Willis Towers Watson: Germany needs new approach to longevity data
Critics of the most recent update to Germany’s mortality tables have been proven right by Willis Towers Watson’s head of actuarial consulting in Germany, Hanne Borst.
A general discount applied to mortality assumptions to account for “socio-economic factors” only fits certain companies, Borst said in a presentation at MCC’s annual occupational pension conference Zukunftsmarkt Altersvorsorge in Berlin.
Last year, the updated tables – traditionally issued by the Heubeck consultancy – had resulted in additional liabilities of up to €10bn for Direktzusage pension plans, which are benefits paid directly from company balance sheets. Heubeck later admitted its data was “inconsistent”.
Heubeck introduced the socio-economic factors to take into account the fact that people with a lower income tend to die earlier than those with higher incomes.
Speaking in Berlin earlier this month, Borst said: “The discount only works in companies that have a balanced workforce with both higher as well as lower earners.”
All other companies would have to be aware that the recent update to the mortality tables could lead to assumptions in their liability calculations that did not match the actual payments, she added.
Borst warned that any update to the mortality assumptions based on currently available general data could only be a best estimate: “And best estimate means exactly that – it is not necessarily closer to the truth.”
Heubeck uses data from the German first pillar system, as no data that might be more relevant for the second pillar is available.
However, Borst pointed out that one of the major differences between the general population and people in occupational pension plans was the share of lower earners. In Germany, people covered by an occupational pension plan are more likely to be from the medium to higher-earner section of the population, she said.
Speaking to IPE last year, Georg Thurnes, chief actuary and board member at Aon Hewitt Germany, said: “To apply the reduction to both higher and lower income earners in retirement will only be meaningful when a company’s staff mirrors a representative sample of the population.”
Fixing broken tables
Borst proposed the introduction of individualised longevity calculations for larger companies as a way of bringing more accurate data to the second pillar.
However, she said this was only feasible for companies employing several thousand people.
“But it gives the companies a much clearer calculation of their liabilities – albeit it will most likely lead to increased liabilities,” she said.
Some major German listed companies have gone down this route, but Borst added that there were strict regulations regarding individualised mortality tables.
Evelyn Stoll, head of the occupational pension department at Volkswagen in Germany, said her company had so far rejected the idea of compiling its own mortality tables as it wanted to avoid the ”efforts of follow-up checks”.
European longevity comparisons
Borst also presented a European comparison showing significant differences in longevity assumptions used for calculating pension liabilities.
While Germany currently expects a 65-year-old male to live another 20 years under the latest Heubeck tables, Austria adjusted its mortality tables (AVÖ) last year to a 23-year life expectancy from the same age.
In the Netherlands and Switzerland the estimate at age 65 is just over 22 years, and France expects these males to live for almost 25 years.
Comparing these assumptions to OECD statistics and life-expectancy calculations shows Germany is closest to the international group’s estimate of 18 years.
The widest gap can be seen in the French longevity tables as the OECD only expects a 65-year-old French male to continue to live another 19.4 years.