Basis Risk: When The Hedge Does Not Track

You hedge a regional bank stock with a financials sector ETF because there's no listed future on that exact company and the ETF is liquid. Most days, the ETF and the stock move together well enough. Then a regional-banking-specific scare hits — a deposit run at a similarly sized peer — and the stock craters while the broad financials ETF, dominated by large diversified banks, barely moves. The hedge didn't fail because you sized it wrong; it failed because the thing you hedged with was never a perfect substitute for the thing you owned. That residual gap is basis risk, and it's present in almost every real hedge to some degree.

Where basis risk comes from

Instrument mismatch. Hedging a single stock with a sector or index proxy, hedging one bond with a different bond, hedging a specific crop with a related-but-different futures contract — anywhere the hedge isn't literally the same asset, there's a gap.

Timing mismatch. The hedge and the position settle, expire, or reset at different times, so their prices can diverge over the intervening period even if they're the same underlying asset.

Location or grade mismatch. Common in commodities — hedging oil delivered in one region with a futures contract that settles delivery in another region, where local supply shocks can move the two prices apart.

Quantifying it

$$\text{Basis} = P_{\text{position}} - P_{\text{hedge instrument}}$$

In words: basis is simply the price gap between what you own and what you're hedging with, and basis risk is the variability of that gap over the life of the hedge — not its level, since a level difference can be scaled away, but its wobble.

Worked example

You're long $8m of the regional bank stock, hedged with $8m short of the financials ETF, using a hedge ratio estimated at 1.0 from a stable historical correlation of about 0.85 between the two.

Over a calm month, the stock and ETF track closely — daily basis moves average under 0.3 percent, and the hedge absorbs about 85 percent of the stock's day-to-day variance, consistent with the correlation. Then the deposit-run scare hits: the stock falls 22 percent in three days while the ETF falls 3 percent. The basis moves 19 percentage points in three days — a size and speed of gap the historical correlation gave no warning of, because correlation measured over calm periods understates how far apart two related-but-different assets can move under an idiosyncratic shock.

P&L on the position: $-8{,}000{,}000 \times 0.22 = -1{,}760{,}000$. P&L on the hedge: $+8{,}000{,}000 \times 0.03 = 240{,}000$. Net loss: $1,520,000 — the hedge covered barely 14 percent of the actual move, not the 85 percent its historical correlation implied, because the shock was specific to a risk the ETF proxy doesn't carry.

Living with it

Basis risk can't be hedged away without moving to a more exact — and usually more expensive or less liquid — instrument, so the decision is really about how much idiosyncratic risk in the specific name you're comfortable leaving unhedged. If the position carries meaningful single-name risk that a sector proxy can't capture (concentrated exposure to one company's own balance sheet, as in the bank example), consider a partial single-name hedge — options on the stock itself, even if less liquid — to cover the tail the proxy can't reach, rather than relying on the proxy alone.