Fees compound too: the long arithmetic of costs

The reframe that makes the number visible

If someone tells you a fund charges one percent a year, your brain files it as small. One percent of anything sounds like a rounding error. That instinct is wrong, and it is wrong for a specific, fixable reason: you are comparing the fee to the wrong thing.

You are comparing it to your **balance**. The fee should be compared to your **return**.

Suppose a portfolio grows at seven percent a year before charges. A one percent annual charge does not take one percent of your money. It takes roughly one seventh of your growth, every year, forever. Say that out loud and the number stops sounding small. If growth were five percent, the same charge takes a fifth. If growth were three percent, it takes a third.

Now add the second reframe. That fee is not deducted once. It is deducted from a balance that would otherwise have been compounding. So you lose the fee, and then you lose the growth the fee would have earned, and then the growth on that growth, and so on for the entire life of the investment. Costs compound against you with exactly the same mathematics that makes compounding attractive in the first place. This is why the level of charges shows up as a standing concern in international work on long-term savings and pension adequacy: it is one of the few variables that reliably and mechanically changes eventual outcomesSource^source.

A percentage charge is not a fee on your money. It is a permanent share of your growth, and the share compounds.

The mechanism, slowly

Let me build it up in stages so the intuition sticks, using deliberately round hypothetical numbers.

**Stage one: one year.** You have 100,000. The market delivers 7 percent, so gross growth is 7,000. A 1 percent charge on the balance takes roughly 1,000. You keep about 6,000. You have given up about 14 percent of the year's growth. Annoying, not alarming.

**Stage two: the missing compound.** That 1,000 is gone. But it was not going to sit still. Had it stayed, it would have grown at 7 percent for however many years remain. Over twenty more years, at 7 percent, roughly 1,000 becomes roughly 3,870. So the true cost of one year's fee, measured at the finish line twenty years later, is not 1,000. It is nearly four times that.

**Stage three: every year.** Now apply the same reasoning to the fee charged in year two, year three, and every year after. Each one is a small amount removed at the time and a much larger amount missing at the end. The losses stack, and the ones taken earliest hurt most because they had the longest to grow.

**Stage four: the compact way to see it.** Rather than tracking every year separately, compare two growth rates. Portfolio A compounds at 7 percent. Portfolio B compounds at 6 percent, because a 1 percent charge is deducted. Over thirty years, one unit of money becomes about 7.6 units at 7 percent and about 5.7 units at 6 percent. The ratio is roughly 0.75. In other words, a single percentage point of annual charge has removed about a quarter of the final value over thirty years.

Read that again. **One percent a year for thirty years costs roughly a quarter of the ending balance.** Not one percent. Not thirty percent of one year. About a quarter of everything.

Push it further and the pattern is stable: over forty years the same one percent removes roughly a third. A two percent annual charge over thirty years removes closer to half. These are approximations, and the exact figures move with the assumed growth rate, but the shape does not change. The relationship between fee level and final value is brutal and non-linear in the direction you would not choose.

Why intuition fails here

Three cognitive traps do the damage.

• **Anchoring on the wrong base.** The fee is quoted against the balance, which is large, making the fee look small. The comparison that matters is against the return, which is much smaller.
• **Linear thinking about a compounding process.** People instinctively multiply: one percent times thirty years equals thirty percent. Compounding does not work by addition, and the true figure is worse in a way that is hard to feel without doing the sum.
• **Invisibility.** Fees are usually deducted inside the fund before the price you see is published. No money leaves your account. No transaction appears. There is nothing to notice, so there is nothing to react to. A charge you would refuse to pay in cash you will pay indefinitely if it is netted off quietly.

The full inventory: seven lines to look for

"The fee" is rarely one number. Before you can compare anything, you need the total. Here is the inventory I would run for any investment product, in the order I would run it.

1. **The fund's own annual charge.** The headline management or ongoing charge figure. This is the number in the advertisement and it is usually not the whole cost.
2. **Transaction costs inside the fund.** When the fund buys and sells, it pays commissions, spreads and sometimes taxes. These come out of the fund's returns and are often disclosed separately or not at all. A fund that turns over its whole portfolio twice a year has real costs here.
3. **The platform, custody or account fee.** What the institution holding the investment charges you for holding it. Sometimes a percentage, sometimes a flat amount, sometimes both.
4. **Advice or wrapper fees.** If someone is advising you or the product sits inside an insurance or savings wrapper, there may be a separate ongoing charge. Wrapper charges are frequently the largest and least visible line in the whole stack.
5. **Entry, exit and early-surrender charges.** One-off costs at the start or end. An exit penalty in the early years is worth identifying before you commit, not after.
6. **Currency conversion spreads.** If you fund the account in one currency and buy in another, the conversion rate applied is a cost, whether or not it is labelled as a fee. This can be a meaningful repeated charge for regular contributions.
7. **Performance fees.** A share of gains above some threshold. Read the threshold carefully, and read whether losses must be recovered before the fee applies again.

Add lines one to four to get an ongoing annual percentage. That is the number to put into the compounding comparison above. Lines five to seven are situational, but line five in particular can dominate if you might need the money sooner than planned.

International securities regulators have devoted considerable standard-setting attention to how these costs are disclosed, precisely because the total is hard for an ordinary investor to assemble from scattered documentsSource^source. In the UAE, checking that a product and the firm offering it are properly licensed, and reading the disclosed charges in the offering documents, is a basic first step before any performance comparison is meaningfulSource^source.

