Buy low-priced stocks. Buy stocks going up. Buy profitable, well-run businesses. Buy stocks nobody else is watching. Four decades of academic research boiled the market down to a handful of measurable characteristics that have paid investors more than the market average, on average, over long periods - and none of them require predicting the future.
Most investors pick stocks one of two ways. They either buy the whole market through an index fund and accept whatever the market-cap-weighted average delivers, or they try to out-guess the market stock by stock, betting on a story, a product, a management team. Factor investing is a third path, and it is the one with the deepest academic paper trail behind it.
The idea starts with a simple observation that has shown up again and again in market data going back to the 1920s: stocks that share certain measurable characteristics have, as a group, outperformed the broad market over long stretches of time. Not every stock with that characteristic. Not every year. But the group, on average, over enough time. Those characteristics are called factors, and five of them have the strongest, most-replicated evidence behind them.
Here is what the research actually says, why each factor is believed to work, where each one has failed, and how to build exposure to them without a PhD or a Bloomberg terminal.
What a "Factor" Actually Is
A factor is a measurable, quantifiable characteristic of a stock that has been shown, across long historical periods and multiple markets, to be systematically associated with returns that differ from the market average. The key word is systematically - a factor is not a single winning trade or a clever narrative, it is a characteristic you could have screened for mechanically, every single day, for decades, and it would have shown up in the data as a persistent tilt in returns.
This distinction matters because it separates factor investing from stock-picking. A stock picker looks at Apple and forms a view about the iPhone cycle. A factor investor does not care what Apple makes - they care whether Apple currently screens as cheap, or as high-momentum, or as high-quality, relative to the rest of the market, and they build a portfolio of many stocks that share that characteristic, expecting the group average to work out even though any individual name might not.
The modern academic foundation traces to Eugene Fama and Kenneth French's 1992 paper, which found that a stock's size (market capitalization) and its book-to-market ratio (a value measure) explained a meaningfully larger share of the variation in stock returns than the market-beta model that had dominated finance theory since the 1960s. That paper effectively founded the field. Every factor covered below has since been documented, debated, and re-tested across different countries, different time periods, and different asset classes - the reason these five, and not dozens of other proposed "factors," have survived that scrutiny.
The 5 Factors With the Strongest Evidence
1. Value — Cheap Stocks Outperform Expensive Ones
The thesis: stocks priced cheaply relative to their fundamentals - earnings, book value, cash flow, sales - have historically outperformed stocks priced expensively relative to those same fundamentals, over long holding periods.
How it's measured: the classic value metrics are price-to-book (P/B), price-to-earnings (P/E), price-to-cash-flow, and price-to-sales. A value screen ranks the market by one or a blend of these ratios and selects the cheapest quintile or decile.
Why it's believed to work: two competing explanations dominate the literature. The risk-based view says cheap stocks are cheap because the market has correctly priced in real distress risk - value stocks tend to be more economically fragile, more leveraged, more exposed to recessions, and the extra return is compensation for bearing that risk, not a free lunch. The behavioral view says the market systematically overreacts to bad news, pushing already-struggling companies' prices below what their actual fundamentals justify, and the value premium is the correction of that overreaction as prices mean-revert.
Where it has struggled: value had arguably its worst decade on record from roughly 2010 through 2020, as a handful of mega-cap growth and technology stocks compounded so far ahead of the rest of the market that traditional value screens badly lagged. That stretch was long and painful enough that "value investing is dead" became a recurring headline for the better part of a decade - shortly before value staged one of its sharpest multi-year reversals on record in 2021-2022 as rates rose and growth multiples compressed.
2. Size — Small Caps Outperform Large Caps
The thesis: smaller companies, measured by market capitalization, have historically delivered higher average returns than large companies over long periods.
How it's measured: market cap itself, typically split into deciles from largest to smallest, with the size premium measured as the return spread between the smallest and largest deciles.
Why it's believed to work: small companies carry real structural disadvantages that show up as risk - less access to capital, less analyst coverage (meaning more potential for the market to simply not have found and priced in real value yet), thinner trading liquidity, and more exposure to a single product line or customer concentration going wrong. The premium is largely explained as compensation for these risks rather than a market inefficiency, though some of the outperformance is attributed to lower analyst coverage leaving more names under-researched and therefore more likely to be mispriced.
