Tail-Risk Hedging

Thinking Tools4.3 Crafting Solutions for the Future State

From the options-hedging practice of institutional finance, with the fat-tail case named by Nassim Nicholas Taleb (The Black Swan, 2007) and set out by Mark Spitznagel (Safe Haven, 2021).

Protects against rare catastrophes by paying a small, steady cost for positions or arrangements that pay off heavily in disaster, then checks they hold when everything fails together.

Thirty years of a manufacturer's value, two crashes, and a hedge that costs a little every year

A normal year adds 7 percent; a crash year, which comes with a 5 percent chance, takes 45 percent off. The hedge costs a share of value every year and pays out in a crash. The upper chart shows both companies; the lower shows the hedged company's value for every 100 of the unhedged one: below 100 it is paying for the hedge, above 100 the hedge has paid. Size the hedge, then cut what it delivers in a crash and watch it turn into pure cost. Illustrative.

1.0%
100%

Company value (both start at 100)

UnhedgedHedged

0: 100: 1072: 114: 1234: 131: 1406: 150: 1618: 172: 18410: 197: 21012: 225: 24114: 133: 14216: 152: 16218: 174: 18620: 199: 21322: 117: 12524: 134: 14326: 153: 16428: 176: 18830: 201Hedged, 0: 100Hedged, : 106Hedged, 2: 112Hedged, : 119Hedged, 4: 126Hedged, : 134Hedged, 6: 142Hedged, : 150Hedged, 8: 159Hedged, : 169Hedged, 10: 179Hedged, : 190Hedged, 12: 201Hedged, : 213Hedged, 14: 205Hedged, : 217Hedged, 16: 230Hedged, : 244Hedged, 18: 259Hedged, : 274Hedged, 20: 290Hedged, : 308Hedged, 22: 296Hedged, : 313Hedged, 24: 332Hedged, : 352Hedged, 26: 373Hedged, : 396Hedged, 28: 419Hedged, : 444Hedged, 30: 47102004006008000246810121416182022242628300: 100: 1072: 114: 1234: 131: 1406: 150: 1618: 172: 18410: 197: 21012: 225: 24114: 133: 14216: 152: 16218: 174: 18620: 199: 21322: 117: 12524: 134: 14326: 153: 16428: 176: 18830: 201Hedged, 0: 100Hedged, : 106Hedged, 2: 112Hedged, : 119Hedged, 4: 126Hedged, : 134Hedged, 6: 142Hedged, : 150Hedged, 8: 159Hedged, : 169Hedged, 10: 179Hedged, : 190Hedged, 12: 201Hedged, : 213Hedged, 14: 205Hedged, : 217Hedged, 16: 230Hedged, : 244Hedged, 18: 259Hedged, : 274Hedged, 20: 290Hedged, : 308Hedged, 22: 296Hedged, : 313Hedged, 24: 332Hedged, : 352Hedged, 26: 373Hedged, : 396Hedged, 28: 419Hedged, : 444Hedged, 30: 4710200400600800024681012141618202224262830

Hedged value for every 100 unhedged

0: 100: 992: 98: 974: 96: 956: 95: 948: 93: 9210: 91: 9012: 89: 8914: 154: 15316: 152: 15018: 149: 14720: 146: 14522: 252: 25024: 248: 24526: 243: 24128: 239: 23630: 23401002003004000246810121416182022242628300: 100: 992: 98: 974: 96: 956: 95: 948: 93: 9210: 91: 9012: 89: 8914: 154: 15316: 152: 15018: 149: 14720: 146: 14522: 252: 25024: 248: 24526: 243: 24128: 239: 23630: 2340100200300400024681012141618202224262830

Year 30 of 30

Unhedged
201
Hedged value for every 100 unhedged
234
Hedged
471

An example

A packaging manufacturer with $400 million in sales and a term loan with tight covenants asked one question at its annual risk review: what single year would end the company? The finance team named three: a deep recession with frozen credit markets, the loss of its only qualified resin supplier, and a fire at the main plant. In the first, revenue could fall 40 percent while the lenders tightened at the same moment.

