Jack Treynor's 1965 Harvard Business Review article proposed judging investment funds by more than their rate of return. As he later described the paper, what question was it trying to answer?

How to evaluate a fund's performance net of its costs, separating what its manager earned from what its investors paid in fees and expenses.

How to evaluate a fund's performance with the market effect subtracted, separating what its manager contributed from what the market did to every fund.

How to evaluate a fund's performance against its peers, separating its standing within its category from the category's own average return.

How to evaluate a fund's performance over a full cycle, separating the manager's lasting results from any single unusually good or bad year.

A fund with a beta of one point four returned exactly the Treasury bill rate over a year in which the market rose strongly. What is the fund's Treynor ratio?

Negative, because a fund with a beta above one fell short of what the rising market paid for that much risk, and the ratio records the shortfall.

About zero point seven one, being one divided by the beta of one point four, because the fund's return and the risk-free rate were equal.

Zero, because the numerator is the return above the risk-free rate, and the fund earned nothing above it for the market risk it carried.

The Treasury bill rate divided by one point four, because the reward the ratio measures is the whole return the fund actually earned.

Over the same year, with a risk-free rate of four percent, Fund C returned sixteen percent with a beta of one point five and Fund D returned eleven percent with a beta of zero point seven five. A client asks which manager did the better job once systematic risk is taken into account. What is the answer?

Fund D, at about nine point three against eight: it earned less in total but more excess return for each unit of beta it accepted.

Fund C, at eighteen against about five point three: its twelve points of excess return multiplied by its higher beta outweigh Fund D's figure.

Fund C, at twelve points of excess return against seven: the measure ranks managers by the return they earned above the risk-free rate.

Neither, at about eight apiece: each fund's excess return rose in proportion to its beta, so both earned the same reward per unit of risk.

A portfolio with a positive beta plots above the Security Market Line, having earned more than the Capital Asset Pricing Model predicts for its beta. How do its Treynor ratio and its Jensen's alpha stand?

Its Treynor ratio is above the equity risk premium, but its alpha may be negative, since alpha is measured against the market index rather than the line.

Its Treynor ratio is above the equity risk premium, and its alpha is positive, both recording return beyond what the model predicts for its beta.

Its alpha is positive, but its Treynor ratio may be below the equity risk premium, since a beta above one dilutes the ratio below the line's slope.

Its Treynor ratio is below the equity risk premium and its alpha is positive, since it earned its extra return by carrying less systematic risk.

Over a year the risk-free rate was two percent and the market returned ten percent. Fund P had a beta of zero point five and Fund Q a beta of two, and each earned one percentage point of Jensen's alpha. How do their Treynor ratios compare?

They are equal, because equal alphas mean equal risk-adjusted performance, and the Treynor ratio and alpha both adjust for the same systematic risk alone.

Fund Q's is the higher, because its larger beta earned it the far larger return above the risk-free rate, and that excess is all the ratio's numerator holds.

Fund P's is the higher, because the same one point of alpha is spread over fewer units of systematic risk and so adds more to each unit's reward.

Fund Q's is the higher, because alpha earned at a high beta was earned on a larger share of the market's movement and so counts for more.

Over a year the market returned two percent while Treasury bills returned five percent. Three portfolios, with betas of zero point six, one and one point six, each earned exactly the return the Capital Asset Pricing Model called for at its beta. What are their Treynor ratios?

Minus five, minus three and about minus one point nine, the market's three-point shortfall being divided by each portfolio's own beta.

Minus one point eight, minus three and minus four point eight, the negative figure growing in step with each portfolio's beta as its exposure rises.

Zero for each, since a portfolio that earned exactly what its beta called for has no reward or shortfall for the ratio to record.

Minus three for each, a negative figure for all three although none of them fell short of the return that its systematic risk called for.

In 2005 Georges Hübner proposed a Generalized Treynor Ratio that carries the measure from one index to several. How is that generalised ratio defined?

As a portfolio's excess return per unit of the simple average of its betas on each index, normalised by the market's own beta of one.

As a portfolio's abnormal return per unit of the standard deviation of its returns, normalised by the same risk figure for the benchmark.

As a portfolio's excess return per unit of the highest of its betas on any one index, normalised by the same risk figure for the benchmark.

Two equity funds each hold several hundred stocks spread across every sector, so that almost none of their company-specific risk remains. Over the same period Fund E ranks above Fund F on the Treynor ratio. What would the Sharpe ratio be expected to show?

The opposite ranking, since the Sharpe ratio penalises the market risk that the Treynor ratio rewards, so the two measures pull apart.

Nearly the same ranking, since once company-specific risk is diversified away a fund's total risk is almost entirely its market risk.

A ranking set by the funds' returns alone, since with company-specific risk gone their standard deviations become identical to each other.

No usable ranking, since the Sharpe ratio is designed for concentrated portfolios and gives unreliable figures for diversified ones.

A fund built to rise when the stock market falls has a beta of minus zero point five. Over a year it returned two percentage points less than Treasury bills. What does its Treynor ratio show?

Minus four, a negative figure that records the shortfall, the two points being divided by the size of the beta whatever its sign.

Plus one, a positive figure because the shortfall is multiplied by the negative beta, which reverses its sign and halves its size.

Plus four, a positive figure that makes the fund look attractive although it returned less than the risk-free rate over the year.

Minus one, a negative figure because the shortfall is scaled by the size of the beta, which halves the reading in either direction.

A manager keeps a portfolio's risk around a fixed target on average but raises or lowers it from month to month according to forecasts of the market. Jensen showed that a single regression over the period still measures the success of that activity. What does the one beta it estimates then represent?

