What Is Negative convexity?

A position or trading strategy has negative convexity when its value bends downwards as the market moves: gains slow down or stop as the market moves in its favour, and losses speed up as it moves against it. The second derivative of value with respect to price (or yield) is negative, so the payoff is concave. It is the opposite of positive convexity and describes the same exposure as short volatility.

The two textbook examples are:

  • Sold options: a sold option has negative gamma. The seller’s gain is capped at the premium received, while the loss grows with the size of the move. See selling volatility.

  • Callable bonds and mortgage-backed securities: when interest rates fall, borrowers refinance and repay early, so the price rises less than an ordinary bond’s would; when rates rise, repayments slow and the security behaves like a longer bond just as its price falls. Hanson (2014) shows that investors hedging this mortgage convexity buy and sell government bonds in the direction of the rate move, which amplifies yield moves.

Investors take negative convexity because they are paid for it. The seller of an option collects a premium, a callable bond pays a higher yield than a non-callable one, and a market maker earns a spread. The result is frequent small gains and rare large losses, a negatively skewed return distribution, and a track record that can look smooth until the first large move. By Jensen’s inequality, a concave payoff loses value as volatility rises, so the position also loses when the market prices in more risk.

Negative convexity in trading strategies and DeFi

  • AMM liquidity provision: the value of a constant-product liquidity provider position grows with the square root of the price ratio, a concave function. The loss against holding is impermanent loss, and a CLMM position is more concave inside its range. Fees are the premium (Milionis et al. 2022).

  • Option-selling vaults: covered calls, cash-secured puts and other option-writing strategies.

  • Market making, grid trading, mean reversion and martingale strategies, which add to positions as the price moves against them.

  • Leveraged yield and carry: recursive looping, cash and carry and funding rate harvesting, which earn steadily and face forced unwinds or liquidation in sharp moves.

Dealers and hedgers who are short convexity and delta hedge have to buy as prices rise and sell as they fall. When many hold the same exposure, their hedging adds to the move.

Negative convexity is found by the same checks as positive convexity: a frown instead of a smile when strategy returns are plotted against market returns, a negative coefficient on the squared market-return term in a Treynor and Mazuy (1966) regression, negative skew, and losses that coincide with spikes in implied volatility. Tail measures such as expected shortfall and risk of ruin describe the risk better than the Sharpe ratio.

Vault strategy categories with this profile include options, liquidity provider, market making, grid trading, martingale-like and lending looping vaults.

Literature and references:

  • Samuel G. Hanson, Mortgage Convexity, Journal of Financial Economics 113(2), 2014.

  • Jason Milionis, Ciamac C. Moallemi, Tim Roughgarden and Anthony Lee Zhang, Automated Market Making and Loss-Versus-Rebalancing, 2022.

  • Peter Carr and Liuren Wu, Variance Risk Premiums, Review of Financial Studies 22(3), 2009.

  • Frank J. Fabozzi, Bond Markets, Analysis, and Strategies, Pearson (negative convexity of callable bonds and mortgage-backed securities).

  • Nassim Nicholas Taleb, Dynamic Hedging: Managing Vanilla and Exotic Options, Wiley, 1997.

See also