Why is there no concern about the state's unfunded pension liabilities?

[MrEarl]

Public service pensions were reformed in 2013, which significantly reduced future pension liabilities.

There's still a hell of a long way to go, imho.

 

[TomEdison]

The State’s public service pension liabilities are actuarially valued every three years. The most recent valuation I can find is for 31 December 2021. There should be a later valuation as of 31 December 2024. I can’t find that it has been published, but on prior form we can expect it in early 2027.

The 2021 valuation put the aggregate value of pension liabilities accrued to date at €175 billion, assuming the policy of linking pension increases to pay for the grade is continued. (If the policy were changed so that pension increases were linked to CPI, the actuarial value of accrued liabilities would fall to €141.4 billion.)

While the figure is large, it matures over a period of about 70 years. According to the speech the Minister made on publication of the actuarial report, the cost of actually paying public sector pensions was about 1% of GDP in 2022. That was expected to increase to about 1.1% of GPD by 2040, but to fall after that, down to about 0.7% of GDP by 2070. This is the result of “measures taken to mitigate costs” (i.e. the introduction of the Single Scheme) playing out over time.

While you could adopt a policy of funding public sector pension liabilities as they accrue, the initial result (and, by “initial”, I mean “lasting for decades”) would be a significant increase in expenditure of tax revenues on public sector pensions, because the Government would have to continue paying the 1% of GDP that it currently pays, plus it would have to find more money to pay into a pension fund for accruing pension obligations. Eventually, as the funded pension obligations matured and the unfunded ones expired, annual expenditure on pensions out of tax revenue would start to fall. Effectively, you’d be front-loading the pension expenditure - paying for pensions as they accrue, rather than as they fall due for payment. There’s no a priori reason to think that this would cost any less overall, in real terms; all other things being equal it would probably cost more, because you’d have the costs of administering the pension fund.

 

[faketales]

While you could adopt a policy of funding public sector pension liabilities as they accrue

I had a fully funded public pension in Canada. One of the conditions was that if returns were not as expected both the employer (a level of government) and the employee may have to increase their contribution. It was defined benefit but not exactly defined contribution. I would argue you shouldn’t really be able to have both.

Regarding the double paying. I fear that in relation to the state pension that’s what people are being asked to do. Young people still need to pay tax to fund it at current levels. However there is an increasing need to fund a private pension / MFF due to a likely drop in the value of the state pension in future.

 

[JimmyB99]

I had a gawk at that valuation report.

https://assets.gov.ie/static/documents/actuarial-review-of-public-service-occupational-pensions-in-ireland-adc1dbbf-f8de-43a0.pdf

There’s no a priori reason to think that this would cost any less overall, in real terms; all other things being equal it would probably cost more, because you’d have the costs of administering the pension fund.

I'd be interested in the views of @Colm Fagan and @Duke of Marmalade of Marmalade here! Bottom line is that it looks to me that GDP is assumed to grow at c. 4% p.a. and, as the report itself says an equity risk premium of 3% p.a. to 4% p.a is a reasonable assumption, it would seem to me that prefunding some of the liabilities exclusively via equities has a greater chance of being cost reducing (versus being cost neutral or cost increasing) in the long-term.

 

[TomEdison]

Regarding the double paying. I fear that in relation to the state pension that’s what people are being asked to do. Young people still need to pay tax to fund it at current levels. However there is an increasing need to fund a private pension / MFF due to a likely drop in the value of the state pension in future.

That's not a bug, it's a feature. And it always has been.

The social insurance retirement pension in Ireland is about 24% of average earnings. (Since most people's earnings tend to rise over the course of their career it's less than 24% of average earnings for a person of retirement age.) The result is that for most people the social insurance pension will not, on its own, provide an adequate income in retirement; it will need to be supplemented with a second-tier, funded occupational or personal pension. This isn't new.

This is much more the case in Ireland than in other countries. Social security pension as a percentage of national average earnings in . . .

