Inefficient Market Hypothesis
That finally does make sense
The famous Efficient Market Hypothesis (EMH) states that asset prices reflect all available information. While this bold statement could be either right or wrong, we won’t argue over it here. Instead, we will discuss common interpretations and conclusions for this statement, and why most of these interpretations and conclusions are totally wrong. Stay tuned.
Probably, the most common interpretation of the EMH is that as long as an asset price reflects all available information, then the asset price is the most accurate estimation possible for the intrinsic asset value. Such interpretations lead to the conclusion that one cannot “beat” the market consistently. However, even if asset prices actually do reflect all available information, this doesn’t necessarily mean they reflect it in the right way. This only means that once you are reading some public information, and here by “public” we mean information many people have read before you, then market reaction to this information is already over, so it is too late to “trade” this information. This doesn’t mean the market reaction was adequate. Moreover, markets are known for their inadequate reactions.
So, what about impossibility to “beat” the market? Once asset prices could incorrectly “reflect” available information, shouldn’t it be possible to beat the market by using a better asset price model? Um… it depends on what do we mean by “beat” here.
Let’s perform a thought experiment.
Imagine, that for some stock the whole future cash flow is known to everyone. In other words, for this particular stock, all market participants know exactly when each future dividend payment and the liquidation payment will be made, and what will be the exact amount of these payments. Effectively, such knowledge turns a stock into a risk-free bond. The pricing model for risk-free bonds is well known, and it quite well corresponds with the actual prices of government bonds that are usually considered risk-free. So, with high level of certainty, we could expect the market price of a stock with known future cash flow to also correspond with the pricing model of risk-free bonds. Let’s call this price the true intrinsic value of the stock.
Now, imagine the future cash flow of a stock is known only to you, but not to anybody else in the market. Thus, you are able to directly calculate the true intrinsic value of the stock, while other participants could only try to estimate it, and you want to use your advantage to “beat” the market.
Important note: you know for sure, that your information regarding the future cash flow of the stock is correct, and this is actually the case, but you cannot prove the correctness of this information to anybody else.
Let the market price for the stock be two times lower than the true intrinsic value, so you buy the stock expecting to quickly double your money. However, as long as other participants don’t actually know that the stock is undervalued, there is no reason for the price to grow, so no quick profit for you.
Being a bit disappointed, you decide to wait for the next dividend payment. You assume, that as long as the market underestimates the future stock cash flow in general, it probably also underestimates the next dividend payment in particular. After a larger-than-expected dividend is paid, the market will admit its mistake and correct the price upwards.
However, exactly at the moment of the dividend payment, the market price drops down by the dividend amount. So the market not only refuses to admit its mistake and correct the price upward, but corrects it downward, effectively offsetting the dividend payment. This phenomenon is known as dividend gap.
Even worse, when a dividend payment is made, both the market price and the true intrinsic value of the stock are reduced by the dividend amount, which makes undervalued stocks even more undervalued (and overvalued stocks even more overvalued). When a stock with the market price of $50 and the true intrinsic value of $100 pays a $10 dividend, its market price drops to $40 and the true intrinsic value drops to $90. Before the dividend, the market price was $50 / $100 = 50% of the true intrinsic value, while after the dividend, it will only be $40 / $90 = 44.4%. Tough luck.
Of course, the market cannot indefinitely ignore the reality, as a stock price cannot go below zero, however, as Maynard Keynes once said: “markets can remain irrational longer than you can remain solvent.”
This thought experiment shows, why buying undervalued (really undervalued) stocks doesn’t help “beating” the market in reasonable time frame. Not because the market valuation is correct, but because the market is stubborn and doesn’t want to neither admit nor correct its mistakes.
But why the market behaves in such a way? In order to explain this we need to consider another common form of the EMH, which states that the asset price at the beginning of a time period is the expected (using all available information) price at the end of the period plus expected (again, using all available information) dividend payments during the period, all discounted according to the time value of money. Mathematically, this could be written as:
This formula is commonly (mis-)interpreted as the market continuously estimates expected future prices and expected future dividends, and updates the price based on these estimations. However, such continuous estimation process would require superhuman intelligence and enormous amount of computational resources, thus it is very unlikely to actually happen. Also, this interpretation implies current prices to depend on future prices witch is a bit weird.
However, this formula could be translated into the following one (leave a comment if you want to see formal derivation):
where
and in this form it has much more sense. The formula could be read as the following: the asset price at the end of a time period equals to the price at the beginning of the period, minus dividends paid during the period, adjusted according to the changes of the time value of money and to the new information that became available during the period, all reverse-discounted.
And this is indeed how the market works! In order to determine the next price, it takes the previous price, subtracts any dividends paid since then (thus dividend gap), adjusts the value based on changed interest rates and new information (yes, only new information is considered), then finally divides by the time discount factor, which is usually less than 1, so the price slightly increases over time.
As long as the next price is calculated by adjusting the previous price, the process accumulates errors, rather than eliminates them. If the previous price was already inadequate, the next price will also be, and even more errors could be introduced by incorrectly factoring in the new information.
OK, if knowing the true stock value doesn’t help, how can insider trading be profitable?
Insiders can actually “beat” the market not because they know more, but because they know earlier. If you know some private information about a company that will never be revealed, you cannot actually benefit from it. However, if you get some public information before this information was published, then you may anticipate the market reaction, and front-run it for your benefit. Remember the “important note” above.
So, what are the conclusion?
One cannot consistently “beat” the market on reasonable time frames using “buy-hold-sell” schema, and the problem is not with “buy” or “hold”, but rather with the “sell” part, as the market doesn’t guarantee fair price at selling. However, this doesn’t apply for the “buy and hold” schema. When you buy an undervalued stock not for selling later, but for holding forever, future prices of the stock doesn’t affect you, so there is no reason to measure your success based on the portfolio market value. Eventually you will “beat” the market by earned yield, rather than by market price.
