Anton Kramer: What does the performance test still assess?

February 10, 2026

Pensionfunds Rules and Legislation

_{This column was originally written in Dutch. This is an English translation.}

By Anton Kramer, Co-Founder of OverRendement

The new performance test has been under consultation in recent months. Because participation in industry pension funds is mandatory, legislators understood early on that performance must be monitored through supervision and testing rather than through competition. But anyone who takes a closer look at the proposed test will find it difficult to avoid the question: what are we actually testing?

Pension funds operate in a finely meshed interplay of social partners, management, administrators, regulators and participants with varying risk preferences. It is therefore logical that funds do not want to be judged on factors over which they have little influence. The result, however, is that more and more relevant decisions are being placed outside the scope of the performance test.

We see this, for example, in illiquid investments, where the benchmark return is equated with the fund return. By definition, there is nothing left to assess. In the liquid portfolio, investment policy is shifting increasingly towards passive investing against tailor-made benchmarks. The real choices – which markets, which shares, which risks – are being moved to the strategic level. And that strategic policy falls just outside the scope of the performance test.

An extreme example illustrates this mechanism. Suppose a pension fund decides to hold a share portfolio with ten Amazon shares and ten Volkswagen shares. The tailor-made benchmark then consists of exactly the same shares. In the performance test, we then measure whether the fund has indeed succeeded in buying ten Amazon shares and ten Volkswagen shares. This will usually be the case: test passed. But the much more relevant question – was the choice of Amazon and Volkswagen a wise one, and would NVIDIA or Tesla, for example, have yielded a higher return? – remains completely out of the picture. Of course, no fund literally chooses such a portfolio, but the underlying principle is exactly the same.

The time frame over which the assessment is made is also important. Pension funds invest for the long term, so the average payout is often twenty years in the future. But if it only becomes apparent after twenty years that the fund has performed poorly on a structural basis, the damage has long since been done. A five-year evaluation period has been chosen. At the same time, we want to separate luck from skill. A fund that achieves a 10% return is not necessarily better than a fund with 5%. Call it professional bias, but we want statistical significance.

I advocate a simple test for the five-year period. For example, by comparing the return with a standard benchmark of equities and long-term bonds or swaps, with the weighting matching the risk profile that the fund has chosen based on a survey of participants. In the long term, the essential question remains: has the pension ambition actually been achieved?

Above all, it is crucial for participants' confidence that the performance test is objective and independent. A test that increasingly disregards decisive choices risks missing its target. In that case, we mainly check whether the fund is properly implementing what it has devised itself, but not whether it is actually pursuing a good pension policy.