Why is no hedge ever perfect?
The corn hedge in the last lesson landed on exactly $4.50 both times because the numbers were chosen to. Real hedges do not land exactly. Three things get in the way, and each has a name worth knowing.
1. Basis risk
A hedge does not lock a price. It locks futures price plus basis:
Effective price = futures price sold + basis at lift, where basis = local cash − futures.
In the example: $4.60 + (−$0.10) = $4.50.
The farmer knew $4.60 back in June. What the farmer did not know was that November's local basis would be −$0.10. Suppose harvest-time basis came in at −$0.25 instead — a local glut, an expensive freight market, an elevator with no room. Then:
- Effective price = $4.60 − $0.25 = $4.35
- Shortfall versus plan: $0.15 × 50,000 = $7,500
— and the futures leg performed flawlessly throughout. Nothing broke.
Hedging does not eliminate risk. It converts price risk into basis risk: a much smaller, more local, more predictable exposure. That conversion is the product being sold, and stating it accurately is the difference between a hedging programme that survives contact with reality and one that gets oversold internally.
2. Cross-hedging
Sometimes no contract exists for what you actually own. A jet-fuel buyer hedges with heating oil or gasoil. A landlocked crude producer hedges with WTI or Brent. A corporate bond portfolio hedges with Treasury futures.
In each case the hedge instrument and the exposure are correlated, not identical, and every point of imperfect correlation is additional basis risk. Cross-hedges work right up until the relationship they depend on stops holding — which tends to be in exactly the conditions that made you want the hedge.
3. The hedge ratio, and the rounding
Contracts = (exposure to hedge ÷ contract size), adjusted for how sensitive the exposure is to the hedge instrument.
Commodity case, straightforward. 50,000 bushels ÷ 5,000 bushels = 10 contracts. But a 47,300-bushel crop is 9.46 contracts, and you may only trade 9 or 10. Round down and 2,300 bushels ride unhedged; round up and you are short 2,700 bushels you do not own — a small speculative position created purely by rounding.
Equity case, with a sensitivity adjustment.
- Portfolio value $1,250,000, beta to the S&P 500 of 1.2
- Index at 5,000, E-mini multiplier $50 → one contract = $250,000 notional
- Contracts = (1,250,000 × 1.2) ÷ 250,000 = 6 contracts short
That neutralises the portfolio's exposure to the market. What survives is everything specific to the individual holdings — which was the point, if the goal was to remove market risk and keep stock selection. But note what it assumes: that beta of 1.2. Betas drift, and relationships between a portfolio and an index change. Hedge ratios are maintained, not set.
The honest summary
A hedge trades a large, unbounded, unpredictable risk for a small, bounded, local one — and charges you the upside for the service. That is genuinely valuable and genuinely limited. Calling it "protection" is how hedging gets mis-sold.
None of this is a recommendation to hedge anything, or to use futures to do so. It is a description of how the mechanism behaves and where it leaks.
In the data
A hedge ratio estimated from history is a beta, and a beta is always measured against something. The two tables below are the same stock, Apple, over the same trailing 120 sessions, measured first against the S&P 500 and then against a Nasdaq-100 fund.
The reference you measure against is the entire hedging question: a ratio computed against the index you do not hedge with is not a smaller error, it is a different one. The window matters too. Each figure is the slope over the 120 sessions ending on its date, so the first and latest readings in each table differ even though nothing about the method changed.
Try it now
- Pick a portfolio value, then take a measured beta instead of assuming one. Treat your portfolio as having Apple's latest beta against the S&P 500, the last reading in the first table above, read the index level off the chart below, and compute the contract count that would neutralise the position at $50 per index point.
- Note the rounding: how much exposure is left over, and in which direction?
- Recompute with beta 0.8, then 1.5. The same portfolio value needs a very different contract count. That sensitivity is the maintenance burden nobody mentions in the pitch.