Does increased hedging lead to decreased price efficiency? The case of VIX ETPs and VIX futures

• Author: Robert I. Webb,Bart Frijns,Adrian Fernandez‐Perez,Alireza Tourani‐Rad
• Date: 01 August 2019
• Published date: 01 August 2019
• DOI: http://doi.org/10.1111/fire.12195

DOI: 10.1111/fire.12195

ORIGINAL ARTICLE

Does increased hedging lead to decreased price

efficiency? The case of VIX ETPs and VIX futures

Adrian Fernandez-Perez1Bart Frijns1AlirezaTourani-Rad1

Robert I. Webb2

1Department of Finance, Auckland University of

Technology,Auckland, New Zealand

2McIntire School of Commerce, Universityof

Virginia, Charlottesville, Virginia

Correspondence

BartFrijns, Department of Finance, Auckland

Universityof Technology,Private Bag 92006,

Auckland1142, New Zealand.

Email:bfrijns@aut.ac.nz

Abstract

We examine the impact of the introduction of VIX exchange-traded

products (ETPs) on the information content and pricing efficiency of

VIX futures. We document that trades in VIX futures have become

less informative and that pricing errors exhibit more persistence

after the introduction of VIX ETPs. In addition, we observe that the

price process of the VIX futures has become noisier over time. These

findings suggest that the introduction of the VIX ETPs had a promi-

nent effect on the properties and dynamics of the VIX futures.

KEYWORDS

market microstructure, price pressure, VIX ETPs, VIX futures

JEL CLASSIFICATIONS

C22, G13

1INTRODUCTION

Exchange-traded products (ETPs) havebecome increasingly popular among retail investors. Issuers of ETPs often use

futures contracts to manage their risk exposure as demand for their ETPs changes. Alexander and Korovilas (2012)

and Asensio (2013) document that excessive liquidity demand from VIX ETPs due to their large-scale hedging activi-

ties could affect VIX futures prices.1This paper examines whether the introduction of ETPs on the VIX changed the

intraday price dynamicsof VIX futures due to hedging pressure by VIX ETPs issuers.

Although VIX futures were introduced by the Chicago Board Options Exchange(CBOE) on March 26, 2004, trading

volume was rather low until the introduction of VIX options on February 24, 2006. However,what really led to the

popularity and a major uptake in VIX futures trading was the introduction of several VIX ETPs. These ETPs, which

track an index based on VIX futures, made the tradingof volatility widely accessible to retail investors. The first ETPs

were the VXX and the VXZ (introduced by Barclays Bank PLCon January 29, 2009), which track the Standard & Poor's

1Białkowski, Dang, and Wei (2016) find the fund flows into different VIX ETP groups have significant impacts on daily changes in the VIX. Bollen, O'Neill,

and Whaley(2017) document that VXX (the most important direct VIX ETP) leads the short-term VIX futures price index in price discovery, and likewise VIX

futures leads VIX index, or as theycall it the “tail is wagging the dog.” Bessembinder, Carrion, Tuttle, and Venkataraman (2016) document a similar effect of

howthe demand in the USO (the ETF on crude oil) has an impact on crude oil futures. Henderson, Pearson, and Wang (2015) find that investor flows in and out

ofcommodity-linked notes (CLNs) are passed to and withdrawn from the futures markets via issuers’ trades to hedge their CLN liabilities.

