The Multi-Venue FX Algo

Issue 09 Q2 2008
Automated Trader Magazine

Ian Smith, Director, and Cameron Mouat, Head of AES FX Trading, Credit Suisse, provide an example of how execution algorithms capture liquidity across multiple trading venues in the foreign exchange market.

The Transaction: A portfolio manager at the London office of a large US-owned asset manager wants to construct an approximate hedge to offset its exposures to the Australasian economies. The Kiwi (New Zealand Dollar) has experienced a 25-year high against the US dollar in recent months due to concerns about a lengthy US recession, but it has encountered resistance at around 0.8210, due to perceived weaknesses in the New Zealand economy. The transaction is aimed in part at offsetting exposures created by buying Kiwi- denominated stocks during the recent NZD surge.

The Challenge: To place a 250m NZD/USD hedge (i.e. selling NZD), using a target price based on analysis of price movements in the last six months, at the beginning of the UK trading day, when changes in liquidity and market sentiment can impact price significantly. As the hedge is imprecise, the trader
may be requested to adjust his position by the portfolio manager, depending on market conditions.

The Algo: An FX algorithm designed to identify and access liquidity across multiple foreign exchange trading venues and streaming price feeds from banks. The algorithm reacts automatically to changing liquidity levels by adjusting the size of the orders it places on venues to avoid pushing the market if volumes in a particular pair are low. The algorithm works in conjunction with a smart order routing (SOR) platform that shifts orders between venues in real time, based on current factors and historical liquidity.

The Trader: The trader is an FX specialist working for a large equity-focused conventional asset manager who wants to complete the deal without impacting the market and understands that this may take time.

7.30am: The Kiwi is trading at 0.7927 against the US dollar as the trader configures his algorithm to sell 250m of NZD for USD. Generally, liquidity in the NZD/USD pair picks up noticeably throughout the London morning and all pairs are currently sensitive to any changes in the domestic or global economic outlook, due to continued concerns over the impact of the credit crunch. The trader sets a limit price of 0.7900 and expects the increasing liquidity throughout the morning to work in his favour. He places the whole transaction through the FX algo and expects it to take liquidity at all the price sources to which it has access (i.e. both trading venues and banks’ natural interest pricing streams). Initially, the algorithm is set to place offers at the minimum level permitted by individual venues (e.g. NZD 1m on Reuters and less for ECNs with less liquidity), to avoid signalling risk while volumes are low, but will rise automatically to take advantage of greater liquidity later in the session. The algorithm will also aggressively take liquidity at prices which are established without pushing the market to new levels. ...

Qualified Registrants & Subscribers Only Content

This article is for qualified registrants and subscribers only, please either log into your account above and then qualify your account or click here to purchase a subscription for Automated Trader Magazine.