Do ETFs Increase Volatility?

• Author: FRANCESCO FRANZONI,ITZHAK BEN‐DAVID,RABIH MOUSSAWI
• Published date: 01 December 2018
• Date: 01 December 2018
• DOI: http://doi.org/10.1111/jofi.12727

THE JOURNAL OF FINANCE •VOL. LXXIII, NO. 6 •DECEMBER 2018

Do ETFs Increase Volatility?

ITZHAK BEN-DAVID, FRANCESCO FRANZONI, and RABIH MOUSSAWI∗

ABSTRACT

Due to their low trading costs, exchange-traded funds (ETFs) are a potential catalyst

for short-horizon liquidity traders. The liquidity shocks can propagate to the under-

lying securities through the arbitrage channel, and ETFs may increase the nonfun-

damental volatility of the securities in their baskets. We exploit exogenous changes

in index membership and find that stocks with higher ETF ownership display sig-

nificantly higher volatility. ETF ownership increases the negative autocorrelation in

stock prices. The increase in volatility appears to introduce undiversifiable risk in

prices because stocks with high ETF ownership earn a significant risk premium of up

to 56 basis points monthly.

EXCHANGE-TRADED FUNDS (ETFS)ARE increasingly popular in financial markets.

Introduced in the early 1990s, today this asset class boasts $2.5 trillion in

assets under management (AUM) in the United States ($3.5 trillion globally),

accounting for about 35% of the volume in U.S. equity markets. Increased

access to liquidity and diversification is undoubtedly the greatest benefit for

investors. Unlike traditional index funds, ETFs provide intraday liquidity, and

thus they can satisfy high-frequency demand for trading. Moreover, investment

∗Itzhak Ben-David is with the Fisher College of Business, The Ohio State University, and

NBER. Francesco Franzoni is with Universit`

a della Svizzera italiana (USI) Lugano and the Swiss

Finance Institute. Rabih Moussawi is with the Villanova School of Business, Villanova University,

and Wharton Research Data Services (WRDS) at the Wharton School, the University of Penn-

sylvania. We are especially grateful to Andrew Ellul, Marco Di Maggio, Robin Greenwood (AFA

discussant), and Martin Oehmke (NBER discussant). We thank Pierre Collin-Dufresne; Chris

Downing; Vincent Fardeau; Thierry Foucault; Rik Frehen; Denys Glushkov; Jungsuk Han; Johan

Hombert; Augustin Landier; David Mann; Rodolfo Martell; Massimo Massa; Albert Menkveld;

Robert Nestor; Marco Pagano; Ludovic Phalippou; Anton Tonev; TugkanTuzun; Dimitri Vayanos;

Scott Williamson; Hongjun Yan;and participants at seminars and conferences at the NBER Sum-

mer Institute (Asset Pricing), Toulouse School of Economics, Insead, HEC Paris, the Cambridge

Judge Business School, Villanova University,USI Lugano, the 4th Paris Hedge Funds Conference,

the 5th Paul Woolley Conference (London School of Economics), the 8th Csef-IGIER Symposium

(Capri), the 5th Erasmus Liquidity Conference (Rotterdam), the 1st Luxembourg Asset Pricing

Summit, the Center for Financial Policy Conference at the University of Maryland, Jacobs Levy’s

Quantitative Financial Research Conference at the Wharton School, the Geneva Conference on

Liquidity and Arbitrage, the 20th Annual Conference of the Multinational Finance Society,the 7th

Rothschild Caesarea Conference, the Swedish House of Finance, the FIRS conference (Toronto),

and SAC Capital Advisors for helpful comments and suggestions. Ben-David acknowledges sup-

port from the Neil Klatskin Chair in Finance and Real Estate and from the Dice Center at the

Fisher College of Business. Franzoni acknowledges support from the Swiss Finance Institute. An

earlier version of this paper circulated under the title “ETFs, Arbitrage, and Shock Propagation.”

The authors do not have any material disclosure to make.

DOI: 10.1111/jofi.12727

2471

2472 The Journal of Finance R

strategies from which retail investors are typically precluded, such as using

leverage and selling short, have become widely available due to ETFs.

One may wonder, however, whether the ease of trade that makes ETFs so

popular has unintended consequences for the securities in the ETFs’ baskets.

The liquidity of ETFs likely attracts high-frequency demand. This demand can

affect the prices of the underlying securities because ETFs and their baskets

are tied by arbitrage.1The ETF-underlying securities may therefore be exposed

to a new layer of demand shocks, which can make the prices of these securities

more volatile. In this paper, we explore this conjecture and its implications for

asset pricing.

