Week 40: VIX and the Volatility Complex

---

1. Why This Is Important

The VIX is the most quoted, least understood number in finance. CNBC says "the fear gauge spiked to 28" and viewers nod, but almost nobody on the receiving end of that headline can tell you what 28 actually means, what it predicts, or — crucially — why every retail product built on top of it tends to bleed money. Volatility is what wags the dog: in a real crisis it is the first thing that moves, and everything else rearranges itself around it. If you don't have a working model of how the volatility complex behaves, you will be the last one to know what just happened to your portfolio.

There are four reasons a serious investor needs to understand VIX cold:

VIX is a price for an asset class, not a forecast: The VIX is a 30-day implied-volatility number, computed model-free from a strip of SPX option prices. It tells you what the option market is charging for protection right now, not what realised vol will actually be. The premium between the two — the variance risk premium — is one of the cleanest, most persistent return sources in markets, and it shows up in every option you trade.

The term structure is normally upside-down from intuition: VIX futures are in contango ~85% of trading days. Three-month implied vol is almost always higher than spot VIX. That single structural fact is why long-vol products (VXX, UVXY) bleed roughly 30%+ per year, and why every "buy volatility for insurance" pitch deserves the same scrutiny as a 4% mortgage in 1981.

Volatility is the cheapest way to be wrong with conviction: Long-vol bets pay off in catastrophes. But carry is brutal — you pay every day the world doesn't end, and you give back the gains within weeks of the bottom because the term structure resets. Most retail "tail hedges" are bought at peak VIX and sold at the trough, which is the exact opposite of how they're supposed to work.

The short-vol trade has eaten more reputations than the long-vol trade: Selling volatility looks like free money for years, until it isn't. Volmageddon (Feb 2018) wiped out the XIV ETN in a single afternoon — a -96% close after years of +30%/yr returns. Vol blow-ups are the reason the entire portfolio rearranges itself when they hit. You cannot understand 2008, 2018, 2020, or 2022 without understanding what was happening in the vol surface those weeks.

This lesson gives you the formula, the term structure, the products, the wreckage record, and a framework for using vol as a tool rather than as a slot machine.

---

2. What You Need to Know

2.1 What VIX Actually Is — The Variance-Swap Formula

The CBOE redefined VIX in 2003 to be model-free. It does not assume Black-Scholes; it does not assume any model. It is the square root of a weighted strip of SPX option prices that replicates a 30-day variance swap.

The formal definition is

$$\text{VIX}^2 = \frac{2}{T} \sum_i \frac{\Delta K_i}{K_i^2} e^{rT} Q(K_i) - \frac{1}{T}\left(\frac{F}{K_0} - 1\right)^2$$

where $T = 30/365$, $K_i$ are the listed SPX option strikes, $Q(K_i)$ is the OTM mid-quote at strike $K_i$, $F$ is the forward, and $K_0$ is the strike just below $F$. The result is annualised — VIX = 20 means the option market is implying a 20% standard deviation of SPX returns over the next year, or equivalently $20/\sqrt{12} \approx 5.8\%$ over the next month.

Two implications most retail traders miss:

• VIX is not "expected volatility"; it is the price of a variance swap. It will systematically be higher than realised vol because option sellers demand a premium for unhedgeable jump risk. The long-run gap is ~3-4 vol points (the variance risk premium). That premium is what every covered-call and put-write strategy harvests.
• Daily expected SPX move = VIX / sqrt(252). A VIX of 20 implies a 1.26% daily standard deviation. A VIX of 40 implies 2.52%. This is the single most useful conversion to memorise.

2.2 The Term Structure — Why Contango Eats VXX

VIX itself is a spot index. You cannot trade it. What you can trade are VIX futures, which settle to a special opening quote of VIX on the third-Wednesday morning of the expiry month. The futures curve almost always slopes upward: front month at, say, 16, second month at 17, third at 17.8. This is contango.

Why? Because the option market knows that vol mean-reverts to ~16-18, but also knows there is some chance of a panic in the next 90 days. The further out you go, the more "tail" you're insuring against. The price the curve charges for that insurance, day after day, is roughly +0.05 to +0.10 vol points per trading day at the front, summing to ~+1.5 vol points/month.

