Unsophisticated investors unnecessarily exposed to short-term market volatility

[Colm Fagan]

“My Future Fund”, the Irish Auto-Enrolment Scheme, starts receiving contributions from January next.
The default investment strategy (to age 51) is to invest in equities. My own pension is still almost entirely in equities (at age 75!). Its market value can fluctuate wildly: it fell by 15.7% in March 2020 and by 12.5% in June 2016; in the two months February and March 2020, its value fell by over a quarter (down 25.5%). I'm (supposedly) an “experienced investor”, yet I find the falls heart-stopping at times. God knows how people with no investing experience will feel - true for the vast majority of AE members.

Now, to introduce an alternative to “My Future Fund” - “My Past Fund”. It operates like an interest-bearing savings account. Between January 2014 and May 2025, the average interest rate on “My Past Fund” was 8.7% a year (compounded monthly); never in the 138 months did the "interest rate" fall below (plus) 4.8% pa. In contrast, monthly returns on my own ARF fund were negative almost 40% of the time.
The progress of “interest rates” on “My Past Fund” is shown in the graph below. In 124 of the 138 months from Jan 2014, the “interest rate” was between +6% and +12% pa.

As you’ve probably guessed, “My Past Fund” is the smoothed version of my own pension fund. Smoothing is per a mechanical formula - a 99% weighting to last month’s smoothed value increased by the expected long-term return and a 1% weighting to current market value. The precise formula, and the progression of actual and smoothed fund (per €1,000 on 1 January 2014, allowing for withdrawals) are shown in the entry dated 3 June 2025 on the Pensions tab of my website, http://colmfagan.ie/ (colmfagan.ie). The spreadsheet shows smoothed value (SV) and market value (MV) crossing 17 times in the period.

I use smoothed values/ returns to keep me sane when markets go crazy, e.g., in March 2020, when the market value fell 15.7%, the smoothed return was an annualised (plus) 4.8% (i.e., +0.4% in the month).
Of course, for me, smoothed returns are only a psychological crutch: they’re not real. However, the light-bulb moment was the realisation that smoothed returns can be made real for AE members. They buy in – always – at SV and they exit – always – at SV. AE means they can’t pile in when SV<MV, nor can they exit in droves when SV>MV.

Smoothed returns on “My Past Fund” show why it’s far better to be in equities than bonds. If my own pension had been invested in bonds (directly or through an annuity), I would have been lucky to get a third of the close to 9%pa I earned over the last 11+ years. AE members could have earned similar returns (or higher, because of lower costs) at deposit account volatility.
Why instead are we exposing unsophisticated investors unnecessarily to full market fluctuations, forgetting that AE is different, that the constraints on contributions and withdrawals make it possible to adopt a different, more inventive, and far better approach?

 

[Dr Strangelove]

Once SV>MV then drawdowns are effectively a cash transfer from one member to another.

This makes AE perpetual - a bit like a sovereign. And it means if there’s a shortfall there is a call on the Exchequer. One of the main reasons for the set-up of AE is to reduce the exposure of the taxpayer to the cost of ageing - not increase it!

AE has many, many flaws but the concept of individual membership with market values is a good one.

 

[Fortune]

My Future Fund”, the Irish Auto-Enrolment Scheme, starts receiving contributions from January next.

I'll believe it when I see it

 

[Fortune]

And it means if there’s a shortfall there is a call on the Exchequer

I'm not sure how that would be the case; the SV would just drop

 

[Colm Fagan]

Once SV>MV then drawdowns are effectively a cash transfer from one member to another.

The transfers are in one direction when SV>MV and in the other direction when SV<MV. Over the long term, it works out fair for everyone - they join and leave at market value (on average, a very important qualifier) but with the big plus that the volatility of short-term fluctuations in market values is completely eliminated. The net result is as shown in the graph, and in the spreadsheet from which the graph is derived, and which I attach to this post.
Of course, it's fair to ask why would people buy in when SV>MV or why would they sell if SV<MV. They do it because it's AE. They wouldn't otherwise. Result: everyone's a winner. They all enjoy the higher long-term equity returns at volatility levels of deposit accounts. It's only possible for AE

 

[Zebedee]

Colm. I appreciate all the work you’ve done on this. It’s a very interesting idea.

