Date of Graduation


Document Type


Degree Type



College of Business and Economics



Committee Chair

Arabinda Basistha

Committee Co-Chair

Jack Dorminey

Committee Member

Stratford Douglas

Committee Member

Brad Humphrey

Committee Member

Adam Nowak


This dissertation is a collection of essays examining current issues in asset price forecasting. The first chapter of this dissertation discusses the relationship between investor sentiment and excess stock return. This essay takes a novel approach in estimating investor sentiment use social media posts. In this study, I construct daily, equity-specific, investor sentiment indexes from Twitter and test the efficient market theory. We use a multinomial inverse regression to build the dictionary of relevant words and phrases for construction of the indexes. We find that our investor sentiment measure has a positive and statistically significant effect on individual stock returns. These findings are robust to different models and specifications. Chapter 2 examines the ability of international sector predictors to forecast US housing price inflation. Under floating exchange rate regimes, the Dornbusch model predicts shocks to domestic or foreign economies will be reflected in exchange rates. When exchange rates are fixed, shocks are likely to affect the net foreign asset holdings. In this study, I examine the role of the exchange rates and the net change in foreign asset holdings in improving US real estate inflation forecasts. I conduct in-sample and out-of-sample comparison of forecasting models relative to an autoregressive baseline model. I find that inclusion of foreign sector variables can improve the US real estate inflation forecasts by up to 40 percent. This improvement is mostly driven by changes in the net foreign asset holdings at longer horizons. The results are robust to samples at the metropolitan level although with different gains. Chapter 3 continues with this line research. Here, I determine the ability of net capital inflows from regions to forecast US housing inflation. Over the last decade, there has been a high correlation between balance of payment measures (Current Account deficits and Net Financial Accounts). The international finance theory has focused on determining the cause of this relationship. Specifically, this theory has found that deregulation in credit markets, accommodative US monetary policy, and fixed exchange rates caused US housing prices and balance of payments measures to move together. In 2015, BEA released new estimates of balance of payments measures in line with international standards, such that now bilateral financial account data has been created. In this study, I use a number of components from bilateral financial account data, to forecast US housing prices. Further, to empirically test the implications of the international finance theory, I use factor analysis methods to create an bilateral financial account index to forecast US housing prices. Overall, I find that many of these measures are able to produce improved forecasts of up to 50 percent.