Now layer on what VXX, UVXY, and similar products actually do: they hold a constant-maturity 30-day VIX-futures position, rolling daily from the front month to the second month. In contango, every roll is a sale at a low price and a purchase at a higher price. The annualised cost of that roll, historically, has been -25% to -40% in calm markets and -10% to -20% on average across a full cycle.

This is not a quirk. It is mathematical. There is a corollary to the rule that volatility leads everything else: the volatility surface is paying you to be short, until the day it isn't.

2.3 The Product Wreckage Record

The full taxonomy of what has happened to retail investors in volatility ETPs:

• VXX (long 1x VIX futures, iPath): Launched Jan 30, 2009. Original VXX matured Jan 2019. Series B started Jan 2018 (overlap year). Cumulative return since 2009 inception, adjusted for the eight reverse splits required to keep the share price above $1: approximately -99.99%. A $1 investment is worth less than $0.0001.
• UVXY (long 1.5x — was 2x — VIX futures): Even worse. Decay is roughly 1.5x VXX's, so you're losing ~45-55%/yr in calm markets. Reverse-split count: 11 and counting.
• XIV (short 1x VIX futures, Credit Suisse ETN): The winner — until Feb 5, 2018. VIX doubled in a single afternoon as systematic vol sellers were forced to cover. XIV's NAV dropped ~96% in two hours. Credit Suisse triggered the acceleration clause and shut it down. Investors who held overnight got pennies.
• SVXY (short 0.5x VIX futures, ProShares): Was 1x until the same Feb 2018 event, when ProShares cut the leverage to 0.5x for survival. Still positive total return, but with much smaller upside than XIV's pre-Volmageddon glory days.
• VIX itself: Cannot be held. Spot VIX is a calculation, not a tradable. Every "VIX" product is really a VIX-futures product, and the futures live on a different curve than the spot.

The general rule: if a product holds a constant-maturity long-vol position, it is a slowly-melting ice cube. Buying it as an "investment" is structurally guaranteed to lose money. Buying it as a trade, sized to a few-day holding period during a confirmed vol spike, can work — but the timing window is brutal.

2.4 VIX Levels and Spike Characteristics

Calibration anchors any sane investor should commit to memory:

• Median VIX (1990-2026): ~16.5. The all-time-low is 9.14 (Nov 2017). The 25th percentile is ~13, the 75th percentile is ~22.
• Calm: 12-15. Bull-market default state. SPX daily moves under 1%.
• Normal: 15-20. Long-run average. Healthy markets with periodic earnings noise.
• Elevated: 20-30. Correction, geopolitical shock, election year.
• Stressed: 30-50. Bear market, banking event, recession scare. Aug 2024 carry-trade unwind, Sept 2022 Fed shock both touched the high 30s.
• Panic: 50-80. Genuine crisis. 2008 (peak ~89), 2010 flash crash (~48), Aug 2011 debt-ceiling/Europe (~48), Feb 2018 Volmageddon (~50 intraday).
• Extreme: 80+. All-time high 82.69 on Mar 16, 2020 (COVID lockdown announcement). Only one other day above 80 in 36 years (2008 Lehman week).

Spike behaviour: VIX rises ~3-5x faster than SPX falls. The 1-day correlation coefficient is around -0.75. The classic relationship is

$$\Delta\text{VIX} \approx -1.0 \times \Delta\text{SPX}\%$$

for normal moves (i.e., a -1% SPX day adds about +1 vol-point to VIX), but with severe convexity at the tails — a -5% SPX day might add +10-15 points, not +5. This convexity is exactly what long-vol payoffs harvest.

2.5 VVIX — Vol of Vol

VVIX is the implied volatility of VIX itself, computed from VIX options. It typically prints 70-110 in normal markets and spikes to 150-200 in crises. VVIX is the cleanest read on how stretched the options market thinks the vol surface is. When VVIX is at 80 and you're getting paid 12% to sell SPX puts, the market is pricing fairly. When VVIX is at 150 and you're getting paid 30%, you are not being overpaid — you are being paid for the very real risk that VIX itself jumps another 50%.

Practical rule: never sell vol when VVIX is rising fast. Even if VIX is already high. The vol-of-vol is the second derivative, and it's where the leverage in the system lives.

