Sie sind auf Seite 1von 9

THE BLOTTER | 28 August 2013 | Volume 4 | Issue 8

The Blotter presents ITGs insights on complex global market structure, technology, and policy issues.

Canadian Order Flow Trends Effect of NASDAQ Outage on Canada


Highlights Message traffic levels trended higher: Level 1 message traffic averaged 25.5MM messages in Q2-2013, 30% higher than Q1 levels. The quality of order flow degraded: Order-to-Trade and Volume Traded-to-Order ratios worsened and the rate of order flow increased. The percentage of long lived orders fell to 11.7% vs 15.4% Q1. Trade volume remained near-flat QoQ, increasing only 1% to average 444MM shares daily.

Contributors Doug Clark Managing Director Dora Lee Vice President

contact Asia Pacific +852.2846.3500 Canada +1.416.874.0900 EMEA +44.20.7670.4000 United States +1.212.588.4000 info@itg.com www.itg.com

From the combination of increased fleeting orders, declining order flow quality, and tell-tale order flow footprint, we find evidence that activity from HFT participants increased during the quarter. Canadian impact from the NASDAQ outage on 22-Aug-2013: We analyze the order flow characteristics, quoted spread, quoted volume and conclude that liquidity in some stocks were minimally affected while others were significantly impacted. Furthermore, we find that trading venues with inverted pricing models experienced a greater loss of order activity.
FIGURE 1: Canadian Lit Market Order Flow During NASDAQ Outage
25 20 Message Count (000's) 15 10 5 0 09:30 10:00 10:30 11:00 11:30 12:00 12:30 13:00 13:30 14:00 14:30 15:00 15:30 CHIX2

Outage Period

NASDAQ interlisted order flow in Canada fell significantly as cross-border strategies were unable provide liquidity due to the US trading halt.

CHIX
Source: ITG

ALPHA

TSX

OMEGA

PURE

TMXS

28 August 2013 | Volume 4 | Issue 8

Message Traffic After hitting a low in Q1-2013, order flow steadily trended higher. Both our Order-toTrade (O:T) and Volume Traded-to-Order (V:O) metrics show a decline in order flow quality. In the following analysis, we explain that a number of factors may be at play. The daily level 1 message traffic gained 30% to average 25.5MM vs 19.5MM messages in Q1 (Figure 2). Quote volatility decreased 3.9%, trade volatility increased 5.5%, and average daily traded volume didnt vary significantly from Q1, rising only 1% to 444MM shares.
FIGURE 2: Volatility and Message Trac (1)
120 Message Trac (Millions) 100 80 60 40 20 0 4-Jan-10 30-Jun-10 29-Mar-11 27-Jun-11 23-Mar-12 21-Jun-12 20-Mar-13 29-Sep-10 29-Dec-10 26-Sep-11 22-Dec-11 20-Sep-12 18-Dec-12 18-Jun-13 1-Apr-10 Q1 Q2 14 Intraday Volatility (bps) Volume Traded-to-Order Ratio 12 10 8 6 4 2 0

THE BLOTTER

Message traffic levels trended up this quarter. Level 1 traffic averaged 25.5MM messages in Q2-2013, 30% higher than Q1 levels.

L1 Message Trac
Source: ITG

Quote Volatility

Trade Volatility

From Figure 3, both O:T and V:O ratios continued the Q1 trend for the worse: the Order-to-Trade ratio increased 28% to average 54 and Volume Traded-to-Order ratio fell 33% to 9.9. The degradation of both the O:T and V:O metrics by such substantial amounts came from overall higher order flow levels without similar gains in the number of trades, nor total shares traded during the quarter. This is a strong indicator that activity from HFT participants has increased. The introduction of CX2 on 21-May-2013 also contributed to increased L1 message traffic levels. Of the lit trading venues, the CX2 proportion of message traffic averaged 11% while its daily market share averaged 4%.
FIGURE 3: Message Trac Ratios for Canadian Lit Markets (1)
140 120 Order-to-Trade Ratio Q1 Q2 25 20 15 10 5 0
4-Oct-10 18-Nov-10 6-Jan-11 22-Feb-11 13-Apr-11 31-May-11 19-Jul-11 2-Sep-11 20-Oct-11 6-Dec-11 24-Jan-12 9-Mar-12 25-Apr-12 12-Jun-12 27-Jul-12 13-Sep-12 30-Oct-12 13-Dec-12 31-Jan-13 19-Mar-13 3-May-13 19-Jun-13

Average Order-to-Trade and Volume Traded-to-Order ratios both degraded in Q2. O:T gained 28% while V:O decreased 33%.

