A thesis that became
a transparency tool

The thesis behind the tool

SFDRLens grew out of my master's thesis in Political Economy at the University of Amsterdam, where I am researching how transparency legislation shapes behaviour in the financial sector — specifically whether mandatory disclosure requirements actually produce the transparency they promise, or whether they generate what might be called compliance theatre: technically correct disclosures that are practically useless.

The Sustainable Finance Disclosure Regulation (SFDR) is a particularly interesting case study for this question. It is one of the most ambitious transparency frameworks ever applied to financial markets, requiring thousands of fund managers across the EU to publish detailed data on the environmental and social impacts of their investments every year. The regulation was designed to empower investors — both institutional and retail — to compare funds and hold managers to account.

My thesis examines SFDR through the lens of political economy: what incentives does the regulatory framework create, who benefits from the current state of disclosure, and what does the gap between regulatory intent and market reality tell us about how transparency legislation functions in practice?

What I found — and what was missing

Early in my research, I needed to do something straightforward: compare how different funds were reporting their Principal Adverse Impact (PAI) indicators — the 14 mandatory metrics that SFDR requires all Article 8 and Article 9 funds to disclose annually. These include things like the GHG intensity of portfolio companies, exposure to fossil fuels, board gender diversity, and violations of international human rights standards.

The data is legally public. Every fund manager subject to SFDR is required to publish PAI statements on their website by 30 June each year. Comparing this data across funds should, in theory, be straightforward. In practice, I quickly discovered that it was almost impossible.

The European Supervisory Authorities themselves acknowledge the problem. Their 2025 annual report on PAI disclosures noted that divergent approaches across fund managers make cross-fund comparability structurally difficult — even when funds are technically compliant with the same regulation.

This gap is not just a research inconvenience. It is a substantive finding about how SFDR is functioning in practice — and it became the empirical core of my thesis.

Why I decided to build it myself

I have always been drawn to the intersection of policy and data. My interest in SFDR comes from political economy — the question of how regulatory frameworks shape incentives and behaviour. But I also have a genuine interest in data science, and I believe strongly that the most compelling research is research that produces something useful beyond the thesis itself.

When I realised that no free, comparable, publicly accessible SFDR disclosure database existed, I faced a choice: work around the gap, or close it. I decided to close it — not because I had all the resources a commercial data vendor has, but precisely because I don't. An independent, academically grounded, non-commercial tool occupies a space that vendors cannot: it has no incentive to favour any fund manager, no subscription to protect, and no commercial relationship that might colour its methodology.

Building SFDRLens also means doing what my thesis argues for: not just analysing the transparency gap, but contributing to closing it. The dataset I am assembling — hand-collected from public SFDR PAI statements, normalised using a documented methodology, and published freely — is itself the primary empirical contribution of my research.

What the project involves

SFDRLens combines three things I find genuinely interesting: regulatory analysis, empirical data work, and building tools that make complex information accessible to people who need it.

Where this is heading

A note on independence

SFDRLens has no commercial relationships with any fund manager, data vendor, or financial institution. It is funded entirely by the time I invest in it as part of my thesis research. The methodology is documented openly, the dataset will be available for download, and every analytical choice is explained and auditable.

I welcome challenge, critique, and collaboration. If you think my methodology is wrong, I want to know. If you work at a fund manager and believe your disclosure has been miscoded, please tell me. If you are a regulator, pension fund trustee, journalist, or fellow researcher who wants to use this data, it is yours.

The only thing I ask is that if you use this data in your own work, you cite it and link back — not for credit, but because transparency about data provenance is the whole point.