Interest Rates and Equity Multiples: When the Link Breaks

What should happen to equity valuations when the Federal Reserve raises interest rates? The standard answer — higher rates compress multiples — is one of the most repeated ideas in market commentary, cited in every earnings-season recap and rate-decision preview. But it's also one of the more historically conditional relationships in investing. Getting it right requires understanding not just the mechanism but the regime in which it operates.

The 2022 experience looked like a textbook confirmation. The Fed raised rates by 425 basis points in nine months, from near zero to 4.50% by December. The S&P 500 fell 19.4% for the year — its worst annual performance since 2008. But look closely at the earnings story: S&P 500 reported earnings actually grew in 2022, finishing the year up roughly 5%. Almost all of the equity market's decline was multiple compression — the market paying less per dollar of earnings, not earning fewer dollars. The discount-rate channel ran exactly as the model predicts.

But it hasn't always worked so cleanly. The 1970s were a decade of sharply rising nominal rates — the 10-year Treasury yield climbed from roughly 6% at the decade's start to more than 15% by 1981 — yet equity multiples didn't follow a clean inverse path. Real equity returns were terrible, but the rate-to-multiple relationship had considerably more noise in it than the standard story suggests. Zoom forward to 2020 and the Fed cut to near zero, multiples expanded to levels not seen since the late 1990s, and the stock market recovered its pandemic losses within months despite a genuine economic catastrophe. The mechanism is real. The conditions that activate it — and the conditions that occasionally override it — are what this framework addresses.

The Discount-Rate Channel

A stock's price is the present value of its future cash flows. That present value is calculated by discounting those flows at a rate that reflects both the opportunity cost of investing elsewhere — the risk-free rate, typically anchored to Treasury yields — and the equity risk premium, the extra return investors demand for taking on equity rather than sovereign risk. These two components sum to the discount rate, and it is this number that connects Federal Reserve policy to stock valuations. Small changes in the Fed's rate path ripple directly into the denominator of every present value calculation in the market.

When the risk-free rate rises, the discount rate rises. When the discount rate rises, the present value of every future dollar of earnings falls. This shows up as multiple compression: a company trading at 20 times forward earnings, all else equal, will trade at a lower multiple when rates are higher. The market doesn't wait for rate hikes to actually land — it reprices as soon as the expected future rate path changes. That's why equity markets often begin selling off weeks before the first hike, and why rate-sensitive sectors move sharply on every Fed statement and CPI print.

The mechanism is amplified for long-duration assets — companies whose value resides primarily in earnings expected years or decades from now. A high-growth technology company projecting most of its present value in years 10 through 30 is far more sensitive to discount-rate changes than a mature business generating stable cash flows today. This duration logic is why the 2022 selloff hit growth stocks hardest: they had the most to lose when the rate used to discount their distant cash flows jumped by more than 4 percentage points. Understanding how interest rates affect stocks at the individual company level follows this exact reasoning.

Running the Numbers: A 1% Rate Change

A simplified perpetuity model makes the mechanics concrete. Assume a company earns $5 per share growing at a steady 4% annually. The Gordon Growth Model gives the multiple as:

P/E = 1 / (discount rate − growth rate)

At a discount rate of 8% with 4% growth: P/E = 25x, implying a $125 stock. Raise the discount rate to 9% — a 1-point increase — and P/E falls to 20x: $100. A 20% decline in fair value with no change to earnings whatsoever.

The effect compounds for higher-growth companies. With 6% growth and an 8% discount rate: P/E = 50x. Add the same 1-point rate increase and P/E drops to 33x — a 33% compression. Amazon's forward P/E collapsed from roughly 65x at the start of 2022 to under 30x by October of that year. Reported earnings expectations for Amazon were barely revised; the price decline was almost entirely discount-rate arithmetic running in real time.

And the model reveals what happens at the zero lower bound: as the discount rate approaches the growth rate, the multiple approaches infinity, and small changes in either variable produce enormous swings. This isn't a mathematical curiosity — it's why 2020 produced the most rapid multiple expansion in decades, and why the subsequent normalization was so violent. Zero-rate regimes don't suspend the discount-rate channel. They stretch it to the point where any reversal hits considerably harder.

Three Historical Analogues

The historical analogue that gets less attention than it deserves is the 1994 'bond massacre.' Greenspan's Fed raised rates from 3% to 6% over 12 months — one of the sharpest tightening cycles in the postwar record — with virtually no advance warning to markets. The 10-year Treasury yield jumped roughly 250 basis points. And yet the S&P 500 fell only about 9% before recovering to finish roughly flat on the year. Why? GDP growth was genuine — real growth ran near 4% in 1994 — and earnings expectations rose as rates did. The discount rate climbed, but so did the growth rate in the pricing model's numerator. The effects partially offset, muting what would otherwise have been a more damaging multiple compression.

