When simple isn’t best


There’s a lot to be said for simplicity in an increasingly complex world, especially when it comes to finance.

We often overestimate people’s ability to understand financial calculations, especially as members of final salary schemes. Just think for a moment that four in five adults have a low level of numeracy (Source National Numeracy UK) – that’s 80% of the adult population!

So, it makes sense that information should be accessible and simple. But there are times when simplifying isn’t the right answer.

Pensions Dashboards problem

Part of the current Pensions Dashboards consultation includes a proposal to simplify the basis of how to calculate an Estimated Retirement Income (ERI).

For defined contribution (DC) there’s a logical move to ensure that the results from a DC calculation on a Dashboard are on a common basis with DC schemes. As things stand under the Dashboards project, the calculation differs, which means that you’d get two different figures. Not good for the customer.

The whole premise of Dashboards is to connect people with their pensions; trust in the accuracy of the numbers served up is central to this objective.

It gets harder when you look at calculating defined benefit (DB) ERIs. Unlike with DC, where  there is a statutory requirement to produce an estimate each year, there is no such requirement for DB.

The ability to calculate an ERI, or indeed any estimated pension, depends on data quality and how good the system is at running automated calculations. Smaller schemes often rely on manual processes running on spreadsheets.

The best equipped schemes will have automated most, but often not all calculations. Some are too hard to do and too expensive to automate.

The dilemma for many DB schemes without the ability to automate calculations is how to tackle the problem. Bite the bullet and spend a lot of money to run an ERI for each member now, or wait until a view request hits then turn around a number within 10 days, which is the Dashboards SLA.

Is it right if simple is wrong?

To estimate a retirement income accurately means applying the correct factors to each tranche of benefits. The proposal to simplify the basis of the ERI calculation is to use a broad brush approach, which would simply apply a rate of inflation on the accrued pension.

The thinking is that this simplified calculation will make it easier for DB schemes to generate a number. I can see some problems with that approach.

Firstly, many calculations will fail because the underlying data isn’t good enough. Changing the calculation basis doesn’t impact this.

Secondly, a simplified basis still requires a calculation; it’s not clear if there would be a significant saving in a simplified approach. The DWP admits this is a concern.

And lastly, the ERI will almost certainly be incorrect. You could argue that if it’s calculated in line with the legislation it complies with requirements, so the scheme is off the hook. But I have three issues with this argument.

  1. The simplified value may conflict with a number supplied by the scheme direct to the member.
  2. Given that pension schemes members aren’t typically actuaries (remember, 80% of adults have a low level of numeracy), let’s not expect them to read or understand the nuances of how the number has been calculated. They will just take it at face value and bank on it.
  3. Whilst the ERI on the simplified calculation basis may comply with Dashboards requirements, what will the trustee think about supplying a number that’s not as accurate as it could be?

In conclusion, yes, we want to strive for simple, but if there’s a choice to be made it is better to be correct!

Click below to read our other Pensions Dashboards blogs:

Further reading

Spence & Partners promotes Tom Pook as Chief Commercial Officer and Graham Newman as Chief Actuary

Press Release
by Alan Collins   •  

DB Schemes overspend £300m each year on unnecessary running costs and inefficiencies - Spence DB Scheme Running Costs Report

Press Release
by Alistair Russell-Smith   •  

Scheme funding has improved – now what?

by Graham Newman   •  

More Insights?