AbleMarkets 20-minute Aggressive HFT Index Helped Beat VWAP by 8% Across Russell 3000 Stocks in 2015

Transcription

1 AbleMarkets 20-minute Aggressive HFT Index Helped Beat by 8% Across Russell 3000 Stocks in 2015 Live out-of-sample demo of the 20-minute aggressive HFT index performance in execution on Canadian dollar (USD/CAD) can be seen here: Abstract AbleMarkets 20-minute Aggressive HFT Index tracks bursts of aggressive HFT activity in either buyerinitiated or seller-initiated trades. When buyer-initiated trades exceed seller-initiated trades by at least 10% in any given 20-minute interval, as reported by the AbleMarkets aggressive HFT index, halt the buy schedule while doubling up on the sell volume schedule. Similarly, when sellerinitiated trades exceed buyer-initiated trades by at least 10% in any given 20-minute interval, as reported by the AbleMarkets aggressive HFT index, halt the sell schedule while doubling up on the buy volume schedule. This simple process helped outperform by 8% on average across all of the Russell 3000 equities in Introduction The subject of high-frequency trading has generated lots of research and even more press. All HFT, however, is not created equal and can be broken down into passive and aggressive, based on their preferred execution methods. Passive HFT comprises strategies that favor limit orders. Such strategies include market-making and other similarly patient trading algorithms that can tolerate the high-risk of orders not being executed that comes with limit-order territory. Aggressive HFT strategies, on the other hand, tend to possess time-sensitive trading signals, or rapidly-decaying alpha, and execute aggressively via market orders or marketable limit orders. While passive HFTs create liquidity and have a generally beneficial effect on the markets, aggressive For more information, please contact Steve Krawciw: or steve@ablemarkets.com 1

2 HFTs remove liquidity, increasing slippage and volatility for other market participants. On average, aggressive HFT comprise just 10-40% of all market activity across markets, yet those numbers often hit 100%. Worst of all, aggressive HFT often comprise toxic flow, and create volatility by eroding the limit order book, as well as driving prices by eroding the limit order book. In normal market circumstances, aggressive HFTs avoid accumulating large positions and tend to be fairly balanced: aggressive HFT buy orders are normally swiftly offset by aggressive HFT sell orders. In some cases, however, aggressive HFT participation becomes lopsided with aggressive HFT buyers dominating aggressive HFT sellers and vice versa. Exhibit 1 illustrates the effect of excessive aggressive HFT participation on a sample limit order book. Excessive aggressive HFT buyer-initiated flow wipes out the asks, or the supply side of the limit order book. As a result, when executing following a burst of aggressive HFT buy activity, buy execution becomes too expensive, while the sell execution is considerably cheaper, since the aggressive HFTs are imminently reversing their position, by strategy design. The opposite effect takes place with the aggressive HFT seller-initiated flow. AbleMarkets research shows that when buyer-initiated trades exceed seller-initiated trades by at least 10% in any given 20-minute interval, the following 20 minutes are characterized by overly expensive buy trades and relatively cheap sell trades. Similarly, when seller-initiated trades exceed buyer-initiated trades by at least 10% in any given 20-minute interval, the market conditions during the following 20 minutes are better suited for buying than for selling. Price you see Price you get Spread you see Spread you get price Exhibit 1. Impact of aggressive HFT on a limit order book. For more information, please contact Steve Krawciw: or steve@ablemarkets.com 2

3 AbleMarkets Aggressive HFT Index AbleMarkets aggressive HFT index is a index probabilistically measuring participation of aggressive high-frequency traders as a proportion of volume traded. The index separately reports the participation of aggressive HFT buyers and aggressive HFT sellers. Specifically, the participation of aggressive HFT buyers is measured as the traded volume initiated by the aggressive HFT buyers over the total buyerinitiated volume within a given time period. Similarly, the participation of aggressive HFT sellers is measured as the traded volume initiated by the aggressive HFT sellers over the total seller-initiated volume within a given period of time. The AbleMarkets aggressive HFT index measures the participation of aggressive HFT in lit markets (exchanges) only, and does not measure aggressive HFT participation in dark pools and other trading venues. The AbleMarkets aggressive HFT index is available for most equities, ETFs, commodities, foreign exchange, futures and options. The 20-minute AbleMarkets aggressive HFT index discussed in this paper is delivered automatically every 20 minutes either via XML or FTP, with open-source scripts for data retrieval provided by AbleMarkets. and AbleMarkets Aggressive HFT Index Overlay Overview Optimal execution refers to the efficient allocation and intraday timing of trades such that a large order, split up into small trades, is traded at a price better or equal to a selected benchmark. Typically, the order-originating entity is a large portfolio manager, such as a hedge fund or a pension fund. The execution brokers can be investment banks, private divisions of hedge funds, and other independent brokers. Execution is a large business, processing close to US$ trillion in daily volume worldwide, at least half of which originates in the United States. Volume-Weighted Average Price () is one of the most commonly used benchmarks in pricing of execution. Whether it is optimal or not, has become a standard benchmark in the execution business due to its intuitive nature, transparency and ease of calculation. Most execution brokers deploy in the following framework: Step 1: for the previous 20 days, compute 1-minute volume buckets corresponding to each minute of the trading day Step 2: average those buckets across the 20 days to create a volume map Step 3: on the day of execution, divide the total amount to be executed into equal trades to be executed each minute, then scale them by the volume map determined in Step 2: when historical volume is high, For more information, please contact Steve Krawciw: or steve@ablemarkets.com 3

