Abstract
Liquidity and term structure estimation: Non-linear state-space framework
by
Sudarshan Kumar
Term structure estimation in the emerging markets has additional complexity of the infrequent trading in many securities and concentration of the liquidity in the few maturity segments. Infrequent trading in many securities reduces the number of bonds with the complete panel data. Additionally, because of the lack of liquidity in many maturity segments, term structure estimation using small number of liquid securities is not practical. Term structure studies in these markets usually ignore these market specific complexities, and borrow the framework directly from developed market. This study extends an existing popular dynamic term structure framework (Dynamic Nelson Siegel) to non-linear state-space setting to incorporate the impact of liquidity on the bond prices. To incorporate heterogeneous Liquidity of the securities in and across maturity segments, study explores two alternative specifications of the liquidity. In the first, liquidity is defined as a function of the observable proxy variables - trading volume, age of the security and duration. In the second framework, study assumes market liquidity as a latent factor and augment it with the three existing latent actors of the Dynamic Nelson Siegel model. Study explores the impact of the liquidity on the bond prices (Liquidity premium) and its measurement error both, and finds strong empirical evidence supporting that. Study empirically evaluates the proposed framework on the Indian government bond market data for the period 2009-2017.
Committee
Workshop
Key Dates
Communication
StatFin Main Webpage