The risk of the market index itself, a beta of exactly one, since the deviations above and below the target cancel against the market.

The highest level of risk the manager reached, since the months of greatest exposure dominate a regression run on the market's returns.

The level of risk in the final months of the period, since the regression weights recent returns more heavily than the older ones.

The average level of risk the manager held around the target, not the risk the portfolio carried in any particular month of the period.

Jensen examined what a single regression of a portfolio's returns on the market's gives when the manager can genuinely forecast general market movements. What did he show about the estimates?

The estimated beta is biased downward and the estimated alpha upward, so the measure tends to overstate the effect of the forecasting ability.

The estimated beta is biased upward and the estimated alpha downward, so the measure tends to understate the effect of the forecasting ability.

The estimated beta is biased downward and the estimated alpha downward too, so the measure tends to hide the forecasting ability altogether.

The estimated beta is biased upward and the estimated alpha upward too, so the measure tends to count the forecasting ability twice over.

In 1966 Treynor and Mazuy extended the usual regression of a fund's excess return on the market's by adding the market's excess return squared as a second explanatory term. What does a positive coefficient on that squared term indicate?

That the fund's beta falls as the market goes up, read as defensive positioning, while the intercept still measures the return from selecting securities.

That the fund's beta rises as the market goes up, read as forecasting skill, while the intercept still measures the return from selecting securities.

That the fund's beta rises as the market goes up, read as borrowing, while the intercept then measures the cost of the money borrowed to fund it.

That the fund's returns are more volatile than the market's, read as concentration, while the intercept then measures its company-specific risk.

In the test proposed by Roy Henriksson and Robert Merton in 1981, a manager is treated as switching a portfolio between two betas. What decides which of the two applies?

Whether the fund's own return exceeds its benchmark's, a high beta applying when it does and a lower beta otherwise.

Whether the market's volatility exceeds its long-run average, a low beta applying when it does and a higher beta otherwise.

Whether the market's return exceeds the risk-free rate, a high beta applying when it does and a lower beta otherwise.

Whether interest rates are rising or falling, a low beta applying while they rise and a higher beta while they fall.

Work by Ingersoll and his co-authors published in 2007, as Georges Hübner summarises it, shows that return-based regressions used to detect a manager's market-timing skill can be gamed. How?

A manager who borrows to raise beta by the same amount in every month can alter the intercept and the timing coefficient in exact proportion.

A manager with access to a complete derivatives market can alter the regression intercept without changing the timing coefficient to a proportional extent.

A manager who rebalances only at each quarter end can alter the measured timing coefficient without making any change to the portfolio's holdings.

A manager with access to a complete derivatives market can alter the measured timing coefficient without changing the intercept to a proportional extent.

An investor comparing two funds sees that Fund G, with a beta of zero point six, has a higher Treynor ratio than Fund H, with a beta of one point one. Fund G's beta was measured against a highly volatile index and Fund H's against a much less volatile one. What should the investor understand before choosing Fund G?

Fund G's low beta may owe much to the volatile index behind it, so choosing it for its higher ratio could take him outside his risk parameters.

Fund G's ratio is the more trustworthy of the two, since a beta measured against a more volatile index is generally the more demanding test of a fund's risk.

Fund H's ratio is the one flattered, since measuring against a less volatile index can inflate a fund's beta and so raises the ratio computed from it.

Neither ratio is affected by the index, since each beta already expresses the fund's movement as a share of its own index's movement.

A 1997 study in the Financial Analysts Journal ran the Treynor and Mazuy regression for a sample of domestic equity funds against the Standard and Poor's 500 index, and then an extended version controlling for the funds' holdings outside that index. What did the two versions find about the funds' market-timing ability?

Positive and significant timing ability in the original version, but negative timing ability once the holdings outside the index were controlled for.

Negative timing ability in the original version, but positive and significant timing ability once the holdings outside the index were controlled for.

Negative timing ability in both versions, the control for the holdings outside the index leaving the original result essentially in place.

Positive timing ability in both versions, the control for the holdings outside the index roughly doubling the size of the original result.

How does the stability over time of a portfolio's beta compare with that of the betas of the individual securities it holds?

It is more stable than its holdings' betas, and fixed for as long as the manager leaves the holdings and their weights unchanged.

It is less stable than its holdings' betas, since the estimation errors in each holding's beta accumulate in the portfolio's own figure.

It is more stable than its holdings' betas, though still liable to change as those betas and the covariances among the holdings change.

It is less stable than its holdings' betas, since the covariances among the holdings shift more than any single holding's beta does.

An analyst expects economic conditions over the coming years to differ from those of the past decade, and expects the change already under way to persist. Which beta is likely to be the more useful for forecasting a fund's risk-adjusted performance?

One calculated over the period most like the long-run average, since economic conditions tend to return to that average over time.

One calculated over the full decade, since a longer run of returns gives more observations and so reduces the error in the estimate.

One calculated over the full decade, since a window spanning both the old conditions and the new averages the two of them fairly.

One calculated over a shorter recent period, since it reflects the conditions that are now expected to continue into the years ahead.

A market-neutral portfolio earned two percentage points above the risk-free rate, with an estimated beta of zero point zero five. What is its Treynor ratio, and how far can the figure be relied on?

Forty, a figure that can be relied on, since so small a beta leaves the ratio almost untouched by whatever the market does.

Forty, a figure that a small error in so small a beta moves a long way, since an estimate of zero point one instead would halve it.

Zero point one, a figure that can be relied on, since the two points of excess return are scaled down by so small a beta.

