Friends, if I could show you only one number before you placed a currency trade, it would not be the price of the pair. It would not be the VIX. It would not even be the dollar index. It would be the US 2-year Treasury yield. The 2-year yield sits at the exact intersection of what the Federal Reserve has done and what the market expects it to do next. It is too short to be dominated by term premium and fiscal anxiety. It is too long to be anchored mechanically to the current fed funds rate. It is the market’s best real-time estimate of the near-term path of US monetary policy. And because monetary policy differentials and exchange rates move together, the 2-year yield is where this monitor reads the rate channel. The Fed’s note documents that association; it does not test the 2-year against other tenors, so nothing in it ranks them. Most retail forex traders have never pulled up a chart of the 2-year yield in their lives. That is the edge.
3.5% Dollar Appreciation Per 1 Percentage Point: The Fed’s Own Finding
In May 2024, the Federal Reserve Board published a FEDS Note titled “Monetary Policy and Exchange Rates during the Global Tightening.” It is the clearest institutional evidence available for how rate differentials transmit to currencies, and it uses the 2-year overnight index swap (OIS) rate as its primary explanatory variable. The full note is available here: www.federalreserve.gov Their finding: a 1 percentage point widening of the 2-year OIS differential between the US and the average of eight advanced foreign economies (Australia, Canada, the euro area, Japan, New Zealand, Norway, Sweden and the United Kingdom) is associated with approximately 3.5% dollar appreciation. The note states an association, not a cause, and this article previously wrote “generates” based on the 2021-2024 global tightening cycle, the largest synchronized monetary policy shift in decades. During that period the note’s decomposition accounts for about half of the 13% advanced-economy dollar appreciation by relatively higher US interest rates, and the other half by risk appetite measures (VIX and high-yield credit spreads) — the two moving with currencies together rather than one driving the other. The sensitivity varied by pair. JPY showed the highest beta at approximately 5.0% appreciation per 1 percentage point OIS widening, reflecting Japan’s massive net foreign investment position and the BoJ’s prolonged ultra-loose policy. EUR, AUD, CAD, NOK, NZD and SEK showed significant but lower sensitivities. GBP is the exception the note names explicitly: “With the exception of the British pound, the slope coefficients are highly statistically significant.” For sterling the OIS coefficient is 0.4, with a p-value of 0.78 and an adjusted R-squared of 0.00 — no measurable relationship in this sample at all. This single finding explains why the 2-year yield deserves more attention than any other tenor in the FX trader’s toolkit. It is the tenor the Fed’s own note chose as its primary explanatory variable, and the one whose association with the dollar it quantifies. The note does not compare tenors, so it does not establish that the 2-year moves currencies more than any other.
Why the 2-Year and Not the 10-Year?
This is a question worth sitting with. Most financial media coverage focuses on the 10-year yield. Most yield curve discussions center on the 2s10s spread. So why does the 2-year yield matter more for FX? The answer lies in what each tenor reflects. The 10-year yield is a composite of expected future short rates, term premium, and inflation compensation. It carries information about fiscal policy, government debt supply, and long-run growth expectations. All of that matters for currencies, but it operates over quarters and years, not weeks. The 2-year yield, by contrast, is almost entirely driven by expectations for Fed policy over the next 8 to 24 months. When the 2-year yield moves, it is because the market is repricing the number or timing of rate cuts, or because an inflation print just changed the expected policy path. These are the repricing events that move currencies in real time.
