How much is enough? Building a number you can defend

Why the popular shortcuts fail

Ask how much you need to retire and you will get a rule: a multiple of final salary, a percentage of pre-retirement income, a round figure that sounds serious. Each of these is a compression of a real calculation, and the compression discards exactly the information that makes your situation yours.

**Salary multiples fail** because they anchor on income rather than spending. Two people earning the same amount can have wildly different retirement needs if one saves 40% of income and the other saves 5%. The saver's post-retirement spending is already much lower and — critically — the saving itself stops at retirement, so their target is smaller on both counts.

**Replacement-rate rules fail** in the other direction. A rule that says you need a given percentage of pre-retirement income implicitly assumes a spending pattern, a housing situation and a state or employer pension backdrop that may not describe you. International pension analysis does use replacement rates, but as a comparative measure across systems, not as a personal target you can lift directly.Source^source

**Round numbers fail** because they are anchors dressed as analysis. A figure that sounds like enough is a feeling.

The alternative is not more complicated, it is just personal: start from what you spend, adjust for what changes, subtract what other sources will provide, and convert the remainder into a capital requirement — as a range, not a point.

Step one: find the honest baseline

Everything downstream depends on this number, and almost everyone gets it wrong on the first attempt, in the same direction: too low.

The reliable method is to work from outflows rather than budgets. Take twelve months of actual bank and card activity. Total money out. Subtract transfers to savings and investments, and subtract debt principal repayments that will be finished before retirement. What remains is your real annual spending, including the irregular items that a monthly budget quietly omits — the annual insurance premium, the flights, the school fee, the dental work, the replaced laptop, the wedding gift.

Twelve months is the minimum because annual items are the whole problem. A three-month sample will understate a real year by a meaningful margin, and understating the baseline understates the target proportionally at every later step.

Two notes on doing this cleanly.

• **Separate one-off distortions from irregular normality.** A house move is a distortion. A holiday is irregular normality. Strip the first, keep the second.
• **Do not sanitise.** The purpose is not a virtuous budget. It is a forecast. If you spend on things you would rather not admit to, they still belong in the number, because you will probably still spend on them.

Step two: adjust for what actually changes

Retirement changes the composition of spending more than the total. Work through the changes in both directions rather than assuming a uniform reduction.

**Falls, usually:**

• Commuting, work clothing, work meals, professional subscriptions.
• Contributions to long-term savings — you are no longer funding the pot, you are drawing from it. For a heavy saver this is the single largest reduction.
• Costs attached to dependents, if they will have become independent.
• Housing costs, if a mortgage completes or you downsize. Note this one only counts if the completion is genuinely before your retirement date.

**Rises, often:**

• Health cover and out-of-pocket medical costs, which tend to increase with age and can rise sharply if you leave an employer scheme that was covering you.
• Time-filling spending. More free hours generally means more spending in the early, active years — travel, hobbies, eating out, family visits.
• Home services you previously did yourself, later in retirement.
• Support for family in either direction.

**Changes shape:**

• Travel and leisure typically follow a curve: high in the first active phase, lower in a quieter middle phase, and then displaced by care-related costs in a later phase. Modelling a single flat spend for thirty years misses this, though it is a defensible simplification if you widen the range instead.

Apply these as adjustments to the baseline, itemised, so that each one is visible and arguable. A target you can defend is a target where someone can ask "why did you assume health costs rise by that much?" and you can point at the line.

Step three: subtract the other sources

Your portfolio does not have to fund everything. Before converting to a capital number, subtract the income streams that will exist regardless.

Common ones to identify:

• **State or national pension entitlements**, where you have them, based on contribution history and residency. These vary enormously by country and by your own record, so use your own statement rather than an average.
• **Employer or occupational pensions**, particularly any defined benefit entitlement, which is income rather than capital and should be netted off directly.
• **End-of-service entitlements.** In the UAE, private sector employees accrue end-of-service benefits set out in labour legislation administered by the Ministry of Human Resources and Emiratisation.Source^source The important structural point for planning is that this is a lump sum tied to length of service, not an indexed income for life — so it adds to capital, not to the income column, and it does not grow with inflation after you receive it.
• **Rental income**, net of maintenance, vacancy and management, not gross.
• **Continuing part-time work**, if you genuinely intend it. Be conservative here; the ability to work is not guaranteed and health is the usual reason it stops.

What remains after subtraction is the **gap**: the annual amount your own capital must produce.

Be careful about counting an asset twice. If you plan to downsize and release equity, that equity is capital you will add to the pot — but the reduced housing cost is already in your adjusted spending. Count the release once, and only if you would genuinely move.

Step four: convert the gap into capital, as a range

Now the arithmetic. A gap of a given annual amount requires capital equal to that amount divided by the rate you assume you can sustainably withdraw, in real terms, over your horizon.

The mechanics, using a hypothetical. Suppose the adjusted spending is 240,000 a year and other sources provide 60,000. The gap is 180,000.

