Correction vs Bear Market: How Markets Behave
A market that only went up would be free money, and there is no free money. The price of owning stocks is that you have to hold them through stretches when they are worth less than you paid, without knowing in advance how much less or for how long. That is not a flaw in the system. It is the system.
Most people accept that in the abstract. What they do not have is a sense of scale. How often does the market fall? How far does it usually go? How long does it take to come back? Which falls are ordinary weather and which ones are something breaking? Without those numbers, every decline feels like the beginning of a catastrophe, because there is nothing to measure it against.
This guide supplies the scale, and it takes it from one specific record: the S&P/TSX Composite, the index that tracks the bulk of the Canadian stock market, across 11,848 trading days from June 1979 to September 2026. Every figure below comes out of that series.
One boundary before the numbers start. This guide is about the market as a whole. A single company can fall 40% in an afternoon on news that has nothing to do with anyone else, and the reasons a particular stock moves are a separate subject, covered in what moves a stock price. Here we are looking at the weather, not at one tree.
The words, and who actually decides them
Start with the vocabulary, because the words sound official and are not.
There is no regulator, exchange or standards body that defines a correction or a bear market. No rule book contains those thresholds. They are market conventions: round numbers that financial journalists and analysts settled on decades ago and now repeat because everyone else does. They describe magnitude, nothing more. Knowing that a decline has crossed 20% tells you how far it has fallen and absolutely nothing about why, or what happens next.
| Term | What people mean by it | What it actually tells you |
|---|---|---|
| Pullback | A fall of less than 10% from a recent high | Almost nothing. The market does this constantly. |
| Correction | A fall of 10% or more | That the decline is now large enough to be widely reported |
| Bear market | A fall of 20% or more | The same, one round number further along |
| Crash | A very large fall in a very short time | That the speed was unusual. There is no threshold. |
| Drawdown | The distance from a previous peak to a later low, at any size | The actual measurement the other four are labels for |
| Volatility | How large the moves are, up or down | The size of the swings, not their direction |
| Bull market | A sustained rise, usually dated from the last bear market low | Where we are in the story, told backwards |
Drawdown is the useful one. It is the plain measurement: how far the market is below its own last record, expressed as a percentage. Correction and bear market are just two lines drawn on that measurement.
How arbitrary those lines are is easy to demonstrate. Between September 1983 and July 1984 the S&P/TSX Composite fell 19.96%. It took nine and a half months to bottom out and another six months to get back to even. By convention that was not a bear market, because it stopped four hundredths of a percentage point short. Anyone who lived through it experienced something indistinguishable from one.
Treat the words as shorthand for scale. Do not treat them as diagnoses.
What the record actually looks like
Here is every day of the last 47 years, drawn as the distance between where the index closed and its own previous record high.
The shape of that chart is the single most useful thing a new investor can internalise, and it is this: the market is almost never at a record.
Over those 11,848 days, the S&P/TSX Composite closed at a new record high on 6.2% of them. It closed below a previous high on 93.8%. It was more than 10% below a previous high on 43.8% of all trading days, more than 20% below on 19.6%, and more than 30% below on 8.8%.
Read that again with a portfolio in mind. If you own Canadian stocks, then close to half the time you look at your account, the market you own will be in what a headline would call a correction or worse. Not because something has gone wrong. That is simply the resting state.
The white space at the top of that chart, the moments at a record, is where almost nobody feels anxious and where almost nobody is buying with any urgency. The deep red canyons are where the news is worst and the buying is best. That inversion is the entire psychological problem of investing, and it is visible before you read a single word about behaviour.
How deep they go, and how long they last
In 47.2 years the index fell 10% or more from an all-time high on 19 separate occasions: one every two and a half years. Eight of those went on to reach 20%: one every 5.9 years.