The European Central Bank hosted a conference where Krohn (2025) presented research on “Demand-Driven Risk Premia in Foreign Exchange and Bond Markets.” The paper provides precise timing data: Treasury demand shocks cause USD depreciation that peaks at approximately 10 days and reverts to insignificance after 14 days. The full paper is available from the ECB: www.ecb.europa.eu This 10-day peak aligns with the front-end of the yield curve, not the long end. When the 2-year yield moves on a data release or Fed communication, the FX market responds within days. When the 10-year yield moves on a fiscal announcement or term premium repricing, the FX response takes weeks to quarters. Both matter, and this monitor reads the 2-year over a 1-to-4-week window. BIS Working Paper 1195 reports the long-end half of that contrast, and the attribution matters: the line that monetary policy effects on longer-term yields “build up gradually over time, beyond a horizon of one month” is that paper’s summary of Brooks et al. (2020), scoped in its own words to “at least prior to the GFC.” Adrian, Gelos, Lamersdorf and Moench go on to find that since the GFC the pattern has changed — “post-GFC tightening shocks only have transitory positive effects on yields, which reverse later.” Either way it is the long end, and nothing cited here measures the front-end leg at the horizon named above: www.bis.org
The 3-Month to 2-Year Spread: Your 1-4 Week FX Signal
Here is where the practical application gets specific. The spread between the 3-month Treasury bill and the 2-year yield, what institutional desks call the “3m2s” spread, is one of the most effective near-term FX signals available. The 3-month bill rate is mechanically pinned to the current fed funds rate. It barely moves between FOMC meetings. The 2-year yield, as we just discussed, moves daily based on data releases, Fed communications, and market expectations. So the 3m2s spread isolates the market’s repricing of expected policy changes, stripped of the anchoring effect of current policy. When the 3m2s spread widens (2-year yield rising relative to 3-month), the market is pricing in fewer cuts or later cuts. This is USD-positive. This monitor reads that repricing over a 1-to-4-week window. That window is our convention for the front-curve channel, not a result carried over from any paper cited here — none of them tests the 3m2s spread at that horizon. When the 3m2s spread compresses or inverts (2-year yield falling relative to 3-month), the market is pricing in more cuts or sooner cuts. This is USD-negative. The same 1-to-4-week transmission lag applies. The academic foundation for this transmission timing is a pair of papers by the same two authors, and this article previously ran them together as one. Meredith and Chinn, Long-Horizon Uncovered Interest Rate Parity, is NBER Working Paper 6797 (November 1998); Chinn and Meredith, Monetary Policy and Long-Horizon Uncovered Interest Parity, is IMF Staff Papers 51(3), 409-430 (2004) — different title, different year, reversed author order. Their finding, in the paper’s own abstract: UIP “has been almost universally rejected in studies of exchange rate movements” using short-horizon data, but “these long-horizon regressions yield much more support for UIP — all the coefficients on interest differentials are of the correct sign, and almost all are closer to the UIP value of unity than to the zero coefficient implied by the random walk hypothesis.” Much more support, then, rather than a claim that parity holds: www.nber.org A related Fed working paper by Chaboud and Wright (2003), “Uncovered Interest Parity: It Works, But Not For Long,” works a narrower question than this article previously implied. In their words they are “going in exactly the opposite direction, examining UIP over extremely short horizons”: the comparison is intraday against daily, and they do not test the one-to-twelve-month range at all. The multi-year result belongs to Meredith and Chinn, whom they cite in their literature review: www.federalreserve.gov How this monitor uses them: the 3m2s spread is read on a 1-to-4-week horizon and the 2s10s on a 1-to-4-quarter one. Both horizons are design choices of ours, chosen to match what each spread is about — near-term policy repricing against structural growth and term premium. No paper cited here tests either spread at either horizon, and an earlier version of this article presented both windows as though the literature established them.
