What Moves the Stock Market?

This paper surveys what is known, what is contested, and what remains unknown about the determinants of stock market returns and risk, and it applies that body of knowledge to conditions in October 2026. It integrates five literatures: market microstructure, theoretical and empirical asset pricing, return predictability, behavioral finance, and macro-finance. It adds a historical analysis of seven major market crises and a dated assessment of the current U.S. market. The long-run evidence is unambiguous on one point. U.S. equities returned 9.8% per year in nominal terms and 6.6% in real terms from 1900 to 2025, against 1.6% real for bonds and 0.5% for bills (Dimson, Marsh, & Staunton [DMS], 2026). That premium is large relative to the predictions of standard consumption-based models, which is the equity premium puzzle (Mehra & Prescott, 1985). Cross-sectional evidence is more fragile than it first appeared. Hundreds of published "anomalies" shrink by roughly half after publication (McLean & Pontiff, 2016), and many fail stricter statistical hurdles (Harvey, Liu, & Zhu, 2016; Hou, Xue, & Zhang, 2020). A core set of factors, including market, value, momentum, profitability, and investment, survives replication across countries (Jensen, Kelly, & Pedersen, 2023). Aggregate predictability from valuation ratios is real at long horizons but weak and unstable out of sample (Goyal & Welch, 2008; Campbell & Thompson, 2008). Applying these lessons to October 2026, the S&P 500 sits near record highs at a cyclically adjusted price-to-earnings (CAPE) ratio near 41.6, second only to December 1999. The top ten stocks hold roughly 39% of the index. The 10-year Treasury yield has risen above 5% for the first time since 2007, and the Federal Reserve resumed rate hikes in September 2026 amid an oil shock tied to the 2026 Iran war. Valuation evidence implies below-average expected long-horizon returns. Very strong earnings growth driven by artificial intelligence investment is the main counterweight. The paper closes with a synthesis that grades the confidence of each major claim and sets out an agenda for future research.

Keywords: equity risk premium; asset pricing; market efficiency; factor investing; return predictability; behavioral finance; market microstructure; financial crises; market concentration; monetary policy

JEL Classification: G10, G11, G12, G14, G18, G41, E44, E52

1. Long-run reward for risk is real and large. U.S. equities earned 6.6% per year after inflation from 1900 to 2025, versus 1.6% for government bonds and 0.5% for Treasury bills (DMS, 2026).
2. The premium is not uniquely American, but the U.S. was unusually lucky. Equities beat bonds and bills in all 21 countries with continuous histories since 1900 (DMS, 2026). The U.S. now accounts for about 62% of world equity market value, so U.S.-only data overstate what a random investor in 1900 should have expected.
3. Standard theory cannot explain the size of the premium. Consumption-based models with plausible risk aversion predict a premium well under 1%, against roughly 6% observed (Mehra & Prescott, 1985). Habit formation, long-run risk, and rare disasters are the leading candidate explanations. None is decisive.
4. Forward-looking premium estimates are lower than history. Damodaran's implied equity risk premium (ERP) for the S&P 500 was 4.20% on October 1, 2026, below the roughly 6% historical premium over bills.
5. Most stock-level "anomalies" are weaker than published. Anomaly returns fall about 26% out of sample and about 58% after publication (McLean & Pontiff, 2016). Roughly 65% of 452 anomalies fail a conventional significance test under more careful construction (Hou et al., 2020).
6. A small core of factors is robust. Market, value, momentum, profitability, investment, and low-risk factors replicate across 93 countries (Jensen et al., 2023). Momentum earned the highest long-run U.S. factor premium in the DMS data, 7.7% per year from 1900 to 2025.
7. Valuation predicts long-horizon returns, not short-horizon ones. CAPE and dividend yield explain a meaningful share of 10-year return variation in-sample, but nearly zero of monthly returns, and most predictors fail out of sample (Goyal & Welch, 2008).
8. Wealth creation is extremely concentrated. About 4% of U.S. listed firms accounted for all net stock market wealth creation from 1926 to 2016, and most individual stocks underperformed Treasury bills over their lifetimes (Bessembinder, 2018).
9. Active management underperforms in aggregate, as arithmetic requires. In 2025, 79% of active U.S. large-cap funds trailed the S&P 500, the fourth-worst year in the 25-year history of the S&P SPIVA scorecard (S&P Dow Jones Indices, 2026).
10. Crashes come from leverage and liquidity, not only from valuation. The worst episodes (1929, 2008) combined overvaluation with credit stress. The fastest (1987, 2020, April 2025) were liquidity and policy shocks that recovered quickly.
11. U.S. market concentration is at a 100-year high. The ten largest S&P 500 stocks hold roughly 39% of the index as of September 2026, up from about 23% in 2019.
12. Valuations are near historic extremes. The CAPE ratio was about 41.6 on October 8, 2026, versus a long-run average near 17. Only the December 1999 peak of 44.2 was higher.
13. Rates have become a headwind. The 10-year Treasury yield reached as high as 5.36% in early October 2026, a 24-year high, and the Federal Reserve raised its policy rate to 3.75%–4.00% on September 16, 2026, its first hike since 2023.
14. Earnings are doing the heavy lifting. FactSet's estimate of third-quarter 2026 S&P 500 earnings growth stood at 29.6% on October 9, 2026, the third straight quarter above 25%. Hyperscaler AI capital spending guidance for 2026 is roughly $700–725 billion.
15. Market plumbing is changing fast. Retail investors account for an estimated 20%–35% of U.S. equity volume. Zero-days-to-expiry options were 59% of S&P 500 index option volume in 2025. The SEC has opened paths to near-24-hour and tokenized stock trading.

1.1 Why the Stock Market Matters

The stock market is the economy's main mechanism for pricing and transferring ownership of productive capital. It performs four functions that matter well beyond Wall Street. First, it lets firms raise equity capital to fund investment without the fixed obligations of debt. Second, it gives savers a way to share in the growth of the corporate sector, which is the core of most retirement systems in developed economies. Third, it produces prices that aggregate dispersed information about firms' prospects, which in turn guide managers' investment and hiring decisions. Fourth, it creates a market for corporate control, which disciplines management through the threat of takeover.

The scale is enormous. The U.S. equity market alone accounts for around 62% of world equity market value (DMS, 2026). Household wealth, pension solvency, university endowments, and sovereign wealth funds all depend on equity returns. Stock prices also feed back into the real economy through a wealth effect on consumption and through firms' cost of capital. Because of this, a sharp decline in stock prices can tighten financial conditions as effectively as a rise in interest rates.

1.2 The Master Question and Eight Research Questions

This paper asks one master question: what determines stock market returns and risk, how well do we understand it, and what does the evidence imply about the market as it stands today? It breaks this into eight research questions (RQs):

Table 1. Research questions and where they are addressed.

1.3 Method

This is a structured narrative survey with a current-conditions analysis. The search strategy and source-selection rules are documented in Appendix D. In brief, peer-reviewed journal articles and primary data sources are preferred. Working papers, central bank publications, and established industry research fill gaps. Financial press is used only for recent events and current figures, and each such figure carries an "as of" date. Where sources conflict on a current number, the paper reports the range and names the conflict rather than picking one silently.

The paper distinguishes four kinds of statements throughout. Historical facts are dated events and measured values. Empirical findings are statistical results from specific studies and samples. Theoretical predictions are implications of formal models. Interpretation is the author's judgment about what the evidence means, and it is flagged with phrases such as "this paper's assessment."

1.4 Contribution

Most surveys address one strand of this literature. This paper connects them. The microstructure literature explains how prices form trade by trade. Asset-pricing theory explains what expected returns should be. Empirical finance tests whether they are. Behavioral finance explains departures. Macro-finance links all of it to policy. Crises show what happens when the pieces fail together. The current-conditions section then tests whether this accumulated knowledge says anything useful about a specific, live market environment, which surveys rarely attempt.

1.5 Roadmap

Section 2 traces the historical development of equity markets. Section 3 covers market structure and microstructure. Section 4 presents asset-pricing theory. Section 5 examines the equity risk premium. Section 6 reviews cross-sectional factors and anomalies. Section 7 covers aggregate predictability and valuation. Section 8 covers behavioral finance. Section 9 covers macroeconomic and monetary drivers. Section 10 analyzes major crises. Section 11 covers participants, passive investing, and concentration. Section 12 covers volatility and derivatives. Section 13 covers research methods and their pitfalls. Section 14 covers regulation. Section 15 assesses the market as of October 2026. Section 16 synthesizes the findings. Sections 17 to 19 cover limitations, future research, and conclusions.

2.1 Origins: Joint-Stock Companies and the First Exchanges

Equity markets grew from a practical problem: long-distance trade required more capital, for longer, and at more risk than any single merchant could bear. The Dutch East India Company (Vereenigde Oostindische Compagnie, VOC), chartered in 1602, solved it by issuing transferable shares to the public. Because investors could not withdraw capital from the company directly, a secondary market emerged in Amsterdam where shares changed hands. That market developed forward contracts, options, and short selling within decades, along with the first recorded complaints about manipulation.

London followed. Share dealing moved from the coffee houses of Exchange Alley to a formal Stock Exchange in 1801. The South Sea Bubble of 1720, in which South Sea Company shares rose roughly tenfold within months and then collapsed, produced the Bubble Act and an early lesson that new financial technology and easy credit can feed speculation.

In New York, 24 brokers signed the Buttonwood Agreement in 1792, setting fixed commissions and agreeing to trade mainly with each other. The New York Stock & Exchange Board was formalized in 1817. For much of the 19th century, U.S. equity markets were dominated by railroad securities. Railroads were 63% of U.S. market value in 1900 and are less than 1% today (DMS, 2026).

2.2 The Regulatory State: 1929 to 1975

The 1929 crash and the Great Depression transformed U.S. market governance. The Securities Act of 1933 required registration and disclosure for new issues. The Securities Exchange Act of 1934 created the Securities and Exchange Commission (SEC), required periodic reporting, and prohibited manipulation. The Investment Company Act of 1940 regulated mutual funds. The underlying philosophy was disclosure rather than merit review: the government would not decide which securities were good, but it would require that investors receive the information to decide for themselves.

For four decades afterward, the structure was stable. The New York Stock Exchange (NYSE) dominated trading in listed stocks through a specialist system, in which one dealer per stock maintained the order book and was obliged to provide liquidity. Commissions were fixed. On May 1, 1975 ("May Day"), the SEC abolished fixed commissions. The same year, Congress passed the Securities Acts Amendments of 1975, which directed the SEC to foster a "national market system" linking competing venues. That mandate shaped the next 50 years of market-structure policy.

2.3 Electronic Trading, Decimalization, and Fragmentation

NASDAQ launched in 1971 as an electronic quotation system for over-the-counter stocks. In the 1990s, academic work revealed that NASDAQ dealers avoided odd-eighth quotes, which kept spreads artificially wide (Christie & Schultz, 1994). That finding led to the SEC's 1996 Order Handling Rules and an antitrust settlement. Electronic communication networks (ECNs) then gained share by matching orders automatically. Regulation ATS (1998) gave alternative trading systems a legal framework.

Decimalization in 2001 replaced fractional prices, typically sixteenths of a dollar, with pennies. Spreads for liquid stocks narrowed sharply. Regulation NMS, adopted in 2005 and implemented in 2007, introduced the Order Protection Rule (Rule 611), which requires venues to avoid trading through better-priced protected quotes elsewhere. Combined with automation, Reg NMS turned equity trading into a competition among many electronic venues linked by high-speed routing. NYSE converted to a hybrid electronic market and then became a for-profit public company.

2.4 High-Frequency Trading and the Speed Race

By the late 2000s, high-frequency trading (HFT) firms had largely replaced human specialists as liquidity providers. Estimates of the HFT share of U.S. equity volume varied widely by definition, but commonly exceeded half of volume in the early 2010s. The speed race led to co-location of servers in exchange data centers, microwave links between Chicago and New York, and a debate about whether nanosecond advantages create social value. Section 3 reviews the evidence.

2.5 The Index Fund Revolution

John Bogle's Vanguard launched the First Index Investment Trust in 1976, following the academic case that most active managers could not beat the market after costs (Jensen, 1968). The first U.S.-listed exchange-traded fund (ETF), the SPDR S&P 500 Trust, launched in January 1993. Index funds and ETFs grew slowly at first and then rapidly. Passive U.S. equity funds overtook active U.S. equity funds in assets in August 2019, and passive funds overtook active funds across all U.S. fund categories at the end of 2023, $13.3 trillion to $13.2 trillion (Morningstar data, as reported by Hargreaves Lansdown, 2024).

2.6 Settlement, Hours, and Tokenization

Settlement cycles shortened from five business days (T+5) to T+3 in 1995, T+2 in 2017, and T+1 on May 28, 2024. Each step reduced counterparty exposure in the clearing system. The GameStop episode of January 2021, when clearing-margin calls forced some brokers to restrict buying, accelerated the move to T+1.

Trading hours are also expanding. The SEC approved applications for extended overnight trading sessions from several venues, and in September 2026 issued a temporary five-year exemptive order allowing qualifying venues to trade tokenized representations of U.S. stocks, provided the tokens represent actual ownership of the underlying shares (SEC, 2026, as reported in trade press). Section 14 discusses these changes.

Table 2. Timeline of equity market development.

3.1 Where Stocks Trade

A U.S. stock can trade on any of roughly 16 registered exchanges, on dozens of alternative trading systems (ATSs, including "dark pools"), or off-exchange through wholesalers that internalize retail orders. Off-exchange trading regularly accounts for close to half of U.S. equity volume. The fragmentation is a direct result of the national market system design: venues compete on fees, speed, and order types, and Reg NMS links them.

Table 3. U.S. equity trading venues.

The closing auction has become especially important. Index funds that track benchmarks priced at the close concentrate their trading there, and a large share of daily volume in many stocks now executes in the final auction.

3.2 Order Types and the Limit Order Book

The two basic order types are market orders, which demand immediate execution at the best available price, and limit orders, which specify a worst acceptable price and may wait. Limit orders supply liquidity and market orders consume it. The collection of outstanding limit orders forms the limit order book. The best bid and best offer across all venues form the National Best Bid and Offer (NBBO). Exchanges also offer hundreds of specialized order types, such as pegged orders, post-only orders, and hidden orders, which matter mostly to professional traders.

3.3 The Bid-Ask Spread and Its Components

The bid-ask spread is the price of immediacy. Microstructure theory decomposes it into three parts.

Order-processing costs cover the fixed costs of operating as a market maker.

Inventory costs compensate a dealer for holding unwanted risk. A dealer who buys more than they sell must carry the position until it can be offloaded and will widen quotes or shade prices to manage that risk.

Adverse selection costs arise because some traders know more than the market maker. Glosten and Milgrom (1985) showed that even a competitive, risk-neutral market maker must set a positive spread when some counterparties are informed, because trades against them lose money on average. The spread is the market maker's charge to uninformed traders to cover losses to informed ones. Kyle (1985) modeled a single informed trader who spreads trades over time to hide information. In Kyle's model, price impact is linear in order flow:

where \(\Delta p\) is the price change, \(x\) is the informed trader's order, \(u\) is noise-trader order flow, and \(\lambda\) ("Kyle's lambda") measures market depth. A smaller \(\lambda\) means a deeper, more liquid market.

Roll (1984) showed that the spread can be inferred from the negative autocorrelation of price changes caused by "bid-ask bounce":

where \(S\) is the effective spread and \(\Delta p_t\) is the price change in period \(t\). The estimator is undefined when the covariance is positive, which often happens in practice.

3.4 Measuring Liquidity

Liquidity has several dimensions: tightness (cost of a round trip), depth (quantity available near the best price), resilience (how quickly prices recover after a trade), and immediacy. No single measure captures all of them.

Table 4. Liquidity measures. \(A\) is the ask, \(B\) the bid, \(M\) the midpoint, \(P\) the trade price, \(D\) the trade direction (+1 buy, −1 sell), \(r_d\) the daily return, \(\text{DVol}_d\) the daily dollar volume, and \(D\) in the Amihud formula the number of days.

Liquidity is priced. Amihud and Mendelson (1986) found that stocks with wider spreads earn higher average returns, compensating holders for trading costs. Amihud (2002) showed that market-wide illiquidity predicts higher future returns, and Pástor and Stambaugh (2003) found that stocks whose returns are more sensitive to aggregate liquidity shocks earn a premium.

3.5 High-Frequency Trading: Evidence For and Against

The evidence on HFT is mixed and depends on what is measured.

Evidence that HFT improves market quality. Hendershott, Jones, and Menkveld (2011) used the 2003 introduction of automatic quote dissemination on the NYSE as a natural experiment and found that algorithmic trading narrowed spreads and made quotes more informative. Brogaard, Hendershott, and Riordan (2014) found that HFTs, on average, trade in the direction of permanent price changes and against transitory pricing errors, which improves price efficiency. Retail transaction costs for liquid stocks fell to historic lows during the HFT era.

Evidence of costs. Budish, Cramton, and Shim (2015) argued that continuous limit order books create an arms race. Because prices of correlated securities move together, any public news creates a mechanical arbitrage opportunity for the fastest trader, which leads to "sniping" of stale quotes. Liquidity providers must widen spreads to cover sniping losses. They proposed frequent batch auctions, which would make speed beyond a threshold worthless. HFT liquidity can also vanish in stress, when it is most needed.

The Flash Crash. On May 6, 2010, U.S. indices fell roughly 9% intraday within minutes and largely recovered within about 20 minutes. Some individual stocks traded at a penny and others at $100,000. Kirilenko, Kyle, Samadi, and Tuzun (2017) found that HFTs did not trigger the crash, which began with a large automated sell program in E-mini S&P 500 futures, but that they amplified it by rapidly trading inventory among themselves ("hot potato" trading) and withdrawing liquidity. The episode led directly to single-stock circuit breakers and the limit-up/limit-down mechanism (Section 14).

3.6 Payment for Order Flow and Retail Execution

Most U.S. retail brokers route customer orders to wholesale market makers, which pay the brokers for the flow. This "payment for order flow" (PFOF) funds zero-commission trading. Wholesalers can profit because retail orders are, on average, uninformed, so trading against them carries little adverse selection. Supporters argue that retail investors receive price improvement over the NBBO and pay no commissions. Critics argue that the arrangement creates a conflict of interest and keeps retail orders out of lit markets where they could improve public price discovery. The SEC proposed an order-competition rule and a Regulation Best Execution in 2022, but formally withdrew both, along with 12 other pending proposals, on June 12, 2025 (Paul Hastings, 2025). PFOF remains legal in the United States and has been banned in the European Union's revised market rules.

