Volatility-of-Volatility Risk

∗
Darien Huang Christian Schlag Ivan Shaliastovich Julian Thimme

February 2018

Abstract

We show that time-varying volatility of volatility is a significant risk factor which

affects the cross-section and the time-series of index and VIX option returns, beyond
volatility risk itself. Volatility and volatility-of-volatility measures, identified model-
free from options data as the VIX and VVIX indices, respectively, are only weakly
related to each other. Delta-hedged index and VIX option returns are negative on
average, and more negative for strategies more exposed to volatility and volatility-of-
volatility risks. Volatility and volatility of volatility significantly and negatively predict
future delta-hedged option payoffs. The evidence is consistent with a no-arbitrage
model featuring time-varying volatility and volatility-of-volatility factors, which are
negatively priced by investors.

∗
Darien Huang is at Cornell University, Christian Schlag is at Goethe University Frankfurt and Research
Center SAFE, Julian Thimme is at Goethe University Frankfurt, and Ivan Shaliastovich is at University of
Wisconsin Madison. We thank Franklin Allen, Luca Benzoni, João Gomes, Mete Kilic, Krishna Ramaswamy,
Scott Richard, Nikolai Roussanov, Anders Trolle, Amir Yaron, Hao Zhou, and seminar participants at Whar-
ton, the 2014 European Finance Association Meeting in Lugano, 2014 Optionmetrics Research Conference,
and the 2014 Asian Meeting of the Econometric Society in Taipei for their comments and suggestions. Schlag
gratefully acknowledges research and financial support from SAFE, funded by the State of Hessen initiative
for research LOEWE. Shaliastovich thanks the Jacobs Levy Equity Management Center for Quantitative
Financial Research, the Rodney White Center, and the Cynthia and Bennett Golub Endowment for financial
support.

Electronic copy available at: https://ssrn.com/abstract=2497759

1 Introduction

Recent studies show that volatility risks significantly affect asset prices and the macroecon-
omy.1 In the data, asset market volatility is often captured by the volatility index (VIX).
Calculated in real time from the cross-section of S&P500 option prices, the VIX index pro-
vides a risk-neutral forecast of the index volatility over the next 30 days. The VIX index
exhibits substantial fluctuations, which in the data and in many economic models drive the
movements in asset prices and risk premia. Interestingly, the volatility of the VIX index
itself varies over time. Computed from VIX options in an analogous way to the VIX, the
volatility-of-volatility index (VVIX) directly measures the risk-neutral expectations of the
volatility of volatility in the financial markets. In the data, we find that the VVIX has sep-
arate dynamics from the VIX, so that fluctuations in volatility of volatility are not directly
tied to movements in market volatility. We further show that volatility-of-volatility risks
are a significant risk factor which affects the time-series and the cross-section of index and
VIX option returns, above and beyond volatility risks. The evidence is consistent with a
no-arbitrage model which features time-varying market volatility and volatility-of-volatility
factors which are priced by the investors. In particular, volatility and volatility of volatil-
ity have negative market prices of risk, so that investors dislike increases in volatility and
volatility of volatility, and demand a risk compensation for the exposure to these risks.

Our no-arbitrage model extends the one-factor stochastic volatility specification of equity
returns in Bakshi and Kapadia (2003). Specifically, we introduce a separate time-varying
volatility-of-volatility risk factor which drives the conditional variance of the variance of
market returns.2 We use the model to characterize the payoffs to delta-hedged equity and
VIX options. The zero-cost, delta-hedged positions represent the gains on a long position in
1
See e.g. Bansal and Yaron (2004), Bloom (2009), Bansal, Kiku, Shaliastovich, and Yaron (2014),
Fernandez-Villaverde and Rubio-Ramı́rez (2013) for the discussion of macroeconomic volatility risks, and
Coval and Shumway (2001), Bakshi and Kapadia (2003), Campbell, Giglio, Polk, and Turley (2012) for
market volatility risks.
2
We use the terms “variance” and ”volatility” interchangeably unless specified otherwise.

Electronic copy available at: https://ssrn.com/abstract=2497759

the option, continuously hedged by an offsetting short position in the underlying asset. As
argued in Bakshi and Kapadia (2003), delta-hedged option payoffs most cleanly isolate the
exposures to volatility risks, and thus are very useful to study the pricing of volatility-related
risks.3 Indeed, we show that the expected payoff on the delta-hedged index option positions
capture risk compensations for both volatility and volatility-of-volatility risks, while for VIX
options, the expected gains primarily reflect the compensation for volatility-of-volatility risk.
Imposing a standard linear risk premium assumption, we can further decompose the risk
premia components of the expected gains on the index options into the product of the
market price of risk, the risk exposure, and the time-varying quantity of each source of risk.
For VIX options, the expected gains do not generally admit a linear factor structure, so we
resort to numerical calibration techniques to support our theoretical arguments.

The model delivers clear, testable predictions for expected option returns and their rela-
tion to volatility and volatility-of-volatility risks. In the model, if investors dislike volatility
and volatility of volatility, so that the market prices of these risks are negative, delta-hedged
equity and VIX option gains are negative on average. In the cross-section, the average re-
turns are more negative for option strategies which have higher exposure to the volatility
and volatility-of-volatility risks. Finally, in the time series, higher volatility and volatility of
volatility predict more negative delta-hedged option gains in the future.

We find that the model implications for volatility and volatility-of-volatility risks are
strongly supported in the cross-section of index and VIX option returns. First, we use
the VIX and VVIX indices to validate our volatility measures, and compare and contrast
their dynamics. In our monthly 2006 - 2016 sample, the option-implied volatility measures
capture meaningful information about the uncertainty in future market returns and market
volatility in the data. Using predictive regressions, we show that the VIX is a significant
predictor of the future realized variance of market returns, while the VVIX significantly
3
For example, unlike delta-hedged positions, zero-beta straddles analyzed in Coval and Shumway (2001)
are not dynamically rebalanced and may contain a significant time-decay option premium component.

Electronic copy available at: https://ssrn.com/abstract=2497759

forecasts future realized variation in the VIX index itself. Including both volatility measures
at the same time, we find that the predictive power is concentrated with the corresponding
factor (i.e., the VIX for market return volatility, the VVIX for VIX volatility). This evidence
confirms that the measured VIX and VVIX indices can indeed separately capture volatility
and volatility-of-volatility movements in the asset markets.

In the time-series, the VVIX behaves quite differently from the VIX, consistent with a
setup of our model which separates market volatility from volatility of volatility. The VVIX
is much more volatile, and is less persistent than the VIX. The correlation between the two
series is 0.27. While both volatility measures share several common peaks, most notably
during the financial crisis, other times of economic distress and economic uncertainty, such
as the Eurozone debt crisis and flash crash in May 2010 as well as the U.S. debt ceiling crisis
in August 2011, are characterized by large increases in the VVIX with relatively little action
in the VIX.

On average, the risk-neutral volatilities of the market return and market volatility cap-
tured by the VIX and VVIX exceed the realized volatilities of returns and the VIX. The
difference between the risk-neutral and physical volatilities of market returns is known as
the variance premium (variance-of-variance premium for the VIX), and the findings of pos-
itive variance and variance-of-variance premium suggest that investors dislike variance and
variance-of-variance risks, and demand a premium for being exposed to these risks.

We next turn to the asset-price evidence from the equity index and VIX option markets.
In line with our model, we consider discrete-time counterparts to the continuously-rebalanced
delta-hedged gains; this approach is similar to Bakshi and Kapadia (2003)4 . Consistent with
the evidence in previous studies, the average delta-hedged returns on out-of-the-money equity
index calls and puts are significantly negative in our sample. The novel evidence in our paper
is that the average delta-hedged returns on VIX options are also negative and statistically
4
We perform a numerical analysis of the model to verify the adequacy of the approximation of the option
delta by the delta computed from the Black-Scholes formula.

Electronic copy available at: https://ssrn.com/abstract=2497759

significant at all strikes, except for out-of-the-money puts and in-the-money calls which are
marginally significant. Estimates of the loss for call options range from -0.25% of the index
value for in-the-money VIX calls to -1.20% for out-of-the-money calls. The negative average
returns on index and VIX options directly suggest that the market prices of volatility and
volatility-of-volatility risks are negative.

We then show that the cross-sectional spreads in average option returns are significantly
related to the volatility and volatility-of-volatility risks. In lieu of calculating exact model
betas, we compute proxies for the option exposures to the underlying risks using the Black
and Scholes (1973) vega and volga. Vega represents an increase in the Black-Scholes value
of the option as the implied volatility increases by one percentage point, and thus provides
an estimate for the exposure of equity options to volatility risks, and of VIX options to
volatility-of-volatility risks. Volga is the second partial derivative of the option price with
respect to the volatility, which we use to measure the sensitivity of the index option price to
the volatility-of-volatility risks. Vega and volga vary intuitively with the moneyness of the
option in the cross-section, and help us proxy for the betas of the options to the underlying
risks. Empirically, we document that average option returns are significantly and negatively
related to our proxies for volatility and volatility-of-volatility risks. Hence, using the cross-
section of equity index options and VIX options, we find strong evidence for a negative
market price of volatility and volatility-of-volatility risks.

Finally, we consider a predictive role of our volatility measures for future option returns.
In the model, expected delta-hedged gains are time-varying and are driven by the volatility
and volatility of volatility (by volatility of volatility for VIX options). In particular, as option
betas are all positive, when the market prices of volatility-related risks are negative, both
volatility measures should forecast future returns with a negative sign. This model prediction
is consistent with the data. The VIX and VVIX significantly negatively predict future index
option returns, and the VVIX is a significant negative predictor of option returns on the

Electronic copy available at: https://ssrn.com/abstract=2497759

VIX. Hence, using the cross-sectional and time-series evidence from the option markets, we
find strong support that both volatility and volatility-of-volatility risks are separate priced
sources of in the option markets, and have negative market prices of risks.

Related Literature. Our paper is most closely related to Bakshi and Kapadia (2003) who
consider the implications of volatility risk for equity index option markets. We extend their
approach to include volatility-of-volatility risk, and bring evidence from VIX options. To
help us focus on the volatility-related risks, we consider dynamic delta-hedging strategies
where a long position in an option is dynamically hedged by taking an offsetting position in
the underlying. Delta-hedged strategies are also used in Bertsimas, Kogan, and Lo (2000),
Cao and Han (2013) and Frazzini and Pedersen (2012), and are a standard risk management
technique of option traders in the financial industry (Hull (2011)). In an earlier study, Coval
and Shumway (2001) consider the returns on zero-beta straddles to identify volatility risk
sensitive assets. Zhang and Zhu (2006) and Lu and Zhu (2010) highlight the nature and
importance of volatility risks by analyzing the pricing of VIX futures. Also notably our
analysis suggests that variance dynamics are richer than that of the square-root process
typically considered in the literature — these findings are consistent with the results of
Christoffersen, Jacobs, and Mimouni (2010) and Branger, Kraftschik, and Völkert (2016).

In a structural approach, Bollerslev, Tauchen, and Zhou (2009) consider a version of the
Bansal and Yaron (2004) long-run risks model which features recursive utility and fluctua-
tions in the volatility and volatility of volatility of the aggregate consumption process. They
show that in equilibrium, investors require compensation for the exposure to volatility and
volatility-of-volatility risks. With preference for early resolution of uncertainty, the market
prices of the two risks are negative. As a result, the variance risk premium is positive on av-
erage, and can predict future equity returns. Bollerslev et al. (2009) and Drechsler and Yaron
(2011) show that the calibrated version of such a model can account for the key features of
equity markets and the variance premium in the data. Our empirical results are consistent

Electronic copy available at: https://ssrn.com/abstract=2497759

with the economic intuition in these models and complement the empirical evidence in these
studies.

Finally, it is worth noting that in our paper we abstract from jumps in equity returns,
and focus on diffusive volatilities as the main drivers of asset prices and risk premia. For
robustness, we confirm that our predictability results are robust to controlling for jump risk
measures such as the slope of the implied volatility curve, realized jump intensity (Barndorff-
Nielsen and Shephard (2006) and Wright and Zhou (2009)), and risk-neutral skewness (Bak-
shi, Kapadia, and Madan (2003)). Hence, we argue that the VIX and VVIX have a significant
impact on option returns even in the presence of stock market and volatility jumps; we leave
a formal treatment of jumps for future research. Reduced-form models which highlight the
role of jumps include Bates (2000), Pan (2002), and Duffie, Pan, and Singleton (2000), among
others. Muravyev (2016) documents that inventory risk faced by market-makers also affects
option prices.

Our paper proceeds as follows. In Section 2 we discuss our model which links expected
delta-hedged equity and volatility option gains to risk compensations for volatility and
volatility-of-volatility risk. In Section 3, we consider an illustrative numerical example to
show the relation between VIX option gains and the volatility-related risks. In Section 4,
we describe the construction of both the model-free implied variance measures and high-
frequency realized variance measures, and summarize their dynamics in the time-series. We
show that the implied variances have a strong ability to forecast future realized variance.
Section 5 provides the empirical evidence from option prices by empirically implementing
the delta-hedged option strategies in our model. Section 6 presents robustness tests for al-
ternative measures of variance, as well as robustness of the results in the presence of jump
risks. Section 7 concludes the paper.

Electronic copy available at: https://ssrn.com/abstract=2497759

2 Model

In this section we describe our model for stock index returns, as well as for equity and
volatility option prices. Our model is an extension of Bakshi and Kapadia (2003) and features
separate time-variation in the market volatility and volatility of volatility. Both volatility
risks are priced, and affect the level and time-variation of the expected asset payoffs.