Zero point one, a figure that a small error in so small a beta moves a long way, since an estimate of zero point one would double it.

A fund runs a strategy whose payoffs are deliberately skewed rather than spread evenly about their mean. What limits the use of its Treynor ratio and its Jensen's alpha as guides to its performance?

Both rest on the risk-free rate staying constant through the period, and the instruments that produce skewed payoffs move with that rate.

Both rest on the fund's holdings being priced every day, and the instruments that produce skewed payoffs are often priced only when closed out.

Both rest on the pricing model's mean-variance assumption, which confines them to strategies expected to have normally distributed returns.

Both rest on the fund carrying no borrowed money, and the instruments that produce skewed payoffs create obligations beyond its capital.

A consultant compares two managers of US equity funds. Both Treynor ratios were calculated against the S&P 500, but Manager S's covers the five years ending last December and Manager T's covers the five years ending a decade earlier. Manager T's ratio is the higher. What should the consultant conclude from that?

That Manager T did better, since the ratio already adjusts each fund for its market risk and so puts results from different periods on the same footing.

That Manager S did better, since managers are ranked on their latest risk-adjusted results and the older ratio belongs to a span that no longer applies.

That Manager T's lead should be confirmed with the Sharpe ratio, since a ratio built on total risk is unaffected by the measurement period.

Nothing about which manager did better, since Treynor ratios computed over different measurement periods cannot be meaningfully ranked against each other.

An adviser plans to show a fund's Treynor ratio in an advertisement. The SEC staff's March 2025 FAQ gave the Sharpe ratio, the Sortino ratio and other similar metrics as examples of portfolio characteristics. On whether such a characteristic is performance under Rule 206(4)-1, where does that leave the adviser?

With a determination of its own to make on the facts, since the rule does not define performance and the staff took no position on any characteristic.

With a requirement to treat it as performance, since the staff stated that every characteristic computed from a portfolio's returns counts as performance for the rule's purposes.

With a requirement to treat it as performance, since the rule's own definition of performance names risk-adjusted ratios among the figures it covers.

With a need to obtain the staff's view in advance, since the staff decides case by case which of the characteristics count as performance under the rule.

Suppose a portfolio characteristic shown in an adviser's advertisement, such as a Treynor ratio, is not performance within the meaning of Rule 206(4)-1. How do the rule's performance provisions bear on the way it is presented?

They reach it in part: the gross-and-net requirement still applies to it in full, although the separate requirements on time periods do not.

They do not reach it: its presentation falls outside the scope of the performance provisions in paragraph (d) of the rule altogether.

They reach it through paragraph (d)'s presentation rules, which govern portfolio characteristics in just the same way as performance.

They reach it only where it is shown to retail investors, since paragraph (d) governs characteristics in advertisements to retail investors.

An adviser relies on the SEC staff's March 2025 answer on portfolio characteristics to show a gross Treynor ratio, and in the same advertisement shows the strategy's gross internal rate of return without a net figure. How far does that answer help with the internal rate of return?

Fully, provided the figure is labelled a portfolio characteristic rather than performance wherever it appears in the advertisement.

Fully, since internal rate of return is a characteristic computed from returns and the answer treats it in the same way as the ratio.

Not at all, since the answer's positions do not apply to internal rate of return, however that figure is labelled in the advertisement.

Partly, since the answer's positions reach internal rate of return for private funds but not for separately managed accounts.

Under the SEC staff's March 19, 2025 answer, an adviser may display a gross Treynor ratio without the corresponding net Treynor ratio, and the staff would not recommend enforcement action, if certain conditions are met. Which of the following is one of them?

The ratio is accompanied by the adviser's fee schedule, from which a reader can work out the net figure.

The ratio is accompanied by the same ratio for the strategy's benchmark index, calculated on a gross basis.

The ratio is calculated from a composite of every account managed to the strategy rather than one account.

The ratio is clearly identified as having been calculated without the deduction of fees and expenses.

Treynor and Black, in 1973, showed how to allocate among components each judged by its differential return against a benchmark. If resources are allocated optimally, how should a component with a negative mean differential return be held?

Not at all, left out of the portfolio, since a holding proportional to a negative mean differential return is simply set to zero.

In a negative amount, sold short, the holding being proportional to its mean differential return divided by the variance of that return.

In a negative amount, sold short, the holding being proportional to its mean differential return divided by its standard deviation.

At its benchmark weight, neither added to nor reduced, since a negative mean shows that the analysis has no edge on that component.

Treynor and Black measured each component by the ratio of its mean differential return to the standard deviation of that return. With resources allocated optimally, how does the whole portfolio's figure relate to its components' figures?

The square of the portfolio's ratio equals the average of its components' squared ratios.

The square of the portfolio's ratio equals the square of the sum of its components' ratios.

The square of the portfolio's ratio equals the sum of the squares of its components' ratios.

The square of the portfolio's ratio equals the largest of its components' squared ratios.

Treynor and Black reported a squared figure, the square of a component's mean differential return divided by its standard deviation. William Sharpe, discussing the paper in 1994, preferred the unsquared measure. What did he object to in the squared figure?

It changes with the currency in which the returns are measured, so the same manager's figures cannot be compared across markets.

It ranks components in a different order from the unsquared ratio, even where every one of the mean differential returns is positive.

It cannot be calculated for a component whose mean differential return is negative, so those components drop out of the figure.

Its values are all positive, even where the mean differential return is negative, so it obscures important information about performance.

