Investors historically earned, or expect to earn, an extra return for owning stocks instead of relatively safer fixed-income assets such as government bills or bonds. Known as the equity risk premium (ERP), it compensates them for the uncertainty of equity ownership, including the possibility that share prices can fall sharply, dividends can change and the return ultimately received is not contractual.
In practice, the ERP is also a valuation tool. It helps investors, analysts and companies form a required return on equity and estimate a cost of equity. The essential question is straightforward: after allowing for what a comparatively low-risk bond offers, how much additional return does the stock market appear to offer for taking equity risk? The CFA Institute describes the premium as the historical or expected excess return of equities over fixed-income assets.
Equity risk premium: the excess return investors demand from stocks
“Premium” can sound like an observable interest rate, but the ERP is not a single rate printed on a market screen. It is an estimate. That distinction matters because the estimate can be backward-looking, based on returns that have already occurred, or forward-looking, based on prices and expectations today.
A historical ERP asks how much stocks returned above bonds or bills over a selected past period. It can offer perspective on the long-run reward investors have received for bearing equity risk. But its result depends on choices including the start and end dates, the equity market used, the fixed-income benchmark and whether returns are measured arithmetically or geometrically. Past returns are evidence, not a contractual promise about the future.
An expected ERP instead asks what return investors require or appear to be pricing into equities now. This is the version commonly used in a required-return framework. In simplified form:
Required return on equity = risk-free rate + equity risk premium
The equation is a framework rather than a guarantee of performance. A company or index may face risks beyond those reflected in a broad market measure, and different valuation methods may make further adjustments. Still, it captures the basic economic logic: investors generally require a higher expected return from a residual claim on a business than from a benchmark government bond.
Why a low premium can make stocks expensive relative to bonds
Stocks can look expensive relative to bonds when the expected extra return for owning them becomes small. One practical forward-looking comparison starts with the S&P 500’s forward earnings yield and subtracts the expected real yield on a 10-year Treasury. The Federal Reserve Board has used that spread as a practical measure of expected equity risk premium and relative stock-bond valuation.
A narrowing spread means investors are receiving less expected compensation, by this measure, for holding equities rather than long-term Treasuries. That can be consistent with richer equity valuations, greater risk appetite, or both. It does not establish that a market decline is imminent. Rather, it says the valuation cushion relative to the selected bond benchmark has become thinner.
The comparison is relative, not absolute. A stock market can have a positive expected return while still appearing less attractive against bonds than it did earlier. Conversely, bond yields can fall enough to widen the spread even if share prices have not declined. Looking only at an index level or only at a Treasury yield misses the relationship between the two.
There is also an important difference between an earnings yield and a bond yield. A Treasury yield is tied to specified payments and repayment terms, subject to interest-rate and inflation effects and, for non-government issuers, credit risk. Equity earnings are neither a coupon nor cash automatically paid to shareholders. They are a starting point for estimating the cash flows a business may generate over time.
From stock prices to earnings yield to the stock-bond spread
The mechanics are easier to see by reversing the familiar price-to-earnings ratio. Earnings yield is earnings divided by price. If an index trades at 20 times expected earnings, its forward earnings yield is 1 divided by 20, or 5%. The reciprocal relationship is useful because it puts an equity valuation measure into percentage terms that can be compared with a bond yield.
A basic sequence is:
- Obtain a forward price-to-earnings ratio for the equity index.
- Convert it to a forward earnings yield by dividing 1 by the P/E ratio.
- Select a long-term Treasury benchmark and determine whether the comparison uses a nominal or expected real yield.
- Subtract the Treasury yield from the earnings yield to obtain a simple stock-bond spread.
For illustration only, suppose an index has a forward P/E of 25. Its forward earnings yield is 4%. If the relevant expected real 10-year Treasury yield is 2%, the simple spread is 2 percentage points. If the index price rises while expected earnings do not, the P/E rises, the earnings yield falls, and the spread narrows. If the Treasury yield rises with the earnings yield unchanged, the spread narrows as well.
That arithmetic makes clear why a shrinking spread does not have one cause. It may reflect higher stock prices, lower expected earnings, higher bond yields, or a combination. Interpretation requires checking which inputs moved and whether the earnings forecasts and Treasury-yield measure are comparable.
The 10-year U.S. Treasury is widely used because it is a long-term government-rate benchmark and has readily available history. The Federal Reserve Bank of St. Louis’ FRED series provides the 10-year Treasury constant-maturity yield at daily, weekly, monthly and annual frequencies. A nominal 10-year yield is not interchangeable with an expected real yield, however. A comparison should name the yield concept being used.
