An algorithm cannot lie



An algorithm cannot lie; it will tell you what you set it to tell you with the data you fed it. It has no choice, it is entirely deterministic;- and yet algorithms can help you lie.

Let’s take a step back, what is an algorithm?

“a process or set of rules to be followed in calculations or other problem-solving operations, especially by a computer”.

A few years back, I saw a problem.It was not (then) an immediate problem. Even now it is not an immediate problem, but I see it becoming an increasing problem as years pass by.

The problem was that while over a million employers will be required to choose a workplace pension , only a small number will have the independent means or access to an adviser – to choose in an informed way.

The problem was not that there isn’t a way of telling good from bad – or more specifically “suitable from unsuitable”, the problem was (and is) that getting people up to speed with the dynamics of that choice is extremely expensive.

The OFT pointed out firstly that pension plans are extremely complicated machines and secondly that people are poor buyers of these pensions. As the years have gone by, the number of in-house experts, IFAs and pension savvy business advisers has decreased while the number of employers has increased.

We saw this coming when in 2013 we constructed our business plan


The inflection point denoted by the orange marker was the point we thought demand for information would exceed advice. At the point of writing (September 16th) we are seeing a huge number of decisions being taken by employers staging this quarter , next quarter and (worryingly) last quarter.

The algorithm of pension choice

Over the past three years we have seen over 4000 employers choose a workplace pension using an algorithm we built in early 2013 and have refined since our launch that autumn.

We capture data from employers relating to their workforce which tells employers their cashflow liabilities and providers the potential profitability of offering a workplace pension. We capture more data which allows employers to see the attractiveness to them as employers of contracting with each workplace pension provider. Most ambitiously of all, we show employers the likelihood of each provider giving good pension outcomes to staff over their lifetimes.

Our algorithm cannot lie but can the market?

In a perfect market, our algorithm would be managed by our customers. There would be no need for advice as employers would balance the future needs of their staff against the exigencies of running their businesses in an optimal way.

This of course assumes that employers care equally for their staff’s long-term future as for short term profitability.  In a perfect market,we would simply run all employers through our system with minimal risk to ourselves because we were providing a monitor of decision making.

But this isn’t a perfect market. The major market distorter is NEST, the Government’s own scheme, which has spent getting on for £500m of tax-payers money creating a pension scheme which is (currently) free for employers to use and breathtakingly easy to set up and run.

We have set the algorithm within Pension PlayPen to reflect the enormous bargain employers get from surfing all this Government money and getting a free ride from NEST.

We have recently seen NEST’s ratings adjusted downwards by our independent provider analysts at First Actuarial. For the first time, NEST does not warrant a perfect score for the durability of its current offering.

There is real doubt over how the debt that NEST has accumulated will be repaid (other than through price increases to employers). There is concern about NEST’s investment strategy and concern that NEST has a poor range of options for people retiring from it. Our algorithm has been adjusted not by what the market is saying (the market is flocking to NEST), but what First Actuarial is saying (an increasing disquiet with NEST).

Pension PlayPen supports NEST’s ambitions to expand its market presence through aggregation. The link takes you to our submission to NEST’s current consultation. We were first out of the block saying that any expansion plans must include a clear statement on how and when money borrowed from the tax-payer would be repayed.

Or is beta better?

Anyone using can still use NEST as a top rated provider, but it is no longer a no-brainer. First Actuarial have exercised their right to intervene with conviction – breaking market beta and guiding employers towards a variety of provider solutions.

There are many, inside and outside Government, who would prefer NEST to become not just the default provider for the huge numbers of employers still to stage, but a “safe harbour”. Charlotte Clarks’ famous statement that “no employer could be blamed for choosing NEST” is an argument for just that.

Auto-enrolment operates on a default contribution structure, why not a default provider?

Who is running the algorithm of choice?

The Auto-Enrolment regulations are plain – employers choose the workplace pension for their staff. But who runs the algorithm of choice, the Government or the private sector. The market is distorted because the State wishes to be both provider and to run the algorithm of choice.

They can get away with this – and direct choice to NEST – only if there is no credible alternative. Paradoxically, organisations such as Pension PlayPen which intervene and get employers to think about what they are doing when choosing a pension for their staff- are competing not just with NEST but Charlotte Clark’s guidance (itself a primitive algorithm).

In as much as Charlotte is head of the DWP’ private pension strategy unit, we must take this seriously.

I have in the past – semi seriously – suggested that the Government nationalises The serious side of this comment is that it would formalise what Government what to do , which is to see the market behave in an orderly way. But the comic side of my comment is that the exuberance, wit and panache of our site would disappear as soon as it was subject to state intervention.

The Government relies on the private sector to deliver private pensions and the apparatus that supports them. This is precisely why it is not building the pension dashboard.

The Government should not be intervening in pension’s choice. Only £14m of the £460m + owed by NEST to the taxpayer can be written off as a grant to meet NEST’s public service obligation. The rest of the money is an overspend by NEST which will have to be reclaimed.

I find it very worrying that – despite the need for private interventions -such as that described above, the Pensions Regulator still refuses to promote those providing algorithms that help employers choose workplace pensions.

It is time that the Government stopped tinkering with MAF and fit and proper person rules and promoted real choice in the market, allowing its own pension scheme to be put in proper context.

The Government is trying to run the algorithm of choice through statements such as Charlotte’s and through the feeble choose a pension pages of tPR’s website but the are failing the market in doing so.

If is time the DWP and tPR accepted that the arbiters of what’s good  cannot be the providers of NEST.

noisy nest


About henry tapper

Founder of the Pension PlayPen, Director of First Actuarial, partner of Stella, father of Olly . I am the Pension Plowman
This entry was posted in NEST, Pension Freedoms, pensions and tagged , , , , , , . Bookmark the permalink.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s