The Inflation-Expectations Feedback Loop
Here is a finding from the academic literature that is genuinely counterintuitive and that most retail traders get backwards. Engel and Wu published “Exchange Rate Models Are Better Than You Think” in 2024 as NBER Working Paper 32808. Their model links exchange rate changes to real interest rates, expected inflation, trade balances, global risk measures, and liquidity demand. The full paper is available from the University of Wisconsin: users.ssc.wisc.edu One of their key findings: higher US inflation leads to dollar appreciation when monetary policy is credible. This is the opposite of the textbook prediction that higher inflation should weaken a currency through purchasing power erosion. The reason it works in reverse is that markets price in the Fed’s reaction function. When inflation runs hot, the market expects the Fed to hold rates higher for longer, which widens rate differentials, which strengthens the dollar. This creates a feedback loop that runs directly through the 2-year yield. Hot CPI print comes in. Market reprices the Fed path. 2-year yield rises. Rate differential widens. Dollar strengthens. All within hours of the data release. The reverse is equally true. Soft inflation print. Market prices in earlier cuts. 2-year yield falls. Rate differential compresses. Dollar weakens. Again, hours. Engel and Wu’s finding is that this mechanism has made exchange rate models fit dramatically better since the 1990s, with fit increasing “almost monotonically”
as central banks have adopted credible inflation targeting. Before inflation targeting, the reaction function was uncertain, so inflation data did not reliably transmit to rates and currencies. Now it does. The models describe the data because the central banks are predictable — a result about fit rather than about out-of-sample forecasting. This is also why the 2-year yield is the right tenor to watch. It captures the Fed’s reaction function in real time. The 10-year yield is contaminated by term premium, fiscal expectations, and global savings flows. The 2-year yield is clean.
What the 2-Year Yield Tells You Right Now That Price Cannot
Here is the practical reality that makes the 2-year yield indispensable. Currency pairs can move for many reasons: geopolitical headlines, positioning squeezes, month-end flows, risk-on/risk-off rotations. Some of these moves are noise. Some are signal. The 2-year yield helps you distinguish between the two. If EURUSD drops 50 pips and the US 2-year yield did not move, that is probably positioning or flow-driven noise. It is likely to revert. If EURUSD drops 50 pips and the US 2-year yield rose 5 basis points on the same session, that is a rate-driven move. It has structural backing and is more likely to follow through. If EURUSD drops 50 pips and the US 2-year yield actually fell, you have a divergence. Price is moving one way while the rate channel is pointing the other. That is either a leading signal that the pair will reverse, or a sign that a non-rate channel (energy, geopolitics, risk appetite) is temporarily dominant. Either way, it demands your attention. This cross-referencing is what institutional desks do every day. They do not just look at the pair. They look at the pair in the context of the rate environment. And the rate environment starts with the 2-year yield. The FRED Blog published a visualization of the same channel at a different tenor in February 2026 — Chien and Bloodworth plot the US dollar against the euro alongside the US 10-YEAR Treasury minus the German government bond, not the 2-year spread this article is about. Read it as the long-horizon companion rather than as the same chart. It is also worth reading for its counter-example: after 2 April 2025 US yields rose sharply against German yields, which “in theory … would have implied a stronger US dollar. But instead, the dollar depreciated”: fredblog.stlouisfed.org For free, authoritative, daily-updated 2-year yield data, the Federal Reserve Economic Data (FRED) database provides the H.15 constant maturity series (DGS2). This is the institutional gold standard: fred.stlouisfed.org
The Rate Structure Panel: Where This Comes Together
The entire transmission from the 2-year yield to individual currency pairs runs through a predictable sequence. The 2-year moves. The rate differential shifts. The dollar adjusts. Individual pairs follow, each with their own sensitivity based on their bilateral rate spread and their specific transmission channels. This is what the US Rate Structure panel on the 4xForecaster FX Monitor is built to display. The 3-month, 2-year, and 10-year yields with their daily deltas. The front curve spread (3m2s) and term structure spread (2s10s). Not because the numbers are interesting in themselves, but because they tell you whether the rate channel — the layer this monitor treats as structural — is supporting or undermining the directional bias on every pair downstream. The 2-year yield is the hinge point. Everything above it in the transmission chain (volatility regime, risk appetite) determines whether the rate channel is operating normally or being overwhelmed. Everything below it (DXY direction, individual pair biases) follows from what the 2-year yield and its differentials are doing. It is, without qualification, the most important number in forex. And now you know why. This is Part 3 of the Macro-to-FX Transmission Series from 4xForecaster. Next: Reading the Yield Curve for Currency Direction: The Spreads That Actually Matter.