Hedging with futures: methodology for contracts, selection, and costs
Hedging with futures answers which equity index and sector futures, and how many contracts, can hedge the equity risk of a portfolio, what that costs, and how much risk remains afterward. This methodology describes what the code behind the Hedging page does.
Which contracts are available
The list of futures comes from the weekly Traders in Financial Futures report of the US regulator CFTC (futures only), and for commodities from the Disaggregated report of the same agency. The report gives each market's name, exchange, and open interest on Tuesday; it is published on Fridays and only covers markets with at least 20 reportable traders. It contains no prices.
- What is fixed in the code: sector or index, region, contract size, currency, and a fund that stands in for the returns of the underlying. The contract size deliberately does not come from the CFTC text, because that text contains occasional errors; the text is only used as a cross-check.
- Open interest: by default only contracts with at least 10,000 open contracts count; the threshold can be changed. If your own number of contracts exceeds 1% of open interest, you get a notice.
- Your own contracts: European contracts are not in the CFTC report. You enter them yourself with name, region, sector, contract size, currency, and proxy; they take precedence over any other choice.
- Futures prices are entered by you. Without a price the number of contracts stays empty; the page still calculates all weights, ratios, and the risk.
How the page chooses the hedge
Optimizer: from the allowed instruments, the page picks the hedge ratios that leave the portfolio's equities with the lowest volatility, based on a region plus sector model built from today's holdings. The look-through breaks every fund down into equities by region and sector for this. The instruments come from a ready-made set (all, several asset classes, bonds only, commodities only, world, USA, Europe, Asia) or from your own selection; in your own selection, a group without a check mark stays open. Limits are possible for the hedge as a whole (by default at most 100% of portfolio value), the net equity share afterward, the remainder per sector and per region, each single sector and country, each instrument, each group of instruments, and the maximum number of instruments for equities. If the limits cannot be met, the page says so instead of inventing a solution.
For comparison, the full hedge by rule: each region and sector cell goes to the matching future, first to one of your own contracts, otherwise to the region's sector future with enough open interest, otherwise to the region's broad index future; otherwise it stays open. If a sector's amount is smaller than half a contract, it goes through the broad future, because a number of contracts rounded to zero hedges nothing. The result shows the volatility of this rule next to that of the optimizer.
Without holdings data: if the funds' holdings lists are missing, the page loads them. If there are none, it works from returns: a regression of the portfolio's weekly returns on the returns of the proxies, with short sales only (no negative ratios). The selection goes step by step; each additional future has to reduce the variance by at least 1% of the starting value, at most three are preset (up to six can be chosen). The regression needs at least 60 common weeks.
- minimum variance hedge ratio for a single future (the portfolio's beta to the underlying)
- weekly returns of the portfolio and the underlying
Holdout and history: volatility before and after the hedge is shown for the whole period and for the last 52 weeks, with hedge ratios estimated only from the time before. A history chart shows what would have been left of the portfolio if the instruments had been sold at today's amounts the whole time, and next to it with adjustments over time. There, futures sold short earn the US dollar money market rate on the hedged amount. Scenarios show the portfolio and the hedge in historical crises and in assumed moves.
From hedge ratio to number of contracts
- share of portfolio value that the instrument hedges (from the optimizer or the rule)
- value of one contract in its currency; the amount is converted into that currency first
The number of contracts is rounded to whole contracts in the usual way. For the broad US index there is optionally the micro contract at one tenth of the size, so smaller portfolios can be matched more closely. Contract numbers can also be set by hand; the page then calculates the risk with those numbers.
Costs and tied-up capital
Once there is a number of contracts, costs come in. All defaults are assumptions you can change:
- Margin: 6% of contract value (exchanges set it for each contract).
- Margin calls: for a 10% rise in the underlying; that much cash should be on hand, because a short future settles losses daily when prices rise.
- Interest minus dividends: the futures price contains interest until expiration minus expected dividends (preset at 1.5% a year). The seller receives this difference over the life of the contract.
- Trading and rolling: 2.5 units of the contract currency per contract and trade (commission and half the bid-ask spread), rolling four times a year for quarterly contracts.