Canada: 37.1%
US: 39.7%
Germany: 42.1%
France: 56.6%
Spain: 65%
Portugal: 75.5%

There's another slew of countries that have low social security pensions, but mandatory second tier pensions (usually funded). These include:

Denmark: 31.6% plus 43.1%
Netherlands: 28.7% plus 46.0%
UK: 28.7% plus 17%

I suppose we're technically now in the second group, since we have introduced a mandatory second-tier pension in the form of MFF. I don't know of any modelling about what an MFF pension will look like, as a percentage of national average earings, for someone who has a full career of contributing to MFF, but that won't be a real world question until about 40 years have passed.

So as you can see we're a bit of an outlier here - strikingly low sociall insurance pension, and (until now) no mandatory second-tier pension. The consequences of this include:

- We pay much lower social insurance contributions + mandatory second-tier contributions than in most other developed countries

- The "looming demographic crisis" (due to rising life expectancy and falling birthrates driving up the cost of providing social security pensions) looms somewhat less for us than it does for many other developed countries

- We have a greater need than most other countries for voluntary second-tier pensions to supplement the social security pension. As of 2025, 67% of the workforce has second-tier pension coverage.

 

[TomEdison]

Bottom line is that it looks to me that GDP is assumed to grow at c. 4% p.a. and, as the report itself says an equity risk premium of 3% p.a. to 4% p.a is a reasonable assumption, it would seem to me that prefunding some of the liabilities exclusively via equities has a greater chance of being cost reducing (versus being cost neutral or cost increasing) in the long-term.

I'm way out of my comfort zone here, but I think the equity risk premium is not the extent to which return on equities will outpace GDP growth; it's the extent to which the return on equities will outpace the return on government bonds.

A funded pension scheme won't be 100% invested in equities; it'll be in a pool of different asset classes (equities; bonds; property; etc) judged by almost infinitely wise actuaries to be appropriate to the liability profile of the fund membership. Let's call that a mixed fund.

Will pensions provided by a scheme that taxes the country to pay contributions to such a mixed fund cost taxpayers more or less, in real terms, than pensions provided by a scheme that simply taxes the country to pay the pensions as they fall due? I think the answer depends on whether the real net return on that mixed fund will be higher or lower than real GDP growth (treating GDP as a measure of the country's tax capacity). And we can't model that without knowing the asset allocation of the hypothetical mixed fund.

 

[joe sod]

Well that puts to bed all this talk about means testing the state pension. If the state even after paying todays pensioners and other liabilities out of prsi receipts still has a 6.2 billion surplus and then uses that money for current spending and other stuff rather than saving it in a fund for future pension liabilities. They don't have a leg to stand on if a hard left party attempted to means test it after they spent the prsi receipts on other spending like bailing out RTE etc.

 

[ClubMan]

Great posts @TomEdison - they certainly deal with some of my confusion on the matter. Thanks.

 

[JimmyB99]

it's the extent to which the return on equities will outpace the return on government bonds.

In between meetings, so very briefly - the above is a reasonable definition but not inconsistent with my point! I never said that the ERP was a function of GDP growth although I can see how my shorthand could have given that impression! [Indeed, there may well be an inverse/bizarre relationship between GDP growth and the risk free return.] I also didn't get into the options cited for the base risk free return in that report or the fact that the ERP quoted is well and truly at the conservative end of the range.

Also, again, I never said that all the liabilities would all be in an all equity fund - that would be financially and politically too volatile and not even practical.

prefunding some of the liabilities exclusively via equities

The actuaries quoted in my post have done extensive work in this area - i.e. the merits of equity investment in pension policy in the long haul. For example, @Colm Fagan made massive public effort to demonstrate these merits in relation to mandatory DC scheme.

