Toby Nangle recently wrote “How to (more) properly compare bond yields across markets” (non-gift link). The story behind the article is straightforward: commentators are going back to their old habit of comparing the raw yields on 10-year bonds and making assertions about what this means about implied credit quality. As Nangle’s article notes, this is not a good idea, since bond yields embed rate expectations.

Since I do not have the derivatives data to dig into current pricing, I will just offer some basic principles.

When Raw Yields Matter

If you want to speculate on currency movements over long horizons, you can buy foreign currency bonds.

For example, in the past I bought U.S. Treasurys when I felt that the Canadian dollar was relatively expensive and Canadian governments were trading with yields less than Treasurys. Even if I was wrong about the Canadian dollar, the yield cushion on Treasurys meant that the U.S. dollar had to depreciate by the amount of the yield spread over a multi-year horizon. For two countries that used to be major trading partners and with similar inflation profiles, trend depreciation seemed unlikely.

But if you are a portfolio manager working after the 1990s, taking currency bets is the job of the forex team, not fixed income. The short-term volatility of currencies dwarfs that of the relative performance of bond returns. (In the long-term, the belief that currencies revert to purchasing power parity diminishes the currency effect on returns.)

“Normal Hedging” — Short-Term Lending

You have a currency risk in your portfolio if you are net long that currency in terms of asset values. The “easiest” way to get a hedged exposure to a foreign bond without taking currency risk is to borrow short-term in that currency, and buy that bond. That is, call up your friendly broker and ask for a $100 million loan so that you can buy a $100 million bond. Although some entities can arrange such loans, in most cases the broker would just laugh. You need a better structure, and a fixed-floating interest rate swap gives you the economic exposure of such an arrangement, but with more easily managed counterparty risks.

To rephrase: an interest rate swap has no initial investment, and thus changes to currency values have no effect on the value of the swap. Eventually, the swap will make/lose money, and the net present value of this profit/loss will be needed to be converted to your local currency, but the exposure is to the profits, not the notional value of the swap.

However, the payoff of the swap is the fixed leg rate (equivalent to the bond yield) versus the floating rate of the currency. I.e., the equivalent to the floating rate on a loan to buy the bond. This is not an apples-to-apples comparison to the outright purchase of a bond — which has no financing cost embedded in its return.

The way in which real money investors/issuers can lend/borrow across currencies is the cross-currency basis swap market. A cross-currency swap (link to earlier article) is a linked set of borrowings in two currencies packaged into a single derivative structure. If I am a Canadian investor with C$100 million that I want to plough into a 10-year gilt, I can enter into a 10-year cross-currency basis swap where I lend C$100 million for 10 years at the Canadian floating rate, and borrow the equivalent amount of pounds (based on the spot exchange rate), and I then buy a 10-year gilt. The interest payments I get are:

Net interest = 10-year gilt rate (fixed) - (floating GBP rate) + (floating CAD rate) + (basis).

The “basis” in the previous sentence is the fixed spread attached to the basis swap. Cross-currency basis trading occurs in an opaque wholesale market, and the imbalances in supply/demand for currencies in the market results in a spread being attached to the contract to allow the market to clear.

If you look at that structure, it looks almost identical to entering into a GBP swap, and investing in short-term CAD paper as collateral. In terms of basic payoffs, this is correct — this is basic financial engineering. However, the basis spread reflects balance sheet constraints for international investing.

To the extent there is a lesson, it is this: once you hedge out currency risk, each currency operates as its own separated world. The payoff that matters for fixed coupon bonds is versus floating rate. Which means the absolute level of the bond yield does not matter, just its carry.

This caught some old school investors that used to advocate long Australia/short Japan trades at the end of the 1990s/early 2000s. Australian yields were the highest in the developed world, towering over “unsustainably low” (lol) Japanese yields. Unfortunately for any fixed income manager who could not take forex risk, the Australian curve was deeply inverted, and the Japanese curve often was the steepest in the G7 — they got eaten alive by carry. (If they could take currency risk, they did OK, until they didn’t. The yen went down the escalator, and up the express elevator.)

Comparisons using Currency Swaps

If you want to compare government yields across currencies, you largely end up looking at the cross-currency basis swap spreads. If the basis spread is large, it can be cheaper for a issuer to go into a foreign currency and issue bonds, even if the spread in the foreign currency is wider than it pays in local currency. This is a standard tactic for large corporations and even Canadian provinces.

As such, financing costs implied by cross-currency basis swaps are telling us mainly about conditions in that market, and not the governments’ credibility.

Fixed Currency Swaps

Nangle’s article does the analysis for fixed-fixed currency swaps. Unless things have radically changed (which I doubt), these are not a widely traded instrument. The base instrument are floating/floating for the very good reason that the main users of the swaps live in a floating rate world. A fixed-fixed currency swap is just a combination of 3 liquid derivatives: a floating/floating cross-currency basis swap, and two fixed-floating interest rate swaps in the two currencies.

The end result of such a swap is just telling us about the cross-currency basis swap spread, and the two spreads embedded in the interest-rate swaps. (Since swaps and bonds are not fungible — swaps have no up-front investment — there is a spread between a cash bond and the corresponding swap rate.)

What if I Want to Know About Credit Risk?

If you want to isolate credit risk, you have two options.

1. You can use credit default swaps (CDS). The credibility of CDS protection on developed floating currency sovereigns is low.

2. You need to compare bond yields to another cash instrument that has no perceived default risk. (Since swaps are not a cash instrument, they are not a valid comparator.)

Unless you are in the euro area (with multiple sovereigns sharing the same currency), you are going to find that the only curve with no perceived default risk is the central government curve. The only time you see other bonds trading through the government curve is for tax reasons, or some other stupid technicality.

Concluding Remarks

If it were possible to easily outperform bond benchmarks by buying developed countries’ bonds that have higher yields, you would actually hear stories of people actually succeeding with that strategy. (Buying developing market bonds is a different story, it can be successful.)

Appendix: Fed Meeting

A small rate cut Wednesday seems to be on the cards. Although a lot of ink will be spilled about the Fed kowtowing to Trump, I do not think even a string of 25 basis point rate cuts would make much of a difference. I think it will only be perceived to be a policy error if the labour market tightens, and conditions there are somewhat muddled.

Email subscription: Go to https://bondeconomics.substack.com/ 

(c) Brian Romanchuk 2024