Taxes and Bond Returns

By Rakesh Bali

E-mail Story Print Story

Capital losses can be deducted except in the case of “wash sales.” A wash sale is defined as a sale within 30 days of the acquisition of a substantially similar property. This rule does not apply to dealers and excludes acquisitions by gift bequest or inheritance. The general deduction procedure is as follows: First, losses and gains are netted within each category. Short-term losses should be deducted from short-term gains, and long-term losses should be deducted from long-term gains. If all losses are absorbed in this fashion, the relevant tax rates apply. Net losses in either tax category can be applied against the other. Note that this would reduce the tax shield from short-term losses for investors whose marginal tax rate exceeds 15%. Remaining net capital losses can be deducted from other income, up to $3,000 per year. If net capital losses exceed $3,000, they can be carried forward to future years.

In addition, for tax purposes, interest income and the bond’s coupon payments are not equivalent. Similarly, capital gain or loss is not equal to the appreciation or decline in the market price of the bond. Both interest income and capital gains or losses are determined by the bond’s tax basis. The basis of a bond initially equals the total purchase price. Current tax law equates the basis to the present value of future coupons and the principal, with the bond’s yield to maturity at purchase serving as the discount rate. For a par bond, this produces a basis that remains constant during the holding period. For a premium bond, the basis declines (i.e., is amortized) until it reaches the principal value at maturity. For a discount bond, the basis increases at the rate of the accrued market discount until it reaches the principal value at maturity. This increment may be included in income on a ratable basis. With the exception of municipal bonds (discussed below), it is irrelevant for tax purposes whether the discount is a market discount or an original issue discount (OID), bonds issued at prices substantially below par. Interest income in any year is defined as the annual coupons plus the change in the basis over that year. Realized capital gain or loss is defined as the difference between the sale price and the current basis.

An important difference in the taxation of interest income and capital gains is the investors’ ability to defer taxes on capital gains until they are realized. The only deferrable taxes on interest income occur when a bond is purchased at market discount, in which case part of the interest income (the accrued market discount) can be deferred.

Although the definitions of interest income and capital gains on tax-exempt municipal bonds are identical to the definitions for taxable bonds, their taxation is different. Not all of the interest income from municipal bonds is tax exempt. Interest paid by the issuer of a municipal bond is always tax exempt regardless of whether it is in the form of coupon payments or in the form of an OID. The accrued interest generated upon disposition by a market discount, however, is always taxable. Because buyers of premium municipal bonds do not pay taxes on coupon income, they do not receive any tax breaks on the amortization of the premium either. Exhibit 1 summarizes the differences in the taxation of returns on premium bonds, market discount bonds, and OID bonds for both taxable and tax-exempt municipal bonds.

Tax-efficient mutual funds have only just begun appearing in the marketplace. Little research exists in the area of the after-tax returns of bond mutual funds. Some research indicates poor after-tax performance of equity funds. More work is needed in the development of appropriate after-tax benchmarks. Investors also need to be disciplined and should have longer investment horizons to avoid paying the higher income tax rate and generating trading costs.

Tax-Induced Bond-Trading Strategies

Research has demonstrated that a simple buy-and-hold strategy is not optimal when personal taxes are taken into consideration. Optimal tax-induced trading may include realizing a capital loss, deferring capital gain, changing long-term status to short-term status (via sale and repurchase), realizing future losses in the short term, raising the basis above par (via sale and repurchase), and deducting the amortized premium from ordinary income.

Example: Short-term loss realization. Consider a three-year maturity par bond paying 10% percent annual coupons that is bought by an investor in the 28% income tax bracket with a three-year investment horizon. Because the bond is purchased at par, its basis will equal $1,000 throughout the investor’s holding period. If the investor holds the bond until maturity, $28 is the tax on annual coupons. The after-tax yield is 7.2%.

Assume that at the end of the first year the yield rises to 15%, resulting in a decline in market price to $918.71. The bondholder could realize a short-term capital loss by selling the bond and repurchasing. If he chooses the latter, a capital loss of $81.29 is realized, generating tax savings of $22.76 and reducing the basis to $918.71. At maturity, the bondholder pays taxes on the accrued interest of $81.29, which increases taxes at the end of Year 3 by $22.76. Incremental after-tax cash flows caused by the sale are shown in Exhibit 2. This is effectively an interest-free loan of $22.76, which the bondholder repays at maturity. This strategy should be pursued only if it is profitable once the transaction costs are considered.