A worked comparison over thirty years

Take a deliberately simple hypothetical, and keep every assumption identical except cost.

Suppose you put aside 2,000 a month for thirty years. Assume the underlying investments return 7 percent a year before charges, every year, with no volatility, because this example is about isolating one variable and not about predicting markets.

• **Version A, total ongoing charges of 0.25 percent.** Net growth rate 6.75 percent.
• **Version B, total ongoing charges of 1.0 percent.** Net growth rate 6.0 percent.
• **Version C, total ongoing charges of 2.0 percent.** Net growth rate 5.0 percent.

You contribute the same amount in all three: 720,000 over the period.

At 5 percent, monthly contributions of 2,000 for thirty years grow to roughly 1.66 million. At 6 percent, roughly 2.0 million. At 6.75 percent, roughly 2.31 million.

Now look at what those gaps represent. Version C ends with roughly 650,000 less than Version A. Your total contributions across the whole thirty years were 720,000. **The cost difference between a cheap arrangement and an expensive one approaches the entire sum of money you saved.** You did the same work, took the same risk, endured the same downturns, and handed over close to a year-for-year equivalent of your own contributions in the difference.

Now express it as a share of the growth rather than a share of the total, which is the honest way to see it. In Version A, growth above contributions is about 1.59 million. In Version C it is about 940,000. The extra 1.75 percentage points of charge consumed about 41 percent of the investment growth. That is the real price tag, and it is nowhere near "1.75 percent".

Three refinements that make it more realistic

The clean example above overstates precision in three ways worth naming.

• **Real returns are not smooth.** Volatility changes the path but not the direction of the fee effect. Charges are typically levied on the balance regardless of performance, so in bad years you pay for losses.
• **Inflation eats into all three versions equally.** If you want the numbers in today's purchasing power, subtract inflation from every growth rate. The _relative_ gap between the versions persists, and in real terms it gets proportionally larger, because a fixed charge is a bigger share of a smaller real return.
• **Higher-cost does not automatically mean worse gross performance.** Version C might, in principle, be a strategy that earns more before charges. The point of the example is not that expensive funds always lose. It is to size the hurdle: Version C's manager must beat Version A's by 1.75 percentage points per year, every year, for thirty years, just to draw level.

What is worth paying for, and what is not

None of this argues for zero cost at any price. Some charges buy something real. The discipline is to know which.

**Usually worth paying for:**

• **Access you could not otherwise get.** Some exposures are genuinely unavailable without a fund structure, and the fee is the price of entry rather than a claim of skill.
• **A structure that solves a real problem.** A wrapper that provides a needed legal, tax or estate function may earn its charge, provided you can state the function in one sentence.
• **Behaviour management, honestly priced.** If an adviser stops you from selling everything in a panic once per decade, that is worth more than the fee. The honest version of this argument comes with a clear, separately disclosed fee, not a charge buried in a product.
• **Genuine operational quality.** Reliable custody, clean reporting, and the ability to actually get your money back are not free, and the cheapest provider is not always the one that delivers them.

**Rarely worth paying for:**

• **A percentage charge on money that is simply sitting there.** Fees scale with your balance, but the work of holding a large balance is not proportionally larger than holding a small one.
• **Layered charges for the same exposure.** Paying a wrapper fee, then a fund-of-funds fee, then the underlying funds' own fees is paying three times for one decision.
• **Complexity you did not ask for.** Structured products with several embedded costs frequently deliver a payoff you could approximate with two cheap holdings.
• **Past performance.** You cannot buy last decade's return, but the fee you pay for the story is very much this decade's.

A practical routine

If you do nothing else, do this once a year. It takes under an hour.

1. **Write down every charge you pay, using the seven-line inventory.** Not the headline number, the total.
2. **Express the total as a share of a plausible long-run return.** If your total charge is 1.5 percent and you assume a 6 percent long-run return, you are paying a quarter of your growth. Write that sentence down.
3. **Estimate the thirty-year cost.** As a rough guide, each percentage point of annual charge removes somewhere near a quarter of the final value over thirty years. Two points, close to half.
4. **Identify the single largest line.** It is usually the wrapper or advice layer, not the fund. Address the largest line first; shaving basis points off the smallest is theatre.
5. **Check for exit penalties before doing anything.** Sometimes the cheapest route out of an expensive product is to stop contributing rather than to surrender it. Sometimes it is not. Get the actual figures.
6. **Confirm licensing and disclosure.** Verify the product and provider against the relevant regulator's registers, and read the charges section of the offering documents in full rather than the summary sheet.

The summary worth remembering

Compounding is usually taught as an optimistic idea: money grows on itself, so start early. That is true. But compounding is a mechanism, not a friend, and it applies with equal force to anything expressed as a persistent annual percentage.

A charge of one percent a year does not cost you one percent. Over an investing lifetime it costs you something closer to a quarter of your final wealth, and two percent costs something closer to half. Those are not marketing exaggerations; they fall straight out of the same exponential arithmetic that makes long-term investing worth doing at all.

The good news is that this is one of the few variables in investing you can control precisely, in advance, without predicting anything. You cannot know next decade's returns. You can know, to the basis point, what you are paying for them.

_This article is financial education, not financial advice. All figures are hypothetical illustrations, not projections or expected outcomes, and do not take account of your personal circumstances. Consider seeking advice from a licensed professional before making decisions about your money._