Treynor Ratio | Financial Regulation Courses

SERIES 65 | FINANCIAL REGULATION COURSES

The Treynor ratio — also called the reward-to-volatility ratio — is a risk-adjusted performance measure that evaluates how much excess return a portfolio or investment strategy generates per unit of systematic risk, calculated by dividing the portfolio's excess return above the risk-free rate by the portfolio's beta — the measure of its sensitivity to broad market movements.

Introduced by Jack Treynor in a 1965 Harvard Business Review article, How to Rate Management of Investment Funds, and grounded in his foundational work on the Capital Asset Pricing Model — work he began in the late 1950s, independently of and roughly contemporaneously with William Sharpe, John Lintner and Jan Mossin, and circulated in manuscript before the formal CAPM publications of Sharpe and Lintner — the Treynor ratio embodies the central insight of modern portfolio theory that only systematic risk deserves compensation in an efficiently diversified portfolio, because unsystematic risk can be eliminated through diversification at no cost and therefore should not command any additional expected return.

By using beta rather than total standard deviation as its risk measure, the Treynor ratio deliberately excludes unsystematic risk from the performance evaluation — rewarding portfolio managers for returns generated per unit of the market risk they cannot avoid while ignoring the diversifiable risk they should have eliminated.

The Treynor ratio falls within the Series 65 examination's coverage of risk-adjusted returns — the NASAA content outline effective June 12, 2023 lists risk-adjusted returns among its performance measures and names alpha, beta and the Sharpe ratio, though not the Treynor ratio itself — and it is one of the three primary risk-adjusted performance measures alongside the Sharpe ratio and Jensen's alpha, the critical point being the distinction between using beta versus standard deviation as the risk denominator and the implications of that choice for when each measure is most appropriate.

What Treynor Set Out to Measure

Treynor's 1965 article, How to Rate Management of Investment Funds, set out a concept of performance evaluation that went beyond the rate of return. In his own later description of the paper, the question was how to evaluate portfolio performance with the market effect subtracted — how to separate what a fund's manager had contributed from what the general market had done to every fund over the same period. That is why the measure that bears his name divides a portfolio's reward by its sensitivity to the market rather than reporting the portfolio's return alone, and why a fund that simply rode a rising market with a high beta, adding nothing of its own, earns no better rating than the market itself.

Deborah Kidd, writing for CFA Institute in 2011, records a further twist in the history: the ratio named after Treynor is not exactly the ratio he intended to develop. His index was built to measure reward to volatility, a term that was later misinterpreted as beta. Georges Hübner, introducing a generalisation of the measure, describes Treynor's original interpretation as the ratio of a portfolio's abnormal excess return — Jensen's alpha — to its systematic risk exposure.

Treynor's work on the Capital Asset Pricing Model itself had a long unpublished life. Richard Roll, writing for the American Finance Association, records that Treynor wrote down the basic ideas of the model in the late 1950s, working independently of and roughly contemporaneously with William Sharpe, John Lintner and Jan Mossin, and that his own paper on the model circulated among specialists for decades before it was published. Michael Jensen's 1968 study of mutual fund performance cited both Treynor's 1965 article and his then unpublished manuscript on the pricing of risky assets.

The Formula — Structure and Components

The Treynor ratio formula is straightforward — a two-component expression that captures excess return per unit of systematic risk.

Treynor ratio equals the portfolio return minus the risk-free rate divided by the portfolio beta.

T equals R sub p minus R sub f divided by beta sub p.

Each component of the formula carries specific meaning that must be understood precisely for examination purposes.

The portfolio return — R sub p — is the total return of the portfolio being evaluated over the measurement period, expressed as a percentage. This is the actual realised return including both income and capital appreciation — the total return as defined in the Total Return entry of this dictionary.

The risk-free rate — R sub f — is the return available on a risk-free investment over the same period — empirically proxied by the yield on short-term United States Treasury bills as covered in the Treasury Bill entry. Subtracting the risk-free rate from the portfolio return produces the excess return — the return above and beyond what the investor could have earned with no risk at all. This excess return is the reward for bearing risk — and the Treynor ratio measures how efficiently the portfolio has converted systematic risk into this excess reward.

The portfolio beta — beta sub p — is the measure of the portfolio's systematic risk — its sensitivity to movements in the broad market portfolio. A portfolio with a beta of one point five moves approximately fifty percent more than the market in both directions — gaining fifteen percent when the market gains ten percent and losing fifteen percent when the market loses ten percent. A portfolio with a beta of zero point eight moves approximately eighty percent as much as the market. Beta is the denominator of the Treynor ratio — the unit of systematic risk against which excess return is measured.

A Worked Example — Computing the Treynor Ratio

A concrete numerical example makes the Treynor ratio calculation immediately clear and provides the template for solving any examination question requiring the computation.

Portfolio A has produced an annual return of fourteen percent. The risk-free rate for the period is three percent. Portfolio A has a beta of one point two. The Treynor ratio equals fourteen minus three divided by one point two — equalling eleven divided by one point two — equalling nine point one seven.

Portfolio B has produced an annual return of twelve percent over the same period. The same risk-free rate of three percent applies. Portfolio B has a beta of zero point eight. The Treynor ratio equals twelve minus three divided by zero point eight — equalling nine divided by zero point eight — equalling eleven point two five.

Despite Portfolio A producing a higher absolute return — fourteen percent versus twelve percent — Portfolio B has a superior Treynor ratio — eleven point two five versus nine point one seven — because Portfolio B achieved its return with substantially less systematic risk. Portfolio A required a beta of one point two — taking on twenty percent more systematic risk than the market — to produce its fourteen percent return. Portfolio B required only a beta of zero point eight — twenty percent less systematic risk than the market — to produce twelve percent. On a per-unit-of-systematic-risk basis Portfolio B is the superior performer — it generated more excess return for each unit of market exposure it accepted.

