EXCHANGE RATES, FINANCIAL CONSTRAINTS, AND INVESTMENT IN THE U.S. MANUFACTURING SECTOR

This dissertation investigates the impact of the real value of the dollar on investment in U.S. manufacturing using both industry-level and firm-level panel data. Although there is a large literature on investment functions, relatively few studies have estimated exchange rate effects on investment. This study begins by estimating models based on Campa and Goldberg (1999) with more recent industry-level data (for 1976-2005) and more advanced econometric methods. Their finding of a significant positive effect of dollar appreciation on investment through the channel of lower costs of imported inputs cannot be confirmed. However, there is robust evidence for negative effects of the real value of the dollar via the channel of export competitiveness for all industries and ones with high markup rates and import penetration. Alternative estimation techniques are used to correct for serial correlation, heteroskedasticity, and cross-sectional dependence, and as a robustness check for the results. Industry-specific, trade-weighted exchange rates are also used as another sensitivity test, and the results show significant negative coefficients on changes in the export-weighted real dollar index for the whole sample and three out of four industry subsamples.This dissertation then provides estimates of the impact of the real exchange rate on investment at the firm level in U.S. manufacturing for 1995-2010, using a modified accelerator model that controls for the user cost of capital and financial (liquidity) constraints. The dissertation corrects biases in previous micro-level measures of user cost and shows that using chain-weighted user costs eliminates the endogeneity that is found with time-varying weights. Although the trade-weighted real dollar indexes are not significant in the entire sample, there are significant negative effects of dollar appreciation in subsamples defined by high degrees of export orientation, imported input use, import penetration, and all three combined, especially using the indexes weighted by total trade and by imports. Results for whether the dollar index is significant in levels or growth rates vary, indicating some ambiguity about whether the exchange rate affects desired capital stocks, liquidity constraints, or simply the rate at which firms adjust toward their desired capital stocks.