Skip to main content
eScholarship
Open Access Publications from the University of California

UC Irvine

UC Irvine Electronic Theses and Dissertations bannerUC Irvine

Essays in Macroeconometrics

  • Author(s): LUSOMPA, AMAZE
  • Advisor(s): Swanson, Eric
  • et al.
Abstract

My dissertation consists of 3 chapters that cover two major themes. The first theme is to provide

more accurate, robust, and efficient statistical methods to estimate the impact shocks have on the

economy. The second theme is estimating the time-varying and nonlinear impact shocks have on

the economy.

What is the impact of a stimulus package? What is the impact of the Federal Reserve increasing

the fed funds rate? What are the sources of business cycle fluctuations? These are some of the

most important questions in macroeconomics, and to answer them it is often necessary to

estimate impulse responses. Over the past decade, Local Projections have been increasingly used

to estimate impulse responses. Local Projections are linear regressions that estimate impulse

responses directly. The first chapter of my dissertation, “Local Projections, Autocorrelation, and

Efficiency”, introduces a more accurate and robust method to estimate impulse responses using

Local Projection. These methods can be employed using Bayesian or frequentist methods. I also

extend Local Projections to be able to estimate time-varying impulse responses.

The second chapter of my dissertation, “Identifying Government Spending Shocks, Fiscal

Foresight, and Time-Varying Parameters”, focuses on estimating the government spending

multiplier, and testing the multiplier for time variation and nonlinearity. Ramey (2011, 2019)

document the wide range of estimates for multipliers in both empirical studies and dynamic

simulations. This chapter makes several contributions to the estimation of fiscal multipliers.

First, this chapter points out that the current identification schemes for estimating the multiplier

are not identifying exogenous changes in government spending, which has caused previous

estimates of the multiplier to be biased. I also show that the way the impulse responses are

estimated cause additional bias in the multiplier estimates.

The third chapter of my dissertation, “A Kalman Filter Test for Structural Instability”, focuses on

testing parameter instability in regression models. Researchers are often interested in whether

there is parameter instability in regression models. There are several studies that show a

significant amount of macroeconomic and financial time series exhibit parameter instability

These changes can occur for many reasons such as policy changes, technological evolution,

changing economic conditions, etc. There are a multitude of methods available to test for

parameter instability, but it has been demonstrated that the most popular tests for parameter

instability do not perform adequately. That is, even when there is parameter instability, these tests

have trouble detecting the instability. In this chapter, I propose a more powerful test to detect

parameter instability.

Main Content
Current View