2.1 Dynamics of Equity and Equity Option Prices

Under the physical measure (P) the stock index price St evolves according to:

dSt p
= µ(St , Vt , ηt )dt + Vt dWt1 ,
St
√ (2.1)
dVt = θ(Vt )dt + ηt dWt2 ,
√
dηt = γ(ηt )dt + φ ηt dWt3 ,

where W i (i = 1, 2, 3) are the Brownian motions which drive stock returns, the stock return
variance, and the variance of variance, respectively. The Brownian components can be
correlated, i.e., we assume dWti dWtj = ρi,j dt for all i 6= j. Vt is the variance of instantaneous
returns, and ηt is the variance of innovations in Vt . Note that the drift of the variance Vt only
depends on itself, and not on the stock price St or the volatility of volatility ηt . Similarly,
the drift of the volatility of volatility ηt is a function of the volatility of volatility ηt .

Under the risk-neutral measure (Q) the stock price St follows a similar process, but with
the drifts adjusted by the risk compensations for the corresponding risks:

dSt p
ft1 ,
= rf dt + Vt dW
St
√ (2.2)
dVt = (θ(Vt ) − λVt )dt + ft2 ,
ηt dW
√ f3
dηt = (γ(ηt ) − ληt )dt + φ ηt dW t .

Electronic copy available at: https://ssrn.com/abstract=2497759

 f i represent Brownian motions under the risk-neutral measure
In this representation, the W
Q. λVt captures the compensation for variance risk, and ληt reflects the compensation for
innovations in the stochastic variance of variance. If investors dislike variance and variance-
of-variance risks, the two risk compensations are negative. In this case, the variances have
higher drifts under the risk-neutral than under the physical measure.

Let Ct (K, τ ) denote the time t price of a call option on the stock with strike price K and
time to maturity τ. Assume the risk-free rate rf is constant. To simplify the presentation,
we further abstract from dividends. While we focus our discussion on call options, the case
of put options follows analogously. Given the specified dynamics of the stock price under
the two probability measures, the option price is given by a twice-differentiable function C
of the state variables: Ct (K, τ ) = C(St , Vt , ηt , t). By Itô’s Lemma,

∂C ∂C ∂C
dCt = dSt + dVt + dηt + bt dt, (2.3)
∂S ∂V ∂η

for a certain drift component bt .

The discounted option price e−rf t Ct is a martingale under Q and thus has zero drift. We
use Itô’s Lemma again to obtain that:

∂C ∂C
∂C
St rf + θ(Vt ) − λVt + (γ(ηt ) − ληt ) + bt − rf Ct = 0. (2.4)
∂S ∂V ∂η

This implies that:

 
∂C ∂C ∂C
bt = rf Ct − St − (θ(Vt ) − λVt ) − (γ(ηt ) − ληt ). (2.5)
∂S ∂V ∂η

Let Πt,t+τ stand for the delta-hedged option gain for call options:

Z t+τ Z t+τ  
∂C ∂C
Πt,t+τ ≡ Ct+τ − Ct − dSu − rf Cu − Su du. (2.6)
t ∂S t ∂S

Electronic copy available at: https://ssrn.com/abstract=2497759

The delta-hedged option gain represents the gain on a long position in the option, continu-
ously hedged by an offsetting short position in the stock, with the net balance earning the
risk-free rate.
R t+τ
Combining equations (2.3) and(2.5) and noting that Ct+τ = Ct + t
dCu , we see that
the delta-hedged gain for call options is given as

Z t+τ Z t+τ  
∂C ∂C
Πt,t+τ ≡ Ct+τ − Ct − dSu − rf Cu − Su du
t ∂S t ∂S
Z t+τ Z t+τ Z t+τ Z t+τ (2.7)
V ∂C η ∂C ∂C √ 2 ∂C √ 3
= λu du + λu du + ηu dWu + φ ηu dWu .
t ∂V t ∂η t ∂V t ∂η

Since the expectation of Itô integrals is zero, the expected delta-hedged equity option
gains are given by

Z t+τ  Z t+τ 
∂C η ∂C
Et [Πt,t+τ ] = Et λVu du + Et λu du . (2.8)
t ∂V t ∂η

The expected option gains depend on the risk compensation components for the volatility
and volatility-of-volatility risks (λVt and ληt ), and the option price exposures to these two
∂C ∂C
sources of risks ( ∂V and ∂η
).

For tractability, following the literature let us further assume that the risk premium
structure is linear:

λVt = λV Vt , ληt = λη ηt , (2.9)

where λV is the market price of the variance risk and λη is the market price of the variance-
of-variance risk. We can further operationalize (2.8) by applying Itô-Taylor expansions
(Milstein (1995)). This gives us a linear factor structure (see details in Appendix A):

Et [Πt,t+τ ]
= λV βtV Vt + λη βtη ηt . (2.10)
St

Electronic copy available at: https://ssrn.com/abstract=2497759

The sensitivities to the risk factors are given by:

∞ ∞
X τ 1+n X τ 1+n
βtV = ΦVt,n > 0, βtη = Φηt,n > 0, (2.11)
n=0
(1 + n)! n=0
(1 + n)!

where ΦVt,n and Φηt,n are positive functions which depend on the moneyness of the option and
∂C ∂C
on the sensitivities ∂V
and ∂η
, respectively. Hence, the expected payoff on the delta-hedged
option position combines the risk compensations for the volatility and volatility-of-volatility
risks. The two risk compensations are given by the product of the market price of risk,
the exposure of the asset to the corresponding risk, and the quantity of risk. In particular,
options are positive-beta assets with respect to both volatility and volatility-of-volatility
risks. Hence, if investors dislike volatility and volatility-of-volatility risks so that their market
prices of risks are negative, the expected option payoffs are negative as well.

2.2 Dynamics of VIX Option Prices

The squared VIX index for a time to maturity τ is the annualized risk-neutral expectation
of the quadratic variation of returns from time t to t + τ , given by:

Z t+τ 
1
V IXt2 = EQ Vs ds . (2.12)
τ t t

Given our model assumptions, the VIX index is a function of the stock market variance:
V IXt = V IX(Vt ). For example, in a linear model where the variance drift θ(Vt ) is linear in
Vt , the squared VIX is a linear function of the stock market variance Vt .

Let Ft be the time t price of a VIX futures contract expiring at t + τ . Under the
assumptions of no arbitrage and continuous mark-to-market, Ft is a martingale under the

10

Electronic copy available at: https://ssrn.com/abstract=2497759

risk-neutral measure Q:

Ft = EQ Q
t [V IXt+τ ] = Et [V IX(Vt+τ )] .
(2.13)

Under our model structure, the futures price F is a function of the market variance Vt and
volatility of volatility η. Under economically plausible scenarios, the futures price is monotone
in the two volatility processes.5 Knowing Ft and ηt is sufficient for Vt , so we can re-write the
   
economic states Vt ηt in terms of Ft ηt .

Let Ct∗ be the time t price of a VIX call option, whose underlying is a VIX forward
contract. The option price is given by a twice differentiable function C ∗ of the state variables,
so that Ct∗ (K, τ ) = C ∗ (Ft , ηt , t). By Itô’s Lemma:

∂C ∗ ∂C ∗
dCt∗ = dFt + dηt + b∗t dt, (2.14)
∂F ∂η

for a drift component b∗t .

Under the risk-neutral measure Q, the discounted VIX option price process e−rf t Ct∗ is a
martingale, so it must have zero drift:

∂C ∗ Q ∂C ∗
D [Ft ] + (γ(ηt ) − ληt ) − rf Ct∗ + b∗t dt = 0, (2.15)
∂F ∂η

where DQ [Ft ] is the infinitesimal generator applied to F , i.e., the conditional expectation of
dFt under Q.
5
See also Zhang and Zhu (2006), Lu and Zhu (2010), and Branger et al. (2016) for VIX futures pricing
models.

11

Electronic copy available at: https://ssrn.com/abstract=2497759

 This implies that

∂C ∗ η ∂C ∗ ∂C ∗ Q
b∗t = rf Ct∗ + λ − γ(ηt ) − D [Ft ]
∂η t ∂η ∂F
(2.16)
∂C ∗ η ∂C ∗
= rf Ct∗ + λ − γ(ηt ),
∂η t ∂η

where the second line follows from the fact that Ft is a martingale under Q, so that its drift
is equal to zero.

Combining (2.16) with (2.14), we obtain the equation for the delta-hedged VIX option
gain:

t+τ Z t+τ
∂C ∗
Z
Π∗t,t+τ = ∗
Ct+τ − Ct∗− dFs − rf Cs∗ ds
t ∂F t
Z t+τ ∗ Z t+τ ∗
(2.17)
∂C η ∂C √
= λs ds + φ ηs dWs3 .
t ∂η t ∂η

The delta-hedged VIX option gain in Equation (2.17) is the counterpart to the delta-hedged
equity option gain in Equation (2.7). The difference comes from the fact that the short stock
position serving as the hedge in the case of equity options is funded at rf , while for VIX
futures the hedging position is zero cost.

Taking expectations, we can derive the corresponding expected gain on delta-hedged VIX
options:

Et Π∗t,t+τ
 
1 t+τ
Z  ∗ 
∂Cs η
= Et λ ds. (2.18)
Ft Ft t ∂ηs s

Similar to index options, the expected gain on delta-hedged VIX options depends on the
market price of volatility-related risks and the option exposure to these risks. However, in
contrast to index options, at least for short maturities τ , VIX options are mainly exposed to
∂Cs∗
volatility-of-volatility risks. Their exposure to these risks, captured by ∂ηs
, is expected to be
positive, so that the average level of the expected gains on VIX options and its dependence

12

Electronic copy available at: https://ssrn.com/abstract=2497759

on the volatility-of-volatility factor is governed by the market price of the volatility risks λη .
Unlike the expected gains for index options in (2.10), the expected gains for VIX options do
not generally admit a linear factor structure because the option price is no longer homoge-
neous in the underlying asset value. We provide a numerical example below to support our
arguments, and then consider the relative importance of the risks in the data.

3 Illustrative Numerical Example

3.1 Calibration

We calibrate the model to match key asset-pricing moments in the data. The Q-dynamics
of the model are given in Equations (2.2) and (2.9). As it is common in the literature we
choose θ(Vt ) = κV (V̄ − Vt ) and γ(ηt ) = κη (η̄ − ηt ). For simplicity, we assume that the three
Brownian motions W 1 , W 2 and W 3 are uncorrelated.

We choose parameter values for rf , κV , V̄ , λV , κη , η̄, λη , and φ. The parameters can be

found in Table 1. We choose rf and V̄ to match the unconditional level of risk-free rate and
the variance of the S&P500 index in our sample period, respectively. We choose λV , η̄, and κV
to target the level, the variance, and the persistence of V IX 2 , respectively. These moments
can be computed in closed form, because V IX 2 is linear in the state variables, as shown in
Appendix B.1. Finally, we calibrate λη , φ, and κη to match the level, the variance, and the
persistence of V V IX 2 . We run a simulation study to compute model-implied moments of
V V IX 2 , since they are not available in closed form.

13

Electronic copy available at: https://ssrn.com/abstract=2497759

3.2 Expected gains on delta-hedged VIX options

To calculate model-implied delta-hedged option gains, we calculate prices of VIX futures

and VIX options on a grid, as described in Appendix B.1. We use Equation (2.18) and
numerically solve the integral
t+τ
∂Cs∗ η
Z  
Et λ ds. (3.1)
t ∂ηs s

Details about the numerical procedure are given in Appendix B.2.

Figure 1 shows expected delta-hedged option gains divided by the futures price as a
function of η for different levels of the variance Vt . For the upper graph we choose λη = −5,
i.e. we assume that the market price of volatility-of-volatility risk is negative. For the lower
graph we assume λη = 5. For the strategies shown here, we always choose at-the-money
options. When the investor sets up a strategy in state (V0 , η0 ), the strike price of the option
he buys is equal to the VIX futures price in state (V0 , η0 ), i.e., to Ft (V0 , η0 , τF ).

Given a negative λη , we find that the expected gains are all negative.6 This is because
the derivatives of the VIX options with respect to η are positive for all grid points and for all
times to maturity. We also find that expected option gains are larger in absolute terms for
higher values of η. Since we truncate the process for V at zero, this effect is more pronounced
for low values of V : A high volatility of volatility not only increases the value of the VIX
option because the option price is convex in V , but also because it increases the upside
potential of the V IX without increasing the downside potential, due to the lower bound at
zero. Expected option gains are not exactly linear in ηt since the derivative of VIX option
prices with respect to η is not constant but decreasing in η (see Figure 2).

With a positive λη expected delta-hedged option gains are positive. In this case, all VIX
option prices are still increasing in the volatility of volatility. We can conclude that the
6
Quantitatively, the expected delta-hedged VIX option gains normalized by the VIX futures price, com-
puted at the average values of the volatility states, are about -0.1% for two-day returns, or -1.5% time-
aggregated to a monthly horizon. These values are close to the empirical estimates presented in Section
5.

14

Electronic copy available at: https://ssrn.com/abstract=2497759

sign of the expected option gain is pinned down by the sign of the market price of volatility
of volatility risk, and that the expected delta-hedged VIX option gains are increasing, in
absolute terms, in the volatility of volatility.

3.3 Delta computations

A key input into the computations of delta-hedged option returns is the delta of the option.
A typical approach in the literature is to approximate the unknown ”true” deltas by the
Black-Scholes implied ones; see Bakshi and Kapadia (2003), Duarte and Jones (2007), and
our subsequent analysis. In this section, we use our numerical model to compute and compare
the “true” and Black-Scholes deltas, and evaluate the adequacy of the approximation for the
VIX options.

