Model and Empirical Analysis for Municipal Bond Yield Spreads

Description

One of the most enduring issues in financial research is whether taxes are an important factorin asset pricing. Many theoretical models have been developed to understand the role of taxes inasset pricing. Substantial efforts have also been devoted to empirical tests on this issue. Despitethese efforts, the effects of taxes are not fully understood and there is still considerable room tounderstand the role of taxes on asset pricing.The tax-exempt bond market provides an excellent laboratory to assess the importance oftaxes in the determination of asset values. The well-known Miller hypothesis stipulates that thetax-example bond yield equals the net after-tax equivalent taxable bond yield. Unlike thecorporate/Treasury bonds, investors have no need to pay personal income for the couponpayment from municipal bonds. In equilibrium, the after-tax return of municipal bonds should bethe same as the return from the corporate/Treasury bonds with the same maturity and quality, andthe tax rate implied by this relation will be the equilibrium marginal tax rate that wedges the twomarkets.However, empirical results are not consistent with this expectation. There is evidence thatmunicipal bond yields are much high compare with yields from Treasuries of the same maturity.The relatively high yields from municipal bonds imply a lower expected tax rate. In particular,for municipal bonds with long-maturity, the implied tax rate is much lower.A large literature has attempted to explain the relation between the yields of taxable and taxexemptbonds by considering various factors. However, none of these factors seem to be able toprovide a satisfactory explanation for the muni puzzle (Green (1993) and Chalmers (1998)).In this project, we plan to: (i). extend the municipal bonds pricing model by incorporatingdefault and liquidity factors. The model will accommodate stochastic interest rates and the taxoffsettingeffect of investment expenses which Green (1993) has found important. A uniquefeature of the model is that it builds in a structure for municipal bond liquidity based on thesearch-based theory of Duffie, Garlenau and Pedersen (2005, 2006) and Ericsson and Renault(2006). This approach has a distinct advantage over the factor or intensity approach which maybe confounded by other missing factors in the model due to a lack of structure in thespecification in the liquidity intensity process in a reduced-form model. (ii). Based on theextended model, we will do empirical tests to estimate related parameters and try to explain theyield relation between taxable and tax-exempt bonds.