At the center of the growing digital economy is data. Data is to the 21st century what oil was to the 20th century. In every industry, it are the companies that can use data effectively that succeed. And investing is no different.
In their search for alpha generating ideas, investment managers are increasingly turning to sources of alternative financial data. But what is alternative data and how does it give fund managers an edge?
- How fund managers generate alpha
- What is alterative financial data?
- An early example of alterative data
- Types of alterative data
- Is alternative data legal?
- Analyzing alternative data
- Investing with alternative financial data
- Alternative data providers
How fund managers generate alpha
The returns generated by investors can be classified as either alpha or beta. If a fund earns the same return as a benchmark index, it has earned beta. Passive investment funds like exchange traded funds typically aim to earn the beta for a market, sector or industry by simply tracking an index. If a fund outperforms a benchmark index, it is earning alpha. This is the excess return above the benchmark.
Actively managed funds and hedge funds typically aim to earn alpha and charge higher fees for doing so. But, beating a benchmark is not easy, and fund managers need an edge to do so. Buying a stock that is a good investment often isn’t enough, because the stock has to actually outperform the market. One of the best ways to pick stocks that outperform is to be right when others are wrong.
What is alterative financial data?
Knowing more than the rest of the market is one of the most reliable ways to be right when others are wrong. Traditionally, investors relied on economic indicators, company filings and price data. But this data is widely available and so offers no information edge. Alternative financial data includes any data set that gives investors unique knowledge about an economy, company, industry, or sector. Typically, this type of data is not widely available, so it gives investors an information advantage.
An important characteristic of alterative data is the fact that it can provide insights long before traditional sources of data become available. Companies typically report their financial results on a quarterly basis, and only weeks after each quarter ends. Official economic data is published quarterly, monthly, or weekly – but also sometimes after the period being measured. By contrast alterative data can be gathered and processed on an hourly or daily basis, and sometimes in real time.
Any data that can help a fund manager generate alpha is valuable. A 2020 study found that 53% of hedge funds were using some form of alternative financial data. Fund managers now spend an estimated $1.7 billion a year on alternative data. Alternative data is also used by startup investors. Venture capital and private equity funds often invest in companies that have little revenue and no profits. Alternative financial data helps them estimate market size and potential demand for a product.
Alternative data is also used elsewhere in the financial services industry. Insurance companies and banks, for example, both use alternative data to help them identify fraud.
An early example of alterative data
Long before social media and location data was widely available, investment analysts needed to be creative. Analysts would spend their weekends driving from one shopping mall to the next estimating the occupancy of parking lots. Later they paid people to count the number of consumers entering individual stores. Or they would estimate hotel occupancy by counting the number of hotel rooms with lights on at night. These crude methods could give analysts an idea of how well a company was performing long before earnings and revenue data were published.
Types of alterative data
Over the past two decades the amount of data being generated has grown exponentially. And, unlike earlier examples most of this data is gathered and stored automatically. In the future, even more data will be generated. Two of the investment megatrends we identified are 5G technology and the internet of things.
Both technologies will result in more connected devices generating even more data. The big data industry has emerged to assist with gathering, sorting and analyzing these massive data sets. These are some of the alternative financial data sources that can be useful to fund managers.
There are currently over 2,000 operational satellites orbiting earth. Many of these satellites are continually broadcasting images back to earth. This means activity anywhere in the world can be monitored on a day-to-day basis. Short sellers often use satellite images to uncover fraud. By analyzing aerial photos of factories, mines, and ports they can a see if the actual level of activity corresponds to the numbers being published by the company.
Energy analysts use satellite images in several ways. They can monitor the number and activity of oil rigs in parts of the ocean. This information can be used to estimate production levels as well as when new oil and gas wells come online. Analysts can also track the movement of oil tankers to estimate oil production and demand in different regions. Various methods can be used to determine how full oil tanks are too.
Similar insights can be gained by tracking the location and movements of cargo ships. By comparing this information to images from previous periods, a country’s imports and exports can be estimated. Freight movements can also be monitored by tracking cargo ports and railway lines. This data is often used to estimate inventory and production levels for commodities like coal, iron ore and copper.
The agricultural commodities market is another area where satellite photos are useful. Analysts use them to estimate crop size and the timing of harvests. They can also be used to estimate the extent of damage caused by storms. Satellite images can also be used in more obscure and creative ways. In early 2020, satellite images showed that pollution levels had fallen in parts of China. This indicated a massive drop-in industrial activity, which was only reflected in official GDP numbers months later.
Location data is gathered by apps on smartphones as well as by other connected devices. App owners often sell this data, but only after personal information has been removed. This process is known as anonymizing data. With the right software, CCTV camera footage can also be used to count vehicle and foot traffic. Analysts can use this data to monitor the number of people visiting shopping districts and malls – and even individual stores and restaurants.
The advantage of location data is that it is automatically gathered in real time. Large amounts of data can also be analyzed with artificial intelligence software to find hidden patterns that are statistically significant. Some hedge funds have managed to monitor the location and flight paths of corporate jets. They use this to work out what M&A deals may occur in the near future.
