Victor Haghani on Death of Random Walk, and Passive, Buybacks, and LTCM

Victor Haghani's paper "Who Killed the Random Walk," accepted into the Journal of Investment Management, builds a multi-agent model with three investor types: value investors, static investors, and extrapolators. From a reasonable parameter set the model reproduces excess volatility, stochastic volatility, trends, and booms and busts, addressing roughly seven or eight persistent market puzzles. Haghani acknowledges the model has many parameters, which limits how definitive its conclusions can be.

The biggest stock market puzzle Haghani identifies is why equity prices are roughly twice as volatile as long-term earnings, referencing Shiller and Campbell work from the late 1980s. He argues that explaining this excess volatility by invoking a moving discount rate merely reframes the puzzle rather than solving it. Static investors such as target-date fund holders and the Norwegian oil fund maintain fixed allocations like 60/40 and rebalance without adjusting for expected returns. Extrapolators form return expectations from recent history, modeled as the past five years with exponential decay, and survey data supports this behavior, with the effect especially strong in cryptocurrency markets.

Haghani considers momentum the biggest challenge to the random walk, noting Eugene Fama called it the mother of all anomalies despite being a father of the efficient markets hypothesis. Momentum exists across asset classes including natural gas, 30-year bonds, and Bitcoin. The Turtles experiment demonstrated that traders who knew nothing succeeded primarily by cutting losses and letting profits run. Momentum persists because the capital pool dedicated to momentum strategies is much smaller than the capital controlled by extrapolators who create the effect. Haghani predicts that if momentum investors accumulated enough capital to have significant price impact, the strategy would stop working. Momentum is also psychologically difficult because it means buying high and selling higher, causing investors to abandon it during drawdowns. Renaissance Technologies' core fund, which produced the best hedge fund returns ever, was driven substantially by momentum strategies.

Haghani draws a sharp distinction between passive stock selection and passive asset allocation. Market-cap-weighted index funds are fine for most investors and are mistakenly criticized largely by people whose businesses have been hurt by outflows from active management. When index funds buy, they buy from active investors who collectively hold the market portfolio, so there is no net price impact, and roughly 95 percent of stocks remain reasonably priced by fundamentals. Passive asset allocation, meaning holding a fixed allocation like 60/40 regardless of changing conditions, is a separate and genuinely problematic practice. Modern portfolio theory and the capital asset pricing model never prescribed passive asset allocation and expected investors to actively position based on expected return, risk, and preferences. Haghani argues passive asset allocation is responsible for market booms and busts and has already produced a boom without the bust yet.

Haghani describes the overall level of the US stock market as looking ridiculous right now even if individual stock prices within it look roughly reasonable. The consensus long-term expected return for US equities from institutions including Vanguard, BlackRock, and Goldman is approximately 6 percent, while 30-year US Treasury yields are around 5 percent, leaving only roughly 1 percent equity risk premium. He argues that valuing equities on last year's or next year's operating earnings is misleading and that cyclically adjusted long-term earnings are the appropriate basis, since US corporate earnings as a share of GDP have historically been around 6 to 8 percent and cannot sustainably exceed that. Non-US equities offer healthier long-term expected returns relative to safe assets. Elm's current allocation to US equities is below 30 percent against a baseline weight of around 40 percent, the portfolio is slightly overweight non-US equities, and within fixed income it holds mostly treasury bills because TIPS and nominal bonds have negative momentum.

US stock buybacks exceeding one trillion dollars per year have materially pushed up stock prices because the market is not perfectly elastic. Elm's model estimates that one trillion dollars per year of buybacks moves the market up by approximately 3 to 4 percent. When a company buys back shares, index funds sell those shares back and redeploy the cash into the broader market, meaning buybacks lift the whole market rather than only the repurchasing company's stock. Haghani also cites the scarcity of fundamental value investors, a light IPO calendar, capital-light businesses, and extrapolator behavior as additional contributors to elevated US stock prices. He warns that if AI capital expenditure declines and returns on that spending prove poor, S&P 500 earnings could turn massively negative once markdowns and depreciation are included.

The biggest lesson Haghani draws from LTCM is about personal risk taking and skin in the game. He argues that running a leveraged pool of standalone capital may be a bad business structure compared to doing relative value trading within larger institutions, and that relative value trades tend to have fatter tails than delta one trades. LTCM did not run tight stop losses, which was a key divergence from more successful funds, and tight stop losses bring positive exposure to momentum into strategies. He believes LTCM probably could have survived if it had not been so publicly known to be in a large drawdown and likely unwinding positions, noting that other relative value hedge funds survived 1998 though with large losses.