Convenient dollars, inconvenient Treasuries?

Kevin Yin is an economics PhD student and a Canadian business columnist.

The global financial crisis was a weird time in many ways. Not least of which was the fact that a chunk of America’s exorbitant privilege seemed to vanish, even though the US dollar seemed to become more dominant by some measures.

Until 2010, US Treasuries were so desirable that they often commanded a “convenience yield”. This meant that people were willing to hold safe dollar assets like Treasuries at yields below a default-free interest rate in return for this extra “specialness”.

Conventional wisdom argues that global demand for the dollar and Treasuries should move tightly together. If the world needs dollars, it needs Treasuries because they are a good place to store those dollars. If the world needs Treasuries, it needs dollars to buy those Treasuries, so the reasoning goes.

But one measure of these convenience yields suggests a messier link even before the financial crisis, and one that might have been severed since. As economists Wenxin Du, Ritt Keerati, and Jesse Schreger document in a recent working paper, while Treasury convenience fell negative after 2010, dollar convenience grew or persisted. Close readers of Alphaville may remember the first half of this point from an earlier blog.

The authors measure convenience using deviations from covered interest parity on government bonds (red) and interbank benchmark rates (blue).

The Decoupling of USD and UST Premia: median government covered interest parity deviations and median benchmark CIP deviations for G10 currency countries at the 10-year tenor © Du, Keerati & Schreger 2025

Covered interest parity says that the return on a US dollar asset should equal the return of a synthetic version of that asset, at least in theory.

A synthetic 10-year Treasury can be created by selling dollars for euros, buying a 10-year German Bund, then entering FX contracts to swap euro coupons and principal payments back to dollars. If Treasuries yield less than the similarly safe copy, it suggests there is something convenient about that asset. Right now, Treasuries do not yield less — and so convenience yields are negative.

Du et. al. do the same exercise for the dollar by replacing bond rates with risk-free rates in short-term interbank markets (SOFR, €STR etc.). At longer maturities, they include interest rate swaps that turn those fluctuating short-term rates into fixed long-term rates. The overarching idea here is that you’re now comparing risk-free cash-like returns instead of bond returns.

In doing this, the authors find that the longer the maturity of the bond, the greater the decoupling between these convenience yields. At 30-year tenors the decoupling is very large. At 3-month maturities, it takes until well after COVID-19 for Treasuries and the dollar to decouple at all.

The Decoupling of USD and UST Premia: median government covered interest parity deviations and median benchmark CIP deviations for G10 currency countries at the 3-month tenor

Why is this happening? On the Treasury side, convenience yields look like they’ve turned negative, at least in part, because US debt has expanded sharply. This is the explanation that Du and her co-authors favour, as well as researchers like Jiang, Richmond, and Zhang (2025).

But what about the dollar side? The blue line in the charts above is basically flat before 2008, suggesting that the dollar just wasn’t convenient, or maybe the deviations are missing something about convenience. Moreover, the measure becomes positive precisely when the Treasury convenience yield dips negative. The timing suggests some kind of relationship between the two.

How to square the circle

One way to think about these joint facts is that these covered interest parity deviations capture both (a) non-pecuniary “convenience” benefits of dollar assets, and; (b) frictions in financial markets. When it is harder to intermediate in FX markets, hedging FX risk can become expensive and thus dollars can look even more convenient. Such frictions increased substantially after Basel III in 2010.

Many have pointed out that regulations implemented at that time made balance-sheet space more costly, limiting dealer banks’ ability to intermediate both Treasuries and currency swaps. In another paper, Du, Tepper, and Verdelhan (2018) argue that:

CIP deviations increase towards the quarter ends, as banks face tighter balance sheet constraints and renewed investors’ attention due to quarterly regulatory filings. […] The term structure of short-term CIP deviations suggest that banking regulations have a causal impact on asset prices.

And so, the convenience of dollars, which may have shown up in actual yield differences before, now appears in covered interest parity deviations because it suddenly costs something to create synthetic dollars (the balance-sheet space of intermediaries).

The exact features of Treasuries that made them convenient — their usefulness in repo markets as collateral — could also make them inconvenient when balance sheets are constrained because they still need to be held for those uses.

The decoupling even appears to be spreading to shorter parts of the yield curve. Convenience yields of 3 month Treasury bills recently decoupled from those of the dollar, which should be troubling for the Treasury Department. It means that at almost all maturities, markets are exacting higher borrowing costs on the US government despite demand for the dollar.

Du and her co-authors haven’t yet published the data for 2026, but it is hard to imagine the latest developments in Treasury markets have improved the situation.