Sand in the Chips? Evidence on Taxing Transactions in Modern - - PowerPoint PPT Presentation

Sand in the Chips? Evidence on Taxing Transactions in Modern Markets

Jean-Edouard Colliard Peter Hoffmann HEC Paris ECB Market Microstructure - Confronting Many Viewpoints - December 10, 2014

The views expressed here are the authors’ and do not necessarily reflect those of the ECB or the Eurosystem.

Road map

Introduction Methodology and data Average impact The role of high-frequency market-making Institutional trading Conclusion

Context

◮ Renewed interest for financial transactions taxes (FTT)

since the onset of the crisis:

◮ Budget deficits ◮ Public discontent with financial sector ◮ 11 countries committed to a European FTT

◮ “Historical” evidence rather negative, but...

◮ No micro evidence ⇒ few/no links to economic

mechanisms

◮ Poor data (no counterfactuals, emerging markets, etc.) ◮ Market structure has changed (e.g. HFT)

UK Stamp Duty Revenue/Volume

This paper

◮ Evidence on the French FTT (August 2012)

◮ Stamp duty on ownership transfers ◮ Exemptions for liquidity provision ◮ Tax on domestic non-MM HFT activity

◮ Diff-in-diff with other Euronext stocks ◮ (Lit) market quality

◮ Role of MM exemption ◮ Trader types (HFT, nonHFT, Mixed)

◮ Institutional trading and impact on different clienteles

Results

◮ On average, the FTT had a negative, but muted impact

• n market quality

◮ Drop in volume (−10%), depth, resiliency, price

efficiency

◮ No effect on bid-ask spread and volatility

Results

◮ On average, the FTT had a negative, but muted impact

• n market quality

◮ Drop in volume (−10%), depth, resiliency, price

efficiency

◮ No effect on bid-ask spread and volatility

◮ Is there any role for the MM exemption?

Results

◮ On average, the FTT had a negative, but muted impact

• n market quality

◮ Drop in volume (−10%), depth, resiliency, price

efficiency

◮ No effect on bid-ask spread and volatility

◮ Is there any role for the MM exemption?

◮ Stocks w/ HFT MM: no impact ◮ Stocks w/o HFT MM: ր vola, price impact, spreads

Results

◮ On average, the FTT had a negative, but muted impact

• n market quality

◮ Drop in volume (−10%), depth, resiliency, price

efficiency

◮ No effect on bid-ask spread and volatility

◮ Is there any role for the MM exemption?

◮ Stocks w/ HFT MM: no impact ◮ Stocks w/o HFT MM: ր vola, price impact, spreads

◮ More informed market orders, decrease in liquidity

provision

◮ Exempted HFT MMs continue providing liquidity

⇒ structure of liquidity provision matters for tax design

Results

◮ To wash out the effect of HFT, we analyze institutional

trading (changes in portfolio holdings)

◮ Average impact of −20% ◮ Larger reduction in holdings of French stocks and trading

volume for:

◮ funds with high turnover (Amihud and Mendelson

(1986))

◮ index funds

Literature - Theory

◮ Keynes (1936), Tobin (1978), Stiglitz (1989), Summers

and Summers (1989), Schwert and Seguin (1993)

◮ Potential Rationales for taxing (some) trading activity:

◮ Noise traders (DeLong et al. (1990b)) ◮ Speculators (DeLong et al. (1990a), Di Maggio (2013)) ◮ Intermediaries (Menkveld and Yueshen (2013)) ◮ Short-termism (De Long, Shleifer, Summers, and

Waldmann (1990b))

◮ Net effect determined by changes in market composition

◮ Kupiec (1996), Song and Zhang (2005), Bloomfield et

• al. (2009)

◮ Empirical challenge: need disaggregated evidence

Literature - Empirics

◮ Empirics: Roll (1989), Umlauf (1993), Campbell and

Froot (1994), Jones and Seguin (1997), Hau (2006), Baltagi et al. (2006), Pomeranets and Weaver (2012), Liu and Zhu (2009), Deng et al. (2014), etc.

◮ Empirical challenges

◮ no disaggregated data ⇒ harder to link to theory ◮ emerging markets or pre-2000s ◮ no X-section, no control groups

◮ Other papers on French FTT confirm average impact:

Meyer et al. (2014), Haferkorn and Zimmermann (2013), Capelle-Blancard and Havrylchyk (2013), Coelho (2014), Becchetti et al. (2013).

