The problem: your positions are not independent
You hold six stocks. Each was researched separately, each was sized separately, each has its own thesis. It feels like six positions.
Then a bad inflation print lands and all six fall on the same morning, by different amounts but in the same direction. The diversification you thought you had did very little, because on days driven by macro news individual company stories stop mattering and almost everything trades as a proxy for the same thing: the market.
This is not a flaw in your stock selection. It is a structural fact about equities. A large share of any single stock's daily variance is explained by the market's move, and that share rises sharply during stress — exactly when you would most like it not to.
So the useful question is not "how many positions do I have?" but "if the index falls 1%, how much money do I lose?" Beta-weighting answers that, by expressing every position in the portfolio as an equivalent amount of a single benchmark.
What beta is, and what it is not
Beta is the historical sensitivity of a stock's returns to a benchmark's returns. It is a regression slope: run the stock's returns against the index's returns over some window, and beta is the gradient of the line.
- Beta 1.0 — on average moved with the index, one for one.
- Beta 1.6 — on average moved 1.6% for each 1% of index move. A high-beta growth name.
- Beta 0.5 — moved half as much. A utility or a consumer staple.
- Beta below zero — moved against the index. Rare and usually unstable.
Three things beta is emphatically not, all of which get assumed anyway:
It is not a constant. Beta is estimated over a window, and it changes as the window moves. A two-year beta and a three-month beta for the same stock can differ by half a point. Whichever you use, you are using a number from the past to size a hedge for the future.
It is not a measure of total risk. Beta captures only the part of a stock's movement explained by the index. Everything company-specific — the earnings miss, the recall, the fraud — is invisible to it. A beta-neutral portfolio is not a riskless portfolio; it has simply removed the market component.
It is not a prediction. A beta of 1.4 says how a stock behaved relative to the index, not how it will behave next Tuesday.
The method, in three steps
The arithmetic is straightforward. The care goes into the inputs.
- Convert each position to index-equivalent exposure. Multiply the position's market value by its beta. A $20,000 holding with a beta of 1.4 behaves, on market-driven days, like $28,000 of the index.
- Add them up. Sum the beta-weighted values across every position, signed — shorts are negative. The total is your net index-equivalent exposure: the single number that says how much index you effectively own.
- Size the hedge to offset it. Divide the exposure you want to remove by the value of one unit of your hedging instrument.
index-equivalent exposure = Σ (position value × beta) hedge size = exposure to remove ÷ value per hedge unit
Worked example: hedging a six-stock portfolio
A $180,000 book, beta-weighted to the S&P 500
| Position | Market value | Beta | Index-equivalent |
|---|---|---|---|
| Semiconductor A | $34,000 | 1.75 | $59,500 |
| Software B | $28,000 | 1.30 | $36,400 |
| Bank C | $31,000 | 1.15 | $35,650 |
| Industrial D | $26,000 | 1.05 | $27,300 |
| Staples E | $33,000 | 0.55 | $18,150 |
| Utility F | $28,000 | 0.40 | $11,200 |
| Total | $180,000 | — | $188,200 |
The headline is the gap between the two totals. The book holds $180,000 of stock but carries $188,200 of index exposure — an effective portfolio beta of 1.05. It is slightly more aggressive than simply owning the index, despite two-thirds of the names being defensive by reputation.
Look at where the exposure actually sits. Semiconductor A is 19% of the capital and 32% of the market risk. Utility F is 16% of the capital and 6% of the risk. By dollars the book looks evenly spread; by risk it is a concentrated bet on one high-beta name.
Sizing the hedge. With SPY at $580, a full hedge is:
188,200 ÷ 580 = 324 shares of SPY, short
That would leave the book roughly market-neutral: a 1% index fall costs about $1,882 on the stocks and gains about $1,882 on the short.
Hedging with futures instead of shares
Shorting an ETF is simple but ties up capital and needs borrow. Index futures are the usual alternative, and the arithmetic is the same with a different denominator: the notional value of one contract.
| Contract | Multiplier | Notional | Suits |
|---|---|---|---|
| E-mini S&P (ES) | $50 | $270,000 | Books of roughly $250k and up |
| Micro E-mini (MES) | $5 | $27,000 | Most private portfolios — ten times finer |
For the $188,200 of exposure above:
ES: 188,200 ÷ 270,000 = 0.70 contracts MES: 188,200 ÷ 27,000 = 6.97 → 7 contracts, short
One ES contract would be a 43% over-hedge — you would be net short the market while believing yourself neutral. Seven MES contracts land within half a percent of the target. This granularity problem is the single most common reason a hedge does not behave as expected, and it is why the micro contracts exist.
You usually do not want a full hedge
A full hedge removes the market component entirely. What remains is the company-specific performance of your six names relative to the index — which is a legitimate strategy, but it is a different one from the strategy you presumably chose these stocks for.
Fully hedged, your portfolio makes money only if your stocks beat the index, and makes nothing at all from the market rising. If your actual view is "I want to keep owning these, but I am nervous about the next fortnight", a partial hedge expresses that far better.
| Hedge | Exposure removed | SPY shares | MES contracts | Resulting stance |
|---|---|---|---|---|
| 25% | $47,050 | 81 | 1.7 → 2 | Still clearly long; takes the edge off. |
| 50% | $94,100 | 162 | 3.5 → 3 or 4 | Half the market risk, half the market upside. |
| 100% | $188,200 | 324 | 7.0 → 7 | Market-neutral. A pure stock-selection bet. |
And a hedge has a running cost. Short ETF positions pay borrow and any dividend; futures carry roll costs and margin. A hedge left on indefinitely is a slow, permanent drag on returns, which is an argument for treating it as a temporary position with a reason to exist and a condition for coming off — both written down when you put it on.
The assumption that fails when you need it most
Beta-weighting rests on one assumption: that the historical relationship between your stocks and the index continues to hold. Here is where that assumption breaks, in rough order of how much damage it does.
- Beta rises in a crash. Correlations converge toward 1 during severe selloffs. A portfolio hedged on calm-market betas of 1.05 can behave like a beta of 1.3 during the event, so the hedge covers less than it was sized to cover — and it under-covers precisely on the day it was bought for.
- It does not touch single-stock risk. A perfectly beta-neutral book still loses the full amount when one holding drops 30% on its own news. The hedge removes the market, not the company.
- Sector concentration hides inside the aggregate. A book of six technology names beta-weighted to the S&P is hedged against the S&P, not against technology underperforming the S&P. Six names, one sector, one story — the market hedge does not see it.
- The beta figure itself is a choice. Different windows, different data providers and different benchmarks produce meaningfully different betas for the same stock. Use one source and one window consistently, and treat the resulting exposure as an estimate with a range rather than a figure.
None of that makes the exercise pointless. Even an imperfect beta-weighted number is far better than the implicit assumption most people run on, which is that six positions equals six independent risks. Knowing you carry $188,200 of index exposure on a $180,000 book — and that a third of it sits in one stock — is worth having, whatever the error bars on each individual beta.