The company's existing comfort, customers spread across food, medical, and consumer goods, failed the test of a bad year: in a credit freeze all three cut orders together. So the team bought protection aimed at those scenarios. It took a committed credit line split across two banks, a second resin supplier qualified at a higher price, business-interruption cover on the plant, and a small position in long-dated options that pay when credit spreads blow out. The board set the cost at 0.9 percent of sales a year, written into the budget as a standing line, and agreed never to cancel it after a quiet year.

Why it works

Averages hide the scenarios that matter most. A company can survive many ordinary losses and still be ended by one large one, and once it is gone the good years that follow do it no good. Paying a small, known amount every year to cut off the ruinous outcome trades a slightly lower average for a much higher chance of being around to collect it.

The second insight is where most hedges fail: correlation. Protection that works when one thing goes wrong often fails when everything goes wrong at once, which is exactly when it is needed. Diversified customers stop diversifying in a recession; a credit line from a struggling bank vanishes with the bank. Testing each hedge against the joint disaster, the year when everything fails together, is what separates a hedge from a comforting story.

The thinking behind it

Nassim Nicholas Taleb Nassim Nicholas Taleb, a former options trader, argued in The Black Swan (2007) that rare, high-impact events dominate outcomes and that models built on ordinary variation understate them. His practical conclusion is to stop forecasting the rare event and instead arrange exposure so that being wrong about it is survivable, with a small known cost traded for protection against the open-ended loss.

Mark Spitznagel Mark Spitznagel, who founded the hedge fund Universa Investments, set out the case in Safe Haven (2021): the right protection against crashes can raise a portfolio's long-run growth as well as lower its risk, because avoiding the deep loss matters more to compounding than the small yearly cost. It is an investor's argument from practice, with a foreword by Taleb, and not peer-reviewed research.

How to run it

  1. List the years that would end the business. Name the events, alone or together, that would break covenants, empty the cash, or lose the license to operate.
  2. Size each one roughly. An order of magnitude is enough: how much revenue, cash, or capacity goes, and for how long.
  3. Find protection that pays in that scenario. Committed credit, insurance, a second supplier, options, or spare capacity, each tied to a named disaster.
  4. Test every hedge against the joint disaster. Ask whether it still pays when markets, customers, and counterparties all fall together, and drop the ones that do not.
  5. Price it as a standing cost. Set a yearly budget for the protection, as a share of sales or value, and put it in the plan like rent.
  6. Buy before the storm. Protection gets expensive or disappears once trouble is visible, so put it in place in calm years.
  7. Review the list each year. Add new ways to be ended, retire hedges whose scenario has passed, and keep the budget whatever last year did.

Where it goes wrong

The usual failure is canceling the protection after several quiet years, because it looks like money spent for nothing. It shows up as a budget review that calls the hedge waste. The recovery is to keep the cost as a fixed line and judge it against the scenario it covers, never against the last year's outcome.

A second failure is protection that does not hold under correlation: insurance from a carrier that fails in the same catastrophe, or a credit line with a clause that lets the bank walk away in a downturn. Read the terms against the joint disaster. A third is overbuying, until the yearly cost becomes the slow loss it was meant to prevent; size each hedge to the damage it caps.

When to use something else

  • Barbell Strategy. When the question is how to split a whole portfolio or plan between very safe and small speculative bets, and not how to protect one exposure.
  • Wind Tunneling. When the catastrophic scenarios are not yet known and the first job is to run the plan through several hard futures.
  • Factor of Safety. When the risk is a known load on a known system and a margin built into the design is cheaper than protection bought outside it.

Where it sits in SOLVE eX

Go to the source

  • The Black Swan, Random House, 2007
  • Safe Haven: Investing for Financial Storms, Wiley, 2021

Back to the catalog