Managing an ARF for Longevity: A Guide to Retirement Income Strategy
A practical look at how to structure and draw down an Approved Retirement Fund (ARF) in Ireland — what the research actually supports, and where common industry wisdom is more myth than evidence.
The Starting Point: You Don't Get to Choose Whether to Withdraw
Once money moves into an ARF, Revenue imposes a minimum imputed distribution each year, whether or not you actually need the cash:
4% of the fund value per year from the year you turn 61 until the year you turn 70
5% per year from age 71 onward
6% for ARF holders with combined ARF assets over €2 million
This amount is treated as income and taxed accordingly — Income Tax at your marginal rate, USC, and (if under 66) Class S PRSI at 4.2%. Once you hit 66, PRSI generally no longer applies.
This mandatory, non-negotiable annual sale is the single biggest structural difference between managing an ARF and managing a normal investment account — and it's why when the market happens to dip matters more for ARF holders than it does for someone simply accumulating wealth.
The Real Enemy: Sequence-of-Returns Risk
Two retirees can experience the exact same average annual return over 25 years and end up in completely different financial positions, purely because of the order the returns arrived in.
If a downturn hits early in retirement, the retiree is forced to sell a larger number of units to generate the same withdrawal amount — permanently reducing the capital base available to recover when markets eventually rebound. This is sequence-of-returns risk, and unlike market risk generally, it can't be diversified away — it's about timing, not just asset allocation.
For an ARF, this risk is baked in from day one, because the 4–5% withdrawal happens regardless of market conditions.
Getting Invested: Lump Sum vs. Phasing In
A common instinct is to "ease in" — investing a large sum gradually over months rather than all at once. The evidence doesn't support this as the higher-return approach.
Vanguard's landmark 2012 study examined rolling 10-year periods from 1926 across the US, UK, and Australian markets, comparing an immediate lump-sum investment against a 12-month phased entry. The findings:
Lump sum outperformed the phased approach in roughly two-thirds of historical periods
The average outperformance was 1.5–2.4% over the deployment year
Extending the phase-in period to 36 months made lump sum look even stronger — it won in nearly 90% of periods tested
The logic is simple: equity markets rise more often than they fall, so capital sitting on the sidelines is, on average, missing out on gains rather than avoiding losses.
Where phasing in still has a legitimate role: it's a behavioural tool, not a financial one. If a large day-one drop would genuinely cause a client to abandon their plan or panic-sell, a phased entry that keeps them invested and confident may outperform a "textbook optimal" lump sum they can't actually stick with. The research is explicit that this is a psychological trade-off, not a mathematical edge.
Practical takeaway for an ARF: invest the long-term growth portion as a lump sum. There is no evidence base for drip-feeding retirement capital into markets over time.
The Cash Buffer / "Bucket" Strategy — What It Actually Does
The most common structure recommended for decumulation is the bucket approach: segregating funds into short-, medium-, and long-term pots, so that near-term withdrawals are funded from cash and short bonds rather than by selling equities during a downturn.
This is intuitive, and it is the industry-standard explanation given to clients. But the underlying research is more nuanced than the popular version suggests.
The case for buckets
Holding 1–4 years of expected withdrawals in cash/short bonds means a market downturn never forces a sale of depressed equities — the withdrawal is funded from the buffer instead, giving growth assets time to recover.
The case against treating buckets as mathematically superior
Michael Kitces (a widely cited US financial planning researcher) tested this directly, using a retiree starting withdrawals in 1966 — one of the worst historical sequences on record:
A 60/30/10 (equity/bond/cash) bucket strategy with strict "don't sell equities when they're down" liquidation rules was compared against a simple total-return portfolio, rebalanced annually, with withdrawals taken pro-rata across all asset classes.
The two approaches produced identical outcomes at every point in time.
The reason: once a portfolio is rebalanced back to target each year, it makes no difference which "bucket" a withdrawal is nominally drawn from — the ending balance and resulting allocation come out the same either way. Rebalancing alone already ensures that whatever asset class is up gets sold, and whatever asset class is down gets bought — which is a step further than bucket strategies typically go, since buckets avoid selling depressed equities but don't necessarily buy more of them.
The more striking part of the analysis is what happens when the bucket strategy is run without annual rebalancing. Kitces tested this too, and the result wasn't just "slightly worse" — it was dramatically worse. Without rebalancing to replenish the equity sleeve after a downturn, the ongoing inflation-adjusted withdrawals gradually consumed the equity bucket entirely. Once equities were depleted, the retiree was left holding only fixed income for the remainder of retirement, unable to participate in any subsequent market recovery. In other words, a bucket strategy detached from a rebalancing discipline isn't merely suboptimal — it can actively destroy the portfolio's ability to recover from a bad sequence.
(Separately, some industry commentary references academic work by Woerheide and Nanigan suggesting cash buffers can increase portfolio failure rates versus no buffer at all. That claim comes from a different source, not from the Kitces analysis above, and would need to be verified against the original paper before being cited to clients.)
The honest conclusion
A cash buffer's proven value is behavioural, not mathematical — clients who can see a dedicated cash reserve are demonstrably less likely to panic-sell equities during a downturn.
A buffer is not a superior mechanical strategy to disciplined rebalancing — it produces identical results to total-return rebalancing when both are properly rebalanced, and can produce actively worse results if the bucket structure is run without rebalancing, since the equity sleeve can be depleted entirely with no mechanism to buy back in after a downturn.
The risk to avoid: treating the buffer as "safe money set aside forever" rather than a working part of the portfolio that gets replenished (from growth assets, in good years) and drawn down (in bad years) — and critically, ensuring the overall portfolio is still rebalanced back to target allocations regardless of which bucket a withdrawal nominally came from.
Putting It Together: A Framework for an ARF
Size a modest buffer — typically 2–4 years of expected withdrawals — in cash and short-term bonds. This exists to fund the mandatory imputed distribution during downturns, not to sit there indefinitely.
Invest the remainder as a lump sum into a diversified growth allocation appropriate to the client's time horizon, risk tolerance, and other income sources.
Rebalance annually. This is doing more work than most clients realise — it enforces "sell high, buy low" automatically and is arguably more important to long-term outcomes than the buffer itself.
Replenish the buffer opportunistically — top it up from growth assets in strong years; leave it alone and let it absorb the withdrawal in weak years.
Coordinate withdrawal timing with other income — State Pension (from age 66), other pension income, and marginal tax rates. Drawing more than the statutory minimum in lower-income years before the State Pension begins can sometimes be more tax-efficient than deferring everything.
Watch fees. On a multi-decade time horizon, ongoing platform and fund charges compound in exactly the same way returns do — a 1% difference in annual charges is a material drag over 25–30 years.
A Note on the Behavioral Piece
None of the above should be read as "buckets don't matter." They matter enormously for client experience — a visible, dedicated cash reserve is one of the most effective tools for keeping clients invested through a downturn rather than crystallising losses out of fear. The nuance worth carrying into client conversations is simply this: the buffer's job is to protect the plan from the client's own behaviour during a crisis, not to mathematically outperform a well-rebalanced portfolio. Both things can be true, and understanding the distinction helps set the right expectations up front.
This document summarises general research and industry practice for discussion purposes. It does not constitute financial, tax, or investment advice, and should be adapted to each client's individual circumstances, risk profile, and objectives.