Next-generation algos can overcome enduring risks.
Minimising the risks involved in portfolio trading is a fast-evolving art, according to Ian Peacock, global head of execution services, CA Cheuvreux, and Chris Foxall, managing director, Global Portfolio Trading, a CA Cheuvreux & CLSA service.
Recent changes to equity market microstructure - including polluted pre-trade and opaque post-trade data, liquidity fragmentation and the rise of the dark pool and the high-frequency trader - have made achieving best execution a much more complex business for the buy-side trader. Algorithmic and DMA volumes have risen at the expense of the high touch order, but portfolio trading still accounts for 18% of commissions paid by European institutional investors (Greenwich Associates, 2010) in European markets and we believe it should retain its position as a mandatory vehicle in the world of best execution.
Global portfolio trading has changed significantly over the last decade. It used to be that getting the file format correct and the trade settled on time was enough of a distinction to stay relevant. While those requirements are still paramount, the evolution of portfolio trading due to advances in technology and investment strategies provides an opportunity for portfolio trading desks to become increasingly valuable to their buy-side counterparts. Riskbased trading has been around for a considerable amount of time, but is not ideal for most investment managers due to the inherent conflicts that can arise. However, the need to reduce volatility of trading performance has never been greater. With today's powerful analytical tools, the risk factors of any given trade can be identified fairly accurately. Along with the ability to see sector, beta and cash skew, it is now possible to look at the marginable contribution of risk made by each trading component. Trading desks that use this level of risk analysis can make more informed decisions when deciding how to implement a portfolio rebalance, a transition or a benchmark relative cash flow.
The buy-side desk structure can influence the way the risk of a portfolio trade is managed. If a rebalance is split up according to trader's responsibilities, there can be significant unintended consequences. For example, if a desk has sector traders, and each manages a portion of the portfolio trade, they may introduce unnecessary exposures by trading the most liquid names due to their minimal market impact estimates. Typically, traders look at the largest order or the most illiquid as the most risky. However, this is often not necessarily the case. Without understanding how each name increases or reduces the risk of the portfolio trade it is impossible to make a completely informed decision. The same issues occur when buy-side desks are organised based on region. There maybe unnecessary risk introduced at a sector level that could be mitigated by trading the portfolio as a whole, or at least understanding the marginable contribution of risk of each component.
Other complexities of the trading process such as cash management or even shortterm alpha can create unexpected results. Portfolio managers often don't understand why their estimated trading costs bear no resemblance to realised trading costs. This is in part due to the incomplete information passed through to the trading desks or brokers, as well as the way the portfolio trade is implemented. As a solution to this problem, some traders are becoming more aligned to portfolio managers or strategies as well as using portfolio desks. Algorithmic trading, dark pools and multilateral trading facilities have a very important place in the trading landscape. When it comes to portfolio trading, there needs to be a more intelligent method of trade implementation other than purely focusing on reducing implicit and explicit costs. Portfolio-based algorithms are still in their infancy but are providing a good way to negotiate the majority of complexities discussed here. The ability to apply constraints at the portfolio level or choose an aggressiveness level based on estimated risk or alpha content allows the portfolio traders to focus on adding peripheral value to the trade. The goal of a true portfolio algo is to minimise the risk-adjusted trading costs of the execution relative to the arrival price for each line in the portfolio. This algo takes into account the correlation between all stocks in the portfolio. Traditionally, traders use implementation shortfall or other algos to trade each line on the portfolio, which doesn't take into account the correlative effects within the overall portfolio and is therefore sub-optimal. Advanced portfolio algos can accept constraints and modes such as:
a market-sector neutral mode that will govern real-time aggression and keep within portfolio cash balance constraints;
a benchmark choice that will allow liquidation vs. the arrival price or the closing fixing
an 'adaptive' mode which dynamically adapts to price-sensitive benchmark slippage to overn real-time aggression.
An integrated approach
Finding natural liquidity with the ability to re-optimise in real time, as well looking at technical and other factors that can affect the portfolio during implementation, are just a few of the ways that portfolio desks are changing to remain relevant to the buyside. Integrating the portfolio management and construction processes with the trading function is the Holy Grail for investment managers. In addition, they need a global platform to seamlessly execute through a vast network of execution venues, in order to maximise access to pools of natural liquidity worldwide. In response to these needs CA Cheuvreux and CLSA have launched a new co-branded Global Portfolio Trading platform. As the move into passive and enhanced index strategies continues, a global agency model will help clients achieve best execution in an unconflicted manner. Brokers need to focus on the next generation of portfolio algorithms to further this quest, as well as offering a consultative approach to developing specific trading strategie