Instead of futures, the same hedge can be implemented with ETFs. Short sale of an ETF on the sector or index: borrow fee 0.5% a year, no interest on the sale proceeds, margin 50% (customary initially in the US), trading costs 0.05%. Purchase of an inverse ETF with leverage 1, 2, or 3: ongoing costs 0.9% a year and the path effect of daily rebalancing, measured over a selectable holding period (three months by default) across all periods of the price history, at least 30 of them.
Where the risk comes from
Every run shows a risk tree before and after the hedge: equities, government bonds, other bonds, commodities, other, currencies, and the idiosyncratic risk of single securities. It is based on monthly returns of the asset classes in the account currency, at least 24 months. Each source's contribution can be calculated with three methods: Euler (the contributions add up to the total risk), isolated (each source on its own), and incremental (what disappears if the source is removed). Equities can be broken down to single stocks, by region, country, sector, industry, or fund.
In addition to equities, the page can suggest currency futures for the foreign currencies of the holdings, US Treasury futures for the interest rate risk of US dollar bonds (with assumed durations of the contracts), and commodity futures for commodity funds, which are matched to a contract by how closely their weekly returns moved together over the last three years. The price of the currency futures is approximated from the spot rate.
What research shows about hedging
Ederington (1979) applied portfolio theory to hedging: the best hedge ratio is not simply one to one, but the one at which the combined volatility of the position and the future is lowest, and its effectiveness is measured by the share of variance removed. This is exactly how the page calculates without holdings data, and the optimizer also looks for the lowest volatility. Figlewski (1984) showed for S&P 500 futures that basis risk determines how much volatility is left: broad portfolios can be hedged well, small ones or ones that differ strongly from the index much less so. That is why the page also measures effectiveness in held-out weeks and shows the remaining risk. The basics of contract value, margin, and rolling are covered in Hull (2021).
Limitations
- No futures prices: the prices belong to the exchanges; without a price you enter, there is no number of contracts. Returns come from proxies (funds on the same index), not from the futures themselves.
- Only US contracts in the list; European and Asian contracts only as your own entries.
- Estimation error: hedge ratios from the past do not reliably hold in the future; the holdout in the held-out weeks shows how stable they were.
- Cost assumptions: margin, dividends, borrow fees, and trading costs are defaults, not quotes from a broker.
- Not advice: the page calculates examples. Before trading, check contract size, expiration, and margin with the exchange or your broker.
Frequently asked questions
How many futures do I need to hedge my portfolio?
Portfolio value times beta to the index, divided by the value of one contract (contract size times price). A portfolio of 500,000 with a beta of 1 and a contract worth 250,000 gives two contracts. The page calculates this for each future and rounds to whole contracts.
Why is the hedge not perfect?
Because the portfolio and the future's underlying do not move exactly together (basis risk) and because the number of contracts is rounded. The broader the portfolio and the closer to the index, the less volatility remains.
What does hedging with futures cost?
Mainly trading and roll costs and the capital tied up for margin and margin calls. The difference between interest and dividends in the futures price goes to the seller. The page shows all parts with defaults you can change.
Futures or inverse ETF?
Futures are usually cheaper and more precise, but need a futures account and margin calls. Inverse ETFs are easy to buy, charge ongoing fees, and drift away from the multiple of the index return over longer periods because they rebalance daily.
Why does the page show no number of contracts?
Because the app has no futures prices; they belong to the exchanges. Enter the current price of the contract, and the page calculates the number of contracts and the costs. Hedge ratios, remaining risk, and the risk tree are shown even without a price.
Literature
- Ederington, Louis H. (1979): The Hedging Performance of the New Futures Markets. The Journal of Finance 34 (1), pp. 157-170. doi:10.1111/j.1540-6261.1979.tb02077.x
- Figlewski, Stephen (1984): Hedging Performance and Basis Risk in Stock Index Futures. The Journal of Finance 39 (3), pp. 657-669. doi:10.1111/j.1540-6261.1984.tb03654.x
- Hull, John C. (2021): Options, Futures, and Other Derivatives. Pearson, 11th edition (Global Edition).
This page describes how Portfolio Lab calculates; past results are not a forecast. All terms in the glossary.