As a consequence of the pandemic, quant hedge fund managers painfully discovered computer-driven trading decisions based mainly on value investing strategies may no longer be foolproof. Should the quants have relied more on alternative data providers, these hedge funds would perhaps have been able to adjust their long-term positions to the risk realities presented by the pandemic.
Although big data and alternative data are often used interchangeably, there is a fundamental difference between the two concepts. Within the context of hedge funds, big data is helpful for macroeconomic analysis, including establishing the dynamics between currencies, commodities prices, foreign exchange market activities and economic activities specific to a country or at the global level. On the other hand, alternative data, a subset of big data, is information collected from strategic sources relevant to a specific context or microeconomics, even down to the level of individuals, businesses or particular sectors of the economy. Alternative data is qualitative or quantitative financial information sourced from non-financial sources such as mobile application analytics, product reviews or online browsing activity. However, the definition of alternative data is transient because once it is discovered and aggregated (structured) for sale, it becomes mainstream and offers little to zero outperformance value. There are many benefits of using alternative data for hedge funds, which often yield business intelligence with exceptional levels of granularity. Beyond just finding alpha, alternative data sources are also helpful for risk management, corporate development, business intelligence and support decisions around investing in private equity.
Throughout 2020, data shows that discretionary hedge funds outperformed their quants counterparts. For instance, two leading quantitative hedge funds, Renaissance Technologies and Two Sigma registered unprecedented losses though the decline in general industry performance had already begun earlier. Labelled the ‘quant winter’, the period of 2018-2020 saw the industry contend with poor returns and job cuts following the underperformance of quants that had been witnessed for nearly a decade preceding the quant winter. Reports indicate that 55% of quants have posted losses since 2016. On the other hand, 2020 was a good year for discretionary hedge funds. For instance, the Pierre Andurand discretionary fund went short on crude oil following accurate predictions on the effect of lockdowns gained 152%. The China-based Mandarin Offshore fund averaged 28% in returns after assertively buying US-based tech stocks during the bear market at the peak of the pandemic. The positive performance of discretionary hedge funds was owing to strategies and the use of alternative data.
An Alternative Investment Management Association (AIMA) report indicates that quants and discretionary funds apply data differently. Discretionary managers use alternative data and fundamentals to validate their trading decisions on open positions. In contrast, quants use big data and traditional data to feed computer models that make and take trading positions without the human element.
Due to the centrality of data, fund managers have the option of either building an internal data ecosystem or source from alternative data providers when considering data integration strategies. Many find that it is often more cost-effective to use a data provider rather than incur the costs of setting up infrastructure and human resources necessary for their ecosystem. Alternative data providers source and analyse information that is not otherwise publicly available and present it for use to hedge funds.
While considering alternative data providers, hedge funds should seek to work with those capable of providing unstructured, structured or semi-structured data considering the transient nature of this data. The cleaning and analysis of unstructured data are done using machine learning or natural language processing technologies. Ideally, a preferred data provider also may have the capability to develop exclusive unsupervised ML models that feed on unstructured data, to provide a basis for a unique trading thesis or strategy. The models can also be built around specific securities of interest or any other factor relevant to strategy.
Alternative data providers with capacity to provide value in horizon scanning processes can support quantitative, credit, and event-driven funds, among the biggest COVID losers, to avoid the impact of sudden turns in the market and black swan events. Data that can aid horizon scanning can serve as a basis for justifying the increase in exposure for arbitrage volatility funds.
Despite the evident value in alternative data, only 48% of large firms with over US $1 billion in Assets Under Management (AUM) invest in this data. The AIMA report indicates that going forth, both systemic and fundamental managers will streamline alternative data into their processes. But before investment firms can draw value from alternative data sources, they have to contend with the fact that alternative raw data is often expensive, incomplete, irrelevant or unverified. Some alternative data sources gather data outside the General Data Protection Regulation (GDPR) or similar regulations. Alternative data providers can help hedge funds navigate these challenges.
To benefit from alternative data providers, hedge funds should work with data providers with a demonstrable capacity for handling unstructured data. Tools such as AMPLYFI’s DeepInsight, for example, can aid this. The platform provides a comprehensive repository of resources consisting of hundreds of millions of expert documents and sources, covering patents, scientific papers and news articles with thousands of new documents added every day. With such depth of information, DeepInsight can reveal insights on demand, including unexpected connections between entities, ESG factors, events, sentiment, or geographical location.