The Black-Scholes delta for a call option on a futures contract is given by ∆ = e−rf τC N (d1 )
where N denotes the stadnard normal cumulative distribution function, and d1 is given as

log(Ft (V, η)) − log(K) + 12 IV 2 τC

d1 = √ . (3.2)
IV τC

Here, IV denotes the implied volatility of the underlying. We use the observed option and
futures prices and solve the Black Scholes formula for IV .

The Black-Scholes model assumes that the option’s underlying follows a geometric Brow-
nian motion. In our model Vt does not follow such a process. As a consequence, neither the
VIX nor VIX futures prices follow a proportional process. In particular, the variance of the
VIX futures price is largely driven by η, while its level is largely driven by V .

To evaluate the quality of the approximation, we compute the true delta from the model,
∂C ∗
∂F
. The two prices C ∗ and F are both functions of V and η. An innovation in F can be due
to an innovation in V or in η, and these two innovations may lead to different innovations in

15

Electronic copy available at: https://ssrn.com/abstract=2497759

C ∗ . The goal is to set up a portfolio that is neutral to innovations in the variance V , such
that the portfolio is in the end only exposed to innovations in the volatility of volatility. It
is therefore appropriate to consider the derivative in the V -direction only. We use the chain
rule to come up with
∂C ∗ /∂V
CF∗ v = (3.3)
∂F/∂V
∂C ∗
as a proxy for ∂F
. CF∗ v quantifies how much the VIX option price changes after a change
in F that is due to a change in V .

Figure 3 shows both variants of delta as functions of the state variables V and η. The
plotted deltas are for options with a strike price of F (V̄ , η̄). We find that the two variants
of delta have a similar shape, but CF∗ V is closer to 0 when the option is deep in the money
and closer to 1 when the option is far out of the money. The Black Scholes delta of the
at-the-money option is equal to 0.63 while CF∗ V of the at-the-money option is 0.60. In the
empirical part of our paper we use mostly at-the-money options and test if the results are
robust to variations in deltas. The impact of using either delta on model-implied option
gains is analyzed further in Appendix B.3.

4 Variance Measures

4.1 Construction of Variance Measures

The VIX index is a model-free, forward-looking measure of implied volatility in the U.S.
stock market, published by the Chicago Board Options Exchange (CBOE). The square of
30
the VIX index is defined as in Equation (2.12) where τ = 365
. Carr and Madan (1998),
Britten-Jones and Neuberger (2000), and Jiang and Tian (2005) show that V IX 2 can be
computed from the prices of call and put options with the same maturity at different strike

16

Electronic copy available at: https://ssrn.com/abstract=2497759

prices:

"Z #
St∗ ∞
2erf τ
Z
1 1
V IXt2 = Pt (K)dK + Ct (K)dK , (4.1)
τ 0 K2 St∗ K2

where K is the strike price, Ct and Pt are the put and call prices, St∗ is the fair forward price
of the S&P500 index, and rf is the risk-free rate. The VIX index published by the CBOE
is discretized, truncated, and interpolated across the two nearest maturities to achieve a
constant 30-day maturity.7 Jiang and Tian (2005) show through simulation analysis that
the approximations used in the VIX index calculation are quite accurate.

Since February 2006, options on the VIX have been trading on the CBOE, which give
investors a way to trade the volatility of volatility. As of Q3 2012, the open interest in front-
month VIX options was about 2.5 million contracts, which is similar to the open interest in
front-month S&P500 index option contracts.

We calculate our measure of the implied volatility of volatility using the same method
as the VIX, applied to VIX options instead of S&P500 options. The index, which has since
been published by the CBOE as the “VVIX index” in 2012 and back-filled, is calculated as:

Ft ∞
2erf τ
Z Z 
1 ∗ 1 ∗
VV IXt2 = P (K)dK + C (K)dK , (4.2)
τ 0 K2 t Ft K2 t

where Ft is the VIX futures price, and Ct∗ , Pt∗ are the prices of call and put options on the
VIX, respectively.8 The squared VVIX is calculated from a portfolio of out-of-the-money call
and put options on VIX futures contracts. It captures the implied volatility of VIX futures
returns over the next 30-days, and is a model-free, forward-looking measure of the implied
7
More details on the exact implementation of the VIX can be found in the white paper available on the
CBOE website: http://www.cboe.com/micro/vix
8
The official index is back-filled until 2007. We apply the same methodology and construct the index for
an additional year back to 2006. The correlation between our measure of the VVIX and the published index
is over 99% in the post-2007 sample. Our empirical results remain essentially unchanged if we restrict our
sample to only the post-2007 period.

17

Electronic copy available at: https://ssrn.com/abstract=2497759

volatility of volatility.

In addition to the implied volatilities, we can also compute the realized volatilities for the
stock market and the VIX. The construction here follows Barndorff-Nielsen and Shephard
(2004) using high-frequency, intraday data.9 Realized variance is defined as the sum of
squared high-frequency log returns over the trading day:

N
X
2
RVt = rt,j . (4.3)
j=1

Barndorff-Nielsen and Shephard (2004) show that RVt converges to the quadratic varia-
tion as N → ∞. We follow the standard approach of considering 5 minute return intervals.
A finer sampling frequency results in better asymptotic properties of the realized variance
estimator, but also introduces more market microstructure noise such as the bid-ask bounce
discussed in Heston, Korajczyk, and Sadka (2010). Liu, Patton, and Sheppard (2015) show
that the 5 minute realized variance is very accurate, difficult to beat in practice, and is
typically the ideal sampling choice in most applications combining accuracy and parsimony.

We estimate two realized variance measures, one for the S&P500 and one for the VIX.
For the S&P500, we use the S&P500 futures contract and the resulting realized variance
will be denoted RV SP X . For the VIX, we use the spot VIX index and denote the resulting
realized variance of the VIX by RV V IX , which is our measure of the physical volatility of
volatility. For robustness, we also entertain an alternative measure of the realized variance
RV V X∗ which is computed using the 5-minute VIX futures returns. The sample for this
measure is shorter, and starts in July 2012.
9
The data are obtained from http://www.tickdata.com.

18

Electronic copy available at: https://ssrn.com/abstract=2497759

4.2 Variance Dynamics

All of our variables are at the monthly frequency. The implied variance measures are given by
the index values at the end of the month, and the realized variance measures are calculated
over the past month and annualized. The sample for the benchmark measures runs from
February 2006 to December 2016.

Table 2 presents summary statistics for the implied and realized variance measures. While
the average level of the VIX is about 22%, the average level of the VVIX is much higher at
about 88%, which captures the fact that VIX futures returns are much more volatile than
market returns: volatility, itself, is very volatile. V V IX 2 is also much more volatile and
less persistent than V IX 2 , with an AR(1) coefficient of 0.33 compared to 0.81 for V IX 2 .
The VVIX exhibits relatively low correlation with the VIX, with a correlation coefficient
of about 0.27. The mean of realized variance for S&P500 futures returns is 0.024, which
corresponds to an annualized volatility of 15.5%. S&P500 realized variance is persistent
and quite strongly correlated to the VIX index (correlation coefficient of 0.88) and much
more weakly correlated to the VVIX index (correlation coefficient about 0.26). The realized
variance of VIX is strongly related to the VVIX index (correlation of 0.61), and to a much
lesser extent, the VIX index (correlation of 0.18).

In the last row of Table 2 we also consider the realized variance of VIX futures returns,
since the VIX futures contract (not the spot VIX) is the underying asset for VIX options and
there is no simple cost-of-carry relationship between VIX futures and spot VIX. Compared to
spot VIX realized variance, VIX futures realized variance is lower on average, less volatile,
more persistent, and more correlated with VIX. Recall, however, that our data for VIX
futures is from 2012m7 to 2016m12 because of the availability of tick data.

Figure 4 shows the time-series of the VIX and VVIX from February 2006 to December
2016. There are some common prominent moves in both series, such as the a peak during

19

Electronic copy available at: https://ssrn.com/abstract=2497759

the financial crisis. Notably, however, the VVIX also peaks during other times of economic
uncertainty, such as the summer of 2007 (quant meltdown, beginning of the subprime crisis),
May 2010 (Eurozone debt crisis, flash crash), August 2011 (U.S. debt ceiling crisis), and
August 2015 (sell-off driven by the Chinese stock market crash). The movements in VIX
during these events are far smaller than the spikes in the VVIX. The suggests that the VVIX
captures important uncertainty-related risks in the aggregate market, distinct from the VIX
itself.

In Figure 5 we present time-series plots for both S&P500 and VIX realized variances. As
shown in Panel A, the realized and implied variances of the stock market follow a similar
pattern, and S&P500 realized variance is nearly always below the implied variance. There
is a large spike in both series around the financial crisis in October 2008, at which point
realized variance exceeded implied variance. The difference between the mean of V IX 2 and
RV SP X is typically interpreted as a variance premium, which is the difference between end-
of-month model-free, forward-looking implied variance calculated from S&P500 index options
and the realized variance of S&P500 futures returns over the past month. Unconditionally,
the average level of the VIX (22%) is greater than the average level of the S&P500 realized
volatility (15.6%), so that the variance premium is positive, consistent with the evidence in
Bollerslev et al. (2009) and Drechsler and Yaron (2011). This also implies a negative market
price of volatility risk.

Panel B of the Figure shows the time series of the realized and implied volatility of the
VIX index. Generally, the implied volatility tends to increase at times of pronounced spikes
in the realized volatility. The implied volatility is also high during other times of economic
distress and uncertainty, such as May 2010 (Eurozone debt crisis and flash crash), and August
2011 (U.S. debt ceiling crisis). During normal times, the VVIX is above the VIX realized
variance, although during times of extreme distress we see the realized variance of VIX can
exceed the VVIX. The average level of the VVIX (87.8%) is greater than the average level

20

Electronic copy available at: https://ssrn.com/abstract=2497759

of the VIX realized volatility (81.8%), so that the volatility-of-volatility premium is also
positive.10 As we see in Table 2, VIX futures realized volatility is lower than VIX index
realized volatility, so the actual magnitude of the volatility-of-volatility risk premium is even
larger when compared to VIX futures. Similar to our discussion of the variance premium,
this evidence suggests that investors dislike volatility-of-volatility risks, and the market price
of these risks is negative.

In addition to unconditional moments, we can also analyze the conditional dependence

of volatility on volatility of volatility. Specifically, we consider the predictability of future
realized variances by the VIX and VVIX, in spirit of Canina and Figlewski (1993), Chris-
tensen and Prabhala (1998), and Jiang and Tian (2005) who use option implied volatilities to
predict future realized volatilities. Following that approach, we consider both univariate and
multivariate encompassing regressions to assess the predictability of future realized variances
by the VIX and VVIX.

In our main specification, the dependent variable is the realized variance (RV ) over the
next month, for both the S&P500 and the VIX. Univariate regressions test whether each
implied volatility measure (the VIX or the VVIX) can forecast future realized variances;
multivariate encompassing regressions compare the relative forecasting importance of the
VIX and VVIX and whether one implied volatility measure subsumes the information con-
tent of the other. The univariate regressions are restricted versions of the corresponding
multivariate encompassing regression. For the S&P500 the regression is given as

SP X
RVt+1 = β0 + β1 V IXt2 + β2 V V IXt2 + β3 RVtSP X + t+1 . (4.4)

10
Song (2013) shows that the average level of his VVIX measure, computed using numerical integration
rather than the model-free VIX construction, is lower than the average realized volatility of VIX. One of the
key differences between his and our computations is the frequency of returns used in the realized variance
computations. Consistent with the literature, we rely on 5-minute returns to compute the realized variances,
while Song (2013) uses daily returns.

21

Electronic copy available at: https://ssrn.com/abstract=2497759

Similarly for the VIX, we have:

V IX
RVt+1 = β0 + β1 V IXt2 + β2 V V IXt2 + β3 RVtV IX + t+1 . (4.5)

Our benchmark results are presented for all variables calculated in annualized variance units.

The first regression in Panel A of Table 3 shows that the VIX can forecast future re-
alized variance of S&P500 returns. This is consistent with the findings of Jiang and Tian
(2005). The VVIX can also forecast future S&P500 realized variance somewhat, although
the statistical significance is weaker than that of the VIX and the magnitude of the regres-
sion coefficient is several times smaller. In the encompassing regression, we see that the
VIX dominates the VVIX in forecasting future S&P500 realized variance. A one standard
deviation increase in V IX 2 is associated with a 0.6 standard deviation increase in the real-
ized variance of S&P500 returns next month. The coefficient on the VIX does not change
much when we include the VVIX. Including lags of the realized variances themselves do not
materially change the results. These results are consistent with our model specification.

Panel B of Table 3 shows our predictability results for VIX realized variance, which is
our proxy for physical volatility of volatility. The VIX is not significantly related to future
VIX realized variation; in fact, the point estimate is negative. The t-statistic is nearly zero,
and the adjusted R2 is below zero. In contrast, the VVIX is a significant predictor of future
VIX realized variation. The regression coefficient for the VVIX is about 0.6 in a univariate
regression, and is 0.7 in the multivariate regression. A one standard deviation increase in the
current value of V V IX 2 is associated with more than 0.2 standard deviation increase in next
month realized variance of VIX. In Panel C, we use VIX futures realized return volatility as
the proxy for physical volatility-of-volatility and also find that VVIX is a significant predictor
of future realized variance which drives out the VIX in a multivariate regression.