Those eight are worth seeing individually, because the averages hide how different they are from each other.
| Peak | Trough | Fall | Months falling | Months back to even |
|---|---|---|---|---|
| Feb 29, 1980 | Mar 27, 1980 | -22.35% | 0.9 | 3.9 |
| Nov 28, 1980 | Jul 8, 1982 | -43.95% | 19.3 | 9.9 |
| Aug 13, 1987 | Oct 28, 1987 | -31.00% | 2.5 | 69.9 |
| Apr 22, 1998 | Oct 5, 1998 | -31.78% | 5.5 | 13.7 |
| Sep 1, 2000 | Oct 9, 2002 | -49.99% | 25.2 | 38.8 |
| Jun 18, 2008 | Mar 9, 2009 | -49.80% | 8.7 | 63.3 |
| Sep 3, 2014 | Jan 20, 2016 | -24.36% | 16.6 | 12.7 |
| Feb 20, 2020 | Mar 23, 2020 | -37.43% | 1.1 | 9.5 |
The median bear market took 7.1 months to reach its low and 13.2 months to climb back to where it started. The eleven declines that stopped short of 20% were much quicker: a median of 3.1 months down and 3.7 months back.
But look at the spread rather than the median. The 2020 crash took five weeks to fall 37% and nine and a half months to recover. The 1987 crash fell 31% in ten weeks and then took five years and ten months to get back to its old high. Same rough depth, wildly different sentences to serve. Nothing visible at the low told you which one you were in.
Two honest qualifications on that table.
First, these are price levels only. The S&P/TSX Composite as quoted excludes dividends, and a Canadian investor holding those companies collected cash throughout every one of those declines. The month a holder’s total position got back to even therefore came earlier than the month the index did, in every row. The table understates the recovery.
Second, and more important: 19 declines reached 10%, and only 8 carried on to 20%. There is no reliable way to tell, at the moment a market is down 11%, whether you are in one of the eleven that stops there or one of the eight that keeps going. Anyone who tells you otherwise is describing the past. The falls that eventually became the two worst bear markets in the series, in 2000 and 2008, both looked exactly like an ordinary correction on their way through 10%.
A fall inside the year says nothing about the year
The other number people misread is the annual one. A year is reported as a single figure, so an investor comes to imagine years as either good or bad. The lived experience inside them is nothing like that.
Across the 46 calendar years from 1980 to 2025, the S&P/TSX Composite finished 33 of them higher than it started. In the median year, it fell 11.8% at some point along the way.
Twenty-six of the 46 years contained a fall of 10% or more. Thirteen of those 26 still finished higher than they began. Half the time a headline correction arrives, the year it happens in ends up being a perfectly good year.
That is the practical lesson of the chart. A decline in progress carries no information about how the year finishes. The dots and the bars are close to unrelated. An investor who reacts to the dot is reacting to something that does not predict the bar.
Volatility is a measure of size, not direction
People use volatile as a synonym for falling. It is not. Volatility measures how large the daily moves are, in either direction, and the largest up days and the largest down days show up in the same weeks.
Over the last 20 years, 5,020 trading days, the S&P/TSX Composite finished higher on 55.3% of days and lower on 44.6%. It moved more than 1% on 20.8% of days, more than 2% on 5.1%, and more than 3% on just 1.8%. Ninety-one days out of 5,020 moved more than 3%.
Now notice where they sit. The three largest single-day gains in the last two decades were March 24, 2020 at plus 11.96%, October 14, 2008 at plus 9.82%, and March 13, 2020 at plus 9.66%. The three largest single-day losses were March 12, 2020 at minus 12.34%, March 9, 2020 at minus 10.27%, and March 16, 2020 at minus 9.89%.
The best days and the worst days are the same days, in the same fortnights, in the same two crises. Volatility is not the market going down. It is the market disagreeing with itself loudly, and that argument produces violent moves in both directions before it resolves.
