Observed protection
Annualised volatility was 27.7% unhedged and 23.9% with IWM in the two-year test, using hedge sizes estimated before each quarter.
Performance evidence ↓Hedge Lab · NASDAQ · NKSH
Hedge research · Financial Services
NKSH Hedge comparison
National Bankshares Inc
Banks - Regional · USD
Terminal chart ↗Week ending · hedge-v1
Selected comparison · IWM13.8%lower volatility in the test period
7 of 8 quarters improved · Lower cost
Compare all hedges ↓- Unhedged volatility
- 27.7%annualised
- Hedged volatility
- 23.9%annualised · after assumed costs
- Extra annual expense
- 2.90%of long notional · assumed terms
- Current hedge size
- 0.60×short notional / long notional
- Worst-week improvement
- 43.1%same six worst stock weeks
The case for further reviewWorth reviewing for risk reduction
Strategy, evidence and risks ↗Execution not yet verified · Based on the selected hedge and assumptions.
Research summary
Protection and trade-offs
Hedge comparisons are separate from investment ratings. The figures below use 2024-09-13 to 2026-09-11 and the assumptions shown on this page.
Cost and return
The assumed extra expense averaged 2.90% a year of the long position. Net profit was 36.9 percentage points lower than the unhedged position over the test.
Cost breakdown ↓Risk that remains
23.9% annualised volatility remained after hedging. Company earnings, valuation changes and events can move the stock independently of this ETF.
Remaining exposure ↓Borrow availability is unverified unless a current broker observation is shown below. Historical costs are assumptions. Short sales can lose more than the initial position.
Market and sector hedges
Compare the alternatives
Every candidate uses the same dates, account model and cost rates. Each hedge size was estimated using only the preceding year.
| Hedge | Current size | Volatility | Reduction | Maximum drawdown | Net P&L | Extra annual expense | Screen |
|---|---|---|---|---|---|---|---|
| UnhedgedNKSH | — | 27.7% | — | -24.0% | 51.0% | — | Reference |
| SPYUS large companies | 0.23× | 27.2% | 2.0% | -25.2% | 17.5% | 2.56% | Comparison only
|
| QQQNasdaq 100 | 0.03× | 27.5% | 0.9% | -24.9% | 28.0% | 1.26% | Comparison only
|
| IWMSelected | 0.60× | 23.9% | 13.8% | -18.5% | 14.0% | 2.90% | Qualifies |
| XLFFinancials | 0.63× | 25.7% | 7.4% | -36.6% | 10.6% | 3.37% | Comparison only
|
P&L is cumulative profit or loss as a percentage of original long notional. It is not annualised. Maximum drawdown is measured against the simulated account’s own running peak. Worst-week improvement above 100% means the hedge turned the average loss across those weeks into a gain.
Position and expenses
Assumptions and costs
Change the hedge and cost rates to update the entire comparison. A basis point is 0.01%. The long position starts fully funded.
Long position$100,000NKSH
Illustrative short$59,865IWM · current ratio
Gross exposure$159,865long plus short notionals
Net dollar exposure$40,135not a measure of market neutrality
The chosen long position exceeds 1% of the stock’s median weekly turnover. Actual spreads, trading capacity and short availability need separate checking.
No verified broker quote. The borrow rate is an editable assumption. Check availability, fees, collateral and recall terms with your broker.
Expense breakdown
| Expense | Test total | Annual average | USD |
|---|---|---|---|
| Stock borrowing | 5.32% | 2.67% | $5,316 |
| Short proceeds interest (credit) | 0.00% | 0.00% | $0 |
| Long financing | 0.00% | 0.00% | $0 |
| Trading and slippage | 0.96% | 0.48% | $962 |
| Net expense | 6.28% | 3.15% | $6,278 |
Percentages are of original long notional. Interest on short proceeds is a credit. Distributions are embedded in the adjusted-return proxy and are not charged again. Rates are held constant throughout the historical test.