3.7 Maker-Taker Fees and Tick Sizes

Most exchanges pay rebates to limit orders that add liquidity and charge fees to orders that remove it. Critics argue that the rebates distort broker routing toward venues that pay the most rather than those that execute best (Battalio, Corwin, & Jennings, 2016). In September 2024, the SEC amended Regulation NMS to introduce a half-penny tick size for stocks whose spreads are consistently tight, to lower the access-fee cap, and to redefine round lots. Compliance dates for the tick and fee changes were pushed back to November 2, 2026, by an October 2025 exemptive order. Industry groups have sought further delay, and at least one trade report states a further extension to November 2027. This paper could not confirm that report against an SEC order, so the timing should be treated as unsettled.

3.8 Assessment

Markets for large U.S. stocks are more liquid and cheaper to trade than at any time in history for ordinary trade sizes. The costs of modern market structure show up in rare stress episodes, in the complexity of the system, and in the share of trading that never reaches public order books. This paper's assessment is that the main open question is no longer whether electronic markets lower average costs. They do. The open question is whether a fragmented, speed-driven structure is fragile in tail events, and how much price discovery is lost when retail order flow is internalized.

4.1 Present Value: The Starting Point

Every model of stock prices starts from the same identity. A share is a claim on future cash flows, so its price equals the discounted value of what it will pay:

where \(P_t\) is the price at time \(t\), \(E_t[D_{t+k}]\) is the expected dividend (or, more generally, cash distributed to shareholders) in period \(t+k\), and \(r\) is the discount rate, which is the expected return investors require. If dividends grow at a constant rate \(g < r\), the sum collapses to the Gordon (1962) growth model:

The second form is the most useful equation in practical equity analysis. It says the expected return equals the dividend yield plus expected growth. It also reveals why stock prices are volatile: when \(r - g\) is small, small changes in either \(r\) or \(g\) cause large changes in price. Equities behave like very long-duration assets.

All of asset pricing is a debate about \(r\). Why is it what it is, does it vary over time, and why does it differ across stocks?

4.2 The Efficient Market Hypothesis

Fama (1970) defined an efficient market as one in which prices "fully reflect" available information, and distinguished three forms by the information set:

• Weak form. Prices reflect all information in past prices and volume. Technical analysis cannot earn abnormal returns.
• Semi-strong form. Prices reflect all public information. Fundamental analysis of public data cannot earn abnormal returns.
• Strong form. Prices reflect all information, including private information. Even insiders cannot profit.

The evidence broadly supports the weak and semi-strong forms in the sense that prices react quickly to news (Fama, Fisher, Jensen, & Roll, 1969) and most professional managers do not beat the market after fees (Jensen, 1968; Fama & French, 2010). The strong form fails, since insider trading is profitable, which is why it is illegal.

Two conceptual problems complicate any test. First, the joint hypothesis problem: market efficiency can be tested only together with a model of what expected returns should be. If a strategy earns "abnormal" returns, either the market is inefficient or the model of normal returns is wrong. The data cannot distinguish the two by themselves (Fama, 1991). Second, the Grossman-Stiglitz paradox (Grossman & Stiglitz, 1980): if prices were perfectly efficient, no one would pay to gather information, but then prices could not reflect information. Markets must therefore be inefficient enough to compensate the people who make them efficient. This implies an "equilibrium degree of disequilibrium," which is a more realistic target than perfect efficiency.

Lo (2004) proposed the Adaptive Markets Hypothesis as a bridge between efficiency and behavioral views. In this framework, market efficiency is not a fixed property. It varies with the number and type of competitors, the opportunities available, and how quickly participants adapt. Profitable strategies attract capital until they decay, and new ones appear as conditions change. Section 6 shows that this description fits the life cycle of many anomalies well.

4.3 The Capital Asset Pricing Model

The Capital Asset Pricing Model (CAPM), developed independently by Sharpe (1964), Lintner (1965), and Mossin (1966) from Markowitz's (1952) portfolio theory, was the first equilibrium model of expected returns. Its assumptions are strong: investors care only about the mean and variance of portfolio returns, share the same beliefs, and can borrow and lend at a risk-free rate. Under these assumptions, every investor holds the same portfolio of risky assets, the market portfolio, combined with borrowing or lending. In equilibrium:

where \(E[R_i]\) is the expected return on asset \(i\), \(R_f\) is the risk-free rate, \(E[R_m]\) is the expected market return, and \(\beta_i\) measures the asset's sensitivity to market movements. The core insight is that only systematic risk, the part correlated with the market, earns a premium. Idiosyncratic risk can be diversified away at no cost, so it is not rewarded.

Empirical failures. Early tests found the relation between beta and average returns was positive but too flat. Low-beta stocks earned more than the CAPM predicts, and high-beta stocks earned less (Black, Jensen, & Scholes, 1972). Fama and French (1992) found that, once size and book-to-market are controlled for, beta has almost no power to explain the cross-section of U.S. returns from 1963 to 1990. Roll (1977) argued that the CAPM is fundamentally untestable because the true market portfolio includes all wealth, including human capital and private assets, which no index captures. Black (1972) showed that if investors face borrowing constraints, the security market line is flatter, which Frazzini and Pedersen (2014) later developed into the betting-against-beta factor.

The CAPM remains the most widely used model in practice. Graham and Harvey (2001) found that about three-quarters of U.S. CFOs use it to estimate the cost of equity. This persistence, despite weak empirical support, reflects its simplicity and the lack of a consensus replacement.

4.4 Arbitrage Pricing Theory and the Intertemporal CAPM

Ross (1976) derived a multifactor model from a no-arbitrage argument instead of equilibrium. If returns are driven by a small number of common factors plus idiosyncratic noise, then in a large market:

where \(\beta_{ik}\) is the exposure of asset \(i\) to factor \(k\) and \(\lambda_k\) is the risk premium for factor \(k\). The Arbitrage Pricing Theory (APT) does not say what the factors are. Chen, Roll, and Ross (1986) proposed macroeconomic factors: industrial production growth, changes in expected inflation, unexpected inflation, the default spread, and the term spread.

Merton (1973) developed the Intertemporal CAPM (ICAPM), in which long-lived investors also hedge changes in future investment opportunities. Assets that pay off when investment opportunities deteriorate are valuable hedges and earn lower expected returns. The ICAPM gives a theoretical justification for multifactor models, though, like the APT, it does not specify which state variables matter.

4.5 The Stochastic Discount Factor

Modern asset pricing unifies all of these models with one equation (Cochrane, 2005). For any asset with gross return \(R_{t+1}\):

where \(m_{t+1}\) is the stochastic discount factor (SDF), also called the pricing kernel. The SDF is high in bad states of the world, when an extra dollar is most valuable, and low in good states. Rearranging gives:

An asset earns a risk premium if it pays off poorly when \(m\) is high, that is, in bad times. This is the deepest statement in asset pricing: risk premia compensate for poor performance in bad times, not for volatility as such. Every asset-pricing model is a claim about what \(m\) is. In the CAPM, \(m\) is linear in the market return. In factor models, \(m\) is linear in the factors.

In the consumption-based CAPM (Breeden, 1979; Lucas, 1978), \(m\) is the marginal rate of substitution of a representative investor. With power utility and relative risk aversion \(\gamma\):

where \(\delta\) is the subjective time discount factor and \(C_t\) is consumption. Hansen and Jagannathan (1991) derived a lower bound on the volatility of any valid SDF:

The right-hand side is the Sharpe ratio. With a U.S. equity Sharpe ratio around 0.4 to 0.5 on annual data, the SDF must be highly volatile. Consumption growth is smooth, with a standard deviation of only about 1% to 2% per year, so power utility can only match the bound with implausibly high risk aversion.

4.6 The Equity Premium Puzzle and Its Proposed Resolutions

Mehra and Prescott (1985) formalized the problem. Using U.S. data from 1889 to 1978, they observed an average equity premium over short-term bills of about 6%. A standard consumption-based model with reasonable risk aversion could produce a premium of at most about 0.35%. Raising risk aversion enough to match the premium generates a risk-free rate far higher than observed, which is the companion risk-free rate puzzle (Weil, 1989).

Four major families of resolution have been proposed:

1. Habit formation. Campbell and Cochrane (1999) let utility depend on consumption relative to a slow-moving habit level. When consumption falls toward habit, as in recessions, effective risk aversion rises sharply. The model produces a large, time-varying premium and explains why valuation ratios predict returns.
2. Long-run risk. Bansal and Yaron (2004) add a small, persistent component to consumption growth and assume investors with Epstein-Zin preferences care about the timing of uncertainty resolution. Small shifts in long-run growth expectations then have large price effects.
3. Rare disasters. Rietz (1988) and Barro (2006) argue that the premium compensates for the small probability of catastrophic declines in consumption, such as wars, depressions, and expropriations. Barro calibrated disaster probabilities from 20th-century history and found they can account for the observed premium.
4. Behavioral and institutional explanations. Benartzi and Thaler (1995) propose "myopic loss aversion": investors who feel losses more than gains and evaluate portfolios frequently demand a high premium to hold a volatile asset. Survivorship bias also matters. The U.S. market is the most successful in the world over the 20th century, so U.S. data overstate the premium an investor could have expected ex ante (Section 5).

This paper's assessment is that each explanation captures something real, none is accepted as complete, and the puzzle remains one of the defining open problems in financial economics.

4.7 From the CAPM to Multifactor Models

Empirical work since the 1980s has produced a sequence of factor models, each adding factors that explain returns the previous model missed:

Fama-French three-factor model (Fama & French, 1993):

where MKT is the market excess return, SMB ("small minus big") is the return of small stocks minus large stocks, HML ("high minus low") is the return of high book-to-market (value) stocks minus low book-to-market (growth) stocks, \(\alpha_i\) is the abnormal return, and \(\varepsilon_{it}\) is an error term.

Carhart four-factor model (Carhart, 1997) adds momentum, UMD ("up minus down"), the return of recent winners minus recent losers.

Fama-French five-factor model (Fama & French, 2015) adds RMW ("robust minus weak"), the return of high-profitability firms minus low-profitability firms, and CMA ("conservative minus aggressive"), the return of low-investment firms minus high-investment firms:

A striking result is that HML becomes largely redundant once RMW and CMA are included in the 1963–2013 U.S. sample, since value returns are mostly explained by the other factors.

q-factor model (Hou, Xue, & Zhang, 2015) derives similar factors from investment theory. Firms invest more when their cost of capital is low, so low-investment firms should earn higher expected returns. Firms with high expected profitability relative to investment must also have high discount rates. The model uses market, size, investment, and return-on-equity factors and performs comparably to or better than the five-factor model in head-to-head tests.

A central interpretive question runs through all of this work. Are factor premia compensation for risk, consistent with efficient markets and the SDF framework, or are they evidence of systematic mispricing? Section 6 takes up that question factor by factor.

5.1 Definition

The equity risk premium (ERP) is the expected return on stocks in excess of a risk-free rate. It is the single most important number in finance. It drives retirement planning, pension funding assumptions, corporate cost of capital, and the relative pricing of stocks and bonds. Because expected returns cannot be observed directly, the ERP must be estimated, and different methods give different answers. Three approaches dominate: historical averages, forward-looking models implied by current prices, and surveys.

5.2 Historical Estimates

The longest high-quality global data come from Dimson, Marsh, and Staunton, published annually in the UBS Global Investment Returns Yearbook. The 2026 edition covers 35 markets from 1900 to 2025.

Table 5. Long-run U.S. returns. Source: DMS (2026), UBS Global Investment Returns Yearbook 2026, Public Summary Edition.

From these geometric averages, the realized U.S. premium over bills is approximately \(1.066/1.005 - 1 \approx 6.1\%\) per year, and over bonds approximately \(1.066/1.016 - 1 \approx 4.9\%\) per year. These are this paper's calculations from the DMS figures.

Several broader results from the same dataset matter:

• Equities outperformed bonds, bills, and inflation in all 21 countries with continuous histories since 1900.
• Developed markets returned 8.5% per year in U.S. dollars from 1900 to 2025, versus 6.9% for emerging markets. From 1960 to 2025, emerging markets returned 10.9% versus 9.6% for developed markets.
• Equities and bonds have each, on several occasions, lost more than 70% in real terms. A 60/40 equity-bond portfolio has never lost more than 50% in real terms.
• U.S. inflation averaged 2.9% per year since 1900, the third-lowest of the 35 markets.

5.3 Arithmetic Versus Geometric Means

The choice of average matters. The geometric mean is the compound growth rate actually experienced. The arithmetic mean is the average of annual returns and is the correct input for a one-period expected return. For a lognormal return with volatility \(\sigma\), the two are related approximately by:

With U.S. equity volatility around 20% per year, the arithmetic mean exceeds the geometric mean by about 2 percentage points. Much confusion in ERP debates arises from comparing an arithmetic estimate from one source with a geometric estimate from another.

5.4 Survivorship and the "Triumph of the Optimists"

The U.S. market is the most successful major market of the past 125 years. Its share of world equity value rose from roughly 15% in 1900 to around 62% today (DMS, 2026). Researchers using U.S. data are, in effect, studying the winner. Dimson, Marsh, and Staunton (2002) titled their foundational book Triumph of the Optimists to make this point. Markets such as Russia in 1917 and China in 1949 went to zero for domestic investors, and others, including Germany and Japan, suffered near-total losses in war. A world-portfolio premium is therefore a more honest estimate of what investors should expect than a U.S.-only premium. Jorion and Goetzmann (1999) estimated that, across 39 markets over the 20th century, real capital appreciation outside the U.S. was substantially lower than in the U.S., and they argued that survivorship bias inflates U.S.-based estimates.

5.5 Forward-Looking Estimates

Forward-looking methods ask what premium is implied by today's prices and reasonable cash flow forecasts.

Dividend and earnings growth models. Fama and French (2002) estimated the 1951–2000 premium using the Gordon framework. Dividend growth implied a premium of about 2.55% per year and earnings growth about 4.32%, both far below the realized 7.43% (arithmetic). They concluded that much of the high realized return in the late 20th century came from an unexpected decline in discount rates, which pushed up valuations. That windfall is not repeatable.

Implied ERP. Damodaran solves for the discount rate that equates the S&P 500 level with the present value of expected cash returned to shareholders through dividends and buybacks, then subtracts the Treasury yield. His implied ERP has been unusually stable in 2026:

Table 6. Damodaran implied U.S. equity risk premium, 2026. Source: Damodaran Online (accessed October 2026).

Damodaran also changed his methodology in July 2026. After Moody's downgrade of U.S. sovereign debt in May 2025, he no longer treats the U.S. Treasury rate as fully default-free and adjusts the mature-market ERP for the U.S. default spread.

The stability of the implied ERP is notable given large moves in yields. Its rough constancy means that rising Treasury yields have been offset by rising expected cash flows, especially from strong earnings growth, rather than by falling prices.

Surveys. Welch (2000) surveyed academic financial economists and found a consensus 30-year arithmetic ERP of about 7%. Later surveys produced lower numbers. Graham and Harvey's long-running survey of U.S. CFOs has generally found expected 10-year premiums in the range of about 3% to 5% (approx.). Surveys measure beliefs, which may differ from rational expectations, but they matter because beliefs drive actual capital allocation.

5.6 Comparing Estimates

Table 7. Equity risk premium estimates by method.

5.7 What the Current ERP Implies

Two readings of the current environment conflict. The implied ERP of about 4.2% is close to its long-run average in Damodaran's series, which suggests stocks are fairly priced relative to bonds once expected cash flow growth is included. Simple valuation ratios point the other way. The CAPE earnings yield, the inverse of CAPE, was about 2.4% in October 2026 (1/41.6), below a 10-year Treasury yield of roughly 5.1% to 5.3%. By that cruder comparison, stocks offer no premium at all over nominal bonds on trailing, cyclically adjusted earnings.

The two views reconcile only if earnings grow substantially faster than history over the next decade. The implied ERP builds that growth in through analyst forecasts. CAPE does not. This paper's assessment is that the realized premium over the next decade depends heavily on whether the current wave of AI-related earnings growth persists, which is the central question of Section 15.

6.1 The Core Question: Risk or Mispricing?

A cross-sectional "anomaly" is a firm characteristic that predicts returns after adjusting for a benchmark model, usually the CAPM. Every anomaly admits two interpretations. In the risk-based view, the characteristic proxies for exposure to some undiversifiable risk that the benchmark model misses, so the higher return is fair compensation. In the mispricing view, investors systematically misjudge firms with the characteristic, and limits to arbitrage stop sophisticated traders from correcting the error. The distinction matters. A risk premium should persist and carries genuine bad-times exposure. A mispricing should shrink as it becomes known and should be largest where arbitrage is hardest.

6.2 The Major Factors

Size. Banz (1981) found that small firms earned higher returns than large firms after adjusting for beta. The size effect was among the first anomalies, and also among the first to weaken: it was largely absent in U.S. data from the early 1980s through the 2010s. It is concentrated in the very smallest, least liquid stocks and in January. Asness, Frazzini, Israel, Moskowitz, and Pedersen (2018) argued that the size premium re-emerges once one controls for quality, because small stocks are disproportionately "junk." In the DMS long-run data, the U.S. size premium is modest, and in 2025 small caps lagged large caps by about 12 percentage points (S&P Dow Jones Indices, 2026).

Value. Basu (1977) found that low price-to-earnings stocks outperformed. Fama and French (1992) showed that book-to-market subsumed many related measures. Value stocks have high book value relative to price. The risk story holds that value firms are distressed and inflexible, so they suffer more in bad times (Zhang, 2005). The behavioral story holds that investors extrapolate past growth too far, overpricing glamour stocks and underpricing value stocks (Lakonishok, Shleifer, & Vishny, 1994). Value suffered a severe drawdown from 2007 to 2020, the longest in its recorded history. Arnott, Harvey, Kalesnik, and Linnainmaa (2021) attributed much of the decline to a widening valuation spread between value and growth rather than to deteriorating fundamentals, and noted that book value increasingly misses intangible capital such as research, software, and brands.

Momentum. Jegadeesh and Titman (1993) found that stocks with high returns over the past 3 to 12 months continue to outperform over the next 3 to 12 months. Momentum is the most persistent anomaly. DMS (2026) report a long-run U.S. momentum premium of 7.7% per year from 1900 to 2025, and momentum ranked first among factors in 13 of 24 country-decades they examine. Asness, Moskowitz, and Pedersen (2013) documented momentum in eight asset classes, including bonds, currencies, and commodities. Momentum is hard to explain with risk. Its leading explanations are behavioral: investors underreact to news, then overreact as trends persist (Barberis, Shleifer, & Vishny, 1998; Daniel, Hirshleifer, & Subrahmanyam, 1998). Momentum also has a distinctive risk: rare, violent crashes, usually during market rebounds after bear markets, when beaten-down losers surge. Daniel and Moskowitz (2016) documented crashes in 1932 and 2009, the latter costing a momentum portfolio more than 70% in a few months.