So, I'd be interested to hear whether:
(a) he believes investing funds is likely to be cost reducing (i.e. a specific comment on your initial specific claim); and
(b) he would be tempted to put a paper together??!! (my back of the envelope calcs/guesses are that the potential savings are very significant indeed- think investing €1billion, increasing at 4% p.a. getting a net return of, say, 6% versus the assumed GDP return of 4% for 20, 30, 40 years!!)

 

[Duke of Marmalade]

I'd be interested in the views of @Colm Fagan and @Duke of Marmalade of Marmalade here!

First thing to note is that the SIF and the ADL are different animals. Possibly the SIF is misnamed for it is a PayAsYouGo vehicle for social insurance of which the State Pension is the most notable. It swings from positive to negative cashflow with the economic cycle and with unemployment being currently low it is in surplus.
The Accrued to Date Liability of Public Service pensions is a different thing entirely. In line with many (most?) developed countries Public Sector pensions are by and large also on a PAYG basis. Let us consider the macro alternatives of instead adopting the approach to micro pension funding which we are all familiar with.
So we have an ADL of €175bn or 44% of GDP. Where do we invest it? What about Government Bonds? Well if they are Irish GBs that would just be a three card trick.

What about equities? There isn't that capacity in Ireland unless we start to buy up swathes of the economy- i.e. Socialism.

Foreign equities? Whilst Simon plots an escape valve for Irish punters to invest in foreign ETFs, to do so on a national scale would be massively recessionary on pure Keynesian arguments.

So where can we invest that ADL? How about on infrastructure and education etc. which will form the basis for the future taxation needed to pay the liabilities. That's what we do. We could hire a team of accountants/actuaries/economists etc. to put a value on education and infrastructure assets but what's the point? The fact is there is another huge difference between the national pension dynamic and the individual one - in practice the national situation is almost a steady (or increasing) state, a moving target. The individual dynamic switches from accumulation to decumulation on a predefined timescale.

 

[Protocol]

While you could adopt a policy of funding public sector pension liabilities as they accrue, the initial result (and, by “initial”, I mean “lasting for decades”) would be a significant increase in expenditure of tax revenues on public sector pensions, because the Government would have to continue paying the 1% of GDP that it currently pays, plus it would have to find more money to pay into a pension fund for accruing pension obligations. Eventually, as the funded pension obligations matured and the unfunded ones expired, annual expenditure on pensions out of tax revenue would start to fall. Effectively, you’d be front-loading the pension expenditure - paying for pensions as they accrue, rather than as they fall due for payment. There’s no a priori reason to think that this would cost any less overall, in real terms; all other things being equal it would probably cost more, because you’d have the costs of administering the pension fund.

Thanks for this.

You make a good argument about not moving towards funded PS pensions.

Can I ask you: would it not cost less overall due to the investment growth on the possible public service DC funds?


Secondly, can I ask you, if there is no good reason to pre-fund PS pensions, does the same argument apply to State social insurance pensions?

 

[jpd]

Can't see GDP growing consistently at 4% - no country has ever managed that for any meaningful length of time

 

[Protocol]

Well that puts to bed all this talk about means testing the state pension.

The social welfare system in Ireland has two parallel schemes.

The SI schemes are funded by PRSI conts into the SIF.

SOCIAL INSURANCE (not means-tested) SOCIAL ASSISTANCE (means-tested) State Pension contributory Non-con pension Illness benefit Invalidity pension Disability Allowance JSB JSA Carers Benefit Carers allowance Maternity and paternity benefit

When I hear people suggesting to means-test the State Pension, I assume they mean to scrap the SPC and have only the non-con pension?

This would mean a large scale retrenchment of social insurance, and a large reduction in PRSI rates.

I can't see that happening.