Borrow-rate sensitivity
| Borrow rate | Net P&L | Versus unhedged |
|---|---|---|
| 0.0% | 19.4% | -31.6 pp |
| 1.0% | 17.6% | -33.4 pp |
| 3.0% | 14.0% | -36.9 pp |
| 5.0% | 10.5% | -40.5 pp |
| 10.0% | 1.6% | -49.3 pp |
The hedged test earned less than the unhedged position even with no borrowing fee. This is a return comparison; it does not price the value of protection.
Later-period evidence
Performance and drawdowns
52 observations set the first hedge size. Each subsequent quarter uses a fresh estimate from the preceding year. The selected hedge is ranked after these tests.
Accumulated profit and loss
Decline from account peak
Volatility by quarter
Each quarter contains 13 weekly returns, annualised for comparison. Short samples are uncertain.
Hedge size over time
Estimated before each quarter. The current proposed ratio can differ from the last historical test block.
Quarterly test values
| Through | Hedge ratio | Unhedged volatility | Hedged volatility | Reduction |
|---|---|---|---|---|
| 2024-12-13 | 1.15× | 25.6% | 22.9% | 10.5% |
| 2025-03-14 | 0.81× | 28.8% | 21.9% | 23.8% |
| 2025-06-13 | 0.97× | 22.2% | 17.6% | 20.6% |
| 2025-09-12 | 0.85× | 37.3% | 28.7% | 23.0% |
| 2025-12-12 | 0.87× | 32.1% | 32.8% | -2.2% |
| 2026-03-13 | 0.86× | 21.4% | 19.1% | 11.0% |
| 2026-06-12 | 0.75× | 25.3% | 17.9% | 29.3% |
| 2026-09-11 | 0.84× | 29.1% | 27.4% | 5.7% |
Observed losses and illustrative shocks
Downside and rebound risk
A useful hedge must be assessed when the stock falls independently and when the short position rises. Historical observations and hypothetical shocks are shown separately.
The stock’s six worst test weeks
| Week | Unhedged | Hedged | ETF return |
|---|---|---|---|
| 2026-05-15 | -7.26% | -5.57% | -2.31% |
| 2025-04-04 | -6.28% | 3.01% | -9.61% |
| 2025-08-01 | -5.87% | -2.35% | -4.22% |
| 2024-12-27 | -5.46% | -5.68% | 0.21% |
| 2026-04-24 | -5.41% | -5.70% | 0.32% |
| 2024-12-06 | -5.24% | -3.90% | -1.22% |
Includes assumed expenses. Dates are chosen from the unhedged results and reused for the hedge.
Market conditions in the test
| SPY condition | Weeks | Unhedged mean | Hedged mean |
|---|---|---|---|
| Falling market | 44 | -0.31% | 1.00% |
| Rising market | 60 | 1.07% | -0.50% |
| Market falls 2% or more | 10 | -1.58% | 1.28% |
| Market rises 2% or more | 15 | 2.87% | -0.48% |
Average weekly P&L. Conditions overlap. At least three observed weeks are required.
Three-month scenarios
| Scenario | Stock total return | ETF total return | Unhedged P&L | Hedged P&L |
|---|---|---|---|---|
| Both fall | -20% | -15% | -20.2% | -11.8% |
| Both rise | 20% | 15% | 19.8% | 10.2% |
| Stock falls alone | -25% | 0% | -25.2% | -25.7% |
| Hedge rises alone | 0% | 20% | -0.2% | -12.8% |
| Long falls, short rises | -20% | 20% | -20.2% | -32.7% |
| Both unchanged | 0% | 0% | -0.2% | -0.8% |
Explicit assumptions, with no assigned probabilities. These shocks include distributions; they do not represent a price forecast. Costs cover 91 days and one entry and exit.
Exposure after hedging
Remaining risks
23.9% annualised volatility remained after hedging. Company earnings, valuation changes and events can move the stock independently of this ETF.
Company and trading risks
- Business performance
- Results, margins, financing and valuation can change without a matching ETF move. A stock-specific loss remains possible.
- Holdings overlap
- IWM may hold NKSH. Current holdings weights are not verified here. Part of the short may offset the stock itself.
- Currency and sector
- USD quotation does not remove foreign business exposure or ADR currency risk. A sector ETF can also have a different business mix.
- Borrow and margin
- Fees can rise, borrow can be recalled and collateral requirements can change. A hedge can lose money while the long also falls.
