The Stock Fell 20%: Three Questions Before You React

Your stock dropped 20 percent, so what do you do? Today, nothing with your account. A price is a smoke alarm, not a diagnosis: it has gone off, and that is the whole of what it knows. Open three things in this order: the company's filings from the last few days, the chart of its whole industry, and the market calendar. A fall on its own is not information about a company until you know which of three causes produced it. Reacting because the number is red is the fastest way to lose money for no reason.

A price is not a statement about a business#

The instinct is healthy, which is what makes it dangerous. If someone is handing over shares a fifth cheaper than yesterday, they have surely read the filing more carefully than you, or they know somebody in the industry. Everywhere outside a stock exchange, a price really does carry information about the goods.

Except the seller is under no obligation to have a view on the company.

A price records who had to trade, or wanted to trade, on a given day, not what the company earns. A fund redeems units because clients are withdrawing money. A broker closes a position because the margin ran out. An index portfolio sells every company in the sector, the good ones and the weak ones, because that is what its rulebook says. None of those three sellers opened the annual report, and none of them had to. Who is on the other side of your panic, then? Usually a rulebook.

The threshold is borrowed as well. Twenty percent is the line at which a broad index earns the name bear market, as the US market regulator FINRA defines it. The number that triggers panic about one company was written to describe an entire market, where single businesses average each other out. In a single company, 20 percent is sometimes just a week.

So the question "why did it fall" is worth splitting into three, asked in this order:

1. Did the company publish anything in the last few days that changes its numbers?
2. Is it falling alone, or is the whole basket it sits in falling?
3. Are market mechanics running underneath: leverage, redemptions, rebalancing?

The order is not decoration. The first question settles the most and takes the least time, and the other two only make sense once the first one comes back empty.

What the company published this week#

Documents carry dates, which makes this the only one of the three questions with a hard answer. Open the investor relations page and check whether a current report, an earnings release or an outlook landed inside the window of the fall. If one did, read the document, not an article about it.

Meta is the clean example. The fourth-quarter release landed after the close on 2 February 2022, and the shares gave back more than 26 percent the next day. In the release itself the company reported $33.7 billion of revenue, up 20 percent year on year. In the same document it guided the next quarter to $27–29 billion, or 3 to 11 percent growth. Management cut its own growth rate to roughly half, or to a sixth, of what the business had just delivered, depending on which end of the range you take.

The information was not the price but the outlook. It sat in a document anyone could open for free, and it is still sitting there, including after the price bounced.

An outlook is the one place in a results release where a company talks about the future and puts its name to it. Read it comparatively, never nominally: $27 billion on its own says nothing, $27 billion after $33.7 billion says everything. The trap is that on the day of a release you get a dozen articles about the market reaction and none about the range.

Is it falling alone, or is the basket falling#

An empty first question moves you one level up, to the basket. An index or sector fund buys and sells everything that meets its criterion, without pausing to ask which company is better. When clients withdraw money, the fund redeems units and hands back the whole basket at once. Your company falls not because it did anything, but because it carries the right industry code. Same code, same treatment: fund flows drag good companies down too.

Checking takes two minutes: open the charts of three competitors and the industry index over the same week.

I would bet the middle row is the most common, and it is certainly the worst understood. A fund does not buy a company, it buys a basket, so when money moves in or out the whole basket moves with it, even on days when no company in it has given anyone a reason. Hence the weeks with no announcement at all in which the price still travels low double digits; what moves prices between earnings has its own list of causes, and flows sit near the top of it.

The effect is not symmetric over time, either. A fund taking inflows buys most of whatever carries the largest weight, meaning whatever has risen fastest, and pushes that weight higher itself. On outflows the same mechanism runs in reverse, and at a quarterly rebalance it sells companies wholesale because the rulebook says so, not because any of them broke.

So what does that leave you with? This: in a week when the whole industry gave up low double digits, your company's price carries almost no information about the company. It carries information about how much money left the industry.

When someone sells because they have to#

The third question belongs to the days when everything falls, and whatever had risen fastest falls hardest. On those days the seller is not driven by an opinion but by an obligation. A leveraged investor gets a margin call, and if it goes unmet, the firm can force the sale of the securities, sometimes without notice. What gets sold is what can be sold, not what is weak.

The best-documented recent case is the session of 5 August 2024 in Tokyo. The TOPIX banks index had the worst day in its forty-year history, and the Bank for International Settlements puts it down to the unwinding of positions funded in cheap yen and to a jump in risk metrics that forced funds to cut exposure. On the data that supposedly caused the move, the same institution is blunt: "The US news by itself could not be taken as an unequivocal sign of a deteriorating outlook, let alone a looming global recession, and did not warrant such a market reaction."