Implied ERP estimates reverse-engineer the market’s assumptions
The simple earnings-yield spread is a useful proxy, not a complete implied ERP model. ERP cannot be observed directly, so analysts estimate it through historical returns, investor surveys or valuation models. Each approach addresses a related but distinct question.
An implied ERP model begins with the current level of an equity index and assumptions about future cash flows. It then solves for the discount rate that makes the present value of those expected cash flows equal to today’s market price. Subtracting the relevant risk-free rate from that expected return produces the implied premium. Aswath Damodaran of NYU Stern describes this approach as solving for the premium consistent with the current index level and expected future equity cash flows.
This is sometimes called reverse-engineering market expectations. Rather than starting with a preferred premium and calculating a fair value, the analyst starts with the observed market value and asks what return assumption would justify it. The answer is conditional on the assumptions used for earnings, cash distributions, growth and the risk-free rate.
Damodaran’s U.S. estimates illustrate the broader mechanics: an expected S&P 500 return is compared with a dollar risk-free rate, with adjustments involving the U.S. default spread and mature-market risk premium. Those inputs show why an implied ERP should not be treated as a universal market fact. It is a model output built from explicit choices.
What the premium can and cannot tell an investor
ERP is useful when a decision requires a disciplined hurdle rate. An analyst valuing future business cash flows can use an equity risk premium as one component of the discount rate. A market observer can use a stock-bond spread to frame whether equities offer relatively more or less expected reward than a bond benchmark. Corporate-finance estimates of the cost of equity also commonly rely on the concept.
It is not a timing signal with a fixed trigger. A low implied premium or narrow earnings-yield spread may persist, widen because Treasury yields fall, or change because earnings expectations are revised. It does not say that every company is expensive, that every bond is attractive, or that an investor’s appropriate allocation can be inferred from a single number.
Most importantly, a smaller ERP does not make equities as safe as bonds. Stocks generally have greater volatility and greater potential for loss, while bonds typically provide more predictable contractual payments, though bondholders still face risks including issuer default and changing interest rates. The U.S. Securities and Exchange Commission’s Investor.gov distinguishes those characteristics of stocks and bonds.
The premium concerns expected compensation, not a conversion of one asset’s risk into another’s. An investor can conclude that the compensation available for equity risk appears slim and still recognize that the underlying risks of stocks and bonds remain fundamentally different.
Federal Reserve chart showing the spread between the S&P 500 forward earnings-to-price ratio and the expected 10-year real Treasury yield, a practical measure of expected equity risk premium and relative stock-versus-bond valuation. — Source: Federal Reserve Board
Why different ERP estimates can disagree
Disagreement among ERP estimates is normal because the methods do not use the same evidence or answer precisely the same question. A historical estimate summarizes a chosen period. A survey records respondents’ expectations. An implied estimate translates current prices and forecast cash flows into an expected return. None is automatically the definitive number.
Even two implied models can diverge materially. One may use different forecasts for aggregate earnings or cash returned to shareholders. Another may make different assumptions about long-run growth, the appropriate risk-free rate, default spreads or the mature-market premium. Small changes to a long-term assumption can alter the rate that equates present value with a market price.
The treatment of inflation is another source of confusion. A forward earnings yield may be compared with a nominal Treasury yield in one presentation and an expected real Treasury yield in another. Those measures serve different purposes. Mixing them without explanation can make a spread look more precise than it is.
For readers assessing a published ERP figure, the most useful questions are practical: Is it historical, surveyed or implied? What equity market and bond benchmark does it use? Are earnings and cash-flow assumptions forward-looking? Is the rate nominal or real? Clear answers turn an opaque percentage into a usable, appropriately limited valuation input.
Frequently Asked Questions
What is a good equity risk premium?
There is no single permanent “good” figure. The appropriate estimate depends on the method, market, bond benchmark, time horizon and assumptions about future cash flows and returns.
Is the equity risk premium the same as the S&P 500 earnings yield?
No. The earnings yield measures equity valuation. A simple stock-bond spread subtracts a Treasury yield from the earnings yield, whereas a fuller implied ERP model uses expected cash flows and current market prices to solve for an expected return.
Why does a higher Treasury yield pressure the stock-bond spread?
If the forward earnings yield is unchanged, a higher Treasury yield leaves less excess yield from equities over bonds. That narrows the comparison used as a proxy for expected equity compensation.
Can a negative stock-bond spread occur?
Yes. It can occur when the selected Treasury yield exceeds the index’s forward earnings yield. The result should be interpreted carefully because the two measures represent different claims and because the chosen rate and earnings forecasts matter.
Does a low ERP mean investors should sell stocks?
Not by itself. It is a valuation and required-return input, not a stand-alone instruction. Investment decisions also depend on objectives, time horizon, diversification, risk tolerance and the assumptions behind the estimate.
