Tail Risk Hedging Strategies: Put Programs, VIX Calls, the Cost of Protection, and the CalPERS Decision

Tail-risk hedging means paying a regular, known cost, usually for out-of-the-money S&P 500 index puts or VIX call options, in exchange for a large payout if markets crash. The trade-off is stark: the premium is lost in most months, and the payoff comes only in rare, sudden sell-offs. Whether a program is worth it depends on how much it costs relative to what it pays in a crash, how well its timing lines up with real drawdowns, and what else the investor could do with the money.

In short: a rolling put program costs a few percent of the protected amount every year. On Cboe’s S&P 500 5% Put Protection Index (PPUT), a monthly 5% out-of-the-money put overlay turned $1 into $8.08 from mid-1986 to 2018, against $20.85 for the S&P 500, while cutting the worst drawdown from about 51% to about 39%. Much of the research (Litterman 2011, Ilmanen 2012, Israelov 2019) finds that for long-horizon, unlevered investors, holding less equity is usually a cheaper way to reduce drawdowns than buying puts. CalPERS, the largest US public pension, ended its explicit tail-hedging program in late 2019 on those grounds, shortly before the March 2020 crash. Advocates argue a small, highly convex hedge can still improve long-run compounding. The evidence favors neither side unconditionally.

This is the institutional follow-on to protective puts explained. It describes how programs are built and what the evidence shows. It is not a recommendation to buy or sell protection.

The misconception: “puts stop drawdowns”

A put protects a specific level on a specific date. A drawdown is a path. If the market falls 4% a month for four months, a 5% out-of-the-money monthly put never pays, and the hedged portfolio falls further than the unhedged one because it also paid four premiums. Roni Israelov’s study for AQR (“Pathetic Protection: The Elusive Benefits of Protective Puts,” Journal of Alternative Investments, 2019) put it bluntly: “Unless option purchases and their maturities are timed just right around equity drawdowns, they may offer little downside protection. In fact, they could make things worse by increasing rather than decreasing drawdowns and volatility per unit of expected return.”

The arithmetic, for a hypothetical monthly 95% put costing 0.46% of the portfolio (paid by holding slightly fewer shares):

One-month index move	Unhedged	With a monthly 95% put
+3%	+3.00%	+2.53%
0%	0.00%	-0.46%
-2%	-2.00%	-2.45%
-5%	-5.00%	-5.44%
-10%	-10.00%	-5.44%
-20%	-20.00%	-5.44%
-34%	-34.00%	-5.44%
Four months of -4.5% each	-16.8%	-18.3%

The put works perfectly on a single large monthly drop and not at all on a slow slide. That is the core problem professionals design around.

The building blocks

Index puts

The standard tool is an out-of-the-money put on the S&P 500 (SPX options, which are European-style and cash-settled). The design choices are:

Strike (moneyness). Deeper out-of-the-money puts are cheaper but pay only in larger crashes.
Tenor. Short-dated puts have high gamma and pay quickly in a sudden crash, but must be bought 12 times a year. Longer-dated puts cost less per month and decay more slowly, but respond less to a sharp drop.
Roll and monetization rules. When to buy, and crucially when to take profits after a crash. A put that rose tenfold and then expired worthless after a rebound protected nothing.

Cboe publishes two benchmark versions, according to its Cboe S&P 500 Put Protection Indices methodology:

PPUT holds the S&P 500 (dividends reinvested) and buys a one-month put each month on the third Friday, at the “first available strike below 95%” of the index level before 11:00 a.m. ET, paying the volume-weighted average price from 11:30 a.m. The expiring put settles against the Special Opening Quotation. History begins June 30, 1986; the index launched August 3, 2015.
PPUT3M, the Cboe S&P 500 Tail Risk Index, buys 10% out-of-the-money puts on the quarterly cycle (March, June, September, December), at the strike at or above 90% of the index. Its base date is March 19, 2004; it launched August 2, 2021.

VIX calls

VIX calls pay when implied volatility spikes, which usually happens when stocks fall. Cboe’s VIX Tail Hedge Index (VXTH) holds the S&P 500 and each month buys one-month 30-delta VIX calls, with a weight that depends on the forward VIX level. Per Cboe’s index documentation: no calls when the one-month forward VIX is at or below 15, a 1% weight between 15 and 30, a 0.5% weight between 30 and 50, and none above 50. The logic, in Cboe’s words, is to vary the hedge “depending on the likelihood that a ‘black swan’ event is about to occur.” VIX calls are priced off VIX futures, not spot VIX, which is why their behavior in February 2018 and March 2020 depended on the futures curve (see the VIX in depth).

Other tools

Institutions also use put spreads (buy a put, sell a lower-strike put to cut cost while capping the payoff), variance swaps and options on volatility, credit protection, and “implicit” hedges such as long-duration Treasuries, which have historically rallied in many, but not all, equity sell-offs.