Profitability. Novy-Marx (2013) showed that gross profits scaled by assets predict returns about as strongly as book-to-market, and that profitability and value are negatively correlated, so combining them is especially effective. Fama and French (2015) incorporated profitability as the RMW factor. Profitable firms earning higher returns contradicts the intuition that safer firms should earn less, which makes a pure risk interpretation awkward.

Investment. Firms that grow their assets rapidly earn lower subsequent returns (Cooper, Gulen, & Schill, 2008; Titman, Wei, & Xie, 2004). The q-theory explanation holds that firms invest more when their cost of capital is low, so high investment signals low expected returns. The behavioral explanation holds that managers overinvest when their stock is overpriced, and the market fails to anticipate the subsequent disappointment.

Low volatility and betting against beta. Low-beta and low-volatility stocks have earned higher risk-adjusted returns than the CAPM predicts. Ang, Hodrick, Xing, and Zhang (2006) found that stocks with high idiosyncratic volatility earn abnormally low returns. Frazzini and Pedersen (2014) attributed the flat security market line to leverage constraints: investors who cannot borrow buy high-beta stocks to raise expected returns, bidding up their prices. A betting-against-beta (BAB) factor that is long leveraged low-beta stocks and short high-beta stocks earned significant returns across 20 international markets.

Quality. Asness, Frazzini, and Pedersen (2019) defined "quality" as profitable, growing, safe, and well-managed firms, and found that a quality-minus-junk factor earns significant risk-adjusted returns. Investors do pay more for quality, but not as much as they should.

Short-term reversal. Stocks with high returns over the past month tend to underperform the next month (Jegadeesh, 1990; Lehmann, 1990). The effect reflects liquidity provision: traders who absorb temporary order imbalances earn the reversal. It is expensive to trade because it requires high turnover.

Accruals. Sloan (1996) found that firms whose earnings consist largely of accruals, rather than cash, earn lower subsequent returns. Investors appear to treat accrual-based earnings as being as persistent as cash earnings, when in fact they reverse more.

Post-earnings-announcement drift. Ball and Brown (1968) and Bernard and Thomas (1989) showed that stock prices continue to drift in the direction of an earnings surprise for weeks or months after the announcement. This is a direct challenge to semi-strong efficiency, since the information is public. The effect has weakened as trading costs fell and algorithmic traders exploited it.

Table 8. Summary of the major equity factors.

6.3 The Factor Zoo

By the 2010s, researchers had published hundreds of return predictors. Cochrane (2011) called this the "factor zoo." Harvey, Liu, and Zhu (2016) catalogued 316 factors published in leading journals and argued that, because so many hypotheses had been tested on the same data, the conventional significance threshold of a t-statistic of 2.0 was far too low. Accounting for multiple testing, they proposed that a new factor should clear a t-statistic of about 3.0, and concluded that many published factors are likely false discoveries.

The underlying problem is simple. If researchers test 100 random signals at the 5% significance level, about five will appear significant by chance. Publication favors significant results, so the published literature over-represents lucky findings.

6.4 Post-Publication Decay

McLean and Pontiff (2016) studied 97 anomalies and compared their returns in three periods: the original sample, the period after the sample but before publication, and the period after publication. Returns fell about 26% out of sample and about 58% after publication. The 26% decline is an estimate of statistical bias, such as data mining. The additional decline after publication reflects investors learning about the anomaly and trading on it. Decay was larger for anomalies that are cheaper to arbitrage, which supports a mispricing explanation for at least part of the original returns.

6.5 The Replication Debate

Two influential studies reached very different conclusions about the reliability of the anomaly literature.

Hou, Xue, and Zhang (2020) replicated 452 anomalies using consistent methods. They excluded microcaps by using NYSE breakpoints and value-weighted returns, and they found that 65% of anomalies failed to reach a t-statistic of 1.96. With a multiple-testing threshold of 2.78, 82% failed. Many anomalies exist mainly in tiny stocks that are expensive and impractical to trade.

Jensen, Kelly, and Pedersen (2023) asked whether finance has a replication crisis and answered largely no. They studied 153 factors in 93 countries using a Bayesian framework that accounts for the fact that factors cluster into related themes. They found that most factors replicate, that results hold internationally, and that multiple-testing corrections are less damaging once the correlation among factors is modeled properly. The factors cluster into a small number of themes, such as value, momentum, quality, investment, and low risk, and the themes are robust even when individual signals are not.

Chen and Zimmermann (2022) provide open-source code that reproduces about 300 predictors and found that the vast majority of clearly significant original results replicate when the original methods are followed.

The two camps differ less than their headlines suggest. Both agree that microcap-driven signals are unreliable, that individual signals overstate the evidence, and that broad themes are more robust than specific variables. This paper's assessment is that a handful of factor themes are genuine, persistent features of returns, while the long tail of published anomalies is largely noise, microcap effects, or arbitraged away.

6.6 Machine Learning and the Cross-Section

Gu, Kelly, and Xiu (2020) applied machine learning methods, including penalized regression, random forests, and neural networks, to a large set of predictors. Neural networks and tree-based methods roughly doubled the out-of-sample predictive performance of linear models, with gains concentrated in nonlinear interactions among characteristics. Kelly, Pruitt, and Su (2019) developed instrumented principal component analysis (IPCA), in which factor loadings depend on characteristics, and found that a small number of latent factors explains the cross-section better than standard observable factor models.

Three cautions apply. First, monthly out-of-sample predictive R² values for individual stocks remain below 1%. Second, gains are largest in small, illiquid stocks, where trading costs are highest. Third, machine learning methods are especially vulnerable to overfitting when the researcher tries many specifications (Section 13).

7.1 From Random Walk to Time-Varying Expected Returns

Early efficient-markets research assumed that expected returns were roughly constant, so returns should be unpredictable. Shiller (1981) challenged this with a variance-bounds test. If prices equal the present value of future dividends discounted at a constant rate, prices should be less volatile than the subsequent realized dividends. In U.S. data, prices were five to thirteen times too volatile. Either markets were irrational or discount rates varied over time.

The profession eventually accepted the second interpretation, though it remains debated whether discount-rate variation reflects rational risk premia or sentiment. Cochrane (2011) summarized the shift: the variation in valuation ratios corresponds almost entirely to variation in expected returns, not expected cash flow growth.

7.2 The Campbell-Shiller Decomposition

Campbell and Shiller (1988) derived a log-linear approximation linking the dividend-price ratio to future returns and dividend growth:

where \(dp_t = \log(D_t / P_t)\) is the log dividend-price ratio, \(r\) is the log return, \(\Delta d\) is log dividend growth, and \(\rho\) is a constant slightly below 1, approximately \(1/(1 + D/P)\). The identity implies that a high dividend yield must be followed by high returns, low dividend growth, or both. It cannot be followed by neither. Empirically, U.S. dividend yields forecast returns but forecast dividend growth very weakly. Cochrane (2008) called this "the dog that did not bark": if the dividend yield did not forecast returns, it would have to forecast dividend growth, and it does not.

7.3 Predictive Variables

The standard predictive regression is:

where \(r_{t \to t+h}\) is the cumulative return from \(t\) to \(t+h\), \(x_t\) is a predictor, and \(\varepsilon\) is the forecast error.

Table 9. Return predictors and reported explanatory power. R² values are approximate summaries of published results and differ across samples and specifications.

7.4 Why Long-Horizon R² Values Rise, and Why That Can Mislead

R² values rise with the forecast horizon because valuation ratios are highly persistent while short-term returns are noisy. When a slow-moving predictor forecasts a small expected return component each period, that component accumulates across years while noise partially averages out. This is a real effect, and Cochrane (2008) showed that long-horizon results follow mechanically from short-horizon ones.

Two statistical problems inflate the evidence, however. First, overlapping observations: monthly data on 10-year returns contain only about a dozen independent 10-year periods since 1900, so standard errors must be adjusted, and the effective sample is small. Second, Stambaugh bias (Stambaugh, 1999): when the predictor is persistent and its innovations are correlated with return innovations, as is true for valuation ratios, the ordinary least squares slope estimate is biased in finite samples. The bias is approximately:

where \(\phi\) is the autoregressive coefficient of the predictor, \(\sigma_{u\varepsilon}\) is the covariance between predictor and return innovations, and \(\sigma_u^2\) is the variance of predictor innovations. Because \(\hat{\phi}\) is biased downward for persistent series and the covariance is negative for valuation ratios, the slope estimate is biased upward. This makes predictability look stronger than it is.

7.5 In-Sample Versus Out-of-Sample

Goyal and Welch (2008) tested a comprehensive set of predictors and found that almost none beat the historical average return out of sample. Most had negative out-of-sample R² values, and most of their in-sample power came from a few periods, especially the 1973–1975 oil shock. They concluded that the models would not have helped a real-time investor.

Campbell and Thompson (2008) responded that simple, economically motivated restrictions, such as setting forecasts to zero when the predicted premium is negative and constraining slope signs, produce positive out-of-sample R² values. They are small, around 0.5% per month, but Campbell and Thompson showed that even this level of predictability can increase a mean-variance investor's returns meaningfully, because monthly returns are so noisy that small forecast improvements are valuable.

This paper's assessment is that both papers are correct. Valuation has real information about long-horizon expected returns, but the relation is too noisy and unstable to support confident market timing at short horizons.

7.6 Valuation in October 2026

The cyclically adjusted price-to-earnings (CAPE) ratio divides the price by the 10-year average of inflation-adjusted earnings. The averaging smooths the business cycle.

Table 10. CAPE readings. Current values are daily estimates from data trackers and are subject to revision.

At about 41.6, the CAPE is the second-highest reading in more than 140 years of data. Historically, starting CAPE levels in the high 30s and above have been followed by low and sometimes negative real returns over the following decade. The 2000 starting point, at a CAPE of about 44, was followed by negative real S&P 500 returns over 2000–2010.

Several arguments suggest CAPE overstates current overvaluation.

1. Accounting changes. Since the 2000s, accounting rules require immediate write-downs of goodwill and expensing of intangible investment, depressing reported earnings relative to economic earnings. Siegel (2016) argued that this biases CAPE upward relative to history, and that alternative earnings measures such as national-income profits produce lower readings.
2. Payout policy. Firms now return more cash through buybacks and retain less, which changes the relation between earnings growth and prices.
3. Composition. The index today is dominated by high-margin technology and software firms with lower capital intensity than the industrial firms of earlier decades.
4. Interest rates. For much of the 2010s, low real rates justified higher multiples. This argument has weakened sharply in 2026 as the 10-year yield has risen above 5%.

The counterargument is that these adjustments explain some of the gap but not all. Even under generous adjustments, the U.S. market is expensive by historical standards. This paper's assessment is that the evidence implies below-average expected real returns over a 10-year horizon with wide uncertainty. It does not imply that a decline is imminent. High CAPE readings persisted for years in the late 1990s and from 2017 onward. As a timing tool, CAPE has a poor record.

8.1 Two Pillars

Behavioral finance rests on two pillars (Barberis & Thaler, 2003). The first is investor psychology: real investors depart from rational expectations and expected-utility maximization in systematic, predictable ways. The second is limits to arbitrage: rational traders cannot always correct the resulting mispricings, because arbitrage is costly and risky. Both are needed. Psychological biases alone do not move prices if arbitrageurs can trade against them freely. Limits to arbitrage alone produce no mispricing unless someone is making errors.

8.2 Prospect Theory and Loss Aversion

Kahneman and Tversky (1979) proposed prospect theory as a descriptive model of choice under risk. Its key features are:

• Reference dependence. People evaluate outcomes as gains and losses relative to a reference point, often the purchase price, rather than as final wealth.
• Loss aversion. Losses hurt roughly twice as much as equivalent gains feel good.
• Diminishing sensitivity. The value function is concave for gains, producing risk aversion, and convex for losses, producing risk seeking.
• Probability weighting. People overweight small probabilities and underweight moderate and large ones.

The value function is commonly written as:

where \(x\) is a gain or loss, \(\alpha\) and \(\beta\) are curvature parameters, and \(\lambda > 1\) is the coefficient of loss aversion. Tversky and Kahneman (1992) estimated \(\alpha \approx \beta \approx 0.88\) and \(\lambda \approx 2.25\).

Probability weighting has direct market implications. Investors who overweight small chances of large gains overpay for lottery-like stocks with positively skewed returns. Barberis and Huang (2008) showed this can explain the low average returns of IPOs, out-of-the-money options, and high-idiosyncratic-volatility stocks, connecting prospect theory to the low-risk anomaly in Section 6.

8.3 Overconfidence and Excessive Trading

Investors tend to overestimate the precision of their information and their skill. Overconfidence predicts excessive trading. Barber and Odean (2000) studied 66,465 households at a discount broker from 1991 to 1996. The 20% of households that traded most earned 11.4% per year net of costs, against 17.9% for the market. Men traded 45% more than women and earned lower net returns as a result (Barber & Odean, 2001). The main cost of overconfidence is not bad stock picks but the transaction costs of trading too often.

8.4 The Disposition Effect

Investors sell winners too early and hold losers too long. Shefrin and Statman (1985) named this the disposition effect. Odean (1998) documented it in brokerage data: investors realized gains at a rate about 1.5 times their rate of realizing losses, even though selling losers is better for taxes. The effect is consistent with prospect theory, since investors are risk-seeking in the loss domain, and with reluctance to admit a mistake. Grinblatt and Han (2005) showed that the disposition effect can generate momentum: if investors sell winners early, good news is incorporated into prices too slowly, which produces continuation.

8.5 Attention, Herding, and Extrapolation

Retail investors buy stocks that catch their attention, such as those in the news, with extreme returns, or with high volume (Barber & Odean, 2008). Because selling only involves stocks one already owns, attention mainly affects buying, which creates temporary price pressure that later reverses. Barber, Huang, Odean, and Schwarz (2022) found that Robinhood users concentrated purchases in stocks on the app's "top movers" lists, and that intense herding episodes predicted negative subsequent returns.

Investors also extrapolate. Survey expectations of future stock returns rise after strong past returns, even though, under rational models, high past returns usually mean higher valuations and lower expected returns (Greenwood & Shleifer, 2014). Extrapolative expectations are a key ingredient in modern bubble models (Barberis, Greenwood, Jin, & Shleifer, 2018).

8.6 Investor Sentiment

Baker and Wurgler (2006) built a composite sentiment index from the closed-end fund discount, NYSE turnover, the number and first-day returns of IPOs, the equity share of new issues, and the dividend premium. When sentiment is high, subsequent returns are low for stocks that are hard to value and hard to arbitrage, such as small, young, unprofitable, high-volatility, and non-dividend-paying firms. Stambaugh, Yu, and Yuan (2012) found that many anomalies are stronger following high-sentiment periods, mainly through their short legs. This is consistent with overpricing being harder to correct than underpricing because of short-sale constraints.

8.7 Noise Traders and Limits to Arbitrage

De Long, Shleifer, Summers, and Waldmann (1990) showed that irrational "noise traders" can survive and affect prices even when rational arbitrageurs are present. The key is noise-trader risk: an arbitrageur who buys an underpriced asset risks that noise traders become even more pessimistic before prices converge. Arbitrageurs with short horizons, or who face margin calls, cannot wait it out.

Shleifer and Vishny (1997) made this concrete. Real arbitrage is done by specialized professionals managing other people's money. If a mispricing worsens, the arbitrageur shows losses, investors withdraw capital, and the arbitrageur is forced to liquidate at the worst time. Arbitrage is therefore weakest exactly when mispricing is largest. The collapse of Long-Term Capital Management in 1998 was a textbook case.

Other limits include short-selling costs and constraints, the absence of perfect substitutes for many securities, and fundamental risk, since even a correctly identified mispricing can be overwhelmed by real news.

8.8 Case Studies of Persistent Mispricing

Royal Dutch and Shell. Royal Dutch Petroleum and Shell Transport and Trading were legally bound to split cash flows 60/40. In an efficient market, Royal Dutch shares should always trade at 1.5 times Shell shares. Froot and Dabora (1999) documented deviations from parity of as much as 35%, lasting years. The deviation correlated with the markets where each share mainly traded, suggesting that location-specific investor demand moved prices.

3Com and Palm. In March 2000, 3Com sold about 5% of its Palm subsidiary in an IPO and announced it would distribute its remaining Palm shares to 3Com holders. Each 3Com share would receive about 1.5 Palm shares. On the first day of trading, Palm's price implied that the rest of 3Com had a large negative value. Lamont and Thaler (2003) documented the anomaly and showed that it persisted for months, because Palm shares were extremely difficult and expensive to short.

GameStop, January 2021. Retail investors coordinating on the Reddit forum WallStreetBets drove GameStop shares from under $20 at the start of January 2021 to an intraday high near $483 on January 28. Short interest had exceeded 100% of the float. Short sellers, including hedge fund Melvin Capital, suffered large losses. Several brokers restricted purchases after clearinghouse margin calls spiked. The SEC's staff report (SEC, 2021) found that the rally was driven mainly by positive sentiment and buying, not by short covering alone, and highlighted the concentration of risk in clearing. The episode demonstrated noise-trader risk in reverse: coordinated retail buying imposed losses on professional arbitrageurs who were, on fundamentals, arguably correct.

8.9 Bridging the Views

The modern synthesis is that the efficient-markets and behavioral views are not opposites. Prices are approximately efficient for most securities most of the time, because arbitrage works well enough when it is cheap. Mispricing is concentrated where arbitrage is hard: small, illiquid, hard-to-short, hard-to-value stocks, and market-wide bubbles, where there is no close substitute to hedge with. Lo's (2004) Adaptive Markets Hypothesis formalizes this as an evolutionary process. Greenwood, Shleifer, and You (2019) tested Fama's challenge that bubbles cannot be identified in advance. They found that sharp industry price run-ups do not on average predict lower returns, but run-ups accompanied by high volatility, rising issuance, and young firms do predict substantially higher crash probabilities. Bubbles are thus partly predictable, which is evidence against strict efficiency.

9.1 The Interest Rate Channel and Equity Duration

Interest rates affect stock prices through the discount rate in the present-value formula. Equity duration, the sensitivity of price to discount rates, is high. Using the Gordon model, duration is approximately:

With \(r - g\) around 4%, duration is about 25 years, longer than most government bonds. Growth stocks, whose cash flows are concentrated far in the future, have especially long duration. This is why rising rates in 2022 hurt growth and technology stocks more than value stocks.

