Global Bond Carry and Value (Cross-Country Rates)

Thesis (edge)

Bond markets in different countries offer different rewards for taking the same basic risk, which is lending money for a long time. Australia might pay you well above its cash rate to hold a ten-year bond, while Japan pays you almost nothing. Nothing forces those two to be equally attractive, and a global investor can lean toward the one that pays and away from the one that does not.

There are two related signals here. Carry asks what you earn if nothing changes. It is the yield you collect, plus the roll-down, which is the gain you get as a bond ages and slides down a upward-sloping curve toward lower yields and higher prices. Value asks whether a market's yield is high or low relative to its own history once you strip out inflation. Together they answer a simple question: which government bond market is currently paying the most for the risk you take, and which is paying the least?

The trade is constructed as long the attractive markets and short the unattractive ones, so that a global rise or fall in interest rates largely cancels out. What remains is a bet on relative value across countries rather than on the direction of rates, which is a very different and generally less crowded exposure than simply being long duration.

Where it works (regimes)

The strategy earns steadily when monetary policy is diverging and when curves are upward sloping, because that is when carry and roll-down are meaningful and when the cross-country differences are wide. The mid-2000s and the 2015 to 2019 period, where the Federal Reserve was tightening while Europe and Japan were not, offered exactly this.

It struggles when everything converges. When every developed central bank pins rates at zero and buys bonds directly, the differences between markets shrink to noise, and the model is ranking markets on distinctions too small to trade profitably after costs.

It can be badly hurt by direct intervention. When the Bank of Japan committed to yield curve control, it explicitly held the ten-year JGB yield at a level a value model would have called absurdly expensive. Being short that market on the grounds that it was expensive was correct in theory and a loser in practice for years, because a central bank with a printing press can outlast you.

Signals

• Carry: yield plus roll-down. The key insight for a newcomer is that these are two separate things. A bond can offer a high yield with almost no roll-down if the curve is flat, and a modest yield with strong roll-down if the curve is steep at that maturity. What you actually earn if the world stands still is the sum of the two.
• Value: the real yield, meaning the nominal yield minus expected inflation, compared to that country's own history. A market yielding 4 percent with 1 percent inflation is genuinely generous. A market yielding 4 percent with 6 percent inflation is not paying you at all, it is quietly taking from you.
• Volatility adjustment: bond markets differ in how much they move. Rank on carry per unit of volatility, otherwise the model will systematically prefer the most volatile market and mistake risk for reward.
• Combination: score each market on carry and on value, average the two ranks, and take the top and bottom groups. When the two disagree, the position is smaller, which is the correct behaviour.

Portfolio construction

The critical discipline is duration matching. Do not think in contracts, think in interest rate sensitivity. Being long ten Bund contracts and short ten Treasury contracts is not a hedged position, because the two instruments have different durations and different volatilities. Convert everything into a common measure of how much you lose per one basis point move, and construct the book so the long and short legs cancel at the global level.

Hedge the currency. This point is repeated in every serious treatment of the strategy for a good reason: an unhedged long position in an Australian bond is mostly a bet on the Australian dollar, and the FX volatility is several times larger than the rates volatility. Without a currency hedge you are not running a rates strategy, you are running a badly disguised FX strategy.

Rebalance monthly. Set a portfolio volatility target and scale the whole book to hit it.

Risk model

The first risk is that the market-neutral construction is less neutral than it looks. In a genuine global bond selloff, such as 2022, every developed market falls, and the long leg falls more than the short leg protects, because the markets with high carry often have longer duration and more sensitivity. Neutrality on paper is not neutrality in a crisis.

The second risk is intervention. These are the markets where central banks are the largest and least price-sensitive participants. Quantitative easing, yield curve control and direct purchases can hold a market at a level no fundamental model would endorse, and can do so for years.

The third risk is the correlation to carry strategies elsewhere. Bond carry, FX carry and credit carry all tend to lose money in the same weeks, because they are all being paid to take the same underlying risk, which is that funding stays cheap and nothing bad happens. If you run several carry books you very likely have one large position, not several small independent ones.

Stress test against 1994, 2013 and 2022, which are the three episodes where a rapid, global repricing of rates punished exactly this kind of book.

Costs & implementation

Government bond futures are among the most liquid instruments in the world, and direct trading costs are very low. That is the good news, and it makes the strategy realistically implementable at meaningful size.

The complications are mechanical rather than economic. Futures must be rolled every quarter, and the roll has its own cost and its own quirks, particularly in the Treasury contract where the cheapest-to-deliver bond can change and shift the contract's effective duration underneath you. Currency hedges must be rolled too. Different markets have different holidays, so a global book is never fully synchronized.

None of this is fatal, but a backtest that ignores roll mechanics and currency hedging costs is overstating the result, sometimes substantially.

Failure modes

• Leaving the currency exposure unhedged, which turns a subtle rates strategy into a crude FX bet.
• Comparing yields across countries without adjusting for inflation, which makes high-inflation markets look attractive when they are not.
• Ignoring duration and thinking in contract counts, which quietly makes the book directionally long or short rates.
• Believing the strategy is uncorrelated with other carry trades. It is not.
• Shorting a market because it is expensive while a central bank is actively buying it. Being right eventually is not a business model if you are stopped out first.
• Backtesting only on the forty-year bond bull market, where being long duration made money regardless of the signal.

Our Notes & Suggestions

The single most valuable thing this strategy offers is diversification from equity risk, but only if it is built honestly. Once the currency is hedged and the duration is matched, the return stream genuinely looks different from a stock portfolio, which is rare and worth something.

Start simple. Six liquid markets, carry defined as yield plus roll-down, a basic real yield value score, monthly rebalancing, currency hedged, duration matched. Get that working and understood before adding term premium models or curve positioning on top. The sophisticated version is not obviously better than the simple version, and it is far easier to break.

Finally, run the backtest separately across the disinflation decades, the zero rate decade and the recent tightening cycle. If the strategy only works in one of them, you have learned something important about what you are actually being paid for.