Which means an institution whose profession is describing markets stated after the fact that the news did not explain the reaction. The balance sheets of the banks in that index looked exactly the same that evening as they had on Friday. Nobody consulted them. Leverage that runs out of collateral has its own mechanics: a margin call nobody met.

None of this tells you what to do. It tells you what cannot be read out of a price in a week like that, which is usually the more valuable half.

What to do in the hour after the fall#

Three questions sound sensible right up to the moment you are looking at a red screen, so write them down as an order of operations. The whole thing fits in an hour:

1. Open the company's investor relations page and check whether it published anything inside the window of the fall. Read the document, not the article about it.
2. Compare the company's chart with three competitors and the industry index over the same week.
3. Check the calendar: a rate decision, macro data, expiries, an index rebalancing date.
4. Go back to the assumptions you wrote before you bought, and mark which ones this event actually damages.
5. Write down the date and the sentence that will tell you whether you were right or wrong. Then come back to that sentence on the rhythm of the filings rather than the headlines, the way you would with any company you follow.

Step five is the one people skip, and it is the one that separates an investor from somebody living through a market. Without that written sentence, six months later you remember only the feeling of the day, and the feeling always ratifies whatever you did.

Should you buy the dip#

This is the question the reader usually arrives with, so here is the plain answer.

Say you put in $10,000, the price fell 20 percent, and $8,000 is left. You add another $10,000 at the lower price. Your average cost drops by roughly a tenth, while the amount riding on this one thesis grows from $10,000 to $18,000, close to double. Cheaper, and twice as exposed.

Your average cost is a number about you, not about the company. The market cannot see it and will not come back to it out of politeness.

Improvising every call is the first step to hell in the market, because a decision taken without a criterion cannot be graded afterwards: you never learn whether it worked or whether you were lucky. I think a decision to add is worth exactly as much as the condition written down before the fall. Without one it is not a decision but a reaction that happened to be given the name of a strategy.

A good decision and a bad quarter#

That reaction rests on one quiet assumption: that a correct decision produces a visible result quickly. If it did, the winner would simply be whoever brought the most capital, and there would be nothing else to the job. Over short horizons the link is far looser. You can be right about the business and watch a loss for three quarters running, because all three mechanisms in this piece happened to run against you.

So a process is graded by the process, not by the quarter. The question is not how much the price lost. It is whether you had written assumptions when you bought, whether you checked them against documents, and whether any of them has just stopped being true. Three yeses plus a loss on the account means a bad quarter behind a good decision.

The reverse error feels pleasant, which makes it worse. You buy for no reason at all, the price rises, and you walk away convinced you have an eye for this. The result ratified a decision nobody made, so next time you commit more and skip the note. Luck is not a method, and it runs the other way too: a loss after a well-argued purchase can scare somebody off a method that was working. For a single company I would count in years rather than quarters before a result says anything about the decision, and only if you spend those years checking the assumptions rather than the price.

The second and third questions are answered in Taufolio by Today's move: it explains what shifted your portfolio's value over the last few days, including when what shifted was the whole sector. All of it sits inside Monitoring, which costs no credits.

When the fall really does break the thesis#

There is a version of this story in which I am the one who is wrong, and it deserves an honest hearing. Sometimes the market genuinely does know first: a contract quietly lapses, an anchor customer starts shopping elsewhere, somebody in the supply chain sees the order book before any shareholder does. The price can then slide for weeks ahead of the document that confirms it. There is no way to settle that on the day of the fall.

Since it cannot be settled, one thing is left: write down what exactly would have to show up in the numbers for you to call the thesis broken, and put a date on it. A margin down for the third quarter running. Customer losses in the segment that was supposed to grow. A second consecutive cut to the outlook. Those are events with documents behind them, so they carry dates and can be checked. Conditions for breaking a thesis get written before the fall, not on the day the screen turns red.

I do not know whether any particular fall is the start of a lasting deterioration. Nobody knows that on the day, and anyone who says otherwise is selling confidence. What I do know is when I will find out: at the next set of results, against a line I wrote down earlier.

The three questions in this piece have two homes in Taufolio. Breakthrough news speaks up only when a company publishes something capable of changing its business, which answers the first question and the second. The Investment thesis holds the conditions you set while things were calm, and it is what you return to on the day of the fall. Neither costs credits; how the two fit together is easiest to see on the product page.

A smoke alarm sounds the same for a fire as for burnt toast, and nobody holds that against it. It has one job: to make you go and check the kitchen.