- Relationship changes
- Current correlation is 0.36. The approximate 95% beta interval is 0.17 to 1.02. These estimates are uncertain.
Overlap between hedge alternatives
| Correlation | SPY | QQQ | IWM | XLF |
|---|---|---|---|---|
| SPY | 1.00 | 0.94 | 0.74 | 0.60 |
| QQQ | 0.94 | 1.00 | 0.68 | 0.38 |
| IWM | 0.74 | 0.68 | 1.00 | 0.54 |
| XLF | 0.60 | 0.38 | 0.54 | 1.00 |
Latest 52 weekly ETF returns. Values close to +1 mean similar observed movements. Combining these hedges may provide less diversification than their different names suggest. This does not measure investor crowding.
Review conditions
Changes to watch
Recheck the comparison after new weekly data, company results, a material price correction or a change in borrowing terms.
Recheck if the rolling beta interval crosses zero or the correlation falls below 0.25.
The screen requires at least six improved quarters. A new failing quarter can remove the opportunity.
Check the current borrow rate, available quantity and recall terms with the broker before using a short.
A hedge does not repair an investment case. Read any published valuation separately.
Audit record
Source dates and chart values
Edition 49a879bb-aa6e-4f2b-a80d-95092bf172d5 · Source through 2026-09-11 · Built 15 Sep 2026, 09:22 BST · Adjusted weekly closes. Comparison calculations and assumptions are included below.
All weekly chart values
| Week | Unhedged P&L | Hedged P&L | Unhedged drawdown | Hedged drawdown | Hedge ratio |
|---|---|---|---|---|---|
| 2024-09-13 | 0.00% | 0.00% | 0.00% | 0.00% | — |
| 2024-09-20 | -0.72% | -3.42% | -0.72% | -3.42% | 1.151× |
| 2024-09-27 | -4.29% | -6.80% | -4.29% | -6.80% | 1.151× |
| 2024-10-04 | -3.61% | -5.57% | -3.61% | -5.57% | 1.151× |
| 2024-10-11 | -4.01% | -7.16% | -4.01% | -7.16% | 1.151× |
| 2024-10-18 | -1.52% | -7.02% | -1.52% | -7.02% | 1.151× |
| 2024-10-25 | -5.82% | -7.94% | -5.82% | -7.94% | 1.151× |
| 2024-11-01 | -4.02% | -6.25% | -4.02% | -6.25% | 1.151× |
| 2024-11-08 | 4.32% | -8.04% | 0.00% | -8.04% | 1.151× |
| 2024-11-15 | 4.56% | -3.21% | 0.00% | -3.21% | 1.151× |
| 2024-11-22 | 5.35% | -7.68% | 0.00% | -7.68% | 1.151× |
| 2024-11-29 | 7.98% | -6.62% | 0.00% | -6.62% | 1.151× |
| 2024-12-06 | 2.74% | -10.52% | -4.85% | -10.52% | 1.151× |
| 2024-12-13 | 5.40% | -5.11% | -2.39% | -5.11% | 1.151× |
| 2024-12-20 | 1.65% | -5.28% | -5.86% | -5.28% | 0.812× |
| 2024-12-27 | -3.81% | -10.97% | -10.92% | -10.97% | 0.812× |
| 2025-01-03 | -5.68% | -13.63% | -12.65% | -13.63% | 0.812× |
| 2025-01-10 | -10.49% | -15.73% | -17.10% | -15.73% | 0.812× |
| 2025-01-17 | -0.74% | -9.26% | -8.07% | -9.26% | 0.812× |
