Liquidity in ETF Markets. CNRS EDHEC Business School

Transcription

1

2 Liquidity in ETF Markets Laurent DEVILLE CNRS EDHEC Business School (joint with iha. Cl Calamiaand F. Riva)

3 ETFs and liquidity ETFsare often presented tdas a low cost alternative ti to traditional index mutual funds for institutional and retail investors who face low TER high liquidity but incur trading costs Studies on ETFs liquidity essentially focus on US ETFs liquidity of ETFs is higher than the liquidity of the underlying stocks portfolio (Hedge and Mc Dermott, 2004) new ETFs tend to track less liquid indexes than mutual funds (Guedj and Huang, 2009) We investigate the relationship existing between the liquidity of an ETF and that of its benchmark index both in Europe and the US over a long period

4 The development of ETFs Exchange Traded dfunds in US markets kt SPDR introduced in 1993 on AMEX spectacular growth in trading volume after launching of QQQ in 1999 most actively traded equity securities on U.S. stock exchanges In Europe exchanges started trading ETFs in 2000 (London, Frankfort then Paris) some large providers are subsidiaries of investment banks development of new replication techniques based on swaps As of the end of December providers issuing morethan3,000etfs totaling 6,612 listings Asset under Management still only US$ 1,350.9 billion but capturing more new assets than Mutual Funds

5 Equity ETFs markets

6 ETFs and market stability Developments of the ETF markets kt raise issues on whether they are innocuous for market stability Flash Crash of May 2010 CFTC and SEC reports 68% of canceled trades involved ETFs severe price dislocations for ETFs 2011IMFGlobalFinancialStabilit Financial Stability report and Financial Stability Board 2011 Potential financial stability issues arising from recent trends in Exchange Traded dfunds (ETFs) complexity lack of transparency of new types of ETFs (synthetic, leverage and inverse) use of securities lending by ETFs providers potential low liquidity during a crash or when they replicate low liquidity assets

7 Liquidity: what really matters? ETF markets kt are designed dso as to ensure stable tbl and efficient provision of liquidity creation/redemption process specific market making rules Industry says that the real liquidity of an ETF is equal to that of its underlying stocks liquidity should be equivalent for all ETFs on the same index What about conventional microstructure issues in ETFs? market making order flow fragmentation competition

8 Illustration of the impact of the introduction of competitors After 4 years of monopoly for Lyxor CAC 40 ETFs, two competitors were introduced in 2005 the spreads of the incumbent then significantly decreased

9 OUTLINE 1. Explaining liquidity in ETFs secondary markets 2. Data and liquidity measures 3. Liquidityidi in US and European equity ETFs 4. The determinants of liquidity 5. Conclusions

10 1. EXPLAINING LIQUIDITY IN ETFS SECONDARY MARKETS

11 Original operational structure Source: Ramaswamy (2011)

12 Market making Market making in ETFs closely l resemble market kt making in index options Market makers can hedge their inventory by taking opposite positions in index futures contracts (when they exist) the underlying stock basket Market makers can also directly unwind their inventory through the creation or redemption of ETFs generally in kind (against the basket portfolio) for full replication ETFs generally in cash or against futures for swap based ETFs

13 Relationships between ETFs and stocks Does the advent of ETFs impact trading and market quality with regard to related assets? migrationof of liquidity trading from the stock market to the ETF market could deter the liquidity of individual securities ETFs could facilitate and increase arbitrage activity and enhance both the liquidity of the underlying stocks and the efficiency of the derivatives markets Although concerns are essentially raised by the development of the markets empirical evidence is limited to the period surrounding the introduction of ETFs

14 Relationships between ETFs and stocks (cont d) Hamm (2010) and Ben David, Franzoni and Moussawi (2011) extend the analysis of the impact of ETFs to a large number of ETFs over long periods investigate the impact of ETFs on the stocks held restrict their attention to the US market focus on specificdimensions of liquidity Main results there is a positive association between the availability of ETFs and the adverse selection on the individual stocks arbitrage activity between ETFs and the underlying stocks increases the risk of contagion across assets, especially for less liquid securities

15 Motivation As ETFs markets kt develop and evolve, we need to understand the liquidity of equity ETFs and its determinants through more and broader evidence from international samples for multiple liquidity measures This study is a first step in that direction uses an international ti sample extracts the determinants of liquidity for a subsample of ETFs documents differences across ETFs, index types, markets and replication types in liquidity measures

16 2. DATA

17 Dataset 1 We use a large sample for our univariateanalysis i i (from BlackRock ETF Landscape 2011) all European and US equity ETFs (style and leveraged ETFs excluded) In the United States 688 ETFs on 667 indexes issued by 25 companies In Europe 774 ETFs on 558 indexes listed on 22 exchanges issued by 32 companies 60.6% of which achieve replication with a swap based structure as of 2011 on average, an ETF is cross listed on 4 different exchanges (ranges from 1 to 15) Highly segmented markets 45 common indexes and 5 common issuers