Rates are not the only channel. Higher rates also raise firms' borrowing costs, slow the economy, and reduce expected earnings growth. Rate increases that reflect a stronger economy can coincide with rising stock prices, while increases that reflect inflation fears usually do not.

9.2 Inflation and Equities

Equities are claims on real assets, so in theory they should hedge inflation over the long run. Empirically, U.S. stocks have performed poorly during periods of high and rising inflation. Fama (1981) proposed a proxy explanation: high inflation signals weak future real activity, and it is the weak activity, not inflation itself, that depresses stocks. Modigliani and Cohn (1979) proposed an inflation-illusion explanation: investors wrongly discount real cash flows with nominal interest rates, which undervalues stocks when inflation is high and overvalues them when inflation is low. DMS (2026) find that across 35 countries since 1900, real equity and bond returns were highest when economic growth was high and inflation was low.

The stock-bond correlation also depends on the inflation regime. From about 2000 to 2021, U.S. stock and bond returns were negatively correlated, so bonds hedged equity risk. In high-inflation regimes, such as the 1970s and 2022, the correlation turns positive, because inflation hurts both assets at once. DMS (2026) note that equity-bond correlations have risen, making diversification harder. The 2026 environment, with an energy-driven inflation shock and yields above 5%, is consistent with the positive-correlation regime.

9.3 Monetary Policy Announcements

Bernanke and Kuttner (2005) used federal funds futures to isolate the unexpected component of Federal Reserve rate decisions. A surprise 25-basis-point cut in the federal funds rate was associated with about a 1% increase in broad U.S. stock indices. Most of the effect came through lower expected future excess returns, that is, a lower risk premium, rather than through higher expected dividends or lower real rates.

Lucca and Moench (2015) found that from 1994 to 2011, the S&P 500 earned large average excess returns in the 24 hours before scheduled Federal Open Market Committee (FOMC) announcements, which they called the pre-FOMC announcement drift. These pre-announcement windows accounted for a large fraction of the total equity premium over that period. The effect has been weaker in more recent data, possibly because it became widely known. Savor and Wilson (2013) documented more broadly that stocks earn higher returns on days of major macroeconomic announcements, consistent with investors being compensated for bearing announcement risk.

9.4 Quantitative Easing

After the policy rate reached zero in December 2008, the Federal Reserve purchased large quantities of Treasury and mortgage-backed securities. The Fed's balance sheet grew from under $1 trillion before 2008 to about $9 trillion at its 2022 peak. Quantitative easing (QE) raised stock prices through several channels: lower long-term yields, a portfolio-rebalancing effect that pushed investors into riskier assets, and a signaling effect that rates would stay low. Isolating QE's effect on equities is difficult because purchase announcements coincided with other policy actions and crisis developments. Quantitative tightening, the reversal of QE, began in 2022.

9.5 Earnings and the Business Cycle

Over long horizons, stock prices track earnings. Corporate earnings are much more cyclical than GDP. S&P 500 operating earnings fell by more than half from peak to trough in 2008–2009. Recessions are the main driver of bear markets. Of the major U.S. bear markets since World War II, those accompanied by recessions have been deeper and longer on average than those without.

Profit margins are a key current variable. U.S. corporate profit margins reached record levels in the 2020s, helped by the high margins of the largest technology firms. Whether margins mean-revert, as they traditionally have, or remain structurally higher because of the changing composition of the index, is a central uncertainty for valuation.

9.6 The 2022 Tightening, the 2024–2025 Easing, and the 2026 Reversal

The post-pandemic policy cycle is an unusually clear case study of monetary transmission.

• 2022. With CPI inflation peaking at 9.1% in June 2022, the Fed raised the federal funds rate from near zero in March 2022 to 5.25%–5.50% by July 2023, the fastest tightening since the early 1980s. The S&P 500 fell about 25% from its January 2022 peak to its October 2022 trough. Long-duration growth stocks fell further.
• 2023–2024. Inflation receded, rate hikes paused, and AI enthusiasm drove a concentrated rally. The Fed began cutting in September 2024 and reached 4.25%–4.50% by December 2024.
• 2025. After holding through mid-2025, the Fed cut three times, in September, October, and December, ending 2025 at 3.50%–3.75% (E*TRADE from Morgan Stanley, 2025). The S&P 500 returned about 17.9% in 2025, its third straight double-digit year.
• 2026. The Iran war that began in late February 2026 and the closure of the Strait of Hormuz produced an oil shock. Brent crude traded in a range of roughly $61 to $138 per barrel from January through September, by one summary, and settled near $106 on October 13. U.S. headline CPI inflation was 3.4% year over year in August 2026, and the core personal consumption expenditures (PCE) price index ran above 3% throughout 2026. Kevin Warsh succeeded Jerome Powell as Fed chair in May 2026. On September 16, 2026, the FOMC voted 12–0 to raise the target range by 25 basis points to 3.75%–4.00%, the first hike since July 2023. The median projection called for one more hike in 2026.

Bond markets moved sharply. The 10-year Treasury yield rose from about 4.42% at the end of June 2026 to about 5.19% in late September, its highest close since July 2007. It reached as high as 5.36% in early October, a 24-year high, before easing to roughly 5.14% on October 15 as investors sought safety during an equity sell-off.

The striking feature of 2026 is that equities rose despite this. The S&P 500 was up roughly 13% to 14% year to date in mid-October, and set a record close of 7,818.93 on October 6. That combination, rising real and nominal yields alongside record stock prices, is unusual. It indicates that markets have been pricing exceptionally strong earnings growth rather than falling discount rates. Section 15 evaluates whether that pricing is justified.

10.1 A Framework

Kindleberger's classic account, building on Minsky, describes a recurring sequence: a displacement (a new technology or policy shift) creates real profit opportunities, credit expands, prices rise and attract speculators, euphoria sets in, insiders begin to sell, a trigger causes distress, and panic follows (Kindleberger & Aliber, 2005). Two variables explain most of the difference between a painful correction and a systemic crisis. The first is leverage: losses on borrowed money force sales and transmit stress to lenders. The second is liquidity: when market makers and arbitrageurs withdraw, prices overshoot. Valuation determines how far prices can fall. Leverage and liquidity determine how fast and how far the damage spreads.

10.2 The Crash of 1929 and the Great Depression

Causes. The 1920s combined real technological change, including electrification, automobiles, and radio, with rapid growth in margin lending, where investors could buy stocks with as little as 10% down. Investment trusts, often themselves leveraged, layered on further borrowing. Brokers' loans to investors rose sharply in 1928–1929.

Crash and decline. The Dow Jones Industrial Average peaked at 381.17 on September 3, 1929. It fell about 23% over Black Thursday (October 24) and Black Monday and Tuesday (October 28–29). After a partial rebound in early 1930, it declined almost continuously to 41.22 on July 8, 1932, a cumulative loss of about 89%. The Dow did not regain its 1929 nominal peak until November 1954, about 25 years later. Including dividends and deflation, the real recovery was faster, roughly a decade, but the experience still dominates popular memory.

Policy response and interpretation. Monetary policy tightened when it should have eased, about 9,000 banks failed, and the money supply contracted by about a third (Friedman & Schwartz, 1963). Galbraith (1954) emphasized speculative excess. Romer (1990) argued that the crash itself depressed spending by increasing uncertainty about future income. The modern consensus is that the stock market crash did not cause the Depression alone. The banking collapse and policy errors that followed turned a severe recession into a catastrophe. The lessons shaped the Securities Acts, deposit insurance, and the Fed's later willingness to act as lender of last resort.

10.3 Black Monday, 1987

Causes. U.S. stocks rose about 40% in the first eight months of 1987. Rising interest rates, a falling dollar, and trade-deficit concerns built pressure. A new strategy, portfolio insurance, mechanically sold index futures as prices fell to replicate a put option.

Crash. On October 19, 1987, the Dow fell 22.6% in one day, still the largest one-day percentage decline in its history. The S&P 500 fell about 20.5%. From its August 25 peak to its December 4 trough, the S&P 500 fell about 33.5%.

Policy response. Before markets opened on October 20, the Fed announced its readiness to serve as a source of liquidity and encouraged banks to lend to securities firms. Markets stabilized. The S&P 500 regained its prior peak by July 1989, less than two years later.

Interpretation. The Brady Commission (1988) found that portfolio insurance and index arbitrage transmitted selling between futures and cash markets, which were not equipped to absorb it. There was no major news on October 19 that could justify a 20% revaluation, which is difficult to reconcile with strict efficiency. The crash led to the first market-wide circuit breakers. It also had a lasting effect on options markets: before 1987, implied volatility was roughly the same across strike prices, and afterward, out-of-the-money puts became permanently more expensive. This "volatility smirk" reflects investors' persistent fear of crashes.

10.4 The Dot-Com Bubble, 2000–2002

Causes. The commercialization of the internet was a genuine displacement. Investors extrapolated rapid growth for any company associated with it, and IPOs of unprofitable firms soared on their first day. The S&P 500 CAPE reached 44.2 in December 1999, its all-time high.

Decline. The NASDAQ Composite peaked at 5,048.62 on March 10, 2000 and fell about 78% to 1,114.11 on October 9, 2002. It did not regain its peak until April 2015. The S&P 500 fell about 49% from its March 24, 2000 peak to the October 9, 2002 trough and did not regain its prior peak until May 2007.

Interpretation. Ofek and Richardson (2003) linked the timing of the collapse to the expiration of IPO lockups, which suddenly increased the supply of shares and relaxed short-sale constraints. Brunnermeier and Nagel (2004) found that hedge funds rode the bubble rather than shorting it, consistent with limits to arbitrage. Bubbles do not always reflect useless technology. The internet transformed the economy, but most early investors still lost money, because they overpaid. DMS (2026) note the same pattern for railroads: a transformative industry can make poor investments when expectations run ahead of profits.

The dot-com bust was a valuation crash with relatively little leverage, and the accompanying 2001 recession was mild. It is the closest historical analogue to current debates about AI (Section 15).

10.5 The Global Financial Crisis, 2007–2009

Causes. A U.S. housing boom financed by subprime mortgages was packaged into securitized products held throughout the financial system, often by highly leveraged institutions using short-term funding. When house prices fell, losses spread through repo markets and asset-backed commercial paper, causing a run on the shadow banking system (Brunnermeier, 2009; Gorton & Metrick, 2012).

Decline. The S&P 500 peaked at 1,565.15 on October 9, 2007 and fell about 57% to 676.53 on March 9, 2009. Lehman Brothers failed on September 15, 2008. The VIX volatility index closed above 80 in November 2008. The S&P 500 did not regain its prior peak until March 2013.

Policy response. The response was the largest in history to that point: the Troubled Asset Relief Program (TARP), Fed emergency lending facilities, near-zero rates from December 2008, QE, bank stress tests in 2009, and a temporary ban on short selling in financial stocks in September 2008.

Interpretation. The GFC was a leverage crisis. Equity losses were severe because the financial system that supported credit to the whole economy was impaired. Researchers also found that the 2008 short-sale ban reduced liquidity without stopping price declines (Beber & Pagano, 2013). Post-crisis reforms, chiefly the Dodd-Frank Act and Basel III capital rules, were designed to prevent a repeat.

10.6 The COVID-19 Crash, 2020

Causes. A pandemic shut down large parts of the global economy in March 2020. In bond markets, even U.S. Treasuries became illiquid as investors scrambled for cash.

Decline. The S&P 500 fell about 34% from February 19 to March 23, 2020, the fastest decline of that magnitude in history. The VIX closed at 82.69 on March 16, 2020, a record close. Trading was halted by market-wide circuit breakers four times in March 2020.

Policy response. The Fed cut rates to zero, restarted QE without a size limit, and created facilities to buy corporate bonds, its first such intervention. Congress passed fiscal packages worth trillions of dollars. The S&P 500 regained its peak by August 18, 2020, about five months after the trough.

Interpretation. Baker, Bloom, Davis, Kost, Sammon, and Viratyosin (2020) found that no previous infectious disease outbreak, including the 1918 influenza pandemic, had produced comparable daily stock market moves, and attributed the difference to policy responses and the information environment. Ramelli and Wagner (2020) showed that firms with exposure to China and with high leverage or low cash fell most, consistent with investors rapidly pricing in real economic and financial channels. The speed of the recovery shows how strongly central bank backstops now influence equity markets.

10.7 The 2022 Rate Shock

Causes. Inflation rose to multidecade highs, and the Fed responded with rapid hikes. Unlike most bear markets, this one was driven primarily by the discount rate rather than by earnings collapse.

Decline. The S&P 500 fell about 25% from January 3 to October 12, 2022. The NASDAQ fell about a third. A traditional 60/40 portfolio had one of its worst years in a century, because bonds fell alongside stocks.

Recovery. The S&P 500 regained its January 2022 peak in January 2024. The recovery was led by a narrow group of large technology stocks associated with AI after the release of ChatGPT in November 2022.

Interpretation. 2022 demonstrated equity duration in practice. Long-duration assets, including unprofitable growth stocks, speculative technology, and cryptocurrency, fell most. It also ended the low-inflation, negative stock-bond correlation regime that had made bonds a reliable equity hedge for two decades.

10.8 The April 2025 Tariff Shock

Causes. On April 2, 2025, the U.S. administration announced broad "reciprocal" tariffs, which it called "Liberation Day." The tariffs were far larger than markets expected. China retaliated two days later.

Decline. The S&P 500 fell about 21% intraday from its February 19 high of 6,147.43 to an intraday low of 4,835.04 on April 7, 2025. On a closing basis the decline was just under 20%, so the index did not officially enter a bear market (S&P Dow Jones Indices, 2025). The two-day loss on April 3–4 was among the largest in post-war history. Retail investors bought heavily: JPMorgan estimated that retail accounted for a record 35% of U.S. equity volume during the April volatility.

Recovery. On April 9, 2025, the administration announced a 90-day pause on most reciprocal tariffs, and the S&P 500 rose about 9.5% in one day, one of its largest daily gains on record. The index recovered most of its losses by the end of April and reached new highs by late June 2025.

Interpretation. The episode was a pure policy shock. It was severe and fast, but it reversed quickly once the policy changed. It is a reminder that in 2025–2026, government trade and geopolitical decisions became first-order drivers of equity risk, as they were in 1973–1974. DMS (2026) note that geopolitical events produced three of the six worst episodes for global equities since 1900, though their regressions find no general relationship between a geopolitical threat index and future returns.

10.9 Comparison

Table 11. Major U.S. equity market declines. Price-index levels, excluding dividends. Sources: S&P Dow Jones Indices; historical index data; S&P Global commentary for 2025. Durations are approximate.

10.10 Cross-Episode Lessons

1. Leverage determines severity. The two deepest declines, 1929 and 2008, involved leveraged financial systems. The valuation-driven 2000 bust was severe for technology stocks but did not threaten the financial system.
2. Policy determines duration. Recoveries were slow when policy was tight or slow to respond (1929–1932) and fast when central banks and governments acted aggressively (1987, 2020, 2025).
3. Speed is increasing. The 2020 and 2025 declines unfolded in weeks, compared with months or years earlier. Faster information flow, electronic trading, and leveraged products compress the timeline.
4. Valuation sets the downside. High starting valuations amplified the 1929 and 2000 declines and lengthened recoveries.
5. The "Fed put" is a double-edged sword. Repeated rescues have shortened bear markets, but they may encourage risk-taking and reduce the market's self-correcting capacity. In 2026, with inflation above target and the Fed hiking, the scope for a rapid rescue is narrower than in 2020.

11.1 Who Owns U.S. Stocks

U.S. equity ownership has shifted over decades from directly held household portfolios to intermediaries. Households still own a large share of U.S. equities when indirect holdings through mutual funds, ETFs, and retirement accounts are included, but the share held directly in brokerage accounts fell for decades before rebounding with the zero-commission era. Foreign investors hold a substantial and rising share. Pension funds have shifted from equities toward bonds and private assets as plans matured. Hedge funds hold a small share of assets but account for a much larger share of trading.

11.2 Active Versus Passive

Index funds and ETFs have taken an increasing share of fund assets for three decades. Passive U.S. equity funds overtook active U.S. equity funds in August 2019 ($4.27 trillion versus $4.25 trillion), and passive funds overtook active funds across all U.S. fund categories at year-end 2023 ($13.3 trillion versus $13.2 trillion) (Morningstar data, as reported in trade press). In 2025, active U.S. equity funds had net outflows of about $32 billion, led by large-cap funds.

The shift reflects consistent evidence that active management underperforms after fees.

Table 12. Active large-cap fund underperformance. Source: S&P Dow Jones Indices, SPIVA U.S. scorecards (2025, 2026).

In 2025, the S&P 500 beat the S&P MidCap 400 by about 10 percentage points and the S&P SmallCap 600 by about 12 points. Active managers typically hold smaller stocks than the index and underweight the very largest companies, so years of mega-cap dominance are especially difficult for them.

11.3 The Arithmetic of Active Management

Sharpe (1991) made an argument that requires no empirical data. Before costs, the average actively managed dollar must earn the market return, because active and passive investors together hold the market, and passive investors earn exactly the market return. After costs, which are higher for active management, the average active dollar must underperform the average passive dollar. Underperformance by active funds in aggregate is therefore guaranteed, not an empirical surprise.

The question is whether some managers have persistent skill. Fama and French (2010) found that, net of fees, the distribution of mutual fund alphas looks like what one would expect from pure luck, with little evidence that the best performers are skilled. Berk and Green (2004) offered a different interpretation. Skilled managers exist, but investors chase performance and allocate more money to them. Larger funds face diminishing returns to scale, so in equilibrium, skilled managers capture their skill as fees and net-of-fee alphas are zero. In this view, the absence of net outperformance does not imply the absence of skill.

11.4 Does Passive Investing Distort Prices?

As passive ownership grows, an important question is whether markets remain efficient. Several channels of concern have been studied.

Index inclusion effects. Shleifer (1986) found that stocks added to the S&P 500 rose about 3% on announcement, even though inclusion conveys no fundamental information. Harris and Gurel (1986) found the effect partly reversed. This suggested that demand curves for stocks slope downward, contrary to the textbook assumption of perfect substitutes. The inclusion effect has shrunk in recent decades as traders anticipate it.

The inelastic markets hypothesis. Gabaix and Koijen (2021) estimated that the aggregate U.S. stock market is highly price-inelastic. Many institutions, such as balanced funds with fixed equity-bond allocations, rebalance toward fixed targets rather than responding to prices. Their estimates imply that investing $1 into the stock market raises its aggregate value by about $5. If correct, fund flows, including flows into passive funds, can have large effects on aggregate prices, and the share of price movements explained by fundamentals is smaller than standard models imply. The hypothesis is influential but contested, since estimating demand elasticities is difficult.