To test this hypothesis, one must first understand the difference between the

conjectured mechanism and the effect that other institutional investors can

have on asset prices (e.g., Coval and Stafford (2007), Lou (2012), Vayanos and

Woolley (2013)). Similar to the effect of mutual fund or hedge fund flows on

stock prices, the demand for ETF shares exerts pressure on the prices of the

underlying securities. What makes ETFs distinct is that they allow investors to

access the market continuously and at a low trading cost. ETFs may therefore

potentially attract more high-frequency demand than other institutional port-

folios, including traditional index funds. Moreover, although active managers

have discretion in the extent to which they track their benchmark, which gives

them control over the price impact of their trades, arbitrage trading between

ETFs and the underlying securities is mechanical and therefore may lead to

a bigger price impact. Indeed, ETFs track specified indices, which means that

ETF arbitrageurs have almost no discretion in terms of the timing and compo-

sition of their trades. As a result, although the channel of shock propagation

that we conjecture is not qualitatively different from the price impact that

other institutional investors could trigger as a response to flows, the frequency

with which ETFs operate and the passive nature of their strategies make ETF

arbitrage a distinct mechanism that warrants separate attention.

The natural alternative hypothesis to our conjecture is that ETFs provide a

buffer for demand shocks that would otherwise hit the underlying securities’

prices. Grossman (1989) makes this point about the introduction of futures,

which in this respect are analogous to ETFs. He argues that the existence of

futures entails additional market-making power to absorb the impact of liq-

uidity shocks, which reduces volatility in the spot market (see also Danthine

(1978), Turnovsky (1983)). To disentangle the two hypotheses, we need to de-

termine the direction of the link between the presence of ETFs and stock-level

volatility.

We start our empirical analysis by providing suggestive evidence that ETFs

attract short-term investors. ETFs are significantly more liquid, on average,

than the basket of underlying securities in terms of bid-ask spread, price

1If the ETF price deviates from the net asset value (NAV) of the portfolio holdings because of

a demand shock, arbitrageurs trade the underlying securities in the same direction as the initial

shock to the ETF price. As a result, the underlying securities can inherit the shocks that occur in

the ETF market.

Do ETFs Increase Volatility? 2473

impact, and turnover. Consistent with theories positing that short-horizon

clienteles self-select into assets with lower trading costs (Amihud and Mendel-

son (1986)), using trade-level data we find that institutions holding ETFs have

a significantly shorter horizon than those holding the underlying securities.

We next conduct our main test of the effect of ETF ownership, the fraction

of a stock’s capitalization that is held by ETFs, on stock-level volatility. Using

ordinary least squares (OLS) regressions, we find that this relation is positive

and significant. In particular, a normal shock to ETF ownership shifts the

volatility of the median stock in the S&P 500 to a place that is between the

55th and 64th percentiles.

Although the OLS regressions control for observable stock characteristics

and include stock fixed effects, ETF ownership may be endogenous. To address

this concern, we rely on the natural experiment provided by the annual re-

constitution of the Russell indexes (Appel, Gormley, and Keim (2015), Chang,

Hong, and Liskovich (2015)). Exploiting the mechanical rule that allocates

stocks between the Russell 1000 and 2000 indexes, we use the index-switching

event as an instrument for ETF ownership. This test confirms that the effect

of ETF ownership on volatility is positive and statistically significant. Specifi-

cally, the shift in ETF ownership that follows the index reconstitution pushes

the median stock’s volatility up to the 65th percentile of the distribution. The

instrumental variable (IV) estimates somewhat exceed their OLS counterparts,

which suggests a negative omitted variable bias in the latter. However, the lo-

cal nature of the experiment that we use for identification does not allow us

to generalize these magnitudes to the full sample. We corroborate the validity

of the exclusion restriction by showing that switching indexes has a stronger

effect in months when ETF ownership is greater, and for stocks with a higher

ratio of ETF ownership to mutual fund ownership (index or active) and hedge

fund ownership.

Next, we take on the task of separating the hypothesis that the observed

increase in volatility reflects the improvement in price discovery brought about

by ETFs from the hypothesis that the increase in volatility arises from non-

fundamental demand for liquidity, which ETFs attract. In the former case, the

increase in volatility would be a result of stock prices responding more promptly

to fundamental news, as in Andrei and Hasler (2015). In the latter case, the

increase in volatility would amount to noise.

To separate these alternatives, we rely on the premise that liquidity shocks

subsequently revert, whereas fundamental information leads to permanent

price changes. First, we estimate predictive regressions of stock returns as a

function of ETF flows at the stock-day level. We find that most of the contem-

poraneous stock-price effect of ETF flows reverts over the next 40 days, in line

with the view that the demand shocks in the ETF market translate into nonfun-

damental price changes for the underlying securities. Second, using variance

ratios, we find that ETF ownership leads stock prices to deviate from a random

walk at the daily frequency and that the autocorrelation of returns becomes

more negative. These findings support the conjecture that ETFs add noise to

stock prices.