ESG Training InstituteVisit Website →
← Back to blog

SFDR: Make Principal Adverse Impacts Audit Ready for Reporting Leads

September 1, 2026
SFDR: Make Principal Adverse Impacts Audit Ready for Reporting Leads

Principal adverse impacts are the negative sustainability effects that Regulation (EU) 2019/2088 requires financial market participants to measure, disclose, and act on. If your organization falls in scope, you must publish an annual PAI statement built on the Annex I template, calculated from four quarterly data snapshots, ahead of the June 30 deadline. Start collecting Q1 data early in the year, not waiting until May.

TL;DR:

  • Mandatory indicators include greenhouse gas emissions, biodiversity impact, gender pay gaps, anti-corruption policies, and exposure to controversial weapons.
  • Firms must average data from four quarterly snapshots over a calendar year and disclose coverage rates and estimation methods transparently.
  • Larger firms must publish entity-level PAIs, while product disclosures relate to individual holdings, both due by June 30 annually.
  • Regulators seek evidence of quantifiable, time-bound actions linked to indicator improvements, not just generic narratives.
  • Building a sustainable, repeatable PAI reporting process requires dedicated data management, governance sign-off, and specialized training.

Table of Contents

What are principal adverse impacts under SFDR?

Principal adverse impacts, or PAIs, are the measurable negative effects that an investment decision has on the environment or on people. SFDR built the concept around double materiality: it is not enough to ask whether climate risk affects your portfolio's returns; you also have to disclose how your portfolio affects the climate, workers, and communities. That second direction, outward from the investment toward the world, is what PAI reporting captures.

The SFDR Delegated Regulation organizes these negative environmental effects and social harms into a fixed architecture, not a menu firms invent themselves. The regulation defines multiple adverse impact indicators split across tables, including a set of mandatory key indicators that every firm considering PAIs must report at entity level. Tables 2 and 3 hold the optional indicators, covering additional environmental and social themes.

The 18 mandatory indicators fall into three practical groups:

  • Climate and environment: greenhouse gas emissions and intensity, exposure to fossil fuel companies, non-renewable energy consumption, energy intensity of high-impact sectors, activities affecting biodiversity-sensitive areas, emissions to water, and hazardous waste ratio.
  • Social and employee matters: violations of UN Global Compact principles and OECD Guidelines, lack of processes to monitor compliance with those principles, unadjusted gender pay gap, board gender diversity, and exposure to controversial weapons.
  • Governance and human rights: anti-corruption and anti-bribery policies, alongside indicators specific to sovereign and real estate investments where relevant.

Opting into Tables 2 and 3 is not a one-time gesture. Once you select an optional indicator, you commit to tracking and disclosing it going forward, and reversing that choice later invites supervisory questions about why the data disappeared. Firms should weigh that long-term operational commitment carefully before adding indicators beyond the minimum one environmental and one social indicator required when you elect to consider PAIs at all.

Who has to disclose, and what belongs in the Annex I statement?

SFDR splits the obligation into two layers, and mixing them up is one of the most common early mistakes.

  1. Entity-level disclosure (Article 4). This is a firm-wide statement covering how the organization considers adverse impacts across all its investment decisions. Firms of larger size, typically those with many employees or parent undertakings of large groups, must publish it. Smaller firms operate under "comply or explain": they can state they do not consider PAIs, but that explanation has to be specific and defensible, not a boilerplate line buried in a disclaimer. The Joint ESAs 2024 report flags vague non-consideration statements as a recurring supervisory concern.
  2. Product-level disclosure (Article 7). Article 8 and Article 9 products must state how each individual financial product considers PAIs, referencing the same indicator set but tied to that specific product's holdings.
  3. Timing and averaging. The disclosure covers a full calendar year and must be published by June 30 of the following year. The underlying figures are not year-end snapshots. They are the average of values observed on March 31, June 30, September 30, and December 31, giving you four data points to average per indicator.
  4. Template sections. The Annex I statement itself has a fixed shape: a summary of the entity's approach, a table of the mandatory and any optional indicators with historical figures, a narrative on policies to identify and prioritize adverse impacts, a description of engagement policies, references to relevant international standards, and a year-over-year comparison once you have prior data to show.

EIOPA maintains editable product templates that map directly onto these fields, which saves you from formatting the layout from scratch.

Calculating PAIs: averaging, project rules, and netting

The four-quarter average is the mechanical heart of PAI reporting, but the details around it trip up even experienced teams. If your custodian reports holdings on trade date while your fiscal year runs on a different cycle, that misalignment can quietly shift which positions get captured in a given quarter, so document your snapshot dates explicitly.

Some indicators apply at the issuer level and others at the project level, and confusing the two produces distorted results.

  • Issuer-level indicators, like carbon emissions and board gender diversity, attach to the company as a whole regardless of which specific asset or bond you hold.
  • Project-level indicators, such as land artificialization or serious accident rates tied to a specific facility, require you to trace the impact to the actual asset financed, not just the parent entity's aggregate footprint.
  • Netting for long and short positions follows specific guidance: you generally cannot simply net a short position against a long one to shrink your reported adverse impact. The ESAs Q&A on the SFDR Delegated Regulation sets out how to present netted exposures so the disclosure reflects actual portfolio risk rather than an artificially flattered number.
  • Coverage and proxies. Not every investee reports every indicator. When data is missing, firms typically substitute estimates or sector proxies, and that substitution has to be disclosed, not smoothed over.

