ETF Competition and Market Quality

• Author: Ryan L. Davis,Travis Box,Kathleen P. Fuller
• DOI: http://doi.org/10.1111/fima.12246
• Published date: 01 September 2019
• Date: 01 September 2019

ETF Competition and Market Quality

Travis Box, Ryan L. Davis, and Kathleen P. Fuller∗

This paper examines competition between exchange tradedfunds (ETFs) that hold nearly identical

portfolios of securities. We provide evidence that incumbent-fund liquidity is negatively affected

when a new ETF is added to an asset class. The degradation in liquidity is even more severe

whenever both funds follow the same benchmark.We also document a decline in primary-market

activity for the incumbent ETFs after the arrival of new competitors. Furthermore, increasingthe

number of funds in an asset class does not put downwardpressure on fund management fees. Thus,

the deterioration in market quality may not be offset by decreasingcosts of fund ownership.

Exchange traded funds (ETFs) provide access to a large diversified portfolio of securities that

can easily be traded throughout the day. From 1997 to 2017, the aggregate total net assets (TNA) of

passively managed,long-only ETFs listed in the United States has grown from $2.2 billion to just

under $3 trillion, whereas the total number of funds has expanded from 18 to 1,204. As new ETFs

have poured into the marketplace, the likelihood that multiple funds compete within the same

asset class has risen considerably. Our paper considers how the degree of competition between

passively managed ETFs affects their secondary-market liquidity, primary-market activity, and

the management fees paid by investors.

When multiple ETFs track a similar benchmark, their intrinsic values derive from nearly

identical portfolios of underlying assets. If the degree of substitutability between competing funds

is high, ETF sponsors may lower management fees to lure market share away from their rivals.

Although investors should respond positivelyto any potential decline in expenses, the addition of

a new ETF to an asset class may also reduce the share of secondary-market order flow to each of

the incumbent funds. Barclay and Hendershott (2004) and Hendershott and Jones (2005) provide

evidence that order flow fragmentation maylead to wider bid-ask spreads and lower overall depth.

Similarly, Pagano (1989) argues that order flow consolidation leads to an improvement in the

liquidity of a market. Thus, if the addition of a new fund splits trading between similar ETFs,

incumbent funds may experience a reduction in liquidity, whereby market participants experience

wider bid-ask spreads and lower overall depth.1Conversely, several empirical studies have found

improvements in liquidity after markets become more fragmented (e.g., Bessembinder, 2003b;

We appreciate the helpful comments of Bing Han (Editor), an anonymous referee, Michael Goldstein, Marc Lipson,

Andrew Lynch, Robert Van Ness and David Stowe, as well as seminar participants at Auburn University, University of

Alabama at Birmingham, the 2016 Magnolia Conference,the 2016 Southern Finance Association Annual Meetings, and

the 2017 Midwest Finance Association Annual Meetings.

∗Travis Boxis an Assistant Professor of Finance in the College of Business at Clemson University, SC and was formerly

an Assistant Professor of Finance in the College of Business at the University of Mississippi in University, MS. Ryan

Davis is an Assistant Professor of Financein the Collat School of Business at the University of Alabama at Birmingham

in Birmingham, AL. Kathleen Fuller is an Associate Professor of Financein the College of Business at the University of

Mississippi in University, MS.

1Malamud (2016) suggests that the introduction of new ETFs may lead to a decrease in liquidity for existing ETFs

through a demand substitution effect. Cherkes, Sagi, and Stanton (2009) develop a model that shows the creation of a

new closed-end fund reduces both the demand in and the liquidity benefit provided by other funds within that particular

sector. In addition, Corwin and Coughenour (2008) find that when New York Stock Exchange specialists allocate their

attention to stocks with increased trading activity,the liquidity of the other stocks in their por tfolio decreases. Similarly,

Financial Management •Fall 2019 •pages 873 – 916

874 Financial Management rFall 2019

Boehmer and Boehmer, 2003; O’Hara and Ye, 2011). In addition, Malamud (2016) suggests that

the introduction of new ETFs may also improve liquidity in the underlying securities, which in

turn could boost the secondary-market liquidity of existing ETFs.

Our results, based on a sample of 1,275 ETFs from 1997 to 2017, indicate that the entry of a new

competitor increases intraday effective spreads for the incumbent funds in the same asset class

by 0.49 basis points relative to ETFs with similar levelsof trading volume, authorized participant

activity, stock returns, and fund flows.2Whenever the entrant also tracks the same benchmark as

the incumbent fund, the decline in secondary-market liquidity for the incumbent funds becomes

even more extreme, with effective spreads widening by 9.05 basis points relative to funds with

similar characteristics. Following Glosten (1987), wedecompose effective spreads into a realized

spread component and a price impact component to determine whether the increase in transaction

costs leads to greater revenue for liquidity providers, greater losses for liquidity demanders due

to adverse selection, or both. Although our results indicate that adverse selection can explain

some of the losses experienced by liquidity demanders, an increase in the proportion of surplus

captured by liquidity providers accounts for most of the widening observed in effective spreads.

Unlike common stocks, ETFs have an open-ended structure via the share creation and redemp-

tion process that facilitates arbitrage between the fund and its underlying portfolio of securities.3

Thus, the primary market for ETFs acts as an additional layer of liquidity that allows large insti-

tutional investors to trade ETFs in the primary market by exchanging the underlyingpor tfolio for

new ETF shares, or in the secondary market by exchanging cash for the ETF, depending on where

trading costs are lower. For a more complete picture of the effects of competition on liquidity,

we use the frequency with which ETF shares are created or redeemed to capture primary-market

liquidity. Specifically, we use the coefficient of variation for each fund’s shares outstanding to

approximate the level of activity in the primary market for a particular ETF. Although we observe

only minor changes in the volatility of shares outstanding following the entry of a competing

fund into an asset class, we find a significant reduction in the coeffi cient of variation for shares

outstanding when the entrant also follows the same benchmark. Thus, the introduction of a nearly

perfect substitute affects both the primary and secondary market for shares of the incumbent funds.

