The 10-Year, the Equity Risk Premium and a World Drowning in Debt

All of modern finance rests on a single idea: that there exists a riskless asset. A minimum, certain, guaranteed return that serves as the floor upon which everything else is priced. The 10-year US Treasury. The famous 10Y. The “risk-free rate.” The number that shows up, openly or quietly, in every DCF, every CAPM model, every investment-banking spreadsheet on the planet.

And here’s the 2026 problem: that floor is moving. Not because people have stopped trusting the dollar, nor because the United States is about to default tomorrow — nothing that dramatic. It’s moving because the debt arithmetic in the developed world has, for the first time since the 1940s, become genuinely uncomfortable. And that changes everything: what the S&P is worth, what your house is worth, what gold is worth, and most of all, what it costs to finance literally anything.

First, the boring bit: what the risk-free rate actually is

The concept is simple. If I can lend money to the US government and get it back with certainty — because the US government prints its own currency and has never, ever defaulted — then the yield on that bond is the minimum that any other investment must beat. If a company wants my money, it has to offer more than the 10Y. If a house wants my capital, it has to net out above the 10Y. If a stock wants my conviction, it has to promise returns above the 10Y, with enough excess to compensate for the risk of dropping 40%.

The 10Y is, literally, the baseline. The zero on top of which every risk premium is added. Which is why, when it rises, the floor under everything else rises too — and when it rises sharply, risk assets have to re-price downward to stay competitive. This isn’t opinion. It’s arithmetic.

Where it sits today: the new 4–5% normal

For thirteen years — from 2009 to 2022 — the 10Y lived in an induced coma between roughly 1.5% and 3%. The Fed pinned it down with QE, hoovering up bonds by the truckload. That was the era of free money. Anything yielded more than the bond, so anything went up. Property, equities, crypto, pixelated-monkey NFTs — everything. The tide was high, and every boat floated.

Since 2023, the 10Y has settled into a very different band: between 4% and 5%. May 2026 finds us near 4.3–4.5% — down a touch from the 2024 highs, but well above the “normal” that the entire millennial-and-Gen-Z investor generation has known their adult lives. To someone who started investing in 2010, this 10Y feels high. To someone who started in 1985, it feels normal — even low. It’s a reminder that “normal” depends a lot on when you opened your eyes.

And here’s the interesting bit: this 4–5% level is probably not transitory. It’s not “the Fed will cut and we’ll all go back to 1.5%.” There’s a structural reason the 10Y is staying high, and it’s called debt.

The elephant in the room: the debt mountain

Some numbers, plain, so you can see what we’re actually looking at. Public debt-to-GDP across the major developed economies, 2026 (approximate):

• Japan: ~250%
• Italy: ~140%
• United States: ~125% (and rising)
• France: ~115%
• United Kingdom: ~100%
• Spain: ~105%
• Germany: ~65% (the disciplined one in class, until recently)

For perspective, in the year 2000 the United States was at 55% debt-to-GDP. The UK was at 35%. So in twenty-five years we’ve doubled, tripled, or more, the size of public debt relative to GDP. And we did it while rates were rock-bottom — which masked the cost.

But now rates aren’t rock-bottom. And the bill has started to arrive. In 2026, US Treasury interest expense exceeds the defence budget. Read that again. The United States pays more to refinance its debt than to run the largest military on Earth. In the UK, debt interest is the third largest budget line, ahead of education. In Italy, it’s one of the largest line items, on a par with healthcare. This isn’t a technical detail — it’s the fiscal reality of the developed world, right now.

How does this feed into the 10Y? Very directly: the US Treasury has to issue trillions of dollars every year to roll old debt and fund new deficits (around 6% of GDP). That gigantic bond supply pushes prices down and yields up. When there’s plenty of paper and limited appetite, you pay up to place it. It’s the oldest market on Earth and it works like any other.

“Risk-free” isn’t quite risk-free anymore

Here’s the intellectual heresy of the moment: the 10-year Treasury is still nominally “risk-free” — the government will give you your dollars back at maturity, no debate there. But there are three risks the original CAPM model simply ignored, and that in 2026 you really can’t ignore anymore.

First, duration risk. Buy a 10Y at 4.3% and watch yields rise to 5.5% — that bond drops 8-10%. Your “safe” money lost 10% in mark-to-market. You get it back if you hold to maturity, fine, but you lose opportunity and purchasing power for a decade. That happened in 2022 — the worst year for bonds in forty. “Risk-free” on the coupon is not “risk-free” on the price.

Second, inflation risk. Bonds pay in nominal dollars. If inflation eats that 4.3% — and between 2021 and 2023 we had years with CPI running 7-9% — your real return was deeply negative. “You’ll get your dollars back” and “you’ll get your purchasing power back” are two very different statements.

