Inverted Yield Curve: What It Means for Currency Traders

A yield curve inverts when shorter-dated government debt pays more than longer-dated debt. That much is agreed everywhere. What is almost never stated is which two maturities the writer had in mind, and the choice is not cosmetic.

Different spreads invert on different dates, sometimes months apart, and Federal Reserve staff research has argued that the one attracting most of the attention is the one carrying least of the information.

That makes the first task a narrow one. Before an inversion can mean anything, a reader has to know which two numbers produced it, why those two, and what part of each is a policy expectation rather than something else. This page is about that, and it hands the currency mechanics to the pages that already carry them.

What an Inverted Yield Curve Is, and Which Curve

A yield curve plots what a government pays to borrow at each maturity, from a few weeks out to thirty years. Normally the line rises: lending for longer earns more. An inversion is the state where some part of that line slopes the other way, and shorter borrowing costs more than longer borrowing.

Notice the words some part. The curve is a whole set of maturities, and it does not tip over as a single object. One segment can invert while the rest of the line still rises. So the sentence the yield curve inverted is not a complete statement. It becomes one only when it names the two points being compared.

This matters immediately, because the segments do not move together and they do not cross zero on the same day. A reader who takes one commentator saying the curve inverted in March and another saying it inverted in August is usually not looking at a contradiction. They are looking at two different spreads, each described as though it were the only one.

It is also worth being clear about where the curve comes from. No institution publishes it as a decision. It is assembled from the yields at which government bonds are actually trading, which means an inversion is an observation about a market rather than an announcement by anybody. Nothing has been declared when a curve inverts; a set of prices has reached a particular arrangement.

Which Spread You Are Actually Reading

Three measures dominate published commentary, and they are not interchangeable.

The first is the 10-year yield minus the 2-year yield, commonly written as the 2-10 spread. It is the one most news coverage means, and Federal Reserve staff have noted the volume of attention it draws in its own right.

The second is the 10-year yield minus the 3-month bill. It uses the same long end and a much shorter front end, so it responds to different things.

The third is the near-term forward spread described below, which uses no long maturity at all.

Read the middle column and the disagreement stops being mysterious. Two of the three measures are anchored at the ten-year point, and a ten-year yield is not a pure statement about interest rates. Part of it is what the market expects policy to average over the decade, and part is the extra compensation lenders demand for tying money up that long.

That second part is the term premium, and it moves for reasons with no connection to monetary policy: demand from pension funds and foreign central banks, the supply of new issuance, changes in how safe long bonds feel relative to everything else. Any inference about policy drawn from a spread anchored at ten years is drawn through all of it.

Before quoting an inversion, then, the first question is not what it predicts. It is which two numbers were subtracted.

The Spread Federal Reserve Staff Research Found More Informative

In 2018 two Federal Reserve staff economists published a FEDS Note examining exactly this question. Their finding was not that the long-term spread is uninformative. It was stronger than that: for predicting recessions, long-term spread measures are statistically dominated by an alternative they named the near-term forward spread.

The construction matters as much as the result. That spread is built entirely from Treasury yields at maturities shorter than two years. No ten-year point enters it, which is precisely why it does not inherit the term premium problem. What it isolates instead is the direction the market expects near-term policy to take, and the authors read a negative value as the market expecting policy to ease.

A 2022 reprise of the same work put the comparison in the sharpest available terms. Someone already tracking the near-term measure, the authors report, would have learned nothing further about the economy ahead from the widely quoted 2-10 spread, and nothing further from the occasions on which it turned negative. Not weaker information. None the other measure did not already carry.

Two things follow for anyone reading curve commentary. The first is that the widely quoted measure is not the one this research supports, so an argument resting on a 2-10 inversion is resting on the weaker of the two available readings.

The second is more useful. Because the near-term measure is an expectation of the policy path, it is expressed in the same terms as the quantity that actually moves an exchange rate. A long-term spread is not, which is why it takes a longer chain of reasoning to get from one to the other, and why more can go wrong along the way.

This is one strand of research rather than a settled consensus, and it is presented here as what a named source found, on a named date, with the reasoning it gave.

What an Inversion Says About a Currency, and What It Does Not

A currency pair is a relative price, so a single country curve is the wrong unit of analysis for it. What can matter is the gap between two countries expected policy paths, and a curve that inverts in one country while the other stays upward sloping is describing a widening of that gap rather than an event in its own right.

The mechanism connecting yields to exchange rates, and the conditions under which it stops working, is not this page. Our page on intermarket analysis sets it out, including when the relationship inverts, and it should be read before any curve reading is turned into a currency view.

There is a second reason to be careful, and it is the one most easily missed. By the time a curve segment inverts, the expectations producing it are already in the price of every instrument that prices off those rates. The distinction between what a market has already discounted and what a committee has projected is developed on our page about the Fed dot plot, and the same reasoning applies here without needing to be repeated.

Where the rate gap does reach a trading account directly is through the financing charged or paid on a position held overnight, and our explanation of the carry trade covers how that arrives.

What an Inversion Cannot Tell You

It carries no timing. The spread is a comparison of prices available now, and nothing in the arithmetic contains a date for anything that might follow.

It is not a forecast in any case, because it is a price. It records what buyers and sellers of government debt agreed to accept for each maturity, which is a description of the present rather than a claim about the future.

It carries no magnitude either. A spread can sit a long way below zero or barely below it, and the arithmetic of the comparison says nothing about whether that difference is large in any economically meaningful sense.

And it says nothing about whether a downturn has begun or ended. Recessions are dated after the fact against monthly economic data, and that process is covered on our page about the phases of the economic cycle. A curve reading is not a substitute for it, and neither one confirms the other.

Frequently Asked Questions

What does an inverted yield curve mean?

It means that at some point along the maturity range, shorter-dated government debt is paying more than longer-dated debt. It describes a relationship between two specific maturities at one moment, so the statement is incomplete until those two maturities are named.

Which yield spread should a currency trader watch?

Federal Reserve staff research in FEDS Notes argues for the near-term forward spread, which is built only from Treasury maturities shorter than two years and reads as the expected direction of near-term policy. That is the quantity expressed in the same terms as the driver of an exchange rate, unlike a spread anchored at the ten-year point.

Does an inversion mean a currency will fall?

No. An inversion is a set of prices in one country debt market and carries no direction for any exchange rate. A currency pair responds to the difference between two countries expected policy paths, and that difference is already reflected in current prices.

Does the 2-10 spread add anything once the near-term forward spread is known?

According to the 2022 update of the Federal Reserve staff research, no. A reader already tracking the near-term measure, the authors report, would have learned nothing further about the economy ahead from the 2-10 spread, and nothing further from the occasions on which it turned negative.

Is the yield curve a timing signal?

No. The spread is a comparison of yields available at one moment and contains no date. Any timing attached to it comes from the person quoting it rather than from the measure.