Road map

Introduction Methodology and data Average impact The role of high-frequency market-making Institutional trading Conclusion

The policy experiment

◮ France introduced an FTT on August 1st, 2012 ◮ Restricted to French stocks with market cap ≥ 1 bln e ◮ Purchases are taxed at 0.2% (ownership transfers) ◮ Several exemptions (market making, primary market) ◮ Additional tax on HFT (traders residing in France) ◮ HFT defined as activity below a threshold of 0.5 seconds

Identification strategy

◮ Standard diff-in-diff ◮ Control group: non-French stocks traded on Euronext ◮ Identification assumption: common trends ◮ Standard checks: placebo DiD, visual inspection ◮ Allow for differences in short-/long-run impacts

E(yi,t | i, t) = αi + γt + βAugDAug

i,t

+ βSep/OctDSep/Oct

i,t ◮ We focus on Sep/Oct (August “polluted” by seasonality)

Identification: FTT and HFT Tax

29 stocks 85 stocks 18 stocks 32 stocks 1 bln. EUR French Dutch FTT+HFT Tax HFT Tax only FTT only

Data

◮ Sample period: Jun 1st - Oct 31st (109 days) ◮ Stocks traded on Euronext with “sufficient liquidity” ◮ Drop: Banks (Crisis), PT (Crisis), BE (change in local

FTT)

◮ 117 above 1 bln EUR (85 vs. 32), 47 below 1 bln EUR

(29 vs. 18)

◮ Trades, quotes, LOB changes (TRTH) ◮ Trader group IDs (only FR): HFT, nonHFT, Mixed

(Source: EUROFIDAI/AMF)

◮ Institutional Portfolios (Factset)

Road map

Introduction Methodology and data Average impact The role of high-frequency market-making Institutional trading Conclusion

Intraday price range minus pre-tax average - 2012

Mar 1 May 1 Jul 1 Aug 1 Sep 1

0.5 0.0 0.5 1.0

Trading volume (log) minus pre-tax average

9.9 33.

June 1 July 2 August 1 September 3 October 1

0.8 0.6 0.4 0.2 0.0 0.2

Regression results - June/July vs. Sep/Oct

Variable/Group French > 1 bln French < 1 bln Volume

• 0.104**

0.040 (-2.46) (0.45) Volatility 0.365 1.854 (0.41) (1.14) Effective Spread 0.029 0.243 (0.16) (0.30) Price Impact 0.201

• 0.365

(1.28) (-0.77) Depth

• 10.761***
• 4.805

(-2.85) (-1.45) Resiliency

• 0.018*

0.021** (-1.90) (2.08) AR 0.007**

• 0.004

(2.09) (-0.74)

Discussion

◮ Overall, the FTT’s impact is rather muted

◮ Spreads and volatility are unchanged ◮ Economically small reductions in resiliency and price

efficiency

◮ Result on depth potentially driven by reduced demand for

liquidity (e.g. Parlour and Seppi (2003))

◮ Consistent with a positive role of safeguarding liquidity

provision

◮ Market making exemption ◮ Ownership transfers

Road map

Introduction Methodology and data Average impact The role of high-frequency market-making Institutional trading Conclusion

Two groups of stocks

◮ SLP stocks: special program of Euronext, rebates for

HFTs providing liquidity if they subscribe to the program.

◮ Non SLP stocks: smaller stocks without this feature,

HFT less prevalent.

Variable/Group SLP Non SLP > 1 bln % volume % l.o. % volume % l.o. HFT 27.5 27.3 16.9 3.7 Mixed 56.4 55.4 55.7 65.2 Non HFT 16.0 17.3 27.3 31.0

◮ Dutch SLP stocks as a control for French SLP stocks ◮ Few control stocks non SLP > 1 bln EUR

⇒ stocks < 1 bln EUR as control for non MM stocks

Four groups of stocks

• ●

■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■

◆ ◆ ◆ ◆ ◆ ◆ ◆ ◆ ◆ ◆ ◆ ◆ ◆ ◆ ◆ ◆ ◆ ◆ ◆ ◆ ◆ ◆ ◆ ◆ ◆ ◆ ◆ ◆ ◆ ◆ ◆ ◆ ◆ ◆ ◆ ◆ ◆ ◆ ◆ ◆ ◆ ◆ ◆ ◆ ◆ ◆ ◆ ◆ ◆ ◆ ◆ ◆ ◆ ◆ ◆ ◆ ◆ ◆ ◆ ◆ ◆ ◆ ◆ ◆ ◆ ▲ ▲ ▲ ▲ ▲ ▲ ▲ ▲▲ ▲ ▲ ▲ ▲ ▲ ▲ ▲ ▲ ▲ ▲ ▲ ▲