Example: Increasing the basis above par. Given the conditions above, assume that at the end of the first year the yield drops to 5%, resulting in an increase in the market price to $1,092.97. The bondholder could realize long-term capital gain of $92.97 by selling the bond and repurchasing it. This would generate a tax bill of $18.59 (at a 20% tax rate). Why would a bondholder decide to increase his tax bill? The transaction will also increase the basis of the investment to $1,092.97 and would allow the bondholder to deduct the corresponding value of the amortization from the coupon income (taxed at 28%). Exhibit 3 presents incremental after-tax cash flows from the transaction. The transaction amounts to the bondholder’s extending a loan of $18.59, which the government repays in two unequal annual installments of $12.70 and $13.33. The internal rate of return of this tax-free loan is 25.45%. If the strategy of increasing the basis above par is not profitable due to the transaction costs, then by default the strategy of deferring capital gains dominates.

Taxes and Bond-Market Equilibrium

The above examples show that changes in interest rates might allow an investor planning to hold a bond until maturity to increase the present value of the after-tax cash flows through tax arbitrage. As a result, the value of a bond for a taxable investor should be higher by the value of the tax-timing option. Whether or not the value of the tax-timing option is reflected in the bond prices depends on the tax status of the marginal bondholder.

A comparison of municipal and taxable bonds provides insight into the impact of taxes on bond prices. A comparison of short-maturity tax-exempt bonds and Treasury bonds reveals that tax-exempt yields are lower than taxable yields by 25%–35%. This empirical evidence suggests that the marginal bondholder is taxable and the market equilibrates on an after-tax basis. The estimated value, however, does not necessarily represent the marginal tax rate of the marginal investor.

First, it is possible that investors with different marginal tax rates prefer different bonds, in terms of their tax status, maturity, or coupon rate. Consequently, different tax-clienteles may be “setting” prices for different types of bonds. Schaefer (“Taxes and Security Market Equilibrium,” in Financial Economics: Essays in Honor of Paul H. Cootner, 1982) examined a capital market with redundant securities, where investors face heterogeneous marginal tax rates and no frictions other than taxes, and found that equilibrium is impossible because arbitrage opportunities exist. Introducing frictions such as trading costs and short-selling constraints can preclude arbitrage and generate equilibrium.

Second, investors may optimally follow complex tax-induced trading strategies, which would differ across taxable and tax exempt instruments. Thus, both the tax clientele and the tax-timing option imply that the ratio of taxable and tax-exempt yields would not equal the tax rate of the marginal investor. Nevertheless, taxes have an important effect on bond prices that investors should be aware of.

Tax-exempt bonds allow one to defer taxes indefinitely. Taxes on Treasuries are deferred at least until the end of the current tax quarter (when an estimated tax payment will be made). Consistent with the prediction that intra-year changes in the tax-exempt-to-taxable-yields ratio should reflect changes in the tax deferral, one study (S. J. Hochman, O. Palmon, and A. P. Tang, “Tax-Induced Intra-Year Patterns in Bond Yields,” Journal of Finance 48, 1993) found that the yields ratio exhibited a sawtooth pattern varying between 0.55 and 0.58. Green and Odegaard (“Are There Tax Effects in Relative Pricing of U.S. Government Bonds?,” Journal of Finance 52, 1997) examined the effect of the 1986 Tax Reform Act using a sample of pre- and post-1986 prices of U.S. Treasury bonds. Their estimates of the implicit tax rates are positive (mean estimated tax rate is 16.05%), and significant for the pre-1986 era, but are close to zero for the period following the act. The authors attribute this difference to changes in both the tax laws and the institutional environment. By redefining market discount as interest income and requiring linear amortization of the market premium, the 1986 act reduced the magnitude of tax effects on bond valuation. In addition, the rise of derivatives markets and the increased variety of Treasury bonds in this period enhanced the ability of symmetrically taxed institutions to arbitrage away any tax premium.

Tax-exempt prime-grade municipal bonds trade at a premium relative to taxable Treasuries; however, this premium declines with maturity. Several explanations of this puzzle have been proposed. First, it is possible that higher default risk and the call features of long-term municipals may explain the relation. Research (C. Trczinka, “The Pricing of Tax-Exempt Bonds and the Miller Hypothesis,” Journal of Finance 37, 1982) has found, however, that the puzzle persists when the municipals are compared to corporate bonds of similar credit rating. J.M.R. Chalmers’ “Default Risk Cannot Explain the Muni Puzzle” (Review of Financial Studies 11, 1998) examined a sample of municipal bonds secured by the U.S. government and found that long-term yields on these bonds were higher than predicted by theory. Because the bonds are effectively risk-free and noncallable, the author ruled out default risk and the call feature as sources of the long-term-yield differential.

