Date of Award
Doctor of Philosophy
This dissertation is a collection of three chapters on inflation dynamics and money demands. Chapter 1 tests the forward-looking New Keynesian Phillips curve using a novel panel data set for the 50 U.S. states from year 1977 to 2005. Consistent with Gali and Gertler (1999), our results support a linkage between inflation and real unit labor cost, and reject a linkage between inflation and output gap. We also address several important econometrics issues in the empricial studies. Our tests on model identification and instruments validity reveal that compared with the model with real unit labor cost, the GMM estimators in the model with output gap are more sensitive to the choice of instruments. Also, we find that the unit labor cost has stronger persistence than the output gap, and that these two variables have almost opposite dynamic cross correlations with inflation. We conclude that the observed high autocorrelation properties of U.S. inflation-as measured by the sum of AR coefficients-is well described by the forward-looking New Keynesian Phillips curve. In the second chapter, we extend the pure forward-looking New Keynesian Phillips curve to a hybrid model. We adopt a dynamic panel data model by adding a lagged inflation variable to the explanatory variables. We find relative larger weights of future inflation than the lagged inflation. This finding confirms the forward looking behavior in theory and it is also consistent with our results from the pure forward-looking model estimation. Furthermore, we obtain more evidence of dominant forward-looking behavior by using the principal components based instruments. Our results show that principal components based methods produce more precise estimates with a substantial decrease in all three estimated standard errors. We obtain more evidence of dominant forward-looking behavior across all regressions. By comparing two groups of the Kleibergen-Paap Wald F rk statistic (KP statistic), we find that using principal components is a good option to overcome the weak identifications. This finding is consistent with Bai and Ng (2010) and Kapetanios and Marcellino (2010). However, contrast with our earlier findings, in the hybrid model, the identification of the parameter of the real marginal cost becomes a problem. The third chapter investigates the long-run money demand using a panel data set for the 50 U.S. states from year 1977 to 2005. Regional heterogeneity as well as the cointegration and cross-section correlation properties of panel data are considered in great detail. Contrary to previous studies in the field, we adopt panel data techniques with nonstationary and cointegrated variables which controls for dynamics, non-stationarity, parameter heterogeneity and unobserved time-varying heterogeneity. The empirical results reveal an income elasticity close to 0.7 and an interest semi-elasticity around -0.02 and these two parameter values match closely with the empirical estimates by Ball (2001). Furthermore, it is found that the magnitude of the estimates of error correction term is much less than unity (around 0.05), which suggests that the adjustment time of U.S. money demand to return to its long-run equilibrium may be rather long. Compared to a standard homogeneous panel model of money demand function, our results obtained from heterogeneous panel model estimation indicate that the heterogeneity across states is important. It shows that the observed instability of money demand functions in aggregate U.S. studies could be explained by inappropriate aggregation across heterogeneous states. After accounting for regional heterogeneity, the estimates of income elasticity for the U.S. money demand function are clearly less than one.
This dissertation is only available for download to the SIUC community. Current SIUC affiliates may also access this paper off campus by searching Dissertations & Theses @ Southern Illinois University Carbondale from ProQuest. Others should contact the interlibrary loan department of your local library or contact ProQuest's Dissertation Express service.