Learn about market efficiency (EMH) debate — Stock-Market notes

Overview

The Efficient Market Hypothesis (EMH) is the central battleground in finance: can you beat the market, or is all public information already reflected in prices? This debate shapes investment strategy, academic research, and how we understand price discovery.

Figure — Learn about market efficiency (EMH) debate

Core Concepts

The Three Forms of EMH

Why this hierarchy? Each form is stronger (more restrictive) than the last. If semi-strong holds, weak must hold (past prices are a subset of public info). The hierarchy lets us test efficiency at different levels.

Deriving the Random Walk from First Principles

Question: Why does efficiency imply a random walk?

Derivation:

Step 1: Define the pricing condition on total return. Investors are paid via price appreciation and dividends $D_{t+1}$ . If the market requires return $r$ , the efficient price today satisfies: $P_t = \frac{E_t[P_{t+1} + D_{t+1}]}{1 + r}$ This is present value on the total payoff (price plus cash flow). Why include $D$ ? EMH prices total returns; ignoring dividends would mis-state what investors actually earn.

Step 2: Rearrange for the expected total payoff: $E_t[P_{t+1} + D_{t+1}] = P_t(1 + r) \implies E_t\!\left[\frac{P_{t+1} + D_{t+1} - P_t}{P_t}\right] = r$ The expected total return equals the required return $r$ —no predictable excess return exists.

Step 3: Introduce the realized price. Writing the realized total payoff as its expectation plus a surprise: $P_{t+1} + D_{t+1} = P_t(1+r) + \epsilon_{t+1}$ where $\epsilon_{t+1}$ is a surprise—news that wasn't known at time $t$ . Why a surprise? Because if it were predictable, it would already be in the time- $t$ expectation.

Step 4: Characterize $\epsilon$ . By definition of efficiency, $\epsilon_{t+1}$ is:

Zero mean: $E_t[\epsilon_{t+1}] = 0$ (no predictable component)
Uncorrelated with time- $t$ information: $Cov(\epsilon_{t+1}, I_t) = 0$ for any info $I_t$ available at $t$

Step 5: Connect the two random-walk forms.

If the required return $r = 0$ and dividends are zero, we recover the pure random walk (no drift): $P_{t+1} = P_t + \epsilon_{t+1}$ .
If $r > 0$ , prices trend upward on average: a random walk with drift. The drift is just compensation for the time value of money and risk, not a predictable arbitrage.

Formula Random Walk Implications

No drift (fair game): $\;P_{t+1} = P_t + \epsilon_{t+1}, \quad E_t[\epsilon_{t+1}] = 0$ With drift (required return $r$ , total payoff): $\;P_{t+1} + D_{t+1} = P_t(1 + r) + \epsilon_{t+1}, \quad E_t[\epsilon_{t+1}] = 0$ Log-return form (approximate): for small returns, $\ln\!\frac{P_{t+1}+D_{t+1}}{P_t} \approx r + \tilde{\epsilon}_{t+1}$ , where $\tilde{\epsilon}_{t+1} \approx \epsilon_{t+1}/P_t$ is a proportional (percentage) shock. Note this is an approximation: a fixed price-level shock $\epsilon$ is not exactly proportional to $P_t$ , so the log form holds only for small $\epsilon/P_t$ . Consequence: No trading rule based on $I_t$ can predict $\epsilon_{t+1}$ . Past prices, news, analyst reports—all useless for forecasting surprises.

The Case For EMH: Evidence

The Case Against EMH: Anomalies

Why these matter: If efficiency held, these patterns would be arbitraged away. The fact they persist suggests limits to arbitrage or behavioral biases.

Behavioral Explanations: Why Efficiency Fails

Key Behavioral Mechanisms

Overconfidence: Traders overestimate their ability, leading to excessive trading and mispricing of hard-to-value stocks.
Representativeness heuristic: Investors see patterns in random data ("hot hand fallacy"), driving momentum.
Limits to arbitrage: Even if a stock is mispriced, arbitrageurs face:
- Fundamental risk: The mispricing could worsen before correcting
- Noise trader risk: Irrational traders might push prices further away
- Horizon constraints: Hedge funds with short lockup periods can't wait for long-term mean reversion
- Borrowing costs: Shorting overpriced stocks is expensive or impossible

Why this matters: Without free, riskless arbitrage, mispricings persist. This is the Limits to Arbitrage argument (Shleifer & Vishny, 1997).

The Synthesis: Adaptive Markets Hypothesis

Example: The momentum premium has shrunk post-1990s as more funds adopted momentum strategies. This is consistent with adaptive efficiency—anomalies are real but get arbitraged away as they become known.

Practical Implications

If you believe...	Then you should...	Why?
Strong EMH	100% passive indexing	No point trying to pick stocks or time the market
Weak anomalies exist	Factor investing (tilt toward value, momentum, quality)	Capture premiums that persist due to behavioral biases
Markets are exploitable	Active management + short-term trading	Try to identify mispricings before others
Adaptive efficiency	Hybrid approach: passive core + tactical tilts	Recognize that alpha is rare and decays, but not impossible

Common Mistakes & Steel-manning

Connections

Factor Investing Basics: Value, momentum, and quality factors exploit semi-strong inefficiencies
Behavioral Biases in Trading: Overconfidence, herding, and loss aversion drive anomalies
Limits to Arbitrage: Why mispricings persist even when known
Index Funds vs Active Management: EMH's main practical consequence
Random Walk Theory: Mathematical foundation of weak-form efficiency
Fama-French Three-Factor Model: Market, size, value — the risk-based benchmark
Carhart Four-Factor Model: Adds momentum to the three-factor model

Recall Feynman: Explain to a 12-Year-Old

Imagine you're trading Pokémon cards. The Efficient Market Hypothesis says: "Every card's price is already fair, because thousands of kids are constantly checking prices and trading to get the best deal."

