I’m not sure if I’ve mentioned this in the blog, but once upon a time, I almost became an economist. Law school won out over graduate school largely because I was tired of calculus. So, I’ve always had a soft spot for the SEC’s Division of Economic and Risk Analysis – and last month at PLI’s “SEC Speaks” conference, I was reminded that this Division will have a significant role in Chair Atkins’ ambitious rulemaking agenda. Here are seven things I learned about DERA from that event:

1. DERA is a substantial operation. The division has approximately 170 people – economists, data scientists, statisticians, lawyers, and other professionals – with the majority holding PhDs in economics, finance, or related fields. It supports economic policy, examination, enforcement, analysis, data management, and risk analysis. It’s a “high impact” division.

2. DERA wants data for rulemaking. The panel emphasized that comment letters on upcoming rule proposals will be most persuasive if they contain data. DERA uses data from various sources for the economic analyses that inform Staff recommendations and Commission decisions – and data points from companies and other industry participants can reveal information that public filings do not. There’s a whole guide about how the economic analysis works for rulemaking. Ideally, submit data in structured format so that DERA can easily analyze it.

3. The Statistics & Data Visualizations page is a hit. I recently blogged about this page. Launched in August at Chairman Atkins’s direction, the statistics “white page” provides centralized, interactive statistics covering capital formation, market participants, market activity, and investors. Since launch, it has attracted 40,000+ views and 28,000+ users!

4. Structured data makes AI tools work better – but only if the data is clean. DERA’s Office of Disclosure designs “taxonomy” – that’s the list of tags that makes filings machine-readable. With more market participants using AI, it’s important to note that the models are only as good as the data they are trained on. Structured data with standard taxonomy provides the context that makes AI tools effective. The SEC’s own AI task force has worked on projects such as analyzing the tonality of narratives in filings.

5. Scaling errors are still a problem. We’ve blogged about DERA’s data quality reminders on fixing scaling errors, but apparently the problem continues. Other common errors include using outdated tags, creating custom tags for standard disclosures, and discrepancies between different parts of the same filing. While lawyers often don’t have much visibility into the detailed tagging process, you can remind filing agents to use public validation rules to check data quality before submitting a filing.

6. DERA’s data sets are getting more popular. DERA’s Office of Disclosure publishes 15 free datasets on its website that are regularly updated. In the past year, downloads of the “Financial Statements and Notes” data set has doubled – it’s one of the most downloaded items on sec.gov.

7. DERA’s economic analysis is also critical to enforcement actions. The Division helps establish whether material non-public information would cause a stock price change – e.g., through event studies – and identifies suspicious trading patterns – e.g., by figuring out the odds of making successful trades. They gave an interesting illustration of a marble jar that helped me wrap my head around how they find suspicious trades. I’ll spare you the details – but will definitely be passing it along to my 5th grade mathlete.

– Liz Dunshee