2.6 The Vol Risk Premium and How To Harvest It

The empirical fact that has built more careers than any other in derivatives: implied vol > realised vol, on average, by ~3-4 points/year in SPX. That gap is the variance risk premium (VRP). It is real, persistent, and explains the long-run return of every short-vol strategy from PUTW (cash-secured puts on SPX) to JEPI (SPX covered-calls + ELN) to SVXY itself.

Three ways to harvest VRP without blowing up:

Cash-secured puts on SPX or sector ETFs — capped downside, capped upside, sized to portfolio. Covered in Week 28.

Covered calls on positions you'd hold anyway — gives up the right tail, keeps the body. Covered in Week 27.

SPX iron condors with defined max loss — short both wings, long protective wings further out. Max-loss is known on entry. Covered in Week 30.

What does not work for the long-run investor: long VXX, long UVXY, short SVXY, or any structure where the worst-case loss exceeds the position size. The barbell is the right framing: you can run a 90% beta portfolio and a 5-10% short-vol sleeve, but you cannot run a 95% short-vol portfolio. The blowup days take 2-3 years of carry to recover, and the carry isn't large enough to justify the leverage.

2.7 The 2018 Volmageddon Case Study

Feb 5, 2018, was the textbook example of vol-tail-wags-dog. The setup: SPX had been grinding higher with realised vol below 6% for months. VIX was pinned at 9-11. Short-vol products had attracted ~$2 billion in retail AUM. Pension funds and family offices were quietly running "vol-control" mandates that systematically sold more vol as realised vol stayed low — the same risk-parity logic that worked in calm regimes.

Then SPX fell 4% in the regular session. VIX spiked from 17 to 37 by the close. The XIV/SVXY products had to cover their short futures positions into a thin after-hours market. Their forced buying drove VIX futures even higher. By 4:15pm ET, XIV's NAV implied a 96% loss. Credit Suisse pulled the plug.

The lesson is not "don't sell vol." The lesson is: leverage on vol is fatal because vol itself is the leverage. A 1x short-vol position is already a leveraged bet on the variance risk premium. Stacking 1x fund leverage on top of that creates 2x effective leverage on a series with kurtosis of 30+. The math doesn't survive.

---

3. Common Misconceptions

"VIX measures actual market volatility." No — it measures expected (implied) volatility from option prices. The market is paying a premium for unhedgeable risk; that premium typically runs 3-4 vol points above realised.

"A high VIX means stocks will fall." It means stocks just fell, usually. VIX is more of a coincident-to-lagging indicator than a leading one. The forecast power of VIX for forward 30-day SPX returns is weak; its forecast power for forward realised vol is moderate.

"VXX is a long-volatility ETF; if vol goes up, VXX goes up." Only over short horizons. Held for a year, VXX loses to VIX by 25-40%/yr because of contango roll and compounding drag.

"I'll buy VXX as a hedge for my portfolio." Buying VXX today and rolling it for a year typically costs 30%+ of notional. A protective put on SPX usually costs 2-4% for similar tail protection. The put is almost always cheaper insurance.

"XIV blew up because of the algorithms." XIV blew up because it was a 1x short-vol product offering 1x leverage on a series with 30+ kurtosis. The algorithms were proximate, not root.

"VIX of 50 means a 50% market crash is coming." VIX of 50 means the option market expects a 14.4% one-month standard deviation of SPX returns. Big move expected, direction unspecified.

"Buying volatility before earnings is smart because IV always rises." IV rises into the print, but typically crushes immediately after — the IV-crush is the entire reason short-strangle strategies work around earnings.

"Selling SVXY is the same as selling vol." SVXY is short 0.5x VIX futures. It is not pure short-vol; it is short the futures curve, which embeds contango. Same direction, different greeks.

"VIX has been getting structurally lower since 2010." It hasn't. The unconditional median is roughly stable since 1990. What's changed is the frequency of micro-spikes (1-2 day) versus prolonged 30+ regimes.

"I should buy VXX at low VIX because it's cheap." Low VIX is a state, not a price level. A VIX of 12 with a steep contango can decay just as fast as a VIX of 18 with a flat curve. Look at the roll cost, not the absolute level.

---

4. Q&A Section

Q1: How do I actually use VIX in day-to-day decisions? A: Three uses. (1) Position sizing — when VIX > 25, cut new long-equity sizing in half versus when VIX < 15. (2) Options pricing — if you're about to sell a put and VIX < 14, you're underpaid; wait. If VIX > 30 and you're already long stock, that's premium income waiting to be collected via covered calls. (3) Regime — VIX > 30 for more than 5 sessions is a confirmed stress regime; rebalance across the four tranches, don't fight it.

