The Capital Market Line: Adding a Risk-Free Asset

So far the best portfolios you could buy lived on a curve: the efficient frontier, that graceful arc of risky-asset mixes where each point squeezes the most return out of its risk. It’s a beautiful curve. It’s also about to get demoted.

Because we’ve quietly been assuming you can only hold risky things. Drop that. Let people park money in something that barely moves — a short-term Treasury bill — and the whole picture snaps from a curve into a straight line. One line that beats the entire frontier at almost every point. That line is the Capital Market Line (CML), and once you see why it’s straight, you’ll understand why every textbook portfolio is really just two ingredients: one risky basket and a pile of cash.

The risk-free asset: the boring anchor everything pivots on

Picture an asset that pays you a known return with essentially no surprises — a short-term government Treasury bill held to maturity is the classic stand-in. You hand over cash now, you get a fixed amount back in three months, and a solvent government is about as close to certain to pay as finance gets.

Two numbers define it, and both are extreme:

• Volatility ≈ 0. The return is locked in, so there’s no spread of outcomes. Standard deviation $\sigma_f \approx 0$.
• Correlation ≈ 0 with risky assets. A fixed payout doesn’t bob up and down with the stock market — it doesn’t bob at all. So its covariance with anything is essentially zero.

We’ll call its return the risk-free rate, written $r_f$. On a risk–return chart (volatility on the x-axis, expected return on the y-axis), the risk-free asset is the one point sitting flat on the return axis: zero risk, height $r_f$. That single anchor point is what bends the whole story.

Mixing risk-free with one risky portfolio: why it’s a straight line

Here’s the magic trick, and it’s pure algebra. Take any risky portfolio — call its expected return $\mathbb{E}[R_{risky}]$ and its volatility $\sigma_{risky}$ — and split your money: a fraction $m$ goes into the risky portfolio, the rest, $1-m$, into the risk-free asset.

What’s the combined expected return? Just the weighted average:

$\mathbb{E}[R_p] = (1-m)\,r_f + m\,\mathbb{E}[R_{risky}] = r_f + m\big(\mathbb{E}[R_{risky}] - r_f\big).$

What’s the combined volatility? Normally combining two assets means an ugly variance formula with a covariance term. But the risk-free asset has zero variance and zero covariance with everything, so all those terms vanish and you’re left with just:

$\sigma_p = m\,\sigma_{risky}.$

Look at what happened. Both $\mathbb{E}[R_p]$ and $\sigma_p$ are linear in $m$ — straight-line functions of the same dial. As you turn $m$ from 0 to 1 and beyond, the point $(\sigma_p, \mathbb{E}[R_p])$ slides along a straight line that starts at the risk-free point $(0, r_f)$ and passes through the risky portfolio. No curve. The curvature on the old frontier came entirely from covariance between two risky assets; kill one asset’s variance and the bend disappears.

Worked example

Let the risky portfolio return 10% with volatility 16%, and let $r_f = 3\%$. Put 50% in each ($m = 0.5$):

Halving your exposure to the risky basket halved its volatility (16% → 8%) and pulled the return exactly halfway between $r_f$ and the risky return (3% → 6.5%). Both moved in lockstep along a line. That straight, proportional trade-off is the entire point.

The tangency portfolio: the steepest line you can draw

You can draw a line from the risk-free point to any risky portfolio on the frontier. Some lines are steep (great return per unit of risk), some are shallow (lousy). Which risky portfolio should you actually combine with cash?

The one that gives the steepest line. Geometrically, pivot a ruler at the risk-free point $(0, r_f)$ and rotate it upward until it just touches the efficient frontier — touches it at a single point without crossing above. That kissing point is the tangency portfolio, and the line to it is steeper than the line to any other frontier portfolio.

Why steepest? Because the slope of the line from $r_f$ to a portfolio is exactly its Sharpe ratio — excess return per unit of volatility, $(\mathbb{E}[R] - r_f)/\sigma$. The tangency portfolio is, by construction, the risky portfolio with the highest achievable Sharpe ratio. Any other risky portfolio you mix with cash gives you a flatter line, which means less return for the same risk. So tangency wins, full stop.