The empirical evidence suggests that fluctuations in the volatility of volatility are not

22

Electronic copy available at: https://ssrn.com/abstract=2497759

directly related to the level of the volatility itself. This is consistent with our two-volatility
model specification in Section 2. In many reduced form and structural models, the volatility
of volatility is directly linked to the level of the volatility. For example, Heston (1993) models
volatility as following a Cox, Ingersoll, and Ross (1985) square-root process. In that case, the
level of volatility itself should forecast future realized volatility of volatility. The evidence in
the data does not support this assumption, and supports richer dynamics of the volatility
process with separate movements in the volatility of volatility.

5 Evidence from Options

In this section we analyze the implications of equity and VIX option price dynamics for the
pricing of volatility and volatility-of-volatility risks in the data. Our model suggests that the
market prices of volatility and volatility-of-volatility risks determine the key properties of the
cross-section and time-series of delta-hedged equity and VIX option gains. Specifically, if the
market prices of volatility and volatility-of-volatility risks are negative, the average delta-
hedged equity and VIX option gains are also negative. In the cross-section, the average
returns are more negative for option strategies which have higher exposure to the volatility
and volatility-of-volatility risks. Finally, in the time series higher volatility and volatility of
volatility predicts more negative gains in the future. We evaluate these model predictions in
the data, and find a strong support that both volatility and volatility-of-volatility risks are
priced in the option markets, and have negative market prices of risks.

23

Electronic copy available at: https://ssrn.com/abstract=2497759

5.1 Delta-Hedged Option Gains

We consider discrete-time counterparts to the continuously-rebalanced delta-hedged gains in

Equations (2.7) and (2.17):

N −1 N −1
X
X τ
Πt,t+τ = Ct+τ − Ct − ∆tn Stn+1 − Stn + rf (∆tn Stn − Ct ) ,
| {z } N
option gain/loss |n=0 {z } |n=0 {z }
delta hedging gain/loss risk-free rate
N −1 N −1
(5.1)
X τ X
Π∗t,t+τ = Ct+τ
∗
− Ct∗ − rf Ct∗ .


∆tn Ftn+1 − Ftn −
| {z } N
option gain/loss |n=0 {z } |n=0 {z }
delta hedging gain/loss risk-free rate

∂Ctn
∆tn indicates the option delta at time tn , e.g., for a call option, ∆tn = ∂Stn
, and N is the
number of trading days in the month. This discrete delta-hedging scheme is also used in
Bakshi and Kapadia (2003) and Bertsimas et al. (2000).

At the close of each option expiration, we look at the prices of all options with non-zero
open interest and non-zero trading volume. We take a long position in the option, and
hedge each day using the ∆ according to the Black-Scholes model, with the net investment
earning the risk-free interest rate appropriately.11 To minimize the effect of recording errors,
we discard options that have implied volatilities below the 1st percentile or above the 99th
percentile. All options have exactly one calendar month to maturity; S&P500 options expire
on the third Friday of every month, while VIX options expire on the Wednesday that is 30
days away from the third Friday of the following month.

Table 4 shows average index and VIX delta-hedged option gains in our sample. We
separate options by call or put, and group each option into four bins by moneyness to obtain
Π
eight bins for both S&P500 and VIX options. The column S
gives the delta-hedged option
11
This requires an estimate of the implied volatility of the option, which may require an option price. We
use implied volatilities directly backed out from market prices of options whenever possible; if an option does
not have a quoted price on any intermediate date, we fit a cubic polynomial to the implied volatility curve
given by options with quoted prices, and back out the option’s implied volatility. This is similar to typical
option position risk management done by professional traders.

24

Electronic copy available at: https://ssrn.com/abstract=2497759

 Π
gain scaled by the initial index level, and the column C
gives the delta-hedged option gain
scaled by the initial option price, which can be interpreted more readily as a “return” in the
traditional sense.

Panel A of Table 4 shows that on average out-of-the-money delta-hedged S&P500 call op-
tions have significantly negative returns. Likewise, delta-hedged put options on the S&P500
also have significantly negative returns at all levels of moneyness. To guard against potential
outliers in option returns, we show that the results are similar using medians rather than
average values; in percentages, delta-hedged index put options are negative about 80% of
the time. This evidence is largely consistent with Bakshi and Kapadia (2003). In the model,
negative average returns on delta-hedged index puts imply that volatility and/or volatility-
of-volatility risks have a negative market price of risk. S&P500 option gains display mild
positive serial correlation, which we will account for in our time-series predictive regressions
in the later sections.

Panel B of Table 4 shows average returns for delta-hedged VIX options. The average
delta-hedged VIX option returns are negative and statistically significant in all bins except
for out-of-the-money puts and in-the-money calls, which are marginally significant. Call
options lose more money as they become more out of the money, regardless of whether we
are scaling by the index or by the option price. Estimates of the loss for call options ranges
from -0.25% of the index value for in-the-money VIX calls to -1.20% of the index value for
out-of-the-money VIX calls. When viewed as a percentage of the option price, delta-hedged
VIX calls return about -1% per month at the money, and -20% out-of-the-money. The call
option returns are negative 60% to 80% of the time. The results for VIX put options are
similar. In the model, negative average returns for delta-hedged VIX options imply that
investors dislike volatility-of-volatility risks, and are systematically paying a premium hedge
against increases in the volatility of volatility. This suggests that the price of the volatility-
of-volatility risk is negative. VIX option gains exhibit weak negative serial correlation, which

25

Electronic copy available at: https://ssrn.com/abstract=2497759

we also account for in our time-series predictive regressions in the later sections. For both
the S&P500 and VIX, the delta-hedged option gains are quite volatile.

In the next section, we provide further direct evidence by controlling for the exposures
of the delta-hedged option positions to the underlying risks.

5.2 Cross-Sectional Evidence

As shown in the previous section, the average delta-hedged option gains are negative for
S&P500 and VIX options. Our model further implies (see Equations (2.10) and (2.18))
that options with higher sensitivity to volatility and volatility-of-volatility risks should have
more negative gains. To compute the estimates of option exposures to the underlying risks,
we follow the approach of Bakshi and Kapadia (2003) which relies on using the Black and
Scholes (1973) model to proxy for the true option betas.

Specifically, to compute the proxy for the option beta to volatility risk, we consider the
vega of the option:

r
∂C τ − d21 d2
1
=S e 2 ∝ e− 2 , (5.2)
∂σ 2π
h   i
1 S σ2
where d1 = √
σ τ
log K + rf − q + 2
τ , q is the dividend yield, and σ is the implied
volatility of the option. This approach allows us to compute proxies for the exposures of
equity options to volatility risks, and of VIX options to the volatility-of-volatility risks.

To illustrate the relation between the moneyness of the option and the vega-measured
exposure of options to volatility risks, we show the option vega as a function of option
moneyness in Figure 6. Vega represents an increase in the value of the option as implied
volatility increases by one percentage point. Higher volatility translates into higher future
profits from delta-hedging due to the convexity effect; hence both call and put options have

26

Electronic copy available at: https://ssrn.com/abstract=2497759

strictly positive vegas. Further, since the curvature of option value is the highest for at-the-
money options, at-the-money options have the highest vega in the cross-section, and thus the
largest exposure to volatility risks. An alternative way to proxy for the option sensitivity to
volatility risks is to use the “gamma” of the option, which represents the second derivative
∂2C
of the option price to the underlying stock price: Γ = ∂S 2
. As shown in Figure 6, the shape
of the vega and gamma functions are almost identical, hence, the implied cross-sectional
dispersion in volatility betas by moneyness are very similar as well.

To capture the sensitivity of option prices to the volatility of volatility, we compute the
Black and Scholes (1973) second partial derivative of the option price with respect to the
volatility, which is known in “volga” for “volatility gamma”. Volga is calculated as:

r
∂ 2C
   
τ − d21 d1 d2 ∂C d1 d2
=S e 2 = , (5.3)
∂σ 2 2π σ ∂σ σ

√
where d2 = d1 − σ τ . Figure 7 shows the plot of volga as a function of the moneyness of
the option. Volga is positive, and exhibits twin peaks with a valley around at-the-money.
At-the-money options are essentially pure bets on volatility, and are approximately linear
in volatility (see Stein (1989)). Therefore, the volga is the lowest for at-the-money options.
Deep-out-of-the-money options and deep-in-the-money options do not have much sensitivity
to volatility of volatility either, since for the former it is a pure directional bet, and for
the latter the option value is almost entirely comprised of intrinsic value. Options that are
somewhat away from at-the-money are most exposed to volatility-of-volatility risks.

Table 5 shows our cross-sectional evidence from the regressions of average option re-
turns on our proxies of options’ volatility and volatility-of-volatility betas. Panel A shows
univariate and multivariate regressions of delta-hedged S&P500 option gains scaled by the
index on the sensitivities of the options to volatility and volatility-of-volatility risks. Our

27

Electronic copy available at: https://ssrn.com/abstract=2497759

encompassing regression for delta-hedged S&P500 options is:

i
Πit,t+τ
GAIN St,t+τ = = λ̃1 V EGAit + λ̃2 V OLGAit + γt + it,t+τ . (5.4)
St

Since each date includes multiple options, as in Bakshi and Kapadia (2003) we allow for a
date-specific component in Πit,t+1 due to the option expirations, i.e., we include time-fixed
effects represented by γt . Conceptually, our approach is related to Fama and MacBeth (1973)
regressions. Instead of estimating risk betas in the first stage, due to the non-linear structure
of option returns, we measure the exposures from economically motivated proxies for the risk
sensitivities.12

The results in Panel A show that both volatility and the volatility of volatility are priced
in the cross-section of delta-hedged S&P500 option returns. While univariate estimates for
vega are insignificantly different from zero, univariate estimates for volga are significantly
negative at -0.005 with a t-statistic of -4.28. In multivariate regressions, once we control
for the effect of volga, the loading on vega becomes signicantly negative as we expect from
the volatility premium. This demonstrates that especially in a period where we observe
large spikes in volatility, it is important to control for exposure to volatility-of-volatility to
properly see the impact of exposure to volatility on delta-hedged S&P500 option returns.

Panel B of Table 5 presents cross-sectional results for delta-hedged VIX options. As Equa-
tion (2.17) shows, delta-hedged VIX options are mainly exposed to volatility-of-volatility
risks, and the vega for VIX options captures the sensitivity of VIX options to innovations
in the volatility of volatility. The coefficient on vega of -1.23 is negative and statistically
significant. Thus, in the cross-section of both S&P500 options and VIX options, we find
strong evidence of a negative price of volatility-of-volatility risk.
12
Song and Xiu (2016) demonstrate an alternative method of estimating risk sensitivities nonparametrically
using local linear regression methods.

28

Electronic copy available at: https://ssrn.com/abstract=2497759

5.3 Time-Series Evidence

In the model, time-variation in the expected delta-hedged option gains is driven by V and η,
and the loadings are determined by the market prices of volatility and volatility-of-volatility
risk. We group options into the same bins as we used for average returns in Table 4, and
average the scaled gains within each bin, so that we have a time-series of option returns
for each moneyness bin. To examine the contribution of both risks for the time-variation in
expected index option payoffs, we consider the following regression:

i
Πit,t+τ
GAIN St,t+τ = = β0 + β1 V IXt2 + β2 V V IXt2 + γGAIN St−τ
i
+ ui + it+τ , (5.5)
St

where we include fixed effects ui to account for the heterogeneity in the sensitivity of options
in different moneyness bins to the underlying risks. We regress the delta-hedged option gain
scaled by the index from expiration to expiration on the value of the VIX and VVIX indices
at the end of the earlier expiration; in other words, we run one-month ahead predictive
regressions of delta-hedged option returns on the VIX and VVIX. We include lagged gains
to adjust for serial correlation in the residuals, following Bakshi and Kapadia (2003).

Panel A of Table 6 shows the regression results for the index options. The univariate
regression of delta-hedged S&P500 option gains on V IX 2 yields a negative coefficient which
is statistically significant, consistent with Bakshi and Kapadia (2003). V V IX 2 also predicts
future S&P500 option gains with a negative coefficient, and in a multivariate regression on
both VIX and VVIX we find both loadings are negative and statistically significant.

It is important to point out that the true loadings on the volatility factors are time-
varying, and generally depend on the underlying volatility and volatility-of-volatility states
(see, e.g., Equation (2.11)). As shown in Bakshi and Kapadia (2003), these loadings are in-
sensitive to volatility for at-the-money options. However, away from the money, the loadings
may depend on the variance risk factors, so that the relation between delta-hedged gains and

29

Electronic copy available at: https://ssrn.com/abstract=2497759

the variance-related risk factors is non-linear. To guard against finding a spurious coefficient
on the VVIX, we repeat the exercise using only at-the-money options, and find very similar
results.

Panel B of Table 6 shows the corresponding evidence for VIX options. The VVIX nega-
tively and significantly predicts future VIX option gains. This is consistent with a negative
market price of volatility-of-volatility risk. The results remain very similar for only at-the-
money options.

6 Robustness

6.1 Alternative Variance Specifications

Our results for the predictability of realized by implied variance are robust to alternative
specifications of volatility. Specifically, we consider regressing in volatility units or log-
volatility units, rather than variance. The robustness regressions follow the form:

q
x
RVt,t+1 = β0 + β1 V IXt + β2 V V IXt + t+1 (6.1)
q
x
ln RVt,t+1 = β0 + β1 ln V IXt + β2 ln V V IXt + t+1 (6.2)

where x refers to SP X or V IX.

In Table 7, we see that the point estimates and significance are very close to our baseline
specification in variance units.

30

Electronic copy available at: https://ssrn.com/abstract=2497759

6.2 Sensitivity to Jump Risk Measures

The evidence in our paper highlights the roles of the volatility and volatility-of-volatility
factors, which are driven by smooth Brownian motion shocks. In principle, the losses on
delta-hedged option portfolios can also be attributed to large, discontinuous movements
(jumps) in the stock market and in the market volatility. In this section we verify that
our empirical evidence for the importance of the volatility-related factors is robust to the
inclusion of jump measures considered in the literature.

