The Yield Curve

What this article does

If you've made it through Bond Yields 101, you know what a single Treasury yield is. This article is about something more powerful — what happens when you line up ALL the Treasury yields side by side, from 1-month bills to 30-year bonds.

That line of yields is called the yield curve, and its shape is arguably the single most reliable economic forecasting tool ever invented. Not perfect. But more accurate than most economists, more reliable than most surveys, and freely available to anyone who knows how to read it.

By the end of this article, you will:

• Understand what the yield curve is and why its shape matters
• Know what "inversion" means and why every recession since 1955 has been preceded by one
• Recognise the difference between bull steepening, bear steepening, bull flattening, and bear flattening
• Understand the 2022-2024 inversion (longest and deepest on record) and what happened next
• Know which curve measure to actually watch (the 2s10s vs the 3m10y)
• Be able to read the curve as a forecasting tool, not just a chart

This is article 3 of 9 in the Lighthouse foundation series. Bond Yields 101 explained how individual yields work. This article is about the system of yields and what it tells you.

Picture the curve

Imagine a chart. The horizontal axis shows Treasury maturities — 1 month, 3 months, 6 months, 1 year, 2 years, 5 years, 10 years, 30 years. The vertical axis shows yield. You plot a dot for each maturity at its current yield, then connect the dots.

That's the yield curve. The yield curve plots Treasury yields across all maturities on a single chart, from 1-month bills to 30-year bonds. Its shape — upward-sloping, flat, or inverted — encodes the market's expectations about future growth, inflation, and Federal Reserve policy.^[1]

In a normal environment, the curve slopes upward — short-term yields are lower, long-term yields are higher. There's a logical reason: lending money for 30 years carries more uncertainty than lending for 3 months. Inflation could spike, the borrower could default, opportunity costs could rise. Investors demand a premium for taking on that uncertainty. Term Premium.

A typical "normal" curve might have:

• 3-month bill at 4.0%
• 2-year note at 4.2%
• 10-year note at 4.5%
• 30-year bond at 4.8%

A gentle upward slope. The shape says: "We expect things to be fine, with maybe a little more inflation or risk later than now."

But the curve doesn't always look like that. Sometimes it's flat. Sometimes it slopes downward. And when it slopes downward — when short rates exceed long rates — something unusual is happening.

The inversion: when the curve flips

When short-term yields rise above long-term yields, the curve is inverted. Yield curve inversion

This is weird. It's saying: "Lend money to the US government for 30 years and you'll earn LESS than lending for 3 months." Why would anyone accept that?

Because the buyers of long-dated bonds are betting that future short rates will be even lower than today's short rates. They're locking in today's higher long-term yield because they expect short rates to drop.

And short rates drop in one main scenario: when the Federal Reserve cuts them. The Fed cuts when the economy weakens.

An inverted yield curve is the bond market saying: "We think the Fed will cut rates soon — because we think the economy is going to slow down."

That's it. That's the whole signal. The curve isn't predicting recession directly — it's predicting Fed cuts, which are usually a response to recession.

Why this is the most reliable recession signal in financial history

Here's the genuinely remarkable fact: Every US recession since 1955 has been preceded by an inversion of the 2-year/10-year Treasury yield curve. The lead time from inversion to recession start has ranged from 6 months to 24 months, averaging around 14 months.^[2]

That's not "most" recessions. That's not "with some exceptions." Every single one. Six decades of inversions, six decades of recessions following them, with one notable false alarm we'll come back to.

Compare that to other forecasting tools:

• Economists' GDP forecasts: famously bad at calling recessions
• Stock market: drops during recessions but rarely warns in advance
• Consumer surveys: lagging indicators, not leading
• Fed Chair speeches: deliberately vague, by design

The yield curve doesn't have an opinion. It's a price set by millions of bond traders who collectively wager trillions of dollars based on what they actually expect to happen. When that collective wager flips negative — when those traders are paying a premium to lock in long rates because they expect short rates to fall — they're rarely wrong.

The lead time matters

Here's the part most retail investors miss: inversion does not mean "recession next week."