• At a sustainable withdrawal assumption of 3%, the capital requirement is 6,000,000.
• At 3.5%, it is roughly 5,140,000.
• At 4%, it is 4,500,000.

Three things to notice.

First, **the spread is enormous**. Half a percentage point of assumption moves the target by close to a million in this example. That sensitivity is the honest headline of the whole exercise: the target is not a fact, it is a function of an assumption you cannot verify in advance.

Second, **this is why you should carry a band rather than a point**. Report the target as "roughly 4.5 to 6 million, depending on withdrawal assumption and horizon" and plan against the middle while knowing where the edges are. A single number invites false precision and, worse, invites a false sense of completion when you hit it.

Third, **the withdrawal rate must be a real rate**, meaning it already accounts for inflation increases to your withdrawals. Sustained inflation reduces the purchasing power of a fixed income stream, so a plan that withdraws a nominal amount unchanged for thirty years is not a plan to maintain a standard of living.Source^source Every figure in this exercise should be in today's purchasing power, consistently.

Longer horizons demand lower rates

A withdrawal assumption is not a universal constant; it depends on how long the money must last. A 20-year horizon tolerates a materially higher rate than a 40-year one, because the arithmetic of depletion is far more forgiving over a shorter period. Retiring earlier does not just mean less time to accumulate — it means a lower sustainable withdrawal rate on whatever you accumulated, so the target rises from both directions at once.

Step five: stress test the four fragile assumptions

A defensible number is one you have tried to break. Four assumptions carry most of the risk.

1. **The return assumption.** Test the plan with returns two percentage points lower than your base case for the first decade of drawdown. If the plan collapses, it was not a plan; it was a hope with a spreadsheet.
2. **The sequence of returns.** The order matters far more when withdrawing than when contributing. A poor stretch in the first few years of drawdown forces you to sell more units to fund the same income, permanently shrinking the base. Test what happens if the first three years are bad and how you would respond — typically by holding a cash buffer of one to three years of withdrawals so you are not a forced seller, or by having a pre-agreed rule to reduce discretionary spending temporarily.
3. **Inflation.** Test a period of persistently higher inflation. This hits hardest where income sources are fixed in nominal terms, such as a lump sum entitlement or a non-indexed annuity.
4. **Longevity and health.** Plan to a long life, not an average one. Averages include people who die early, which is the wrong tail to plan against. Then separately consider a late-life care scenario, which is a different, lumpier cost than ordinary living expenses and is the most common reason a plausible plan fails at the end.

The output of stress testing is not a new number. It is a set of pre-decided responses: what you would cut, what buffer you hold, at what point you would adjust. Deciding those in advance is worth more than another decimal place on the target.

What the number does not tell you

• **It does not tell you when you can stop.** The target is a level of capital in today's terms. Reaching it depends on contributions, time, returns and costs, and the last two are not yours to control.
• **It does not tell you the money will last.** A sustainable withdrawal assumption is a probability statement about a range of outcomes, not a guarantee about yours.
• **It does not cover a catastrophic health event or long-term care.** Those are typically handled through insurance, family arrangements or a dedicated reserve, not through the ordinary income calculation.
• **It does not account for your tax position, residency, or the rules of any specific pension or savings vehicle**, all of which vary by jurisdiction and change over time.
• **It is not a competitive score.** Someone else's number reflects their spending and their other sources. Comparing targets is comparing lifestyles, not comparing progress.

Re-anchoring: the part most people skip

A target built once and never revisited decays quickly, because the input that drives it — your spending — moves with your life. The maintenance is light.

Once a year, in the same month:

1. Re-run the twelve-month outflow total. This is the only step that must be redone from data.
2. Re-apply the adjustments, changing any that no longer hold.
3. Re-check the other sources, particularly any pension statement or entitlement.
4. Recompute the band.
5. Compare against actual accumulated capital and note the gap in years of contribution, not just currency.

That last step is the useful one. Expressing the shortfall as "about six more years at the current contribution rate" tells you something you can act on, in a way that "we are 1.8 million short" does not.

Two rules make the annual review honest. First, do not revise the target downward simply because it feels far away; revise it downward only when the spending baseline genuinely falls. Second, when income rises, increase contributions before the spending baseline absorbs the increase — otherwise the target rises with the lifestyle and the gap never closes.

The short version

Build the number in this order: twelve months of real outflows, itemised adjustments for what changes at retirement, subtraction of guaranteed and semi-guaranteed sources, conversion of the remaining gap into capital using a range of withdrawal assumptions in real terms, then stress tests that produce pre-decided responses rather than new numbers.

You will end up with something like "between X and Y in today's money, assuming a horizon of Z years, with a buffer plan for a bad first decade". It is longer than a rule of thumb and less satisfying than a round figure. It is also the only version you can defend when someone — including you, at 3am — asks where it came from.

Source^source: OECD.

Source^source: UAE Ministry of Human Resources and Emiratisation.

Source^source: International Monetary Fund.