Implementation shortfall (IS) is the most used benchmark for portfolio managers and yet it is one of the most poorly designed algorithms by most brokers, according to Hitesh Mittal, Founder & CEO, BestEx Research.
One of the reasons IS algorithms, including liquidity-seeking algorithms, often fall short is due to the “noisy nature of their measurement,” he said.
The primary factors contributing to higher IS, namely market impact costs, spread costs, rapid alpha decay during order execution, and adverse selection costs, can be challenging to discern for a small order pool, he added.
“It’s often unclear whether the IS resulted from natural price fluctuations of the underlying instruments or from these cost factors,” he told Traders Magazine.
“When measurement is lacking, the algorithm’s implementation is often inadequate,” he added.
Furthermore, some cost factors are inversely correlated: minimizing spread costs necessitates more liquidity-providing execution strategies, but these can increase adverse selection, Mittal said.
Several other factors unique to each portfolio manager, such as risk aversion to missing the benchmark, the alpha profile of the trades, and the size of the order for which they’re using the algorithm, mean that “different implementations will be optimal for different traders,” he explained.
“While minimizing IS is the goal, IS algorithms may not always be the best choice for buy-side firms,” he argued.
“They are certainly not the right choice for buy-side firms unless they understand the internal design of the algorithm, the rationale behind the broker’s design choice, and have assurance that the crucial aspects such as execution speed are optimal for their specific use case,” he added.
Mittal said that the design of IS algorithms needs to be highly customized and tailored to meet the unique needs of each portfolio manager.
He explained that several factors contribute to these needs – the size of orders that can vary from as small as half a percent of the Average Daily Volume (ADV) to exceeding multiple days of ADV, the differing volatility of underlying instruments, and the diverse ways in which the algorithms are utilized.
Some buy-side firms may employ algorithms in a purely low-touch manner, entrusting the entire order to the algorithm, whereas others may opt for a hybrid approach where traders are actively engaging with orders, seamlessly swapping them in and out of algorithms, he said.
“In that case, the algorithms are used more as a tool rather than being the core driver of the strategy and only receive a portion of the order,” Mittal said.
He further said that the nuances of a portfolio manager’s investment approach also significantly impact the design of IS algorithms.
For instance, high turnover portfolio managers are more concerned about maintaining lower average costs as opposed to focusing on the execution risk associated with an individual order, he said.
“On the other hand, portfolio managers who deal with rapid alpha decay would require algorithms that can keep pace and trade swiftly,” he added.
According to Mittal, while algorithms like Volume Weighted Average Price (VWAP) or Percentage of Volume (POV) offer traders control over defining the order’s duration and participation, IS algorithms demand further optimization.
“They need to cater to the diversity of orders and portfolio managers,” he said.
“Sell-side firms have a crucial role to play in this scenario,” he said.
“They need to engage closely with buy-side firms to grasp their objectives and understand the distinct characteristics of their orders,” he added.
This knowledge is essential to optimize the behavior of their IS or liquidity-seeking algorithms, thus ensuring they are best suited to their clients’ unique circumstances, Mittal said.
According to Mittal, to address this issue, BestEx Research has developed a first of its kind framework to allow traders to design their own custom implementation shortfall algorithm that fits their profile best.
“Adaptive Optimal IS is not just an algorithm, it’s a comprehensive, flexible framework that we’ve integrated into our AMS’s Strategy Studio,” he said.
“This allows for the creation of an Implementation Shortfall (IS) strategy that’s perfectly tailored to the specific needs of each client,” he said.
Usually, brokers can customize their strategies, according to Mittal, but this process can take several months and is typically reactive, in response to a client’s request rather than a proactive design decision.
The Adaptive Optimal IS framework enables the design of such strategies in a day, instead of months, avoiding these makeshift solutions, he said.
He stressed that a key component of this framework is the capacity for A/B testing, which allows for direct comparison of strategies and helps determine which design decisions work best for each firm and trader.
As for other algos, Mittal said that issues are inevitable.
“It’s important for buy-side firms to challenge sell-side firms to disclose the inner workings of their algorithms and question their design assumptions. Trade planning is just one part of the execution strategy; the models for placing internal limit orders are equally critical,” he said.