The Ontario Securities Commission’s investor education service makes the related point about why the argument gets loud, in its explanation of how market volatility affects your investments: prices swing on new information, on world events, and on shifts in sentiment, and the biases that amplify them have names. Herd behaviour is doing what everyone else is doing rather than what your own information supports. Confirmation bias is reading each new fact as proof of what you already believed. Overconfidence bias is mistaking a run of luck for skill. All three are stronger during the weeks when the moves are largest, which is exactly when a decision costs the most.
Why a Canadian’s bear market is not an American one
This is where the generic version of this article stops being useful to you, because a bear market is not a global event with one number attached. It happens to an index, and the index you own decides what you experience.
Here are five episodes measured on both the S&P/TSX Composite and the S&P 500, peak to trough within each window.
| Episode | S&P/TSX Composite | S&P 500 |
|---|---|---|
| Dot-com bust, 2000 to 2002 | -50.0% | -48.9% |
| Global financial crisis, 2008 to 2009 | -49.8% | -49.6% |
| Oil crash, 2014 to 2016 | -24.4% | -14.2% |
| Pandemic crash, 2020 | -37.4% | -33.7% |
| Rate shock, 2022 | -17.6% | -25.4% |
The first two rows are the ones everyone remembers, and in those the two markets fell together almost exactly. Global crises are global.
The interesting rows are the third and the fifth. In the collapse in oil prices that ran from late 2014 into early 2016, the Canadian market fell 24.4%, a full bear market by the conventional definition, while the American market fell 14.2% and most US investors remember a rough patch rather than a bear market. In 2022, the direction of the surprise reversed: the Canadian market fell 17.6% and never technically reached bear territory, while the US market fell 25.4%.
The cause is composition. The S&P/TSX Composite is concentrated in banks, energy producers and miners, so a collapse in commodity prices lands on a large share of the index at once. The American market carries far more of its weight in technology, so a repricing of technology valuations, which is what 2022 largely was, hurts there and passes over Canada more lightly.
Two consequences follow, and they are the reason this matters to you rather than being trivia.
The first is that headlines are usually about the American market, so when you read that stocks are in a bear market, the number in the headline is often not the number in your account. Check the index you actually own.
The second is that concentration cuts both ways, and it is an argument for knowing what sits inside your holdings. If most of your portfolio is Canadian, the thing most likely to produce your personal bear market is a commodity shock or something that hits the Canadian banks, because that is where the index’s weight is. That is worth knowing before it happens rather than during. It is also the practical case for holding companies whose earnings do not all depend on the same handful of drivers, which is what the blue chip end of the market is usually reached for.
The mistake, and what it costs
The mistake is not owning stocks during a fall. Everyone does that. The mistake is selling during one.
It feels like the responsible thing. The market is down 20%, the news is uniformly bad, and getting out looks like limiting the damage until things calm down. The problem is arithmetic, and the arithmetic is brutal.
Ten thousand dollars tracking the index for those 20 years and never touched became $29,933. The same $10,000, absent for only the five best days out of 5,020, became $19,319. Absent for the ten best days, $14,509: less than half the result, from being out of the market for ten days in twenty years. Absent for the best 30 days, $6,605, which is less than the amount originally invested.
Now the fact that turns this from a curiosity into a warning. Of the twenty best single days in those two decades, nineteen happened while the index was 10% or more below its previous high.
The best days are not scattered through the calm periods. They are packed inside the frightening ones. The best day of the entire twenty years, March 24, 2020, came the day after the pandemic low, in a week when the news was as bad as it has ever been. October 14, 2008, the second best day, sits in the middle of the financial crisis.
So the manoeuvre does not work, and it fails for a structural reason rather than through bad luck. Selling to avoid the worst days requires being out during precisely the stretch that contains the best days, and you have to be right twice: once about when to leave and once about when to return. In March 2020 the second decision came due immediately. The low was March 23, and the largest single gain of the entire twenty years landed the following morning.
The specifically Canadian version of the cost
If the selling happens inside a registered account, there is a second bill on top of the investment one, and it is a bill many people do not know exists until they have paid it.