Specifically, we consider three measures of jump risks, which we construct for the S&P500
and the VIX. Our first jump measure corresponds to the slope of the implied volatility curve:

SLOP E SP X = σOT
SP X SP X
M − σAT M ,
(6.3)
V IX V IX V IX
SLOP E = σOT M − σAT M.

The OT M contract for the S&P500 options is defined as a put option with a moneyness
closest to 0.9, and for VIX options as a call option with a moneyness closest to 1.1. In
both cases, the AT M option has moneyness of 1. These slopes are positive for both index
and VIX options. Positive slope of the index volatility smile is consistent with the notion
of negative jumps in market returns (see e.g. Bates (2000), Pan (2002), Eraker, Johannes,
and Polson (2003)), while the fact that the implied volatility curve for VIX options slopes
upwards is consistent with positive volatility jumps (Drechsler and Yaron (2011) and Eraker
and Shaliastovich (2008), among others). In this sense, these slope measures help capture
the variation in the market and volatility jumps in the economy.

Our second jump measure incorporates the whole cross-section of option prices, beyond
just the slope of the smile. It is based on the model-free risk-neutral skewness of Bakshi

31

Electronic copy available at: https://ssrn.com/abstract=2497759

et al. (2003):

erf τ Wt,t+τ − 3µt,t+τ erf τ Vt,t+τ + 2µ3t,t+τ

SKEW (t, t + τ ) =  rτ 3/2 , (6.4)
e f Vt,t+τ − µ2t,t+τ

where Vt,t+τ , Wt,t+τ , Xt,t+τ are given by the prices of the volatility, cubic, and quartic con-
tracts, and µt,t+τ is function of them. Importantly, these measures are computed model-free
using only observed option prices. The details for the computations are provided in the
Appendix C.

Finally, our third measure of jump risks is based on high-frequency index and VIX data,
rather than option prices. It corresponds to the realized jump intensity, and relies on the
bipower variation methods in Barndorff-Nielsen and Shephard (2004), Huang and Tauchen
(2005), and Wright and Zhou (2009). Specifically, while the realized variance defined in (4.3)
captures both the continuous and jump variation, the bipower variation, defined as

 M
X
π M
BVt = |rt,j−1 ||rt,j | (6.5)
2 M −1 j=2

measures the amount of continuous variation returns. Hence, we can use the following test
statistic to determine if there is a jump on any given day:

RVt −BVt
RVt
Jt = q , (6.6)
θ QVt
M
max 1, BV 2
t

π 2


where θ = 2
+ π − 5, and QVt is the quad-power quarticity defined in Huang and Tauchen
(2005) and Barndorff-Nielsen and Shephard (2004). The test statistic is standard normally
distributed. We flag the day as having a jump if the probability exceeds 99.9% both for
index returns and for the VIX. These cut-offs imply an average frequency of jumps of once
every two months for the index, and about three jumps a month for the VIX. This is broadly
consistent with the findings of Tauchen and Todorov (2011), who find that VIX jumps tend

32

Electronic copy available at: https://ssrn.com/abstract=2497759

to happen much more frequently than S&P500 jumps. Over a month, we sum up all the
days where we have a jump, and we define our jump intensity measure on a monthly level
as:

T −1
1X
RJt = Jt+i ,
T i=0

where T is the number of trading days in the given month.

We use the jump statistics to document the robustness of the link between the volatility
and volatility-of-volatility factors and options gains. We consider a regression of the form

i
GAIN St,t+τ = β0 + β1 V IXt2 + β2 V V IXt2 + β3 JU M Pt + γGAIN St−τ
i
+ ui + it+τ (6.7)

where JU M Pt is one of the above jump risk proxies. We use index jump measures for index
gains, and VIX jump measures for VIX gains.

Table 8 displays our results. Both for S&P500 and VIX options, controlling for SLOP E
does not change the ability of V IX and V V IX to predict future delta-hedged option gains.
Both factors are still significant, and the point estimates βˆ1 and βˆ2 remain largely unchanged.
SLOP E itself is not significant at conventional levels for S&P500 options but significantly
positive for VIX options. The risk-neutral skewness also does not affect the predictive ability
of V IX and V V IX, although, again, it is significant for VIX options. The two estimates
of β3 have the correct signs, since skewness is negative for S&P500 options and positive for
VIX options; this is broadly similar to the findings of Bakshi and Kapadia (2003). Finally,
we control for realized jump intensity RJ and we see a similar result where the statistical
significance of V IX and V V IX as well as their point estimates are largely unchanged. RJ
seems to marginally predict future S&P500 option gains; for VIX options, however, RJ does
not seem to predict future VIX option gains.

Hence, our evidence suggests that the VIX and VVIX have a significant impact on op-

33

Electronic copy available at: https://ssrn.com/abstract=2497759

tion returns even in the presence of stock market and volatility jumps. We leave a formal
treatment of jumps for future research.

6.3 Sub-samples

To further investigate robustness of our results, we split our data into two sub-periods of
roughly equal length, a pre-2012 and a post-2012 period. Tables 9 and 10 show the results in
the two sub-periods. As we can see from the tables, the sub-period results are very similar
to the full sample results in Table 4, and show large negative average S&P500 and VIX
option gains, consistent with investors pricing volatility and volatility-of-volatility risks with
negative market prices of risk.

6.4 Delta computations

We have used delta hedging to insulate our option returns from changes in the direction
of the underlying, so that we can better focus on the volatility and volatility-of-volatility
risks. We use Black-Scholes deltas to determine the appropriate hedge. While we have used
simulation to examine the potential impact of using Black-Scholes deltas on our results, we
also provide additional empirical robustness in a manner similar to Coval and Shumway
(2001) by perturbing the deltas used in the option hedging. In Tables 11 and 12, we set
the deltas to 0.95 and 1.05 times the value from the Black-Scholes model, respectively. We
see that our results are not sensitive to the choice of using Black-Scholes deltas for option
hedging and are quite robust.

34

Electronic copy available at: https://ssrn.com/abstract=2497759

7 Conclusion

Using S&P500 and VIX options data, we show that a time-varying volatility of volatility is
a separate risk factor which affects option returns, above and beyond market volatility risks.
We measure volatility risks using the VIX index, and volatility-of-volatility risk using the
VVIX index. The two indices, constructed from the index and VIX option data, capture
the ex-ante risk-neutral uncertainty of investors about future market returns and VIX in-
novations, respectively. The VIX and VVIX have separate dynamics, and are only weakly
related in the data: the correlation between the two series is under 30%. On average, risk-
neutral volatilities identified by the VIX and VVIX exceed the realized physical volatilities
of the corresponding variables in the data. Hence, the variance premium and variance-of-
variance premium for VIX are positive, which suggests that investors dislike variance and
variance-of-variance risks.

We show the pricing implications of volatility and volatility-of-volatility risks using op-
tions market data. Average delta-hedged option gains are negative, which suggests that
investors pay a premium to hedge against innovations in not only volatility but also the
volatility of volatility. In the cross-section of both delta-hedged S&P500 options and VIX
options, options with higher sensitivities to volatility-of-volatility risk earn more negative
returns. In the time-series, higher values of the VVIX predict more negative delta-hedged
option returns, for both S&P500 and VIX options.

Our findings are consistent with a no-arbitrage model which features time-varying market
volatility and volatility-of-volatility factors. The volatility factors are priced by the investors,
and in particular, volatility and volatility of volatility have negative market prices of risks.

35

Electronic copy available at: https://ssrn.com/abstract=2497759

A Delta-Hedged Equity Options
 0
The state vector is xt = St Vt ηt . Under the linear risk premium structure, λVt = λV Vt and

ληt = λη ηt . Note that since Ct is homogeneous of degree 1 in the underlying St and the strike price
∂C ∂C
K, ∂V and ∂η are also homogeneous of degree 1 in St and K. Define a pair of functions:

∂Ct
g1 (xt ) = λVt
∂Vt
(A.1)
∂Ct
g2 (xt ) = ληt
∂ηt

We can re-write Equation (2.8) as:

Z t+τ  Z t+τ 
Et [Πt,t+τ ] = Et g1 (xu )du + Et g2 (xu )du . (A.2)
t t

Define operators L and Γ such that:

∂[.] ∂[.] ∂[.] ∂[.]

L[.] dt = µt St dt + θ(Vt )dt + γ(ηt )dt + dt
∂S ∂V ∂η ∂t
1 ∂ 2 [.] 1 ∂ 2 [.] 1 ∂ 2 [.]
+ 2
[dSt , dSt ] + 2
[dVt , dVt ] + [dηt , dηt ]
2 ∂S 2 ∂V 2 ∂η 2
(A.3)
∂ 2 [.] ∂ 2 [.] ∂ 2 [.]
+ [dSt , dηt ] + [dSt , dVt ] + [dVt , dηt ]
∂S∂η ∂S∂V ∂V ∂η
 
∂[.] p ∂[.] √ ∂[.] √
Γ[.] = St Vt , ηt , φ ηt .
∂S ∂V ∂η

Then, for u > t, Itô’s Lemma implies that:

Z u Z u
0
g1 (xu ) = g1 (xt ) + Lg(xu0 )du + Γg(xu0 )dWu0 .
t t

The integral in the first expectation on the right-hand side of Equation (A.2) becomes:

Z t+τ Z t+τ  Z u Z u 
0
g1 (xu )du = g1 (xt ) + Lg(xu0 )du + Γg(xu0 )dWu0 du
t t t t
1 1
= g1 (xt )τ + Lg1 (xt )τ 2 + L2 g1 (xt )τ 3 + ... + Itô Integrals
2 6
∞
X τ 1+n
= Ln g1 (xt ) + Itô Integrals,
(1 + n)!
n=0

36

Electronic copy available at: https://ssrn.com/abstract=2497759

and likewise for the second integral in (A.2). We can use this to re-write (A.2) as:

Z t+τ  Z t+τ 
Et [Πt,t+τ ] = Et g1 (xu )du + Et g2 (xu )du
t t
∞ ∞ (A.4)
X τ 1+n X τ 1+n
= Ln [g1 (xt )] + Ln [g2 (xt )] .
(1 + n)! (1 + n)!
n=0 n=0

Note that g1 (xt ) = α1 (Vt , τ ; K)St , and g2 (xt ) = α2 (ηt , τ ; K)St . By Lemma 1 of Bakshi and Kapadia

(2003), Ln [g1 (xt )] and Ln [g2 (xt )] will also be proportional to St , which implies that:

Ln [g1 (xt )] = λV Vt ΦVt,n St ∀n

Ln [g2 (xt )] = λη ηt Φηt,n St ∀n.

Therefore, we have:

∞ ∞
X τ 1+n X τ 1+n
Et [Πt,t+τ ] = Ln [g1 (xt )] + Ln [g2 (xt )]
(1 + n)! (1 + n)!
n=0 n=0

= St λV βtV Vt + λη βtη ηt ,
 

which implies that:

Et [Πt,t+τ ]
= λV βtV Vt + λη βtη ηt , (A.5)
St

where the sensitivities to the risk factors are given by:

∞
X τ 1+n
βtV = ΦV > 0
(1 + n)! t,n
n=0
∞ (A.6)
τ 1+n
βtη = Φη > 0.
X
(1 + n)! t,n
n=0

∂Ct ∂Ct
The betas are positive since ∂Vt > 0 and ∂ηt > 0.

37

Electronic copy available at: https://ssrn.com/abstract=2497759

B Delta-Hedged VIX Options

Unlike the expected gains for index options in (2.10), the expected gains for VIX options do not

generally admit a linear factor structure because the option price is no longer homogeneous in

the underlying asset value. We provide a numerical example below to investigate the patterns in

delta-hedged VIX option gains and how they are related to the volatility of volatility.

B.1 Simulation study for model calibration

The squared V IX can be calculated in closed form within the model and is given by

V IXt2 = A + B Vt (B.1)

where the coefficients A and B are

" V
# V
κV V̄ 1 − e−(κV +λ )τ 1 − e−(κV +λ )τ
A= 1− , B= (B.2)
κV + λV (κV + λV )τ (κV + λV )τ

To calculate the time series moments of V V IX 2 , we proceed as follows: We consider a grid of

states (Vt , ηt ), where the grid spans values between 0 and 3 times the steady state of both variables.

The grid points serve as starting values for N paths of the state variables that we draw according to

a discretized version of the dynamics in Equation (2.2). In particular, since ηt follows a square-root

process, we draw ηt+1 given ηt from a non-central χ2 -distribution

(1 − e−(κη +λη )∆t )φ2 2 4κη e−κη ∆t

 
n n n
ηt+∆t = χd η (ωη,t+∆t ) (B.3)
4(κη + λη ) (1 − e−(κη +λη )∆t )φ2 t

for each n = 1, . . . , N , with d degrees of freedom where

4κV κη + λη
d= η̄. (B.4)
φ2 κV + λV

38

Electronic copy available at: https://ssrn.com/abstract=2497759

The increments in V are approximately Gaussian with

 
n
Vt+∆t = N Vtn + [κV V̄ − (κV + λV )Vtn ]∆t, ∆t ηtn (ωV,t+∆t
n
). (B.5)

To ensure non-negative realizations of the variance, we truncate the normal distribution at zero.

To make the simulation results more reliable in small samples N , we use the same pseudo random
n
numbers (ωη,t+∆t n
, ωV,t+∆t ) for each grid point.

