Convexity-Adjusted Duration Mismatch

What Is Convexity-Adjusted Duration Mismatch?

Convexity-Adjusted Duration Mismatch (CADM) quantifies the gap between the duration and convexity profiles of a portfolio's assets and liabilities simultaneously. Standard duration-matching ensures that a 1% parallel shift in the yield curve produces equal dollar changes on both sides of a balance sheet. But when assets carry embedded optionality — such as mortgage-backed securities, callable bonds, or structured notes — their effective duration changes as rates move, creating a nonlinear residual risk that linear duration metrics cannot capture. CADM combines the first-order (duration) and second-order (convexity) sensitivities into a single mismatch metric, expressed in dollar terms per unit of rate change squared.

Formally, CADM = (D_A – D_L) + 0.5 × (C_A – C_L) × Δr, where D denotes modified duration, C denotes convexity, and Δr is the scenario rate move. A positive CADM implies the asset side gains more (or loses less) than liabilities as rates rise — a favorable but often costly hedge position. Negative CADM signals that liability sensitivity outpaces assets in large rate moves, which is the classic underfunded pension or S&L crisis scenario.

Why It Matters for Traders

Fixed income portfolio managers, bank ALM desks, and insurance companies use CADM to stress-test hedges under large, nonparallel rate moves. A portfolio perfectly duration-matched on paper can have severe CADM if the asset side is loaded with negative convexity instruments like agency MBS. When rates rise sharply, MBS prepayments slow and effective duration extends precisely when the manager needs it to shorten — the convexity hedge fails at the worst moment. This dynamic was central to the March 2023 regional bank stress, where duration-matched but convexity-mismatched held-to-maturity portfolios produced mark-to-market losses that exceeded equity buffers.

For macro traders, monitoring aggregate CADM across the banking system offers an early warning for convexity hedging flows — large-scale selling of Treasuries or receiver swaptions triggered by duration extension in MBS books.

How to Read and Interpret It

• CADM near zero: Portfolio is insulated from nonlinear rate moves; hedges are robust across a wide range of scenarios.
• CADM < –50bps DV01-equivalent: Material mismatch; a 100bp rate shock will expose the portfolio to losses beyond what duration hedges cover. Institutional desks typically trigger rehedging protocols at this threshold.
• CADM sharply negative during rising-rate regimes: Watch for forced convexity buying (receiver swaptions, long Treasury futures) which can amplify rate rallies — a self-reinforcing dynamic known as the convexity bid.
• Compare CADM against regulatory capital buffers: if the dollar mismatch exceeds Tier 1 capital by more than 15–20%, the institution faces solvency risk under stress scenarios.

Historical Context

The savings and loan crisis of the late 1980s and early 1990s is the canonical CADM failure. US thrifts funded long-duration fixed-rate mortgages (asset duration ~10–12 years) with short-term deposits (liability duration ~0.5 years), creating a CADM of roughly –9 years before convexity adjustment. When the Fed raised the Fed Funds Rate from 11% to 20% between 1979 and 1981, the convexity gap widened further as mortgage prepayments collapsed, extending asset duration to 15+ years. The result was over 1,000 thrift failures and a government bailout exceeding $160 billion. A more recent example is March 2023, when Silicon Valley Bank's $91 billion HTM portfolio had an estimated duration of 5.6 years against near-zero duration liabilities, and negative convexity from MBS holdings amplified losses to ~$15 billion on a $16 billion equity base.

Limitations and Caveats

CADM assumes a parallel yield curve shift; bear steepeners or bull flatteners can produce very different outcomes from those the metric predicts. Convexity estimates for MBS depend heavily on prepayment model assumptions, which can be wrong during unprecedented rate regimes. The metric also ignores basis risk between hedge instruments (swaptions vs. physical Treasuries) and liquidity constraints that prevent timely rehedging. Finally, CADM is a static snapshot — dynamic hedging costs and transaction friction are not embedded.

What to Watch

• Fed MBS reinvestment policy and how quantitative tightening affects the aggregate negative convexity overhang in bank portfolios.
• MOVE Index levels above 120, which signal environments where convexity hedging flows are most likely to distort Treasury markets.
• Bank ALM disclosures for duration gap data, particularly for institutions with HTM portfolios exceeding 25% of assets.
• Mortgage application and refinancing data (MBA index) as a real-time proxy for MBS convexity extension or compression.