Every US recession since 1955 has been preceded by an inversion of the 2-year/10-year Treasury yield curve. The lead time from inversion to recession start has ranged from 6 months to 24 months, averaging around 14 months.^[2] Average lead time is around 14 months — over a year. Sometimes the gap is shorter (6 months in 1973), sometimes longer (24 months before the 2008 crisis).

This matters because retail investors often see "yield curve inverted" headlines and panic-sell. By the time the recession actually arrives, they've missed 6-18 months of further market gains. The bond market saw it coming. They sold too early.

The yield curve tells you the regime is changing. It does not tell you when. Read it as a smoke detector, not a stopwatch.

Which curve measure should you watch?

There are several "yield curves" depending on which short maturity you compare against the 10-year. The two most-watched:

The 2s10s spread — 10-year yield minus 2-year yield. Most cited in financial media. Tracked daily on FRED as series T10Y2Y. FRED tracks the 10-year minus 2-year Treasury spread daily as series T10Y2Y, with continuous data going back to 1976. The series turns negative whenever the curve inverts, providing a clean binary signal.^[3]

The 3m10y spread — 10-year yield minus 3-month bill yield. Slightly less famous, but Federal Reserve research suggests it's actually the more reliable predictor. The 3-month/10-year spread is sometimes considered an even more reliable recession indicator than the 2s10s. Federal Reserve research by Estrella and Mishkin (1998) established this measure as a leading recession signal with similar lead times.^[4]

For Lighthouse purposes, the 2s10s is the cleanest measure to watch. It's what every financial commentator references. It captures the same signal as 3m10y about 90% of the time. When they disagree, it's usually a sign of something idiosyncratic happening at the very short end (Fed policy mechanics, money-market technicals).

When you see "the yield curve inverted" in headlines, they almost always mean 2s10s.

The 2022-2024 inversion: a record-breaker

The most recent yield curve story is also the most extreme one. The 2022-2024 yield curve inversion was the longest and deepest on record. The 2s10s spread first inverted in July 2022 and remained inverted for over 700 consecutive days, reaching a peak inversion of -108 basis points in July 2023.^[5]

The Fed started hiking aggressively in March 2022 to fight inflation. Short rates rose fast. Long rates rose more slowly because bond markets started anticipating eventual rate cuts.

The 2s10s first inverted in July 2022. It stayed inverted for over 700 consecutive days — the longest inversion on record. At its deepest, the 2-year was yielding 108 basis points (1.08 percentage points) more than the 10-year. That's also the deepest inversion since the early 1980s Volcker era.

By every historical pattern, this should have been screaming "recession imminent." And yet... the US economy didn't enter NBER recession during that period.

That created a lot of "the yield curve is broken!" takes in financial media. It's not broken. What it actually showed:

1. The Fed kept rates very high for very long 2. Markets correctly anticipated eventual cuts 3. The expected slowdown was unusually delayed 4. Fiscal stimulus, COVID savings drawdown, and labour market dynamics extended the cycle

The signal worked. It just had a longer lead time than usual. The curve started uninverting in late 2024, which historically is the more dangerous moment — see next section.

The un-inversion is the dangerous bit

This is the most counter-intuitive part of yield curve analysis: the inversion itself isn't the recession trigger. The un-inversion is.

The yield curve uninverts (returns to positive slope) before recessions actually begin. The combination of inversion followed by rapid steepening is a stronger recession signal than inversion alone — it's the un-inversion that often coincides with the recession's start.^[6]

Think about it. The curve inverts when the bond market starts pricing in future Fed cuts. It stays inverted while the Fed holds rates high. The curve un-inverts (returns to positive slope) when the Fed actually starts cutting — which happens because the economy has weakened enough to require cuts.

So the sequence is typically: 1. Curve inverts → bond market warning 2. Curve stays inverted for 6-24 months → economy still running, but cracks forming 3. Fed starts cutting → curve un-inverts via short rates falling 4. Recession often begins around this point

If you watched only "inversion" you'd panic too early. If you watched only "uninversion" you'd be roughly on time. The smartest macro practitioners watch both — inversion as a warning, un-inversion as a confirmation.

Steepening and flattening — the curve in motion

The curve isn't static. It moves daily, and the way it moves carries signal.

