The Federal Reserve publishes two Treasury yield curves every day. One is the ordinary nominal curve. The other is the inflation-indexed curve, the yields on Treasury Inflation-Protected Securities, whose principal rises with the consumer price index.

Subtract the second from the first and you have the breakeven inflation rate: the average annual inflation over that maturity at which an investor would be indifferent between the two bonds. It is not a survey or a forecast. It is the rate at which real money is willing to take the other side.

What it says right now

On the Fed's H.15 release for August 20:

The 5-year nominal Treasury yielded 4.39% and the 5-year TIPS 2.05%, a breakeven of 2.34%. At 10 years, 4.69% against 2.35%, a breakeven of 2.34%. At 30 years, 5.23% against 2.95%, a breakeven of 2.28%.

So the bond market expects inflation of a little under 2.4% a year for the next five years, the same for the next ten, and slightly less for the next thirty. Against a Fed target of 2%, that is a modest overshoot, and the curve is close to flat, meaning no expectation that it gets worse over time.

Why that is awkward for the week's story

This has been a week of debasement talk. The Treasury doubled its buybacks of long-dated debt on Wednesday and yields rose anyway. Robin Brooks of Brookings warned that the US is "playing with fire" and invoked the yen's long decline. Gold rose, bitcoin had its best week since 2023, and Ray Dalio told his followers to hold 10% to 15% of a portfolio in gold. The trade has a name now.

If that thesis were being priced, it would appear here first. A market that genuinely expected the currency to be inflated away would demand a much larger premium to hold a bond promising fixed dollars, and the 30-year breakeven would be well above the 5-year rather than slightly below it.

It is not. Whatever is driving gold and bitcoin this week, the largest and most liquid market in the world is not pricing an inflation problem.

What the numbers do show

Something else, and arguably more important: real yields are high.

The 10-year TIPS yield of 2.35% is the return an investor gets after inflation, guaranteed by the US government. For most of the decade after 2008 that number was near zero and often negative. A real yield above 2% is a genuinely restrictive setting, and it is the actual reason mortgages are near 6.65%, technology companies are paying wider spreads to fund data centres, and the equity market fell on a day when yields rose.

The 30-year real yield at 2.95% says the same thing more strongly at the long end. That is what the buyback was trying to address, and it is what has not moved.

So the correct reading of this week is probably not that investors fear inflation. It is that they are demanding more compensation for lending to the US government for a long time, given the deficit and the volume of issuance ahead, and that is a different problem with different remedies. Inflation expectations respond to a central bank. Term premium responds to fiscal policy.

How to use this

Anyone can check it. The H.15 release is published every business day, the inflation-indexed rows sit below the nominal ones, and the subtraction takes a moment.

It is a useful discipline whenever a market narrative gets loud, because it is one of the few numbers in finance that reflects what people are doing rather than what they are saying. When the breakevens move, the debasement story will have evidence behind it. Until they do, the evidence is in gold and bitcoin, two assets that also rise for reasons that have nothing to do with the price level.

Nothing here is a view on where any of these rates go next, or advice on what to hold. It is a description of what the market is currently pricing, which is a question of arithmetic rather than opinion.