4 execute the proportionally larger trade; then execute a smaller trade when the volume is historically low. Step 4: Execute the scaled trade program determined in Step 3. Adding logic based on the AbleMarkets aggressive HFT Index (AHFT) does not radically change the nature of. Instead, the AHFT logic is an incremental improvement to that generates considerably positive results. The basic AHFT- overlay algorithm works as follows: Step 1: record the AbleMarkets aggressive HFT Index values for buyer and seller-initiated trades for the past 20 minutes. Step 2: Compare the AHFT buyer and AHFT seller values for each financial instrument. Then: - When AbleMarkets Aggressive HFT Buyer Index exceeds AbleMarkets Aggressive HFT Seller Index by 10% or more, halt all further buying in the given instrument and double up on selling for the next 20 minutes in anticipation of the aggressive HFT reversing their strategies. - When AbleMarkets Aggressive HFT Seller Index exceeds AbleMarkets Aggressive HFT Buyer Index by 10% or more, halt all selling in the given instrument and double up on buying for the next 20 minutes. The appendix provides a practical example for deployment of AbleMarkets Aggressive HFT Index in generation of AHFT execution overlay. Out-of-Sample AHFT Overlay Performance Over 2015 alone, the AHFT overlay delivered 8% lower than cost for execution of buy orders, and 8% higher than price for execution of sell orders, on average for all Russell 3000 constituents. The AHFT overlay improved execution for the 2884 constituents of the entire Russell 3000 index. Among the 116 of the Russell 3000 stocks that have underperformed with the AHFT overlay, the average underperformance was 0. 09% higher cost for execution of buy orders and -0.1% lower cost of execution of sell orders, summed over the entire 2015, a marginal amount given the gains of the majority of Russell 3000 constituents. All the test results were performed strictly out of sample. For a live out-of-sample demo of execution in Canadian dollar foreign currency exchange rate, USD/CAD, please visit For some securities, the added returns are phenomenal. Exhibit 1 shows the performance of the 20- minute Aggressive HFT-based strategy over TWAP and on NYSE:A over January- May 2015 period (more extensive results are available, fees may apply). The overlay works as follows: whenever the 20- minute AHFT Sell activity exceeds the 20-minute AHFT Buy activity by 10% or more, the strategy doubles up on buying, as prescribed by the underlying strategy, and halts all selling for 20 minutes. Conversely, whenever the 20-minute AHFT Buy activity exceeds the 20-minute AHFT Sell activity by 10% or more, the strategy doubles up on selling, and halts all buying for 20 minutes. The results shown for NYSE:A For more information, please contact Steve Krawciw: or steve@ablemarkets.com 4

5 Percent gain of AHFT overlay over TWAP and Able Markets over the first 5 months of 2015 include nearly 120% gain over traditional, and close to 100% gain over TWAP. What are the key differences in performance of AHFT overlay among the Russell 3000 stocks? Exhibits 3 and 4 summarize results of the cross-sectional analysis of 2015 performance of AHFT overlay on the Russell 3000 securities vis-à-vis several factors: share price, market capitalization, dividend yield and market risk, or beta. As Exhibits 3 and 4 show, the key determining factors of AHFT performance appear to be: 1) market capitalization and 2) dividend yield of stocks. Furthermore, decreases in either market capitalization or dividend yield among the Russell 3000 stocks tend to positively affect performance of the AHFT Overlay. In other words, AHFT overlay is more likely to lower the cost of buy execution in stocks with smaller market cap and lower dividend yield. Similarly, AHFT overlay is more likely to raise the price of sell execution in stocks with smaller market cap and lower dividend yield. The majority of the AHFT overlay performance, however, cannot be explained by these four common factors, as indicated by a great statistical significance of the intercept. 140% 120% 100% 80% Improvement of TWAP and with AHFT Overlay 60% 40% 20% Gain of AHFTBuy vs Buy Gaino f AHFTSell vs. Sell 0% -20% Exhibit 2. Performance of AHFT on NYSE:A. For more information, please contact Steve Krawciw: or steve@ablemarkets.com 5

