Provinzial regains P&C profitability after Bernd's wreckage in 2021

25 May 2023

Provinzial Konzern, which stomached one of the heaviest claims bills of €1.6bn ($1.7bn) from storm Bernd in 2021, welcomed its combined ratio improving 28 points last year to a profitable 98.5%.

Germany's second-largest public insurer said in full-year results revealed this week that non-life claims costs had fallen 25% compared with 2021, and notwithstanding 2022 winter storms and the spring storm Emmelinde costing it about €400m, plus "significant impacts" from inflation and higher building costs.

Property and casualty premiums exceeded €4bn for the first time, up €171m from 2021, and chair of Provinzial's board Wolfgang Breuer heralded exceeding Germany's market-wide expansion for the eighth straight year.

Although unit-linked life policy sales grew in tough investment markets, more broadly the underwriter noted a "significant slowdown, market-wide" in single-premium life policy sales.

It was a trend Provinzial could not avoid, and which it attributed to "the rising interest rates [leading to] notably lesser demand".

Overall the group's life premiums fell 17.8%, year on year, to €2.2bn, and Provinzial said "no reversal of the trend was visible" early in 2023 "due to the general condition of markets".

"We want to convince our customers that life insurance is a reliable and solid instrument for [saving] for the future. It alone can provide for an encompassing cover of existential risks – in particular in uncertain and inflationary times," Breuer said.

Provinzial is awaiting BaFin's consideration of fusing its Provinzial Rheinland Lebensversicherung and Provinzial NordWest Lebensversicherung life units - it hopes in 2024 - to drive efficiencies.

The insurer also noted applying artificial intelligence to classify text documents, "to optimise workflows and automate processes," and "intensively pursuing various application possibilities of technology around ChatGPT". The insurer is examining how large language models can be applied to recognise and segment text and images, for example.