Normalised WIA inflow: what the 42nd-week sickness notifications tell us

Anyone tracking inflow into the WIA, the Dutch long-term disability scheme, usually looks at benefit awards or at the number of new benefits in UWV's Kwantitatieve informatie reports. Since 2020, advance payments have been included in new benefits, and since late 2022 benefit awards have excluded the simplified assessments for persons aged 60 and over. Both series therefore give a distorted picture. In this article I estimate the inflow in a way that does not depend on how quickly UWV can assess claims: from the 42nd-week sickness notifications.

Why the published series are distorted

Benefit awards understate the inflow

Since 1 October 2022, employees aged 60 or over at the end of the two-year waiting period can receive a simplified claims assessment. In principle this leads to a WGA 80-100 benefit, the WGA benefit for persons assessed as 80 to 100% incapacitated whose incapacity is not considered permanent. UWV counts these assessments as completed claims, but not as awarded IVA or WGA claims. There were 9,332 such assessments in 2023 and 10,281 in 2024. During the first eight months of 2025 the arrangement was largely suspended; it was reintroduced on 1 September 2025. From January to April 2026 there were 4,210.

In addition, the backlog at UWV is delaying decisions. The number of pending WIA assessments rose from 15,646 at the end of 2019 to 30,343 at the end of 2024, 44,144 at the end of 2025 and 45,579 at the end of April 2026.

New benefit counts overstate the inflow

When UWV cannot decide in time, the claimant receives an advance payment. That counts as a new benefit, even if the claim is later rejected or the person recovers in the meantime. Advance payments within new WGA benefits rose from 24,322 in 2024 to 36,903 in 2025. Reopened earlier benefits also count as new benefits.

Both series are therefore driven by UWV's assessment capacity, each in its own way.

Normalised inflow

The 42nd-week notification is the report an employer makes to UWV once an employee has been sick for 42 weeks. Roughly a year later the WIA claim follows, and after the assessment the benefit. The estimate follows that chain with fixed ratios:

normalised inflow(t) = notifications(t − 16 months) × claim ratio × award ratio

• Claim ratio: the number of WIA claims divided by the 42nd-week notifications of 12 months earlier, averaged over claim years 2019 to 2022: 62.8%.
• Award ratio: the share of claims that leads to a decision times the share of decisions that is an award, averaged over 2018 to 2022: 62.9%. Growth in pending assessments counts as decisions still to come, so a growing backlog does not depress the ratio.

The base period predates the 60-plus arrangement. A lag of 16 months gives the best fit, both with new WGA and IVA benefits and with benefit awards: a correlation of year-on-year changes of 0.67 and 0.71 respectively, against at most 0.45 for other lags.

Normalised inflow shows what the WIA inflow would be without advance payments, without the 60-plus arrangement and without the UWV backlog, at the ratios of the base period. It does not predict that all of that inflow will actually lead to a benefit.

How well does the estimate fit?

The best test is a comparison with benefit awards including the simplified assessments for persons aged 60 and over. That series counts every simplified assessment as an award and is therefore an upper bound.

In 2023 and 2024 the estimate fits well. In 2025 awards were 9% lower. That coincides with the months in which the 60-plus arrangement was suspended, and with pending assessments growing by almost 14,000 that year. It is not a fixed pattern: in 2022 the gap was largest, with the smallest growth in pending assessments of 2020 to 2022.

Looking at individual four-month periods, a seasonal pattern emerges. January to April falls below the estimate almost every year: by 6.9% on average in 2020 to 2024, while the other periods are in line with the estimate on average. The estimate uses a single fixed lag and does not capture that pattern. The gap of 8.4% in January to April 2026 is consistent with it.

Where does the rise come from?

The rise in normalised inflow comes entirely from the 42nd-week notifications, since the ratios are fixed. To see whether that is mainly due to more employees, I divided the notifications by the number of employees according to Statistics Netherlands (CBS).

From 2019 to 2025 notifications rose by 50,800. That rise breaks down into three parts: more employees (15%), a different mix by sex (about 1%) and more notifications per employee (about 84%). Higher labour participation by women explains little: the number of female employees grew by 9%, against 5% for men.

UWV changed the age bands for notifications from 2024 onwards. Age bands after 2023 are therefore not fully comparable with earlier years, so the comparison by age runs up to and including 2023.

Notifications per employee rise most among the youngest employees. The group under 25 is small and consists partly of school pupils and students with part-time jobs, so its high percentage carries less weight. By age, too, the mix explains little: from 2019 to 2023, 15% of the rise is due to more employees, about 1% to a different age structure and 84% to more notifications per employee. On balance, the ageing workforce makes almost no difference, because the group aged 45 to 55 shrank while the group aged 55 to 65 grew.

The figures show who is more often on long-term sick leave, not why. The diagnoses in the UWV figures do not settle this: for about half of the new benefits in 2025 the diagnosis is unknown, probably largely because they are advance payments.

Outlook

Because the estimate uses notifications lagged by 16 months, normalised inflow up to mid-2027 can already be calculated from notifications that are known.

For January to April 2027 the estimate ranges from 21,300 with a lag of 20 months to 23,900 with the award ratio of 2023 to 2025, including the 60-plus assessments. If that period falls short by 6.9%, as January to April did on average in 2020 to 2024, awards including 60-plus assessments would come to around 21,500, against 19,835 in January to April 2026. That seasonal average is based on only five years, ranging from −0.6% to −16.8%.

What this means for insurers

Relying on published awards understates inflow; relying on new benefit counts, which include advance payments, overstates it. In my view, normalised inflow is a better yardstick for underwriting, and a benchmark for IBNR reserves, than the published series. For products that follow UWV's decision, the backlog may lengthen the reporting delay, so IBNR based on the historical reporting pattern may be too low.

Depending on the product and cover, further adjustments are then needed:

• Self-insured employers for the WGA (eigenrisicodragers). According to the minister, benefits following a simplified assessment are paid from the general disability fund (Arbeidsongeschiktheidsfonds). According to HR Praktijk, they are not charged to employers, including self-insurers. That lowers their inflow for as long as the arrangement runs. The arrangement is temporary, so an A/E of around 100% says little about the price afterwards.
• WIA excess and WGA gap covers. These generally follow UWV's decision. Under the simplified assessment, persons aged 60 and over are awarded a WGA 80-100 benefit more often than under a full assessment.
• Advance payments. Whether and when a cover pays out during an advance payment, and what happens if the claim is then rejected, differs by product.

Limitations

• The ratios are fixed at the level of 2018 to 2022. If claim behaviour or assessment practice changes structurally, actual inflow will differ.
• Safety-net claimants, such as agency workers and unemployed persons on sickness benefit, are included in WIA claims and benefits but not in the 42nd-week notifications. The claim ratio assumes their share stays stable. In recent years that share was between 45% and 57% of new benefits. It was higher in 2023 and 2024, in 2024 partly because UWV classified persons aged 60 and over as safety-net claimants. The figures per employee also cover employees only.
• The claim ratio is measured with a 12-month lag and applied with a 16-month lag. With rising notifications, that slightly lowers the estimate.
• The 16-month lag is an average. It is based on 18 to 21 observations.
• The test counts every simplified assessment for persons aged 60 and over as an award. The UWV figures show that not every assessment leads to a decision, so actual awards are somewhat lower than in the table.
• CBS counts employees aged 15 to 75. That approximates the number of persons insured under the WIA, but is not the same. A notification is not a person: someone can be notified more than once. The group aged 65 to state pension age has been left out.
• The UWV figures contain no income data. Whether the rise is concentrated among higher earners cannot be determined from them.

Reproducibility