We calculate the prices of VIX futures that mature in t + τF . We assume a time to maturity τF
τF
of one month and choose ∆t = 30 . For a given parameterization and a particular vector of initial

state variables (Vt , ηt ), we simulate the distribution of Vt+τF and ηt+τF . We then calculate the time

t futures price by
N
1 Xq n
Ft (Vt , ηt , τF ) = A + B Vt+τ (B.6)
N F
n=1

The next step is calculate a grid of prices of VIX options that also mature in one month. These

options are written on VIX futures that have a time to maturity of one month at the maturity of

the option. To do so, we simulate state variables under Q as described above and calculate the VIX

futures prices at maturity of the option using the grid of futures prices calculated earlier. Since the

terminal values of the drawn state variables are between the points of the futures prices grid, we

interpolate (or extrapolate if a state variable is outside the grid) the futures prices. The resulting
n
futures price of the n-th path given initial values of (Vt , ηt ) is denoted by Ft+τ (Vt , ηt , τF ). The call
C

and put prices are

N
1 −rf τC X n
Ct∗ (Vt , ηt , τC , K) = e (Ft+τC (Vt+τC , ηt+τC , τF ) − K)+ ,
N
n=1
(B.7)
N
1 X
Pt∗ (Vt , ηt , τC , K) = e−rf τC n
(K − Ft+τ C
(Vt+τC , ηt+τC , τF ))+
N
n=1

where K denotes the strike price. We use a grid of strike prices that ranges across values between

0.4K̄ and 1.6K̄, where K̄ is the steady state futures price Ft (V̄ , η̄, τC + τF ).

To calculate V V IX 2 , we use VIX option prices calculated above and proceed similarly to the

39

Electronic copy available at: https://ssrn.com/abstract=2497759

procedure that is applied to data. We numerically solve the integral

Ft (Vt ,ηt ,τC +τF )

2erf τC h
Z
1 ∗
V V IXt2 (Vt , ηt ) = P (Vt , ηt , τC , K)dK
τC 0 K2 t
Z ∞ (B.8)
1 ∗ i
... + 2
C t (Vt , ηt , τC , K)dK
Ft (Vt ,ηt ,τC +τF ) K

using our grid of strike prices as discretization. Futures prices Ft (·, ·, τC + τF ) with longer maturity

are calculated as described above.

Finally, we draw one long path of the state variables under P. For that purpose, we proceed as

described above but use the P-dynamics of the state variables:

(1 − e−κη ∆t )φ2 2 4κη e−κη ∆t

 
0
ηt+∆t = χd η t (ωη,t+∆t ), (B.9)
4κη (1 − e−κη ∆t )φ2

with d degrees of freedom where

4κη
d= η̄ (B.10)
φ2

and
 
0
Vt+∆t = N Vt + κV (V̄ − Vt )∆t, ∆t ηt (ωV,t+∆t ). (B.11)

Using our grid of V IX and V V IX, we can now study the time series properties of the model-implied

quantities and compare them with the empirical moments.

B.2 Numerical procedure to solve Equation (3.1)

∂Cs∗
To solve the integral in Equation (3.1) numerically, we approximate the continuous path of ∂ηs by

a discretization with step size ∆t:

N I−1 ∗ n n
t+τ
∂Cs∗ η 1 X X ∂Ct+i∆t (Vt+i∆t , ηt+i∆t , τC − i∆t, K) η
Z  
Et λs ds ≈ λt+i∆t ∆t, (B.12)
t ∂ηs N ∂ηt+i∆t
n=1 i=0

where I∆t = τ . We implement this equation in the following way: First, we compute grids of

VIX option prices according to the procedure outlined in Appendix B.1 also for shorter times to

40

Electronic copy available at: https://ssrn.com/abstract=2497759

maturity, in particular, τC − i∆t for i = 0, . . . , I. Second, we approximate the derivative of C ∗ with

respect to η using finite differences, i.e., at the grid point (VmV , ηmη ), we use

∂C ∗ (VmV , ηmη ) C ∗ (VmV , ηmη +1 ) − C ∗ (VmV , ηmη −1 )

≈ . (B.13)
∂ηmη ηmη +1 − ηmη −1

Third, we again simulate innovations in the state variables under P. After each simulation step,

we end up in states which are between two grid points or even outside the grid. Because we do not

have derivatives at these points, we proceed as follows. We start with the calculation of expected

option gains over the last time step from t + τC − ∆t to t + τC . This gives us a grid of expected

option gains over this short period. We then simulate increments in the state variables in the period

from t + τC − 2∆t to t + τC − ∆t, starting again with the usual (V, η)-grid and add up expected

gains over this period with expected gains over the final period given the state that is drawn using

the scheme that approximates the P dynamics (see Section 3.1). We do so by interpolating the grid

produced before. We use the same procedure and iterate back until time t.

B.3 Spread in trading strategy using different deltas

Instead of using Equation (2.18) to calculate expected option gains, one could alternatively simulate

realized gains from the trading strategy and take the average across realizations. For this purpose,

we would have to calculate deltas of the VIX options and set up a delta hedged portfolio. This

procedure, however, has the disadvantage that we need much larger samples to obtain stable results,

because of the fluctuations in the Brownian motion W3 that are averaged out when taking the

expectation (see Equation (2.17)). While this task is very costly in terms of computation time, one

can look at realized delta hedged option gains path by path and study the difference between gains

when different deltas are used for setting up the hedge portfolio.

We now compare the difference between gains from our trading strategy when using the Black-

Scholes delta or the “true” delta CF∗ V (see Section 3) for setting up the hedge portfolio. Even with

noisy estimates of the expected gains, the difference between both gains is informative because we

41

Electronic copy available at: https://ssrn.com/abstract=2497759

use the same paths of the state variables in the simulation of the two strategies. Thus, the sampling

error widely cancels out when considering the difference.

Figure B.1 shows the difference between delta hedged option gains when using either delta. For

moderate values of V the difference is close to zero. When V is high, the strategy that uses the

Black-Scholes delta yields less negative returns than the strategy that uses CF∗ V . In the light of

this result, we can conclude that using Black-Scholes deltas is a rather conservative strategy. In

our empirical exercise, we would expect even more pronounced gains when using CF∗ V , which is,

however, not observable in the data.

Figure B.1: Difference Between Delta-Hedged Option Gains

The figure plots the difference between the delta-hedged option gains using the model-implied versus the Black-Scholes delta.
All options are at-the-money.

42

Electronic copy available at: https://ssrn.com/abstract=2497759

C Risk-Neutral Skewness

The prices of the volatility, cubic, and quartic contracts Vt,t+τ , Wt,t+τ , Xt,t+τ are given

∞ 2(1 − log SKt ) St 2(1 + log SKt )

Z Z
Vt,t+τ = Ct (t + τ ; K)dK + Pt (t + τ ; K)dK,
St K2 0 K2
∞ 6 log SKt − 3(log SKt )2
Z
Wt,t+τ = Ct (t + τ ; K)dK
St K2
6 log SKt + 3(log SKt )2
Z St
− Pt (t + τ ; K)dK,
0 K2
12(log SKt )2 − 4(log SKt )3
Z ∞
Xt,t+τ = Ct (t + τ ; K)dK
St K2
12(log SKt )2 + 4(log SKt )3
Z St
+ Pt (t + τ ; K)dK,
0 K2

rf τ rf τ rf τ
e e e
and µt,t+τ = erf τ − 1 − 2 Vt,t+τ − 6 Wt,t+τ − 24 Xt,t+τ .

To construct these measures, we use out-of-the-money options to mitigate liquidity concerns.

Following Shimko (1993), each day we interpolate the Black-Scholes implied volatility curve at

the observable strikes using a cubic spline, and then calculate option prices to compute the above

moments. We construct these measures for both S&P500 options and VIX options. Our implied

volatility slope and risk-neutral skewness measures are calculated using options with the same

maturity as our test assets.

43

Electronic copy available at: https://ssrn.com/abstract=2497759

References

Bakshi, Gurdip, and Nikunj Kapadia, 2003, Delta-hedged gains and the negative market
volatility risk premium, Review of Financial Studies 16, 527–566.

Bakshi, Gurdip, Nikunj Kapadia, and Dilip Madan, 2003, Stock return characteristics, skew
laws, and the differential pricing of individual equity options, Review of Financial Stud-
ies 16, 101–143.

Bansal, Ravi, Dana Kiku, Ivan Shaliastovich, and Amir Yaron, 2014, Volatility, the macroe-
conomy, and asset prices, Journal of Finance 69, 2471–2511.

Bansal, Ravi, and Amir Yaron, 2004, Risks for the long run: A potential resolution of asset
pricing puzzles, Journal of Finance 54, 1481–1509.

Barndorff-Nielsen, Ole, and Neil Shephard, 2004, Power and bipower variation with stochas-
tic volatility and jumps, Journal of Financial Econometrics 2, 1–37.

Barndorff-Nielsen, Ole, and Neil Shephard, 2006, Econometrics of testing for jumps in finan-
cial economics using bipower variation, Journal of Financial Econometrics 4, 1–30.

Bates, David, 2000, Post-’87 crash fears in the s&p 500 futures option market, Journal of
Econometrics 94, 181–238.

Bertsimas, Dimitris, Leonid Kogan, and Andrew Lo, 2000, When is time continuous?, Jour-
nal of Financial Economics 55, 173–204.

Black, Fischer, and Myron Scholes, 1973, The pricing of options and corporate liabilities,
Journal of Political Economy 81, 637–654.

Bloom, Nicholas, 2009, The impact of uncertainty shocks, Econometrica 77, 623–685.

Bollerslev, Tim, George Tauchen, and Hao Zhou, 2009, Expected stock returns and variance
risk premia, Review of Financial Studies 22, 4463–4492.

44

Electronic copy available at: https://ssrn.com/abstract=2497759

Branger, Nicole, Alexander Kraftschik, and Clemens Völkert, 2016, The fine structure of
variance: Pricing vix derivatives in consistent and log-vix models, Working paper.

Britten-Jones, Mark, and Anthony Neuberger, 2000, Option prices, implied price processes,
and stochastic volatility, Journal of Finance 51, 621–651.

Campbell, John, Stefano Giglio, Christopher Polk, and Robert Turley, 2012, An intertem-
poral capm with stochastic volatility, Working paper.

Canina, Linda, and Stephen Figlewski, 1993, The informational content of implied volatility,
Review of Financial Studies 6, 659–681.

Cao, Jie, and Bing Han, 2013, Cross-section of option returns and idiosyncratic stock volatil-
ity, Journal of Financial Economics 20, 1–10.

Carr, Peter, and Dilip Madan, 1998, Towards a theory of volatility trading, in Option Pricing,
Interest Rates, and Risk Management, 417–427.

Christensen, Bent, and Nagpurnanand Prabhala, 1998, The relation between implied and
realized volatility, Journal of Financial Economics 50, 125–150.

Christoffersen, Peter, Kris Jacobs, and Karim Mimouni, 2010, Volatility dynamics for the
s&p500: Evidence from realized volatility, daily returns, and option prices, Review of
Financial Studies 23, 3141–3189.

Coval, Joshua, and Tyler Shumway, 2001, Expected option returns, Journal of Finance 56,
983–1009.

Cox, John, Jonathan Ingersoll, and Stephen Ross, 1985, A theory of the term structure of
interest rates, Econometrica 53, 385–407.

Drechsler, Itamar, and Amir Yaron, 2011, What’s vol got to do with it, Review of Financial
Studies 1–45.

45

Electronic copy available at: https://ssrn.com/abstract=2497759

Duarte, Jefferson, and Christopher Jones, 2007, The price of market volatility risk, Working
paper.

Duffie, Darrell, Jun Pan, and Kenneth Singleton, 2000, Transform analysis and asset pricing
for affine jump-diffusions, Econometrica 68, 1343–1376.

Eraker, Bjorn, M. Johannes, and N.G. Polson, 2003, The impact of jumps in returns and
volatility, Journal of Finance 53, 1269–1300.

Eraker, Bjørn, and Ivan Shaliastovich, 2008, An equilibrium guide to designing affine pricing
models, Mathematical Finance 18, 519–543.

Fama, Eugene, and James MacBeth, 1973, Risk, return, and equilibrium: Empirical tests,
Journal of Political Economy 81, 607–636.

Fernandez-Villaverde, Jesus, and Juan F. Rubio-Ramı́rez, 2013, Macroeconomics and volatil-

ity: Data, models, and estimation, in Advances in Economics and Econometrics: Theory
and Applications, (Cambridge University Press).

Frazzini, Andrea, and Lasse Pedersen, 2012, Embedded leverage, Working paper.

Heston, Steven, 1993, A closed-form solution for options with stochastic volatility with
applications to bonds and currency options, Review of Financial Studies 6, 327–343.

Heston, Steven, Robert Korajczyk, and Ronnie Sadka, 2010, Intraday patterns in the cross-
section of stock returns, Journal of Finance 65, 1369–1407.

Huang, Xin, and George Tauchen, 2005, The relative contribution of jumps to total price
variation, Journal of Financial Econometrics 3, 456–499.

Hull, John, 2011, Options, Futures, and Other Derivatives (Prentice Hall, 8th edition).

Jiang, George, and Yisong Tian, 2005, The model-free implied volatility and its information
content, Review of Financial Studies 18, 1305–1342.

46

Electronic copy available at: https://ssrn.com/abstract=2497759

Liu, Lily, Andrew Patton, and Kevin Sheppard, 2015, Does anything beat 5-minute rv? a
comparison of realized measures across multiple asset classes, Journal of Econometrics
187, 293–311.

Lu, Zhongjin, and Yingzi Zhu, 2010, Volatility components: The term structure dynamics
of vix futures, Journal of Futures Markets 30, 230–256.

