Essays on U.S. state-level financial wealth data and consumption data

A new direction in emerging research is to use geographical variations to address questions in macroeconomics that have not been resolved using aggregate data. This dissertation consists of three essays that examine geographically disaggregated U.S. data that are key to central questions in macroeconomics. The first one describes a new panel dataset for the financial wealth of U.S. states, as constructed from anonymous proprietary account-level records of geographic wealth holdings. The essay provides evidence that the new dataset is a reliable measure of stock wealth growth at the state level. It shows that the state-specific stock wealth growth is significantly correlated with the growth of local average annual income. Nevertheless, the essay finds insignificant correlations between financial wealth and both average and median housing wealth. Additionally, there is evidence that the state- level poverty rate, sex ratio, and state income from lottery sales are all associated with state-specific financial wealth growth.;The second essay constructs a state-level measure of consumption spending, one that improves significantly on existing data sources. It compares the new dataset with four other retail sales measures, and discusses their respective advantages and disadvantages as a proxy for state consumption. The essay provides evidence that the two most popular measures of retail sales involve some mysterious prior assumptions, which could contaminate studies of consumption behaviors. It also shows that the new dataset constitutes a superior measure of consumption growth. The most persuasive evidence comes from those states in which taxable retail sales are measured directly by the government.;The third essay combines the two datasets to provide new estimates of the effects on consumption from changes in the stock and housing wealth. The paper finds large but sluggish housing wealth effects. The estimated impact on consumption of a one dollar change in the housing wealth that took place two years prior is about 6 cents. However, the data show no evidence of significant stock wealth effects. Additionally, we find mixed evidence for statistically significant differences between housing and stock wealth effects.