100 80 60 40 20 0

Order-to-Trade Ratio (O:T)


Source: ITG

Volume Traded-to-Order Ratio (V:O)

28 August 2013 | Volume 4 | Issue 8

THE BLOTTER

This points to lower efficiency for CX2, since the venue contributes 2.75x more to message traffic than to shares traded. Table 1 below ranks the relative efficiency of the Canadian lit markets. A messages-to-marketshare ratio greater than 1 means that the venue is less efficient: it is producing more than its share of messages than its proportion of trade volume. One explanation for a high ratio may be greater HFT participation at the affected venues. Order Duration

Table 1: Relative Efficiency of Canadian Lit Markets


Ratio of % L1 Messages to % Marketshare TSX ChiX Pure Alpha CX2 TMXS Omega
Source: ITG

0.51 1.33 1.67 1.99 2.75 2.87 3.69

For traders who use depth-of-book information, order duration is an important metric to help measure the accessibility of the quoted price and volume. The order duration metric measures the quality of liquidity in the order book and is a relevant consideration for traders. Short-lived orders contribute to noise and cloud the true liquidity available in the order book. As the lifetime of orders approaches the time-constant of the trading infrastructure (the time it takes for information to travel between the exchange and trading terminal), the chance of getting filled against these fleeting orders become increasingly remote. Figure 4 below shows the distribution of fleeting orders (orders persisting for <60sec) in the Canadian lit markets. Consistent with L1 message traffic, total level 2 order activity also increased in Q2-2013. The order duration demographics this quarter show that the proportion of long-lived orders (>60sec) fell to 11.8% from 15.4% in Q1. The percentage of fleeting orders in the sub-10ms range fell while orders in both the 10ms-to-100ms and 100ms-to-1s time ranges increased 11% and 18% respectively.
FIGURE 4: Percentage of Fleeting Orders in the Canadian Lit Markets(2)
100% 100

Canadian lit markets ranked by relative efficiency. A messages-to-marketshare ratio greater than 1 means the venue is less efficient and is generating more than its proportion of message traffic.

80%

80

In Q2-2013, total order activity increased 4.3% and order activity accelerated: the percentage of long-lived orders >60s decreased 23% QoQ.

60%

60

40%

40

20%

20

0%

Oct-10

Dec-10

Feb-11

Apr-11

Jun-11

Aug-11

Oct-11

Dec-11

Feb-12

Apr-12

Jun-12

Aug-12

Oct-12

Dec-12

Feb-13

Apr-13

Tot. Num. Of Orders (R) Orders <=1sec


Source: ITG

Orders <=10ms Orders <=10sec

Orders <=100ms Orders <=60sec

Jun-13

Total Number of Orders (Daily Avg, Millions)

Percentage of Orders (%)

28 August 2013 | Volume 4 | Issue 8

By examining the order flow footprint, we find that a number of market participants showed a distinct increase of fleeting orders. Combined with higher total order flow and worsened market quality metrics, we have further evidence to indicate that activity from HFT participants has increased during the quarter. Table 2 below summarizes order duration for three time brackets on a typical day at the end of Q2-2013 to highlight the demographics of lit order flow across the Canadian lit venues. Points to note: Across all venues, the percentage of long-lived orders decreased Pure and Omega saw the greatest percentage increase of fleeting orders
THE BLOTTER

CX2 and TMXS rank highest with the greatest percentage of fleeting orders (in excess of 91%)
Table 2: Order Duration Across Canadian Lit Markets(2)
% of Orders with Duration Below 1 Second Alpha Chi-X TSX Pure Omega TMXSelect CX2(3)
Source: ITG

10 Seconds 71.2% 74.5% 70.2% 64.9% 70.6% 77.8% 78.2%

60 Seconds 87.4% 89.6% 86.4% 83.9% 86.4% 91.3% 92.9%

43.1% 48.1% 42.6% 41.3% 43.7% 52.7% 53.0%

Impact of NASDAQ Outage on Canada Over the last few years, Canadian exchanges has experienced its share of system outages(4,5). As market participants, we have first-hand knowledge of the downstream effect on equities trading on the other venues. If a major US trading venue stops trading, what is the impact on Canadian equities? After technical interruptions to the securities information processor data feed (SIP, a national consolidated tape of all US exchanges) earlier on, NASDAQ experienced a further technical problem that halted trading on NASDAQ-listed stocks in the US between and ~12:14h and ~15:25h on 22-Aug-2013.
FIGURE 5: Canadian Lit Market Order Flow During NASDAQ Outage
25 Outage Period 20
Message Count (000's)

During the outage period, the level of order activity for interlisted NASDAQ tickers fell substantially in the Canadian lit markets.

