Answer

What a Treasury yield is (answer first)

The starting point for how to read Treasury yields is the distinction that a yield is not a rate the central bank sets. A Treasury yield is the market yield implied by the price of a U.S. Treasury security: the annualized return you would earn holding it to maturity, backed out from its market price. When daily trading moves the price, the yield moves with it. As a rough reading, it is the expected average of future short rates over the life of that maturity, plus a term premium that compensates for holding a longer instrument.

Seen this way, it becomes clear why the policy rate (the federal funds target range or a similar overnight rate) and a long yield such as the 10-year do not have to match. If the market prices future rate cuts, the long yield can sit below the current policy rate; if it prices future hikes or inflation, the long end rises. Lining up yields by maturity gives the yield curve, and separating its level, slope and curvature is the subject of this lesson.

Every yield, spread and breakeven figure shown here is illustrative educational data. It is not a real market value, forecast or trading recommendation. Cross-country FX rate differentials are covered in the interest rate differentials and FX lesson, and the gold and real-yield relationship in the gold and real yields lesson. This article stays on U.S. curve and real-yield foundations. For the wider picture, see the macro analysis guide.

Decomposition

From the policy rate to the Treasury yield: a conceptual decomposition

Viewing a long yield as a single lump makes it hard to see why it moved. The figure below shows a conceptual decomposition: starting from the policy rate, layer on the market’s expected average of short rates and a term premium to arrive at the Treasury yield. This is not a precise model; it is a way to separate the drivers.

This split matters because the same “yield rose” can mean different things. Whether the main driver is higher expected short rates (a hawkish repricing) or a wider term premium (supply, uncertainty) changes how it is likely to spill into other assets. Par yields, spot rates, forward rates, the policy rate and OIS are strictly different objects, but this article works mainly with par-yield levels and differences (close in spirit to the Treasury’s constant-maturity yields) and simplifies the exact derivation of forwards.

Shape

Level, slope, curvature and the four shapes

Splitting the curve three ways keeps it clear. Level is the overall height of the curve (are rates broadly high or low), slope is the difference between short and long maturities (upward or downward), and curvature is whether the middle (the belly) bulges above or dips below the ends. From slope, four representative shapes follow.

• Normal: upward-sloping. Short < long.
• Flat: nearly level. Small gap between short and long.
• Inverted: downward-sloping. Short > long.
• Humped: the middle sits above the ends, forming a peak.

The SVG below overlays four shapes on the same maturity axis (3M, 2Y, 5Y, 10Y, 30Y). It distinguishes them by line style and label as well as color. Every value is an illustrative educational example, not a real curve or forecast.

The key point is that these are names for shapes, not signals about the future. An inverted curve has preceded past recessions, but the timing and depth cannot be confirmed from shape. Keep shape as background information to read alongside other macro series.

Data definitions

The illustrative dataset and aligning dates and units

Before computing any spread, decide what, when and in which unit you are comparing. Mixing values from different observation times or business days breaks the meaning of the difference. This article fixes one illustrative educational dataset up front and reuses the same values in the prose, tables, figures and worksheet. Maturities are 3M, 2Y, 5Y, 10Y and 30Y; the unit is percent (%); and spreads are in basis points (bp, where 1bp = 0.01%).

With real data, record for each series the publisher, series or market code, unit, frequency, reference period, observation time, release time, time zone, seasonal-adjustment status, revision policy and retrieval date. In particular, keep the observation date, period end, release date, retrieval date and revision date as separate fields, and never use a value in a historical analysis before it was actually published (avoiding look-ahead bias). Level, difference, percent change, YoY, percentile, z-score and basis point all mean different things, so do not conflate them. This point-in-time discipline is covered in depth in the macro regime and point-in-time lesson.

When joining series of different frequencies (daily yields with weekly or monthly indicators, for example), state the aggregation rules, week boundaries, holidays, missing values, forward filling, duplicates and time zones. Forward filling an economic series does not make a later release historically available, so handle it with care.

Calculation

Computing 2s10s, 3m10y and 5s30s and converting to basis points

A spread is the yield of a longer maturity minus the yield of a shorter one. Below, the symbolic form, variables, units, substitution, result and interpretation are kept separate.

• Symbolic form: 2s10s = y(10Y) − y(2Y), 3m10y = y(10Y) − y(3M), 5s30s = y(30Y) − y(5Y)
• Variables: y(t) is the nominal yield at maturity t (%). The unit is %, and the spread is in bp (= difference × 100).

Substituting the values from Table 1 gives the following.

Interpretation: this illustrative curve is inverted at the front (3m2y = −45bp, 3m10y = −25bp) while the belly-to-long segment is upward (2s10s = +20bp, 5s30s = +65bp). So it is not a pure, uniformly downward inverted curve; it is a mixed shape that is inverted at the short end and normal at the long end. Reading only 2s10s and concluding “upward, so normal” would miss the front-end inversion. Limits: compare closing values from the same observation time and business day, and remember that a spread is only a difference and carries no information about level or speed of change.

Real yields

Comparing nominal, real and breakeven at matched maturities

To separate whether a yield move is a “real-rate” story or an “inflation-expectations” story, line up the nominal and real yields at the same maturity. A real yield is a nominal yield with the effect of inflation removed; in the U.S., the TIPS yield is the standard observed proxy. Approximate breakeven inflation is the difference between the two.