Price efficiency and comovement. Wurgler (2010) argued that index membership causes stocks to move together more than fundamentals justify. Some studies find that higher passive ownership is associated with less informative prices, as fewer investors research individual firms. Others find the effect small, because the remaining active investors become more informed on the margin. The Grossman-Stiglitz logic implies an equilibrium. As passive share grows, the returns to active research rise until enough active investors remain.

Governance. The three largest index fund managers are collectively the largest shareholders of most S&P 500 companies. This raises questions about common ownership and anticompetitive effects, which are debated empirically, and about whether passive managers have incentives to monitor companies.

11.5 Fewer Listed Companies and the Rise of Private Markets

The number of U.S. listed companies has fallen substantially since its 1996 peak. Estimates differ by data source: roughly 8,000 to 8,800 in 1996 versus about 4,000 to 5,500 recently, a decline of 38% to 50% depending on the count. Doidge, Karolyi, and Stulz (2017) documented the "U.S. listing gap" and attributed it to fewer new listings and more delistings through mergers. Contributing factors include the cost of public-company compliance, the abundance of venture capital and private equity, and changes in the economy that favor scale.

The result is that more of the economy's growth is captured in private markets before companies go public, if they go public at all. Bloomberg Intelligence counted nearly 800 U.S. companies valued above $1 billion that remained private as of 2025. Public-market investors therefore have less exposure to the early growth phase of new firms. This may begin to reverse. OpenAI confirmed on June 8, 2026 that it had confidentially submitted a draft IPO registration statement to the SEC, and Anthropic reportedly made a confidential filing on June 1, 2026 (press reports, June 2026). This paper did not find confirmation of listing dates. If those listings occur, they would substantially enlarge the public market for AI.

11.6 Skewness and Concentration of Wealth Creation

Bessembinder (2018) studied all U.S. common stocks from 1926 to 2016. Most individual stocks, about four out of every seven, had lifetime buy-and-hold returns below those of one-month Treasury bills. The best-performing 4% of listed companies accounted for all net wealth creation in excess of Treasury bills. The reason is skewness: individual stock returns compound multiplicatively, so a few extreme winners produce enormous wealth while the typical stock disappoints. The finding explains why diversified index funds are hard to beat. An undiversified investor is likely to miss the few stocks that matter.

11.7 Concentration Today

U.S. market concentration is at levels not seen in at least a century. DMS (2026) state that U.S. equity market concentration was at its highest level in at least 100 years at the end of 2025.

Table 13. S&P 500 concentration. Secondary sources differ in method and date. Figures are approximate.

Concentration has three implications. First, the S&P 500 is less diversified than its 500 names suggest. A large share of its risk is the risk of a handful of firms exposed to a common theme, AI. Second, active managers who underweight the largest stocks will tend to underperform during concentration-increasing periods and outperform during reversals. Third, concentration by itself does not predict poor returns. Historically, periods of high concentration, such as the 1960s "Nifty Fifty," were followed by mixed outcomes. What matters is whether the largest firms' earnings justify their valuations. In contrast to 1999, the largest firms in 2026 are among the most profitable companies in history, which is the strongest argument that today's concentration is not purely speculative.

12.1 Stylized Facts of Volatility

Equity return volatility has several robust empirical properties:

1. Volatility clustering. Large moves tend to follow large moves, of either sign. Calm periods persist, and so do turbulent ones.
2. Fat tails. Daily returns have far more extreme observations than a normal distribution predicts. The 22.6% decline of October 19, 1987 would be essentially impossible under normality.
3. The leverage effect. Volatility rises more after price declines than after equal-sized increases (Black, 1976). Falling prices raise firms' financial leverage, and fear amplifies uncertainty.
4. Mean reversion. Volatility spikes decay over weeks to months toward a long-run average.
5. Correlation asymmetry. Correlations among stocks, and among international markets, rise in downturns (Longin & Solnik, 2001; Ang & Chen, 2002). Diversification fails partly when it is most needed.

12.2 GARCH Models

Engle (1982) introduced the autoregressive conditional heteroskedasticity (ARCH) model, and Bollerslev (1986) generalized it to GARCH. The GARCH(1,1) model is:

where \(r_t\) is the return, \(\mu\) is the mean, \(z_t\) is an independent shock with mean zero and variance one, \(\sigma_t^2\) is the conditional variance, and \(\omega > 0\), \(\alpha \geq 0\), \(\beta \geq 0\) are parameters. The sum \(\alpha + \beta\) measures persistence. For daily equity data it is typically close to 0.98 or 0.99, so volatility shocks decay slowly. The long-run variance is \(\omega / (1 - \alpha - \beta)\).

To capture the leverage effect, the GJR-GARCH model of Glosten, Jagannathan, and Runkle (1993) adds an asymmetric term:

where \(\mathbb{1}[\cdot]\) is an indicator equal to one when the previous shock was negative and \(\gamma > 0\) measures the extra impact of bad news. Nelson's (1991) EGARCH models the log of variance, which guarantees positivity and also allows asymmetry. Engle shared the 2003 Nobel Memorial Prize in Economic Sciences for the ARCH framework.

12.3 Implied Volatility and the VIX

Option prices embed the market's forecast of future volatility. The Cboe Volatility Index (VIX) measures the 30-day expected volatility of the S&P 500 implied by a strip of S&P 500 index option prices across strikes. Since 2003, it has been computed from a model-free formula related to the price of a variance swap:

where \(T\) is time to expiration, \(K_i\) is the strike of the \(i\)-th out-of-the-money option, \(\Delta K_i\) is the interval between strikes, \(R\) is the risk-free rate, \(Q(K_i)\) is the midpoint of the option's bid-ask quote, \(F\) is the forward index level, and \(K_0\) is the first strike below \(F\). The VIX is \(100 \times \sigma_{\text{VIX}}\). Whaley (2000) popularized it as the "investor fear gauge."

Table 14. Notable VIX readings. Sources: Cboe historical data; press reports.

12.4 The Variance Risk Premium

Implied volatility usually exceeds subsequently realized volatility. The difference is the variance risk premium (VRP):

where \(\mathbb{Q}\) denotes the risk-neutral expectation implied by option prices and \(\mathbb{P}\) denotes the real-world expectation. Investors pay a premium for protection against volatility spikes because spikes coincide with market crashes. Carr and Wu (2009) documented large, persistent negative returns to buying variance. Bollerslev, Tauchen, and Zhou (2009) found that the VRP predicts aggregate stock returns at horizons of a few months, better than traditional valuation ratios at those horizons. Selling volatility therefore earns a positive average return, but with severe left-tail risk, like "picking up nickels in front of a steamroller."

12.5 Volmageddon, 2018

By early 2018, exchange-traded products that were short VIX futures had attracted billions of dollars, earning steady returns from the variance risk premium during an unusually calm 2017. On February 5, 2018, the VIX more than doubled in one day. Because these products had to buy VIX futures to rebalance after losses, their hedging demand pushed futures higher in a feedback loop. The largest such product, an exchange-traded note, lost more than 90% of its value in one day and was subsequently terminated. The episode illustrated how a popular strategy, crowded into a small market, can create the very spike it is exposed to.

12.6 The Zero-Days-to-Expiration Boom

Options expiring on the same day they trade (zero days to expiration, 0DTE) have transformed index options markets since daily S&P 500 expirations were introduced in 2022. In 2025, SPX options averaged 3.9 million contracts per day, and 0DTE options averaged a record 2.3 million contracts per day, 59% of total SPX volume. The share reached a monthly record of 62.4% in August 2025 (Cboe, 2025). Cboe estimates that retail traders account for roughly half of 0DTE SPX volume.

The financial-stability debate centers on whether dealers who sell 0DTE options and hedge them dynamically could amplify intraday moves. When dealers are short options ("short gamma"), they must sell as prices fall and buy as prices rise, which reinforces trends. When dealers are long options ("long gamma"), their hedging dampens moves. Evidence so far is mixed. Cboe and some academic work find that 0DTE positioning is fairly balanced between buyers and sellers, which limits net dealer gamma. Skeptics argue that the risk is concentrated in rare, large moves that have not yet occurred. During the April 2025 sell-off, the 0DTE share dropped sharply, from 64% on April 1 to 36% on April 7, suggesting that traders shifted to longer-dated protection when volatility spiked.

12.7 Correlation in Crises

When markets crash, correlations among stocks rise toward one, and international markets fall together. Longin and Solnik (2001) used extreme value theory to show that correlation increases in bear markets but not in bull markets. The practical implication is that diversification within equities protects against firm-specific risk but much less against systematic crashes. Diversification across asset classes works better, but even the stock-bond relationship can break down when inflation is the shock, as 2022 and 2026 show. DMS (2026) find that a 60/40 portfolio has never lost more than 50% in real terms over 1900–2025, even though stocks and bonds each have.

13.1 Portfolio Sorts

The workhorse method in empirical asset pricing sorts stocks into portfolios by a characteristic, such as deciles of book-to-market, and compares average returns across portfolios. A long-short portfolio, long the top decile and short the bottom decile, gives the "spread" return. Choices about breakpoints, weighting, and rebalancing matter. Equal weighting overweights microcaps, which are numerous but small. Value weighting reflects investable returns. Using NYSE breakpoints prevents tiny NASDAQ and AMEX stocks from dominating the extreme portfolios. Hou, Xue, and Zhang (2020) showed that these choices alone determine whether many anomalies appear significant.

13.2 Time-Series Factor Regressions

To test whether a portfolio earns abnormal returns relative to a factor model, regress its excess returns on factor returns:

where \(F_{kt}\) are factor returns. If the model is correct, \(\alpha_p = 0\). Gibbons, Ross, and Shanken (1989) developed the GRS test of whether the alphas of a set of portfolios are jointly zero.

13.3 Fama-MacBeth Cross-Sectional Regressions

Fama and MacBeth (1973) proposed a two-step procedure. Each period, run a cross-sectional regression of returns on characteristics or estimated betas:

Then average the slope estimates over time and compute standard errors from their time-series variation:

The method handles cross-sectional correlation in returns. When betas are estimated in a first stage, standard errors must be corrected for the errors-in-variables problem (Shanken, 1992).

13.4 Standard Errors

Financial data are typically autocorrelated and heteroskedastic, especially with overlapping long-horizon returns. Newey and West (1987) developed a heteroskedasticity- and autocorrelation-consistent (HAC) covariance estimator:

where \(\hat{\Gamma}_j\) is the sample autocovariance of the score at lag \(j\) and \(L\) is the number of lags. For panel data, standard errors clustered by firm and time are now standard (Petersen, 2009).

13.5 Event Studies

Event studies measure the price reaction to specific information, such as earnings, mergers, or regulatory changes (Fama et al., 1969; MacKinlay, 1997). Estimate a normal-return model, such as the market model, over a pre-event window. Then compute abnormal returns in the event window:

where \(AR_{it}\) is the abnormal return and \(CAR\) is the cumulative abnormal return. Event studies are among the most reliable tools in finance because short windows minimize the joint-hypothesis problem: over a few days, the expected return is small relative to the news effect, so the choice of benchmark model matters little.

13.6 Research Pitfalls

Table 15. Common pitfalls in empirical finance research.

13.7 Backtest Overfitting and the Deflated Sharpe Ratio

Bailey, Borwein, López de Prado, and Zhu (2014) showed that, with enough trials, it is easy to find a strategy with an impressive in-sample Sharpe ratio purely by chance. If a researcher tries \(N\) independent strategies with zero true Sharpe ratio, the expected maximum in-sample Sharpe ratio grows roughly with \(\sqrt{2 \ln N}\). Bailey and López de Prado (2014) proposed a "deflated Sharpe ratio" that adjusts for the number of trials, the length of the sample, and the skewness and kurtosis of returns. The practical lesson is that a backtest is only as credible as the researcher's record of how many alternatives were tried.

The Sharpe ratio itself is defined as:

where \(R_p\) is the portfolio return. Annualizing a monthly Sharpe ratio multiplies it by \(\sqrt{12}\) if returns are independent over time.

13.8 Machine Learning: Promise and Limits

Machine learning methods (Gu et al., 2020) can capture nonlinear interactions and handle hundreds of predictors. They also amplify every pitfall above. Financial data have a low signal-to-noise ratio, are non-stationary, and offer relatively few independent observations compared with fields such as image recognition. Large language models trained on text introduce a further risk: if a model's training data include information from after the forecast date, backtests suffer from a subtle form of look-ahead bias that is difficult to detect. This paper's assessment is that machine learning has improved cross-sectional prediction modestly and measurably, but the most reliable gains come from disciplined out-of-sample design rather than from model complexity.

14.1 The Foundations

U.S. securities regulation rests on statutes passed in response to crises.

• Securities Act of 1933. Requires registration and full disclosure for public offerings and creates liability for misstatements.
• Securities Exchange Act of 1934. Creates the SEC, requires periodic reporting by public companies, regulates exchanges and broker-dealers, and prohibits fraud and manipulation. Rule 10b-5 under the Act is the basis for most securities fraud and insider trading enforcement.
• Investment Company Act and Investment Advisers Act of 1940. Regulate mutual funds and investment advisers.
• Sarbanes-Oxley Act of 2002. Passed after the Enron and WorldCom accounting frauds. It requires CEO and CFO certification of financial statements, strengthens auditor independence, creates the Public Company Accounting Oversight Board, and requires internal control assessments (Section 404). Critics argue that compliance costs contributed to the decline in U.S. listings (Section 11.5).
• Dodd-Frank Act of 2010. The main post-GFC reform. It created the Financial Stability Oversight Council, the Volcker Rule restricting proprietary trading by banks, central clearing and reporting requirements for most over-the-counter derivatives, and enhanced oversight of systemically important institutions.

14.2 Market-Structure Rules

Regulation NMS (2005). Established the order protection rule, access rules, and market data rules that define today's linked, fragmented market (Section 3).

Regulation SHO (2005). Governs short selling, including "locate" and close-out requirements to curb abusive naked short selling. A 2010 amendment (Rule 201) restricts short sales at or below the best bid when a stock has fallen 10% in a day.

Circuit breakers. After 1987, exchanges introduced market-wide trading halts. The current system, revised in 2012, halts trading across U.S. equity markets for 15 minutes if the S&P 500 falls 7% (Level 1) or 13% (Level 2) from the prior close before 3:25 p.m., and for the rest of the day at 20% (Level 3). The Level 1 halt was triggered on March 9, 12, 16, and 18, 2020.

Limit-up/limit-down (LULD). After the 2010 Flash Crash, the SEC approved a mechanism that prevents trades in individual stocks outside price bands around a recent reference price and pauses trading when a stock stays at a band. The plan took effect in 2013.

Consolidated Audit Trail. The SEC adopted Rule 613 in 2012 to create a comprehensive database tracking every order across U.S. equity and options markets. Implementation has been slow, contentious, and costly. Its scope and funding have been revisited repeatedly.

Settlement. T+1 settlement took effect on May 28, 2024.

14.3 Europe: MiFID II

The European Union's revised Markets in Financial Instruments Directive (MiFID II), effective January 3, 2018, introduced extensive pre- and post-trade transparency, caps on dark-pool trading, requirements to "unbundle" payments for research from trading commissions, and best-execution reporting. The research unbundling rules were associated with reduced analyst coverage of smaller companies, and the United Kingdom and EU have since relaxed them. The EU's later review restricted payment for order flow.

14.4 Insider Trading Enforcement

U.S. insider trading law is largely judge-made, based on Rule 10b-5's antifraud prohibition. Liability requires trading on material, nonpublic information in breach of a duty. The "classical" theory covers corporate insiders, and the "misappropriation" theory covers outsiders who misuse confidential information. Courts have repeatedly refined the requirement that a tipper receive a "personal benefit." In 2022, the SEC tightened Rule 10b5-1 trading plans, which allow insiders to trade on preset schedules, by adding cooling-off periods after plan adoption.

14.5 The Current Regulatory Agenda (2025–2026)

Since 2025, under Chair Paul Atkins, the SEC has shifted toward deregulation and market modernization.

Table 16. Recent U.S. regulatory actions affecting equity markets. Sources: SEC releases and law-firm summaries (Cooley, 2026; Skadden, 2026; Paul Hastings, 2025); trade press for exchange approvals.

Semiannual reporting. Supporters argue that quarterly reporting encourages short-termism and that less frequent reporting would lower costs and encourage more firms to list. Opponents cite evidence that more frequent reporting improves price efficiency and reduces information asymmetry, which lowers the cost of capital. The record comment volume indicates how contested the proposal is.

Tokenization and 24-hour trading. Tokenized stocks, digital tokens on a blockchain that represent ownership of actual shares, could allow continuous trading and faster settlement. Risks include fragmentation of liquidity across more venues and hours, thin overnight markets prone to large price moves, and operational and custody risks. The SEC's exemptive approach allows experimentation with conditions rather than a permanent rule change.

14.6 Assessment

Regulation has tracked crises closely: 1933–1934 after the crash, Sarbanes-Oxley after Enron, Dodd-Frank after 2008, circuit breakers after 1987, and LULD after 2010. This paper's assessment is that post-2008 reforms made the banking system substantially more resilient, as its performance in 2020 showed. The risks have migrated toward nonbank institutions, leveraged products, private credit, and market plumbing, where regulatory coverage is thinner. The 2025–2026 agenda prioritizes capital formation and innovation over new constraints. Whether that balance proves wise will likely be tested in the next stress episode.

All figures in this section are as of October 15, 2026, unless another date is given. Current-market data come mainly from financial press and data trackers. Where sources conflict, ranges are reported. Nothing in this section is investment advice.

15.1 Index Levels and Returns

Table 17. U.S. market dashboard, October 2026.

The 2026 path has been volatile. Stocks fell after the Iran war began in late February and Brent crude rose above $100 per barrel. The S&P 500's year-to-date drawdown bottomed on March 30, 2026 at about 8.9%. A ceasefire in early April and very strong first-quarter earnings drove a recovery. The index then rose to records through the summer and early autumn despite rising yields, a renewed rate hike, and a breakdown of the ceasefire in July. In the second week of October, AI-related stocks sold off, oil rose again, and the September CPI report was hotter than expected. Sources disagree on the exact October 14 close, with figures ranging from about 7,720 to 7,806, but agree that the index remained within roughly 1% to 2% of its record.