| 2025-01-24 | -0.06% | -9.80% | -7.44% | -9.80% | 0.812× |
| 2025-01-31 | 0.11% | -8.89% | -7.29% | -8.89% | 0.812× |
| 2025-02-07 | -3.39% | -12.27% | -10.53% | -12.27% | 0.812× |
| 2025-02-14 | -2.66% | -11.57% | -9.85% | -11.57% | 0.812× |
| 2025-02-21 | -6.37% | -12.39% | -13.29% | -12.39% | 0.812× |
| 2025-02-28 | -4.15% | -9.06% | -11.24% | -9.06% | 0.812× |
| 2025-03-07 | -6.36% | -8.02% | -13.28% | -8.02% | 0.812× |
| 2025-03-14 | -7.71% | -8.21% | -14.53% | -8.21% | 0.812× |
| 2025-03-21 | -7.16% | -8.39% | -14.02% | -8.39% | 0.973× |
| 2025-03-28 | -7.89% | -7.58% | -14.70% | -7.58% | 0.973× |
| 2025-04-04 | -14.17% | -4.57% | -20.51% | -4.57% | 0.973× |
| 2025-04-11 | -17.92% | -10.09% | -23.98% | -10.09% | 0.973× |
| 2025-04-18 | -15.14% | -8.48% | -21.41% | -8.48% | 0.973× |
| 2025-04-25 | -13.75% | -11.13% | -20.12% | -11.13% | 0.973× |
| 2025-05-02 | -11.67% | -12.30% | -18.19% | -12.30% | 0.973× |
| 2025-05-09 | -11.09% | -11.95% | -17.66% | -11.95% | 0.973× |
| 2025-05-16 | -8.95% | -14.24% | -15.68% | -14.24% | 0.973× |
| 2025-05-23 | -9.85% | -11.82% | -16.51% | -11.82% | 0.973× |
| 2025-05-30 | -10.30% | -13.53% | -16.92% | -13.53% | 0.973× |
| 2025-06-06 | -5.34% | -11.87% | -12.33% | -11.87% | 0.973× |
| 2025-06-13 | -9.17% | -14.38% | -15.89% | -14.38% | 0.973× |
| 2025-06-20 | -11.38% | -17.01% | -17.93% | -17.01% | 0.847× |
| 2025-06-27 | -4.35% | -12.56% | -11.41% | -12.56% | 0.847× |
| 2025-07-04 | 5.88% | -5.37% | -1.94% | -5.37% | 0.847× |
| 2025-07-11 | 4.46% | -6.32% | -3.26% | -6.32% | 0.847× |
| 2025-07-18 | 0.82% | -10.25% | -6.63% | -10.25% | 0.847× |
| 2025-07-25 | 2.41% | -9.49% | -5.16% | -9.49% | 0.847× |
| 2025-08-01 | -3.46% | -11.84% | -10.60% | -11.84% | 0.847× |
| 2025-08-08 | 3.32% | -7.24% | -4.32% | -7.24% | 0.847× |
| 2025-08-15 | 1.07% | -12.16% | -6.40% | -12.16% | 0.847× |
| 2025-08-22 | 8.98% | -7.17% | 0.00% | -7.17% | 0.847× |
| 2025-08-29 | 11.59% | -4.73% | 0.00% | -4.73% | 0.847× |
| 2025-09-05 | 11.01% | -6.29% | -0.52% | -6.29% | 0.847× |
| 2025-09-12 | 7.77% | -9.79% | -3.42% | -9.79% | 0.847× |
| 2025-09-19 | 10.06% | -9.51% | -1.37% | -9.51% | 0.875× |
| 2025-09-26 | 9.21% | -9.82% | -2.13% | -9.82% | 0.875× |
| 2025-10-03 | 4.08% | -16.63% | -6.73% | -16.63% | 0.875× |
| 2025-10-10 | -0.58% | -18.48% | -10.91% | -18.48% | 0.875× |
| 2025-10-17 | 3.54% | -16.49% | -7.22% | -16.49% | 0.875× |
| 2025-10-24 | 4.71% | -17.53% | -6.16% | -17.53% | 0.875× |
| 2025-10-31 | 2.96% | -18.21% | -7.73% | -18.21% | 0.875× |
| 2025-11-07 | 7.05% | -12.53% | -4.07% | -12.53% | 0.875× |
| 2025-11-14 | 7.61% | -10.52% | -3.57% | -10.52% | 0.875× |