Bull steepening — long rates fall slower than short rates fall. The curve gets steeper because the short end is dropping faster. Bull steepening (long rates falling slower than short rates) typically occurs during Fed rate-cut cycles. Bear steepening (long rates rising faster than short rates) typically occurs during periods of strengthening growth or rising inflation expectations.^[7] Usually happens during Fed cutting cycles. "Bull" because falling rates are good for bond prices.

Bear steepening — long rates rise faster than short rates rise. The curve gets steeper because the long end is climbing. Usually happens when growth and inflation expectations rise. "Bear" because rising rates are bad for bond prices.

Bull flattening — long rates fall faster than short rates fall. The curve flattens because the long end is dropping. Usually happens during recession or risk-off events when investors rush into long-dated Treasuries for safety.

Bear flattening — short rates rise faster than long rates rise. The curve flattens because the short end is climbing. Usually happens during Fed hiking cycles. Persistent bear flattening often precedes inversion.

These four patterns describe almost any curve movement you'll see. Steepening, Flattening.

The pattern matters: a flattening curve from short-end rises (bear flatten) is a hiking cycle in progress. A flattening curve from long-end falls (bull flatten) is risk-off panic. The chart looks similar; the macro implication is opposite.

The one false positive: 1998

History gets one asterisk. There has been one notable false positive in the modern era: the 1998 yield curve inversion that did not lead to a recession. It occurred during the LTCM/Asian financial crisis flight-to-safety period and reversed quickly, with no subsequent NBER recession until 2001.^[8]

In 1998, the 2s10s briefly inverted. There was no subsequent recession until 2001 — well outside the typical lead time window. What happened?

The 1998 inversion was driven by a flight-to-safety rush during the LTCM hedge fund collapse and Russian financial crisis. Foreign investors and hedge funds rushed into long-dated Treasuries for safety, depressing long yields. The Fed, seeing the financial stress, cut rates rapidly — which depressed short yields too. The whole curve moved, but the relative shift caused brief inversion.

The signal was real (financial stress) but the cause was idiosyncratic (an external crisis), not a slow buildup of domestic recession risk. The Fed's response stabilised things and the cycle continued.

The lesson: every signal has noise. The yield curve has been right 100% of the time on directional warnings since 1955, but the magnitude and timing have variance. Don't treat any single signal as gospel. Watch it in context.

How to actually read the curve

Practical guidance for using the curve as a forecasting tool:

1. Watch the 2s10s as your primary measure. It's clean, daily, available on FRED. Lighthouse tracks it directly.

2. Note inversions, but don't act on them in isolation. Inversion = "regime is changing, take care." Not "sell everything."

3. Watch for un-inversion. If the curve has been inverted for 12+ months and starts steepening rapidly — particularly bull-steepening (short rates falling) — pay attention. That's often the actual recession trigger.

4. Cross-check with credit spreads. If the curve is inverted AND credit spreads are widening, the signal is much stronger. If the curve is inverted but spreads are calm, the lag is likely longer.

5. Don't fight the Fed. When Fed policy is the dominant force on the curve (during clear hiking or cutting cycles), curve shape reflects policy more than market expectations of growth.

6. Read steepening/flattening in pairs. Don't just say "curve flattened today." Say "bear flattening" or "bull flattening." The mechanism matters more than the direction.

The Lighthouse takeaway

If you remember nothing else from this article, remember:

The yield curve is a chart of all Treasury yields by maturity. Its shape encodes the bond market's expectations of future Fed policy and economic conditions. A normal curve slopes upward. An inverted curve (short rates above long rates) has preceded every US recession since 1955. The 2s10s spread (10-year minus 2-year) is the cleanest measure to watch. The un-inversion that follows a long inversion is often the actual recession trigger, not the inversion itself. Watch in context, not isolation.

The Lighthouse dashboard tracks the 2s10s directly. The next time you see "yield curve inversion" headlines, you'll understand what's actually being said — and what's not. The bond market has been more right than wrong about recessions for sixty-plus years. Worth paying attention.

The next article in the foundation series is How the Fed Actually Works — because to understand why the curve moves, you need to understand the entity at the controls of the short end.

Test your understanding