Where it has struggled: the size premium has been the weakest and least consistent of the five factors over the last three decades specifically, with long stretches - including most of the 2010s - where large caps, particularly mega-cap technology names, meaningfully outperformed small caps. Some researchers now treat "pure" size as a weaker standalone factor than it was originally documented to be, and argue it works best combined with value or quality rather than held alone.
3. Momentum — Recent Winners Keep Winning
The thesis: stocks that have performed well over the trailing 6 to 12 months tend to continue outperforming over the following few months, while recent losers tend to continue underperforming.
How it's measured: typically trailing total return over a 6-12 month lookback window, sometimes excluding the most recent month to avoid short-term reversal effects. A momentum screen buys the top-performing decile and, in long-short institutional strategies, shorts the bottom decile.
Why it's believed to work: momentum is widely considered the factor with the strongest evidence for a behavioral rather than risk-based explanation. The leading theory is investor underreaction - when genuinely good news comes out about a company, the market does not immediately price it in fully; instead the price drifts upward gradually as the information spreads and gets absorbed, which is exactly the pattern a momentum strategy captures. This underreaction has been documented across nearly every market and asset class studied, which is part of why momentum has one of the largest and most persistent premiums in the academic literature.
Where it has struggled: momentum is also the factor most prone to sudden, violent reversals known as momentum crashes - the most cited example is the sharp snapback in 2009 coming out of the financial crisis, when the stocks that had fallen the most during the crash (and were therefore the worst "momentum" names) violently outperformed as the market rallied, inflicting some of the worst short-term losses on momentum strategies in the historical record. Momentum requires more frequent rebalancing than value or quality, which also means higher trading costs eating into the premium in practice.
4. Quality — Profitable, Well-Run Businesses Outperform
The thesis: companies with high profitability, stable earnings, low debt, and efficient capital use have historically outperformed lower-quality companies, particularly on a risk-adjusted basis.
How it's measured: there is no single universal quality metric - common components include return on equity (ROE), return on invested capital (ROIC), gross profitability (gross profit divided by total assets, a metric popularized by researcher Robert Novy-Marx), low debt-to-equity, and stable or growing earnings over time. Most quality screens blend several of these into a composite score.
Why it's believed to work: the behavioral explanation is that investors chase exciting, story-driven, speculative companies and systematically underprice boring, consistently profitable ones - a well-run industrial or consumer staples company grinding out steady ROE does not generate headlines the way a speculative growth story does, even when its underlying economics are stronger. There is also a risk-based component: high-quality companies with strong balance sheets are genuinely more resilient in downturns, which shows up empirically as smaller drawdowns during market stress relative to lower-quality peers.
Where it has struggled: quality is the most consistent of the five factors and has the smallest historical drawdowns, but that consistency comes at a cost - it also tends to lag badly in the earliest, sharpest stages of a bull market rally off a bottom, when the most beaten-down, lowest-quality, highest-risk stocks typically bounce hardest and fastest, exactly the opposite of what a quality screen selects for.
5. Low Volatility — Lower-Risk Stocks Outperform on a Risk-Adjusted Basis
The thesis: stocks with lower price volatility (measured by beta or standard deviation of returns) have historically delivered returns comparable to - or in some periods better than - higher-volatility stocks, despite carrying meaningfully less risk. This is often called the "low-volatility anomaly" because it directly contradicts the core finance-theory prediction that higher risk should be compensated with higher expected return.
How it's measured: typically trailing beta (sensitivity to overall market moves) or trailing standard deviation of returns over a lookback window, usually 1-3 years. A low-volatility screen selects the lowest-volatility decile or quintile of the market.
Why it's believed to work: the leading explanation is a structural one - many institutional investors are prohibited or discouraged from using leverage, so instead of leveraging up a low-risk portfolio to hit a target return, they buy higher-beta stocks directly to try to boost returns, which structurally bids up the price (and depresses the future return) of high-beta stocks relative to low-beta ones. There is also a behavioral overlay: individual investors are drawn to volatile, lottery-like stocks with a small chance of a huge payoff (the same psychology behind lottery ticket purchases), systematically overpaying for that volatility and underpricing the boring, stable alternative.
Where it has struggled: low-volatility strategies structurally lag in strong, sustained bull markets, since by design they are underweight the highest-beta names that lead a rally - the tradeoff is smaller gains in up markets in exchange for smaller losses in down markets, which only shows up as an advantage over a full market cycle rather than any single year.
