| This thesis has two main underlying themes: providing a theoretical rationale for the existence of real growth effects of currency crises that explicitly accounts for the role of uncertainty regarding future domestic policies in the formulation of consumption and investment decisions; and distinguishing between apparently observationally equivalent theories of currency crises empirically, with the help of a unique database on gross external assets and liabilities.; In the theoretical section of this thesis, the objective is to demonstrate that there is a rationale to the fear of floating exhibited by policy makers, by showing that a currency crisis can have real short-term costs in terms of economic growth. The crisis and its adverse growth effects are sourced in uncertainty regarding future fiscal deficits, and expectations that these will be financed through seignorage revenues. This introduces volatility into exchange rate and domestic inflation expectations, which enters negatively into the capital accumulation process. The adverse effects on consumption though not necessarily on domestic investment can be partially smoothed if residents are allowed to undertake some domestic transactions in terms of foreign money, and to hold foreign assets. The main contributions of this chapter are to introduce growth into a crisis framework and to derive the contingency plans for consumption and investment in a manner consistent with the stochastic nature of the state of the economy.; In the empirical section of this thesis, I test across alternative, apparently observationally equivalent theories of currency crises, using a unique database on gross external assets and liabilities not readily available previously to researchers. I find evidence in support of insurance-based moral hazard models, as opposed to the liquidity argument. I also provide empirical evidence in support of the model developed in chapter two, by implementing a bi-variate probit model to control for simultaneity in growth and crisis occurrence and using volatility in domestic asset demands as a key predictor. The volatilities are conditional variances generated through the implementation of an ARCH(p,q) model and are also generated as moving averages of standard deviations. |