 

[Fortune]

I think the answer depends on whether the real net return on that mixed fund will be higher or lower than real GDP growth

It will also depend on the ability of the government of the time (under fund and no-fund scenarios) to raise sufficient taxes to pay the pensions. I suppose demographics will play a part in that too

 

[Itchy]

the actuarial gap would have grown in the meantime.

it has very little to do with the annual receipts and payments

As @Duke of Marmalade point out the ADL and SIF are two entirely different animals. The C&AG periodically reviews the monitoring of the SIF and its actuarial review https://www.audit.gov.ie/media/qaelo00r/17-actuarial-review-of-the-social-insurance-fund.pdf

So where can we invest that ADL? How about in infrastructure and education etc. which will form the basis for the future taxation needed to pay the liabilities. That's what we do. We could hire a team of accountants/actuaries etc. to put a value on education and infrastructure assets but what's the point? The fact is there is another huge difference between the national pension dynamic and the individual one - in practice the national situation is almost a steady (or increasing) state, a moving target. The individual dynamic switches from accumulation to decumulation on a predefined timescale.

The discussions about the management of the states liabilities are largely misdirected when contributors conceptualise the management of their own personal finances and view the states actions through that lens. A point well made by here. We have the ISIF as a contributor to addressing the liability (but it is by no means the sole measure), the wider economic actions of the state and its performance need to be seen in totality. Additionally, the valuations are highly sensitive tot he assumptions, some of which are political i.e. the State are in control of them!

 

[TomEdison]

Remember, all pensions systems are mechanisms from transferring wealth from workers, who produce it, to retirees, who consume it. The difference between a funded and an unfunded pension scheme is smaller than you might think. Even with a funded pension scheme, the fund largely invests in equities and government bonds. And what are they? Government bonds are a claim on future tax receipts; as interest and/or redemption proceeds fall due governments tax businesses and/or employees to pay the amounts required. And equities, essentially, are the right to claim a share of the future profits that businesses and employees will generate. The government isn’t involved in the transfer of funds from businesses/workers to retirees, but the transfer still has to happen.

So the funded versus unfunded pension scheme argument is basically a debate as to which of two mechanisms is the better way to achieve the same transfer. “Better” could mean the lowest cost way, but there are other characteristics of a pension system that we might also value. Both systems have their pros and cons; the unfunded mechanism is undoubtedly cheaper to operate - the overheads are very low, compared to the typical overheads of funded schemes - but it is subject to political risk. The funded mechanism is better at recognising costs as they accrue, but can deliver uncertain results, and forces us to a choice about whether investment and/or mortality risks will be born by employers or by retirees.

We actually use both systems - the social insurance scheme is unfunded while occupational/pension schemes are funded and there’s a view that this hybrid approach is actually the wisest, because diversity in funding arrangements makes for a more resilient system overall.

 

[kickstart]

I feel like I've sat in on a particularly interesting economics lesson today.

 

[Duke of Marmalade]

Government bonds are a claim on future tax receipts; as interest and/or redemption proceeds fall due governments tax businesses and/or employees to pay the amounts required. And equities, essentially, are the right to claim a share of the future profits that businesses and employees will generate.

Great point! I love these macro insights.

while occupational/pension schemes are funded

In the beginning we had benevolent pensions from the likes of Guinness. Whilst apparently benevolent they were also a big attraction to prospective employees and presumably Guinness could pay lesser salaries. Thus more a case of remuneration deferral than benevolence.
These days it is only the State who have the covenant to follow this approach.
A whole tax based scaffolding has been erected to support this remuneration deferral in the private sector and central to this is the separate security of a fund.

 

[TomEdison]

Yes. But in the private sector the insolvency or winding up of the employer is always a possiblility, so private sector pensions have to be funded - the employer has to set aside funds to pay the pensions that have been promised. That way they'll still be paid even if the employer is no longer around.

This isn't an issue when the person promising the pension is the State.

Whilst apparently benevolent they were also a big attraction to prospective employees and presumably Guinness could pay lesser salaries.

Though they didn't. Guiness salaries were well above the average even without the pension scheme.

 

[jpd]

It's wrong to say

This isn't an issue when the person promising the pension is the State

Not quite right - States can renege on their promises but it is rare enough, but not unknown