• Symbolic form: BEI(t) ≈ y_nominal(t) − y_real(t) (same maturity t)
• Substitution (10Y): 4.05 − 1.75 = 2.30 → 10-year breakeven ≈ 2.30%

The figure below splits the 10-year nominal yield of 4.05% into two bands: a real 1.75% and a breakeven 2.30%. Every unit is in percent.

In practice, breakeven embeds an inflation risk premium and a liquidity premium, so read it as a gauge rather than “the market’s exact inflation forecast.” For the same nominal yield, a lower real yield implies a higher breakeven, and vice versa. Analysis built around the real yield, such as the gold and real-yield relationship, can be developed further in the gold and real yields lesson.

Context

The same spread can mean different things

A spread number cannot be read in isolation. The same 2s10s of +20bp can mean different things at different absolute levels, speeds of change and policy regimes. Below are two illustrative educational scenarios.

Both scenarios show 2s10s at +20bp, but A sits near 4% with a high policy rate, while B sits near 1% in an accommodative phase. The same slope carries a different implication once level, speed of change and policy phase differ. That is exactly why you should read the spread, the level, real yields and other macro series together. Do not leap from a single statistic to a price direction; pair it with counter-conditions (for example, whether the front-end inversion resolves or the real yield moves) and additional confirming data.

Educational worksheet

Yield-curve and real-yield worksheet

Enter nominal and real yields across maturities and this educational worksheet computes the 2s10s, 3m10y and 5s30s spreads, approximate breakevens, a curve-shape label and any custom two-tenor spread. It does not judge the future economy or price direction. The defaults are the illustrative data from Table 1. Inputs are computed in your browser only; nothing is sent or saved.

Even with JavaScript disabled, the static result from the defaults (Table 1) is: 2s10s = +20bp, 3m10y = −25bp, 5s30s = +65bp, 10-year breakeven = 2.30%, shape = partial inversion (front inverted, long end normal). The formulas are identical to the body text (Table 2 and nominal minus real).

Limits

What shape can and cannot tell you

The easy stumble in reading the curve is to load too much meaning into shape. Keeping these distinctions in mind avoids over-interpretation.

• An inverted curve does not fix the timing of a recession: past lead cases are known, but the lag and magnitude before any reversal vary, and there are exceptions.
• Correlation does not prove causation: the curve and another asset can appear to move together without one causing the other.
• An extreme value does not guarantee a reversal: even a historically extreme spread need not snap back soon.
• A spread contains no level or speed: the same difference means different things at different levels and speeds of change (Table 3).
• Simplifying assumptions: this article is par-yield-centric and omits the exact differences among spot, forward and OIS.

With these in mind, the curve is not a “standalone conclusion indicator” but background information to cross-check against other series. Testing lead and lag with a time shift is covered in the lead-lag analysis lesson, and turning several drivers into scenarios in the macro scenario analysis lesson.

Practical check

A practical checklist for comparing curves

These are checks for comparing curves on real data. They are for aligning the basis of comparison, not for making trading decisions.

• Are the yields you compare closing values from the same observation time and business day (align time and time zone)?
• Are nominal and real lined up at matched maturities (breakeven is a same-maturity difference)?
• Have you kept spreads consistently in bp and not confused them with the level (%)?
• Have you decided how to handle holidays, missing values and revisions, and checked whether forward filling reflects information “known at the time”?
• Have you looked beyond 2s10s to the front (3m2y, 3m10y) and the long end (5s30s)?
• Have you avoided leaping from one spread to a price direction, and prepared counter-conditions and additional data?
• Have you aligned rounding across the prose, tables and figures, and explained any rounding differences?

Workbench

The workbench workflow and plan scope

Everything above can be followed on real data using the SG Group Macro Research Workbench. The design is graduated: confirm the basics for free and consider higher tiers when you need them. Feature names and scope can change, so this article does not fix them; confirm the current details on the plans page.

This tool mechanically organizes and visualizes public macro data and data you load locally; it does not provide lot, margin, trading-cost, spread, swap, P/L, trading signals or personalized investment advice. Related calculations are handled separately in the lot-size calculation guide and the trading cost calculation guide, and strategy testing in the TradingView backtesting and robustness guide. Other lessons can be found from the English article index.

When to use each

Use cases and caveats for 2s10s, 3m10y and 5s30s

The segment you can read changes with the spread you choose. The table below organizes the use cases and caveats. Each is an observation of shape, not a forecast or trading judgment.

Reading several spreads together makes a mixed shape such as “inverted at the front, normal at the long end” (this article’s illustrative example) harder to miss. If you want to line up rates against COT positioning, the COT percentile and z-score lesson shows how to read positioning extremes.

Summary

Summary: the core answer and your next step

The first thing to grasp about how to read Treasury yields is that a yield is not the policy rate itself, but a market yield reflecting the average of expected short rates plus a term premium. Split the curve into level, slope and curvature, and compare 2s10s, 3m10y and 5s30s in basis points. Compute approximate breakeven as nominal minus real (TIPS), and lining them up at matched maturities lets you separate whether a yield move is a real-rate story or an inflation-expectations story.

In practice, five habits keep you from going badly wrong: (1) align the same observation time and business day, (2) compare nominal and real at matched maturities, (3) do not confuse bp with %, (4) look beyond 2s10s to the front and the long end, and (5) do not leap from one spread to a price direction. An inversion or a spread is background information; it does not fix the timing of a recession or a price direction.

Related

Related reading