Web scraping, which refers to gathering information from websites, has been happening for a long time. But the amount of data, much of it user generated has exploded in the last decade. Social media sites like Facebook, Twitter and Reddit are a wealth of information generated by the users themselves. Review sites like TripAdvisor, Yelp and Zomata can also be mined to find out what consumers think of various products. The same applies to the reviews posted on sites like Amazon.
Information from these sites can be used in various ways. The number of posts or reviews referring to a product or service can be used to gauge sales and brand awareness. Natural language processing software is also getting better at decoding content and converting it into useful metrics. User generated data can be used in other ways too. For example, when a company publishes a press release, the language used can be compared to the reaction of consumers or investors.
Other data sources
Lots of other unique data exists that can give investors an edge. Here are just a few examples:
- Email receipts from ecommerce stores can be obtained from some email service providers. Before the receipts are sold to data providers, all personal information is removed, but details of the transaction including product SKU codes remain. This data can be used to determine sales trends for retailers and manufacturers.
- Certain credit and debit card transactions can be obtained provided they are anonymized. Useful data includes the value and volume of transactions, as well as their location.
- Download data for iOS and Android apps is widely available. In some cases, usage data can also be obtained or estimated. This can give investors an idea of how consumers spend their time.
- Government agencies, city councils and municipalities are often obligated to make certain records available to the public. Records like housing permits, vehicle registrations and company liquidations can all be of value when combined with other data.
- New auto insurance contracts often need to be registered. This data can be used to track vehicle sales.
Is alternative data legal?
Some information that could be considered alterative data is illegal to act upon. Company insiders, as well as service providers like auditors and lawyers, often have information about a company that would give them an unfair advantage. This is inside information and trading with this information is insider trading, which is of course illegal. For this reason, insiders are not allowed to trade a stock during sensitive periods.
Some categories of alternative data fall into a grey area. The issue is not so much the advantage it gives traders as the way it is obtained. Even if data like email receipts or location data is anonymized, some people believe obtaining the data is still an invasion of privacy. There have already been a few class action lawsuits over the issue.
Whether or not alternative data gives investors an unfair advantage is difficult to define, prove or regulate. So legal challenges are likely to concentrate on the privacy issue. Given that large tech firms are under a lot of scrutiny, it is possible that some types of data being used today will not be available in the future.
Analyzing alternative data
Before new data can be incorporated into the investment process, it needs to be turned into insights. Data is first “cleaned” and structured so that it can be analyzed, either manually or by algorithms. The type of alternative data analysis needed usually depends on how granular the data is. If there are just a few data points, they can be studied manually.
For example, an analyst might use satellite images of a factory to verify or disprove a company’s claims about its growth or production. A large number of trucks coming and going at a factory would suggest business is good. Satellite photos might also show the factory’s facilities being expanded to accommodate growing demand. However, if satellite images suggest little activity at a factory when the company CEO is pitching a good story, the market might be too bullish on the stock.
The above example only has a few data points. More often, large quantities of data are generated. Large data sets can be analyzed systematically to find patterns and relationships between the data and the stock price. This can often be done in real time too. Once the data has been organized, it will often be combined with traditional data and stock prices, and then a machine leaning algorithm will be used to identify predictable patterns.
Lehner Investments uses this approach to identify trading opportunities using sentiment analysis and other data for the Data Intelligence Funds. Data is first gathered from financial news and social media websites. Natural language processing and artificial intelligence is then used to give securities a big data score. These scores are then combined with other data to find tradable patterns. You can read more about this approach to combining big data and artificial intelligence here.
While a lot of this type of analysis is done manually, software is quickly taking over. Artificial intelligence software is being developed to find relationships between what is happening on the ground and how market prices react later.
Investing with alternative financial data
There is a lot of room for creativity when it comes to incorporating alternative financial data in the investment process. Certain data sets tend to be suited to different investing strategies. Data that reflects market sentiment can assist with market timing. This type of data can also be combined with other strategies that exploit behavioral finance insights. Qualitative data can be used to augment factor investing if it can be converted into useful metrics.
ESG investing is also becoming an important niche in the portfolio management industry. There are numerous data sets that can be used to construct ESG scores. When it comes to stock picking, alternative financial data can also be used to confirm conclusions reached through fundamental analysis.
And, valuation metrics can be tested against any data that points to revenue trends. Using alternative data is typically more appropriate for active investing and algorithmic trading strategies. However as quantitative investing evolves, it may also be used to create passive investing strategies.
Nothing in the world of investing is certain. This applies to using alternative data too. There are occasions where new data can be misleading for one reason or another. So, it always needs to be cross referenced and confirmed. No investment strategy is immune from market volatility. Thus, portfolio hedging, asset allocation and other techniques should be used to manage and reduce risk.
Alternative data providers
The alternative financial data industry is growing rapidly and there are now as many as 450 data providers selling unique data sets. Companies like Quandl and Dawex act as marketplaces for alternative data. Data is more valuable if very few people have access to it. So, unfortunately for retail investors it is usually priced so that its reach is limited. However, some data sets can be replicated, and ultimately market forces will result in that data becoming affordable to the average investor.
Conclusion: Using alternative financial data for investing
As more data is generated, alternative financial data will no doubt become more widely used. Ultimately most investment funds will need to use this data to remain relevant. But, in the future, the way data is used will be just as important as the data itself. So, it’s likely that the use of artificial intelligence will be just as important as data to the investment industry.