Duration measures a bond's sensitivity to interest rates, expressed in years. The simplest way to understand it: a bond with a 5-year duration falls roughly 5% in price for every 1% rise in rates, and gains roughly 5% for every 1% fall. Duration matching means bringing the asset and liability sides of your balance sheet to approximately the same sensitivity number. Example: a pension fund has an obligation to pay retirement benefits in 20 years (liability duration ~20 years); it should therefore hold long-term bonds with a similar ~20-year duration, not short-term bills. When rates move, the asset value and liability present value shift in the same direction by similar magnitudes, and they offset — leaving the fund's net position unaffected. This is the basic logic of immunization — making the balance sheet "immune" to interest rates.
Silicon Valley Bank's 2023 collapse is the most famous modern textbook case of duration mismatch. SVB's problem: its liabilities (deposits) were mostly demand or short-term deposits — duration near zero — withdrawable at any time. Its assets (investment portfolio) were heavily loaded with long-term Treasuries and agency MBS with durations of 5–10 years. When the Fed rapidly hiked in 2022, SVB's long-term bond holdings fell sharply in value, but its deposit liabilities didn't decline at all — because they remained on-demand, worth face value in dollars. When tech-sector depositors panicked and ran, SVB was forced to sell long-term Treasuries to meet withdrawals, turning paper losses into real losses, and the bank failed immediately. The lesson: zero credit-risk assets (US Treasuries) can still bankrupt an institution through duration mismatch. In the RWA world, DeFi lending protocols face a similar risk — borrowing short at floating rates to fund long fixed-income positions; when rates spike, borrowing costs exceed asset yields.
Duration matching has three practical applications for tokenized fixed-income investors. First, cash-flow timing: if you expect to need funds in three years (buying a home, starting a business), buy tokenized Treasuries with a duration close to three years, not 10-year bonds. The 10-year may look higher yielding, but when you must sell at year 3, if rates have risen you sell at a loss — duration matching eliminates that risk. Second, rate-direction positioning: if you believe rates are about to fall, buy longer-duration tokens (bigger price gain); if you expect rates to rise, shorten duration (less price loss). Third, DeFi strategy duration awareness: borrowing USDC on Aave to buy tokenized long bonds means borrowing at a floating rate (duration ≈ zero) while the asset has a fixed income and long duration (possibly 5–10 years) — this is actively taking on rate risk, and the cost is severe when rates reverse.
Advanced technique: dollar duration converts duration into "for every 1% rate move, how many dollars do I gain or lose?" Formula: dollar duration = duration (years) × bond market value × 0.01. Example: you hold $1 million in tokenized Treasuries with a duration of 4 years; a 1% rate rise produces a paper loss of approximately 4 × $1M × 0.01 = $40,000. This figure lets you compare rate risk across different assets and use rate futures or interest-rate swaps to hedge. In the RWA world, mature on-chain rate-hedging tools are still scarce (Pendle offers some fixed-rate markets); most RWA investors can only passively reduce rate risk by shortening duration — choosing products with shorter maturities — rather than actively hedging with derivatives. This is one of the capability gaps between RWA fixed income and the traditional fixed-income toolkit.
A small Taiwan insurance company in 2026 is considering allocating part of its funds to tokenized fixed income. It has a batch of annuity policy liabilities maturing in 5 years (liability duration ~4.5 years). Option 1: buy a tokenized money market fund (duration ≈ zero, ~4.5% annualized) — severe duration mismatch; if rates fall significantly in 5 years, reinvestment at maturity is at lower rates and today's yield is not locked. Option 2: buy a 5-year tokenized Treasury (duration ~4.5 years, ~4.8% annualized) — high duration match; in 5 years the tokens mature precisely when the policy obligations are due, and the 4.8% yield is locked now regardless of future rate moves. Option 3: buy a 10-year tokenized bond (duration ~8 years, ~5.2% annualized) — mismatch in the other direction; must sell at market price in 5 years, at a loss if rates have risen. It chooses option 2, using duration matching to eliminate rate risk. It gives up option 3's extra 0.4%, but that premium was compensation for rate risk, which doesn't fit the insurer's risk management requirements.
The core trade-off in duration matching is certainty of preservation vs opportunity cost. Perfect duration matching makes your net value immune to rates, but it also means giving up the opportunity to actively adjust based on rate direction: if rates fall sharply after you've matched, you can't shift into longer duration to capture more capital gains. Active duration management (actively betting on rate direction) can generate outperformance but also carries greater rate-prediction risk. For most RWA individual investors: with on-chain rate-derivative tools still immature, choosing a duration close to the time horizon of your planned cash use is the most practical "passive duration management" approach — no precise calculation needed, just the basic logic of "I need this money in 3 years, so I buy the 3-year product."