A previous post discussed the economic distortion caused by deferral of capital gains tax, another, why market-to-market cannot fix this distortion. This post shows how the tax code tries, but does not entirely succeed, in addressing this issue with regard to bonds.
A basic bond is fully described by four, fixed, stated parameters:
- The issuer. The person (individual, corporate, or otherwise) who is obligated to the bond's holder.
- The par value. An amount the issuer is obligated to pay to the holder.
- The due date. The date on which the issuer is obligated to pay the par value to the holder. The time from original issue of the bond to the due date is called its term, the time from the present to the due date, its maturity.
- The coupon. An additional amount the issuer must pay to the holder every period to the holder until the due date. This is typically stated as an annual percentage of the par value, or interest rate.
From this simple structure, a few consequences follow:
First, the market price of a bond, while dependent on its parameters, is not stated or fixed and may vary widely over the course of a bond's life, though of course always converging on the par value as the due date approaches.
Second, note that the market price of a bond must depend on the assessed risk that the issuer will default before the due date. This assessment is by necessity fallible and dependent on many known and unknown variables. But for issuers, such as those with a lot of verifiable assets compared to the amount of outstanding bonds, this risk is sufficiently low as to be treated as effectively zero.
Another set of risk-free bonds are those issued by sovereigns in a currency the sovereign controls. The sovereign can always service the bonds at the trivial cost of just printing the currency they owe and will therefore presumably never have a hard default. (There may still be a soft default where the sovereign essentially negates its obligations by inflating the currency, but that is a separate risk shared by all bonds issued in a currency, regardless of the issuer).
Third, apart from the default risk (or for risk-free bonds), the market price at any time can readily be calculated with a bit of simple arithmetic based on its parameters and a market-wide discount rate, which is typically the expected inflation rate over the remaining term plus some margin which varies only very slowly if at all.
Given all of these facts, it is easy to say how a clever investor could exploit the deferral distortion: buy long-maturity bonds with zero (or low) coupon. Such bonds can be bought at a heavy discount off the par value, so all (or most) of the return will come only at maturity when the bond is effectively sold back to the issuer at par value. This type of return offer two distortionary benefits to the investor:
The first is that the return will consist mostly out of the appreciation between purchase and sale and hence, in principle, qualifies for the treatment at often-lower capital gains tax rates than the ordinary tax rates assessed on interest income. This problem is relatively easy to address by changes in the tax code.
The second, larger and harder, problem is that of deferral which, as discussed in the previous posts, reduces the effective tax rate to close to zero in the sufficiently-long run, regardless of what the statutory rate is. This benefit is so large that it is hard to understand why any private party would issue any long bonds with a coupon higher than zero.
Indeed this loophole is sufficiently obvious that even the tax code has attempted to address it through an unbeloved, complicated set of rules known as Original Issue Discount (OID) rules. Under these rules, bonds which are first issued at a price more than de minimis below their par value, so-called OID bonds, are subject to special requirements. Chief among these is that the holder of an OID bond must pay regular income tax on imputed, "phantom" interest payments on the bonds, which the holder of course never receives. These taxes are then offset at maturity or sale against the capital gains taxes that would otherwise be owed.
The OID rules largely foreclose the exploitation of deferral in the bond context, even as they raise the same liquidity concerns as discussed with respect to the mark-to-market approach.
The reason that the OID rules only mostly solve the problem is that they are keyed off a one-time event: the price of the bond at issuance. If subsequent to issue the market price of a bond falls, for example, because inflation expectations have increased, a clever investor can buy them unencumbered by OID rules and still enjoy a (nearly-tax-)free lunch by deferring the gain.