The cost of protection

The price of a put is set by implied volatility, and equity index puts carry skew: downside strikes trade at higher implied volatility than at-the-money options. Using Black-Scholes with a 4% rate and a 1.3% dividend yield, the cost of a one-month 95% put on $100 of index:

Implied vol of the 95% put	Cost per month	Cost if rolled 12 times
15%	0.22%	2.6%
18%	0.40%	4.8%
20%	0.54%	6.4%
25%	0.93%	11.2%
30%	1.38%	16.5%

And a quarterly 90% put, rolled four times a year: 1.7% a year at 18% implied volatility, 2.5% at 20%, and 4.8% at 25%. These are gross premiums; payoffs offset part of them. Two features matter. Cost rises steeply as implied volatility rises, which means protection is cheapest when it seems least needed and most expensive just after a shock. And a monthly program pays the full term structure of implied volatility, including the premium that option buyers pay on average over realized volatility, described in the variance risk premium.

CalPERS’ chief investment officer gave the board a practitioner’s estimate on April 20, 2020: “explicit options-based tail-risk hedging strategies have an expected cost of three to five percent of the protected notional amount, which is about half of the expected return from the underlying asset being hedged,” against an expected global equity return of about 6% to 7%.

The evidence

The PPUT record

Wilshire Associates’ 2019 study for Cboe, Options-Based Benchmark Indexes: Performance, Risk and Premium Capture (June 1986 to December 2018), compared PPUT with the S&P 500 over 32.5 years.

June 30, 1986 to Dec 31, 2018	S&P 500	PPUT
Growth of $1	$20.85	$8.08
Compound annual return (our calculation from the growth figures)	9.8%	6.6%
Maximum drawdown	-50.95%	-38.92%
Drawdown began	November 2007	June 2007
Months from peak to trough	16	21
Recovered	March 2012	January 2013

The overlay cut the worst drawdown by about 12 percentage points and reduced volatility, but it cost about 3.2 percentage points of return a year, in line with the CalPERS estimate, and PPUT took longer to recover from its worst drawdown than the index did. Israelov’s study, using the same index, found that a simple “divested” portfolio that held less equity and more cash delivered a similar return with better drawdown control.

A simulation: puts versus simply holding less stock

To isolate the mechanism, we simulated 20,000 thirty-year paths of monthly index returns (expected return 7% a year, realized volatility 16%, fat tails from a Student’s t distribution with four degrees of freedom, cash at 4%), and compared buy-and-hold, a monthly 95% put program, and portfolios that hold less stock. The put is priced at 19% implied volatility in the base case (a volatility premium of 3 points) and at 16% in a “fair” case with no premium.

Strategy	Median compound annual return	Median worst drawdown	5th percentile worst drawdown
100% equity, unhedged	5.9%	-41.0%	-63.8%
Monthly 95% puts, priced at 19% IV	3.3%	-40.7%	-63.3%
Monthly 95% puts, priced at 16% IV	5.7%	-33.7%	-52.5%
80% equity, 20% cash	5.8%	-32.0%	-51.3%
70% equity, 30% cash	5.6%	-27.3%	-44.5%

Three findings, all consistent with the published research:

With a volatility premium, the put program loses about 2.6 points a year and barely improves the typical worst drawdown, because most large drawdowns in this model build over several months, and a monthly put pays only on moves larger than 5% within a month.
Even with fair pricing, the put program is roughly matched by holding 80% equity, which delivers a similar return and a similar drawdown with no options at all.
The model is monthly, so it ignores intra-month crashes and the ability to monetize a hedge at its peak, both of which favor active tail programs. Real results depend heavily on those choices. Treat these figures as an illustration of the mechanism, not as a forecast.

The academic debate

Litterman (“Who Should Hedge Tail Risk?”, Financial Analysts Journal, 2011) argued that tail insurance is so expensive that long-term investors should consider being on the other side, and closed: “the next time someone knocks on the door selling a tail-risk insurance product, they should ask for a two-sided market.”
Ilmanen (“Do Financial Markets Reward Buying or Selling Insurance and Lottery Tickets?”, Financial Analysts Journal, 2012) concluded that “the empirical evidence is unambiguous: Selling insurance and selling lottery tickets have delivered positive long-run rewards in a wide range of investment contexts,” while buying financial catastrophe insurance has delivered poor long-run rewards.
Israelov (2019) found that protective puts are inefficient at reducing drawdowns relative to reducing exposure, because of path dependence and the volatility risk premium.
The case for tail hedging rests on portfolio construction rather than on the hedge’s stand-alone return. Mark Spitznagel of Universa Investments argues in Safe Haven (Wiley, 2021) that a small allocation to a highly convex payoff can raise a portfolio’s long-run compound growth by allowing it to rebalance into equities after a crash and to hold more equity overall. That argument depends on achieving very large payoffs from small premiums, which a 5% out-of-the-money monthly put does not deliver; it is a claim about specialized, deeply out-of-the-money, actively monetized programs, and its performance figures come from the managers themselves.

Taken together: the plain, rolling put overlay has a poor long-run record per unit of protection; specialized programs can do better or worse depending on design, cost, and monetization, and their results are hard to verify independently.