And third, the newest and most uncomfortable: financial repression risk. When a government has so much debt it can’t afford to pay high rates for too long, it starts finding creative ways to keep yields artificially low. Capital controls. Forcing pension funds and banks to hold domestic bonds. Doing QE “only when needed” (which conveniently turns out to be most of the time). The bond still pays, but you cover the difference with lower real returns. Japan has been doing this for thirty years. Europe for fifteen. The US is starting to flirt with it.

The equity risk premium: the oxygen of the stock market

Enter the other star of this piece: the Equity Risk Premium (ERP). In Spanish, prima de riesgo de la renta variable. Conceptually, it’s the heart of why anyone buys stocks at all.

The ERP is, plainly, the extra return you expect for owning stocks instead of the risk-free asset. The basic formula: ERP = expected stock return − risk-free rate. If the 10Y is at 4.3% and you think the S&P will deliver 8% nominal long-term, your ERP is 3.7%. That’s the “tip” the market pays you for stomaching volatility, recessions, and the possibility of spending ten years underwater if you enter at the wrong time.

Historically, the US ERP has run between 4% and 6%, depending on the period and the methodology (Damodaran has been publishing it for decades). In the 1970s, with sky-high rates, the ERP briefly went negative — bonds yielded so much that stocks didn’t compensate. In 2000, at the dot-com peak, the ERP also dropped to very low levels. And in 2009, at the bottom of the crisis, it was huge — the market was paying enormously for anyone willing to take risk. The ERP isn’t a constant of nature. It rises and falls with the cycle.

Where the ERP sits in May 2026: frankly worrying

Let’s do the maths by hand, the way I prefer. The S&P 500, in May 2026, trades at a forward P/E of roughly 22–23x. The inverse of that is the earnings yield: 1 / 22.5 ≈ 4.4%. So S&P stocks are “yielding” about 4.4% in earnings-on-price terms.

The 10Y sits at 4.3%. The “napkin” ERP — earnings yield minus 10Y — is 0.1%. Ten basis points. That means: literally nothing. The S&P, by this rough but useful measure, is paying you essentially zero extra to take equity risk.

Caveat: that’s the crudest version of the ERP. If you use a more sophisticated model with expected growth, buybacks and so on, you can get a number between 1% and 2.5%. Damodaran has been publishing implied-ERP estimates for years; in 2026 it sits around 2–2.5%, depending on the month. Either way — simple or sophisticated — today’s ERP is well below the long-run average of 4–5%.

Translation: the market is asking you to take equity risk for a tip. Not a small tip — a historically miserable tip. That fact alone is the single most important thing for understanding the next few years.

So why is anyone still buying stocks?

Good question. Three answers, none of them entirely reassuring.

First: habit. An entire generation of managers and retail investors grew up in the ZIRP era. “Buy stocks, always, on any dip” was the lesson of fourteen years. That habit doesn’t break overnight. A lot of money still flows into the S&P out of inertia, not analysis.

Second: index concentration. Today’s S&P 500 is really the S&P 7 — Apple, Microsoft, Nvidia, Alphabet, Amazon, Meta and a couple of others now make up more than 35% of the index. These companies have brutal margins, net-cash balance sheets, and real growth. Looked at individually, their “quality” justifies high multiples. The problem is that the other 493 names are paying for the party that those few are throwing.

Third: no one has a comfortable alternative. Bonds yield the same as the S&P’s earnings yield, but without growth — just the coupon. Real estate is expensive and expensive to finance. Gold has been climbing for five years (you caught that one, I hope). Cash at 4–5% is attractive but still loses to inflation. When everything is expensive, people stay with what they know. And what they know is “S&P going up.”

What this means for the next few years

If we accept that (a) the 10Y stays structurally between 4% and 5% because of debt, (b) the ERP is near zero, and (c) developed-world governments cannot — politically — cut spending or raise taxes meaningfully, then there are a handful of mathematical consequences, not opinions.

One: future S&P returns will be lower than the last fifteen years. Not because companies are worse, but because you’re entering at high multiples with high rates. Goldman, JP Morgan, Vanguard and almost every serious projection model now have the S&P at 3–6% nominal annual returns over the next decade. A long way from the 13% we’ve grown used to since 2009.

Two: long bonds, for the first time in a decade, are attractive. Buy a 10Y at 4.3%, see yields drop to 3.5% (plausible in a recession), and you take the coupon plus a 6–7% capital gain. If yields rise to 5%, you take a knock but keep the coupon. Reasonable asymmetry. For the first time since 2009, bonds are a serious asset class again, not just “the place you park money while waiting for opportunities.”

Three: gold and real assets are insurance against financial repression. If governments end up taking the easiest path — moderate persistent inflation to erode the debt — then gold, well-financed real estate, and cheap commodities appreciate in real terms. This is the move emerging-market central banks have been making since 2022, buying gold at record pace. They know something the traditional 60/40 portfolio still doesn’t want to hear.

