The Shock Absorber

Four hundred basis points in under five years

Five years ago, lending to the United States government for ten years guaranteed you a loss of purchasing power. Today it guarantees you a gain of a third.

That is the whole of it. Four hundred and one basis points on the real risk-free rate in under five years, turning a guaranteed erosion into a guaranteed compounding, with no credit risk and no reinvestment assumption. Lock in the thirty-year at its 9 October close of 3.32% and $100 becomes $266.76 in real terms over its life.

This is not a cyclical level. The series behind this article runs monthly from January 2000 to September 2026 — 319 observations, and the whole of the record for inflation-protected Treasuries of any useful length. In the 287 months from July 2002 to August 2026, exactly one carried a higher ten-year real yield than today's. It was October 2008.

A note on why I quote the real yield rather than the nominal, since the nominal is what gets headlined. On a nominal basis the ten-year is back to roughly where it stood in March 2002. On a real basis it is close but not quite: the real yield at the March 2002 month-end was 3.32% against today's 2.92%. Forty basis points is not a wide gap, and it does not change the conclusion — today's real yield is a generational level by any reading. What it does show is that the two measures are not interchangeable. A nominal yield can be flat against its own history while the return an investor actually keeps has moved, because the inflation compensation inside it has changed. That is a small point here and a large one elsewhere, which is why the real yield is the measure used throughout.

And alongside it, the S&P 500 sits near its record. A four-point rise in the real risk-free rate appears, on the face of it, to have cost the equity market nothing.

It has cost a great deal. The cost simply has not been paid where most people look for it.

October 2008 is the only comparable month, and it is not reassuring

The single month in the last twenty-four years with a higher real yield than today's deserves a closer look, because the two situations could hardly be less alike.

In October 2008 the real risk-free rate was higher than it is today, by 21 basis points. And yet equities offered nearly six points of compensation on top of it, because the market was priced at eleven times forward earnings. Today, with a marginally lower real yield, equities offer 2.36 — forty per cent of what they offered then.

What "eleven times" means in plain terms is worth spelling out, because it is the whole comparison. In October 2008 a dollar of expected annual company earnings cost an investor $11.03. Today the same dollar of earnings costs $18.93 — seventy per cent more. The company profits being bought are no cheaper than they were; the price attached to them is far higher. That is what "equities were being handed away" means, and why they are not being handed away now.

October 2008 was a dislocation rather than a trend. The real yield was 2.24% in September that year and 2.54% by November; it spiked for a month on forced liquidation and deflation pricing, then subsided. Worth noting too that the forced selling was in inflation-protected paper itself, which trades thinly in a crisis — so the 3.14% print probably overstates the real return genuinely available that month. If so, today's reading is closer to the highest in twenty-four years than to the second highest.

But the comparison stands either way, and it is the point. The last time a buyer of inflation-protected Treasuries could lock in this much guaranteed real return, the alternative was unusually cheap. This time it is not.

The long end is selling off for real reasons, not inflation ones

Before going near equities, it is worth decomposing the curve itself, because the answer is unambiguous and most commentary gets it backwards.

All figures are the close of business on 9 October 2026. The thirty-year breakeven is derived as the nominal yield less the real yield.

The long end is steeper than the belly by 36 basis points. All of it, and more, is real yield. Inflation compensation is inverted — the market prices less inflation over thirty years than over ten, and that subtracts eight basis points from the steepening rather than adding to it.

Whatever is driving investors to demand more to hold thirty-year paper than ten-year paper, it is not a fear of inflation. It is the price of duration itself: supply, term premium, the willingness to hold a very long real liability. That matters for everything that follows, because it means the discount rate applied to long-dated equity cash flows is rising for reasons that have nothing to do with nominal growth — and therefore nothing that rising nominal earnings would naturally offset.

One further consequence. At the September month-end the thirty-year real yield was 3.33%, forty-one basis points above the ten-year's 2.92%. Every spread in this article is measured against the ten-year, for consistency with the twenty-six-year history. Measured against the thirty-year on the same date, the equity market's compensation falls from 2.36 to 1.95. A reader weighing a thirty-year bond should understand that the comparison against that instrument is less favourable to equities than the one presented throughout. The 1.95 figure cannot be ranked against the historical series, which is built on ten-year yields — it is a like-for-like reading on one date, not a percentile.

Whether a retail investor should be looking at a thirty-year bond at all is a fair question, and the honest answer is that most do not buy one directly. They take the exposure through a fund. The two are not the same instrument, and the difference that matters most is not the one usually cited. A bond's interest-rate sensitivity falls as it ages: buy a thirty-year today and in ten years you own a twenty-year, a shorter instrument carrying correspondingly less duration, whether you ever intend to hold it to maturity. A long-dated bond fund rolls its holdings to stay within its maturity band, so its duration does not run down at all. The bondholder's risk shrinks with time automatically; the fund holder's does not. Part 3 takes this up properly.

Every point of the index's gain came from earnings. There was no re-rating.

The three are not independent and checking that they tie is worth a moment. An index level is simply its earnings multiplied by the multiple paid for them, so the index's gain must equal the gain in earnings multiplied by the change in the multiple:

• Forward earnings rose from 222.36 to 404.27, a factor of 1.818
• The multiple fell from 21.43x to 18.93x, a factor of 0.883
• 1.818 × 0.883 = 1.606, against an actual index ratio of 7,651.54 ÷ 4,766.18 = 1.605

The residual difference is rounding in the published figures; on the unrounded series the two tie exactly. Earnings did all the work and then some, and the multiple gave a little of it back.