This comparison illustrates the core function of the Treynor ratio — it adjusts raw returns for the systematic risk taken to achieve them, revealing which portfolio manager has generated superior risk-adjusted value rather than simply superior raw returns. A manager who produces higher returns by taking more systematic risk has not necessarily demonstrated superior skill — they may simply be leveraging market beta. The Treynor ratio filters out this beta-driven return and isolates the return generated per unit of market risk.

The Treynor Ratio and the Security Market Line

The Treynor ratio has a precise geometric interpretation within the Capital Asset Pricing Model framework — one that connects it directly to the Security Market Line discussed in the Security Market Line entry of this dictionary.

The Security Market Line plots the expected return of any asset or portfolio against its beta — the relationship that the CAPM predicts should hold in equilibrium for any efficiently priced asset. A portfolio plotting exactly on the Security Market Line has a Treynor ratio equal to the equity risk premium — the market return minus the risk-free rate divided by the market beta of one — which is simply the slope of the Security Market Line.

A portfolio with a positive beta that plots above the Security Market Line — meaning it earned a higher return than the CAPM predicts for its level of beta — has a Treynor ratio greater than the equity risk premium. This superior Treynor ratio is evidence of positive alpha — the portfolio manager has generated excess return above what the level of systematic risk alone would predict. A positive-beta portfolio plotting below the Security Market Line has a Treynor ratio less than the equity risk premium — evidence of negative alpha — the portfolio underperformed what its beta alone would have predicted.

This geometric interpretation connects the Treynor ratio directly to Jensen's alpha — the third major risk-adjusted performance measure alongside the Treynor ratio and Sharpe ratio. Jensen's alpha measures the vertical distance of a portfolio above or below the Security Market Line — the absolute dollar or percentage excess return above the CAPM prediction. The Treynor ratio measures the slope of the line from the risk-free rate through the portfolio's risk-return point — portfolios with higher Treynor ratios have steeper slopes and lie further above the Security Market Line relative to their beta.

The Treynor Ratio and Alpha per Unit of Beta

Jensen's alpha is the portfolio's return less the risk-free rate and less beta multiplied by the market's excess return over the risk-free rate. Dividing alpha by beta therefore gives exactly the amount by which a portfolio's Treynor ratio exceeds the market's own excess return over the same period. It follows that two portfolios with the same alpha do not have the same Treynor ratio: the one with the lower beta spreads that alpha over fewer units of systematic risk and ranks higher. A fund earning one percentage point of alpha at a beta of zero point five has a Treynor ratio two points above the market's excess return, while a fund earning the same alpha at a beta of two stands only half a point above it.

The same arithmetic governs a period in which the market falls short of the risk-free rate. A portfolio with zero alpha earns exactly the risk-free rate plus its beta multiplied by the market's realised excess return, so its Treynor ratio equals that realised excess return whatever its beta. When the market has lagged Treasury bills, every such portfolio shows a negative Treynor ratio, although none of them has fallen short of the return its systematic risk called for. Jensen made the contrasting claim for his own measure in 1968: if the pricing model is valid, the market conditions of the evaluation period have no effect on alpha.

Georges Hübner's Generalized Treynor Ratio, published in the Review of Finance in 2005, extends the single-index measure to several indexes. It is defined as the abnormal return of a portfolio per unit of premium-weighted average systematic risk, normalised by the premium-weighted average systematic risk of the benchmark, and it preserves the geometric and analytical properties of the original measure. Hübner reports from simulations that its rankings of portfolios are more precise and more stable than those given by Jensen's alpha or the information ratio.

The Treynor Ratio Versus the Sharpe Ratio — The Critical Examination Comparison

The distinction between the Treynor ratio and the Sharpe ratio is the central comparison among the risk-adjusted performance measures. Both ratios measure excess return per unit of risk — but they define risk differently and are therefore appropriate in different analytical contexts.

The Sharpe ratio uses total risk — measured by the standard deviation of portfolio returns — as its denominator. Total risk encompasses both systematic risk and unsystematic risk — the portfolio's beta-driven market sensitivity plus the company-specific, sector-specific, and strategy-specific volatility unique to that portfolio. The Sharpe ratio penalises a portfolio for all sources of volatility regardless of whether they are diversifiable or not.

The Treynor ratio uses only systematic risk — measured by beta — as its denominator. It ignores unsystematic risk entirely — neither penalising nor rewarding the portfolio for the portion of its volatility attributable to non-market factors.

The analytical consequence of this difference is that the two ratios produce different rankings for portfolios that differ in their degree of diversification — and understanding which ratio is more appropriate in a given context is the examination-critical insight.

For a well-diversified portfolio — one that has eliminated substantially all unsystematic risk through holding a large number of positions across many sectors and asset classes — total risk and systematic risk are nearly identical. When unsystematic risk has been diversified away the portfolio's total volatility is almost entirely attributable to its market exposure — its beta. In this case the Treynor and Sharpe ratios produce nearly identical rankings of portfolio performance and either is an appropriate measure. For institutional investors who hold many separate manager mandates — each covering a different asset class or strategy — each individual mandate is one component in a broader diversified total portfolio. The relevant risk measure for evaluating each individual manager is their systematic risk contribution to the total portfolio — because the unsystematic risk of each mandate is diversified away when the mandates are combined. The Treynor ratio is therefore the appropriate measure for evaluating individual managers within a multi-manager institutional framework.