The Capital Market Line: the line that dominates the curve

That steepest line — from $r_f$ through the tangency portfolio and onward — is the Capital Market Line. In a market where everyone agrees on the inputs, the tangency portfolio is the market portfolio $M$ (the whole basket of risky assets, value-weighted), so we write the CML as:

$\mathbb{E}[R_p] = r_f + \frac{\mathbb{E}[R_M] - r_f}{\sigma_M}\,\sigma_p.$

Read it as a recipe: start at the risk-free rate, then earn an extra slope × your volatility. And that slope, $\dfrac{\mathbb{E}[R_M] - r_f}{\sigma_M}$, is precisely the Sharpe ratio of the market portfolio — the price the market pays you for taking on a unit of risk.

Slide the allocation dial below and watch the marker travel the CML. Notice the muted curved frontier underneath the straight line: the CML lies above the frontier everywhere except where it touches at the market portfolio. That’s the punchline — for any level of risk you pick, the CML offers at least as much return as the old curve, and usually more. The risk-free asset didn’t just add an option; it dominated the entire efficient frontier of risky-only portfolios.

Two-fund separation: everyone holds the same risky basket

Here’s the result that makes finance professors grin. If the CML is the best menu, then every rational investor holds some mix of just two things: the risk-free asset and the one tangency (market) portfolio. Nobody needs a bespoke pile of stocks — your stock picks are identical to everyone else’s (the market portfolio); the only personal decision is how much to put in it versus cash. This is two-fund separation.

So risk appetite no longer changes what risky assets you own — it only changes the dial $m$:

• Conservative? You lend. Keep $m < 1$: some money in T-bills, the rest in the market portfolio. You sit on the CML left of the market point — lower risk, lower return. Lending money to the government is the risk-free leg.
• Aggressive? You borrow. Push $m > 1$ by borrowing at the rate $r_f$ and plowing the loan into the market portfolio — that’s leverage. You sit on the CML right of the market point — higher risk, higher return, same Sharpe slope.

Worked Sharpe ratio

The slope of the whole CML is just the market’s Sharpe ratio. With $\mathbb{E}[R_M] = 10\%$, $r_f = 3\%$, $\sigma_M = 16\%$:

$\text{Sharpe} = \frac{10\% - 3\%}{16\%} = \frac{7}{16} \approx 0.44.$

That 0.44 is the exchange rate of the entire market: every extra percentage point of volatility you take on buys you about 0.44 extra percentage points of return — whether you get there by lending ($m<1$) or borrowing ($m>1$). Pick your spot on the line; the rate per unit of risk never changes.

Pitfalls: where the clean line gets messy

The CML is gorgeous on paper. Reality sands off some corners — and a careful investor names them:

• You can’t borrow at the true risk-free rate. Governments borrow at $r_f$; you don’t. Your borrowing rate is higher, so the right half of the CML (the leverage segment) is actually flatter — a kink at the market point, not a single straight line. Borrowed return per unit of risk is worse than lent return per unit of risk.
• The “risk-free” asset isn’t truly risk-free. Inflation can erode the real return, and when a short bill matures you face rollover (reinvestment) risk at an unknown future rate. Certain in nominal dollars ≠ certain in purchasing power.
• The tangency portfolio depends on noisy estimates. It’s computed from expected returns, volatilities, and correlations — all estimated from past data and notoriously unstable. Small errors in the inputs can swing the “optimal” portfolio wildly, so the real-world tangency point is fuzzier than the crisp dot on the chart.
• Leverage cuts both ways. Going right of the market point ($m>1$) magnifies losses just as much as gains, and a leveraged position can trigger a margin call — being forced to sell at the worst possible moment. The line says “more return for more risk”; it doesn’t warn you about getting liquidated.

Key Takeaways

Next up: CAPM — if everyone holds the market, what return should a single asset earn? Enter beta and the Security Market Line.