What Moves Stock Prices When the Company Said Nothing at All

What moves stock prices on a day when the company has announced nothing? Not its earnings. Over days and weeks the price is set by algorithms, by correlations between companies and by flows of money, and the other side of your order is almost never another investor. A move without news always has a reason. It is rarely your company.

Four days a year and all the rest#

Start with the version that is true. A company is worth the cash it will earn over the coming years, and what it earns today shows up in a filing. Filings arrive once a quarter. Four times a year you get new information about a business you part-own.

If that were the whole story, the price would move in steps. Publication, a jump up or down, then three flat months because nothing new is known. New publication, new step.

It sounds sensible. Except that the price does not wait for the filing: it moves in every session, sometimes by several percent, through weeks in which nothing inside the company changed that anyone could name.

No company's chart looks like that.

What actually changes inside a company over three months? It signs contracts, opens stores, lays people off, loses tenders. All of it happens slowly and quietly, and it only becomes visible from outside as numbers, on publication day. A factory does not become two percent less productive between Tuesday and Wednesday.

So the question is who sets the price on the other two hundred and fifty days.

Who is on the other side of your order#

When you click buy, you picture your order meeting somebody's sell order. That was true when an exchange was one trading floor. Today the same security trades in parallel on many venues at once, and your order enters plumbing you cannot see from the tap.

"The current price" sounds like one number. It is a record of the fact that somewhere, on one of those venues, somebody just traded at that level. A company has one listing venue, but its shares change hands in the same second in many places, at prices that differ by fractions of a cent. Your chart shows one of them, picked and smoothed by somebody along the way.

In a 2014 literature review the SEC noted that estimates of the share held by high frequency trading firms then typically exceeded 50% of total volume in US-listed equities, and called that trading "a dominant component of the current market structure" (SEC literature review). Which means roughly every second trade has somebody who read the filing on one side and, on the other, somebody who never intends to open it.

You can see the timescale in how the market tried to level it. The Investors Exchange runs every order through a 350-microsecond delay, designed to take the head start away from the fastest participants (SEC paper). The advantage this whole game is played for fits inside under a three-thousandth of a second.

Then there is the tick, the smallest amount by which somebody's order can be improved. For years that was a whole cent. From the first business day of November 2025 the US regulator allowed a second increment of half a cent for the tightest-spread securities, which doubles the number of places where somebody can step in front of your order (SEC, minimum pricing increments).

I think this is the single most useful thing a private investor can understand about the market: over one day you are not racing another person, you are racing the speed of light in a cable. That race is not won with a faster connection or an earlier alarm.

Why your commission is lower than it should be#

If a machine is going to stand on the other side of your order, it has to receive the order first. It receives it from your broker.

The mechanism is called payment for order flow. Instead of sending your order to a public exchange, the broker routes it to a firm that will fill it itself, and is paid for the routing. Part of that payment comes back to you as a lower commission. So you see cheap and fast. You do not see the price the trade might have got somewhere else.

The trace of this mechanism is on your own trade confirmation. Look at how many decimal places your broker reports in the execution price. Four means the order was filled in a fraction of a cent, and that fraction is the price improvement the executing firm handed back above the quote. Two means you are getting the same trade rounded to the whole cent. Neither version tells you what that order would have got on a public exchange.

The EU decided that an intermediary cannot look for the best price for the client and the best payment for itself at the same time. Regulation 2024/791 came into force on 28 March 2024, and member states that previously allowed such payments have to phase them out by 30 June 2026 (EUR-Lex).

So the same mechanism is standard in the United States and is being wound up in the EU. It is not often that the answer to "is this allowed" depends on which side of the ocean your account sits.

Half the volume never reaches the chart#

An order that never went to a public exchange never reaches the tape either, and the tape is where the volume bar under your chart comes from. Private venues, known as dark pools, exist so that a large order can be filled without being shown to everyone first.