For a concentrated or undiversified portfolio — one that holds few positions or that concentrates in specific sectors — unsystematic risk is substantial and has not been eliminated. The portfolio's total volatility includes significant non-market risk that the investor is bearing without compensation according to the CAPM framework. In this case the Sharpe ratio is more appropriate because it captures all the risk the investor is actually experiencing — both the systematic risk and the uncompensated unsystematic risk that the concentrated portfolio has failed to diversify away.

The examination rule is clean and memorable — use the Treynor ratio for well-diversified portfolios where unsystematic risk has been eliminated, use the Sharpe ratio for concentrated or undiversified portfolios where total risk is the relevant measure of the investor's actual risk experience.

Limitations of the Treynor Ratio

Like all quantitative performance measures the Treynor ratio has specific limitations that investment advisers must understand to apply it appropriately and avoid drawing misleading conclusions from it.

The beta dependency is the most significant limitation — the Treynor ratio's output is only as reliable as the beta estimate used in the calculation. Beta is estimated from historical return data relative to a specified market index — typically using a regression of monthly returns over a trailing window of several years — Morningstar's published methodology for its fund statistics, for example, uses the trailing thirty-six months. Historical beta estimates are imprecise — they carry estimation error, they assume the portfolio's systematic risk exposure has been stable over the estimation period, and they are sensitive to the choice of market index used as the benchmark. A portfolio that has changed its strategy, sector allocation, or leverage level during the estimation period will have a historical beta that does not accurately reflect its current systematic risk — making the Treynor ratio calculated from that historical beta an unreliable measure of current risk-adjusted performance.

The benchmark dependence means that Treynor ratios are only meaningfully comparable when computed against the same market index over the same measurement period. A portfolio's Treynor ratio computed against the S&P 500 will differ from its Treynor ratio computed against the Russell 2000 — because the beta estimate changes with the benchmark. Comparing Treynor ratios across portfolios that used different benchmarks for their beta estimation is analytically invalid.

The backward-looking limitation applies to all historical performance measures — a high historical Treynor ratio does not guarantee superior future risk-adjusted performance. Past risk-adjusted returns reflect the specific market conditions, portfolio composition, and manager decisions during the measurement period and may not persist in different future environments.

The negative beta edge case deserves specific mention — a portfolio with a negative beta — one that tends to rise when the market falls — produces a negative denominator in the Treynor formula. If the numerator is also negative — the portfolio underperformed the risk-free rate — a negative divided by a negative produces a positive Treynor ratio that erroneously appears attractive. If the numerator is positive — the portfolio outperformed the risk-free rate — a positive divided by a negative produces a negative Treynor ratio that erroneously appears unattractive. For portfolios with negative betas the Treynor ratio formula produces results that are arithmetically correct but practically misleading — the Sharpe ratio or other measures are more appropriate for such portfolios.

A Stable Beta and the Market Timer

A Treynor ratio computed from one regression over the measurement period treats the portfolio's beta as a single fixed number. Michael Jensen, setting out the regression framework in 1968, noted that this assumes the risk level of the portfolio is stationary through time, and that the assumption need not hold, because a portfolio manager can change the risk level of the portfolio very easily — by switching from more risky to less risky equities, by changing the distribution of the portfolio's assets between equities, bonds and cash, or by borrowing funds and investing them in equities — and may do so deliberately in trying to outguess the movements of the market.

Jensen showed that the regression still measures the success of such timing activity so long as the manager attempts, on average, to maintain a given target level of risk, with the risk actually carried at any moment varying around that target according to the manager's forecasts of the market. The single beta estimated from the period is then an estimate of that average risk level, not of the risk the portfolio carried in any particular month.

The estimate is not neutral, however. Jensen showed that if the manager has no ability to forecast general market movements, the regression gives an unbiased estimate of the manager's ability to select undervalued securities; but if the manager can forecast market movements, the estimated beta is biased downward and the estimated alpha is biased upward, so the measure tends to overstate the effect of the timing ability. For a portfolio whose return exceeded the risk-free rate, a beta estimate biased downward also raises the Treynor ratio computed from it, since the same excess return is divided by a smaller number.

Testing for Timing Skill — Treynor and Mazuy, Henriksson and Merton

Because a single beta cannot by itself separate selection from timing, two return-based regressions have been used on a regular basis to detect market-timing skill. The first, proposed by Treynor and Mazuy in 1966, adds to the regression of the fund's excess return on the market's excess return a second term in the square of the market's excess return. The coefficient on that squared term measures the convexity of the fund's exposure to the market portfolio: if it is positive, the manager gradually increases beta as the market goes up, which is read as evidence of timing ability, while the intercept continues to measure the return from selecting securities.

The second, proposed by Roy Henriksson and Robert Merton in 1981, is piecewise rather than curved. It treats the manager as switching the portfolio between two betas depending on the sign of the market's excess return — a high beta when the market return exceeds the risk-free rate and a lower beta otherwise — so that a good market timer shows greater market exposure in rising markets than in falling ones.

Both tests can be gamed. As Georges Hübner summarises the work of Ingersoll and his co-authors published in 2007, a manager who has access to a complete derivatives market can easily alter the measured market-timing coefficient without affecting the regression intercept to a proportional extent. A convex pattern in a fund's returns can therefore come from derivative positions as well as from any genuine ability to forecast the market. Robert Merton's interpretation of the switching model rests on the cost of replicating a protective put on the market, and Hübner applies the same option-replication reasoning to the Treynor and Mazuy regression, using both its linear and its quadratic coefficients to price the cheapest option portfolio with the same convexity.