15

10

0
09:30 10:00 10:30 11:00 11:30 12:00 12:30 13:00 13:30 14:00 14:30 15:00 15:30

CHIX
Source: ITG

ALPHA

TSX

OMEGA

PURE

TMXS

CHIX2

Typical

28 August 2013 | Volume 4 | Issue 8

In this section, we use this natural experiment to examine what happens in Canada when a major US trading venue shuts down. We analyze the order flow characteristics, quoted spread, quoted volume and conclude that liquidity in some stocks were minimally affected while others were significantly impacted. Furthermore, we find that trading venues with inverted pricing models experienced a greater loss of order activity. Of the universe of NASDAQ-listed stocks, we look at 30 that are interlisted with Canada. Figure 5 above charts the level 2 order flow activity on 22-Aug-2013 for interlisted NASDAQ securities trading on the Canadian lit venues. Compared to the typical message count, Aug 22nd started the day slightly below average and drops off significantly during the outage. The sharp decline in activity highlights the level of cross-border activity that is normally present for these symbols. Absent these intermediaries, the remaining activity between 12:14h and 15:25h likely reflects natural order flow in these symbols.
FIGURE 6: Percentage Distribution of Order Flow During NASDAQ Outage
20% 100% 5%

THE BLOTTER

80%

Distribution of order flow to the Canadian lit markets for interlisted NASDAQ symbols. Order activity to TMX Select, CX2, and Omega fell significantly

Message Count

60%

Outage Period

40%

20%

0% 09:30 10:00 10:30 11:00 11:30 12:00 12:30 13:00 13:30 14:00 14:30 15:00 15:30 CHIX2

CHIX
Source: ITG

ALPHA

TSX

OMEGA

PURE

TMXS

The next question to address is: what are the preferred venues that receive order flow from arbitraging intermediaries? Without a hedging destination, intermediaries for the interlisted symbols shut down their strategies. As a result, any change observed in order flow distribution amongst the lit markets will highlight the preferred venues. Figure 6 shows the distribution of order activity for the Canadian lit markets. During outage period, order flow contribution from TMX Select, CX2, and Omega fall from over 20% to below 5%. This reveals that venues with an inverted pricing model receive a substantial level of their order activity from intermediary trading activity. With a loss of trading participants in the Canadian markets during the outage, we now look at how liquidity for the interlisted NASDAQ symbols were affected - did the spread and quoted volume significantly change? Our analysis finds two major categories of stocks in the NASDAQ-interlisted set of symbols: (A) those having limited spread change and (B) those with significantly affected spread. Symbols such as BB, PAA, and SSO fall into category (A) while examples of category (B) stocks include NDQ, RGL, and SHS.

28 August 2013 | Volume 4 | Issue 8

Figure 7 compares the intra-day characteristics of quoted spread for the stocks in categories (A) and (B). Once the NASDAQ trade halt was announced, the average spread for category (A) stocks increased 5% from 58bps to 61bps while category (B) stocks increased 25% from ~150bps to ~190bps. What caused this discrepancy? We uncover two characteristics of category (A) vs (B) stocks to help explain the difference. First, we find that the majority of symbols in category (A) have significant Canadian volume. Figure 8 illustrates the daily volume distribution between Canada and the US for stocks in both categories (A) and (B). The majority of category (A) stocks, marked with an asterisk (*), are on the left (larger % of Canadian volume), while most of category (B) stocks are on the right (larger % of US volume). With a significant percentage of trading volume normally in Canada, the narrow spread of category (A) stocks is likely supported by Canadian-based order flow. Second, from the diminished order rate found in Figure 5, the order flow footprint of the remaining participants is consistent with natural order flow. In the case of category (A) stocks, the level of Canadian-based natural order flow was sufficient to maintain a competitive spread even in the absence of cross-border strategies. For category (B) stocks in Figure 8, the majority of trading occurs largely in the US. With US participants in short supply (whether natural or intermediary) the spread widened significantly as the remaining activity on the Canadian exchanges was unable to provide competitive liquidity.
FIGURE 7: Top-of-Book Spread for Category (A) and (B) Stocks
500 Outage Period 400

THE BLOTTER

The spread profile is significantly different for NASDAQ-interlisted symbols that have Canadiandominant natural participants (A) vs. US-dominant intermediary participants (B)

Spread (bps)

300 200 100 0 09:30 10:00 10:30 11:00 11:30 12:00 12:30 13:00 13:30 14:00 14:30 15:00 15:30

PercSpread (A)
Source: ITG

PercSpread (B)

Turning our attention to quoted volume, we also find evidence that category (A) stocks were supported by dominantly-Canadian natural flow. Figure 9 compares the aggregate bid/ask volume for both stock categories. During the outage, category (B) stocks lost as much as 75% of the pre-halt volume while category (A) stocks lost only 25% of the quoted volume. Without contributions from US intermediary flow, liquidity was adversely affected in Canada for (B) stocks.