12 13 14 15 16 17 18 19 Log Volume 5 10 15 20

Median cancellation time(ms)

• FR SLP

■ NL SLP

◆ FR Non SLP ▲ NL Non SLP

Impact of the FTT on SLP and non SLP stocks

Variable/Group SLP Non SLP Log Volume

• 0.032
• 0.225***

(-0.69) (-3.46) Volatility 0.479 2.612** (0.51) (2.29) Range

• 0.069

0.272** (-0.62) (2.11) Effective Spread 0.175 1.144** (1.58) (2.23) Price Impact 0.181 1.805*** (1.37) (5.65) Depth

• 12.750***
• 2.265

(-2.72) (-1.19) Resiliency

• 0.008
• 0.027***

(-0.82) (-3.03) AR 0.009** 0.008 (2.24) (1.52)

Impact on price range - SLP stocks

Impact on price range - Non SPL stocks

0.29

• 0.031

July 2 August 1 September 3 October 1

Price range - Non SLP

Wrap-up

◮ More adverse impact on stocks without HFT

Market-makers

◮ Volume: tax-exempt market-making activity

clouds impact on volume

◮ Liquidity: higher adverse selection ◮ Two potential mechanisms

◮ Change in trading population, informed/uninformed ◮ Change in behavior or trading strategy

Gains on market orders and limit orders

Variable Horizon HFT Mixed Non HFT Price impact 10s 5.46 3.62 3.14 5min 6.12 5.16 4.26 30min 6.53 5.95 4.59 Realized spread 10s 4.25 2.15 2.06 5min 3.19 1.17 0.64 30min 2.51 0.85

• 0.08

Impact of the FTT on non SLP stocks

Variable/Group Log volume HFT

• 0.336**

(-2.18) Mixed

• 0.266***

(-3.29) Non HFT

• 0.017

(-0.19)

Note: not representative of HFT for larger stocks.

Impact on make/take decision

Group/Share (%) Limit orders Market orders HFT

• 0.53
• 6.71***

(-1.55) (-4.34) Mixed

• 5.42***

3.07** (-3.76) (2.20) Non HFT 5.96*** 3.63** (5.90) (2.48)

Impact on revenues of market and limit orders

Measure/Account type HFT Mixed Non HFT Price impact 2.040*** 2.341*** 1.870*** (4.46) (4.17) (3.43) Realized spread

• 1.677*
• 1.130***
• 1.037**

(-1.81) (-3.28) (-2.41)

HFT Volume

• 29. %
• 11. %

July 2 August 1 September 3 October 1

Log HFT Volume

Mixed Volume

• 23. %
• 26. %

July 2 August 1 September 3 October 1

Log Mixed Volume

Non HFT Volume

• 1.7 %
• 22. %

July 2 August 1 September 3 October 1

Log Non HFT Volume

Discussion

◮ Support for both mechanisms

◮ ց HFTs ⇒ less informed market orders ◮ ց Mixed ⇒ limit orders sent by less informed agents ◮ All trader types focus on more informed trades

◮ Evidence consistent both with:

◮ FTT (relatively) encouraging more informed strategies /

discouraging liquidity provision

◮ Market-making exemption being efficient at protecting

liquidity provision in SLP stocks

◮ Additional evidence on composition effects

⇒ data on institutional trades

Road map

Introduction Methodology and data Average impact The role of high-frequency market-making Institutional trading Conclusion

Impact on long-term investors

◮ Aggregate market activity clouds the impact on end

investors

◮ Large chunk of volume is short-term and tax-exempt ◮ Difficult to filter out the impact on the taxpayer

= end investor

◮ Use trading volume estimated from changes in portfolio

holdings (Factset) Volumef

i,t = |xf i,t − xf i,t−1| × pi,t

◮ evaluate the average effect on institutional trading

activity

◮ investigate differences across investor types (clienteles)

◮ DID: Q4 or Q3 2012 vs. Q2 2012

Tax revenues

◮ Aggregate market activity clouds the impact on end

investors

◮ We identify 20% of the total tax revenues

⇒ tax on long-term investors

◮ Break-down by country:

8.45 5.75 3.59 1.33 0.74 0.7 0.47 0.28 0.25 0.87

US LU FR GB DE IE SE BE CH Other

Mutual funds - Summary statistics

Variable Mean

• St. Dev.