If Charizard is worth $100 and you see one for$ 80, you buy it instantly. But so do 50 other kids. Within seconds, the seller raises the price to $100 because of demand. You can't get a "deal" because the crowd is too smart and too fast.

But here's the twist: sometimes most kids believe Charizard will be super rare next year, so they all buy, pushing the price to $120 even though it's "really" worth$ 100. That's a behavioral bias—everyone's excited together. Or maybe some kids notice that cards that went up last month tend to keep going up this month (momentum), so they ride the trend.

The EMH debate is: "Are the kids so smart that prices are always fair?" vs. "Do the kids make predictable emotional mistakes?" The truth is somewhere in between. Prices are mostly fair most of the time, but if you're clever (or lucky), you might spot the mistakes.

Active Recall Flashcards

#flashcards/stock-market

What does the Efficient Market Hypothesis (EMH) claim about asset prices?

Prices fully reflect all available information, making it impossible to consistently beat the market using that information.

What is the weak form of EMH?

Prices reflect all past price and volume data, so technical analysis cannot generate excess returns.

What is the semi-strong form of EMH?

Prices reflect all publicly available information (news, earnings, reports), so fundamental analysis cannot generate excess returns.

What is the difference between a random walk with and without drift?

Without drift:

P_{t+1}=P_t+\epsilon_{t+1}

(fair game,

r=0

, no dividends). With drift:

P_{t+1}+D_{t+1}=P_t(1+r)+\epsilon_{t+1}

, where the upward trend is just compensation for time value and risk, not a predictable arbitrage.

Why must EMH pricing use total returns, not just price changes?

Because investors are paid via price appreciation AND dividends; ignoring dividends mis-states the actual return, so the pricing condition uses

E_t[P_{t+1}+D_{t+1}]/(1+r)

Name two anomalies that challenge EMH.

(1) Momentum: past winners continue to outperform; (2) Post-earnings announcement drift: prices adjust slowly over weeks after earnings surprises.

Which factors are in the Fama-French three-factor model, and which model adds momentum?

Fama-French (1992) = market, size (SMB), value (HML). Carhart (1997) adds momentum as a fourth factor.

What is the "limits to arbitrage" argument?

Even if mispricings exist, arbitrageurs face fundamental risk, noise trader risk, and borrowing costs, so they can't fully eliminate anomalies.

What does the Adaptive Markets Hypothesis claim?

Market efficiency is dynamic and context-dependent; anomalies exist but get arbitraged away as they become known, and efficiency varies over time and across assets.

Why does EMH imply indexing is optimal?

If prices fully reflect information, active management cannot systematically beat the market, so low-cost passive indexing captures market returns without wasting fees.

What is overconfidence bias and how does it violate EMH?

Investors overestimate their ability to pick winners, leading to excessive trading and mispricing of hard-to-value stocks, which should not happen if markets were fully efficient.

What is the January effect?

Small-cap stocks historically have earned abnormally high returns in January, a seasonal anomaly inconsistent with weak-form efficiency.

Concept Map

Hinglish (regional understanding)

Intuition Hinglish mein samjho

Hinglish (regional understanding)

Intuition Hinglish mein samjho

Dekho, EMH ka core idea bilkul simple hai - agar market efficient hai, toh saari available information already price mein reflect ho chuki hoti hai. Socho tumne suna ki Apple kal ek naya product launch karega, aur tum shares kharidne jaate ho. Lekin ruk jao - agar tumhe pata hai, toh hazaaron analysts ko bhi pata hai! Woh log already khareed chuke honge, aur price already us news ko dikha raha hoga. Matlab jab tak tum action loge, tab tak bahut late ho chuka hoga. Yehi market ka magic hai - itne saare smart log ek doosre ko outsmart karne ki koshish karte hain ki paradox ban jaata hai: yeh competition hi consistently jeetna almost impossible bana deta hai.

Ab is idea ko maths mein daalte hain toh "random walk" nikalta hai. Kyunki koi bhi predictable pattern nahi bacha, aane wala price aaj ke price plus ek surprise (epsilon) ke barabar hota hai, jahan us surprise ka expected value zero hai - matlab koi bhi future move predict nahi kar sakta. Agar required return $r$ ko include karo (time value of money aur risk ke liye compensation), toh price upar trend karta hai average mein - isse "random walk with drift" kehte hain. Yahan yaad rakho: yeh drift arbitrage nahi hai, yeh sirf tumhare paise ka fair reward hai. EMH ke teen forms bhi hain - weak (past prices se prediction nahi), semi-strong (public info se edge nahi), aur strong (insider info bhi kaam nahi karti).

Yeh matter kyun karta hai? Kyunki yeh poora debate decide karta hai ki tum apna paisa kaise invest karoge. Agar market efficient hai, toh active management (mehnat karke stocks pick karna) waste of time hai - simple passive index fund mein daal do aur relax karo. Regional student ke liye yeh bahut practical lesson hai: expensive fund managers ko paisa dene se pehle sochо ki kya woh sach mein market ko consistently beat kar sakte hain? Zyadatar research bolti hai - nahi. Toh EMH samajhna tumhe overconfidence se bachaata hai aur smart, disciplined investing sikhata hai.

Test yourself — Factor & Behavioral Finance

Connections