By Francesca Campanelli, Chief Commercial Officer at Axyon AI
Volatility in the markets caused by the COVID-19 pandemic continues to present fund managers with significant operational and performance-related challenges. Returns have been below expectations for most, while traditional quantitative models have largely failed to understand the new world and the new underlying rules. To improve the generation of alpha, fund managers will need to rapidly change their approach to how they manage risk and predict future market conditions.
Without effective information, processed at the speed at which markets are moving, investment managers could be operating ineffectively.
Out of sync
The quality of data and how it is used is absolutely crucial for managers today.
In fact, a recent report from Deloitte, which examined various financial analyst reports published between 1st January and 6th March, found that just a small percentage mentioned COVID-19. Of those reports that did reference the virus, 28% were negative, 52% were positive and 20% were neutral. Clearly analysts did not see the outbreak as acting as a bearish trigger for large-cap value stocks.
Yet, during February news stories revealed that many businesses were facing a direct impact from the virus, and business metric data revealed a rapid decline in consumer activity. The application of advanced analytics to these data sets may have provided more real-time insight into the risks associated with Covid-19 for the stock market.
However, these new, more-timely inputs require investment decision makers to adapt too and, in many cases, it needed to be processed with new analytical techniques, including AI-powered technology.
COVID-19 has also completely changed the investment landscape. The crisis has almost certainly sped up the speed of change within the sector, but it has also shifted the markets in an unpredictable direction. In other words, the financial markets can and are reacting in an anomalous way.
For example, in February 2020 at the start of the pandemic, Axyon AI’s anomaly detection system signalled the oil market had reached an unprecedented level of 100% anomaly. By March, oil prices had fallen to a 17-year low, with prices recording the biggest daily loss since the end of the Gulf War in 1991. Just a few weeks later, the same warning of anomalous data appeared on our system regarding a large number of markets when COVID-19 was spreading. In these situations, fund managers had to react immediately to manage the risk and make sure investments were protected.
However, fund managers were held back by traditional risk management models that weren’t set up to deal with black-swan events such as the pandemic. These models are built around strong assumptions on the behaviour of underlying assets, measuring normal distribution patterns on linear scales. While this means they’re effective during times of relative stability, they cannot cope with the flood of chaotic data into their systems caused by record levels of volatility. This made it extremely difficult for fund managers to gain an accurate view of predicted market changes.
Now, a second wave of coronavirus is emerging in economies across the globe. Yet, many fund managers are still continuing to use outdated models that leave them vulnerable to another market fall.
Preparing for the future
While nobody can predict the unpredictable, advanced AI models can detect anomalies ahead of time and be a powerful way of anticipating future events in the markets which are considered unprecedented, such as the COVID-19 pandemic.
Unlike traditional risk models, this type of AI system is completely agnostic about markets and any associated risks, meaning it can be trained just to detect anomalies in the data. However, AI is trained to raise the alarm when the structure in the data is anomalous, which might be a sign of an upcoming unpredictable event.
AI models are also flexible when it comes to reading the reality of the situation at hand, and without any pre-ordered rules. These systems are able to change depending on real data, being able to exploit huge volumes of data and computing power.
In particular, deep learning is a leading-edge technology in artificial intelligence that could benefit fund managers the most. Deep learning can allow fund managers to capture more complex, non-linear patterns in asset behaviour, unlock expanded uses of alternative data, and allow algorithms to adapt to changing market conditions in real-time.
Because of this, fund managers can gain valuable time to adjust risk and protect investments, while not worrying about the chaotic data flooding in and whether it could impair the model.
These AI-powered tools can be used by fund managers to enhance current strategies or even build one from scratch that is truly adapted to the new landscape. By using a new strategy, managers would be able to diversify the generation of alpha in an existing investment process or, importantly in times like these, allow managers to mitigate risk.
The COVID-19 pandemic is an opportunity for new technologies to prove their merits and show that AI and machine learning can offer a better way to make the most of data and adapt to the new normal. Traditional portfolio models have had their limitations exposed during this unprecedented time and aren’t fully fit for purpose in the new landscape. It’s crucial that fund managers are able to recognise this and adapt before the possibility of another market crash happens and they feel the full brunt.
In contrast, AI-powered solutions are flexible to changing market developments and can alert fund managers about a possible market crash. Deep learning especially is a tool that can help fund managers generate alpha through their superior predictive capabilities. Those managers which embrace these tools as part of their investment strategy will be able to lead from the front using data to better protect investors and generating alpha in uncertain and challenging times like these.