| 2025-11-21 | 8.88% | -8.62% | -2.43% | -8.62% | 0.875× |
| 2025-11-28 | 8.79% | -13.64% | -2.51% | -13.64% | 0.875× |
| 2025-12-05 | 9.67% | -13.53% | -1.72% | -13.53% | 0.875× |
| 2025-12-12 | 22.29% | -2.03% | 0.00% | -2.03% | 0.875× |
| 2025-12-19 | 25.14% | 1.52% | 0.00% | 0.00% | 0.860× |
| 2025-12-26 | 23.28% | -0.61% | -1.48% | -2.09% | 0.860× |
| 2026-01-02 | 20.74% | -2.30% | -3.51% | -3.75% | 0.860× |
| 2026-01-09 | 23.31% | -3.73% | -1.46% | -5.17% | 0.860× |
| 2026-01-16 | 24.67% | -4.26% | -0.38% | -5.69% | 0.860× |
| 2026-01-23 | 29.88% | 1.21% | 0.00% | -0.30% | 0.860× |
| 2026-01-30 | 30.71% | 3.66% | 0.00% | 0.00% | 0.860× |
| 2026-02-06 | 37.65% | 8.78% | 0.00% | 0.00% | 0.860× |
| 2026-02-13 | 35.59% | 7.34% | -1.50% | -1.33% | 0.860× |
| 2026-02-20 | 36.30% | 7.45% | -0.98% | -1.22% | 0.860× |
| 2026-02-27 | 34.06% | 6.20% | -2.61% | -2.37% | 0.860× |
| 2026-03-06 | 32.62% | 8.18% | -3.66% | -0.56% | 0.860× |
| 2026-03-13 | 32.83% | 9.81% | -3.50% | 0.00% | 0.860× |
| 2026-03-20 | 29.17% | 7.29% | -6.17% | -2.29% | 0.750× |
| 2026-03-27 | 28.80% | 6.61% | -6.43% | -2.92% | 0.750× |
| 2026-04-03 | 30.99% | 6.23% | -4.84% | -3.27% | 0.750× |
| 2026-04-10 | 33.32% | 5.52% | -3.15% | -3.90% | 0.750× |
| 2026-04-17 | 34.56% | 2.55% | -2.25% | -6.61% | 0.750× |
| 2026-04-24 | 29.14% | -3.14% | -6.18% | -11.80% | 0.750× |
| 2026-05-01 | 31.24% | -1.80% | -4.66% | -10.58% | 0.750× |
| 2026-05-08 | 34.15% | -0.25% | -2.54% | -9.16% | 0.750× |
| 2026-05-15 | 26.89% | -5.82% | -7.82% | -14.24% | 0.750× |
| 2026-05-22 | 29.30% | -5.49% | -6.07% | -13.94% | 0.750× |
| 2026-05-29 | 30.31% | -5.92% | -5.33% | -14.33% | 0.750× |
| 2026-06-05 | 29.20% | -4.81% | -6.14% | -13.31% | 0.750× |
| 2026-06-12 | 33.60% | -3.47% | -2.95% | -12.09% | 0.750× |
| 2026-06-19 | 30.76% | -7.32% | -5.01% | -15.60% | 0.844× |
| 2026-06-26 | 37.44% | -1.90% | -0.16% | -10.67% | 0.844× |
| 2026-07-03 | 34.49% | -4.27% | -2.30% | -12.82% | 0.844× |
| 2026-07-10 | 32.76% | -5.60% | -3.56% | -14.03% | 0.844× |
| 2026-07-17 | 34.62% | -3.23% | -2.20% | -11.87% | 0.844× |
| 2026-07-24 | 34.76% | -2.32% | -2.10% | -11.04% | 0.844× |
| 2026-07-31 | 45.56% | 8.43% | 0.00% | -1.25% | 0.844× |
| 2026-08-07 | 49.27% | 9.09% | 0.00% | -0.66% | 0.844× |
| 2026-08-14 | 49.98% | 8.76% | 0.00% | -0.96% | 0.844× |
| 2026-08-21 | 47.01% | 7.16% | -1.98% | -2.42% | 0.844× |
| 2026-08-28 | 46.50% | 7.78% | -2.32% | -1.85% | 0.844× |
| 2026-09-04 | 49.26% | 10.42% | -0.48% | 0.00% | 0.844× |
| 2026-09-11 | 50.98% | 14.04% | 0.00% | 0.00% | 0.844× |
Read the hedge playbook · NKSH / IWM
NKSH / IWM
The case, the trade-offs and the work still to do.
