Bond Convexity: Why Duration Alone Doesn’t Tell the Whole Story

If you’ve spent any time with fixed income concepts, you’ll know that duration is the go-to measure for a bond’s interest rate sensitivity. A bond with a duration of 5 years loses roughly 5% in price for every 1% rise in yield, and gains roughly 5% for every 1% fall. Useful, intuitive, and widely used. But here’s the problem: that relationship is only approximately true, and the approximation gets worse precisely when it matters most during large, rapid interest rate moves.

The reason duration is only an approximation is that it assumes the price-yield relationship is linear. It isn’t. The actual relationship between a bond’s price and its yield is a curve, and that curvature is what convexity measures. Understanding convexity properly means understanding why that curve exists, what it implies for bond pricing, and why perhaps counterintuitively convexity is almost always a good thing for bond investors.

The Price-Yield Relationship: A Curve, Not a Line

Start with something you already know: bond prices and yields move in opposite directions. As yields rise, bond prices fall; as yields fall, bond prices rise. This much is straightforward from the discounting mechanics of bond valuation.

What’s less intuitive is the shape of that relationship. If you graphed a bond’s price against a continuous range of yields, you wouldn’t get a straight line, you’d get a curve that bows outward toward the investor. Technically, it bows toward the upper-left on a price-yield diagram, which means for a given change in yield, the price gain from a yield decrease is larger than the price loss from an equal yield increase.

That outward bow is convexity. And it has a specific implication that matters enormously for fixed income portfolios: a bond with more convexity performs better than a bond with less convexity in both rising and falling rate environments, assuming equal duration. It gives more price appreciation when rates fall and cushions more price loss when rates rise. This is why convexity is considered uniformly desirable investors are essentially getting a benefit on both sides of the rate movement.

Why the Curve Exists: The Mathematics Behind the Bow

Duration, as a price sensitivity measure, is derived from the first derivative of the price-yield function; it measures the slope of the price-yield curve at the current yield level. The problem is that a first derivative captures only the local linear approximation. As you move away from the current yield level, the slope itself changes, because the price-yield relationship is curved rather than straight.

Convexity captures the second derivative, the rate of change of the duration measure itself as yields change. A positive convexity value means that as yields fall, duration increases (the bond becomes more sensitive to further rate moves); as yields rise, duration decreases (the bond becomes less sensitive to further rate moves). This self-reinforcing property is precisely what creates the beneficial asymmetry described above: when rates fall and prices rise, the bond’s sensitivity increases, amplifying further gains; when rates rise and prices fall, the bond’s sensitivity decreases, cushioning further losses.

Duration vs. Convexity: A Two-Term Price Approximation

The practical way this enters the CFA curriculum is through the full price change approximation formula, which uses both duration and convexity together to estimate how a bond’s price will change for a given yield movement.

The approximate percentage price change can be expressed as:

% Price Change ≈ (−Modified Duration × Δy) + (½ × Convexity × Δy²)

The first term modified duration multiplied by the yield change is the linear, duration-based estimate. The second term one-half times convexity times the squared yield change is the convexity correction. This correction is always positive, regardless of whether yields rise or fall, because Δy is squared. Rising yields mean Δy is positive, Δy² is still positive, so convexity adds back some of the price loss. Falling yields mean Δy is negative, Δy² is still positive, so convexity adds to the price gain.

This is the mathematical expression of the beneficial asymmetry: convexity always works in the investor’s favor, adding a positive correction to the linear duration estimate no matter which direction yields move.

A Worked Numerical Example

Let’s make this concrete. Suppose a bond has a modified duration of 7 and a convexity of 80. Yields change by 100 basis points (1%) in both directions.

For a 100 bps yield increase (Δy = +0.01):

Duration-only estimate: −7 × 0.01 = −7.00%

Convexity correction: +½ × 80 × (0.01)² = +½ × 80 × 0.0001 = +0.004 = +0.40%

Full approximation: −7.00% + 0.40% = −6.60%

For a 100 bps yield decrease (Δy = −0.01):

Duration-only estimate: −7 × (−0.01) = +7.00%

Convexity correction: +½ × 80 × (−0.01)² = +0.40% (still positive)

Full approximation: +7.00% + 0.40% = +7.40%

Notice the asymmetry: the full price loss on a 100 bps rise is 6.60%, but the full price gain on a 100 bps fall is 7.40% the gain exceeds the loss by 0.80% even though the yield move is equal in magnitude. That 0.80% differential is the convexity benefit working exactly as described.

Also notice that the convexity correction is the same in both cases 0.40% because Δy² is always positive regardless of direction. Convexity makes a symmetric positive contribution to price change estimates, always improving the outcome relative to the duration-only estimate.

Now imagine yields move by 200 bps instead of 100 bps. The convexity correction scales with Δy², so it becomes ½ × 80 × (0.02)² = ½ × 80 × 0.0004 = 1.60% four times larger than the 100 bps correction. This is why convexity matters most during large yield moves: its contribution grows with the square of the rate change, making it increasingly material as volatility increases.