The 1999-2000 tightening shows the other side. The Fed raised rates 175 basis points from mid-1999 through May 2000. The Nasdaq fell over 78% from its March 2000 peak to its trough in October 2002. Rate hikes alone didn't cause that — the collision of tightening with starting valuations that assumed discount rates staying low indefinitely made the outcome severe. At 80x earnings, a 1-point rate increase is not a headwind. It's an existential repricing. The lesson: starting valuation determines how much damage any given rate move can actually do.

And 2022: 425 basis points in nine months, starting multiples elevated but not dot-com extreme, earnings holding up. The S&P 500's trailing P/E went from roughly 24x at year-start to below 18x by October. Clean, mechanical, almost textbook. The difference between 1994, 2000, and 2022 wasn't the mechanism — it was the starting conditions each cycle brought to the table.

When the Link Breaks

The 1970s are the most instructive counterexample, and the most frequently misread. Nominal rates rose sharply throughout the decade, eventually reaching extraordinary levels under Volcker — the federal funds rate peaked above 20% in 1981. The Shiller P/E did compress, falling from around 15x in 1972 to below 8x by the late 1970s. On the surface, the model held.

But here's where I find the framework less precise than the clean arithmetic implies: the earnings channel was running simultaneously, in the opposite direction. High inflation meant nominal earnings grew strongly — companies reported higher earnings in nominal dollars even as real profits were flat or declining. The P/E compression was partly masked by inflated reported earnings, making the rate-to-multiple signal considerably messier in real time than it appears in retrospect. Real equity returns were deeply negative for the decade. But an investor relying mechanically on 'rates rising means multiples compressing' as a trading signal would have found the 1970s frustrating and ambiguous. Stagflation breaks the link because it simultaneously inflates the numerator, partially offsetting the denominator effect.

The 2020-2021 period is the second breakage point. At near-zero rates, the pricing model becomes extremely sensitive to small changes in either the discount rate or growth expectations. Equity multiples expanded to Shiller P/E levels above 38x in late 2021 — a level only reached at the peak of the late-1990s bubble. Normal-regime models, calibrated to the range of monetary policy as we've usually known it, don't adequately capture zero-bound non-linearities: additional easing below zero in nominal terms has limited mechanical effect, so the sensitivity concentrates at the moment of reversal. Financial repression extended the low-rate regime long enough for multiples to reach levels implying extraordinary growth for decades.

Stagflation distorts in one direction. Zero-bound regimes distort in another. The channel doesn't break permanently in either case — but it does stop behaving the way the clean model predicts.

The Equity Risk Premium in Context

One practical use of this framework is tracking the equity risk premium — the spread between the earnings yield (the inverse of P/E) and the risk-free rate. When that spread is wide, equities offer meaningful compensation for bearing equity risk over Treasuries. When it narrows toward zero, the cushion protecting current multiples against a rate surprise has compressed. Not because the market must immediately correct — thin equity risk premiums can persist for years, as the late 1990s demonstrated — but because the mechanism that has historically driven multiple compression is working with less headroom.

As of early May 2026, the S&P 500's forward earnings yield is running close to parity with the 10-year Treasury yield — an equity risk premium of roughly 0.02%, one of the lowest readings in the modern data series, per J.P. Morgan's mid-year 2026 research. The Fed held its target range at 3.50–3.75% through its first three meetings of the year, having cut by 1.75 percentage points cumulatively in 2024 and 2025. Forward P/E on the index is running near 26.5x.

There is no precise ERP level that signals imminent danger — if there were, the signal would be arbitraged away long before retail investors could act on it. But a thin premium does mean that a material upside surprise in rates would work through the discount-rate channel with less cushion than historical averages suggest. A 50-basis-point re-pricing of rate expectations would, mechanically, have a larger effect on current multiples than the same move would at an equity risk premium of 3–4%. Not a timing call. A regime observation: the yield curve currently embeds rate stability, and the valuation model is priced for it.

Key Takeaways

The discount-rate channel is durable but context-dependent. Four observations for applying the framework without mistaking it for a formula:

• A 1% rise in the discount rate compresses multiples more for high-duration, high-growth businesses — 33%+ at a 50x starting multiple — than for stable, low-growth ones (roughly 20% at 25x). Starting valuation determines how much damage any rate move inflicts.
• The link breaks in stagflation, because rising inflation simultaneously inflates nominal earnings, partially offsetting the denominator effect. It also breaks at the zero lower bound, where normal-range sensitivity gives way to extreme non-linearity concentrated at the moment of reversal.
• The 1994, 2000, and 2022 cycles each confirm the mechanism while illustrating different starting conditions. 1994 was cushioned by genuine earnings growth. 2000 was amplified by extreme starting valuations. 2022 ran close to textbook.
• An equity risk premium near zero, as in early 2026, means current multiples assume rate stability. A material upside rate surprise would work through the discount-rate channel with less cushion than in normal-ERP environments.