A second explanation for Chalmers’ results is the tax-timing option, the value of which increases with maturity. The tax-timing option increases the value of a taxable bond, and more so for longer-term bonds. It is not clear, however, whether the tax-timing option affects bond prices significantly. R.H. Litzenberger and J. Rolfo (“Arbitrage Pricing, Transaction Costs and Taxation of Capital Gains,” Journal of Financial Economics 13, 1984) examined this issue by focusing on three government bonds with the same maturity. The value of a tax-timing option attached to a single bond is lower than the value of an option attached to a bond portfolio with a promised stream of cash flows. This could exist only if trading costs or short-selling constraints prevent arbitrage by tax-exempt investors. The results supported the tax-timing option hypothesis and showed that its effect on bond prices is bounded by the no-arbitrage condition for tax-exempt investors.

The 1986 act began a phase-out of the interest paid on borrowings other than a residence. An analysis of the components of household debt before and after the tax act (D.M. Maki, “Household Debt and the Tax Reform Act of 1986,” American Economic Review 91, 2001) found that households moved their portfolios away from consumer debt into mortgage debt, supporting the contention that taxes affect consumer behavior.

Taxes and Corporate Debt Financing Behavior

Taxes also affect the behavior of the corporate, municipal, and government entities that issue bonds. This affects the supply side of the equation and thus bond returns. One advantage of debt financing for corporations is the tax-deductibility of interest payments. Companies with net operating losses have zero tax liability, and losses can be carried back or carried forward to reduce tax payments.

Empirical evidence is consistent with this theoretical prediction. Graham (“Debt and the Marginal Tax Rate,” Journal of Financial Economics 51, 1996) found that companies with high marginal tax rates were more likely to issue new debt. Similarly, Manzon et al. (“Evidence on the Effect of Taxes on Firms’ Decisions to Retire Debt Early,” Journal of Financial Research 19, 1996) showed that companies with low marginal tax rates were more likely to retire their debt early. They also found that sometimes companies retired their debt at a loss when such a loss could be immediately deducted from their taxable income. Graham (“How Big Are the Tax Benefits of Debt?,” Journal of Finance 55, 2000) also found that companies with large market capitalization, high liquidity, and high profitability used debt conservatively. The tax benefit of debt was found to be about 10% of firm value and about 4% net of personal taxes.

Research (R. Varma and D. Chambers, “The Role of Financial Innovation in Raising Capital: Evidence from Deep Discount Debt Offers,” Journal of Financial Economics 26, 1990) has examined the stock market effects of announcements of new issues of OID bonds in 1981–1982. Because OID bonds are issued with coupon rates well below the market yields and issue prices that are below par, the coupon payments represent only part of the return; the balance is paid along with the principal at maturity.

Varma and Chambers hypothesized that the impact of OID bond issuance on shareholder wealth may differ from the impact of par bond issuance. Until July 1982, OID bond issuers could deduct a fixed straight-line portion of the original issue discount from their income every year. This rule ensured that the present value of the tax shield provided by OID bonds was higher than that provided by par bonds. A decision to issue an OID bond might signal that managers expect to have enough earnings to utilize extra tax shields. Investors react favorably and push the stock price up. The authors found positive significant (0.8%) two-day abnormal returns upon announcements of OID bonds. These abnormal returns disappeared after 1982.

In the 1980s, numerous high-yield bonds were issued at substantial discounts to face value (high-yield discount obligations; HYDO) paying low coupons or none at all, instead making payment-in-kind, PIK, distributions. These bonds were used to finance many highly leveraged transactions. Under the prevailing tax law, the accrued market discount was deductible as an interest payment, even though the payment was made much later at maturity. The Revenue Reconciliation Act of 1989 made a certain portion of this discount deductible only at maturity, and a certain portion disallowable depending upon the yield of the bond at issue compared to the Treasury’s yield of similar maturity. The issuance of such bonds diminished greatly after the 1989 act.

Active Strategies

Differential taxation of interest income, as well as long- and short-term capital gains, creates incentives for investors to follow tax-induced trading strategies. Such trading strategies might allow investors to increase their after-tax returns more than a simple buy-and-hold strategy. Companies have tended to overuse existing holes in the Tax Code, which has led to legislation changing the tax structure.