The Market Index Behind the Beta

Every Treynor ratio is a ratio to a beta, and every beta is a beta against a particular index. Deborah Kidd, writing for CFA Institute, cites Richard Roll's 1978 finding that even small changes in the proxy used for the market had large effects on risk-adjusted ratios, and draws the practical consequence: a portfolio with a low beta measured against a highly volatile index could have a higher Treynor ratio than it would if the ratio were based on a higher beta measured against a less volatile index, so an investor who selects the portfolio with the higher Treynor ratio and lower beta without knowing the index behind it might inadvertently choose a portfolio that falls outside his or her risk parameters.

A low beta is also not the same thing as low volatility. Morningstar's published methodology for its fund statistics, which regresses a fund's monthly excess return over Treasury bills on the excess return of the index for the fund's broad asset class, observes that a low beta means only that the fund's index-related risk is low. A specialty fund investing mainly in gold will usually have a low beta and a low R-squared, because its performance is tied more closely to gold prices than to the overall stock market, and it might fluctuate wildly while its beta stays low — so its excess return divided by that beta says little about the risk its investors actually bore.

The index matters to tests of timing as much as to the ratio. A 1997 study in the Financial Analysts Journal found that the original Treynor and Mazuy regression, run against the Standard and Poor's 500 index, showed negative market-timing ability for a sample of domestic equity funds, while an extended version controlling for the funds' holdings outside that index found positive and significant timing ability for 633 funds over 1984 to 1994.

Forecasting From a Historical Treynor Ratio

The Treynor ratio is generally used to analyse past performance, and any insight investors hope to gain from it into future performance depends to a large degree on beta. Beta is often thought of in a forward-looking sense, yet it is based on historical price movements and its predictability is limited. Portfolio betas are inherently more stable than the betas of the individual securities a portfolio holds, but they are still subject to change as those underlying betas and the covariances among the holdings change over time.

The length of the estimation window matters. A beta calculated over a shorter period can have more relevance as a forecasting tool than a beta calculated over multiple years if current economic conditions differ from the past and are anticipated to persist. Beta measures the degree to which a portfolio's returns are influenced by the same factors that influence the market return — the market does not itself cause the portfolio's returns — and for that reason Barr Rosenberg and James Guy demonstrated in 1976 that beta is more useful as a forecasting tool when projected fundamental factors are included in the beta regression. A portfolio with a heightened sensitivity to interest rates, for example, will have a higher beta going forward if increased uncertainty about interest rates is expected to have a significant effect on market volatility.

When an investor asks whether a past Treynor ratio is likely to persist, Kidd's advice is to evaluate whether the period under measurement differs from current economic and market forecasts, and whether the manager is expected to employ the same strategy going forward. A portfolio's beta indicates how much risk the manager took relative to the market; the Treynor ratio assesses whether the manager generated excess return for taking that risk.

Where the Ratio Breaks Down — Near-Zero Betas and Asymmetric Returns

The formula divides by beta, so it is undefined for a portfolio whose beta is exactly zero and unstable for one whose beta is close to zero. A market-neutral portfolio built to carry almost no market exposure can show an enormous Treynor ratio from a modest excess return — an excess return of two percentage points over a beta of zero point zero five gives a ratio of forty — and a small error in so small a beta estimate moves the ratio a long way, since an estimate of zero point one instead would halve it. Where such a portfolio's returns are also poorly explained by the index, Morningstar's methodology treats the beta itself as less reliable, because a high R-squared generally indicates a more reliable beta figure and a low one does not.

The ratio also inherits the assumptions of the Capital Asset Pricing Model from which it is drawn. Kidd notes that the model's mean-variance assumption limits the Treynor ratio and Jensen's alpha to strategies that are expected to have normally distributed returns, and that they are not useful for asymmetrical return strategies — strategies whose payoffs are deliberately skewed rather than spread symmetrically about their mean. Both measures have also been criticised for the model's assumptions about beta and costless transactions, and for its reliance on a single factor to explain excess return.

For a portfolio in either position the formula still returns a number, so nothing in the figure itself warns the reader that the beta beneath it is near zero or that the returns behind it are skewed; the limitation has to be checked from the portfolio's beta, its R-squared and the nature of its strategy.

The Treynor Ratio in the Fiduciary Framework

Investment advisers owe their clients a fiduciary duty under the Investment Advisers Act of 1940, and broker-dealers recommending securities transactions or investment strategies to retail customers must meet the care obligation of Regulation Best Interest at 17 CFR 240.15l-1, which requires reasonable diligence, care and skill to understand the potential risks, rewards and costs of a recommendation — so a professional who evaluates and compares portfolio managers or investment strategies on behalf of clients should use analytically appropriate performance measures. Presenting only raw returns without risk adjustment — or using the wrong risk-adjusted measure for the portfolio's diversification level — may fall short of these duties of care by providing incomplete or misleading performance information.

An investment adviser evaluating multiple mutual fund managers for inclusion in a client's diversified portfolio should use the Treynor ratio rather than the Sharpe ratio for the individual manager comparison — because each manager's mandate is one component of a larger diversified portfolio and the relevant risk measure is systematic risk contribution rather than total volatility. An adviser evaluating a single concentrated stock portfolio against a diversified benchmark should use the Sharpe ratio — because the client is bearing the full total risk of the concentrated portfolio and needs a measure that captures both the systematic and unsystematic components of that risk.

Is a Treynor Ratio Performance Under the Marketing Rule

An investment adviser that shows a Treynor ratio in an advertisement must consider the SEC's Marketing Rule, Rule 206(4)-1 under the Investment Advisers Act of 1940, whose performance provisions sit in paragraph (d) of the rule. The rule does not define performance, so an adviser may need to decide whether a particular portfolio or investment characteristic is performance at all. In a frequently asked question posted on March 19, 2025, the staff of the SEC's Division of Investment Management gave as examples of such characteristics yield, coupon rate, contribution to return, volatility, sector or geographic returns, attribution analyses, the Sharpe ratio, the Sortino ratio and other similar metrics — a description that on its terms reaches a risk-adjusted ratio such as the Treynor ratio.