6 Exhibit 3: Cross-sectional regression of AHFT performance in buy execution on Russell 3000 equities. Dependent variable: AHFT Buy Independent variables: Market price, market capitalization, dividend yield, beta Coefficients Standard Error t Stat Intercept Market price 6.87E E Market capitalization 6.3E E Dividend Yield Beta Adjusted R 2 : 1.2% Exhibit 4: Cross-sectional regression of AHFT performance in sell execution on Russell 3000 equities. Dependent variable: AHFT Sell Independent variables: Market price, market capitalization, dividend yield, beta Intercept Market price Market capitalization Dividend Yield Beta Adjusted R 2 : 1.3% Coefficients Standard Error t Stat E E E E Conclusion AHFT Overlay delivers a powerful easy to implement add-on to any -benchmarked execution system. It is easy to deploy and to monitor, either with human eye or electronically. While the results of the analysis suggest that the AHFT overlay performs better on stocks with lower market capitalization and lower dividend yield, the majority of the Russell 3000 stocks perform well out of sample. For more information, please contact Steve Krawciw: or steve@ablemarkets.com 6

7 Appendix: A Practical Example It is 10:20 AM. The schedule, in number for the next 20 minutes is as shown in Exhibit 1. # # 10: : : : : : : : : : : : : : : : : : : : Exhibit 1. A sample execution schedule. AbleMarkets Aggressive HFT Index reports the following aggressive HFT activity in the last 20-minutes of trading in AAPL: Case 1: Aggressive HFT Buy: 43%, Aggressive HFT Sell: 31%. Action: double up on schedule for AAPL sell orders; halt execution of AAPL buy orders, as shown in Exhibit 2. schedule, # of shares, 10: : : : : : : : : : : : : : : : : : : : Exhibit 2. AHFT-Modified execution schedule following AHFT buyers exceeding AHFT sellers. For more information, please contact Steve Krawciw: or steve@ablemarkets.com 7

8 Case 2: Aggressive HFT Buy: 11%, Aggressive HFT Sell: 22%. Action: double up on schedule for AAPL buy orders; halt execution of AAPL sell orders, as shown in Exhibit 3. 10: : : : : : : : : : : : : : : : : : : : schedule, # of shares, Exhibit 3. AHFT-Modified execution schedule following AHFT sellers exceeding AHFT buyers. For more information, please contact Steve Krawciw: or steve@ablemarkets.com 8

9 Case 3: Aggressive HFT Buy: 25%, Aggressive HFT Sell: 23%. Follow the program, execution schedule 1. 10: : : : : : : : : : : : : : : : : : : : schedule, # of shares, Exhibit 3. AHFT-Modified execution schedule following AHFT sellers comparative to AHFT buyers. For more information, please contact Steve Krawciw: or steve@ablemarkets.com 9

10 Legal Disclaimer Following AHFT- execution overlay strategy involves a degree of risk. There is no guarantee that the AHFT- execution overlay performance objective will be achieved. The past performance of the AHFT- execution overlay should not be construed as an indicator of future performance the AHFT- execution overlay strategy. Presentations to potential clients may contain numerous forward looking statements which only reflect the opinion of AbleMarkets and may not turn out to be accurate. About AbleMarkets Aggressive HFT Index AbleMarkets Aggressive HFT (AHFT) overlay is a streaming index delivering 20-minute averages of aggressive HFT activity across Russell 3000 equities in the U.S. lit markets. Russell 3000 instruments comprise 90% of the U.S. equity trading. The AbleMarkets Aggressive HFT index is a probabilistic framework based on nearly a decade of proprietary research of AbleMarkets. About AbleMarkets AbleMarkets is a leader in market microstructure analytics, co-founded by Irene Aldridge, an expert in high-frequency trading (author of High-Frequency Trading: A Practical Guide to Algorithmic Strategies and Trading Systems, Wiley, 2nd edition). ABLE Markets has pioneered the space since 2007, so that today you can deploy a range of reliable tools to increase portfolio returns, reduce execution costs and reduce risks by using our market microstructure intelligence. Delivered seamlessly, our world-class quantitative and high-frequency research is used by high- and low-frequency fund managers, brokers, exchanges and family offices. For more information, please contact Steve Krawciw: or steve@ablemarkets.com 10