Design and Estimation of Quadratic Term Structure Models

We consider the design and estimation of quadratic term structuremodels. We start with a list of stylized facts on interest ratesand interest rate derivatives, classified into three layers: (1)general statistical properties, (2) forecasting relations, and (3)conditional dynamics. We then investigate the implications of eachlayer of property on model design and strive to establish amapping between evidence and model structures. We calibrate atwo-factor model that approximates these three layers ofproperties well, and show that a flexible specification for themarket price of risk is important in capturing the stylizedevidence in forecasting relations while factor interactions areindispensable in generating the hump-shaped dynamics of bondyields.