a latent topic in the literature. In the first essay, we investigate government spending multipliers using a two-regime model and impulse response functions with fully endogenous regimes. While short-run multipliers vary depending on business cycle fluctuations, we find little evidence that medium or long-run multipliers vary between expansions and recessions. The reason for state dependence found in the literature is the constant-regime assumption used to create impulse response functions. Importantly, a fiscal policy shock has little effect on the duration of a recession. In the second essay, we show that the multiplier does not depend on the monetary policy rule. What we find is that the monetary policy rule itself changes after a government spending shock and converges quickly to a similar regime regardless of the initial condition. This rapid change in monetary policy leaves the multiplier unaffected by the initial monetary policy regime. An exception to this characterization of monetary policy occurs when nominal interest rates are stuck at zero. We analyze the multiplier at the zero-lower bound and find that the multiplier exceeds one. The third essay re-introduces expectile regression. In some cases where OLS assumptions are violated, an expectile regression estimator is also the BLUE for the mean regression: we give three examples. Expectile regression is the BLUE for quantile regression coefficients in special cases where they are equal. But expectiles can be used in some models where quantiles are not helpful, such as binary response models. In those cases, expectile regression is the new best option. The fourth essay dispels misinformation from the literature. Two different likelihood models have been suggested for estimating expectiles as a maximum likelihood estimator. After comparison, it becomes clear that they are not the same and only one of these models is appropriate for that purpose.

Name thesis:degree_name

Ph.D.

Level thesis:degree_level

Dissertation

Discipline thesis:degree_discipline

Economics

Grantor

University of Illinois at Urbana-Champaign

Year dc:date

2022

Author dc:creator

• Philipps, Collin S

Contributors dc:contributor

• Shin, Minchul
• Bernhardt, Dan
• Amir-Ahmadi, Pooyan
• Deltas, George

• Fiscal Policy

Statement dc:rights

• Copyright 2021 Collin Philipps

Language dc:language

en

Handle dc:identifier

http://hdl.handle.net/2142/112992

OAI identifier oai:identifier

oai:www.ideals.illinois.edu:2142/112992