15.2 Earnings: The Core of the Bull Case

The single most important fact about the 2026 market is the strength of corporate earnings. FactSet estimated on October 9, 2026 that S&P 500 earnings grew 29.6% year over year in the third quarter of 2026. That would be the third consecutive quarter of growth above 25%. In a typical season, actual growth exceeds quarter-end estimates as companies beat expectations. FactSet's ten-year average beat effect would put final third-quarter growth above 35%. That projection, not yet an outcome, is based on historical patterns.

Earnings growth of this size is rare outside recoveries from recessions. It is concentrated in technology and AI-related sectors, and in energy, where higher oil prices raised profits. The FactSet forward 12-month P/E ratio fell from 22.5 in December 2025 to 19.9 in late March 2026, even as prices were roughly flat, because earnings estimates rose faster than prices. This is the key distinction between 2026 and 1999. In 1999, prices rose far faster than earnings. In 2026, earnings have kept pace with or outrun prices for much of the year, though the trailing, cyclically adjusted measure (CAPE) remains near record highs because it averages ten years of lower past earnings.

15.3 The AI Investment Boom

The largest U.S. technology companies are investing at an unprecedented scale in AI infrastructure: data centers, specialized chips, and power. Based on first-quarter 2026 guidance as reported in financial press, combined 2026 capital spending for Amazon, Microsoft, Alphabet, and Meta was roughly $700–725 billion, up about 75% or more from roughly $410 billion in 2025. Individual guidance was about $200 billion for Amazon, about $190 billion for Microsoft, $180–190 billion for Alphabet, and $125–145 billion for Meta. Some of the increase reflects higher component prices rather than added capacity. Analysts at several banks project combined spending above $1 trillion in 2027. These figures are based on early-2026 guidance and may have been revised in subsequent quarters.

This spending is revenue for chipmakers, memory producers, networking firms, and power suppliers, which is why Nvidia's market capitalization reached about $5.7–5.8 trillion in early October 2026, having first crossed $5 trillion on April 24, 2026. Memory maker Micron entered the S&P 500's top ten by weight in 2026. Cloud revenue growth has been rapid. Reported first-quarter growth was 63% for Google Cloud and 28% for Amazon Web Services, and cloud providers report large contracted backlogs.

15.4 The Bubble Debate

The central question for U.S. equities is whether AI-related valuations are justified. The arguments are presented here on both sides.

The case that this is not a bubble, or not mainly one:

1. Profits are real. The leading firms are among the most profitable companies in history and fund most of their investment from operating cash flow. In 1999, many leading internet stocks had no earnings.
2. Demand is visible. Cloud providers report contracted backlogs, and executives say they are capacity-constrained. Spending responds to orders already received.
3. Valuations on forward earnings are elevated but not extreme. Forward P/E ratios for the index, around 20 to 23 over the past year, are high relative to history but well below 1999-2000 levels on a comparable basis.
4. Earnings growth is broadening. Third-quarter 2026 growth estimates rose across sectors, including energy and financials.
5. Productivity potential. If AI produces a broad productivity boom, current earnings may understate long-run cash flows, as CAPE critics argue.

The case that risk is elevated:

1. Valuation history. The CAPE of about 41.6 is the second-highest on record. Every previous reading above 30 was eventually followed by a large decline, though often years later.
2. Capital intensity and free cash flow. Heavy capital spending converts high-margin, asset-light software businesses into capital-intensive infrastructure businesses. Reported free cash flow at some hyperscalers declined in early 2026, and some firms have turned to large bond issues to fund spending. Capital-intensive industries historically earn lower returns on capital.
3. Circularity and concentration. A large share of AI revenue flows among a small group of interconnected firms: chipmakers sell to cloud providers, which sell to AI developers, which are partly funded by the cloud providers and chipmakers. A slowdown at any point could cascade. The top ten stocks are about 39% of the index, so a sector-level disappointment would hit the whole market.
4. The discount rate is moving the wrong way. Long-duration growth assets are most sensitive to rates. A 10-year yield above 5% is a material headwind that did not exist during the 2023–2025 phase of the rally. The September 2026 rate hike signals that the Fed's priority is inflation, not supporting asset prices.
5. Historical pattern of technology booms. DMS (2026) and T. Rowe Price (2025) both note that transformative technologies, including railroads and the internet, created enormous economic value while early investors often lost money because they overpaid. T. Rowe Price describes the current situation as two overlapping booms: a near-term one supported by measurable demand, and a longer-term speculative one.

This paper's assessment is that the evidence supports neither confident extreme. The current market does not resemble 1999 in profitability or in forward valuation. It does resemble the late 1990s in investment intensity, concentration, investor enthusiasm, and trailing valuation. The historical record suggests that the most likely outcome is lower-than-average long-horizon returns from current levels, with a meaningful probability of a large drawdown if AI earnings growth slows while yields stay high. The timing of any such adjustment cannot be forecast reliably.

15.5 Rates, Inflation, and the Federal Reserve

The macro backdrop has deteriorated for equities in 2026. Inflation, already above target before the war, was pushed up by the energy shock. Headline CPI was 3.4% year over year in August 2026, and the Cleveland Fed's nowcast projected about 3.6% for September. Core inflation was lower, with core CPI near 2.4% to 2.6% depending on the report. The FOMC's September 2026 projections show PCE inflation at 3.7% for 2026, falling to 2.3% in 2027, with a median projection of one more 25-basis-point hike in 2026.

The rise in long-term yields above 5% reflects higher inflation expectations, a larger term premium tied to large federal deficits and the May 2025 Moody's downgrade of U.S. sovereign debt, and a hawkish Fed. As Section 9 explains, equity duration makes this a direct valuation headwind. That stocks have risen despite it shows how much the market is relying on earnings growth.

15.6 Oil, Geopolitics, and Trade

The 2026 Iran war is the dominant geopolitical event of the year. Based on press and reference sources, the war began on February 28, 2026, the Strait of Hormuz was effectively closed in early March, and U.S. crude rose more than 35% in the first full week of March, its largest weekly gain since futures began trading in 1983. A Pakistan-brokered ceasefire in April was followed by renewed hostilities in July. As of mid-October 2026, shipping through the strait remained far below pre-war levels, and the UAE's state oil company estimated that full flows would not resume until 2027. Brent settled near $106 per barrel on October 13, 2026. Gasoline averaged about $4.33 per gallon in mid-October, compared with about $3.08 a year earlier.

For equities, the oil shock has three effects. It transfers profits toward energy producers, which have hit record highs. It raises inflation, which pushes up yields and keeps the Fed hawkish. It slows growth in energy-importing economies: the IMF's July 2026 update projected world GDP growth of 3.0% for 2026. The 1973–1974 oil shock, which DMS (2026) identify as one of the six worst global equity episodes since 1900, is the cautionary analogue. The 2026 market reaction has been far milder so far, which the ECB attributed partly to expectations that the shortage is temporary.

Trade policy remains a source of risk after the 2025 tariff shock, though it has been overshadowed by energy in 2026.

15.7 Retail Participation and Market Plumbing

Retail investors are a larger force than at any time since the late 1990s. Retail share of U.S. equity volume was about 18% in 2024, and estimates for 2025–2026 range from about 20% to 36% depending on the source and method. JPMorgan estimated a record 35% during the April 2025 volatility. Retail inflows into U.S. stocks reached a record of roughly $308 billion in 2025, according to Reuters as cited in secondary sources. Retail traders also account for about half of 0DTE index option volume.

Retail investors have been rewarded for buying dips during 2020–2026, which may have reinforced the behavior. Section 8's evidence on attention-driven trading and extrapolation suggests that this participation adds to momentum on the way up and could add to volatility in a sustained decline. The structural changes in Section 14, including extended trading hours and tokenized shares, will likely increase retail access further.

15.8 Global Markets Versus the United States

Non-U.S. stocks outperformed the U.S. by a wide margin over the year to early 2026. An ETF tracking the MSCI ACWI ex-USA index returned about 35.8% over the 12 months to late February 2026, against 13.0% for the S&P 500, and was up about 10.4% year to date in late February when the U.S. was roughly flat. The Iran war reversed much of that lead in March, as energy-importing regions such as Europe and Asia were hit harder than the United States, a net energy exporter. The global MSCI ACWI, including the U.S., was up about 13.9% year to date in early October 2026. This paper could not find a reliable current figure for non-U.S. stocks alone.

The long-run case for international diversification remains strong. DMS (2026) find that across 32 markets from 1974 to 2025, investors in most countries would have achieved better risk-adjusted returns by investing globally rather than domestically, especially with currency hedging. Non-U.S. markets trade at substantially lower valuation multiples than the U.S., partly because of sector composition and lower expected growth.

15.9 Summary Assessment

Applying the evidence from Sections 4–14:

• Valuation (Section 7): The CAPE implies below-average 10-year real returns. Confidence in the direction is moderate, and confidence in timing is low.
• Risk premium (Section 5): The implied ERP of about 4.2% is near its average, conditional on analysts' strong earnings forecasts being broadly right.
• Macro (Section 9): Rising yields, rising inflation, and a hiking Fed are headwinds that historically compress multiples.
• Concentration (Section 11): Index risk is unusually dependent on one theme.
• Crisis framework (Section 10): Leverage in the public equity market does not appear extreme, and the banking system is better capitalized than in 2008. A valuation-driven correction similar to 2000–2002 is a more plausible risk than a 2008-style systemic crisis. Private credit and AI-related debt financing are areas where leverage is harder to observe.

16.1 Reconciling Risk-Based and Behavioral Views

The long debate between efficient-markets and behavioral finance has largely converged on a middle position. Prices respond quickly to public news, and most investors cannot profitably exploit public information after costs. In that operational sense, markets are efficient. Expected returns, however, vary over time and across stocks in ways that are partly rational compensation for risk and partly the result of systematic investor errors that arbitrage cannot fully eliminate. The question is not whether markets are efficient, but how efficient they are, for which assets, and when.

The evidence also suggests a pattern. Risk-based explanations fit the aggregate equity premium and the value premium reasonably well. Behavioral explanations fit momentum, post-earnings drift, the low-risk anomaly, and bubbles better. Many factors probably reflect both.

16.2 Answers to the Research Questions

RQ1 (market structure). Electronic, fragmented markets have reduced trading costs for ordinary trades to historic lows. Price formation is fast and generally accurate. The costs are fragility in rare stress events, complexity, and reduced public price discovery from internalized retail flow.

RQ2 (equity risk premium). The historical U.S. premium is about 6% over bills and 5% over bonds, geometric, from 1900 to 2025. That is larger than standard theory explains and larger than investors should expect today. Forward-looking estimates are closer to 4%, and simple valuation measures suggest the premium may be lower still.

RQ3 (cross-section). A small set of factor themes is robust: value, momentum, profitability, investment, and low risk. Most of the hundreds of published anomalies are weaker than reported, concentrated in microcaps, or have decayed after publication.

RQ4 (predictability). Aggregate returns are modestly predictable at long horizons from valuation ratios. Short-horizon predictability is very weak, and most predictors fail out of sample. Valuation is useful for setting long-run expectations, not for market timing.

RQ5 (behavior and limits to arbitrage). Investor biases are well documented and affect prices where arbitrage is costly or risky. Bubbles are partly identifiable in advance from price acceleration, volatility, and issuance, but their timing remains unpredictable.

RQ6 (monetary policy). Monetary surprises move stock prices mainly through the equity risk premium. Inflation regimes determine whether bonds hedge stocks. In 2026, the high-inflation, positive-correlation regime has returned.

RQ7 (crises). Leverage determines the depth of crises, policy determines their length, and declines are getting faster. Post-2008 reforms strengthened banks, while risk has migrated to nonbank and market-based channels.

RQ8 (structural change). Passive investing is now the majority of U.S. fund assets, concentration is at a 100-year high, retail participation is near records, and AI investment dominates earnings and capital spending. These forces increase the market's dependence on a few firms and a single theme.

16.3 Confidence Ratings

Table 18. Confidence in major claims.

16.4 What Remains Unknown

The deepest open questions are: what fundamentally drives the equity premium; how much of the cross-section is risk versus mispricing; whether the shift to passive investing will eventually impair price discovery; whether AI produces durable, broad-based earnings growth or a capital-spending cycle that ends in overcapacity; and how new market structures such as overnight trading, tokenization, and 0DTE options behave in a severe crisis.

This paper has several limitations. First, it is a survey, not an original empirical study. It reports published estimates and does not re-estimate them on a common sample, so numbers from different studies are not always directly comparable. Second, it focuses on U.S. equities, with international evidence used mainly for comparison. Third, the current-conditions analysis in Section 15 relies on financial press, data trackers, and secondary summaries for 2026 figures, because official full-year data are not yet available. Several of these sources disagreed with each other, and the paper reports ranges where they did. Fourth, some current figures, including the latest 10-year yield, September 2026 CPI, and current non-U.S. returns, could not be confirmed from primary sources at the time of writing. Fifth, the paper has not been peer reviewed. Several historical figures for older crashes are taken from standard index histories and are approximate. Finally, readers should verify any figure that matters for a decision against the primary source.

1. AI and the equity premium. How do markets price general-purpose technologies during their investment phase? Comparative studies of railroads, electrification, telecommunications, and the internet could test whether capital-spending booms systematically lower subsequent returns for investors in the leading firms.
2. Passive ownership and price discovery. Better identification strategies are needed to estimate how rising passive ownership affects the informativeness of prices, the cost of capital, and corporate governance.
3. Demand-based asset pricing. The inelastic markets hypothesis and the broader demand-system approach (Koijen & Yogo, 2019) need further testing, including the effects of flows into retirement accounts and target-date funds on aggregate valuations.
4. Market structure under stress. How do overnight trading, tokenized securities, and short-dated options behave during a large shock? Regulators and academics should plan data collection before the next crisis.
5. Intangibles and valuation. Accounting measures of book value and earnings increasingly miss intangible investment. Better measures could change estimates of value premia and the meaning of CAPE.
6. Inflation regimes. The return of positive stock-bond correlation raises questions for portfolio construction that the 2000–2021 data cannot answer.
7. Geopolitical risk. The DMS finding of no general predictive relation between geopolitical threat indices and returns, alongside the severity of a few geopolitical episodes, calls for models of rare but extreme geopolitical shocks.
8. Retail investors and social media. Finer data on retail trading, options use, and social-media coordination could clarify when retail flows stabilize markets and when they destabilize them.
9. Machine learning and replication. Standards for pre-registration, out-of-sample reporting, and disclosure of the number of trials would improve the credibility of machine learning results in finance.
10. Private markets. As more value is created before companies go public, research must address how private-market returns, valuations, and risks interact with the public market.

A century and a quarter of data teaches a few durable lessons. Stocks have rewarded patient investors with real returns far above bonds and bills, in every major market with a continuous history, but with drawdowns that can last decades and destroy most of an investment in real terms. The size of that reward is still not fully explained by economic theory. The U.S. experience was exceptionally favorable, and investors should not assume it will repeat at the same scale.

Beyond the market factor, a few broad patterns, including value, momentum, profitability, investment, and low risk, have persisted across time and countries. Most of the hundreds of other published patterns have not. Aggregate returns are predictable in a weak, long-horizon sense from valuation, but not in a way that supports reliable market timing. Investor psychology matters most where arbitrage is hardest and during episodes of collective enthusiasm or fear. Crises are deepest when leverage is high and shortest when policy responds quickly.

Applied to October 2026, these lessons give a mixed picture. On one side are exceptional earnings growth, highly profitable market leaders, and an implied risk premium close to normal. On the other side are a CAPE near the all-time high, record concentration in one technological theme, 10-year yields above 5% for the first time since 2007, a central bank raising rates into an energy shock, and heavy retail and leveraged-options participation. The weight of historical evidence points to lower long-run returns from current levels than investors have experienced recently, with elevated risk of a large drawdown if AI-related earnings disappoint while rates stay high. History offers no reliable guide to when that might happen, or whether strong earnings growth will instead allow valuations to normalize gradually.

The most robust practical implications are also the oldest: diversify broadly, including internationally; keep costs low; match risk to the time horizon; and treat any confident short-term forecast, including this paper's assessments, with skepticism.

Each figure below is specified so that it can be reproduced from public data. The web version of this paper draws Figures 3, 4, 5, and 7 from the tables cited. It also draws simplified versions of Figure 1, showing ending values only, and Figure 2, showing selected readings only, because the full monthly series were not downloaded for this paper.

Figure 1. Growth of $1 invested in U.S. asset classes, 1900–2025 (nominal, log scale). Horizontal axis: year, 1900 to 2025. Vertical axis: cumulative value of $1 invested at the start of 1900, logarithmic scale from $1 to $1,000,000. Series: equities (ending $124,854), long-term government bonds ($284), Treasury bills ($69), and the consumer price index ($38). Source: DMS (2026), Table 5.

Figure 2. CAPE ratio, 1881–2026. Horizontal axis: month, January 1881 to October 2026. Vertical axis: cyclically adjusted price-to-earnings ratio, 0 to 50. Single series from Robert Shiller's online data, with a horizontal reference line at the long-run mean of about 17 and annotations at September 1929 (about 32), December 1999 (44.2), and October 2026 (about 41.6). Source: Shiller data; Multpl.

Figure 3. Peak-to-trough declines of major U.S. bear markets. Horizontal bar chart. Vertical axis: episode (1929–32, 1987, 2000–02 S&P 500, 2000–02 NASDAQ, 2007–09, 2020, 2022, 2025). Horizontal axis: decline in percent, 0 to 100. Source: Table 11.

Figure 4. Time to recover prior peak after major U.S. declines. Horizontal bar chart, same episodes as Figure 3. Horizontal axis: years from trough to prior peak, 0 to 25. Source: Table 11.

Figure 5. Share of the S&P 500 held by its ten largest constituents. Horizontal axis: date (September 2019, March 2026, December 2025 peak, September 2026). Vertical axis: percent of index market capitalization, 0 to 50. Source: Table 13.

Figure 6. Long-run U.S. factor premia by decade. Grouped bars by decade from the 1900s to the 2020s for size, value, momentum, low volatility, and income factors, plus the full-period average. Vertical axis: annualized premium in percent. Source: DMS (2026), Figure 94 of the UBS Yearbook summary edition. Only the full-period momentum premium of 7.7% is quoted in this paper's text. The decade values should be taken directly from the source.

Figure 7. SPX 0DTE options share of total SPX volume, 2025. Horizontal axis: month. Vertical axis: percent of SPX option volume, 0 to 70. Points: February 56%, May 61%, August 62.4%, full-year average 59%. Source: Cboe (2025).

Figure 8. 10-year Treasury yield and S&P 500 level, January to October 2026. Dual-axis line chart. Left axis: S&P 500 index level. Right axis: 10-year Treasury yield in percent. Annotate the Iran war start (February 28), the March 30 equity low, the September 16 rate hike, and the early-October yield peak of 5.36%. Source: FRED series SP500 and DGS10.

Table C1. Long-run returns across market groups.

Table C2. Recent federal funds rate history. Sources: Federal Reserve; press reports.

Table C3. Growth of zero-days-to-expiration options.

Search strategy. Sources were gathered on October 14 and 15, 2026, from academic journals, central bank and regulator publications, industry research, and financial press. The search covered index levels and returns, valuation measures, monetary policy, Treasury yields, inflation, oil and geopolitics, market concentration, passive and active fund flows, active manager performance, retail trading, options volume, AI capital expenditure, regulation, private markets, and long-run global returns. The UBS Global Investment Returns Yearbook 2026 public summary was consulted directly.

Source selection. Sources were ranked in four tiers: (1) peer-reviewed journals and primary data; (2) working papers, central banks, and regulators; (3) established industry research such as Damodaran Online, the UBS Yearbook, S&P SPIVA, Morningstar, FactSet, and Cboe; and (4) financial press, used only for recent events and current figures. Academic citations are to published articles whose findings are standard in the literature.

Handling conflicts. Where sources disagreed on a current figure, such as the S&P 500 close on October 14, 2026, or the retail share of trading volume, the paper reports a range and names the disagreement. Figures that could not be verified against a primary source are marked "approx." or "as reported."

Calculations. Equity premia in Section 5.2 are this paper's calculations from DMS geometric averages. Durations in Table 11 are approximations from index peak and trough dates.

Limitations of the method. Primary data series such as CRSP and the Kenneth French Data Library were not downloaded and re-analyzed. Several secondary 2026 sources are commentary sites and data trackers, and they are identified as such.

Academic Sources

Amihud, Y. (2002). Illiquidity and stock returns: Cross-section and time-series effects. Journal of Financial Markets, 5(1), 31–56. https://doi.org/10.1016/S1386-4181(01)00024-6

Amihud, Y., & Mendelson, H. (1986). Asset pricing and the bid-ask spread. Journal of Financial Economics, 17(2), 223–249. https://doi.org/10.1016/0304-405X(86)90065-6

Ang, A., & Chen, J. (2002). Asymmetric correlations of equity portfolios. Journal of Financial Economics, 63(3), 443–494. https://doi.org/10.1016/S0304-405X(02)00068-5

Ang, A., Hodrick, R. J., Xing, Y., & Zhang, X. (2006). The cross-section of volatility and expected returns. Journal of Finance, 61(1), 259–299. https://doi.org/10.1111/j.1540-6261.2006.00836.x

Arnott, R. D., Harvey, C. R., Kalesnik, V., & Linnainmaa, J. T. (2021). Reports of value's death may be greatly exaggerated. Financial Analysts Journal, 77(1), 44–67. https://doi.org/10.1080/0015198X.2020.1842704

Asness, C. S., Frazzini, A., Israel, R., Moskowitz, T. J., & Pedersen, L. H. (2018). Size matters, if you control your junk. Journal of Financial Economics, 129(3), 479–509. https://doi.org/10.1016/j.jfineco.2018.05.006

Asness, C. S., Frazzini, A., & Pedersen, L. H. (2019). Quality minus junk. Review of Accounting Studies, 24(1), 34–112. https://doi.org/10.1007/s11142-018-9470-2

Asness, C. S., Moskowitz, T. J., & Pedersen, L. H. (2013). Value and momentum everywhere. Journal of Finance, 68(3), 929–985. https://doi.org/10.1111/jofi.12021

Bailey, D. H., Borwein, J. M., López de Prado, M., & Zhu, Q. J. (2014). Pseudo-mathematics and financial charlatanism: The effects of backtest overfitting on out-of-sample performance. Notices of the American Mathematical Society, 61(5), 458–471. https://doi.org/10.1090/noti1105

Bailey, D. H., & López de Prado, M. (2014). The deflated Sharpe ratio: Correcting for selection bias, backtest overfitting, and non-normality. Journal of Portfolio Management, 40(5), 94–107. https://doi.org/10.3905/jpm.2014.40.5.094

Baker, M., & Wurgler, J. (2006). Investor sentiment and the cross-section of stock returns. Journal of Finance, 61(4), 1645–1680. https://doi.org/10.1111/j.1540-6261.2006.00885.x

Baker, S. R., Bloom, N., Davis, S. J., Kost, K., Sammon, M., & Viratyosin, T. (2020). The unprecedented stock market reaction to COVID-19. Review of Asset Pricing Studies, 10(4), 742–758. https://doi.org/10.1093/rapstu/raaa008

Ball, R., & Brown, P. (1968). An empirical evaluation of accounting income numbers. Journal of Accounting Research, 6(2), 159–178. https://doi.org/10.2307/2490232

Bansal, R., & Yaron, A. (2004). Risks for the long run: A potential resolution of asset pricing puzzles. Journal of Finance, 59(4), 1481–1509. https://doi.org/10.1111/j.1540-6261.2004.00670.x

Banz, R. W. (1981). The relationship between return and market value of common stocks. Journal of Financial Economics, 9(1), 3–18. https://doi.org/10.1016/0304-405X(81)90018-0

Barber, B. M., Huang, X., Odean, T., & Schwarz, C. (2022). Attention-induced trading and returns: Evidence from Robinhood users. Journal of Finance, 77(6), 3141–3190. https://doi.org/10.1111/jofi.13183

Barber, B. M., & Odean, T. (2000). Trading is hazardous to your wealth: The common stock investment performance of individual investors. Journal of Finance, 55(2), 773–806. https://doi.org/10.1111/0022-1082.00226

Barber, B. M., & Odean, T. (2001). Boys will be boys: Gender, overconfidence, and common stock investment. Quarterly Journal of Economics, 116(1), 261–292. https://doi.org/10.1162/003355301556400

Barber, B. M., & Odean, T. (2008). All that glitters: The effect of attention and news on the buying behavior of individual and institutional investors. Review of Financial Studies, 21(2), 785–818. https://doi.org/10.1093/rfs/hhm079

Barberis, N., Greenwood, R., Jin, L., & Shleifer, A. (2018). Extrapolation and bubbles. Journal of Financial Economics, 129(2), 203–227. https://doi.org/10.1016/j.jfineco.2018.04.007

Barberis, N., & Huang, M. (2008). Stocks as lotteries: The implications of probability weighting for security prices. American Economic Review, 98(5), 2066–2100. https://doi.org/10.1257/aer.98.5.2066

Barberis, N., Shleifer, A., & Vishny, R. (1998). A model of investor sentiment. Journal of Financial Economics, 49(3), 307–343. https://doi.org/10.1016/S0304-405X(98)00027-0

Barberis, N., & Thaler, R. (2003). A survey of behavioral finance. In G. M. Constantinides, M. Harris, & R. M. Stulz (Eds.), Handbook of the economics of finance (Vol. 1B, pp. 1053–1128). Elsevier. https://doi.org/10.1016/S1574-0102(03)01027-6

Barro, R. J. (2006). Rare disasters and asset markets in the twentieth century. Quarterly Journal of Economics, 121(3), 823–866. https://doi.org/10.1162/qjec.121.3.823

Basu, S. (1977). Investment performance of common stocks in relation to their price-earnings ratios: A test of the efficient market hypothesis. Journal of Finance, 32(3), 663–682. https://doi.org/10.1111/j.1540-6261.1977.tb01979.x

Battalio, R., Corwin, S. A., & Jennings, R. (2016). Can brokers have it all? On the relation between make-take fees and limit order execution quality. Journal of Finance, 71(5), 2193–2238. https://doi.org/10.1111/jofi.12422

Beber, A., & Pagano, M. (2013). Short-selling bans around the world: Evidence from the 2007–09 crisis. Journal of Finance, 68(1), 343–381. https://doi.org/10.1111/j.1540-6261.2012.01802.x

Benartzi, S., & Thaler, R. H. (1995). Myopic loss aversion and the equity premium puzzle. Quarterly Journal of Economics, 110(1), 73–92. https://doi.org/10.2307/2118511

Berk, J. B., & Green, R. C. (2004). Mutual fund flows and performance in rational markets. Journal of Political Economy, 112(6), 1269–1295. https://doi.org/10.1086/424739

Bernanke, B. S., & Kuttner, K. N. (2005). What explains the stock market's reaction to Federal Reserve policy? Journal of Finance, 60(3), 1221–1257. https://doi.org/10.1111/j.1540-6261.2005.00760.x

Bernard, V. L., & Thomas, J. K. (1989). Post-earnings-announcement drift: Delayed price response or risk premium? Journal of Accounting Research, 27(Supplement), 1–36. https://doi.org/10.2307/2491062

Bessembinder, H. (2018). Do stocks outperform Treasury bills? Journal of Financial Economics, 129(3), 440–457. https://doi.org/10.1016/j.jfineco.2018.06.004

Black, F. (1972). Capital market equilibrium with restricted borrowing. Journal of Business, 45(3), 444–455. https://doi.org/10.1086/295472

Black, F. (1976). Studies of stock price volatility changes. Proceedings of the 1976 Meetings of the American Statistical Association, Business and Economic Statistics Section, 177–181.

Black, F., Jensen, M. C., & Scholes, M. (1972). The capital asset pricing model: Some empirical tests. In M. C. Jensen (Ed.), Studies in the theory of capital markets (pp. 79–121). Praeger.

Bollerslev, T. (1986). Generalized autoregressive conditional heteroskedasticity. Journal of Econometrics, 31(3), 307–327. https://doi.org/10.1016/0304-4076(86)90063-1

Bollerslev, T., Tauchen, G., & Zhou, H. (2009). Expected stock returns and variance risk premia. Review of Financial Studies, 22(11), 4463–4492. https://doi.org/10.1093/rfs/hhp008

Breeden, D. T. (1979). An intertemporal asset pricing model with stochastic consumption and investment opportunities. Journal of Financial Economics, 7(3), 265–296. https://doi.org/10.1016/0304-405X(79)90016-3

Brogaard, J., Hendershott, T., & Riordan, R. (2014). High-frequency trading and price discovery. Review of Financial Studies, 27(8), 2267–2306. https://doi.org/10.1093/rfs/hhu032

Brunnermeier, M. K. (2009). Deciphering the liquidity and credit crunch 2007–2008. Journal of Economic Perspectives, 23(1), 77–100. https://doi.org/10.1257/jep.23.1.77

Brunnermeier, M. K., & Nagel, S. (2004). Hedge funds and the technology bubble. Journal of Finance, 59(5), 2013–2040. https://doi.org/10.1111/j.1540-6261.2004.00690.x

Budish, E., Cramton, P., & Shim, J. (2015). The high-frequency trading arms race: Frequent batch auctions as a market design response. Quarterly Journal of Economics, 130(4), 1547–1621. https://doi.org/10.1093/qje/qjv027

Campbell, J. Y. (1987). Stock returns and the term structure. Journal of Financial Economics, 18(2), 373–399. https://doi.org/10.1016/0304-405X(87)90045-6

Campbell, J. Y., & Cochrane, J. H. (1999). By force of habit: A consumption-based explanation of aggregate stock market behavior. Journal of Political Economy, 107(2), 205–251. https://doi.org/10.1086/250059

Campbell, J. Y., & Shiller, R. J. (1988). The dividend-price ratio and expectations of future dividends and discount factors. Review of Financial Studies, 1(3), 195–228. https://doi.org/10.1093/rfs/1.3.195

Campbell, J. Y., & Shiller, R. J. (1998). Valuation ratios and the long-run stock market outlook. Journal of Portfolio Management, 24(2), 11–26. https://doi.org/10.3905/jpm.24.2.11

Campbell, J. Y., & Thompson, S. B. (2008). Predicting excess stock returns out of sample: Can anything beat the historical average? Review of Financial Studies, 21(4), 1509–1531. https://doi.org/10.1093/rfs/hhm055

Carhart, M. M. (1997). On persistence in mutual fund performance. Journal of Finance, 52(1), 57–82. https://doi.org/10.1111/j.1540-6261.1997.tb03808.x

Carr, P., & Wu, L. (2009). Variance risk premiums. Review of Financial Studies, 22(3), 1311–1341. https://doi.org/10.1093/rfs/hhn038

Chen, A. Y., & Zimmermann, T. (2022). Open source cross-sectional asset pricing. Critical Finance Review, 11(2), 207–264. https://doi.org/10.1561/104.00000112

Chen, N.-F., Roll, R., & Ross, S. A. (1986). Economic forces and the stock market. Journal of Business, 59(3), 383–403. https://doi.org/10.1086/296344

Christie, W. G., & Schultz, P. H. (1994). Why do NASDAQ market makers avoid odd-eighth quotes? Journal of Finance, 49(5), 1813–1840. https://doi.org/10.1111/j.1540-6261.1994.tb04782.x

Cochrane, J. H. (2005). Asset pricing (Rev. ed.). Princeton University Press.

Cochrane, J. H. (2008). The dog that did not bark: A defense of return predictability. Review of Financial Studies, 21(4), 1533–1575. https://doi.org/10.1093/rfs/hhm046

Cochrane, J. H. (2011). Presidential address: Discount rates. Journal of Finance, 66(4), 1047–1108. https://doi.org/10.1111/j.1540-6261.2011.01671.x

Cooper, M. J., Gulen, H., & Schill, M. J. (2008). Asset growth and the cross-section of stock returns. Journal of Finance, 63(4), 1609–1651. https://doi.org/10.1111/j.1540-6261.2008.01370.x

Daniel, K., Hirshleifer, D., & Subrahmanyam, A. (1998). Investor psychology and security market under- and overreactions. Journal of Finance, 53(6), 1839–1885. https://doi.org/10.1111/0022-1082.00077

Daniel, K., & Moskowitz, T. J. (2016). Momentum crashes. Journal of Financial Economics, 122(2), 221–247. https://doi.org/10.1016/j.jfineco.2015.12.002

De Long, J. B., Shleifer, A., Summers, L. H., & Waldmann, R. J. (1990). Noise trader risk in financial markets. Journal of Political Economy, 98(4), 703–738. https://doi.org/10.1086/261703

Dimson, E., Marsh, P., & Staunton, M. (2002). Triumph of the optimists: 101 years of global investment returns. Princeton University Press.

Doidge, C., Karolyi, G. A., & Stulz, R. M. (2017). The U.S. listing gap. Journal of Financial Economics, 123(3), 464–487. https://doi.org/10.1016/j.jfineco.2016.12.002

Engle, R. F. (1982). Autoregressive conditional heteroscedasticity with estimates of the variance of United Kingdom inflation. Econometrica, 50(4), 987–1007. https://doi.org/10.2307/1912773

Fama, E. F. (1970). Efficient capital markets: A review of theory and empirical work. Journal of Finance, 25(2), 383–417. https://doi.org/10.2307/2325486

Fama, E. F. (1981). Stock returns, real activity, inflation, and money. American Economic Review, 71(4), 545–565.

Fama, E. F. (1991). Efficient capital markets: II. Journal of Finance, 46(5), 1575–1617. https://doi.org/10.1111/j.1540-6261.1991.tb04636.x

Fama, E. F., Fisher, L., Jensen, M. C., & Roll, R. (1969). The adjustment of stock prices to new information. International Economic Review, 10(1), 1–21. https://doi.org/10.2307/2525569

Fama, E. F., & French, K. R. (1988). Dividend yields and expected stock returns. Journal of Financial Economics, 22(1), 3–25. https://doi.org/10.1016/0304-405X(88)90020-7

Fama, E. F., & French, K. R. (1989). Business conditions and expected returns on stocks and bonds. Journal of Financial Economics, 25(1), 23–49. https://doi.org/10.1016/0304-405X(89)90095-0

Fama, E. F., & French, K. R. (1992). The cross-section of expected stock returns. Journal of Finance, 47(2), 427–465. https://doi.org/10.1111/j.1540-6261.1992.tb04398.x

Fama, E. F., & French, K. R. (1993). Common risk factors in the returns on stocks and bonds. Journal of Financial Economics, 33(1), 3–56. https://doi.org/10.1016/0304-405X(93)90023-5

Fama, E. F., & French, K. R. (2002). The equity premium. Journal of Finance, 57(2), 637–659. https://doi.org/10.1111/1540-6261.00437

Fama, E. F., & French, K. R. (2010). Luck versus skill in the cross-section of mutual fund returns. Journal of Finance, 65(5), 1915–1947. https://doi.org/10.1111/j.1540-6261.2010.01598.x

Fama, E. F., & French, K. R. (2015). A five-factor asset pricing model. Journal of Financial Economics, 116(1), 1–22. https://doi.org/10.1016/j.jfineco.2014.10.010

Fama, E. F., & MacBeth, J. D. (1973). Risk, return, and equilibrium: Empirical tests. Journal of Political Economy, 81(3), 607–636. https://doi.org/10.1086/260061

Frazzini, A., & Pedersen, L. H. (2014). Betting against beta. Journal of Financial Economics, 111(1), 1–25. https://doi.org/10.1016/j.jfineco.2013.10.005

Friedman, M., & Schwartz, A. J. (1963). A monetary history of the United States, 1867–1960. Princeton University Press.

Froot, K. A., & Dabora, E. M. (1999). How are stock prices affected by the location of trade? Journal of Financial Economics, 53(2), 189–216. https://doi.org/10.1016/S0304-405X(99)00020-3

Gabaix, X., & Koijen, R. S. J. (2021). In search of the origins of financial fluctuations: The inelastic markets hypothesis (NBER Working Paper No. 28967). National Bureau of Economic Research. https://doi.org/10.3386/w28967

Galbraith, J. K. (1954). The great crash, 1929. Houghton Mifflin.

Gibbons, M. R., Ross, S. A., & Shanken, J. (1989). A test of the efficiency of a given portfolio. Econometrica, 57(5), 1121–1152. https://doi.org/10.2307/1913625

Glosten, L. R., Jagannathan, R., & Runkle, D. E. (1993). On the relation between the expected value and the volatility of the nominal excess return on stocks. Journal of Finance, 48(5), 1779–1801. https://doi.org/10.1111/j.1540-6261.1993.tb05128.x

Glosten, L. R., & Milgrom, P. R. (1985). Bid, ask and transaction prices in a specialist market with heterogeneously informed traders. Journal of Financial Economics, 14(1), 71–100. https://doi.org/10.1016/0304-405X(85)90044-3

Gordon, M. J. (1962). The investment, financing, and valuation of the corporation. Richard D. Irwin.