What Drives Higher Convexity?

Understanding which bonds carry higher convexity helps connect the concept to real portfolio management decisions.

Longer maturity bonds have higher convexity than shorter maturity bonds with otherwise similar characteristics, because the price-yield relationship is more curved for bonds with cash flows distributed far into the future. The discounting of distant cash flows amplifies the curvature in the relationship.

Lower coupon bonds and particularly zero-coupon bonds have higher convexity than high coupon bonds of similar duration. A zero-coupon bond has all its cash flow at maturity, creating maximum curvature in the price-yield relationship.

Lower yield environments tend to produce higher convexity, because the price-yield curve is more steeply curved at lower yield levels than at higher ones.

Callable bonds have negative convexity over certain yield ranges, a critical exception worth understanding. When yields fall significantly, the probability of the issuer calling the bond rises, which caps the bond’s price appreciation potential. The price-yield curve for a callable bond actually bends the wrong way in the low-yield region: rather than the price accelerating upward as yields fall, the price is capped near the call price, producing negative convexity. This is why callable bonds typically offer higher yields than otherwise comparable non-callable bonds investors require compensation for the negative convexity they’re accepting.

Mortgage-backed securities are the most commonly discussed real-world example of negative convexity in the CFA curriculum. As yields fall, prepayment speeds accelerate homeowners refinance at lower rates which shortens the effective duration of the MBS and caps price appreciation, creating the same negative convexity dynamic as callable bonds. When yields rise, prepayments slow, duration lengthens, and price losses deepen. This extension risk combined with negative convexity is precisely why MBS carry spread above comparable Treasury securities.

Money Convexity: From Percentage to Rupee Terms

The convexity figure used in the percentage price change formula gives a percentage sensitivity. For portfolio management, analysts often want the dollar or rupee value impact, which requires translating convexity into money terms.

Money convexity (also called convexity in rupee or dollar terms) is simply the bond’s convexity multiplied by its full price per unit:

Money Convexity = Convexity × Full Price

And the rupee price change contribution from convexity for a given yield move is:

Rupee Convexity Contribution = ½ × Money Convexity × Δy²

This framing is particularly useful when comparing bonds with different face values or when thinking about convexity contribution across an entire portfolio. A large position in a bond with modest convexity might contribute more rupee convexity to a portfolio than a small position in a high-convexity bond, and the portfolio manager needs to think in position-weighted rupee terms rather than just per-bond percentage terms.

Convexity in Portfolio Management: Why It Matters

At the portfolio level, convexity isn’t just a risk measurement tool it’s a source of value in volatile rate environments. All else equal, a portfolio with higher convexity will outperform a portfolio with lower convexity when rates move significantly in either direction.

This creates a genuine trade-off that fixed income portfolio managers navigate constantly. Higher convexity bonds long-maturity governments, zero-coupon bonds typically offer lower yields than lower convexity bonds with similar duration. Investors pay for convexity in the form of a lower yield the convexity premium. Whether that premium is worth paying depends on the portfolio manager’s view on rate volatility: if rates are expected to be volatile, high convexity is worth its cost; if rates are expected to be stable, the convexity premium is a drag on returns that may not earn its keep.

This is the fundamental convexity trade-off: you pay for it through a lower initial yield, and it pays back through superior price performance in volatile rate environments. A stable rate environment rewards low-convexity, high-yield strategies; a volatile rate environment rewards high-convexity strategies even at lower initial yields.

Exam Perspective: What to Lock In

For CFA Fixed Income, a handful of points deserve clear anchoring. Convexity measures the curvature in the price-yield relationship, specifically the second-order price sensitivity that duration, as a linear approximation, misses. The full price change approximation adds the convexity term (½ × Convexity × Δy²) to the duration term (−Modified Duration × Δy), and the convexity term is always positive regardless of yield direction, creating beneficial asymmetry. Higher convexity means better price performance in both rising and falling rate environments, all else equal gain more when rates fall, lose less when rates rise. Convexity increases with longer maturity, lower coupon, and lower yield levels. Callable bonds and MBS exhibit negative convexity in certain yield ranges because price appreciation is capped when rates fall significantly. And the convexity premium the lower yield investors accept on high-convexity bonds is the price of that beneficial asymmetry, justified when rate volatility is high but a performance drag when rates are stable.

Final Thoughts

Convexity is one of those fixed income concepts that becomes more valuable the more volatile the interest rate environment. In a stable rate world where yields drift gently in a narrow band, convexity is a theoretical nicety present in the pricing formula, largely invisible in returns. In a world where rates move sharply and unexpectedly, convexity becomes the difference between a bond portfolio that holds up and one that underperforms materially.

The core insight to carry forward is the beneficial asymmetry: convexity always adds to price performance, on both the upside and the downside of rate moves, because the mathematical correction it provides is always positive. Duration tells you approximately how sensitive a bond is to rate changes. Convexity tells you the good news that the true sensitivity is actually better than duration alone suggests, in every direction.