The logic is sound. Say a pension fund has to sell $2 billion of stock, not because it stopped believing in the company but because it has pensions to pay. If it posted that order publicly, everyone would see how much it had to sell and would hold off buying. The seller would kill its own price by announcing it, and the company would take a markdown that has nothing to do with the company.

The volume you see under the candle is now a meter reading for a pipe carrying under half the water. FINRA summed up 2025 with the observation that, measured in shares traded, over-the-counter volume – private venues together with the firms filling retail orders in-house – exceeded on-exchange volume for the first time (FINRA).

A conclusion drawn from volume alone is a conclusion drawn from the smaller half.

Separately there is after-hours trading. When the main venues close, trading continues, so the reaction to results published after the bell begins without most participants. In the morning you see the result of that and call it an opening gap.

That market runs by its own rules, because very few people are in it. Thin liquidity means the same number of shares moves the price far more than it would midday. Say plus eight percent at ten in the evening and minus two by noon the next day: not a contradiction, but two measurements of the same event taken at two different depths of the market.

An algorithm sells one company because another stumbled#

That leaves the question of why machines move a price on a day when nothing happened. There are as many reasons as there are people writing algorithms, but one pattern explains a surprising share of the moves that look senseless from outside.

It is pairs trading.

An algorithm watches two companies in the same industry that have moved together for years. Say the correlation is 80%. When one falls on its own news and the other stays put, the correlation breaks, and the algorithm bets on it reverting: it sells the one that stayed put.

Say the first is Coca-Cola and the second is Pepsi. A story about Coca-Cola's chief executive comes out in the morning. By lunchtime Pepsi is down, though nothing changed at Pepsi and nobody wrote a line about it. An investor looking only at the Pepsi chart concludes the market is random. The market was not random. It was about something other than what that investor thought.

The second layer is keyword scanners. Algorithms do not wait for an analyst: they sweep releases, wires and company filings for phrases that have historically preceded declines. A cluster of negative wording inside one hour is enough to close a position before anyone has read the whole thing.

The simplest version of this runs on a single word. Say a program reads an earnings release, finds revenue below forecast, and sells without checking by how much, or why, or what the company expects next quarter. By the time you open the same release and see that it was half a percent and one delayed shipment, the price is four percent lower. The document says one thing, the price says another, and for the next few hours the price is louder.

The third layer is algorithms trading against other algorithms. When a strategy visibly makes money, programs appear that copy it without understanding why. When it stops working, programs appear that trade against it. The move you see at 15:40 can be the third floor of that spiral.

And here is the part that is hardest to accept. The people who built these systems do not always know why their algorithm sold at that particular moment: a model learns across dozens of data feeds and decides in a fraction of a second, with the explanation arriving afterwards, if it arrives at all.

If the person who built the machine does not know the reason, your odds of reading it off a chart are known.

I also do not believe a better model changes this. A program that predicted tomorrow's price would have to model the decisions of tens of thousands of other programs running in the same second, working from the same sliver of data you have. More compute does not substitute for information nobody can see.

How much of that move is information#

This has been measured. Four researchers decomposed the return variance of US stocks and checked how much of it carries anything at all (Brogaard, Nguyen, Putniņš, Wu, Review of Financial Studies 2022).

Noise alone weighs almost four times as much in a price as everything happening to the economy and the indices combined: 31% against 8%. Roughly one unit in three of your company's volatility says nothing about the company and nothing about the market.

Read that table comparatively, not nominally. Thirty-one percent noise does not mean every third day is empty. It means that across all the volatility your company generates in a year, roughly that much cannot be attributed to information about anything. It is spread very unevenly: there is a lot of information on results day and almost none on an ordinary Thursday.

The same paper carries good news. Since the mid-1990s the noise share has fallen markedly and the firm-specific information share has risen. The market is getting more accurate about what it prices. It is not getting any calmer over a single day.