Q2: Why don't more investors hold long-vol products as portfolio insurance? A: Because they're the most expensive insurance ever invented. Carry is -25% to -40%/yr. SPX puts cost 2-4%/yr for similar tail coverage. Buying VXX as insurance is like buying a fire-insurance policy that costs 30% of your house's value every year — even if it pays off in a fire, you'd have been better off with a $300/yr policy.

Q3: Can I time the VIX? A: Mean-reversion of level is real (median ~16-18). But the timing is brutal — VIX can stay elevated for months in a bear market. A naive "sell vol when VIX > 30" rule worked in 2010, 2011, 2018, 2020, 2022. It would have buried you in 2008 (VIX > 30 for 7 months) and could bury you in any genuine systemic crisis.

Q4: What's the difference between VIX and VVIX? A: VIX is 30-day implied vol of SPX. VVIX is 30-day implied vol of VIX. VVIX tells you how stretched the option-on-VIX market is — i.e., how convex the surface has become. Rising VVIX with rising VIX = panic compounding. Rising VIX with falling VVIX = mean-reversion likely.

Q5: Why did the original VXX get delisted in 2019? A: It had a 10-year maturity (issued Jan 2009 as an iPath ETN, not an ETF). Barclays redeemed it at NAV in Jan 2019. iPath rolled it into VXX Series B (launched Jan 2018, overlap design). The two are nearly identical economically — same constant-maturity 30-day VIX-futures methodology.

Q6: Is the short-vol trade dead after Volmageddon? A: No, but it's smaller. The XIV-style 1x ETN structure is gone. Today's short-vol exposure lives in: (1) covered-call ETFs (JEPI, JEPQ, QYLD), which are diluted versions of the trade; (2) cash-secured-put strategies (PUTW); (3) SVXY at 0.5x leverage. Total AUM in short-vol ETPs in 2026 is ~$50B, half the pre-2018 peak.

Q7: How do I read the VIX term structure to time entries? A: The classic gauge is VX1/VX2 ratio (front-month / second-month VIX futures). When ratio < 1.0 (backwardation), you're in stress — not a time to short vol. When ratio > 1.0 with normal contango (~0.95), you're in calm; selling defined-risk puts is reasonable. Tools like VIX Central or the CBOE term-structure page show this live.

Q8: What's the relationship between VIX and credit spreads? A: Tight, but with a lag. VIX is the "fast" risk indicator (intraday, options-driven). HY credit spreads (BAML HOAS) are the "slow" indicator (daily, dealer-driven). A spike in VIX without a corresponding move in HY spreads is usually noise; a coincident move in both is a regime shift. Cross-reference Week 33's HY-spread chart.

Q9: Should I include any vol exposure in a long-only retirement portfolio? A: Direct vol exposure, no. Indirect, yes — through covered-call ETFs (10-15% sleeve max) or PUTW-style cash-secured-put funds. These harvest VRP without the leverage. Run it as a barbell, with the vol risk taken as defined-risk option premia rather than as a long-VXX position.

Q10: How much of my portfolio should ever be long VXX or UVXY? A: For most investors: 0%. For a sophisticated tactical trader running a defined event thesis (e.g., "FOMC meeting tomorrow, vol is at 12, term structure is flat"), maybe 0.5-1.5% of NAV for a 2-7 day hold. Never as a strategic position. Never sized so a -50% day on the position would matter.

Q11: Why did the all-time-high VIX print on Mar 16, 2020 and not Mar 9 (the actual market low)? A: VIX measures next 30-day expected vol. By Mar 16, the option market was pricing in a full lockdown scenario for the following month — that's why the high print came after the worst single-day SPX drops. VIX is more reliably backward-looking than people assume.

Q12: Is there any reliable way to predict VIX spikes? A: No. The only structural predictor is VIX itself being abnormally low for an abnormally long time, which raises the probability of a spike but says nothing about timing. The vol-control deleveraging mechanism that triggered Volmageddon required a SPX move of >2% to start; you can't forecast that move with VIX. If you could, you wouldn't need VIX.