Notwithstanding higher secondary-market transaction costs, ETF investors could still benefit

from the arrival of a new fund into an asset class if incumbent funds respond by lowering

management fees. For a comprehensive view of investor expenses over time, we calculate the

value-weighted average of fund management fees during each year of our sample for all major

investment types. Despite the dramatic increase in aggregate TNA, along with a growingnumber

of categories with competing funds, a typical ETF investor in 2017 faces roughly the same value-

weighted management fees as they did in 1997.4Narrowing our focus to individual funds, we

analyze the frequency of yearly fee changes and find that investor expenses are unchanged from

in our setting, ETF investor attention maybe redirected by the arrival of new ETFs to a particular area causing the trading

environments of incumbent ETFs to deteriorate.

2Our sample includes 1,275 ETFs across all years but only 1,204 in 2017.

3Should an ETF trade at a premium, for example, a specially designated institutional investor known as an authorized

participant (AP) could purchase the underlying securities, exchange the portfolio for new ETF shares from the fund

sponsor, and then sell the newlycreated shares on the secondar y market.In practice, an AP could simultaneously buy the

underlying shares and short the ETF.At the end of the trading day, the AP would exchange the underlying portfolio for

ETF shares and then close the short ETF position.

4Our results appear to contradict the reporting of the popular press. Barron’s (“ETF Fee Wars Ramp Up,” November

21, 2015) and the Wall Street Journal (“A New Low-Cost Leader among Stock ETFs,” November 10, 2015) reported

that BlackRock, the largest provider of ETFs, cut management fees by more than half in one of their ETFs. Almost

immediately,Charles Schwab cut fees for their competing ETF. Box, Davis, and Fuller (2018) examine this high-profile

fee war and conclude that this case is not representative of the overall ETF industry and that the fees associated with

Box, et al. rETF Competition and Market Quality 875

the previous year across 67.0% of the observations in our sample. For ETFs in categories that

experience an increase in competition, we find that incumbent funds maintain their fees while

entrants charge higher fees during the following year. Ultimately, we find no evidence that fund

sponsors consistently lower fees in the face of competition. Therefore, the reported increase in

effective spreads does not appear to be offset by a decrease in the cost of ETF ownership.

A novelty of our study is that we examine liquidity in a context that is nearly devoid of

competition between market centers. Unlike previous studies that consider competition between

exchanges for order flow in the same securities (e.g., McInish and Wood, 1992; Bessembinder,

2003b; Boehmer and Boehmer, 2003; O’Hara and Ye, 2011), our analysis is unique because

we focus on the addition of new securities that happen to be identical in nearly every way.

In addition, we believe that our study makes two important contributions to the existing ETF

literature. First, we highlight a previously unknown consequence of the ETF industry’s growth

whereby incumbent-fund liquidity decreases after new ETFs enter the market. Second, despite

recent headlines in the popular press and the findings of comparable studies in the mutual fund

literature, we find little evidence that ETF sponsors engage in price competition by reducing

management fees.

I. Conceptual Underpinnings

A. ETFs, the Financial System, and the Market Quality of Underlying Securities

A variety of researchers and regulators have raised specific concerns about the pricing and

trading of these investment vehicles. Chief among these concerns is that ETFs may contribute to

instability when markets are most vulnerable. Ramaswamy (2011) suggests that the complexity

of some ETFs may increase systematic risks in the financial system. Likewise, the Securities and

Exchange Commission (SEC) highlighted the role of ETFs in its official reviews of the May 6,

2010, Flash Crash and the August 24, 2015, stock market plunge.5,6 Others have argued, however,

that such associations between financial market stability and the rise of ETFs may be somewhat

speculative. Madhavan and Sobczyk (2015) show that large net asset value discounts observed in

bond ETFs during periods of stress, such as the financial crisis, reflect efficient pricing, and that

ETFs are not a source of additional volatility or systematic risk.

The literature also cautions that the intraday liquidity of ETFs might attract short-term traders

who introduce noise into the underlying security prices through the arbitrage of fund shares.

Broman and Shum (2016) document that there is, in fact, precisely this type of liquidity clientele

among institutional investors. Malamud (2016) introduces a theoretical model that demonstrates

how the creation and redemption mechanism of ETF shares serves as a shock propagation channel

that allows temporary demand shocks to leave an enduring impact on constituent security prices.

Ben-David, Franzoni, and Moussawi (2018) observe that ETFs increase the nonfundamental

volatility of the securities in their, and Da and Shive(2018) f ind that ETF ownership is associated

with higher realizations of comovement between their underlying holdings. For fixed-income

ETFs, Dannhauser (2017) shows that the liquidity of constituent securities falls as ETF ownership

most funds “are being held steady or even climbing” (“Feeson Mutual Funds and ETFs Tumble toward Zero,” Wall Street

Journal, January 26, 2016).

5“Findings Regarding Market Events of May 6, 2010,” prepared by the staffs of the Commodity Futures Trading

Commission (CFTC) and SEC, available at: https://www.sec.gov/news/studies/2010/marketevents-report.pdf.

6“Equity Market Volatility on August 24, 2015,” prepared by the staff of the SEC, available at: https://www.sec.gov/

marketstructure/research/equity_market_volatility.pdf.