Data through 2026-09-11 · Test 2024-09-13 to 2026-09-11 · USD
Research conclusion
Worth reviewing for risk reduction
The selected hedge meets the historical protection and consistency checks at these cost assumptions. It is a starting point for reviewing risk control, not evidence of an expected investment return.
01
The strategy
The structure keeps a long position in NKSH and adds a short position in IWM to offset part of its broader market or sector movement.
Shorting means borrowing and selling the ETF, then buying it back later. A fall in the ETF helps the short; a rise costs money. The long position still needs its own investment case.
- Long · NKSH
- $100,000Fully funded starting position
- Short · IWM
- $59,8650.60× long notional · latest estimate
- Gross exposure
- $159,865Long plus short · not cash required
No linked company valuation is available in this comparison. The reason to own the long must be established separately.
The 104-week test uses sizes fitted before each quarter. The current illustration uses the latest 52-week fit. It is not a position instruction, a market-neutral designation or a forecast.
02
The evidence for protection
The possible benefit is risk reduction while retaining the long. No expected return or persistent trading edge has been established.
| Check | Observed result | Research rule |
|---|---|---|
| VolatilityMeets rule | 13.8% reduction27.7% unhedged; 23.9% hedged. Annualised from the same weekly test. | At least 10% lower |
| Difficult weeksMeets rule | -5.92% → -3.37%Same six worst unhedged weeks. This average is not a maximum loss. | A better average outcome |
| ConsistencyMeets rule | 7 / 8 quartersHedge sizes were estimated before each quarter; short samples remain uncertain. | At least six improved quarters |
| Current relationshipMeets rule | 0.36 correlationApproximate 95% beta interval 0.17 to 1.02. This describes estimation uncertainty, not a limit on losses. | Correlation ≥ 0.25; beta interval above zero; size ≥ 0.10× |
Quarter values
| Quarter through | Volatility reduction |
|---|---|
| 2024-12-13 | 10.48% |
| 2025-03-14 | 23.80% |
| 2025-06-13 | 20.60% |
| 2025-09-12 | 22.98% |
| 2025-12-12 | -2.15% |
| 2026-03-13 | 11.04% |
| 2026-06-12 | 29.30% |
| 2026-09-11 | 5.73% |
Above zero means lower volatility. Eight observations are not eight independent proofs of an edge. The ETF is selected after the comparisons; the current active-stock universe and selection process can bias the apparent results.
Terms used in the playbook
- Volatility
- The size of the weekly swings, expressed on an annual scale. It measures variation, not the probability or maximum size of a loss.
- Drawdown
- The decline from the simulated account’s previous peak. A hedge can lower volatility while making this decline deeper.
- Correlation and beta
- Correlation describes how closely weekly returns moved together, from −1 to +1. Beta estimates the size of the long’s move relative to the ETF. Neither establishes a cause or guarantees a future offset.
- Notional and gross exposure
- Notional is the dollar size of a position. Gross exposure adds the long and short sizes. It is different from the cash invested, the collateral required or the loss possible.
03
Upside, downside and cost
Protection works when the short offsets a long-position loss. It can surrender gains in a broad rally and add to losses when the long falls while the ETF rises.
Unhedged test P&L$50,982
Hedged test P&L$14,039
Extra annual expense$2,896
P&L is the total across the two-year test, after assumed costs. Expense is an annual average at the same long notional. These are different measures; their difference is not the price of an insurance contract.
Rates used: 3.00% annual borrow, 0.00% interest on short proceeds, 0.00% long financing, and 10.0 basis points per dollar traded each way. At the current short size, a one-percentage-point increase in annual borrow adds about $599 a year if that size is maintained.
