Dissertations, Theses, and Capstone Projects

Date of Degree


Document Type


Degree Name





Thom B. Thurston

Committee Members

Sangeeta Pratap

Christos I. Giannikos

Subject Categories

Econometrics | Finance | Macroeconomics


Macroeconomics, Monetary Economics, Financial Economics, Econometrics, Quantitative Economics


Chapter 1: (A Quantitative Analysis of Interest on Reserves and Reserve Requirements) - I construct a medium scale DSGE model with financial frictions both on the demand (entrepreneurs) and supply (banks) sides of credit to study the costs and benefits of fixed/time-varying minimum reserve requirements and interest paid by the Fed on reserves.The results can be summarized as follows: (1) An optimal time-varying minimum reserve requirement generates substantial welfare gain when compared with a fixed minimum reserve requirement when no interest is paid on reserves. (2) Paying interest on reserves is substantially welfare inferior to a policy with no interest on reserves no matter whether the Fed has a fixed or time-varying minimum reserve requirement. However, paying no interest on reserves is associated with a highly elevated volatilities of the macroeconomic and financial variables, and substantially raises the entrepreneur’s default rate. (3) If the Fed pursues a policy of paying interest on reserves, it is optimal to set interest on reserves 1 percentage point below the policy rate in the steady state implying that no interest is paid on reserves.(4) The frequency of the policy rate hitting the zero lower bound (ZLB) is substantially higher under the no interest on reserves policy as compared to when the Fed pays interest on reserves. (5) From a welfare perspective, it is optimal to set interest on excess reserves equal to interest on required reserves. Setting interest on excess reserves equal to interest on required reserves, however, is associated with an increased macroeconomic and financial volatility. If the goal is to minimize macroeconomic and financial volatility, it is optimal to pay interest only on required reserves and pay no interest on excess reserves.

Chapter 2: (Downward Wage Rigidity, Credit Frictions, and Macroeconomic Dynamics) - This paper studies the interactions between the financial frictions and the downward nominal wage rigidity (DNWR) constraint, and their implications for monetary policy in a medium-scale New Keynesian model augmented with a financial sector. We find that a positive credit shock that tightens a bank’s balance sheet constraint and lowers loan supply, may cause the DNWR constraint to bind, and this amplifies the responses of both macroeconomic and financial variables as compared with the no DNWR environment. The presence of the DNWR constraint dampens the macroeconomic and financial volatility and improves the welfare as compared with the no DNWR scenario. In the context of our model featuring both the zero lower bound (ZLB) on the policy rate, and the DNWR constraint,we find that it is optimal to have a zero steady state inflation from the welfare perspective.However, having a positive steady state inflation decreases the ZLB incidences, the frequency of hitting the DNWR constraint, and the unemployment rate volatility. Finally, we discuss whether monetary policy should respond to the unemployment rate. We observe that the Taylor rule augmented with the unemployment term is welfare superior to the standard Taylor rule. However, there is a tradeoff. The augmented Taylor rule generates the higher volatility for macroeconomic and financial variables. Moreover, the frequencies of hitting ZLB and DWNR constraints, are also higher.

Chapter 3: (Deep Habits in Credit Markets - An Estimated DSGE Model) - I build a medium scale New Keynesian DSGE model with a banking sector where banks are subject to time-varying capital requirement regulation. A rich set of financial shocks are embedded into the model to study the role of deep habits in credit markets in the propagation of the effects of both macro and financial shocks on the endogenous variables. Based on the quarterly data, 1986Q1-2007Q3, using a Bayesian DSGE framework, I provide an estimate of the deep habits parameter for the U.S. economy. The estimated parameter is 0.94. Such a large estimate of the deep habits parameter implies a very high degree of attachment of the borrowers to their lending banks. I find that deep habits considerably amplify the effects of financial shocks on both macroeconomic and financial variables. Moreover, the presence of deep habits, does amplify the effects aggregate shocks (technology shock, investment shock)but not as much as in the case of financial shock. However, the effects of aggregate shocks on the financial variables are significantly amplified. Thus, the existence of deep habits in the model works as a financial accelerator. In addition, I find that transmission of monetary policy to the real economy is more pronounced in the model with deep habits in credit market. Lastly, I studied optimal macroprudential policy. I find that time varying capital requirement amplifies the effects of the shocks to the endogenous variables compared to the fixed capital requirement. In addition, optimal time varying capital requirement is welfare superior to the fixed capital requirement.

Chapter 4: (Nominal GDP Growth Rate Targeting vs. Taylor Rule: Implications for Welfare and Macro-financial Stability) - Opinion in favor of nominal GDP growth targeting rule recently has been rising in monetary policy literature (e.g., Ireland, 2020; Garin et al., 2016; Beckworth and Hendrickson, 2020). Calibrated small scale New Keynesian DSGE models have tended to support the proposition that nominal GDP growth rate targeting will provide welfare improvement over those obtained with alternative strategies such as inflation targeting and Taylor rules with standard parameter settings. This paper examines that proposition in the context of a medium-scale New Keynesian model having a financial sector with frictions. In the context of that model our results indicate that nominal GDP growth rate targeting does poorly as compared with the optimal Taylor rule and a wide range of Taylor rule settings leading up to the optimal (representative agent utility maximizing) Taylor rule. The presence of the financial sector reveals the effects of different monetary strategies on the volatility of key financial variables. We find that the choice of a nominal GDP growth targeting rule over the other strategies will increase financial variable volatility. Interestingly, the model reveals that the ”smoothing” parameter in the Taylor rule actually reduces policy rate volatility as it is raised from 0 to 1, and up to a point reduces financial variable volatility generally as well. Our paper provides evidence that, in the DSGE framework, adding the financial sector has a material effect on the rank ordering of policy frameworks according to their welfare gain. Whereas the nominal GDP growth targeting rule finishes first in the literature without financial markets, it finishes last in this model (or next to last, depending on the values of three financial friction parameters in the model).

Chapter 5: (Assessing the Taylor-Type Interest Rate Rules) - This paper assesses different Taylor-type interest rate rules in a medium-scale DSGE model with segmented financial markets, closely patterned after the model of Sims and Wu (2021). Specifically,we consider the following Taylor-type interest rate rules: (1) a standard Taylor rule, (2) a Taylor rule designed to target the price level, (3) a Taylor rule designed to target average inflation, (4) a forward looking Taylor rule and (5) a backward looking Taylor rule. We compare optimized interest rate rules with each other from the perspective of welfare and volatility. We also examine which rules are more effective in reducing ZLB frequency and ZLB duration. We find that the price level targeting rule generates the smallest welfare cost as compared with the Ramsey solution. The volatilities of the endogenous variables under the price level targeting rule are closest to the ones found under the Ramsey solution.We also find that ZLB frequency is lowest under average inflation targeting, whereas ZLB duration is lowest under price level targeting. We show that a lagged interest rate term in the Taylor rule is effective in dampening macroeconomic and financial volatility and helps to reduce the frequency of the policy rate hitting ZLB. Impulse response analysis shows that the paths of endogenous variables under the Ramsey policy are most closely approximated under the price level targeting rule.