The 5 Factors at a Glance
| Factor | What It Measures | Why It's Believed to Work | Biggest Weakness |
|---|---|---|---|
| Value | Cheap price relative to fundamentals (low P/B, P/E, P/CF) | Risk compensation + market overreaction to bad news | Multi-year underperformance stretches (e.g. 2010-2020) |
| Size | Smaller market capitalization | Illiquidity risk, less analyst coverage | Weakest, least consistent premium of the five |
| Momentum | Strong trailing 6-12 month returns | Investor underreaction to news, gradual price drift | Sharp, violent reversals ("momentum crashes") |
| Quality | High ROE/ROIC, low debt, stable earnings | Market underprices "boring" consistent profitability | Lags in sharp early-bull-market rallies |
| Low Volatility | Low beta / low return variance | Leverage-averse institutions bid up high-beta names | Structurally lags in strong bull markets |
Why Factor Premiums Exist at All
If a factor premium is real, persistent, and public knowledge - published in academic journals for decades - a fair question is why it has not simply been arbitraged out of existence by investors piling in to capture it. The honest answer from the research is a mix of two things, and reasonable researchers disagree on the split.
The risk-based explanation holds that factor premiums are not a free lunch at all - they are the market's price for bearing real economic risk that shows up disproportionately during bad times. Value and small-cap stocks are, on average, more financially fragile and more exposed to distress in a recession; their extra long-run return is simply compensation for the extra pain investors sometimes have to absorb holding them through downturns. Under this view, a factor premium can be entirely genuine and still never be arbitraged away, because arbitraging it away would require someone willing to hold that extra risk without extra compensation - which nobody rational will do indefinitely.
The behavioral explanation holds that markets are not fully efficient because the humans participating in them are not fully rational, in specific, documented, repeatable ways: overreacting to bad news (creating value opportunities), underreacting to good news (creating momentum opportunities), and chasing exciting story stocks over boring profitable ones (creating quality and low-volatility opportunities). Under this view, factor premiums persist because exploiting them requires real behavioral discomfort - value investing means buying stocks everyone else is avoiding, momentum means buying stocks that already look "expensive," and both require holding through the multi-year stretches where the factor underperforms. That discomfort is a genuine behavioral barrier to entry even when the information is fully public.
In practice, most factor researchers today believe both mechanisms play a real role, in different proportions for different factors - momentum leans more behavioral, size leans more risk-based, and value and quality sit somewhere in between.
Factor Cyclicality: The Part Most Investors Get Wrong
Every one of the five factors above has a genuinely positive long-run premium in the historical data. Every one of them has also gone through extended, multi-year stretches of meaningfully underperforming the broad market - stretches long enough to make an investor seriously question whether the factor still works at all.
This is the single most important practical reality of factor investing, and it is also the reason most individual investors who try factor investing abandon it at close to the worst possible time. Value's lost decade from roughly 2010 to 2020 is the textbook example - a factor with 90+ years of positive evidence behind it lagged so persistently, for so long, that "value is dead" became a mainstream thesis, complete with structural explanations for why the old relationship no longer applied. Then value reversed sharply in 2021-2022, delivering a chunk of its entire decade of missing returns in roughly 18 months, right as most of the investors who had given up on it were fully out.
Momentum's failure mode looks different but is equally costly - rather than a slow multi-year fade, momentum tends to fail suddenly and violently, most dramatically in 2009 when the stocks that had cratered hardest during the financial crisis snapped back the fastest, directly punishing momentum strategies that were, by construction, positioned against exactly those names.
The practical takeaway from decades of factor research is not "pick the best factor" - it is that factor timing (trying to hold a factor only during its good years and rotate out during its bad ones) has proven extremely difficult even for professional investors with full-time research teams. The more durable approach that shows up repeatedly in the practitioner literature is diversifying across multiple factors with genuinely different behavior, rather than betting the whole portfolio on any single one.
Combining Factors: Why Diversification Works Here Too
Because each factor fails at different times and for different underlying reasons, blending factors that have low correlation with each other reduces the risk of any single factor's bad stretch dominating a portfolio's return, without giving up the long-run premium any individual factor provides.
Quality + Momentum is one of the most frequently cited combinations in both academic and practitioner research. A stock that screens well on both is a profitable, well-capitalized business that the market is currently rewarding with rising prices - the combination tends to filter out both the "cheap for a good reason" value traps that a pure value screen can catch, and the speculative, low-quality names that can dominate a pure momentum screen during a risk-on rally.