One clarification on the earnings line, since it bears on how much weight it can carry. Both figures are forward estimates — what the market expected over the twelve months ahead, at each date. The 81.8% is therefore the rise in the level of expected earnings across five years, not a measurement of whether each year's forecast was met. Over a window this long the two are closely related, because an estimate that is persistently missed gets marked down rather than compounding for five years. But they are not the same quantity, and I use the forward series because it is the one that stays internally consistent with the multiple and the index throughout.

This disposes of the simplest bear case. Whatever has happened since 2021, it is not multiple expansion. An investor who bought at the end of 2021 has been paid entirely by the earnings leg and has absorbed a de-rating along the way.

It also disposes of the simplest bull case, and that matters more. A market whose entire five-year return has come from earnings growth rather than from re-rating holds no valuation cushion in reserve. There is no compressed multiple waiting to re-expand. The engine is earnings, and only earnings.

Where the four hundred basis points went

The framework that connects an equity market's earnings yield to the risk-free rate:

Forward earnings yield = real risk-free rate + equity risk premium − expected growth

The intuition is simpler than the notation. The earnings yield is what you are paid today per dollar invested. If the real risk-free rate rises, you need paying more today, so the earnings yield should rise and the multiple falls. If investors grow more willing to bear equity risk, you accept being paid less today. And if earnings are expected to grow faster, you also accept less today, because more is coming later.

Rearranged:

equity risk premium − expected growth = forward earnings yield − real yield

The left side cannot be observed. The right side can. Prices give the earnings yield, the TIPS market gives the real yield, and the difference between them is a single number containing both the premium investors demand and the growth they expect, inseparably combined. Call it the spread.

Three hundred and forty of a four hundred and one basis point shock went into the spread. Eighty-five per cent of the rise in real yields was absorbed there rather than in the price.

The market's apparent indifference to a generational rise in real rates was never indifference. It was absorption.

The objection this framework usually attracts, and why it does not land here

The construction resembles the Fed model — comparing the equity market's earnings yield to the Treasury yield and calling equities cheap whenever the first exceeds the second. That comparison has a well-documented weakness, and it is worth setting out, because it is precisely the weakness the real-yield version does not share.

Its flaw is money illusion. The earnings yield is a ratio of two quantities measured in the same currency at the same moment: dollars of earnings over dollars of price. If the price level rises ten per cent, a company's revenues, costs and earnings rise with it, and so does the price of a claim on that unchanged real business. The same factor sits in the numerator and the denominator and cancels out. A bond behaves differently for one reason only — its coupon is contractually fixed in dollars and cannot scale, so the yield must move instead. Compare the two and the answer tracks inflation rather than value: equities look expensive when inflation is high, cheap when it is low.

I am not claiming inflation has no bearing on equity valuations; it plainly does. Margins compress when input costs outrun output prices, central banks answer inflation with higher real policy rates, and uncertainty about future inflation raises the premium investors demand. The claim is only this: the earnings yield does not rise one-for-one with expected inflation by construction, whereas a nominal bond yield does. Inflation reaches equities through economics; it reaches a nominal yield through arithmetic, before any economics happens. Comparing the earnings yield to a real yield keeps the two straight — which is not merely a theoretical nicety but something visible in the curve data above, where the long-end selloff contained no inflation component at all.

Three caveats survive the correction, and I would rather state them than hide them.

• The earnings yield is only roughly real — companies pass inflation through to prices and profits imperfectly, and with a lag.
• Forward estimates start high — analysts tend to open a year optimistic and revise downwards as it progresses.
• The spread holds two things at once — both the premium investors require and the growth they expect, with no way to separate them from price data. That is why I call it a spread throughout rather than an equity risk premium, and why Part 2 opens on precisely what it cannot tell you.

Using the real yield fixes the biggest problem with this comparison. It does not fix them all.

What comes next

Compensation for owning equities over a guaranteed real return absorbed almost the entire rate shock of the last five years. Part 2 asks what that compression has meant historically — what the policy cycle did to it, why September broke a pattern that had held for thirteen months, and what happened the last time investors were paid this little for owning shares instead of bonds.

Notes on the data

Source. All figures are from Bloomberg. The historical series runs monthly from January 2000 to September 2026, 319 observations.

Two reference dates. Yield levels and the curve decomposition are the close of business on 9 October 2026. All equity analysis — earnings yields, multiples, spreads, index levels and every historical comparison — uses month-end data to 30 September 2026, which is where the consistent series ends. No calculation in this article combines figures from the two dates.

Definitions.

• Forward earnings yield — the inverse of the forward price-to-earnings multiple.
• Forward earnings per share — derived as the index level divided by the forward multiple.
• The spread — the forward earnings yield less the ten-year real yield on Treasury Inflation-Protected Securities.
• Thirty-year breakeven — derived as the nominal yield less the real yield.

Returns. Index returns are price returns and exclude dividends.

Medians. Regime comparisons use medians throughout, for consistency and because the series contains genuine extremes. Means differ immaterially: across the full sample the forward multiple averages 17.0x against a median of 16.7x, and the real yield 1.25% against a median of 1.37%.

Views expressed are those of the author and do not constitute investment advice.