How dirty is your pension fund?

Entry type: Single project

Country/area: Netherlands

Publishing organisation: Pointer (KRO-NCRV)

Organisation size: Small

Publication date: 2022-11-30

Language: Dutch

Authors: Thomas de Beus, Isa Martens, Els Engel, Jerry Vermanen, Wendy van der Waal, Els Engel, Sietze van Loosdregt


Datajournalist Thomas de Beus was the driving force behind the data research, data-analysis, concept and text.

Isa Martens was the intern for this project who helped a lot with verifying the dataset.

Designer Wendy van der Waal worked on the concept and design of the interactive and the TV broadcast.

Newsroom developer Els Engel worked on the data and built the interactive.

Jerry Vermanen helpt with the collecting of the portfolio data.

Editor Sietze van Loosdregt led everything in the right direction and reviewd all the text.

Project description:

No meat, no flying, quick showers. Trying to reduce your personal carbon footprint can be discouraging. Besides: while you may be doing your utmost, financial institutions, including pension funds, are still investing billions in the fossil fuel industry. They can make a big difference. Our data team collected the investment portfolios of the largest pension funds in the Netherlands and linked them to a database of fossil companies. This revealed how many fossil euro’s, and in which fossil companies, the pension funds were investing. This resulted in a new dataset and interactive production which exposed the size of harmful investments.

Impact reached:

One of Dutch largest pension funds PME (metal and technology industry) is putting 286 million euro in fossil investments from 29 companies on sale following our investigation. This includes 230 million in coal investments and 56 million in investments in gas and oil extraction. PME stated being “coal-free” back in 2018 and said it would sell all fossil investments by 2021. But after Pointer’s analyses, it turned out there were still fossil fuel companies in their portfolio. Among which one of the biggest coal mine companies like Glencore and BHP Group. “I found it really frustrating, but also very good that you guys found this. We take our climate ambition very seriously,” said PME’s Executive board member Marcel Andringa.

In a broadcast we created, we showed our findings and examined the two investing tactics at play in the pension world: Divest from fossil fuel companies (PME) or continue with engagement (PFZW). We also confronted PFZW about their failure to adhere to their own exclusion criteria when it comes to coal companies. They changed their exclusion text on their website. The broadcast was well watched by Pointer standards with an average of 312,000 viewers.

This is only the tip of the ice berg. This investigation will go on for at least the first quarter of 2023 in which we try to hold other pension funds to account.

Techniques/technologies used:

The hardest part, like most data stories, was collecting the pension funds’ investment portfolios and then neatly linking them to the Urgewald Global Coal Exit List (GCEL) and Global Oil and Gas Exit List (GOGEL). These are recently published public datasets that screen the activities of almost all the world’s coal, oil and gas upstream and midstream companies.

Each pension fund publishes its investment portfolio in different ways and file formats. Most published in very long PDFs that we scraped and had to capture in a spreadsheet. Others published their data in their own dynamically loaded forms on the website. These therefore had to be scraped with Python and Selenium. For a few, the data could be found via the API. Eventually we had a dataset with the columns: pension fund, company, type of investment (bonds, equities etc.) and market value.

Now we have a list of investments in companies from eight pension funds (we also researched three others, but they did not publish their portfolios). The trick is to match the portfolio company names with the company names that also appear in the GCEL and GOGEL lists. This way we reveal all fossil investments and the value it represents.
To do this, we used a “fuzzy matching” technique to link the three data sets. Then it was still a lot of manual work to verify all the matched company names.

R was used for data analysis.

We built an interactive story where people find out what harmful fossil activities their pension fund is (indirectly) investing in.

For a detailed methodology, please refer to the making-of article that accompanies this research. Also included here in the appendices.

Context about the project:

Avoiding the most disastrous consequences of global warming will require major financial institutions to shift their billions in investment and financing to renewable energy. They reached $632 billion per year for 2019-2020. But we need an increase of at least 6 times that number in annual climate finance to get on track for the goals of the Paris Agreement (UNEP – Emission Gap Report).

We found out through this research how old-fashioned the world of pension funds and asset managers is. They invest billions that can make a big impact, but don’t actually know where their money ends up. Quality data on the activities of fossil fuel companies is scarce. We used Urgewald’s public and thus verifiable datasets that for the first time transparently map current plans for expansion for fossil fuel extraction. THis investigation is only scratching the surface. We know that multiple pension funds are working with incoherent or at leat questionable data that favour the fossil fuel industry.

What also surprised us is that such a detailed screening of pension portfolios did not exist before. Several universities informed us that it has not been done like this prior to our efforts.

That it has not been done before by another journalistic title is due, we think, to several factors. First, it’s only the second year that Urgewald published their lists on Oil and Gas companies. Second, that few news rooms in the Netherlands possess the know how to apply the fuzzy matching technique. In addition, it still took a total of three full days with three people to verify the fuzzy matches. And that’s only to verify the matched data set. Let alone all the research, building the story and online experience.

What can other journalists learn from this project?

That in a relatively boring topic like pension fund investment portfolios, an awful lot can be achieved in terms of impact. But also that in a relatively boring topic, like pension fund investment portfolios, an awful lot can be achieved in terms of impact. A lot of money is involved.

And behind every investment is a company that is somewhere, in our case, extracting fossil fuels where people live who suffer and where nature is destroyed.

Ultimately we (in the Netherlands at least) cannot choose where our pension money gets invested in. Employers decide that. Just as you can switch banks, you can’t vote with one’s feet when it comes to pension funds. The only thing you can do as a retiree or someone who works for their pension, is to make your voice heard. But that starts with knowing how your money is being invested. That, in the end, takes journalists who can take the time to unravel and reveal that.

Project links: