What is the chance of losing money in the share market?

It depends almost entirely on how long you hold. On the worked example, with a 7.00% expected return and 15.00% volatility, the chance of ending with less than you started is 33.90% over one year, 17.66% over five, 9.46% over ten and 3.17% over twenty. The probability falls because the expected return accumulates faster than the uncertainty around it.

Does investing get safer the longer I hold?

In one sense yes and in another no, and both matter. The chance of a loss falls sharply with time. The range of possible outcomes in dollars widens at the same time: on the worked example, $50,000.00 at twenty years sits somewhere between $71,595.69 and $354,292.17 across the middle eighty percent of outcomes. You are less likely to lose and less certain what you will have.

Which matters more, the return or the volatility?

Volatility, for this question. On the worked example a portfolio with 10.00% volatility has a 1.59% chance of a loss over ten years, while one with 25.00% volatility has a 28.65% chance over the same period, and still 21.27% over twenty years. That is why diversification does more for this than patience does.

Is twenty years long enough to be safe?

For a well diversified portfolio the chance of a nominal loss becomes small, at 3.17% on the worked example. For a concentrated one it does not: at 25.00% volatility the twenty-year figure is still 21.27%. Time reduces the risk of a diversified portfolio far more effectively than it rescues a concentrated one.

What model does this use?

A lognormal distribution of cumulative returns, which is the standard textbook approach. Annual log returns are treated as independent and normally distributed, with the mean and standard deviation derived from the expected return and volatility you enter. The probability of a loss is then the chance that the cumulative log return falls below zero.

How accurate is this?

It is a model, and its main weakness is well known: real market returns have fatter tails than a normal distribution allows, so severe outcomes happen more often than the arithmetic implies. Returns also show some tendency to mean revert over long periods, which works the other way. Treat the figures as the right order of magnitude rather than precise probabilities.

Does this account for inflation?

No. Everything here is in nominal terms, so a loss means ending with fewer dollars than you started with. Preserving purchasing power is a higher bar and would show a materially higher probability of failure at short horizons. Enter a real return, meaning your expected return less inflation, if you want the answer in those terms.

Why does this not use historical data?

Because a single history is one sample of what could have happened, and quoting it as a probability implies more precision than it carries. Driving the model from assumptions you set makes those assumptions visible and lets you test how much they matter, which is more useful than a fixed figure derived from one market over one period.

Investment Time Horizon Risk Calculator

Quick answer: On the worked example below, the chance of ending with less than you started is 33.90% over one year and 3.17% over twenty. The range of outcomes widens as the probability falls: at twenty years the middle eighty percent of results spans $71,595.69 to $354,292.17 on $50,000.00. Volatility moves this more than the return does.

The single most useful thing a nervous new investor can see is how the chance of losing money changes with the length of time they hold. Over one year a diversified share portfolio is close to a coin toss weighted slightly in your favour. Over twenty it is not remotely a coin toss, and understanding why that happens makes the difference between panic selling and sitting still. The mechanism is that the expected return accumulates in proportion to time while the uncertainty around it accumulates in proportion to the square root of time, so the return pulls ahead of the noise the longer you wait. This page turns that into numbers using your own assumptions rather than a fixed history, because a single market over a single period is one sample of what could have happened and quoting it as a probability implies more precision than it carries. It also shows the thing that usually gets left out of this argument. As the chance of a loss falls, the spread of possible outcomes in dollars gets dramatically wider, so you become less likely to lose money and less certain what you will actually end up with. Both are true simultaneously. And it makes clear which input is doing the work: volatility, not expected return. A well diversified portfolio becomes fairly safe over long periods while a concentrated one stays genuinely risky no matter how long you hold it, which is an argument for diversification rather than for patience.

Calculate.co.nz is proud to be partnered with realtor.co.nz, a trusted resource for navigating the New Zealand property market. Their Helpful Articles section offers clear, well-structured insights across buying, selling, and building, making complex real estate topics more accessible. With a focus on up-to-date guidance and practical knowledge, they empower Kiwis to move forward with clarity and confidence in a constantly evolving property landscape.

Updated August 2026 Current 2026/27 rates applied.

These are modelled probabilities, not historical frequencies. The calculation fits a lognormal distribution to the return and volatility you enter. Its best-known weakness is that real market returns have fatter tails than a normal distribution allows, so severe outcomes occur somewhat more often than the arithmetic suggests. Returns also show some tendency to mean revert over long periods, which pushes the other way. Read the figures as the right order of magnitude rather than precise odds, and note that everything is nominal: keeping pace with inflation is a higher bar than simply not losing dollars.

Verification & Methodology

Model: cumulative returns are treated as lognormal. Annual log returns are assumed independent and normally distributed, which is the standard textbook approach to this question.
Converting your inputs: from an arithmetic mean m and volatility s, the log variance is σ² = ln(1 + s²/(1+m)²) and the log mean is μ = ln(1+m) − σ²/2.
Probability of a loss over T years = Φ(−μ√T / σ), where Φ is the standard normal cumulative distribution. The square root is why time helps: the expected return grows with T while the uncertainty grows with √T.
The range shows the 10th and 90th percentiles, so the middle 80% of outcomes fall between them. One in ten results sits below the lower figure and one in ten above the upper.
Chance of beating cash compares the same distribution against a fixed after-tax cash return compounded over the period.
No historical data is used. Every figure derives from the assumptions you enter, which makes those assumptions visible and testable rather than hidden inside a data series.
Known limits: real returns have fatter tails than the model allows, so extreme outcomes are understated; returns are not perfectly independent year to year; and volatility itself varies over time rather than staying fixed.
Excluded: inflation, so all figures are nominal; tax and fees, which should be netted off the expected return before entering it; and any contributions or withdrawals during the period.
Not financial advice. Last verified: August 2026.

Your assumptions Expected annual return

% p.a.

After fees. Enter a real return if you want the answer after inflation. Annual volatility

The standard deviation of annual returns. This input moves the answer more than the return does.

The amount Invested

Only used to express the range in dollars. The probabilities do not depend on it.

Comparing against cash Cash return after tax

% p.a.

Your horizon Holding period

years

9.46%

chance of a loss over 10 years

Over one year

33.90%

chance of a loss

Over twenty years

3.17%

chance of a loss

Median at your horizon

$89,237.44

after 10 years

Beats cash

78.42%

of the time over 10 years

The chance of a loss by holding period

Held for	Chance of a loss	Worst 10% below	Median	Best 10% above	Beats cash
1 year	33.90%	$44,308.10	$52,981.92	$63,353.75	59.82%
2 years	27.85%	$43,599.28	$56,141.68	$72,292.21	63.75%
3 years	23.60%	$43,647.43	$59,489.89	$81,082.58	66.67%
5 years	17.66%	$44,785.75	$66,797.25	$99,627.04	71.10%
10 years	9.46%	$50,700.62	$89,237.44	$157,065.57	78.42%
15 years	5.39%	$59,651.22	$119,216.31	$238,260.49	83.23%
20 years	3.17%	$71,595.69	$159,266.42	$354,292.17	86.70%
30 years	1.15%	$106,763.84	$284,250.56	$756,795.39	91.35%

The probability column falls steadily. The dollar columns spread further apart. Both happen at once, and only the first is usually mentioned.

Volatility matters more than the return

Volatility	1 year	5 years	10 years	20 years	Roughly
10.00%	24.86%	6.45%	1.59%	0.12%	A mixed or balanced fund
15.00%	33.90%	17.66%	9.46%	3.17%	A broad share fund
20.00%	39.26%	27.12%	19.45%	11.15%	A single country or sector
25.00%	42.93%	34.51%	28.65%	21.27%	A concentrated portfolio

At 25% volatility a twenty year horizon still carries a 21.27% chance of a loss. Time does not rescue concentration the way it rescues diversification.

What the range means in dollars

Held for	Worst 10% below	Best 10% above	Spread	Top as a multiple of bottom
1 year	$44,308.10	$63,353.75	$19,045.66	1.43x
5 years	$44,785.75	$99,627.04	$54,841.29	2.22x
10 years	$50,700.62	$157,065.57	$106,364.95	3.10x
20 years	$71,595.69	$354,292.17	$282,696.48	4.95x
30 years	$106,763.84	$756,795.39	$650,031.55	7.09x

At thirty years the top of the range is seven times the bottom. Long horizons make losses unlikely and outcomes far less predictable.

Why Time Helps At All

The expected return accumulates in proportion to how long you hold. The uncertainty around it accumulates in proportion to the square root of that time.

That mismatch is the whole mechanism. Over one year the noise is large relative to the expected gain, so the outcome is close to a coin toss. Over twenty the expected gain has grown twenty-fold while the noise has grown only about four and a half fold, and the return has pulled clear.

On the worked example that takes the chance of a loss from 33.90% to 3.17%.

Worked Example: 7% Expected, 15% Volatility

One year: a 33.90% chance of ending with less than you started. Roughly one year in three.

Five years: 17.66%, about one period in six.

Ten years: 9.46%, about one in eleven.

Twenty years: 3.17%, about one in thirty-two.

A one-year loss is an ordinary event. A twenty-year loss, on these assumptions, is not.

The Part That Usually Gets Left Out

Time makes losses unlikely. It does not make outcomes predictable, and the second point is routinely dropped from the first.

On $50,000.00 the middle eighty percent of outcomes spans $44,308.10 to $63,353.75 after one year, and $71,595.69 to $354,292.17 after twenty. The top of that range is nearly five times the bottom.

At thirty years it is seven times. So the honest statement is that a long horizon makes you very likely to end up ahead and gives you very little idea by how much, which matters if you are planning around a specific number.

Volatility Does The Work, Not Patience

Change the volatility and the long horizons move dramatically. Change the expected return and they move far less.

At 10.00% volatility the ten year chance of a loss is 1.59%. At 25.00% it is 28.65%, and even at twenty years it is still 21.27%.

That is the case for diversification stated as a probability. A concentrated portfolio does not become safe by being held longer; it stays risky and simply gives the risk more time to express itself. Our index concentration calculator and fund overlap calculator both measure how concentrated a portfolio actually is.

Not Losing Dollars Is A Low Bar

Everything on this page is nominal. A loss means finishing with fewer dollars than you started with, which ignores what those dollars will buy.

Keeping pace with inflation is a materially harder test, and the probability of failing it is higher at every horizon. To see the answer in those terms, enter a real return, meaning your expected return less expected inflation, instead of the nominal one.

Our cash drag calculator makes the same point from the other direction, showing how cash reliably preserves dollars while losing purchasing power.

What This Model Gets Wrong

It is worth being specific rather than waving at uncertainty.

Tails are fatter than the model allows. Severe market falls happen more often than a normal distribution predicts, so the genuinely bad outcomes are understated here.

Returns are not perfectly independent. There is some evidence of mean reversion over long periods, which would make long horizons slightly safer than shown.

Volatility is not constant. It clusters, rising in stressed periods and falling in calm ones, rather than sitting at a fixed level.

Those pull in different directions and none of them is small. The figures are the right order of magnitude, not precise odds.

How To Use This

The practical value is matching a horizon to an investment rather than computing an exact probability.

Money needed within a few years sits at the top of the table, where the chance of a loss is substantial and there is no time to recover. Our first home deposit vs invest calculator covers that case, where a fixed date makes the risk worse still.

Money that will not be touched for decades sits at the bottom, where the question stops being whether you will lose and starts being whether you can leave it alone. Our savings to investment switch calculator works the horizon question in reverse, asking how large a fall a given horizon can absorb.

Related NZ Investment Risk Calculators

Savings To Investment Switch Calculator: the fall a horizon can absorb.
First Home Deposit vs Invest Calculator: short horizons with a fixed date.
Index Concentration Calculator: how concentrated your fund really is.
Fund Overlap Calculator: whether your funds actually diversify.
Cash Drag Calculator: the cost of avoiding this risk entirely.

If you've found a bug, or would like to contact us, or learn more about James Graham and Calculate.co.nz.

Calculate.co.nz is partnered with Interest.co.nz for New Zealand's highest quality calculators and financial analysis.

Calculate.co.nz is the sister site of CalculatorHub.com, the world's largest calculator website by tool count.

All calculators and tools are provided for educational and indicative purposes only and do not constitute financial advice.

Calculate.co.nz is proudly part of the Realtor.co.nz group, New Zealand's leading property transaction literacy platform, helping Kiwis understand the home buying and selling process from start to finish. Whether you're a first home buyer navigating your first property purchase, an investor evaluating your next acquisition, or a homeowner planning to sell, Realtor.co.nz provides clear, independent, and trustworthy guidance on every step of the New Zealand property transaction journey.

Calculate.co.nz is also partnered with Health Based Building and Premium Homes to promote informed choices that lead to better long-term outcomes for Kiwi households.

Calculate.co.nz is hosted in Auckland by SiteHost New Zealand.

All content on this website, including calculators, tools, source code, and design, is protected under the Copyright Act 1994 (New Zealand). No part of this site may be reproduced, copied, distributed, stored, or used in any form without prior written permission from the owner.

About & trust: Why Calculate is NZ's most comprehensive · By the Numbers · How we compare · Editorial standards · How we keep data current · NZ finance glossary · Research & data · Financial literacy NZ · About · Privacy policy · Terms of use

Reviewed and maintained. Last reviewed 2026-08-10 and checked on a twice-monthly cycle against IRD, RBNZ and Stats NZ. How we keep data current.