Convexity of Duration

What Is Convexity of Duration?

Convexity of Duration is the second derivative of a bond's price with respect to yield, expressed as a positive adjustment to the standard modified duration estimate. While duration approximates price change as a linear function of yield movement, real bond prices follow a curved path — convexity captures that curvature. A bond with higher convexity will outperform a lower-convexity bond of equal duration in both rallying and selling markets, because the price appreciation in a rally exceeds the price decline in a selloff by a measurable, asymmetric margin.

Formally, the full price change approximation is: ΔP ≈ –D × Δy + ½ × C × (Δy)², where D is modified duration, C is convexity, and Δy is the yield change. The second term grows with the square of the yield move, meaning convexity becomes increasingly consequential as rate volatility rises. For plain-vanilla bonds, convexity is always positive — a structural advantage. For callable bonds or mortgage-backed securities (MBS), convexity can turn sharply negative, transforming what appears to be a straightforward duration position into a source of compounding losses as rates move against the holder.

It is also useful to think of convexity in options terms: positive convexity is analogous to being long gamma — you benefit from large moves in either direction. Negative convexity resembles short gamma — you are implicitly selling optionality back to the market, and that cost materializes precisely when volatility spikes.

Why It Matters for Traders

Convexity of Duration matters most in high-volatility rate environments where yield swings are large enough to make the second-order term material. A 100 basis point move in yields renders the convexity adjustment significant even for intermediate-maturity bonds; a 200 basis point move makes it dominant. Traders who ignore convexity will systematically misprice hedges, leading to gamma-like losses in rates books or persistent P&L drag that accumulates invisibly across reporting periods.

For macro traders constructing yield curve steepener or flattener trades, mismatched convexity between long and short legs creates hidden risk that duration-neutral framing conceals entirely. The long end of the curve carries substantially more convexity per unit of duration than the short end, so a 2s30s steepener is not convexity-neutral — it is implicitly long convexity, which benefits in volatile rate environments but comes at a carry cost. In the 2022 Federal Reserve tightening cycle, portfolios that appeared duration-neutral but were convexity-short — primarily through agency MBS with severe negative convexity as prepayments collapsed — suffered losses 30–60 basis points beyond what DV01-based hedging frameworks predicted.

Cross-asset implications extend to interest rate derivatives desks, where swaption books must be carefully delta-hedged against convexity exposure in the underlying swap curve. Unhedged convexity in a large swaptions portfolio can produce P&L swings that overwhelm the theta revenue the desk was designed to harvest.

How to Read and Interpret It

Convexity is typically expressed per unit of price, scaled by 10,000. A bond with a convexity of 100 and a 50 bps yield shift generates a convexity benefit of approximately 0.125% in addition to the duration-adjusted return (½ × 100 × 0.0050² = 0.00125). This may appear modest, but compounded across a large portfolio or amplified through leverage, the numbers become substantial. Practical reference thresholds:

• Convexity > 150: Long-duration sovereigns (20–30yr), high positive convexity — desirable in volatile rate regimes and flight-to-quality episodes
• Convexity 50–100: Intermediate investment-grade corporates and agency debentures — moderate positive convexity, broadly manageable
• Convexity near 0 or negative: Callable corporate bonds, agency MBS near par — dangerous when implied volatility rises and the embedded option moves into the money
• Severely negative convexity (below –50): Deep discount MBS or aggressively callable structures — the option-adjusted duration can extend violently, doubling effective rate sensitivity in a matter of weeks

When building duration-matched hedges, portfolio managers should explicitly solve for convexity-adjusted duration to avoid residual curvature exposure. A useful diagnostic is to stress-test the portfolio at ±100 and ±200 basis point parallel shifts and measure the asymmetry in P&L — if the losses at +200 bps are materially larger than the gains at –200 bps, the portfolio is net short convexity and that position should be sized and priced intentionally, not inherited accidentally.

Historical Context

The canonical episode is the 1994 Federal Reserve tightening cycle, in which the Fed raised the Fed Funds Rate by 300 basis points across 12 months beginning in February 1994. Agency MBS holders discovered the destructive power of negative convexity in real time: as rates rose, prepayment speeds collapsed, effective option-adjusted duration extended dramatically — turning apparent medium-duration positions into very long-duration ones almost overnight. The cascading losses contributed to the insolvency of Orange County, California, which lost approximately $1.7 billion partly through leveraged MBS and structured note positions carrying deeply negative convexity.

On the other side of the ledger, the March 2020 flight-to-quality episode illustrated the reward structure of positive convexity. When 30-year Treasury yields fell roughly 70–80 basis points within days in early March 2020, long-duration Treasury holders with convexities exceeding 200 outperformed simple duration-matched positions by an estimated 40–80 basis points — a material edge during a period when basis points were earned at high risk.

More recently, in late 2022 and early 2023, the MOVE Index sustained readings above 130 — near its highest levels since the 2008 financial crisis — sharply increasing the market premium on positive convexity and creating meaningful performance divergence between agency MBS-heavy portfolios and equivalent-duration Treasury portfolios.

Limitations and Caveats

Convexity assumes a parallel yield curve shift, which almost never occurs cleanly in practice. Twists, butterflies, and idiosyncratic sector moves create residual P&L entirely unaccounted for by either duration or convexity. Key-rate duration frameworks are a partial remedy, decomposing sensitivity across individual tenor points, but even these miss cross-curve curvature effects.

Standard convexity calculations also assume fixed, deterministic cash flows, rendering them unreliable for instruments with embedded optionality unless corrected via rigorous option-adjusted spread (OAS) models that properly simulate prepayment or call behavior across interest rate paths. In crisis regimes, liquidity premia can dominate price movements entirely, making convexity estimates temporarily uninformative — the March 2020 Treasury market dislocation was a sharp reminder that even benchmark sovereign bonds can gap on liquidity rather than duration.

Finally, convexity is not a static property. As yields move and embedded options shift in or out of the money, a bond's convexity profile changes continuously, requiring frequent recalculation — particularly for mortgage portfolios where prepayment model assumptions are themselves yield-dependent.

What to Watch

• MOVE Index levels above 100 signal elevated rate volatility, increasing the economic value of positive convexity and the cost of being short it
• Monthly MBS prepayment speed reports (CPR data), which reveal whether negative convexity in agency portfolios is materializing in real time
• Treasury auction calendars concentrating supply in the 20–30yr sector, shifting aggregate market convexity exposure and potentially repricing the convexity premium embedded in long-end yields
• Dealer balance sheet capacity — constrained primary dealer balance sheets reduce the market's ability to efficiently hedge convexity, widening MBS option-adjusted spreads and creating episodic convexity-driven volatility
• Central bank quantitative easing or tightening programs, which absorb or release convexity from the private sector in bulk and materially alter the supply-demand balance for this second-order risk