Alternative Data and NLP in Quant Strategies › Data Ethics and Compliance

Alternative data use is subject to ethical and legal constraints. Regulations including GDPR, CCPA, and securities laws govern both data handling and its use in trading.

Much alternative data comes from individual behavior (credit card transactions, mobile phone locations, social media posts, and web browsing patterns). Even when anonymized and aggregated, this data can raise privacy concerns.

Compliance considerations:

• Consent and purpose limitation: Ensure data was collected with appropriate consent for investment use. Data collected for one purpose typically cannot be used for another.
• Anonymization adequacy: Verify that aggregated data is truly anonymous. Sparse high-dimensional data like credit card transactions can often be de-anonymized.
• Geographic variation: Privacy regulations differ by jurisdiction. GDPR strictly regulates EU personal data. CCPA grants California residents specific privacy rights.

Reputable vendors provide compliance documentation and legal opinions, but you remain ultimately responsible. You must conduct due diligence on data sourcing and review contracts with legal counsel.

Securities laws prohibit trading on material non-public information (MNPI). Alternative data creates gray areas. For example, when does aggregated private sector data become public? If a fund pays for exclusive access to transaction data, does that constitute MNPI?

Legal consensus generally accepts properly sourced alternative data if it doesn't come from breaches of fiduciary duty or contractual obligations. Data scraped from public websites is generally considered public information. Aggregated and anonymized transaction data, properly licensed, is typically permissible.

However, several edge cases remain legally uncertain:

• Corporate insiders who sell data about their own company
• Employees who leak internal metrics
• Hacked or stolen data that becomes available
• Data obtained through deceptive collection practices

Manage these risks with conservative compliance policies, clear documentation of data sources, and legal review of novel datasets.

The concentration of alternative data advantages among well-resourced funds raises questions of market fairness. If only a few large quantitative funds can afford satellite imagery, credit card data feeds, and NLP infrastructure, does this create an uneven playing field that harms market integrity?

Regulators have generally not acted against alternative data usage; however, increased scrutiny seems likely. The European Union's MiFID II requires evidence that algorithmic trading strategies don't disrupt markets. Future regulations may require transparency about alternative data usage.

The GameStop episode also highlighted how social media analysis can intersect with market manipulation concerns. Monitoring social media for coordinated pump-and-dump schemes has become a compliance priority, and funds using social sentiment must distinguish between legitimate information aggregation and trading on manipulated signals.