Term Structure of Interest Rates Changes during International Financial Crisis: The Case of Argentina vs. USA

We bootstrapped spot rates for Argentinean and U.S. federal government debt instruments, and fitted them with smoothing cubic splines, a non-parametric method, to estimate the term structure of interest rates. When estimating the term structure one must decide how close should the data be fitted, considering that the curve should be flexible but should also maintain a certain degree of curve stiffness to identify misspriced securities. Smoothing cubic splines are a helpful tool to deal with this trade-off, since the degree of smoothing can be controlled with the smoothing parameter, which must be set between 0 and 1. Our approach is based on that presented by Fisher, Nychka and Zervos (1995); and to choose the “optimal” smoothing parameter value we applied both generalized cross validation and Reinsch’s (1967) methods. The work analyzes the contagion effects that recent international financial crises such as the “Tequila” Mexican crisis, the Asian crisis, the Russian crisis, and the Brazilian devaluation had on Argentinean and U.S. term structures. It also analyzes what happened during the dates in which the 1999 Argentinean President’s elections took place. We found that Argentinean curves changes were significant during all the crises, especially on short-term maturities. Nevertheless these changes were only temporary, since after some time, the curves went back to similar values and shapes to those that existed before the crises had begun. Finally, we applied a test for splines presented by Silverman (1985), based on Wahba’s (1983) previous results, to analyze if the term structure changes were statistically significant or not. The confidence bands calculated by this method resulted too wide, and consequently they could not discriminate among significant and not significant changes.