According to a well-known proverb, “a little knowledge is a dangerous thing.” In the world of asset management, making an investment decision based on half the facts or incomplete understanding can have alarming consequences. However, with the emergence of big data and advances in technology enabling faster and more comprehensive analytics, many believe it will provide managers with unparalleled amounts of information, facilitating better performance. Or will it?
Making the case for big data is simple. Take a fund management house actively trading listed equities, for example. The decision to buy or sell a specific security will be based on fundamental research, namely an analyst forensically scrutinizing the contents of a company’s public or regulatory filings; account statements; and sell-side brokerage reports. A lot of time is consumed combing through all of this documentation, which in turn leads to heightened costs.
Through sophisticated big data analytics, like artificial intelligence (AI), all of this information could be scrubbed and analyzed in near real time—or real time—allowing for better cost control and resource allocation at asset managers. AI’s processing power is such that it can compute far more information over a shorter time period to a higher degree of precision than any human being. Off the back of such innovation, portfolio managers will be able to execute trades more readily, with far more assurances that the fundamentals underpinning their decisions are strong.
Or so that is how it is meant to work in theory. In practice, the situation is likely to be much more complicated. Any student of history will testify that a lot of the raw materials in circulation, which form the bedrock of academic studies, are often flawed or downright inaccurate. Big data is certainly not bulletproof and having a lot of information should not be confused with having credible information.
Inputting inaccurate data into robotic processing automation (RPA) tools will yield flawed judgement calls, making it critical that managers monitor the technology and carefully validate its results and recommendations. Equally, the mechanics and algorithms underpinning AI software may have anomalies resulting in erroneous analytics, which might include identifying trends that are totally random or coincidental, based on fundamentals that are unsound. Again, it is important managers verify not just data quality on a regular basis, but the effectiveness of the AI software itself.
Service providers including custodian banks, fund administrators and independent technology or research providers are looking for ways to monetize the data they hold. Many are trying to package this information and sell it to clients, a lot of whom may not have the internal capabilities to cleanse or make sense of it on their own. However, fund managers need to ensure the data supplied by providers is accurate, and this can be achieved by thorough back-testing exercises.
Simultaneously, a number of data providers are coming into the market with a limited track record, so it is strongly advisable managers conduct robust due diligence, just as they would on any other counterparty or critical service provider. Accuracy of data is important, but so is its legitimacy and security. Regulators globally are introducing legislation around data security and privacy. Managers must therefore ensure any information they receive is rightfully acquired from providers who are in full compliance with data privacy and protection laws.
A number of experts increasingly refer to data as being an asset class in its own right. Effective and lawful use of accurate big data by fund managers can help generate superior returns or implement meaningful business development improvements. Conversely, substandard big data, deficiencies in AI software and an absence of a human overlay in monitoring analytics is a serious business risk—one that must be avoided if firms are to leverage the swell of information now available to them.
Alternative Investments, Asset Management