Assessing NSE’s Daily Zero Coupon Yield Curve Estimates: A Comparison with Few Competing Alternatives

Abstract

This study assesses and compares, on select criteria of evaluation, the time series of daily term structure estimates provided by the National Stock Exchange (NSE) [using the Nelson-Siegel (1987; NS) methodology] with author’s own estimates of NS, Svensson (1992; SV) and Cox-Ingersoll-Ross (1985; CIR). While no model comes across as best on all criteria of evaluation, NS as estimated by NSE (and in this study) turn out to be the worst of the lot. Wherever CIR comes out to be better than SV, however, the difference is only marginal. As none of the models came out to be best on all days, given the numerical tractability of parsimonious models and availability of relatively cheap computing resources, it is suggested that NSE report estimates based on 3/4 competing specifications and not just using NS, which should be phased out. A suitable alternative exists in SV.