Value + Quality addresses the classic weakness of a pure value screen: a stock can look statistically cheap on P/B or P/E precisely because the underlying business is deteriorating - a genuine value trap. Layering a quality filter (positive and stable ROE, manageable debt) on top of a value screen is specifically designed to separate "cheap and improving" or "cheap and stable" from "cheap because it's structurally broken."
Value + Momentum targets the well-documented behavioral pattern of markets overreacting to bad news and then gradually correcting - screening for stocks that are statistically cheap but have already started to show positive price momentum can catch the early stage of that correction, rather than buying into a value name that is still actively falling (a "falling knife").
The research on multi-factor combinations consistently shows that a well-diversified multi-factor portfolio has historically delivered smoother, more consistent excess returns over a full market cycle than any single factor held in isolation - the same diversification logic that applies across asset classes applies within factors themselves.
Building a Factor Screen Yourself
Factor investing does not require an institutional research desk. The two most accessible paths for an individual investor:
Factor ETFs offer instant diversified exposure to a single factor through a low-cost fund - value ETFs screen the market on metrics like P/B and P/E, momentum ETFs systematically rebalance toward recent price strength, and quality ETFs screen for high profitability and conservative balance sheets. This is the lower-effort path: no individual stock research required, and rebalancing happens automatically inside the fund.
Direct screening means building your own basket using the specific metrics that define each factor:
| Factor | Screen For |
|---|---|
| Value | Low P/E, low P/B, low price-to-FCF relative to sector |
| Size | Market cap below a chosen threshold (small/mid-cap) |
| Momentum | Strong 6-12 month trailing return |
| Quality | High ROE, high ROIC, low debt-to-equity, stable margins |
| Low Volatility | Low beta, low trailing price volatility |
The Stock Alarm Pro screener lets you filter the market directly on these fundamentals and technicals together - for example, combining a valuation filter (low P/E, low price-to-book) with a profitability filter (high ROE, high margins) and a low-debt filter builds a quality-value screen from scratch, in one pass, without needing a separate data provider for each factor. Screening on momentum can be layered in the same pass by filtering for stocks trading above their 50-day and 200-day moving averages with strong trailing returns.
Common Factor Investing Mistakes
Chasing whichever factor performed best last year. Because every factor is cyclical, the factor with the best trailing 1-3 year return is frequently the one closest to a mean-reverting slowdown, not the one about to keep winning - performance-chasing at the factor level suffers from the same behavioral trap as performance-chasing individual stocks or funds.
Abandoning a factor after a rough stretch. Value's decade-long underperformance through 2020 tested the patience of even long-term factor investors, and a meaningful number capitulated and rotated out right before the 2021-2022 reversal. A factor tilt only captures its long-run premium if it survives the bad years, not just the good ones.
Treating factors as guaranteed rather than probabilistic. A positive long-run academic premium describes an average across decades and hundreds of names - it says nothing about what any single stock, or any single year, will do. Factor investing reduces uncertainty relative to picking individual stocks on a narrative; it does not eliminate it.
Over-concentrating in a single factor. A portfolio built entirely around one factor inherits that factor's specific failure mode in full - all of momentum's crash risk, or all of value's decade-long drought risk. The research consistently favors diversifying across factors with genuinely different behavior over betting everything on the single factor with the best-looking backtest.
Ignoring transaction costs on higher-turnover factors. Momentum in particular requires more frequent rebalancing to stay current with which names are actually showing recent strength, and trading costs (and, in a taxable account, short-term capital gains) can meaningfully erode the premium a backtest shows on paper.
Track Factor Signals as They Develop
Factor investing works over years, not days, but the individual signals underneath each factor - a stock crossing above its 200-day moving average, a valuation ratio dropping into a new range, an ROE jump on a fresh earnings report - happen in real time.
Screen the market for factor characteristics with the free Stock Alarm Pro screener - filter by valuation, profitability, momentum, and volatility metrics together to build your own factor baskets.
Explore live markets with no signup required at pro.stockalarm.io/explore to see how factor characteristics are showing up across the market right now.
Once you've built a factor basket worth watching, set a price or percentage-move alert so you catch the moment a name in it breaks out - rather than checking your screen manually every day.
This article is for educational purposes only and does not constitute investment advice. Factor premiums are based on long-run historical averages and are not guaranteed to persist or to apply to any individual stock, fund, or time period. All factors are subject to extended periods of underperformance. Past performance of any factor, strategy, or index mentioned is not indicative of future results. Always conduct your own research or consult a licensed financial advisor before making investment decisions.
