Working Paper: NBER ID: w1184
Authors: George M. Constantinides; Jonathan E. Ingersoll Jr.
Abstract: The assumption that bondholders follow either a buy-and-hold or a continuous realization trading policy, rather than the optimal trading policy,is at variance with reality and, as we demonstrate, may seriously bias the estimation of the yield curve and the implied tax bracket of the marginal investor. Tax considerations which govern a bondholder's optimal trading policy include the following: realization of capital losses, short term if possible; deferment of the realization of capital gains, especially if they are short term; changing the holding period status from long term to short term by sale of the bond and repurchase, so that future capital losses may be realized short term; and raising the basis through sale of the bond and repurchase in order to deduct from ordinary income the amortized premium. Because of the interaction of these factors, no simple characterization of the optimal trading policy is possible. We can say, however, that it differs substantially from the buy-and-hold policy irrespective of whether the bondholder is a bank, a bond dealer, or an individual. We obtain these strong results even when we allow for transactions costs and explicitly consider numerous IRS regulations designed to curtail tax avoidance.
Keywords: bond trading; personal taxes; yield curves; tax implications
JEL Codes: G12; H24
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| tax considerations (H24) | bond trading strategies (G12) |
| bond trading strategies (G12) | bond prices (G12) |
| suboptimal trading policies (F13) | yield curve estimations (E43) |
| tax considerations (H24) | optimal trading strategies (G13) |
| optimal trading strategies (G13) | bond pricing (G12) |