Four: quality matters more than ever. In a zero-ERP environment, the market average is expensive and poorly paid. But there are companies generating 25–30% returns on capital, with clean balance sheets, low debt, real free cash flow, at reasonable multiples. Those still work. The era of free beta is over; the era of stock-picking is back.

Why contrarian value is the only game left in town

Everything above points to a single uncomfortable conclusion: when the market average pays no premium for risk, the only rational way to expect a real return is to not be where the market average is. Full stop. That, without dressing it up, is contrarian value investing.

And I don’t mean “value” in its showroom version — buying an IWD or VTV ETF and sleeping easy. That’s diluted value, value for the masses, value that’s already priced in. Real contrarian value is another animal: paying 60 cents for a dollar of earnings the herd has decided isn’t worth a dollar. It’s owning European banks at 6x earnings while everyone talks about AI. It’s picking up an oil major at 7x cash flow because “oil is finished.” It’s buying Japanese small caps trading below net cash. Stocks you’re embarrassed to mention at dinner.

The maths forces this, it’s not an ideological stance. Indexing in a zero-ERP world gives you 4-5% nominal at 16% volatility. Which is: bond returns with equity risk. That’s not investing — that’s renting anxiety. From 2009 to 2021, the masses did well indexing because the starting valuation was reasonable and rates cooperated. From 2026, that same strategy starts from the worst entry point in forty years. The arithmetic doesn’t forgive.

We said it earlier: 7 stocks are 35% of the S&P. Which means the other 493, by pure maths, are under-allocated. Capital fled them and piled into the Magnificent Seven, and where capital is scarce, prices drop further than fundamentals justify. There — in those boring, forgotten corners with no influencer pumping them — is where the contrarian finds work. Not by hunting for the next Nvidia, but by buying the next Coca-Cola when no one wants it.

Howard Marks puts it best: “you can’t do the same thing as everyone else and expect different results.” In a cheap market, everyone can win. In an expensive one, by definition, someone has to lose — and it’s usually whoever entered last, buying what was already expensive. The contrarian accepts feeling uncomfortable for years in exchange for not being that last buyer. Buffett refused to touch tech in 1999 and people called him senile, washed up, past it. Three years later, after the Nasdaq imploded, they called him a genius. That’s the entire movie.

The ugly part: you’re going to look stupid for a long time. Friends will show you their Nvidia gains, tell you value is dead, that AI changes everything, that this time it’s different. You’ll grit your teeth and hold. And one day — usually a random Tuesday, no warning — the rotation arrives. Ten years of patience recover fifteen years of returns in six months. It happened in 1973, in 2000, and it will happen again. The question isn’t if — it’s when, and whether you’ll have the stomach to keep buying the cheap stuff while everyone around you buys the expensive stuff.

Which is why, in this regime, contrarian value isn’t one of several valid strategies. It is, mathematically, the only one where the calculation gives you a real expected edge. The pure indexer is praying the next decade looks like the last one. The pure growth investor is paying prices that already discount ten years of success. The contrarian value buyer pays below the present value of the cash flows and lets time and mean reversion do the rest. It’s boring. It’s slow. It’s lonely. And it’s the only approach the arithmetic validates in May 2026.

An uncomfortable note on Europe

If you think this is only an American problem, look at Europe honestly. Italy, France, Spain, the UK — all with high debt, low growth, and rates that can’t drop back to zero. The ECB doesn’t have the United States’ favourite tool: a global reserve currency the rest of the world still wants to hold. When the next crisis arrives, Europe will have less room, not more. The Italian 10-year, today around 4.5%, may look cheap in five years.

And Japan is the strangest case study of all: the country that reached this point first and, despite carrying 250% debt-to-GDP, has somehow held it together for thirty years with negative or near-zero rates. How? Because 90% of that debt is held by Japanese savers. Financial repression works if your own country buys its own debt and nobody complains. It’s an option, not a solution.

Practical summary, no fluff

1. The US 10Y is the foundation of everything. Watch it. If it sits at 4.3%, everything else gets priced off that. Its level caps your future returns.
2. Today’s ERP is the most compressed in a generation. Expensive equities and bonds paying almost the same yield means mediocre returns ahead. Accept less, or take more risk consciously.
3. The debt isn’t going anywhere. Developed economies will spend more on interest than on defence, education or health. The political response will be moderate inflation or financial repression. No government is going to make serious cuts.
4. Diversify properly, not for show. Quality stocks, long bonds when yields spike, gold as insurance, some cash at 4–5%. And above all, realistic expectations: we are not going to replay 2010–2021.

The world has entered a new monetary regime, and most portfolios are still built for the old one. That, over time, tends to be a problem. The good news is that the basic concepts — risk-free rate, ERP, discounted cash flows — still work. You just have to apply them again, knowing that the foundation — that 10Y you took for granted — is no longer free.

As always: the FOMC calendar in one hand, the 10Y yield in the other, and a little patience. A zero ERP doesn’t last forever. But when it decompresses, it does it fast — and usually with very little warning.