Financial Review.2019;54:477–500. wileyonlinelibrary.com/journal/fire c

2019 The Eastern Finance Association 477

478 FERNANDEZ-PEREZ ET AL.

(S&P) 500 VIX short- and medium-term futures indexes,respectively. As these ETPs track a VIX futures index, the ETP

provider needs to hedgeits positions bytrading in the VIX futures. Increased demand for ETPs by retail investors led to

such a large increase in the tradingvolume of VIX futures that the demand for VIX futures from ETPs at times exceeded

the open interest in VIX futures (Bollen et al., 2017).2

In this paper, we are interested in investigating the (marketmicrostructure) properties of the VIX futures, focus-

ing on the first- and second-nearby futures contracts. These contracts are by far the most heavilytraded and are also

the most demanded by VIX ETPs, as they are the constituents of the S&P 500 VIX short-term index. Webuild a state-

space model for intraday price dynamics based on Brogaard, Henderschott, and Riordan (2014) and Hendershott and

Menkveld(2014), which we model in transaction time. This model allows us to assess the intraday price dynamics of the

VIX futures and study the informativeness of trades. More importantly, this model allows us to decompose the price

process of the VIX futures into a permanent part (representing the efficient price process) and a transitory part (rep-

resenting temporary deviations from the efficient price or pricing error), assessing the properties of both processes.

Empirically, we estimate our model daily using the Kalman Filter overthe period February 24, 2006 to September 2,

2014.

Our results show that both the price impact (orinformativeness) of order flow for the evolution of the efficient price

and the contribution of order flow to the variance of the efficient price process decrease after the introduction of VIX

ETPs. For the pricing error,we find that order flow, in general, is in the opposite direction of the error (a finding that

is in line with Brogaard et al., 2014). However,this correction of the pricing error has declined, suggesting that there

is less correction occurring in the period after the introduction of the VIX ETPs. In addition, we observe that there is

persistence in the pricing error. This persistence increases overtime, again suggesting that, after the introduction of

the VIX ETPs, pricing errors persist longer. Finally,the efficiency of the VIX futures prices, measured by the ratio of

the variance of the efficient price to the variance of the observed price (efficient price plus transitorycomponent) has

declined, indicating that VIX futures price movementsare less driven by movements in the efficient price and are more

driven by noise. To providefurther evidence of the impact of the introduction of VIX ETPs on the dynamics of VIX

futures, we assess the properties of the VIX futures intraday and find that the persistence in the pricing error declines

toward the end of the trading day in the period before the introduction of the VIX ETPs. However, this persistence

increases toward the end of the trading day after the introduction of the VIX ETPs. This intraday variation may be a

consequence of VIX ETP traders, who rebalance their positions mostly toward the end of the trading day (Alexander

& Korovilas, 2012).

Since the analysis over different subsamples provides only anecdotal evidence of the impact of VIX ETPs on VIX

futures, we conduct an instrumental variables-generalized method of moments (IV-GMM) regression approach to

establish whether it is indeed VIX ETP hedging demand that has caused the reduction in the informativeness of the

VIX futures. The IV-GMM regression indeed supports the idea that the hedging demand of the VIX ETPs adds noise to

the price process of VIX futures, especially toward the end of the trading day.These results hold after controlling for

changes in the underlying, the effect of the algorithmic tradingactivity and a dummy that controls for the technological

changes in futures trading.Therefore, our results suggest that the increase in the noise component of the price process

of VIX futures after the introduction of the VIX ETPs is caused by the impact of their hedging demands.

A related study to ours is that of Bollen et al. (2017) who document that prices of VIX ETPs (VXX in partic-

ular) lead prices of its benchmark, the VIX short-term futures index (SPVXSTR). This price leadership could be

attributed to the relative informativeness/efficiency or market structure of one market over another. In our paper,

we do not consider relative informativeness/efficiency but focus on the VIX futures themselves. We indeed docu-

ment that informativeness of VIX futures has decreased over time, and we attribute this decline to the increase in

hedging demand from the VIX ETPs. We therefore contribute to the literature as these results could not be inferred

from Bollen et al. (2017) as changes in relative informativeness could be due to changes in the numerator or the

denominator.

2The VIX ETP market is large, reaching approximately US$4 billion investedin these ETPs. The market for VIX futures is large as well with over 200,000

contractstraded per month. In fact, VIX derivatives are a major contributor to the profits of the CBOE after options on the S&P 500 index.