I’ve one question about differing levels of cohorts coming to retirement. If for example the first cohort of retirees is very large and SV > MV. Where does the money come from?

 

[Colm Fagan]

Thanks @Zebedee
Firstly, to put your question in context, DSP expects 800,000 joiners at the start (It won't be anything close to that, but I still expect hundreds of thousands of joiners). In the initial years (for the first 30 years, I reckon), contributions from new and existing members will dwarf outgo - gratuities, pensions and death benefits. In those initial (say 30) years, "sales" by retiring/retired members will be more than matched by "purchases" by new and continuing contributors. Transactions between leavers and joiners will all be at SV, which will sometimes be more than, sometimes less than, MV. As you see from my own fund, the smoothing formula is designed to ensure that MV and SV cross over regularly (they did so 17 times in the last 11 years), so transactions take place at MV on average, but without the drama of the massive falls and hikes in MV's from time to time.
The scenario you describe is extremely unlikely before year 30. Given the numbers involved, even a large cohort of retirees in any month is likely to be much less than incoming funds from new and continuing members, bearing in mind also that only 25% of the benefit leaves the fund at retirement; the other 75% is paid gradually throughout the member's retirement.
But cash flows will eventually turn negative. I expect that to happen sometime after year 30. When that time comes, transactions between members and the scheme will be slightly different from what I've shown in the spreadsheet for my own fund.
I assumed that payments to retirees would all come from the fund, e.g., 65.22 outgo in December 2014 per Column D of the spreadsheet attached to post #5 above.
In the real world (assuming the smoothed approach eventually wins the day - unlikely!!), another account, which I call a buffer account, will come into play when cash flows turn negative.
Transfers to the buffer account will start from around year 20, by which time charges on members' accounts are expected to exceed the cost of administering the scheme. The transfer to the buffer account will be the excess of charges over the cost of administering the scheme. The amount of the excess and consequently transfers to the buffer account will increase each year. By the time cash flows turn negative (around year 30), there should be a substantial balance in the buffer account.
When that time comes, exiting members will still receive SV, but the amount withdrawn from the fund will be MV, not SV. If MV>SV, the excess MV-SV will transferred to the buffer account and conversely if SV>MV.
For example, referring to the spreadsheet attached to post #5 above, the outgo in December 2014 is 65.22. If this were the "real" AE scheme, 65.22 (SV) would be paid to the member (me in this case!!) but MV rather than SV would be withdrawn from the fund, i.e., 65.22*1090/1032 = 68.885; 65.22 would be paid to the member; the other 3.665 would be transferred to the buffer account. (I ignored this refinement for my own fund).
If SV>MV, the excess SV-MV will be taken from the buffer account and paid to the member, to bring the total amount paid up to SV. Given that the buffer account is being funded for the previous ten years or so, there should be more than enough in the account to cover any shortfall SV-MV.
The bottom line is that the total value of the AE account at any time equals the MV of all interest of all continuing members (active and retired), nothing more, nothing less, so there's no Ponzi risk: every member of the fund can see that their money is set aside for them in the fund: there is no contamination in the form of money having been "advanced" from the fund to people who've already left.
Sorry if it all sounds complicated. Don't hesitate to come back to me if I haven't explained it well enough.

 

[Zebedee]

Thank you Colm for your quick and detailed reply. I’ll try to digest this and come back to you if I have further questions. Much appreciated.

 

[Colm Fagan]

@Zebedee We're into the honours class at this stage (appropriate, given that it's exam time!).
99.9% of the time, it's fine to extract the actual amount withdrawn (i.e., SV) from the account when there are net leavers - as I did for the example linked to post #5 above. However, in certain extreme situations, this could lead to a death spiral. A death spiral could occur if SV>MV for a prolonged period when net cash flows are negative. This would result in leavers getting more than their fair share, which would impact negatively on NAV for continuing members, increasing still further the gap between MV and SV. Ergo: death spiral. Withdrawing only MV of net withdrawals from the fund prevents this from happening.
As I said, this is an extreme risk. It has no relevance for the example in the attachment to #5, but we must consider the 0.1% risk.