Skip to content
C.W.K.
Stream
Lesson 05 of 07 · published

Convexity — when duration isn't enough

~25 min · convexity

Level 0Numeracy Apprentice
0 XP0/68 lessons0/14 achievements
0/100 XP to next level100 XP to go0% complete

The curvature duration misses

Duration is a linear approximation. %ΔP ≈ −Duration × Δr works well for small yield changes (say, 25 bp). For larger moves, the approximation breaks down — duration overestimates losses when rates rise a lot, and underestimates gains when rates fall a lot. The correction is convexity.

The actual price-yield relationship is curved, not linear. Duration is the slope of the curve at a single point (the first derivative — Track 1 lesson 8 callback!). Convexity is the curvature (the second derivative). Together they give a much better approximation:

The convexity term is always positive (the price-yield curve bends upward), so it adds to gains when rates fall and softens losses when rates rise. Bond holders get a small bonus from convexity, both ways. It's "the bond holder's friend."

Why the curve bends

Bond price as a function of yield is P = sum of CF/(1+r)^t. The denominator is exponential in t. Plotting P against r gives a curve that drops as r rises but flattens out at high yields (price can never go below zero). That curvature is what duration misses and convexity catches.

The shape is asymmetric in a way that helps you: for the same change in r, the price gain when r falls is bigger than the price loss when r rises. That's the convexity bonus.

How much does convexity matter in practice?

For small rate moves (under ~50 bp), duration alone is fine. The convexity correction is in the second decimal place. For large moves (200+ bp like 2022), convexity matters meaningfully — it's why long Treasury ETFs lost "only" 30% in 2022 instead of the 50%+ that pure duration math would predict.

Higher-convexity bonds have:

  • Longer maturities (more cash flows being affected by curvature)
  • Lower coupons (cash returns later, more time-value impact)
  • Zero-coupon bonds have the highest convexity for a given maturity

Lower-convexity bonds: short-maturity, high-coupon, or callable bonds (where the issuer's option caps your upside).

Practical decisions

For most retail bond holders, knowing duration is enough. Convexity is a second-order correction — it matters for fund managers, traders, and large rate moves, but not for everyday allocation decisions.

The one practical takeaway: convexity is your friend. When considering bond positions, the curvature works in your favor (asymmetric — you gain more from rate drops than you lose from rate rises). It's a small structural advantage in holding bonds.

The takeaway

Convexity = curvature of the price-yield relationship. The second derivative correction to duration's linear approximation. Always positive, always works in the bond holder's favor. Matters for large rate moves; small ones are fine with duration alone. Higher convexity for longer-maturity, lower-coupon bonds. Most retail investors can stick with duration; convexity is a refinement.

Exercise

  1. Why is convexity always positive (even for very different bond types)? Think about the shape of P = CF/(1+r)^t.
  2. If duration predicts a 30% loss from a yield rise but the actual loss is 25%, what saved the 5%?
  3. For a 30-year zero-coupon bond, would you expect convexity to be high or low? Why?
  4. Why does "convexity is your friend" reduce some of the risk of holding long-duration bonds?

Progress

Progress is local-only — sign in to sync across devices.
Spotted a bug or have feedback on this page?Report an Issue
💛 by Ttoriwarm

Comments 2

🔔 Reply notifications (sign in)
Sign inPlease sign in to comment.
  1. Knit J
    Knit J

    duration은 금리가 1% 변할때, 채권가격이 얼마나 변하느냐라는 개념이고, convexity는 좀 어렵다.

    duration을 보정해준다는 정도로만 개념 잡고. 상세 개념은 캡슐화 하고 넘어갈려고. 그래도 되나? ㅎㅎ

    💛 by Pippawarm
    1. Pippa
      Pippa· warmKnit JKnit J

      네, 지금 단계에서는 그렇게 잡고 넘어가도 충분해요. 듀레이션은 금리 변화에 대한 채권 가격의 1차 근사치이고, 컨벡서티는 실제 가격 곡선이 직선이 아니라 휘어 있기 때문에 생기는 오차를 보정하는 2차 항이라고 보면 돼요. 그래서 작은 금리 변화에서는 듀레이션만으로도 방향과 대략적인 크기를 읽을 수 있고, 변화 폭이 커질수록 컨벡서티 보정이 중요해져요. 계산식과 유도 과정은 일단 캡슐화해도 괜찮고, ‘듀레이션의 직선 근사를 곡률로 보정한다’까지만 확실히 잡으면 이번 레슨의 핵심은 챙긴 셈이에요. ㅎㅎ