A preview of the PDF is not available
Three Essays in Applied Spatial and Time Series Econometrics
The objective of Chapter 1 is to explain the dynamic relationship between the housing market and the markets for primary building materials in the U.S. by allowing a single structural break that may have risen in response to macroeconomic shocks, specifically the housing market crisis started in 2007. In particular, using a vector error correction model (VECM), Granger–causality test, and impulse response analysis, the dynamics of housing prices and prices of concrete, lumber, plywood, and oriented strand board are investigated in two macroeconomic conditions of the economy: during the presence of structural break as well as before and after it. The data used in the analyses cover monthly time–series observations over the period of 1995-2015. Using the Qu and Perron (2007) methodology in a VECM with certain parameter restrictions shows that there is a strong single structural break in September 2007, which is a close estimation for the start of the Great Recession in December 2007. Thus, the data are split into two segments. The time–series analyses across segments suggest that most of the bidirectional Granger–causalities and dynamic linkages between housing prices and the prices of building materials have weakened or even disappeared after the housing market crisis. Chapter 2 seeks to reveal how the cultural background of societies such as religion affects natural environment while accounting for the possible spatial dependencies between observation characteristics. More specifically, using spatial econometric models and an extensive county–level U.S. data in 2010., this study investigates the impact of religion on the environmental performance of a county after controlling for the other important determinants. The environmental performance is measured using seven criteria air pollutants: CO, NH3, NOx, SO2, PM10, PM 2.5, and VOC. For each air pollutant, the analyses are conducted separately using various spatial models to account for the global and local spillover effects as well as the global diffusion effect inherited in the data. The extensive spatial analyses suggest that there is a strong and positive spatial autocorrelation between observations and religion appears to be associated with environmental performance. Specifically, religions of Judeo–Christian beliefs such as Evangelical Protestants, Black Protestants, Mainline Protestants, Catholics, Orthodox Christians, and Jews exhibit a negative association with environmental performance, whereas Hindus and Buddhists are the only ones that show a positive association. Therefore, I find strong support for White’s Thesis by providing spatial econometric evidence. Chapter 3 examines how the dynamic relationships between rival cryptocurrencies change over time and are affected by shocks. In particular, using a vector autoregressive model (VAR), Granger–causality test, and impulse response analysis, the price dynamics of Bitcoin, Litecoin, and Ripple are investigated by allowing multiple structural breaks. The data used in the analyses cover daily time–series observations over the period of April 2014-July 2015. Using the Qu and Perron (2007) methodology in a VAR shows that there are two strong structural breaks in on November 12, 2015 and September 28, 2016. Thus, the data are split into three segments. The time–series analyses across segments suggest the following results: (1) the Granger–causality from the prices of other coins to Ripple price is gaining strength; (2) the response of each coin to a shock in Bitcoin price is same across segments; (3) in response to a shock in Litecoin price, the impact on Bitcoin price is decreasing over time but the effect on Ripple price is increasing; and (4) the impacts on Bitcoin and Litecoin prices are falling over time in response to a shock in Ripple price.