Milstein, G.N., 1995, Numerical Integration of Stochastic Differential Equations (Kluwer

Academic, Boston).

Muravyev, Dmitriy, 2016, Order flow and expected option returns, Journal of Finance 71,
673–708.

Pan, Jun, 2002, The jump-risk premia implicit in options: Evidence from an integrated
time-series study, Journal of Financial Economics 63, 3–50.

Shimko, David, 1993, Bounds of probability, RISK 6, 33–37.

Song, Zhaogang, 2013, Expected VIX option returns, working paper.

Song, Zhaogang, and Dacheng Xiu, 2016, A tale of two option markets: Pricing kernels and
volatility risk, Journal of Econometrics 190, 176–196.

Stein, Jeremy, 1989, Overreactions in the options market, Journal of Finance 44, 1011–1023.

Tauchen, George, and Victor Todorov, 2011, Volatility jumps, Journal of Business and
Economic Statistics 29, 235–371.

Wright, Jonathan, and Hao Zhou, 2009, Bond risk premia and realized jump risk, Journal
of Banking and Finance 33, 2036–2049.

Zhang, Jin, and Yingzi Zhu, 2006, VIX futures, Journal of Futures Markets 26, 521–531.

47

Electronic copy available at: https://ssrn.com/abstract=2497759

Tables and Figures

Table 1: Model parameters

Variance V̄ κV λV
.19792 2.4982 −4.8

Vol-of-vol η̄ κη φ λη
.0329 13.097 .0694 −5

Risk-free rate rf
.0102

Times to maturity ∆t τ τF τC
1/360 2/360 30/360 10/360

The table shows the calibrated parameters for the model. Variance and vol-of-vol refers to the parameters of the variance and
volatility-of-volatility dynamics, respectively. Times-to-maturity indicate the time intervals at annual frequency.

Table 2: Summary Statistics

variable mean std. AR(1) corr. V IX 2 corr. V V IX 2

V IX 2 0.048 0.053 0.812 1.000 0.269

2
V V IX 0.771 0.214 0.333 0.269 1.000
SP X
RV 0.024 0.050 0.631 0.880 0.258
RV V IX 0.670 0.727 0.182 0.182 0.609
V X∗
RV 0.306 0.181 0.284 0.726 0.640

The table shows summary statistics for the implied and realized variances of S&P500 and the VIX index. The implied variances
 2  2
are computed from the option data: V IX 2 corresponds to V100 IX
and V V IX 2 stands for V 100
V IX
. Realized variances are
annualized and computed using 5-minute data on S&P500 returns (RV SP X ); log VIX index innovations (RV V IX ); and VIX
futures returns (RV V X∗ ). The data are monthly from 2006m2 to 2016m12; VIX futures high-frequency data are from 2012m7
to 2016m12.

48

Electronic copy available at: https://ssrn.com/abstract=2497759

 Table 3: Predictability of Realized Variance Measures

V IX 2 V V IX 2 2
Radj
slope t-stat. slope t-stat.

S&P500 Index
SP X
RVt,t+1 0.604 [4.82] 39.42
0.039 [1.48] 2.01
0.605 [4.53] -0.001 [-0.10] 38.95

VIX Index
V IX
RVt,t+1 -0.607 [-0.57] -0.35
0.612 [3.56] 2.49
-1.384 [-1.34] 0.704 [4.05] 2.67

VIX Futures
V X∗
RVt,t+1 3.215 [3.07] 2.55
0.243 [2.24] 7.19
-0.935 [-0.46] 0.282 [2.34] 5.49

The table shows the evidence from the projections of future realized variances of S&P500 and the VIX index on the current
implied variances. Numbers in brackets indicate Newey-West t-statistics with 6 lags. The data are monthly from 2006m2 to
2016m12; VIX futures high-frequency data are from 2012m7 to 2016m12.

49

Electronic copy available at: https://ssrn.com/abstract=2497759

 Table 4: Delta-Hedged Option Gains

Π Π
S (%) C (%)

moneyness mean t-stat. median % < 0 std. AR(1) mean t-stat. median

S&P500

Call 0.950 to 0.975 0.07 [ 3.23] 0.00 50% 0.55 0.36 1.61 [ 3.71] 0.01
0.975 to 1.000 0.04 [ 2.63] -0.02 52% 0.50 0.19 2.38 [ 3.72] -0.79
1.000 to 1.025 -0.04 [-2.64] -0.05 57% 0.44 0.09 -2.45 [-1.74] -5.90
1.025 to 1.050 -0.08 [-6.51] -0.05 62% 0.38 0.14 -32.98 [-9.00] -22.50

Put 0.950 to 0.975 -0.14 [-7.86] -0.18 81% 0.54 0.35 -19.27 [-9.01] -28.74
0.975 to 1.000 -0.16 [-9.89] -0.19 75% 0.50 0.20 -10.24 [-7.93] -16.89
1.000 to 1.025 -0.22 [-13.17] -0.22 75% 0.48 0.04 -8.77 [-13.99] -10.85
1.025 to 1.050 -0.25 [-12.11] -0.23 78% 0.47 0.05 -5.84 [-12.72] -5.97

VIX

Call 0.800 to 0.900 -0.25 [-1.66] -0.59 62% 2.26 -0.03 -1.38 [-1.54] -3.12
0.900 to 1.000 -0.77 [-3.83] -1.26 70% 2.95 -0.09 -6.76 [-3.80] -12.40
1.000 to 1.100 -0.84 [-3.54] -1.37 71% 3.24 -0.17 -10.03 [-3.34] -16.29
1.100 to 1.200 -1.20 [-5.04] -1.52 77% 3.16 -0.13 -19.01 [-4.81] -25.49

Put 0.800 to 0.900 -0.50 [-1.67] -0.90 64% 2.69 -0.06 -12.76 [-1.55] -25.18
0.900 to 1.000 -0.80 [-3.73] -1.29 71% 3.05 -0.14 -10.91 [-3.08] -21.59
1.000 to 1.100 -0.84 [-3.42] -1.45 70% 3.31 -0.18 -6.13 [-3.17] -11.02
1.100 to 1.200 -1.17 [-4.86] -1.31 77% 3.16 -0.14 -5.53 [-4.77] -6.45

The table shows average delta-hedged option gains on the S&P500 and VIX options across their moneyness. Options have one
month to maturity, are grouped into an equal-weighted portfolio inside the moneyness bin, and are held till expiration. The
delta-hedge is computed using the Black-Scholes formula, with daily rebalancing and the margin difference earning the risk-free
rate. The delta-hedged option gains Π are scaled either by the index or by the option price. The t-statistics are testing the null
that the delta-hedged option gain is equal to zero. The % < 0 column shows the fraction of observations with negative gains.
The data are monthly from 2006m2 to 2016m12.

50

Electronic copy available at: https://ssrn.com/abstract=2497759

 Table 5: Delta-Hedged Option Gains by Volatility Risk Sensitivities
2
Vega = ∂C
∂σ Volga = ∂∂σC2
slope t-stat. slope t-stat.

SPX Options
Πt,t+1
St 0.009 [0.61]
-0.005 [-4.28]
-0.104 [-4.40] -0.013 [-6.21]

VIX Options
Πt,t+1
St -1.23 [-3.35]

The table shows the evidence from the cross-sectional regressions of the delta-hedged S&P500 and VIX option gains on the vega
and volga of the options. Regressions include time fixed effects, and option characteristics are computed using the Black-Scholes
formula. Number is brackets indicate Newey-West t-statistics. The data are monthly from 2006m2 to 2016m12.

51

Electronic copy available at: https://ssrn.com/abstract=2497759

 Table 6: Predictability of Delta-Hedged Option Gains

V IX 2 V V IX 2 GAIN St−1
slope t-stat. slope t-stat. slope t-stat.

SPX Options
Πt,t+τ
St -0.69 [-3.42] 0.24 [4.08]
-0.26 [-2.87] 0.23 [4.16]
-0.50 [-3.41] -0.23 [-2.75] 0.26 [4.20]

SPX Options: ATM only

Πt,t+τ
St -0.53 [-2.81] 0.16 [3.43]
-0.25 [-2.41] 0.18 [3.22]
-0.35 [-2.28] -0.24 [-2.29] 0.19 [3.38]

VIX Options
Πt,t+τ
St -1.01 [-3.38] -0.10 [-7.07]

VIX Options: ATM only

Πt,t+τ
St -0.77 [-4.43] -0.11 [-6.89]

The table shows the evidence of predictability of future delta-hedged S&P500 and VIX option gains by the implied volatility
and volatility-of-volatility measures. The regressions across all monenyness bins include moneyness fixed effects; the regressions
for the ATM options are for the moneyness ranges between 0.975 to 1.025 for the index and 0.9 to 1.1 for the VIX options. Lag
gains are included to correct for serial correlation of the residuals. Number is brackets indicate Newey-West t-statistics. The
data are monthly from 2006m2 to 2016m12.

52

Electronic copy available at: https://ssrn.com/abstract=2497759

Table 7: Predictability of Realized Variance Measures: Alternate Specifications

V IX V V IX ln V IX 2 ln V V IX 2
slope t-stat. slope t-stat. slope t-stat. slope t-stat.

S&P500
q
SP X
RVt,t+1 0.723 [6.37]
0.128 [1.63]
0.732 [5.98] -0.023 [-0.52]

SP X
ln RVt,t+1 0.964 [8.70]
0.528 [1.38]
0.986 [8.26] -0.215 [-1.07]

VIX
q
V IX
RVt,t+1 0.200 [-0.57]
0.687 [5.10]
-0.495 [-1.42] 0.789 [5.15]

V IX
ln RVt,t+1 -0.069 [-0.68]
0.752 [5.18]
-0.161 [-1.57] 0.873 [4.80]

The table shows the evidence from the alternative projections of future realized variances of S&P500 and the VIX index on the
current implied volatility and volatility of volatility measures. The realized and implied variances are modified to be expressed
in standard deviation or log units. Numbers in brackets indicate Newey-West t-statistics with 6 lags. The data are monthly
from 2006m2 to 2016m12.

53

Electronic copy available at: https://ssrn.com/abstract=2497759

Table 8: Predictability of Delta-Hedged Option Gains: Robustness to Jump Measures

V IX 2 V V IX 2 SLOP E SKEW RJ GAIN St−1

slope t-stat. slope t-stat. slope t-stat. slope t-stat. slope t-stat. slope t-stat.

SPX Options
Πt,t+τ
St -0.64 [-2.49] -0.21 [-3.25] -1.44 [-1.02] 0.27 [4.00]
-0.49 [-2.58] -0.23 [-2.92] -0.01 [-0.15] 0.26 [4.06]
-0.49 [-3.44] -0.23 [-2.75] 0.01 [1.81] 0.26 [4.20]

VIX Options
Πt,t+τ
St -0.98 [-3.19] 2.55 [2.60] -0.10 [-7.14]
-0.76 [-2.17] 0.34 [3.64] -0.10 [-7.58]
-1.39 [-3.59] 0.03 [1.12] -0.08 [-3.55]

The table shows the evidence of predictability of future delta-hedged S&P500 and VIX option gains by the implied volatility and
volatility-of-volatility measures, controlling for the jump risk measures. The cross-sectional regressions include moneyness fixed
effects, and lag gains are included to correct for serial correlation of the residuals. The jump measures are as follows. SLOP E is
the slope of the implied volatility smile, calculated as Black-Scholes implied volatility of an out-of-the-money ( K
S
= 0.9) minus
the implied volatility of an at-the-money put option ( K
S
= 1) for S&P500 options, and between a K
S
= 1.1 call option and
an at-the-money call option for the VIX options. RJ is the realized jump variation calculated using high-frequency S&P500
futures tick data for the S&P500 and VIX tick data for the VIX. SKEW is the model-free measure of risk-neutral skewness.
Number is brackets indicate Newey-West t-statistics. The data are monthly from 2006m2 to 2016m12.

54

Electronic copy available at: https://ssrn.com/abstract=2497759

 Table 9: Delta-Hedged Option Gains: Pre-2012 Sub-Sample

Π Π
S (%) C (%)

moneyness mean t-stat. median % < 0 std. AR(1) mean t-stat. median

S&P500

Call 0.950 to 0.975 0.12 [ 2.67] -0.04 54% 0.76 0.40 2.48 [ 2.94] -0.85
0.975 to 1.000 0.04 [ 1.12] -0.02 53% 0.64 0.26 1.61 [ 1.52] -0.94
1.000 to 1.025 -0.08 [-2.92] -0.11 61% 0.58 0.11 -5.83 [-3.00] -11.16
1.025 to 1.050 -0.13 [-4.92] -0.07 63% 0.51 0.12 -28.22 [-5.74] -17.53

Put 0.950 to 0.975 -0.12 [-3.31] -0.25 81% 0.77 0.38 -12.24 [-3.23] -28.10
0.975 to 1.000 -0.19 [-5.98] -0.31 75% 0.66 0.24 -9.36 [-4.68] -18.05
1.000 to 1.025 -0.27 [-8.64] -0.34 72% 0.62 0.05 -9.64 [-9.24] -12.92
1.025 to 1.050 -0.30 [-8.58] -0.30 76% 0.57 0.06 -6.51 [-8.80] -7.33

VIX

Call 0.800 to 0.900 -0.62 [-2.96] -0.65 68% 2.16 -0.10 -3.74 [-3.10] -3.66
0.900 to 1.000 -1.49 [-5.63] -2.01 75% 2.62 -0.21 -13.00 [-5.48] -15.61
1.000 to 1.100 -1.02 [-3.16] -1.46 73% 2.90 -0.10 -11.97 [-2.72] -20.79
1.100 to 1.200 -1.38 [-4.25] -1.60 80% 2.89 -0.08 -22.24 [-3.72] -29.28

Put 0.800 to 0.900 -0.65 [-1.75] -1.37 64% 2.64 -0.10 -14.14 [-1.28] -34.73
0.900 to 1.000 -1.40 [-4.94] -1.96 74% 2.77 -0.21 -19.34 [-3.98] -32.81
1.000 to 1.100 -0.97 [-2.82] -1.56 74% 3.05 -0.13 -7.01 [-2.48] -12.75
1.100 to 1.200 -1.29 [-3.88] -1.38 79% 2.87 -0.07 -6.12 [-3.68] -7.49

The table shows average delta-hedged option gains on the S&P500 and VIX options across their moneyness, in the pre-2012
sub-sample. Options have one month to maturity, are grouped into an equal-weighted portfolio inside the moneyness bin, and
are held till expiration. The delta-hedge is computed using the Black-Scholes formula, with daily rebalancing and the margin
difference earning the risk-free rate. The delta-hedged option gains Π are scaled either by the index or by the option price.
The t-statistics are testing the null that the delta-hedged option gain is equal to zero. The % < 0 column shows the fraction
of observations with negative gains. The data are monthly from 2006m2 to 2011m12.

