The Gateway to Algorithmic and Automated Trading

Anatomy of an Algo: Managing News Volatility

Published in Automated Trader Magazine Issue 32 Q1 2014

Trading a stock on the same day it delivers a major investment update almost inevitably involves having to deal with volatility. Eric Krueger, Head of Electronic and Program Distribution, EMEA and Merrell Hora, Head of Equities Algorithmic Product at Barclays show how it can be successfully accomplished if armed with the right execution tools.

Managing News Volatility

Merrell Hora, Head of Equities Algorithmic Product, Barclays

Merrell Hora,
Head of Equities Algorithmic Product, Barclays

The Scenario:

It is the evening of Wednesday, December 11, 2013, and a London-based hedge fund manager has been tracking Anglo American as its price has declined by more than a third since the start of the year. Although the shares continued to decline after Mark Cutifani took over as CEO in April with a brief to stem cost over runs and rebuild value, for a number of reasons the manager feels that the shares may now be approaching the point of recovery.

For instance, while the company's withdrawal from the Alaska Pebble Mine project was widely cited as an expensive embarrassment, the manager also saw it as a pragmatic response - despite the potential yield from the site - because of the associated costs and political/environmental issues. The manager also views the company's recent appointment of a retired analyst and longstanding critic of the mining sector's poor capital discipline as a non-executive director as similarly pragmatic. More recently, and in the run up to major investor briefing by the company tomorrow, a number of analysts have been reiterating buy ratings and/or flagging Anglo American to outperform the sector.

The manager therefore instructs his head of trading to buy 1 million shares (approximately 22% of the average daily lit venue volume for the previous week) during the following day's trading session, Thursday, December 12, 2013, with a limit price of 1296.

The Assets:

Anglo American ordinary shares

The Challenge:

To buy 1 million Anglo American on Thursday, December 12, 2013. The trader has an Implementation Shortfall benchmark and an absolute price limit of 1296.

The Algo:

Barclays's Hydra strategy combines a number of unique features to minimise footprint while maximising liquidity. In order to minimise its footprint and reduce information leakage, the algorithm uses:

• Parallel order placement that simultaneously sends passive, neutral and aggressive limit orders to avoid leaving a footprint

• Dynamic order placement, rather than a pre-determined schedule, so all order placement decisions are based on real-time market information

• Symbol-based logic which incorporates the results of extensive research at the symbol-by-venue level. The execution logic is customised for each stock, as opposed to the 'one size fits all' model of traditional strategies

In order to maximise liquidity, the algorithm takes advantage of:

• Liquidity opportunities, by detecting block liquidity in dark pools and responding by over sizing allocation to those venues

• Price opportunities, by detecting favourable moves in the price of the stock and reacting by stepping ahead of the current trading schedule

• Dynamic limit prices, by sensing an adverse price movements and dynamically adjusting the limit price to reduce or stop trading

By default, Hydra attempts to maximise executions in dark venues (including the broker's own highly liquid dark pool LX), but if there is insufficient liquidity in dark venues, it will also slice orders to lit venues to maintain a trader's specified aggressiveness level, and will opportunistically take liquidity from lit venues when size is available. The algorithm can be configured to post pegged orders with a range of aggression settings (passive, mid, or aggressive) in dark venues and hidden orders on supported exchanges/ECNs. Depending upon the choice of aggression level, the algorithm may also remove liquidity from non-displayed venues before going to exchanges/ECNs.

However, the algorithm also benefits from being delivered within an electronic trading environment that delivers end-to-end order protection. This includes a dynamic order router that synthesises historical and real-time data and executions to predict liquidity for both passive and aggressive trading on both lit and hidden/dark venues. The order router also delivers high performance access to all major exchanges and dark pools via co-located connectivity and market data using hardware accelerated technology.