"The market fell today" says nothing about your portfolio#

That leaves a sentence you hear daily and that means almost nothing.

The market is not an object you can buy. It is thousands of companies going in different directions and a weighted average in which not every vote counts the same.

The weight is currently spread unusually unevenly. The International Monetary Fund writes plainly that a small set of firms increasingly drives equity markets, and that the concentration measure for two of six major equity markets exceeds its 95th historical percentile (Global Financial Stability Report, April 2026). Those two indices now depend on a handful of firms more than in ninety-five years out of a hundred.

The consequence is unintuitive. An index can rise on a day when most of its companies fall, and the reverse, because what counts is not how many winners there are but how much they weigh. If your company is not among the ones carrying the weight, "the market rose 2%" is information about a few firms you do not own.

A statement about an index is a statement about weight, not about companies.

Flows add their own layer. Buying an index fund means buying the whole basket at once, so your company rises and falls with the fashion for that basket; I take that mechanism apart separately in the piece on flows into ETFs. Macro works the same way and is context rather than a compass: it explains why everything is expensive or cheap, and says nothing about whether your company can hold its prices.

When a move without news does mean something#

All of this reads as though any move without a release can be ignored. It cannot, and that is the most expensive conclusion available from this piece.

First, some information enters the price through trading before it becomes public. That is how the "private information" row in the table above works: 24% of the volatility is knowledge that reaches you late. A 20% drop with no announcement is sometimes the first symptom of something the company will describe three weeks from now.

Second, some moves come from compulsion rather than opinion. When a leveraged investor gets a margin call, they sell what can be sold rather than what they stopped believing in, and that loop has a dynamic of its own that I cover separately. The mirror image is short covering, which can give a rebound the shape of a V before anyone has changed their mind about the business.

Third, in a small company with thin turnover all of this plumbing matters less. There, one order worth a few hundred thousand dollars really can be the news, and it really is worth knowing who placed it.

There is one test, and it fits into a question: can today's move be attributed to a document? If the company or a regulator published something, you have an event and you check it at the source. If nobody published anything, you have a move rather than a fact, and I take that distinction apart in the piece on noise and signal.

The test does not settle whether an event matters. It settles only whether one exists. The work you can actually do starts after that: read the document and check which assumption in your thesis it touches.

Which horizon you have an edge on#

If you lose to physics over a single day, where exactly do you not lose? The answer follows from what sets the price over each stretch of time.

The first three rows need something you do not have: a cable, a data feed or a desk. Row four needs something almost no fund measured on quarterly performance has, which is permission for a year to pass with nothing happening.

That is not a consolation prize. It is a description of the one place where your constraints turn into an edge: nobody forces you to sell, nobody grades you monthly, and you have no client withdrawing money in the worst possible week.

What to do about it this week#

Three habits are enough for this to start saving you time rather than taking it.

1. When a price moves with no release, check the whole sector first. If the sector moved with it, your company is not the story that day.
2. Before you treat a day as information, find the document with that date on it. No document, no event.
3. Read the year off the earnings calendar, not off the chart. A good company's share price can end a year down while nothing in the business broke.

Where to see what actually moved your portfolio#

"Why is my portfolio down 1.4% today" is a fair question. The answer just almost never sits in one company. Today's move therefore looks wider than your holdings, because a portfolio can be moved by politics, by macro data or by an earnings call at a company you do not own. It sits on a tile with a second tab, Portfolio essence, which is what happened across your companies over the past month. Both come with your plan. Breakthrough news, a part of Monitoring, speaks up zero to two times a month, because that is how often something genuinely changes a business. None of the three costs credits, and you can see how they fit together in the product.

I do not know how much of today's move in your company is noise and how much is somebody's knowledge, and I will not find out today. I do know when I will check: in the next filing, which will either support the thesis or take it apart.

So stop asking why the price fell yesterday. Start asking what has to show up in the next filing for you to change your mind about the company. The first question has no good answer. The second has a date.