Gorton, G., & Metrick, A. (2012). Securitized banking and the run on repo. Journal of Financial Economics, 104(3), 425–451. https://doi.org/10.1016/j.jfineco.2011.03.016

Goyal, A., & Welch, I. (2008). A comprehensive look at the empirical performance of equity premium prediction. Review of Financial Studies, 21(4), 1455–1508. https://doi.org/10.1093/rfs/hhm014

Graham, J. R., & Harvey, C. R. (2001). The theory and practice of corporate finance: Evidence from the field. Journal of Financial Economics, 60(2–3), 187–243. https://doi.org/10.1016/S0304-405X(01)00044-7

Greenwood, R., & Shleifer, A. (2014). Expectations of returns and expected returns. Review of Financial Studies, 27(3), 714–746. https://doi.org/10.1093/rfs/hht082

Greenwood, R., Shleifer, A., & You, Y. (2019). Bubbles for Fama. Journal of Financial Economics, 131(1), 20–43. https://doi.org/10.1016/j.jfineco.2018.09.002

Grinblatt, M., & Han, B. (2005). Prospect theory, mental accounting, and momentum. Journal of Financial Economics, 78(2), 311–339. https://doi.org/10.1016/j.jfineco.2004.10.006

Grossman, S. J., & Stiglitz, J. E. (1980). On the impossibility of informationally efficient markets. American Economic Review, 70(3), 393–408.

Gu, S., Kelly, B., & Xiu, D. (2020). Empirical asset pricing via machine learning. Review of Financial Studies, 33(5), 2223–2273. https://doi.org/10.1093/rfs/hhaa009

Hansen, L. P., & Jagannathan, R. (1991). Implications of security market data for models of dynamic economies. Journal of Political Economy, 99(2), 225–262. https://doi.org/10.1086/261749

Harris, L., & Gurel, E. (1986). Price and volume effects associated with changes in the S&P 500 list: New evidence for the existence of price pressures. Journal of Finance, 41(4), 815–829. https://doi.org/10.1111/j.1540-6261.1986.tb04550.x

Harvey, C. R., Liu, Y., & Zhu, H. (2016). … and the cross-section of expected returns. Review of Financial Studies, 29(1), 5–68. https://doi.org/10.1093/rfs/hhv059

Hendershott, T., Jones, C. M., & Menkveld, A. J. (2011). Does algorithmic trading improve liquidity? Journal of Finance, 66(1), 1–33. https://doi.org/10.1111/j.1540-6261.2010.01624.x

Hou, K., Xue, C., & Zhang, L. (2015). Digesting anomalies: An investment approach. Review of Financial Studies, 28(3), 650–705. https://doi.org/10.1093/rfs/hhu068

Hou, K., Xue, C., & Zhang, L. (2020). Replicating anomalies. Review of Financial Studies, 33(5), 2019–2133. https://doi.org/10.1093/rfs/hhy131

Huang, R. D., & Stoll, H. R. (1996). Dealer versus auction markets: A paired comparison of execution costs on NASDAQ and the NYSE. Journal of Financial Economics, 41(3), 313–357. https://doi.org/10.1016/0304-405X(95)00867-E

Jegadeesh, N. (1990). Evidence of predictable behavior of security returns. Journal of Finance, 45(3), 881–898. https://doi.org/10.1111/j.1540-6261.1990.tb05110.x

Jegadeesh, N., & Titman, S. (1993). Returns to buying winners and selling losers: Implications for stock market efficiency. Journal of Finance, 48(1), 65–91. https://doi.org/10.1111/j.1540-6261.1993.tb04702.x

Jensen, M. C. (1968). The performance of mutual funds in the period 1945–1964. Journal of Finance, 23(2), 389–416. https://doi.org/10.1111/j.1540-6261.1968.tb00815.x

Jensen, T. I., Kelly, B., & Pedersen, L. H. (2023). Is there a replication crisis in finance? Journal of Finance, 78(5), 2465–2518. https://doi.org/10.1111/jofi.13249

Jorion, P., & Goetzmann, W. N. (1999). Global stock markets in the twentieth century. Journal of Finance, 54(3), 953–980. https://doi.org/10.1111/0022-1082.00133

Kahneman, D., & Tversky, A. (1979). Prospect theory: An analysis of decision under risk. Econometrica, 47(2), 263–291. https://doi.org/10.2307/1914185

Kelly, B., Pruitt, S., & Su, Y. (2019). Characteristics are covariances: A unified model of risk and return. Journal of Financial Economics, 134(3), 501–524. https://doi.org/10.1016/j.jfineco.2019.05.001

Kindleberger, C. P., & Aliber, R. Z. (2005). Manias, panics, and crashes: A history of financial crises (5th ed.). Wiley.

Kirilenko, A., Kyle, A. S., Samadi, M., & Tuzun, T. (2017). The Flash Crash: High-frequency trading in an electronic market. Journal of Finance, 72(3), 967–998. https://doi.org/10.1111/jofi.12498

Koijen, R. S. J., & Yogo, M. (2019). A demand system approach to asset pricing. Journal of Political Economy, 127(4), 1475–1515. https://doi.org/10.1086/701683

Kyle, A. S. (1985). Continuous auctions and insider trading. Econometrica, 53(6), 1315–1335. https://doi.org/10.2307/1913210

Lakonishok, J., Shleifer, A., & Vishny, R. W. (1994). Contrarian investment, extrapolation, and risk. Journal of Finance, 49(5), 1541–1578. https://doi.org/10.1111/j.1540-6261.1994.tb04772.x

Lamont, O. A., & Thaler, R. H. (2003). Can the market add and subtract? Mispricing in tech stock carve-outs. Journal of Political Economy, 111(2), 227–268. https://doi.org/10.1086/367683

Lehmann, B. N. (1990). Fads, martingales, and market efficiency. Quarterly Journal of Economics, 105(1), 1–28. https://doi.org/10.2307/2937816

Lettau, M., & Ludvigson, S. (2001). Consumption, aggregate wealth, and expected stock returns. Journal of Finance, 56(3), 815–849. https://doi.org/10.1111/0022-1082.00347

Lintner, J. (1965). The valuation of risk assets and the selection of risky investments in stock portfolios and capital budgets. Review of Economics and Statistics, 47(1), 13–37. https://doi.org/10.2307/1924119

Lo, A. W. (2004). The adaptive markets hypothesis: Market efficiency from an evolutionary perspective. Journal of Portfolio Management, 30(5), 15–29. https://doi.org/10.3905/jpm.2004.442611

Longin, F., & Solnik, B. (2001). Extreme correlation of international equity markets. Journal of Finance, 56(2), 649–676. https://doi.org/10.1111/0022-1082.00340

Lucas, R. E., Jr. (1978). Asset prices in an exchange economy. Econometrica, 46(6), 1429–1445. https://doi.org/10.2307/1913837

Lucca, D. O., & Moench, E. (2015). The pre-FOMC announcement drift. Journal of Finance, 70(1), 329–371. https://doi.org/10.1111/jofi.12196

MacKinlay, A. C. (1997). Event studies in economics and finance. Journal of Economic Literature, 35(1), 13–39.

Markowitz, H. (1952). Portfolio selection. Journal of Finance, 7(1), 77–91. https://doi.org/10.1111/j.1540-6261.1952.tb01525.x

McLean, R. D., & Pontiff, J. (2016). Does academic research destroy stock return predictability? Journal of Finance, 71(1), 5–32. https://doi.org/10.1111/jofi.12365

Mehra, R., & Prescott, E. C. (1985). The equity premium: A puzzle. Journal of Monetary Economics, 15(2), 145–161. https://doi.org/10.1016/0304-3932(85)90061-3

Merton, R. C. (1973). An intertemporal capital asset pricing model. Econometrica, 41(5), 867–887. https://doi.org/10.2307/1913811

Modigliani, F., & Cohn, R. A. (1979). Inflation, rational valuation and the market. Financial Analysts Journal, 35(2), 24–44. https://doi.org/10.2469/faj.v35.n2.24

Mossin, J. (1966). Equilibrium in a capital asset market. Econometrica, 34(4), 768–783. https://doi.org/10.2307/1910098

Nelson, D. B. (1991). Conditional heteroskedasticity in asset returns: A new approach. Econometrica, 59(2), 347–370. https://doi.org/10.2307/2938260

Newey, W. K., & West, K. D. (1987). A simple, positive semi-definite, heteroskedasticity and autocorrelation consistent covariance matrix. Econometrica, 55(3), 703–708. https://doi.org/10.2307/1913610

Novy-Marx, R. (2013). The other side of value: The gross profitability premium. Journal of Financial Economics, 108(1), 1–28. https://doi.org/10.1016/j.jfineco.2013.01.003

Odean, T. (1998). Are investors reluctant to realize their losses? Journal of Finance, 53(5), 1775–1798. https://doi.org/10.1111/0022-1082.00072

Ofek, E., & Richardson, M. (2003). DotCom mania: The rise and fall of internet stock prices. Journal of Finance, 58(3), 1113–1137. https://doi.org/10.1111/1540-6261.00560

Pástor, Ľ., & Stambaugh, R. F. (2003). Liquidity risk and expected stock returns. Journal of Political Economy, 111(3), 642–685. https://doi.org/10.1086/374184

Petersen, M. A. (2009). Estimating standard errors in finance panel data sets: Comparing approaches. Review of Financial Studies, 22(1), 435–480. https://doi.org/10.1093/rfs/hhn053

Ramelli, S., & Wagner, A. F. (2020). Feverish stock price reactions to COVID-19. Review of Corporate Finance Studies, 9(3), 622–655. https://doi.org/10.1093/rcfs/cfaa012

Rietz, T. A. (1988). The equity risk premium: A solution. Journal of Monetary Economics, 22(1), 117–131. https://doi.org/10.1016/0304-3932(88)90172-9

Roll, R. (1977). A critique of the asset pricing theory's tests. Part I: On past and potential testability of the theory. Journal of Financial Economics, 4(2), 129–176. https://doi.org/10.1016/0304-405X(77)90009-5

Roll, R. (1984). A simple implicit measure of the effective bid-ask spread in an efficient market. Journal of Finance, 39(4), 1127–1139. https://doi.org/10.1111/j.1540-6261.1984.tb03897.x

Romer, C. D. (1990). The great crash and the onset of the Great Depression. Quarterly Journal of Economics, 105(3), 597–624. https://doi.org/10.2307/2937892

Ross, S. A. (1976). The arbitrage theory of capital asset pricing. Journal of Economic Theory, 13(3), 341–360. https://doi.org/10.1016/0022-0531(76)90046-6

Savor, P., & Wilson, M. (2013). How much do investors care about macroeconomic risk? Evidence from scheduled economic announcements. Journal of Financial and Quantitative Analysis, 48(2), 343–375. https://doi.org/10.1017/S0022109013000136

Shanken, J. (1992). On the estimation of beta-pricing models. Review of Financial Studies, 5(1), 1–33. https://doi.org/10.1093/rfs/5.1.1

Sharpe, W. F. (1964). Capital asset prices: A theory of market equilibrium under conditions of risk. Journal of Finance, 19(3), 425–442. https://doi.org/10.1111/j.1540-6261.1964.tb02865.x

Sharpe, W. F. (1991). The arithmetic of active management. Financial Analysts Journal, 47(1), 7–9. https://doi.org/10.2469/faj.v47.n1.7

Shefrin, H., & Statman, M. (1985). The disposition to sell winners too early and ride losers too long: Theory and evidence. Journal of Finance, 40(3), 777–790. https://doi.org/10.1111/j.1540-6261.1985.tb05002.x

Shiller, R. J. (1981). Do stock prices move too much to be justified by subsequent changes in dividends? American Economic Review, 71(3), 421–436.

Shleifer, A. (1986). Do demand curves for stocks slope down? Journal of Finance, 41(3), 579–590. https://doi.org/10.1111/j.1540-6261.1986.tb04518.x

Shleifer, A., & Vishny, R. W. (1997). The limits of arbitrage. Journal of Finance, 52(1), 35–55. https://doi.org/10.1111/j.1540-6261.1997.tb03807.x

Shumway, T. (1997). The delisting bias in CRSP data. Journal of Finance, 52(1), 327–340. https://doi.org/10.1111/j.1540-6261.1997.tb03818.x

Siegel, J. J. (2016). The Shiller CAPE ratio: A new look. Financial Analysts Journal, 72(3), 41–50. https://doi.org/10.2469/faj.v72.n3.1

Sloan, R. G. (1996). Do stock prices fully reflect information in accruals and cash flows about future earnings? Accounting Review, 71(3), 289–315.

Stambaugh, R. F. (1999). Predictive regressions. Journal of Financial Economics, 54(3), 375–421. https://doi.org/10.1016/S0304-405X(99)00041-0

Stambaugh, R. F., Yu, J., & Yuan, Y. (2012). The short of it: Investor sentiment and anomalies. Journal of Financial Economics, 104(2), 288–302. https://doi.org/10.1016/j.jfineco.2011.12.001

Titman, S., Wei, K. C. J., & Xie, F. (2004). Capital investments and stock returns. Journal of Financial and Quantitative Analysis, 39(4), 677–700. https://doi.org/10.1017/S0022109000003173

Tversky, A., & Kahneman, D. (1992). Advances in prospect theory: Cumulative representation of uncertainty. Journal of Risk and Uncertainty, 5(4), 297–323. https://doi.org/10.1007/BF00122574

Weil, P. (1989). The equity premium puzzle and the risk-free rate puzzle. Journal of Monetary Economics, 24(3), 401–421. https://doi.org/10.1016/0304-3932(89)90028-7

Welch, I. (2000). Views of financial economists on the equity premium and on professional controversies. Journal of Business, 73(4), 501–537. https://doi.org/10.1086/209653

Whaley, R. E. (2000). The investor fear gauge. Journal of Portfolio Management, 26(3), 12–17. https://doi.org/10.3905/jpm.2000.319728

Wurgler, J. (2010). On the economic consequences of index-linked investing (NBER Working Paper No. 16376). National Bureau of Economic Research. https://doi.org/10.3386/w16376

Zhang, L. (2005). The value premium. Journal of Finance, 60(1), 67–103. https://doi.org/10.1111/j.1540-6261.2005.00725.x

Data, Industry, Regulatory, and Press Sources (2024–2026)

Brady Commission. (1988). Report of the Presidential Task Force on Market Mechanisms. U.S. Government Printing Office.

Cboe Global Markets. (2025). SPX 0DTE options jump to record 62% share in August [and related 2025 monthly posts]. https://www.cboe.com/insights/posts/spx-0-dte-options-jump-to-record-62-share-in-august

CNBC. (2026, October 5–6). S&P 500 posts first close above 7,800; A look at the S&P 500's trip to a new record. https://www.cnbc.com/2026/10/06/chart-a-look-at-the-sp-500s-remarkable-and-defiant-trip-a-new-record.html

CNBC. (2026, September 16). Fed rate decision September 2026. https://www.cnbc.com/2026/09/16/fed-rate-decision-september-2026.html

Cooley LLP. (2026, May 11). The SEC's semiannual reporting proposal. https://www.cooley.com/news/insight/2026/2026-05-11-the-secs-semiannual-reporting-proposal-fare-thee-well-quarterly-reporting

Damodaran, A. (2026). Damodaran Online: Implied equity risk premium updates. NYU Stern School of Business. https://pages.stern.nyu.edu/~adamodar/

Dimson, E., Marsh, P., & Staunton, M. (2026). UBS Global Investment Returns Yearbook 2026: Public summary edition. UBS. https://www.ubs.com/global/en/investment-bank/insights-and-data/articles/global-investment-returns-yearbook-2026.html

E*TRADE from Morgan Stanley. (2025, December 10). Fed closes 2025 with cut. https://us.etrade.com/knowledge/library/perspectives/market-happenings/fomc-dec2025-25bps

European Central Bank. (2026, July 27). Energy shock: Why oil and gas prices have risen less than expected [Blog post]. https://www.ecb.europa.eu/press/blog/date/2026/html/ecb.blog20260727~1212bdb8f9.en.html

FactSet. (2026, October 9). Earnings Insight. https://insight.factset.com/

Federal Reserve Board. (2026, September 16). Federal Reserve issues FOMC statement. https://www.federalreserve.gov/monetarypolicy/files/monetary20260916a1.pdf

Hargreaves Lansdown. (2024). Passive eclipses active in US fund market as assets swell to $13.3tn [Reporting Morningstar data]. https://www.hl.co.uk/news/passive-eclipses-active-in-us-fund-market-as-assets-swell-to-133tn

NBC News. (2026, October). Global bond sell-off pushes U.S. Treasury yields to fresh 24-year highs. https://www.nbcnews.com/business/markets/treasury-yields-hit-24-year-highs-global-sell-accelerates-rcna602045

Paul Hastings LLP. (2025, June). SEC withdraws 14 pending rule proposals. https://www.paulhastings.com/insights/client-alerts/sec-withdraws-14-pending-rule-proposals

S&P Dow Jones Indices. (2025). U.S. equities market attributes [Commentary on the April 2025 drawdown]. https://www.spglobal.com/spdji/en/commentary/article/us-equities-market-attributes

S&P Dow Jones Indices. (2026). SPIVA U.S. scorecard: Year-end 2025. https://www.spglobal.com/spdji/en/spiva/article/spiva-us-year-end-2025/

Skadden, Arps, Slate, Meagher & Flom LLP. (2026, September). SEC semiannual reporting proposal draws a record 200,000 comments. https://www.skadden.com/insights/publications/2026/09/insights-september-2026/sec-semiannual-reporting

T. Rowe Price. (2025). Hyperscalers: Is the AI boom dot-com redux? https://www.troweprice.com/financial-intermediary/is/en/thinking/articles/2025/q4/hyperscalers-is-the-ai-boom-dot-com-redux.html

TD Economics. (2026). Every time is different: 2026's energy shock. https://economics.td.com/us-energy-shock-2026

U.S. Securities and Exchange Commission. (2021). Staff report on equity and options market structure conditions in early 2021. https://www.sec.gov/files/staff-report-equity-options-market-struction-conditions-early-2021.pdf

U.S. Securities and Exchange Commission. (2026, May 5). SEC proposes amendments to permit optional semiannual reporting for public companies [Press release 2026-42]. https://www.sec.gov/newsroom/press-releases/2026-42-sec-proposes-amendments-permit-optional-semiannual-reporting-public-companies

YCharts. (2026). S&P 500 annual total return. https://ycharts.com/indicators/sandp_500_total_return_annual

Secondary data trackers consulted for current figures, identified in the text where used: Multpl (CAPE), ChartRow (year-to-date returns, concentration, drawdowns), Axios, Bloomberg, and Al Jazeera (2026 rates and geopolitics), AAA (gasoline prices), and company earnings guidance as summarized in financial press (AI capital expenditure).