Coverage rates change the answer, not just the confidence level. The Joint ESAs 2024 report notes that small shifts in whether investees report a given data point can materially move portfolio-level PAI figures. Publish the coverage percentage alongside each indicator, and state plainly how you derived any estimate. Data availability also varies by indicator: gender pay gap figures and some emissions metrics tend to have thinner reported coverage than others, so treat a low-coverage number with more caution than a well-populated one.

What regulators actually want to see in your "actions taken"

Supervisors have been consistent on this point across multiple review cycles: the weakest part of most PAI statements is the section on actions taken to address adverse impacts. Naming an indicator and reporting a number is the easy half. Showing what you did about it is where most disclosures fall apart.

The Joint ESAs 2024 report is explicit that regulators want quantified, timebound actions with a visible link between what a firm did and how the indicator moved afterward. A generic sentence about "engaging with investees on climate strategy" does not satisfy that expectation.

Credible evidence tends to look like this:

  • Engagement outcomes reported alongside the specific metric that improved as a result, not as a separate unlinked narrative.
  • Divestment decisions tied explicitly to a PAI threshold breach, with the date and rationale documented.
  • Proxy voting records that show votes cast on resolutions directly connected to environmental or social indicators.
  • Supplier or investee remediation metrics, such as a portfolio company reducing hazardous waste output after a documented engagement campaign.

Pro Tip: Build your historical comparison table before you need it. Recording year-over-year figures from your first reporting cycle onward makes every subsequent statement stronger, because regulators specifically look for trend evidence, not a single static snapshot.

Governance signals matter too. Assign named ownership for PAI data quality, tie at least one internal KPI to indicator improvement, and keep a paper trail showing sign-off before publication.

Common pitfalls and a pre-publication checklist

National competent authorities and the ESAs keep flagging the same handful of failures, year after year. Unexplained zeros where a value should exist, generic policy language copied across multiple products, disclosure coverage so thin it undermines the reported average, and missing methodology notes on how estimates were built all show up repeatedly in supervisory reviews, including the EBA's 2025 report.

Run through this checklist before anything goes live:

  1. Confirm scope: entity-level, product-level, or both, and which threshold rules apply to your firm.
  2. Verify data coverage percentages are calculated and disclosed per indicator, not glossed over.
  3. State your methodology in plain language, including every proxy or estimate used.
  4. Document assumptions with enough detail that an auditor could reconstruct your calculation.
  5. Obtain governance sign-off from a named accountable owner before publication.
  6. Cross-check that the PAI statement's product-level figures match what is referenced in pre-contractual disclosures.

Any statement with unexplained zeros, unsourced numbers, or a missing methodology section needs remediation before it goes public. Supervisors treat those gaps as evidence the underlying data process itself is weak, not just the write-up.

Building the workflow that makes PAI reporting sustainable

A one-off spreadsheet exercise will not survive a second reporting cycle. Assign a named data owner, map every holding to its applicable indicators, and lock in your four quarterly snapshot dates at the start of the year rather than scrambling to reconstruct them in June.

A workable operational sequence looks like this:

  • Standardize data sources first: custodian feeds, index provider data, and direct investee disclosures each cover different indicators with different reliability.
  • Replace missing data points with documented estimates only after confirming no primary source exists, and log the substitution.
  • Run calculations on a fixed internal timetable, then route the draft through governance review before it reaches the statement.
  • Archive each year's data and workings, since the historical comparison requirement means you will need last year's figures again.

Building this pipeline takes technical fluency in double materiality assessment methodology and calculation mechanics that most finance teams have not previously had to develop in-house. Structured training closes that gap faster than learning it through supervisory feedback on a live filing.

Why disclosure quality is a risk management issue, not a compliance checkbox

Weak PAI statements do more than invite supervisory letters. They erode the credibility of the sustainability claims your firm makes everywhere else, from marketing to investor decks, because a regulator or a journalist checking your numbers against your rhetoric will find the gap.

The firms getting this right treat PAI reporting as a stewardship signal, not paperwork. Quantified actions tied to indicator movement demonstrate that engagement policies produce real outcomes, which strengthens the case an asset manager makes to institutional allocators asking harder questions every cycle. That case is only as strong as the data pipeline behind it.

The fastest route from generic boilerplate to a defensible, audit-ready statement is investing in staff who understand the calculation mechanics well enough to catch a coverage gap before a supervisor does.

— Ransford

Build PAI reporting capability with a recognized credential

Turning this playbook into a repeatable process inside your organization takes more than a checklist. Esgtraininginstitute's sustainable finance and ESG reporting accreditation is built for reporting leads, risk officers, and assurance practitioners who need to run PAI data governance, calculation, and audit preparation without relying on outside consultants for every cycle.

Esgtraininginstitute

The program maps directly onto the tasks covered here: mapping holdings to mandatory and optional indicators, documenting coverage and proxy assumptions, and building the governance sign-off trail supervisors expect to see. Graduates managing sustainability disclosures across global jurisdictions have already put this framework to work at scale. Visit the accreditation page to review the curriculum and enroll ahead of your next quarterly snapshot deadline.

Where to go for the primary texts

Skip the secondhand summaries where possible and work from these directly:

Sources