Three-month scenarios in dollars
91-day illustrations using the current hedge size, one entry and exit, and the selected cost rates. Both returns include distributions. No probabilities are assigned; these are not best- or worst-case limits.
| Scenario and assumptions | Unhedged | Hedged |
|---|---|---|
| Both fallLong -20% · ETF -15%A falling ETF generates a gain on the short that may offset some of the long loss. | $-20,180 | $-11,758 |
| Both riseLong 20% · ETF 15%A rising ETF loses money on the short. Market-led upside can be surrendered. | $19,780 | $10,224 |
| Stock falls aloneLong -25% · ETF 0%A company-specific or basket-specific loss gets no gross offset from an unchanged ETF. | $-25,175 | $-25,742 |
| Hedge rises aloneLong 0% · ETF 20%The short loses even though the long is unchanged. Cash may be required before either position is closed. | $-200 | $-12,752 |
| Long falls, short risesLong -20% · ETF 20%Both positions lose. The hedge provides no protection against this divergence. | $-20,180 | $-32,732 |
| Both unchangedLong 0% · ETF 0%With no movement, borrowing and trading expenses still matter; any assumed interest credit offsets them. | $-200 | $-767 |
Losses can exceed these illustrations. The weakest displayed scenario produces $-32,732 of hedged P&L. It is not a loss limit. Short losses have no fixed upper bound, and cash can be required before a long-position gain is realised.
04
The simpler alternative
Holding $86,160 in NKSH and leaving $13,840 in cash would have matched the hedged test’s 23.9% annualised volatility, without a short.
| Historical comparison | Volatility | Maximum drawdown | Test P&L |
|---|---|---|---|
| Full long | 27.7% | -24.0% | $50,982 |
| Long plus short | 23.9% | -18.5% | $14,039 |
| 86.2% long, rest cash | 23.9% | -20.9% | $43,926 |
The hedged account earned $29,887 less than this reduced position over the test. The comparison helps assess whether retaining more company exposure justified the short.
Hindsight diagnostic. The reduced position is chosen using the full test-period volatility, so this is not a strategy selected in advance. It keeps the same starting capital, scales long trading and financing costs with position size, resets the smaller long weekly and assumes no interest on cash. Equal historical volatility does not mean equal future risk.
05
Less obvious trade-offs
A smoother path can still lose more
Maximum drawdown was -24.0% unhedged and -18.5% hedged. The hedge made 2 of the six worst unhedged weeks worse. Assess loss size and timing alongside volatility.
Further work Set a loss and cash requirement that the portfolio can withstand, including overnight gaps.
Protection removes some upside too
Hedged P&L was 36.9 percentage points lower over the test. That difference includes the ETF short and extra costs; it is not just the borrowing bill.
Further work Identify the company-specific reason to retain the long. Market or sector upside may be the very exposure the hedge removes.
Dollar balance is not risk balance
The current fit implies 0.60 dollars short per dollar long. Gross exposure is 1.60 times the long position. Neither a balanced dollar amount nor an estimated beta establishes future market neutrality.
Further work Check other sector, style and currency exposures; allow for beta changing after results or corporate actions.
Different hedge names can hide the same risk
Similar ETF returns do not establish independent protection. An ETF may also own the long holding; that overlap has not been verified here.
Further work Inspect current ETF holdings before combining shorts. Check whether the hedge offsets the business exposure you wanted to retain.
06
Practical feasibility
Execution not yet verified
No current broker observation is recorded. The ETF is a research candidate; availability, position capacity, margin and borrowing terms remain unverified.
The long size exceeds 1% of median weekly turnover. Actual trading capacity requires review.
Confirm available borrow quantity, spreads on both legs, margin and collateral, and how the short would be closed after a recall. Short-sale proceeds are not assumed to be freely spendable. Weekly closes cannot establish intraday liquidity or the cash needed through a squeeze.
Do not use a hedge to preserve a broken investment case. If the concern is company-specific rather than market-wide, reducing the long can address a risk the ETF cannot offset.
07
Review plan
- New weekly data
- Recheck the four evidence tests above. A failure calls for review; it is not an automatic trade instruction.
- New results or a company event
- Revisit the reason to own the long and whether the ETF still captures the risk being hedged.
- Borrow fee, margin or availability changes
- Re-run costs and cash requirements. A recall can force the short to close at an unfavourable time.
- Hedge size review
- The test refits every 13 weeks and resizes dollar positions weekly. A different trading schedule needs its own costs and validation.
