Date of Award

December 2019

Degree Type


Degree Name

Doctor of Philosophy



First Advisor

Antu Murshid

Committee Members

Kundan Kishor, Filip Vesely, Jangsu Yoon


Federal Reserve's Mortgage-Backed Securities Purchase Program, Financial Markets, Kalman Filter, Mortgage Spreads, System GMM Estimation, Unconventional Monetary Policy


This dissertation consists of three essays concerning financial economics.

In the first essay, I focus on the role of the life insurance in the financial markets and the factors that drive the development of life insurance industry. The essay examines the causality of the banking sector development on the development of the life insurance market measured by the total premiums received by life insurers using panel estimation for about 90 countries over the period 1996-2010. I employ the dynamic System Generalized Method of Moments estimation technique to resolve the endogeneity problem between the development of the life insurance and the financial development. After controlling for country-specific effects, problems with lagged dependent variable, endogeneity and weak instruments, I find a strong significant link between the development of banking sector and life insurance industry. While the development of banking sector is important for implementing the investment function of life insurers, with the increasing savings type insurance, life insurers face vigorous competition from banks. The results show that the nature of relationship between banks and life insurers can change from complementary to substitute.

In the second and third essays, I study the role of the government in the financial markets, specifically, mortgage-backed securities (MBS) market, during the financial crisis and the impact of its monetary activities on the MBS prices, mortgage markets and the economy overall. The second essay utilizes the vector autoregression with exogenous regressors (VARX) modeling technique to estimate the impact of MBS purchase program's announcements and the change in size of the Federal Reserve's MBS holdings on mortgage spreads. The estimation results of the second essay indicate that Federal Reserve's MBS purchase program has effect on the mortgage spreads through both signaling and portfolio rebalancing channels. The study finds that not only the announcements of the first and the third rounds of Quantitative easing, QE1 and QE3, but also the increase in the size of the Federal Reserve's MBS holdings has negative impact on MBS spreads. In addition to being used to reveal the forecasting ability of the changes in the Federal Reserve's MBS holdings variable, a triangular structural VAR model is used to generate the forecast of MBS spreads, conditional on tapering of Federal Reserve's reinvestments in its MBS portfolio. The third essay examines the quantitative impact of the Federal Reserve's MBS purchase program on the housing market and aggregate economic variables via mortgage spreads. I find that mortgage spread shocks reduce consumption, residential investments and GDP by both economical and statistically significant magnitudes. The estimation results of vector autoregressive model using quarterly data from 1986Q1 to 2017Q4 show that the reduction in mortgage spreads from the Federal Reserve's MBS purchase program has only partially passed through to borrowers due to widening of the primary-secondary mortgage market spread - the spread between mortgage rates and MBS yields - explaining a slow recovery of consumption in post-crisis period. The decline in mortgage rates did not keep pace with this on the secondary rates as the transmission of lower mortgage spreads in the secondary mortgage market to lower borrowing costs in the primary mortgage market has not happened due to a rise in originators' costs and profitability, as well as increases in g-fees.