There is also a matter of weighting. Some facts are known but viewed as uncorrelated or unimportant to the stock price. Facts may be obvious, but wrongly discarded as irrelevant.

I think of markets as a real-time implementation of information theory. There are certain facts that exist, and they are not all known although they may be discoverable. As they become known, or are considered more heavily, those facts bleed into market prices.

In this view, the market is a collective model of the world and thus it is not reflective of all true information.

The model overvalues the confidence and beliefs of people who have money. Sometimes, these people trade stock of companies that deal more with lower-income individuals. Their knowledge of those companies is limited, but over time they may converge toward a better understanding.

Also, sometimes there are errors in interpretation of easily discoverable facts. After 9/11, interest rates went down, but the market didn't price that into auto sales despite 0% interest auto loans being offered to the market.

The market is a model that perpetually converges to a set of facts that are constantly shifting. There is always a gradient (like osmotic pressure) between the model and the reality it represents, so there is always some motion in the market.

And there is always a set of expectations about the future that must be reflected in the model. These expectations have a wide variety of distributions, some are gaussian, others are bimodal, etc. That adds another layer of complexity in getting the model to converge to an appropriate expected value.

Not only that, but several different assets are interlinked. For example, if you buy a large block of call options on a high-volatility name, the market-maker will probably buy stock as a hedge. But in the absence of an upcoming event, he will buy patiently over the course of several trading days in order to minimize the price change resulting from his purchase ("delta impact"). So the market knows that the market-maker traded options, but he has an incentive to hide his hedging activity from other market participants so that they don't front-run him. As a result, it takes time for the purchase of the delta via options to be reflected in the underlying stock -- even though the options trade was printed on the exchange immediately and publicly.

Even if markets were efficient to known information, their price moves do not arrive at equilibrium instantaneously. There is an information gradient -- a kind of osmotic pressure between what is truly happening and what is reflected in the price -- that takes time to normalize itself.

That’s a fascinating explanation. Did you come up with it by yourself?

You are thinking about what is logically efficient, but the market is also concerned about what is emotionally efficient.

A certain proportion of people believe climate change is a hoax, a certain proportion stand to make a lot of money shaping policy as if it were a hoax. With the coronavirus, merited or not the panic is real and the markets are reflecting that panic.

In a world economy driven by consumer spending, how people feel is more real than what is actually real for the market. Therefore, the market is always at its most efficient.