PRICE TO BOOK 3.22 1.14 SIZE 1126 6,315 TURNOVER

• 0.89

1.12 INDEX 0.13 0.34 EURO FUND 0.25 0.43 INTERNATIONAL FUND 0.32 0.47

Impact on holdings Q3-2012 - Q2-2012

Dependent variable: [ln xf

FR,Q3 − ln xf FR,Q2] − [ln xf NL,Q3 − ln xf NL,Q2].

(1) (2) Cons. 0.008

• 0.018

(0.76) (-0.79) TURNOVER

• 0.031***

(-2.63) PRICE TO BOOK

• 0.019*

(-1.80) LOG SIZE

• 0.004

(-0.68) INDEX

• 0.186***

(-5.46) INTERNATIONAL FUND 0.072*** (3.09) EURO FUND

• 0.001

(-0.04)

Impact on volumes Q4-2012 - Q2-2012

Dependent variable: [ln volumef

FR,Q4 − ln volumef FR,Q2] −

[ln volumef

NL,Q4 − ln volumef NL,Q2].

(3) (4) (5) Cons.

• 0.200***
• 0.164**
• 0.157**

(-5.39) (-2.42) (-2.33) TURNOVER

• 0.117***
• 0.105***

(-3.40) (-3.09) PRICE TO BOOK

• 0.175***
• 0.167***

(-5.16) (-4.90) LOG SIZE

• 0.055***
• 0.053***

(-2.72) (-2.66) INDEX

• 0.637***
• 0.561***

(-7.33) (-6.45) INTERNATIONAL FUND

• 0.129
• 0.159*

(-1.51) (-1.86) EURO FUND

• 0.042
• 0.041

(-0.48) (-0.48) DID HOLDINGS 0.411*** (2.75)

Impact on end investors

◮ Other proxies for end investor trading activity

◮ Call auctions (open & close) ◮ Dark pools ◮ Over-the-counter (MIFiD reporting)

Trading mechanism/Coefficient βSep/Oct T-Stat Euronext LOB

• 0.107**

(-2.52) Auction

• 0.167***

(-3.81) OTC

• 0.566***

(-6.49) Dark

• 0.496***

(-2.85)

◮ Overall, results are consistent with a considerable effect

• n end investors

◮ Large variation: technical trades (e.g. Dividends)

+ different clienteles

◮ Same evidence as before, but more indirect

Discussion

◮ Tax severely impacted the most active of the long-term

investors

◮ “Pigovian” nature of this change unclear at best! ◮ Impact hidden when looking only at aggregate data and

short-term activity

◮ Change in market composition, selling and less trading by:

◮ high turnover funds (Amihud and Mendelson (1986)) ◮ index funds (availability of substitutes?)

⇒ impact on the informed/uninformed ratio (Bloomfield et al. (2009))

Road map

Introduction Methodology and data Average impact The role of high-frequency market-making Institutional trading Conclusion

Conclusion

◮ Overall, the FTT’s average impact is relatively muted,

but:

◮ Significant discrepancy across stocks ◮ Adverse selection increases for less liquid stocks ◮ Stocks with HFT MMs more protected against this effect ◮ Underlines the role of the MM exemption

◮ Stamp duty discourages investment strategies with

horizon > 1 day

◮ Reshuffling of shares among end investors to minimize the

impact of the tax

◮ Impact on the clienteles of French companies

Implications for the European FTT

◮ European FTT has much wider scope

◮ Other asset classes (bonds, derivatives...) ◮ No exemptions, all transactions

◮ Protecting liquidity provision seems important ◮ The EC expects to raise 4.8 to 6.5 bln EUR for equities ◮ We estimate that extending the French implementation to

EU-27 would raise 3.4 bln EUR

◮ A more cautious implementation need not forego too

much revenue

References I

Amihud, Y., and H. Mendelson (1986): “Asset pricing and the bid-ask spread,” Journal of Financial Economics, 17(2), 223 – 249. Baltagi, B., D. Li, and Q. Li (2006): “Transaction tax and stock market behavior: evidence from an emerging market,” Empirical Economics, 31, 393–408. Bloomfield, R., M. O’Hara, and G. Saar (2009): “How Noise Trading Affects Markets: An Experimental Analysis,” Review of Financial Studies, 22(6), 2275–2302. Campbell, J. Y., and K. A. Froot (1994): “International Experiences with Securities Transaction Taxes,” in The Internationalization of Equity Markets, NBER Chapters, pp. 277–308. National Bureau of Economic Research, Inc.

