Imperatives.

Responsible finance

What "impact investing" measures (and what it cannot)

Impact reports describe what companies do. Whether the investor's money changed anything is a different, harder question.

Published 2 November 2026 · Responsible finance

Every impact fund publishes numbers: tonnes of CO2 avoided, patients treated, loans made to small businesses, homes built. The numbers are usually real. The question an investor should ask is what they prove. An impact report can describe what the companies in a portfolio did very accurately and still say almost nothing about what the investor's money changed. This article explains why, and how to read the figures with that distinction in mind.

What the label promises

The Global Impact Investing Network (GIIN), the industry's main membership body, defines impact investments as investments "made with the intention to generate positive, measurable social or environmental impact", alongside a financial return that can range from below-market to market rate. It lists four core characteristics: intentionality, the use of evidence and impact data in designing investments, the management of impact performance, and contribution to the growth of the field. The term itself took hold after a 2007 meeting convened by the Rockefeller Foundation.

The market is no longer small. In its 2024 sizing study (23 October 2024), the GIIN estimated impact investing assets under management at about USD 1.571 trillion, managed by more than 3,900 organisations. That figure is a sum of what managers report as impact assets; it is not an independent audit of impact, and the GIIN presents it as an estimate.

Two kinds of impact, often confused

Researchers distinguish two questions that impact reports tend to merge.

Florian Kölbel, Timo Busch and colleagues made this distinction the centre of a widely cited review, "Can Sustainable Investing Save the World?" (Organization & Environment, 2020). They found reasonable evidence that investors can influence companies through shareholder engagement, and much weaker evidence that simply buying or excluding listed shares changes a company's cost of capital enough to alter its behaviour. Paul Brest and Kelly Born had made a similar argument earlier in the Stanford Social Innovation Review (2013): an investment creates real impact only when it provides something the company could not otherwise get, such as capital at concessionary terms, patience, or expertise.

This is the single most useful lens for reading an impact claim. Ask of every figure: is it describing the company, or the investor's role?

What the metrics actually capture

Most impact reporting relies on output and outcome indicators. The GIIN's IRIS+ catalogue, the most widely used reference, offers standardised definitions for metrics such as the number of clients reached, jobs supported, energy generated or greenhouse-gas emissions avoided, so that two funds counting "smallholder farmers served" at least mean the same thing.

The Impact Management Project, a forum that ran from 2016 to 2021 and whose work is now carried on by Impact Frontiers, organised the questions into five dimensions that have become a common vocabulary:

  1. What outcome occurs, and how important is it to the people or planet affected?
  2. Who experiences it, and how underserved were they before?
  3. How much: scale, depth and duration.
  4. Contribution: is the outcome likely better than what would have happened anyway?
  5. Risk: how likely is it that the impact turns out different from what was expected?

The first three are where reports are strongest, because they rest on data the company already collects. The fourth, contribution, requires a counterfactual, an estimate of what would have happened without the company or without the investment. Counterfactuals are expensive: randomised evaluations, comparison groups or careful baselines. For most private-market portfolios they do not exist, and reports fall back on outputs. "Ten thousand loans disbursed" is an output; "household income rose relative to similar households without loans" is an outcome; "and it would not have risen without this fund" is contribution. Each step is harder to prove than the one before it.

The fifth dimension, risk, is the least reported. It covers the chance that a project fails, that benefits are smaller than projected, or that the investment causes harm elsewhere, for example by displacing residents or by locking in emissions.

Where measurement breaks down

Regulators have noticed. In the EU, funds disclosing under Article 9 of the Sustainable Finance Disclosure Regulation must have sustainable investment as their objective, and the European Securities and Markets Authority's guidelines on fund names now restrict the use of words such as "impact" and "sustainable" in fund names to funds that meet minimum thresholds, as our article on what the naming rules require explains. A label is still a disclosure regime, not a certificate of impact.

Frameworks that raise the bar

The Operating Principles for Impact Management, launched in 2019 under the leadership of IFC and now administered by a secretariat hosted by the GIIN, are among the most demanding voluntary frameworks. Their nine principles ask managers to set impact objectives, establish and document their own contribution, assess expected impact before investing, monitor negative effects, consider the effect of exits on sustained impact, and review results. Principle 9 requires signatories to disclose their alignment every year and to arrange independent verification of it at regular intervals. Signing the principles does not prove impact; it does commit the manager to a process an outsider can examine.

Questions to ask before believing an impact report

  1. Company or contribution? Which figures describe what investee companies did, and which describe what the fund changed?
  2. Baselines. For avoided emissions and similar claims, what is assumed to be replaced, and is the method published?
  3. Attribution. Does the fund report full company figures or its pro-rata share?
  4. Who benefits. Are the people reached actually underserved, and how does the fund know?
  5. Misses. Does the report show investments that underperformed on impact, and what was done about them?
  6. Exit. What happens to the impact when the fund sells? Was it sold to a buyer committed to the same mission?
  7. Verification. Has anyone independent checked the process or the numbers, and is their report public?

Many of these overlap with the questions in our checklist for reading a responsible-investment policy. The underlying idea is the one behind accounting for stakeholders: measure what changes for people, not only what was spent on them.

The honest version

Impact investing is not a fraud because its contribution is hard to prove, and it is not proven because its outputs are large. The most credible managers say plainly what they can and cannot measure, separate company impact from their own contribution, and accept lower precision rather than inflated numbers. An investor who asks the questions above will quickly see which reports are doing that.

This article is general information, not investment advice. Figures and rules are as published on the dates given; check the current versions before relying on them.

Sources

Related: fund-name rules · responsible-investment policy checklist · Account for stakeholders.

Independence notice. This domain was used in 2020 by the "Imperative 21" campaign, a coalition of business networks. Imperatives is an independent publication, not affiliated with that coalition or its members. We cover the same questions because they are still open. Read more.