18 Dataset 2 Subsample for a multivariate li i analysis equity ETFs replicating: European National indexes : AEX, IBEX, OMX Stockholm, FTSE MIB, CAC 40, DAX, SMI, FTSE 100, FTSE 250 US National indexes : DJIA, Nasdaq 100, S&P 500 European regional indexes : Stoxx 50 Europe, Euro Stoxx 50, Stoxx Europe 600 and Nikkei and FTSE BRIC ETFs representing 47.4% of AuM in European ETFs in dataset % of AuM in US ETFs in dataset 1

19 Data ETF level l characteristics from Blackrock: benchmark, type of exposure, TER, issuer, primary market and listing date, other listings market data from Bloomberg: daily prices, spreads, volumes and shares outstanding for all listings Index level daily composition, weights and closing values from Bloomberg Stock level market data from Bloomberg: daily prices, spreads and volumes

20 Liquidity measures W k ith l f li idit l l t d We work with low frequency liquidity measures calculated on a monthly basis with daily data (Goyenko and al., 2009) our first level of analysis is given by transaction costs proxies RQS is the relative quoted spread Roll is Roll (1984) effective spread estimator as liquidity providers may compete on prices but not on volumes we compute a price impact proxy Ami. is Amihud (2002) price impact measure we also study the trading volume Volu. is the average daily trading volume aggregated over all listings (in euros) Index liquidity is calculated as the average of measures built at the stock level For European ETFs, liquidity is computed on the primary listing market

21 3. LIQUIDITY IN EQUITY ETFs

22 Univariate analysis We compare mean and median liquidity for Dataset 1 (all equity ETFs) over multiple subsets classifyingetfs by primary listing index style replication type issuer b i i l Wil d t t t f by running simple Wilcoxon and t tests for the difference between medians and means

23 Liquidity by Exchange European ETFs havelower spreadsthanus ETFs In Europe Milan primary lists less expensive ETFs other exchanges exhibit similar spreads In the US Nyse Arca ETFs are cheapest than NASDAQ ETFs These results do not hold for other measures US ETFs have higher liquidity in terms of price impact and trading volumes

24 Liquidity by Index style The size of spreads depends on the index style ETFs replicating developed markets indexes are the cheapest ETFs replicating Emerging market indexes are the most expensive for Global and Developed indexes, sector ETFs have higher spreads surprisingly, in the case of Emerging indices, ETFs replicating national indexes have higher spreads Results qualitatively unchanged for other liquidity measures

25 Liquidity by Replication In the US ETFs using full replication have lower spreads than those using optimizing techniques In Europe there are no significant differences related to the replication technique In terms of price impact the same result hold in the US swap based ETFs have higher price impact than stocks based ETFs

26 Liquidity by issuer q y y (Europe = red dots, US = blue dots)

27 4. THE DETERMINANTS OF LIQUIDITY

28 Multivariate analysis Univariateresults i suggest thatt the liquidity of ETFs depends on the liquidity of their underlying indexes European and US ETFs do not have the same degree of liquidity These results are to be taken with caution selection issues may explain them: it is likely that different sub samples have different distribution of replicated indexes We test these hypotheses on dataset 2 by explaining the liquidity of ETFs by the liquidity of its underlying index and microstrucuture variables with a panel data analysis

29 Explanatory variables Index liquidity measure the liquidity measure for the underlying index constituents portfolio ETF.Volume the average dil daily trading volume of the ETF (in logarithm) Nb.ETFs the number of ETFs in competition on a given index Swap a dummy variable that takes on value 1 for swap based ETFs and 0 for full replication ETFs Leader a dummy variable that takes on value 1 if the ETF is the first listed on this index and 0 otherwise Cross the number of markets on which the ETF is listed Volat the pastmonth benchmarkindex volatility

30 Determinants of Roll spread The liquidity of ETFs significantly depends on its underlying index the higher the index spread, the higher the ETF spread The liquidity of ETFs also depends on the structure of the ETF market more ETFs competing on the same index lead to lower spreads ETFswith higher volume alsoexhibit lower spreads Swap based ETFs have lower spread (but not significant) Additional analysis on European ETFs shows that cross listing does not significantly affect liquidity

31 Determinants of price impact The liquidity of ETFs significantly depends on its underlying index the higher the Index Amihud estimate, the higher the ETF Amihud (ie. the lower the liquidity) The liquidity of ETFs also depends on the structure of the ETF market incumbent ETFs (first to be listed on an index) have lower price impact US ETFs have lower price impact more ETFs competing on the same index lead to increased price impact Swap based ETFs have higher price impact

32 5. CONCLUSIONS

33 Conclusions The equity ETF market ttends to bemore liquidid in Europe than in the US The liquidity of an ETF is the liquidity of its underlying index more liquid indexes are associated with more liquid ETFs The liquidity of an ETF also builds on some characteristics of the ETF market itself ETF trading volume, competition on a given index significantly affect the liquidity the replication strategy does not seem to affect spreads but affects price impact