References II

Capelle-Blancard, G., and O. Havrylchyk (2013): “The Impact of the French Securities Transaction Tax on Market Liquidity and Volatility,” Working paper. Coelho, M. (2014): “Dodging Robin Hood: Responses to France and Italy’s Financial Transaction Taxes,” Discussion paper. De Long, J. B., A. Shleifer, L. H. Summers, and R. J. Waldmann (1990a): “Noise Trader Risk in Financial Markets,” Journal of Political Economy, 98(4), 703–38. (1990b): “Noise Trader Risk in Financial Markets,” Journal of Political Economy, 98(4), 703–38. De Long, J. B., A. Shleifer, L. H. Summers, and R. J. Waldmann (1990c): “Positive Feedback Investment Strategies and Destabilizing Rational Speculation,” The Journal of Finance, 45(2), pp. 379–395.

References III

Deng, Y., X. Liu, and S.-J. Wei (2014): “One Fundamental and Two Taxes: When Does a Tobin Tax Reduce Financial Price Volatility?,” Working Paper 19974, National Bureau of Economic Research. Di Maggio, M. (2013): “Market Turmoil and Destabilizing Speculation,” Working paper. Haferkorn, M., and K. Zimmermann (2013): “Securities Transaction Tax and Market Quality - The Case of France,” Working paper. Hau, H. (2006): “The Role of Transaction Costs for Financial Volatility: Evidence from the Paris Bourse,” Journal of the European Economic Association, 4(4), 862–890. Jones, C. M., and P. J. Seguin (1997): “Transaction Costs and Price Volatility: Evidence from Commission Deregulation,” The American Economic Review, 87(4), pp. 728–737.

References IV

Keynes, J. M. (1936): The General Theory of Employment Interest and Money. Palgrave MacMillan. Kupiec, P. (1996): “Noise traders, excess volatility, and a securities transactions tax,” Journal of Financial Services Research, 10, 115–129. Liu, S., and Z. Zhu (2009): “Transaction Costs and Price Volatility: New Evidence from the Tokyo Stock Exchange,” Journal of Financial Services Research, 36, 65–83. Menkveld, A., and B. Yueshen (2013): “Middlemen Interaction and Its Effect on Market Quality,” Working paper. Meyer, S., M. Wagener, and C. Weinhardt (2014): “Politically Motivated Taxes in Financial Markets: The Case of the French Financial Transaction Tax,” Journal of Financial Services Research.

References V

Parlour, C. A., and D. J. Seppi (2003): “Liquidity-Based Competition for Order Flow,” Review of Financial Studies, 16(2), 301–343. Pomeranets, A., and D. Weaver (2012): “Security transaction taxes and market quality,” Working paper 2011-26, Bank of Canada. Roll, R. (1989): “Price volatility, international market links, and their implications for regulatory policies,” Journal of Financial Services Research, 3, 211–246. Schwert, G. W., and P. J. Seguin (1993): “Securities Transaction Taxes: An Overview of Costs, Benefits and Unresolved Questions,” Financial Analysts Journal, 49(5), pp. 27–35. Song, F. M., and J. Zhang (2005): “Securities Transaction Tax and Market Volatility,” The Economic Journal, 115(506), 1103–1120.

References VI

Stiglitz, J. (1989): “Using tax policy to curb speculative short-term trading,” Journal of Financial Services Research, 3, 101–115. Summers, L., and V. Summers (1989): “When financial markets work too well: A cautious case for a securities transactions tax,” Journal of Financial Services Research, 3, 261–286. Tobin, J. (1978): “A Proposal for International Monetary Reform,” Eastern Economic Journal, 4(3-4), 153–159. Umlauf, S. R. (1993): “Transaction taxes and the behavior of the Swedish stock market,” Journal of Financial Economics, 33(2), 227–240.