Alternative data is big business.  Globally, buy-side firms spent $1.71 billion on it during 2020¹, over half of all hedge funds use it to make investment decisions, and it is increasingly used to validate the sustainability claims of prospective ESG investments. 

In January, the Financial Conduct Authority published a Feedback Statement, Accessing and Using Wholesale Data, following a 2020 Call for Input. The considered firms’ use of alternative data and its potential risks.  Alternative data and its application will inevitably increase, as firms develop systems and technologies that can produce more meaningful, reliable and profitable insights.  In turn, regulatory attention will likely become more focused, and the related risks more clearly identified. 

What is Alternative data?

Alternative data is not easy to define. It is not information from official or corporate sources, like account filings, financial reports or market data.  Typically, it is a form of big data from which a company’s financial health or activity can be assessed.  For instance, satellite imagery could reveal the number of cars parked outside a supermarket and monitor rising or declining customer numbers to predict fluctuating revenues.  Other examples may include mobile phone data, credit card transactions, website traffic, online browsing activity, product reviews, app store analytics, and social media content.²

How is it used and its significance to ESG?

Used with more traditional data sources, alternative data give investors a fuller picture of a prospective investment.  However, typically, the data alone is not enough.  The raw data needs to be processed and analysed, a process which increasingly is performed by AI technologies.  As these technologies improve in sophistication and power, a wider base of data sets could be processed and manipulated to produce more meaningful, valuable and timely insights into businesses. Social media provides a good example.  Whilst it provides an enormous amount of information on the lives and attitudes of consumers, extracting reliable insights is no mean feat.  As well as applying natural language processing techniques, it is also necessary to weed out the junk and any fake information.  Its utility as a data set will be compromised unless one can readily and “efficiently filter the credible from the noise”³. However, little imagination is required to appreciate how valuable alternative data can be if harnessed effectively. 

Another good example of how alternative data can be used concerns supply chains.  Large public companies may have complex and international supply chains.  The use of alternative data can allow investment houses to identify potential trends or future challenges to elements of supply chains.  Having an insight into any single individual supply chain may not present any material advantage, but where one can attain visibility across the whole or significant parts of a company’s supply chain could allow for a significant edge.⁴

As acknowledged by the FCA’s Feedback Statement “alternative data is essential for some managers to understand complex investments such as ESG investments”. For example, investors can access alternative data which uses natural language processing of corporate news flow to provide more in-depth ESG insights than would be available to the broader market. (ESG analytics article)

In part, the utility of alternative data to verify a company’s ESG credentials is borne from a lack of consistent global standards around ESG reporting and concerns around the reliance of ESG scoring.⁵ Companies own ESG performance reports can vary widely in terms of consistency, accuracy and how up to date they are.   To that extent, the use of alternative data may align with the Regulator’s focus on ensuring transparency and accuracy in the ESG investment market, and therefore enhance the protection against greenwashing.⁶ Outside of simply verifying that a company’s ESG credentials comport with their investment mandate and values, investors increasingly see a connection between sustainability, both processes and outcomes, and future profitability.⁷  Big (alternative) data coupled with sophisticated processing are increasingly seen as a means to dig deeper into companies’ actual ESG practices and outputs.

What are the potential regulatory risks?

The FCA’s Call for Input paper captured its provisional concerns about alternative data use.  Its purpose was to “explore whether there are barriers to firms accessing data, or to techniques for analysing data, either of which could act as a driver of harm”⁸ and to “understand whether this dynamic may create information advantages where firms with exclusive access to data or technology can use these to potentially identify market movements ahead of their competitors”.⁹  

Both the Call for Input and Feedback Statement identify several areas of risk presented by alternative data and related technology.  These market risks include barriers to competition¹⁰, as well as privacy and ethical risks. 

Market risks

Many of the associated market risks arise from the increased use of technology, particularly algorithms, to analyse alternative data. These include ‘herding’ when algorithms follow the same market signal, risking flash crashes and inadvertent market abuse when algorithms operate in an unforeseen way.  A central concern is the potential harm to market integrity. The FCA is alive to competition issues, where the unavailability of data leads to information asymmetries¹¹, but could alternative data present a risk of insider dealing?¹²

Under Article 7 of the Market Abuse Regulation, ‘inside information’ is: of a precise nature; not public; relates, directly or indirectly, to an issuer or financial instruments; and, if made public, would likely have a significant effect on the price of the instrument.¹³  The potential classification of alternative data as inside information hinges on whether it is public. Interestingly, some firms have avoided using data sets that are too predictive (i.e. too material) of information normally regarded as inside information, such as quarterly revenue.¹⁴

Factors that indicate information is ‘public’ include where: 1) it is generally available, including through the Internet or some other publication, or where it can be compiled from other information which is generally available; 2) members of the public can obtain the information by observation without infringing rights or obligations of privacy, property or confidentiality.¹⁵  It is irrelevant that the observation or analysis is only achievable by a person with above-average financial resources, expertise or competence.¹⁶  

There is little doubt that alternative data sets are ‘generally available’; the fact that they can only be accessed for a price does not render them non-public.  However, is there a threshold above which the financial and/or technological resources required, whether to buy or process the data set, renders the ‘information’ unavailable in any real or practical sense? The FCA’s Statement acknowledges that cost is a barrier to benefiting from these data sets. 

Privacy and ethical risks

The FCA’s Call for Input paper noted that firms increasingly access alternative data through third-party sources, many of which may be unregulated and/or fall outside the remit of UK data protection laws.  This raises complex privacy risks, for example, in respect of image recognition and user location information, as the Statement notes. Accordingly, regulated entities using alternative data need suitable controls in place, to ensure that their data vendors have sourced the data ethically or in compliance with the General Data Protection Regulation¹⁷, for example, by performing the requisite due diligence. 

Furthermore, in respect of the analytics and algorithms used to process the data, firms must ensure that they have adequate governance and monitoring to protect against biased or unfair outcomes.

The US experience

Enforcement activity in the US suggests that these risks are material.  The SEC has already taken action against App Annie, a company that sells market data on how mobile phone apps are performing (e.g. the number of downloads and the total revenue).  This data was being collected by the company by offering a free service to the apps themselves. The SEC found that, contrary to its representations to its customers, App Annie was using non-aggregated and non-anonymized data from the apps it serviced, in order to make more accurate the market data it was selling.  The SEC order also noted that the company made assurances to its customers that it had processes and internal controls in place to ensure that it was not selling them material non-public information in violation of the federal securities laws.¹⁸   We note however, unlike in the UK, that the definition of non-public in a US context relates to the dissemination of the information, rather than its availability in the UK, “public” means widespread availability, in contrast to the US definition of “widespread dissemination.”

Looking Ahead

The FCA draws few conclusions in the Statement, but plans to keep market developments under review. It has commissioned further research regarding the scale of alternative data usage and data analytics, to better understand the risks and benefits.  Whatever the outcome, these issues are likely to gain traction for two reasons. First, given the reliance on alternative data to assess ESG credentials, its use will rise in tandem with investors’ interest in social and sustainability issues. Second, the FCA increasingly regards itself as a data regulator as much as a financial one, as big data infiltrates all aspects of our lives including financial services.¹⁹  Ultimately, firms will be expected to identify, understand and manage the risks raised by the use of these data sets.

An abridged version of this article was originally published in Thomson Reuters Regulatory Intelligence on 21 March 2022.