This is a research review plan, not an automated entry, exit or rebalancing instruction. Define the investment thesis, acceptable loss, cash reserve and review date before deciding whether the structure fits a portfolio.
Method and limits
Data and calculations
Every result is a historical research comparison. Borrow availability, trading costs and future protection require separate checks.
Research rules
- Universe
- Active US-listed equities quoted in USD. A $1 minimum adjusted close and $5 million median weekly turnover over 13 weeks with valid volume in every week. This is a liquidity screen, not an execution guarantee.
- Candidate hedges
- SPY, QQQ and IWM, plus the stock’s sector ETF when available. These are a small, predefined comparison set. Current sector membership is used.
- History
- 157 consecutive Friday observations: 52 weeks to estimate the first hedge and 104 weeks to test it. No missing weeks are interpolated. Moves above 60% and unresolved corporate-action repair errors require review.
- Hedge size
- Covariance of stock and ETF returns divided by the variance of ETF returns, estimated using the preceding 52 weeks. Re-estimated every 13 weeks and held for the next test block. Negative ratios are set to zero; shorts are capped at 150% of long notional.
- Publication criteria
- At least 10% lower test-period volatility, a better average result in the stock’s six worst weeks, and lower volatility in at least six of eight quarters. Current correlation must be at least 0.25; the approximate 95% beta interval must be above zero; the current short ratio must be at least 0.10.
- Cost categories
- “Higher cost” means the extra assumed expense exceeds 3% a year of long notional. This is a transparent research convention, not a measure of expected profit or a Buy/Sell rating.
Return and cost basis
- Stored prices
- Stored weekly closes are adjusted for corporate actions and distributions. Returns are treated as total-return proxies. They are not an independently reconciled distribution ledger. Dividends are not added or subtracted again.
- Price consistency
- Where daily history exists, weekly returns are compared with the same completed exchange sessions. If differences exceed 0.10 percentage point, a complete daily closing-session history is used for the entire instrument, or the instrument is withheld. Individual missing sessions are never filled. This checks consistency between stored histories; it does not independently verify every market price.
- Account model
- The long position starts fully funded. Each week its dollar notional is reset to the original size; the short is reset to the chosen ratio. Profits and losses accumulate in cash. Charts show P&L as a percentage of original long notional, not a compounded return on a short ETF.
- Expenses
- Default assumptions: 3% annual stock borrow, no interest on short proceeds, no long financing, and 10 basis points (0.10%) per dollar traded each way. Entry, weekly resizing and final close-out are charged. Annual costs use 365.25 days. Taxes, margin changes and recall costs are not modelled.
- Risk measures
- Annualised volatility is sample standard deviation of weekly P&L divided by original long notional, multiplied by √52. Drawdown is the decline from the simulated account’s running equity peak. Both approaches use the same six worst unhedged test weeks for the downside comparison.
- Current borrowing
- Broker observations have a source, date and expiry. They are separate from assumed historical costs. Availability can change or a loan can be recalled. A current quote is not evidence of historical borrow availability.
- Scenario assumptions
- Three-month total-return shocks are illustrations, not forecasts. They include distributions, use the current ratio, and charge one entry, one exit and 91 days of financing and borrow. They do not assume weekly resizing within the shock.
Selection and uncertainty
Each hedge ratio is tested on later observations. The dashboard then selects among those completed comparisons, so its ranking is retrospective. It is not an independently validated trading strategy. The universe contains currently active stocks, and excludes securities without sufficient history. Historical returns can be revised by providers.
The beta interval uses ordinary least squares and ±1.96 standard errors. It assumes independent errors with constant variance. Weekly histories cannot capture intraday margin calls, gaps or the ability to trade at a given price.
Remaining exposure
A sector or market ETF may itself own the stock. Its holdings weight is not currently verified here. Hedging can therefore reduce some of the company exposure as well as broader market exposure. Sector, currency, credit, earnings and valuation risks can remain. Similar ETF returns do not establish crowded ownership.
Short losses can exceed the initial position. A lower historical volatility does not establish a safer future trade. Published company valuations remain separate from this hedge assessment.
Betting Against Beta · Frazzini and Pedersen ↗Research background; this application does not reproduce the paper’s strategy.