Selling a holding inside a TFSA triggers no tax, which is the point of the account. Taking the money out is the problem. The Canada Revenue Agency is explicit that when you withdraw from a TFSA, the amount “will only be added back as available contribution room on January 1 of the next calendar year.” Panic in March and move $30,000 out of your TFSA to cash, and you cannot put that $30,000 back until January 1, whatever the market does in between, unless you happen to have that much unused room already. Putting it back sooner is an over-contribution, and an over-contribution is taxable even when it was an honest mistake. You can check what room you actually have with our TFSA contribution room calculator, and the mechanics of the account are set out in our TFSA guide.
An RRSP is worse, because the exit is taxed. A withdrawal is income in the year you take it, and your institution withholds tax immediately. The CRA publishes the withholding rates on RRSP withdrawals: for residents of Canada, 10% on amounts up to $5,000, 20% on amounts over $5,000 up to and including $15,000, and 30% on amounts over $15,000, with the Quebec rates set at 5%, 10% and 15% plus provincial withholding. The CRA also notes that the amount withheld may not cover what you actually owe once the withdrawal is added to your income for the year, so a further bill can follow at tax time. And the contribution room you used up is gone permanently: an RRSP withdrawal does not restore it. The RRSP guide covers how that room is built in the first place.
Put those together. A frightened sale inside an RRSP can cost the fall, the recovery you were not present for, an immediate tax withholding, a possible top-up bill in April, and the contribution room itself. The market decline was the cheapest part of it.
What actually helps
None of this is an argument for never selling. It is an argument against selling because of a number on a screen. Some specific things do help.
Decide the horizon before you buy, not during the fall. The OSC’s investor education service frames this as the length of time before you need the money back, and the relationship is simple: the longer that period, the more room you have to sit through a decline. Money you need within three years does not belong in stocks, and the reason is in the bear market table above. The median took just over a year to get back to even, and one of them took nearly six.
Know what you own well enough to tell the two situations apart. There is a real difference between a price falling and a business deteriorating. Prices fall for reasons that have nothing to do with the company, as the guide on what moves a stock price sets out at length. When the reason you bought something is still true, a lower price is a lower price. When the reason has stopped being true, that is a different decision and it is worth making deliberately.
Keep buying on a schedule. The OSC names dollar-cost averaging, investing a set amount at regular intervals, as one way to handle volatility. Its mechanical virtue is that it takes the timing decision away from your mood. Fixed amounts buy more units when prices are down, which is when your instincts are working hardest against you.
Do not check daily. More than one day in five moves the index by more than 1%, and almost none of those days mean anything about the years that matter to you. The stock quote you are looking at contains a 52-week range for exactly this reason: it puts today’s number next to the span it has already travelled.
Write down what would make you sell, in advance. A sentence written while calm is worth a great deal on a day the index is down 8%. It also gives you a way to sell that is not panic, which is what makes it credible.
What to take from this
The market’s normal condition is to be below its last record: 93.8% of days. It falls 10% about every two and a half years and 20% about every six. The median bear market takes seven months to bottom and thirteen to recover, but the range around that median is enormous and nothing at the low tells you which one you are in. Most years contain a double-digit fall, and most years still finish up. The largest up days live inside the worst weeks, so the act of avoiding the worst of it is also the act of missing the best of it. And in Canada the exit itself carries costs, in contribution room and in tax, that the fall never charged you.
None of that makes a decline pleasant. It makes it legible, which is the difference between an event you can sit through and one you react to.
Where to go next
- The investing guides index lists this series in order, from what a share is through to reading a company’s filings.
- How the stock market works covers the exchanges, liquidity and market makers behind the index this guide measures.
- What a stock is, and what you actually own is where the series starts, if the share itself is still fuzzy.
- Compounding, and why time in the market beats timing it is the other half of this guide: sitting through the falls above is only worth doing because of what the arithmetic does with the years, and it puts a number on what leaving early costs.