Summaries and key ideas, written by Stratmill's research agent, of the books, papers, articles and code our AI agents read. Each page links to its original.
4,194 documents
The document outlines a two-step generalized method of moments procedure for testing a multifactor asset-pricing model. Using the Fama–French five-factor model and portfolios sorted by size and book-to-market as test assets, it explains how to form excess…
The document asks how to reconcile a normalized pricing-kernel equation from a quality-investing paper with the consumption-based stochastic discount factor, which is usually expressed using a discount factor and the ratio of marginal utilities. The response…
The document offers intuition for the title of John Cochrane’s argument defending return predictability. It connects the phrase to a Sherlock Holmes story in which the absence of an expected response provides a clue. In the paper’s setting, the relevant…
The document examines whether observed return anomalies falsify the efficient market hypothesis (EMH). Its central point is that an anomaly is identified relative to an asset pricing model, so testing it is a joint test of market efficiency and that model. A…
The document gives a practical way to choose the scale constant in a volatility-managed portfolio whose next-period factor return is multiplied by a constant divided by recent realized variance. The proposed procedure begins with a trial value for the…
The note explains the academic distinction between value and growth stocks through valuation ratios rather than a sharp difference in investing philosophy. Value stocks are relatively cheap, with fundamentals such as book value, earnings, sales, or dividends…
The note explains how the Fama–MacBeth procedure uses repeated cross-sectional regressions to estimate risk premia and assess asset-pricing models. In each period, excess returns are regressed on estimated asset betas; the resulting factor-premium estimates…
The document considers whether the price-to-sales ratio can predict broad stock-market movements or serve as a measure of intrinsic value. A questioner reports substantial explained variance from a regression using P/S and suggests that market gains may…
The document asks how the CAPM’s prediction about expected returns can be tested using realized returns. It outlines a two-stage empirical setup: estimate each stock’s market beta from a time-series regression of excess returns on the market excess return,…
The document describes an attempted replication of the large-cap high- and low-prior-return portfolios used in the Fama-French momentum factor. The researcher maps stocks to Russell 1000 constituents and follows a 12-to-2-month return sort with upper and…
The document asks how to interpret a positive or negative intercept when an asset or factor is regressed on a set of other factors. Its example regresses the value factor on market, size, profitability, and investment factors, then asks what a negative,…
The document distinguishes two ways to build low-volatility equity portfolios: ranking stocks by standalone volatility or beta, and minimizing portfolio variance using a covariance matrix subject to constraints. It asks whether optimized minimum-volatility…
The post concerns replicating the cash-flow-to-assets component of Piotroski's F-Score when Compustat's operating cash flow item is unavailable for earlier years. The questioner notes that the desired historical sample begins before the reported start of…
The document explains how to estimate asset exposures to factors and how to interpret factor-attributed returns in an Arbitrage Pricing Theory framework. For a single factor, exposure is the covariance between factor and asset returns divided by the factor’s…
The document discusses using nonnegative matrix factorization (NMF) to identify latent factors in stock data. It describes decomposing a positive input matrix into loadings and factors, and recounts attempts using stock prices and returns shifted to positive…
The document outlines a workflow for constructing and evaluating factor models of asset returns. Begin by selecting factors with an economic rationale, then gather factor and test-asset data. For portfolio-based factors, construction choices include…
The document compares two neural-network asset pricing models that use firm characteristics and macroeconomic variables to estimate time-varying stochastic discount factors and tangency portfolios. It reports differing out-of-sample performance, portfolio…
The document distinguishes testing whether a pricing model’s restrictions hold from comparing a larger model with a restricted version. It uses the CAPM and an expanded cross-sectional return regression to show how intercepts and coefficients on additional…
The document derives the single-index variance decomposition: an asset’s return variance consists of the market-driven component, scaled by beta squared, plus residual variance. It starts with the return model, removes the risk-free constant because…
The document considers whether machine learning can identify firm characteristics that sort stocks into portfolios with predictive returns, and whether it can estimate asset returns in a CAPM-style setting. Its answer focuses on factor discovery: regress…
The discussion considers whether a company’s annual dividend can be modeled as a function of its stock price. The response suggests grounding assumptions in behavioral finance and dividend theory, particularly the idea that investors favor relatively smooth…
The document considers how to combine conflicting valuation signals, such as price-to-book and price-to-earnings readings, when assessing whether a sector is cheap or expensive. It discusses using correlations between valuation measures at one time and…
The document addresses an ambiguity in the definition of information coefficient (IC), a measure used to evaluate the relationship between an investment manager’s forecasts and subsequent returns. Some sources say simply “correlation,” which can reasonably…
The document asks how to attribute a portfolio’s factor-model variance to individual factors. It starts from a covariance decomposition into factor risk and specific risk, then forms position-weighted factor exposures by multiplying asset weights by the…