28 August 2013 | Volume 4 | Issue 8

FIGURE 8: Canadian vs US Daily Trade Volume Distribution for 22-Aug-2013


100% 80% CA to US Ratio 60% 40% 20%

Majority US Volume

Majority Canadian Volume


* WIN * DSG * TGL * EXF FSV * BU * MX * OTC * STB * SW * IM CCT BLD DWI * PAA * BB * SSO SMA MNW SOY * TTH MDZ.A WPT QLT SHS NDQ * AEZ SWI HYG RGL CAN Volume USA Volume Typical

THE BLOTTER
Daily volume distribution between Canada and US for NASDAQ-interlisted symbols

0%

Stocks marked with (*) belong to category (A), otherwise stock is category (B) Source: ITG

FIGURE 9: Top-of-Book Volumes for NASDAQ-Interlisted Stocks


CATEGORY (A) STOCKS 150 100 # Shares (000's) 50 0 -50 -100 CATEGORY (B) STOCKS 60 40 # Shares (000's) 20 0 -20 -40 09:30 10:00 10:30 11:00 11:30 12:00 12:30 13:00 13:30 14:00 14:30 15:00 15:30 Outage Period Outage Period

Canadian Top-of-Book volume profile during NASDAQ outage for interlisted symbols did not vary as dramatically for category (A) stocks that have Canadian-dominant natural participants vs category (B) stocks.

BidVolume
Source: ITG

Oer Volume

28 August 2013 | Volume 4 | Issue 8

One could argue that similar observations could be made when Canadian markets are open on US holidays however the on-off immediacy of the transitions before and after the NASDAQ trade halt allows this event to very closely approximate a pure study of one factor: the effect of US order flow with all-else being equal. In conclusion the ~3-hour NASDAQ outage provided a rare opportunity to gain a variety of insights into interlisted trading on Canadian lit venues: Due to interlisted trading strategies, a trading halt in the US dramatically reduces intermediary order flow on Canadian markets The demographics of orders for venues with inverted pricing models is heavily skewed towards intermediary trade flow
THE BLOTTER

For stocks with sizeable Canadian natural flow, the absence of interlisted strategies did not significantly reduce quoted volume, nor increase spreads For stocks that have lower Canadian traded volume, displayed liquidity worsened due to absent cross-border trading strategies In normal operating conditions, intermediary trading activity does contribute to improved spread and quoted volume for stocks with lower levels of liquidity

28 August 2013 | Volume 4 | Issue 8

Methodology and Notes


(1)

Includes all TSX-listed symbols excluding debentures, warrants, preferred shares, and notes. Message traffic includes level I activity for all exchanges. We define the duration of an order as the time between its first appearance in the order book and any subsequent modification and/or cancellation. Three ranges of fleeting orders are defined: <10ms, <100ms, and <1second. Canadian Market Data Trends - Chi-X Outage on 31-Aug-2011, ITG, September 2011. Canadian Market Data Trends - TSX Trading Disruption on 30-Nov-2011, ITG, December 2, 2011.

(2)

(3)

Chi-X2 data for this quarter spans 21-May-2013 to 28-Jun-2013

(4)

THE BLOTTER

(5)

2013 Investment Technology Group, Inc. All rights reserved. Not to be reproduced or retransmitted without permission. # 82613-17201 The opinions, positions, and/or predictions taken or made in this document reflect the judgment of the individual author(s) and are not necessarily those of ITG. These materials are for informational purposes only, and are not intended to be used for trading or investment purposes or as an offer to sell or the solicitation of an offer to buy any security or financial product. Nothing contained herein should be relied upon as a representation, guarantee, or warranty as to the reasonableness of the assumptions or the accuracy of the sources used by the author(s). These materials do not provide any form of advice (investment, tax or legal). ITG Inc. is not a registered investment adviser and does not provide investment advice or recommendations to buy or sell securities, to hire any investment adviser or to pursue any investment or trading strategy. All trademarks, service marks, and trade names not owned by ITG are the property of their respective owners. Broker-dealer products and services are offered by: in the U.S., ITG Inc., member FINRA, SIPC; in Canada, ITG Canada Corp., member Canadian Investor Protection Fund (CIPF) and Investment Industry Regulatory Organization of Canada (IIROC); in Europe, Investment Technology Group Limited, registered in Ireland No. 283940 (ITGL) and/or Investment Technology Group Europe Limited, registered in Ireland No. 283939 (ITGEL) (the registered office of ITGL and ITGEL is Block A, Georges Quay, Dublin 2, Ireland). ITGL and ITGEL are authorised and regulated by the Central Bank of Ireland; in Asia, ITG Hong Kong Limited (SFC License No. AHD810), ITG Singapore Pte Limited (CMS Licence No. 100138-1), and ITG Australia Limited (AFS License No. 219582). All of the above entities are subsidiaries of Investment Technology Group, Inc. MATCH NowSM is a product offering of TriAct Canada Marketplace LP (TriAct), member CIPF and IIROC. TriAct is a wholly owned subsidiary of ITG Canada Corp.