55

Electronic copy available at: https://ssrn.com/abstract=2497759

 Table 10: Delta-Hedged Option Gains: Post-2012 Sub-Sample

Π Π
S (%) C (%)

moneyness mean t-stat. median % < 0 std. AR(1) mean t-stat. median

Panel A: S&P500

Call Options 0.950 to 0.975 0.03 [ 2.05] 0.01 46% 0.27 -0.07 0.88 [ 2.40] 0.31
0.975 to 1.000 0.05 [ 3.21] -0.01 51% 0.35 -0.14 3.00 [ 3.83] -0.40
1.000 to 1.025 0.00 [ 0.21] -0.03 54% 0.29 -0.08 0.22 [ 0.11] -3.70
1.025 to 1.050 -0.05 [-5.08] -0.04 62% 0.20 0.17 -36.85 [-6.95] -32.35

Put Options 0.950 to 0.975 -0.15 [-13.37] -0.14 81% 0.26 0.02 -24.72 [-10.37] -29.33
0.975 to 1.000 -0.14 [ -9.75] -0.15 75% 0.32 -0.03 -10.96 [ -6.49] -15.59
1.000 to 1.025 -0.17 [-11.76] -0.18 77% 0.31 -0.11 -8.06 [-10.69] -9.39
1.025 to 1.050 -0.20 [ -9.91] -0.21 82% 0.31 -0.01 -5.11 [ -9.87] -5.39

Panel B: VIX

Call Options 0.800 to 0.900 0.07 [ 0.36] -0.25 57% 2.30 -0.05 0.64 [ 0.50] -1.57
0.900 to 1.000 -0.17 [-0.61] -0.87 66% 3.08 -0.08 -1.63 [-0.65] -6.79
1.000 to 1.100 -0.71 [-2.07] -1.21 69% 3.49 -0.23 -8.52 [-2.07] -13.85
1.100 to 1.200 -1.05 [-3.06] -1.32 75% 3.36 -0.19 -16.37 [-3.10] -19.43

Put Options 0.800 to 0.900 -0.25 [-0.50] -0.86 65% 2.79 0.22 -10.54 [-0.85] -20.76
0.900 to 1.000 -0.26 [-0.85] -0.90 68% 3.19 -0.13 -3.43 [-0.68] -12.19
1.000 to 1.100 -0.74 [-2.15] -1.18 66% 3.50 -0.22 -5.46 [-2.06] -8.49
1.100 to 1.200 -1.08 [-3.15] -1.26 76% 3.38 -0.19 -5.08 [-3.15] -5.49

The table shows average delta-hedged option gains on the S&P500 and VIX options across their moneyness, in the post-2012
sub-sample. Options have one month to maturity, are grouped into an equal-weighted portfolio inside the moneyness bin, and
are held till expiration. The delta-hedge is computed using the Black-Scholes formula, with daily rebalancing and the margin
difference earning the risk-free rate. The delta-hedged option gains Π are scaled either by the index or by the option price.
The t-statistics are testing the null that the delta-hedged option gain is equal to zero. The % < 0 column shows the fraction
of observations with negative gains. The data are monthly from 2012m1 to 2016m12.

56

Electronic copy available at: https://ssrn.com/abstract=2497759

 Table 11: Delta-Hedged Option Gains: 0.95 × ∆BS Delta
Π Π
S (%) C (%)

moneyness mean t-stat. median % < 0 std. AR(1) mean t-stat. median

S&P500

Call 0.950 to 0.975 0.07 [ 3.63] 0.04 45% 0.51 0.36 1.70 [ 4.21] 0.81
0.975 to 1.000 0.05 [ 3.29] -0.01 52% 0.46 0.23 2.62 [ 4.38] -0.30
1.000 to 1.025 -0.03 [-2.50] -0.05 57% 0.42 0.07 -2.08 [-1.54] -4.97
1.025 to 1.050 -0.08 [-6.86] -0.05 62% 0.36 0.11 -31.52 [-8.87] -22.31

Put 0.950 to 0.975 -0.14 [-7.24] -0.20 83% 0.61 0.35 -20.09 [ -8.85] -31.90
0.975 to 1.000 -0.17 [-9.01] -0.22 77% 0.59 0.20 -11.02 [ -7.68] -20.48
1.000 to 1.025 -0.23 [-11.36] -0.26 75% 0.60 0.05 -9.34 [-12.36] -13.36
1.025 to 1.050 -0.28 [-10.16] -0.31 74% 0.62 0.06 -6.47 [-10.75] -7.71

VIX

Call 0.800 to 0.900 -0.38 [-2.17] -0.83 61% 2.65 -0.05 -2.13 [-2.03] -4.46
0.900 to 1.000 -0.90 [-4.06] -1.70 69% 3.26 -0.12 -7.80 [-3.94] -14.21
1.000 to 1.100 -0.94 [-3.65] -1.65 73% 3.50 -0.15 -11.05 [-3.36] -20.47
1.100 to 1.200 -1.32 [-5.38] -1.73 79% 3.25 -0.13 -20.90 [-5.11] -29.08

Put 0.800 to 0.900 -0.44 [-1.51] -0.73 63% 2.62 0.07 -11.33 [-1.44] -23.15
0.900 to 1.000 -0.73 [-3.52] -1.03 70% 2.95 -0.14 -10.13 [-2.99] -16.30
1.000 to 1.100 -0.74 [-3.14] -0.79 66% 3.20 -0.18 -5.48 [-2.97] -6.53
1.100 to 1.200 -1.05 [-4.31] -0.86 69% 3.18 -0.12 -4.97 [-4.30] -4.87

The table shows average delta-hedged option gains on the S&P500 and VIX options across their moneyness, when the Black-
Scholes delta is reduced to 0.95 × ∆BS . Options have one month to maturity, are grouped into an equal-weighted portfolio
inside the moneyness bin, and are held till expiration. The delta-hedge is computed using the Black-Scholes formula, with daily
rebalancing and the margin difference earning the risk-free rate. The delta-hedged option gains Π are scaled either by the index
or by the option price. The t-statistics are testing the null that the delta-hedged option gain is equal to zero. The % < 0
column shows the fraction of observations with negative gains. The data are monthly from 2006m2 to 2016m12.

57

Electronic copy available at: https://ssrn.com/abstract=2497759

 Table 12: Delta-Hedged Option Gains: 1.05 × ∆BS Delta
Π Π
S (%) C (%)

moneyness mean t-stat. median % < 0 std. AR(1) mean t-stat. median

S&P500

Call 0.950 to 0.975 0.07 [ 2.68] -0.04 56% 0.64 0.29 1.51 [ 3.03] -0.86
0.975 to 1.000 0.04 [ 1.96] -0.07 56% 0.57 0.13 2.15 [ 2.98] -2.61
1.000 to 1.025 -0.04 [-2.66] -0.08 60% 0.48 0.10 -2.82 [-1.87] -9.11
1.025 to 1.050 -0.08 [-6.06] -0.05 63% 0.40 0.14 -34.43 [-8.96] -25.11

Put 0.950 to 0.975 -0.13 [ -8.46] -0.16 79% 0.48 0.34 -18.45 [-9.05] -25.34
0.975 to 1.000 -0.15 [-10.66] -0.17 72% 0.43 0.19 -9.47 [-7.97] -13.93
1.000 to 1.025 -0.20 [-14.58] -0.19 76% 0.40 0.01 -8.20 [-14.76] -8.48
1.025 to 1.050 -0.22 [-13.33] -0.20 80% 0.38 0.01 -5.21 [-13.55] -5.15

VIX

Call 0.800 to 0.900 -0.11 [-0.77] -0.45 57% 2.26 0.06 -0.63 [-0.71] -2.54
0.900 to 1.000 -0.63 [-3.29] -0.99 67% 2.84 -0.03 -5.73 [-3.37] -9.49
1.000 to 1.100 -0.75 [-3.26] -1.04 68% 3.12 -0.16 -9.01 [-3.17] -12.89
1.100 to 1.200 -1.08 [-4.56] -1.39 77% 3.14 -0.12 -17.11 [-4.38] -21.77

Put 0.800 to 0.900 -0.56 [-1.79] -1.19 67% 2.80 0.06 -14.20 [-1.63] -32.17
0.900 to 1.000 -0.87 [-3.85] -1.60 70% 3.20 -0.13 -11.69 [-3.12] -24.20
1.000 to 1.100 -0.93 [-3.59] -1.94 73% 3.51 -0.17 -6.79 [-3.28] -13.46
1.100 to 1.200 -1.30 [-5.18] -1.72 79% 3.28 -0.14 -6.10 [-5.02] -8.40

The table shows average delta-hedged option gains on the S&P500 and VIX options across their moneyness, when the Black-
Scholes delta is increased to 1.05 × ∆BS . Options have one month to maturity, are grouped into an equal-weighted portfolio
inside the moneyness bin, and are held till expiration. The delta-hedge is computed using the Black-Scholes formula, with daily
rebalancing and the margin difference earning the risk-free rate. The delta-hedged option gains Π are scaled either by the index
or by the option price. The t-statistics are testing the null that the delta-hedged option gain is equal to zero. The % < 0
column shows the fraction of observations with negative gains. The data are monthly from 2006m2 to 2016m12.

58

Electronic copy available at: https://ssrn.com/abstract=2497759

 Figure 1: Model-implied Expected Delta-hedged VIX Option Gains
Panel A: Et [& ]/Ft with 6 2 = -5
# 10-3
0
Vt = V
-0.2
V t = V + sd(V t )
-0.4 V t = V - sd(V t )

-0.6

-0.8

-1

-1.2

-1.4

-1.6

-1.8

-2
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1
2t

Panel B: Et [& ]/Ft with 6 2 = 5

# 10-3
2

1.8 Vt = V
V t = V + sd(V t )
1.6
V t = V - sd(V t )
1.4

1.2

0.8

0.6

0.4

0.2

0
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1
2t

The figures show model-implied expected delta-hedged option gains as a function of the volatility-of-volatility η for different
levels of the volatility state Vt . The upper plot shows gains for a negative market price of volatility-of-volatility risks, λη = −5.
The lower plot shows gains for a positive market price of volatility-of-volatility risks λη = 5. All options are at the money.

59

Electronic copy available at: https://ssrn.com/abstract=2497759

 Figure 2: Model-implied Sensitivities of VIX Option Prices
Panel A: @ C*t / @ 2 t with 6 2 = -5
0.9
Vt = V
0.8 V t = V + sd(V t )
V t = V - sd(V t )
0.7

0.6

0.5

0.4

0.3

0.2

0.1
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1
2t

Panel B: @ C*t / @ 2 t with 6 2 = 5

0.8
Vt = V
0.7 V t = V + sd(V t )
V t = V - sd(V t )

0.6

0.5

0.4

0.3

0.2

0.1
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1
2t

The figure shows model-implied sensitivities (derivatives) of at-the-money VIX option prices to the volatility-of-volatility η
as a function of the state variable η for different levels of Vt . The upper plot shows gains for a negative market price of
volatility-of-volatility risks, λη = −5. The lower plot shows gains for a positive market price of volatility-of-volatility risks
λη = 5.

60

Electronic copy available at: https://ssrn.com/abstract=2497759

 Figure 3: Black-Scholes and Model-implied Deltas

The figure plots the Black-Scholes delta of a VIX option (left panel) and the model-implied delta (right panel) as functions of
the volatility and volatility of volatility state variables V and η.

Figure 4: Volatility and Volatility-of-Volatility Measures

The figure shows the time-series of the volatility and volatility-of-volatility measures. The solid blue line is the VIX and the
dashed red line is the VVIX. Monthly data from 2006m2 to 2016m12

61

Electronic copy available at: https://ssrn.com/abstract=2497759

 Figure 5: Realized and Implied Variance Measures

The figure shows the time-series of the realized and implied variances for the S&P500 (top panel) and the VIX (bottom panel).
The blue solid lines are the realized variances, and red dashed lines are the implied variances. All measures are in annualized
variance units. Monthly data from 2006m2 to 2016m12

62

Electronic copy available at: https://ssrn.com/abstract=2497759

 Figure 6: Vega and Gamma by Moneyness

The figure shows the Black-Scholes vega and gamma of the option by option moneyness for average levels of volatility.

63

Electronic copy available at: https://ssrn.com/abstract=2497759

 Figure 7: Volga by Moneyness

The figure shows the Black-Scholes volga of the option by option moneyness for average levels of volatility.

64

Electronic copy available at: https://ssrn.com/abstract=2497759