The CalPERS decision

CalPERS’ experience is the most public institutional test of the question.

The alternative it chose. In December 2017, according to CIO Ben Meng’s April 2020 remarks, the CalPERS board approved allocations of 15% to factor-weighted equity and 10% to long-duration Treasuries, effective June 30, 2018, as “risk mitigation” segments. CalPERS also ran explicit options-based tail-risk hedges, which press reports identify as mandates with Universa Investments and LongTail Alpha.
The 2019 results. Meng told the Investment Committee that in fiscal year 2018-19, the first year of the risk mitigation segments and “a relative benign market environment,” those segments produced roughly $2 billion of relative value, “over 60 bps at a total fund level,” while the explicit options-based tail-risk strategy “incurred a cost of 7 bps at the total fund level.”
The decision. CalPERS terminated the explicit hedges in late 2019. Meng gave three reasons in his April 2020 remarks: cost (the 3% to 5% of notional quoted above), scalability (“due to lack of depths in the market, it is difficult to gain enough exposure to be meaningful to the size of our portfolio”), and the availability of cheaper alternatives. He cited Litterman (2011) and Ilmanen (2012).
March 2020. The S&P 500 fell 33.9% from its close of 3,386.15 on February 19, 2020 to 2,237.40 on March 23. Meng acknowledged that one article had estimated CalPERS forfeited $1 billion in profit by exiting the strategies, and responded that such analysis “assumed perfect market foresight,” and that “had we known the COVID-19 crisis was coming, we would still have terminated explicit options-based tail-risk hedging strategies.” Board member Margaret Brown said publicly that spring that the board had not been told one of the hedges had been dropped, which became a governance dispute in its own right.

How professionals read the episode depends on what they think the hedge was for. As a stand-alone investment, the program cost money for years and would have paid off spectacularly in a single month. As a portfolio decision, CalPERS argued that long Treasuries and factor tilts provided diversification at lower expected cost and much greater capacity. The two views cannot both be tested on one crash. What the case does show clearly is the governance problem of tail hedging: a program that loses money most of the time is under constant pressure to be cut, and the moment it is cut is, by construction, never known to be the wrong one until afterward. Litterman raised the mirror-image concern in 2011, asking whether management may buy tail insurance to protect itself rather than beneficiaries.

How professionals design a program

Define the liability, not the hedge. A pension with a funding-ratio floor, an insurer with regulatory capital, and a levered hedge fund that faces margin calls have very different reasons to pay for convexity. Litterman’s and Ilmanen’s arguments apply most strongly to unlevered, long-horizon investors with no forced-selling risk.
Set an explicit budget. Programs are typically sized as an annual premium spend (for example, a fraction of a percent of assets) rather than as a hedge ratio, so the cost is known in advance.
Choose strike and tenor to match the threat. Short-dated, deep out-of-the-money puts for sudden crashes; longer-dated puts or put spreads for slower bear markets; VIX calls for volatility spikes.
Write the monetization rules first. When the hedge has risen in value, the program needs pre-agreed rules for selling and rolling, or the gains can evaporate in a rebound.
Compare with the “divested” alternative. The honest benchmark for any tail program is the same portfolio with less equity, as Israelov and CalPERS both did.
Watch the entry price. Because implied volatility drives cost, buying protection after a shock is far more expensive than buying it in calm markets, which tempts programs to start and stop at the wrong times.

What can go wrong

Bleed without a crash. Most months the premium is lost. Over a long calm period, cumulative cost can be large, as the PPUT record shows.
The wrong kind of drawdown. Slow bear markets, or crashes that fall between roll dates, can leave a monthly put program worse off than an unhedged portfolio.
Basis risk. VIX calls depend on VIX futures, which can rise less than spot VIX; index puts do not hedge a portfolio that differs from the index.
Counterparty and liquidity. Over-the-counter hedges carry dealer credit risk, and very large programs face the depth limits Meng described.
Abandonment at the bottom of the bleed. The CalPERS case shows how institutional patience, not the math, often decides whether a program pays off.

The short version

Tail-risk hedging buys crash protection, usually with out-of-the-money index puts or VIX calls, at a known running cost. A monthly 5% out-of-the-money put on the S&P 500 costs roughly 0.2% to 1.4% of the protected amount per month depending on implied volatility, and Cboe’s PPUT index, which runs exactly that program, compounded at about 6.6% a year against 9.8% for the S&P 500 from 1986 to 2018 while cutting the maximum drawdown from about 51% to 39%. Studies by Litterman, Ilmanen, and Israelov conclude that for unlevered long-horizon investors, reducing equity exposure is usually a cheaper drawdown control than buying puts; advocates of specialized convex programs disagree. CalPERS terminated its explicit tail hedges in late 2019, citing cost, scale, and better alternatives, and stood by that decision after the March 2020 crash. For the retail-scale mechanics see protective puts explained; for why option buyers pay a premium on average, see why most options buyers lose.