Brad DeLong's Semi-Daily Journal (2004)

Bond Prices Over the Borgo Pass

As the coach pulled by four night-black horses labored up the road to the Borgo Pass, the fog gathered, and the children of the night howled outside, I used the coach's wifi connection to surf to the Economist's website. There I found that the Buttonwood Tree was shaking its branches in alarm, dropping leaves like rain, as it contemplated the prices of Romanian and Bulgarian government bonds:

Economist.com: ...Buttonwood had another of those I-can’t-believe-what-I’m-seeing moments on Monday morning when idly perusing a piece of research by Credit Suisse First Boston (CSFB). Bonds issued by Bulgaria and Romania, according to a chart in the report, have rallied so fast in recent months that they now yield scarcely a percentage point over the rate at which the healthiest western banks lend to one another, also known as the swap rate.

In a spirit of honesty, Buttonwood confesses to not having known that these two countries issued international bonds, let alone that such things were written about in serious terms by investment bankers. Well, that is his loss, for bonds issued by both countries have performed wonderfully this past year. Spreads over swaps have halved. After wobbling in April, along with just about every other emerging market, Bulgarian bonds have soared to dizzy heights, and the folk at CSFB expect more of the same. Investors, it seems, are convinced that the prospect of both countries joining the European Union in 2007, remote though it is, makes them as rock-solid a credit as you could wish for. Clearly, emerging markets are once again in vogue, and investors’ appetite for risk is back—and about as selective as it was before, which is to say not very.

This is an updated version of Alberto Giovannini's lira-mark convergence trade strategy of a decade ago--an investment strategy that would have made him absolutely filthy rich--ad Long Term Capital Management been better capitalized, and had it better understood how the asset price variance-covariance matrix changes as you approach the bottom of the distribution. These low spreads make it clear to me that lots of people have thought about Alberto Giovannini's ideas over the past decade, and decided that they make sense. But this does not mean that Romanian and Bulgarian bonds are overpriced: even though spreads are small, when long-term bonds are expensive even small changes in interest rate spreads can have very significant effects on bond values. The interest rate spread vis-a-vis first-world governments may be small, but the bond-price spread is still large--hence there is upside.

Let's do a finger exercise:*

Suppose that we are thinking about ten-year annual coupon bonds, with "solid" countries issuing them at 5%, banks swapping with each other at 6%, and "transition" countries issuing them at 7%. Assume that if transition countries default, it won't be in the next three years--that the next three years of interest payments should be discounted at the same rate as "solid" governments. Then the PV today (discounted at the "solid" country rate of 5%) of the next three years' interest payments is 0.1906 for each euro of "transition" bond bought. Assume that if the transition country joins the EU in three years, its debt is then like "solid" country debt. Then the PV today of the value of transition bonds in three years in the good state of the world is 0.9617 for each euro bought.

Observe that the risk that "transition" countries will fail to join the EU is a largely idiosyncratic one that should not be priced (or should not be expensively priced, provided capital markets are working and you can find enough well-diversified investors to lay this idiosyncratic risk off onto). Then we can ask for what values of p (being the probability that the transition country will join the EU in three years) and R (being the residual value of the transition bond in three years in the bad state in which the transition country does not join the EU) is a current spread of two percentage points above "solid" governments warranted. When I do the numbers, my answers are various: if there is only a 60% chance that EU accession will proceed on schedule, then current prices and interest rates are warranted as long as the residual value R in the bad state is 67% of par; if there is a 70% chance of on-schedule accession, then current prices and interest rates are warranted as long as R is 52% of par; and if there is an 80% chance of on-schedule accession, then current prices and interest rates are warranted as long as R is 22% of par.

This is an illustration of one of my brother's principles of practical finance: if in the deal they give you a near-term cash-flow edge, and if the capital appreciation to be expected in the good state when things go as hoped is substantial, and if the good state is reasonably likely, then not just horrible things but truly unbelievably horrible things must happen to the residual value in the bad state for the deal to be a bad ex ante investment. And for reasonably long-duration assets, when interest rates are low, small reductions in spreads do translate into large appreciations in capital values.

*These are NOT real numbers. This is NOT investment advice. I am NOT recommending the purchase of Romanian or Bulgarian bonds. This is merely a finger exercise. This is a use of a current situation to teach an econo-financial principle. This is neither a suggestion to buy or the solicitation of a proposal to sell any securities whatsoever.