The hedge fund world is fast changing. The industry has been long dominated by impressive personalities, famous for resorting to contrarian strategies for success (think Paul Tudor Jones). However, just like many other sectors of the economy, Big Data and AI are now increasingly playing a role in hedge fund management. The era of alternative data is here.
A Background on Hedge Fund Trends
The trend of Big Data and artificial intelligence has been precipitated by the fact that hedge funds’ performances have weakened over the past 15 years. Investors pulled a whopping net $98 billion out of global hedge funds in 2019 and withdrew a further $33 billion in the first quarter of 2020.
An accommodative policy stance by the Federal Reserve, since the 2008 financial crisis, has led to increased correlations between assets like individual stocks and indices. These high correlations have made it difficult to strategise asset allocation and execute long/short strategies. Moreover, increased government stimulus has led to lower yields and stretched equity valuations. As a result, passive investments in stocks and bonds have gained ground.
Hedge funds seem to be staging a comeback in 2021 though, thanks to huge stimulus measures by the Biden administration, with a gain of 4.8% in Q1 2021, the strongest quarter since 2006. Cryptocurrency hedge funds are also rising, outperforming Bitcoin with returns of 117% in Q1 2021. Hedge funds are likely to attract $30 billion from investors in 2021, which will be the first annual net inflow since 2017.
Going Beyond the Traditional Approach
New macro-economic trends are pushing portfolio managers to rethink their positions, to meet investor demand. For instance, there will be new opportunities for hedge funds as investors resort to sector rotation. Cryptocurrency investment has become popular, as has the trend of ESG – environment, social and governance – investing.
A survey by the Alternative Investment Management Association (AIMA) and fund service group SS&C has revealed that 53% of hedge funds, with a combined AUM of $720 billion, reported resorting to non-traditional data sets for new ideas.
The New Approach to Data Analysis
The use of AI and machine learning is playing out across a huge section of investment managers, from traditional fundamental investors to quant traders and even pure AI-driven specialists. They have broadened their approach to all sorts of datasets, including social media chatter, geolocation, satellite images, weather patterns, credit card payments, IoT sensors, and even health data. AI is helping sift through and extract information from websites, bespoke research from unconventional sources, news articles, and more, to provide intelligence via intuitive trading tools. There are myriad ways in which they use this data.
Making Market Predictions or Company Valuations
News analytics is a powerful resource for investors. From breaking headlines to financial commentary, managers can access an aggregate analysis of the market mood. This helps them accurately predict market behaviour. AI enables them to do this on a global scale, bringing additional benefits over traditional news feeds.
Many years ago, hedge funds would send people to huge retail stores to count the number of customers going in and coming out. This was done to make predictions about a retail company or the entire sector in general. Today, alternative data can enable the same thing, on an entirely different scale and level of sophistication.
- Hundreds of apps, including traffic and weather apps, collect geolocation data of wherever people go. Investors can gain insights on footfall traffic to particular retail chains.
- User-level credit card statements (with anonymity) can reflect detailed purchase histories, indicating consumer spending patterns. This can be useful to predict the sales or churn of a specific company.
- Casual chatter on social media platforms can provide non-obvious insights like trader sentiment about an asset, a company, or the entire economy.
- Incorporating satellite data of construction practices or flood risk analysis can provide insight into whether companies are practicing sustainable methods. Alternative data is the key to discovering whether stocks fulfil the ESG criteria.
- Sentiment analysis of employee reviews on job sites can reveal brewing turmoil in a company before its eventual collapse.
The way data is used depends on a firm’s perspective too. Fundamental firms might use them to make human-driven investment decisions. Quantitative funds will combine various datasets and feed them into sophisticated algorithms to gain insights. A trend of automated trading strategies, based on these models is fast gaining ground.
Fundamentally, hedge funds are trying to leverage alternative data to gain a competitive edge and generate “alpha” by making accurate market predictions. However, alternative data is no longer just an “alpha” generator. Data sets can help in risk management too. For instance, hedge funds are tracking how often ticker symbols are being mentioned on Reddit forums, to predict if a short squeeze is imminent.
Operational Efficiency and Strong Customer Relationships
Apart from investment strategies and portfolio optimisation, hedge funds are relying on alternative data sets to drive operational efficiency, accounting, and investor relations.
- AI-driven data analysis is being used to analyse historical trade break data and predict the causes of current trade breaks.
- Speeding up the reconciliation and NAV production process.
- Offering new investment products to clients based on past investment history, their lifestyles, and future goals.
- Offering clients valuable reports, analysis, and other forms of content, based on past online search behaviour.
The Quality of Data Matters
Unique and original datasets matter the most to hedge funds. What they are looking for is history, level of detail, rarity, and breadth.
For instance, breadth of coverage is not just limited to geographical borders. Quant traders might want data specific to the European investor, but they might also want an extensive number of assets to be covered. If data is scarcer and more interesting, they might be happy with 2 to 5 years of historical data, otherwise they would want to access 10 to 15 years of data.
They also care a lot about the legality of data. This means obtaining data through companies operating compliantly all privacy codes and laws.
The race for alternative data sets is gaining pace. As conversations around hedge fund investments gather ground in 2021, funds will look for more rich data points to drive their businesses higher.