The staff took no position on whether any particular characteristic is performance; that remains a determination for the adviser on the facts. The answer states the consequence of the determination: if a characteristic is not performance, its presentation falls outside the scope of the rule's performance provisions in paragraph (d) altogether. Where the adviser concludes that the characteristic is performance, the answer described in the next section supplies the conditions on which the staff would not object to its being shown gross.

The answer is also confined in what it covers. Its positions do not apply to total return, time-weighted return, return on investment, internal rate of return, multiple on invested capital or total value to paid-in capital, however those figures are labelled in an advertisement. A Treynor ratio is a characteristic computed from returns, but the answer's positions give no relief for presenting those returns themselves.

Showing a Gross Treynor Ratio in an Advertisement

A Treynor ratio computed from returns before the deduction of all the fees and expenses a client has paid or would have paid for the adviser's services is, in the staff's terms, a gross characteristic. The March 19, 2025 answer says that the staff would not recommend enforcement action to the Commission under Rule 206(4)-1(d)(1) against an adviser that displays a gross characteristic without the corresponding net characteristic, provided four conditions are met. The provision in question, paragraph (d)(1), bars any presentation of gross performance unless the advertisement also presents net performance with at least equal prominence, in a format designed to facilitate comparison, and calculated over the same time period using the same type of return and methodology — so a gross characteristic that is performance would otherwise need its own net counterpart.

First, the gross characteristic is clearly identified as being calculated without the deduction of fees and expenses. Second, it is accompanied by the total portfolio's gross and net performance, presented consistently with the rule's requirements. Third, that total portfolio gross and net performance is presented with at least equal prominence to, and in a manner designed to facilitate comparison with, the gross characteristic. Fourth, the total portfolio's gross and net performance is calculated over a period that includes the entire period over which the characteristic is calculated.

The fourth condition turns on the periods rather than the figures. A gross Treynor ratio calculated from five years of returns, shown beside the portfolio's gross and net performance for the most recent year only, does not meet it, however clearly the ratio is labelled; the accompanying gross and net performance must cover at least the whole five years from which the ratio was calculated.

Treynor and Black — From Measuring Alpha to Using It

In 1973 Treynor and Fischer Black turned from rating a manager's results to using security analysis to build a portfolio, in a Journal of Business paper titled How to Use Security Analysis to Improve Portfolio Selection. William Sharpe, discussing the paper in 1994, describes its benchmark as in effect a combination of riskless securities and the market portfolio, against which each component's differential return is measured.

Treynor and Black showed that if resources are allocated optimally, the holding of each component is proportional to the ratio of its mean differential return to the square of the standard deviation of its differential return. A component with a negative mean differential return should therefore be held in a negative amount — sold short — rather than simply left out, and the product of the mean return and the optimal holding is always positive.

Under that optimal allocation, the square of the portfolio's ratio of mean differential return to the standard deviation of differential return equals the sum of the same squared figures for its components. Treynor and Black used the term appraisal ratio for the squared figure of a component and the term Sharpe ratio for the squared figure of the whole portfolio. Sharpe, preferring the unsquared measure, observed that a squared ratio has the curious property that all of its values are positive — even those for which the mean differential return is negative — and so obscures important information about performance; he refers to the Treynor and Black measure instead as the Sharpe ratio squared.

Examination Relevance and Key Takeaways

The Treynor ratio falls within the Series 65 examination's coverage of risk-adjusted returns and is best studied in the context of risk-adjusted performance measurement, the CAPM framework, the comparison with the Sharpe ratio, and the appropriate application of each measure for diversified versus undiversified portfolios.

The key points to retain are these.

The Treynor ratio equals the portfolio return minus the risk-free rate divided by the portfolio beta — measuring excess return per unit of systematic risk. It was introduced by Jack Treynor in a 1965 Harvard Business Review article and grew out of his foundational work on the Capital Asset Pricing Model, which he began in the late 1950s independently of the model's other originators. A higher Treynor ratio indicates superior risk-adjusted performance — more excess return generated per unit of systematic market risk accepted. The portfolio with the higher Treynor ratio has produced more reward for each unit of beta exposure regardless of which portfolio produced the higher absolute return.

The critical distinction from the Sharpe ratio is the risk denominator — the Treynor ratio uses beta measuring only systematic risk while the Sharpe ratio uses standard deviation measuring total risk including both systematic and unsystematic risk. The Treynor ratio is most appropriate for well-diversified portfolios where unsystematic risk has been eliminated through diversification — leaving only systematic risk as the relevant risk dimension — and for evaluating individual manager mandates within a multi-manager diversified institutional portfolio. The Sharpe ratio is more appropriate for concentrated or undiversified portfolios where total risk including substantial unsystematic risk is the relevant measure of the investor's actual risk experience.

The Treynor ratio has a precise Security Market Line interpretation — positive-beta portfolios with Treynor ratios above the equity risk premium plot above the SML and have generated positive alpha while positive-beta portfolios with ratios below the equity risk premium plot below the SML and have generated negative alpha. Key limitations include beta estimation error from historical data — sensitivity to the choice of market index benchmark — backward-looking nature that does not guarantee future performance — and analytical breakdown for negative-beta portfolios where the formula produces arithmetically correct but practically misleading results.