Another crucial advantage that Hydra offers is an automated capital commitment feature. This provides instant access to liquidity while transferring execution risk to the broker in an automated and anonymous manner. Barclays will facilitate the balance of an order once it passes a pre-determined percentage complete, typically 70%. Automated capital commitment has been shown to lower market impact, shorten trade duration, and increase the likelihood of completing an order.

Eric Krueger Head of Electronic and Program Distribution, EMEA

Eric Krueger
Head of Electronic and Program Distribution, EMEA

The Trader:

The trader allocated the trade by the firm's head of trading specialises in the mining and heavy industrial sectors. He is a long standing user of Hydra and makes regular use of the capital commitment feature.

The Trade Schedule:

(All times are UK)

7.30am:
The trader opts to use the mid aggression setting and despite recent broker buy notes is expecting that the market will come towards the trade on the open. The trader enables the capital commitment feature. By clicking this, the trader will get access to capital on a pre-determined percentage of the order with Barclays.

8.01am:
As the trader expected, Anglo American gaps down on the open to 1281 on the LSE. Good liquidity on both lit and dark venues allows the algo to accumulate 7,200 shares from lit venues and 2,080 in the dark at an average price of 1279.5.

8.15am:
Liquidity is continuing to build strongly and on lit venues is running at approximately twice the rate of earlier trading sessions during the week. The algorithm picks up a further 14,253 and 4,658 shares on lit and dark venues respectively at an average price of 1285.5.

8.29am:
Steady liquidity conditions, a further 9,398 and 4,658 shares acquired on lit and dark venues respectively at an average price of 1287.

10.33am:
162,550 shares accumulated so far on both dark and lit venues at an average price of 1282. Significant selling pressure on decent volume has developed over the past 40 minutes. The stock is now slipping down to approach the 1270 level, so the trader looks to take advantage of a market coming towards him by increasing the algorithm's aggression setting.

10.36am:
The trader's change of aggression setting has paid off as the market has been trading sideways slightly above the 1270 level for the past three minutes allowing the algorithm to pick up a further 42,955 shares at an average price of 1271.

10.55am:
The stock has been rallying strongly for the past ten minutes and is now approaching the 1290 level. A further 90,875 shares acquired at an average price of 1280.5. The trader dials back the algorithm's aggression setting to maintain price control.

10.57am:
Anglo American has continued rallying and is now approaching 1292.

11.23am:
After a brief dip back below 1290 the stock has strengthened again and is now trading at 1297. A further 69,124 shares acquired at an average price of 1293.5.

11.32am:
No further activity has been possible since the last update as the stock has continued to rally above the limit price and is currently trading at 1309.5.

12.30pm:
The investor presentation starts. Anglo American is still trading well above the limit price at 1300.

12.51pm:
The stock has gradually slipped back over the last 20 or so minutes and is now trading around the limit price.

1.05pm:
Now starting to trade consistently just below the limit price. The algorithm restarts meaningful trading activity.

1.29pm:
Steady price decline since the last update.
A further 37,365 shares acquired since then at an average price of 1294.5.

1.38pm:
A further inconvenient rally has again taken Anglo American above the limit price.
A further 77,250 shares acquired since then at an average price of 1294.

2.18pm:
After continuing to rise steadily for much of the past 40 minutes (to above 1300) the stock is again trading below the limit price.

2.32pm :
The investor presentation has provided an extensive description of the operational measures that the company will undertake to boost profitability and push return on equity employed to more than 15% by 2016. However, the consensus seems to be that there will be no quick fix and the stock has accordingly started to drift further lower to around 1290.5. A further 31,555 shares acquired since then at an average price of 1293.5.

3.17pm:
Further declines have seen the stock fall to 1267. 168,500 shares acquired since the last update at an average price of 1274.5.

3.27pm:
118,355 shares acquired since the last update at an average price of 1273. At this point with barely 60 minutes to the end of the trading session, the order passes the 70% completion threshold, and since the limit is marketable, Barclays facilitates the balance of the order at the market price of 1277.5 resulting in an average price for the trade of 1280.71.