Numerous studies have resorted to parametric models to infer the shape of the term structure of interest rates. Recently, however, it has been shown that nonparametric techniques may be more adequate. This is an empirical study for Chile between December 1992 and April 1998. Monte Carlo simulations, based upon a nonparametric one-factor model, suggest that Chile's downward-sloping term structure could be explained by the mean reversion process in the data. The latter could reflect medium- and long-term goals of monetary policy of the Central Bank of Chile. Some alternative explanations, such as that of the preferred habitats, might be also plausible.