As you pointed out in the post, the version of the EMH you are describing, is unfalsifiable. To me it looks like an obvious misstep by our own rules, so why does it hold so much sway?
It’s exactly the kind of thing we were warned about in the Mysterious Answers sequence: A Semantic Stopsign that stops you from asking the obvious next question. A Curiosity Stopper: because there are Smart People who know better, we are not socially obligated to even look.
Its vagueness makes this EMH feel like a motte-and-bailey argument. Can you tell which part is the bailey and which is the motte? Think about this for at least ten seconds before reading on.
Here’s a guess: “You can’t hope to beat the market, so I don’t have to try, and neither should you!” (bailey). But look at all these anomalies! “The market will price those in now that they are known. They’ll evaporate. (Eventually.)” (motte).
But the vagueness means I can only guess. That’s the nature of such arguments. Maybe the bailey is something else or there’s more to it. What is the EMH defending? We may have to taboo “EMH” to nail that down. I suspect I’m insulting a Sacred Cow. Which one? Maybe “prediction markets are a cool! It’s a general solution to a lot of problems!” (If the stock market isn’t efficient, prediction markets aren’t either.) Or rationalists say “Bets are a tax on bullshit,” so why ain’tcha rich? (Sour grapes. You only make money if you’re lucky. Rationalists don’t buy lottery tickets. Is this an Ugh Field?) Or “Capitalism!” (Applause light! vs. I don’t know, something else?). Why are you attached to an unfalsifiable idea?
Is there a falsifiable version of the EMH? OK, forget Popper, we can do one better: To do a Bayesian update we need both sides of the likelihood ratio. What does the world look like if the EMH is true? If it’s false? Which world does ours look like? That’s your prior. If these cases are indistinguishable, then it’s not a Hypothesis, and the “H” in “the EMH” is a lie. A hypothesis that forbids nothing permits everything. Now what evidence would change your mind?
So let’s try some obvious next questions. If the EMH were true, the market’s moves would be a random walk, with a small bias over time based on growth and interest rates, etc. that explain the buy and hold strategy. Subtract out the bias and the residual should be completely random. Can we distinguish random from nonrandom data? Can you measure its entropy? Is the market random? You know how to use a computer. Look and see!
Can you make a profit from this edge? Why aren’t you?
As you pointed out in the post, the version of the EMH you are describing, is unfalsifiable. To me it looks like an obvious misstep by our own rules, so why does it hold so much sway?
It’s exactly the kind of thing we were warned about in the Mysterious Answers sequence: A Semantic Stopsign that stops you from asking the obvious next question. A Curiosity Stopper: because there are Smart People who know better, we are not socially obligated to even look.
I notice I am confused. If we’re playing by the same epistemological rules, we should not have this kind of disagreement. At least one of us is missing something important. What is missing?
Its vagueness makes this EMH feel like a motte-and-bailey argument. Can you tell which part is the bailey and which is the motte? Think about this for at least ten seconds before reading on.
Here’s a guess: “You can’t hope to beat the market, so I don’t have to try, and neither should you!” (bailey). But look at all these anomalies! “The market will price those in now that they are known. They’ll evaporate. (Eventually.)” (motte).
But the vagueness means I can only guess. That’s the nature of such arguments. Maybe the bailey is something else or there’s more to it. What is the EMH defending? We may have to taboo “EMH” to nail that down. I suspect I’m insulting a Sacred Cow. Which one? Maybe “prediction markets are a cool! It’s a general solution to a lot of problems!” (If the stock market isn’t efficient, prediction markets aren’t either.) Or rationalists say “Bets are a tax on bullshit,” so why ain’tcha rich? (Sour grapes. You only make money if you’re lucky. Rationalists don’t buy lottery tickets. Is this an Ugh Field?) Or “Capitalism!” (Applause light! vs. I don’t know, something else?). Why are you attached to an unfalsifiable idea?
Is there a falsifiable version of the EMH? OK, forget Popper, we can do one better: To do a Bayesian update we need both sides of the likelihood ratio. What does the world look like if the EMH is true? If it’s false? Which world does ours look like? That’s your prior. If these cases are indistinguishable, then it’s not a Hypothesis, and the “H” in “the EMH” is a lie. A hypothesis that forbids nothing permits everything. Now what evidence would change your mind?
So let’s try some obvious next questions. If the EMH were true, the market’s moves would be a random walk, with a small bias over time based on growth and interest rates, etc. that explain the buy and hold strategy. Subtract out the bias and the residual should be completely random. Can we distinguish random from nonrandom data? Can you measure its entropy? Is the market random? You know how to use a computer. Look and see!
Can you make a profit from this edge? Why aren’t you?