Concentration vs Diversification for Value Investors
Warren Buffett has called diversification "protection against ignorance" and argued that it "makes very little sense for those who know what they're doing." Charlie Munger was more direct: "The idea of excessive diversification is madness." On the other side, Benjamin Graham recommended holding 10 to 30 stocks, Walter Schloss routinely held over 100, and Peter Lynch's Magellan Fund held over 1,000 positions at its peak. These are not fringe opinions from mediocre investors. These are the most successful value investors in history, and they held diametrically opposed views on concentration versus diversification. The disagreement is not a matter of one side being right and the other wrong. It is a matter of different investment approaches requiring different portfolio structures, and the right answer depends entirely on the investor's strategy, skill level, and temperament.
The Mathematical Framework
The mathematics of diversification are well established and worth reviewing briefly, because they set the boundaries of the debate.
A single stock carries two types of risk: systematic risk (market risk, which cannot be diversified away) and idiosyncratic risk (company-specific risk, which can be diversified away). As an investor adds uncorrelated positions to a portfolio, idiosyncratic risk declines while systematic risk remains constant. The rate of risk reduction follows a diminishing curve: the first few positions eliminate the most risk, and additional positions beyond 20-30 add diminishing marginal benefit.
Research by Statman (1987) and others has estimated that approximately 90% of achievable diversification benefits are captured with 20-30 stocks, and approximately 95% with 50 stocks. Beyond 50 positions, additional diversification produces negligible risk reduction. This is the empirical basis for the common recommendation of 20-30 stocks as a "well-diversified" equity portfolio.
The expected return side of the equation is more nuanced. In a concentrated portfolio, each position contributes significantly to total returns, both positively and negatively. A 10-stock portfolio where the best performer doubles and the worst declines 50% has a very different return than a 100-stock portfolio with the same best and worst performers but smaller position weights. Concentration amplifies the impact of stock selection skill (alpha) but also amplifies the impact of bad luck or poor analysis.
The Kelly Criterion, developed by John Kelly at Bell Labs in 1956 and adapted for investing by Edward Thorp, provides a mathematical framework for optimal position sizing. The Kelly formula says that the optimal bet size is proportional to the edge divided by the odds. An investor with a large informational edge should concentrate heavily; an investor with a small or uncertain edge should diversify broadly. This is the mathematical formalization of Buffett's intuition: if an investor truly knows more about a company than the market does, they should bet big. If their edge is uncertain, they should spread their bets.
The Case for Concentration
The concentrated value investors, Buffett and Munger chief among them, make several arguments.
Information advantage scales with focus. An investor who follows 15 companies can know each one deeply: the competitive dynamics, management incentive structures, customer relationships, capital allocation history, and balance sheet nuances. An investor who follows 150 companies necessarily knows each one superficially. If the purpose of active investing is to exploit information advantages, concentration maximizes the return on research effort.
Buffett's investment in American Express during the 1964 Salad Oil Scandal illustrates this point. Buffett spent weeks in Omaha counting American Express charge slips at restaurants and shops, directly observing that consumer usage of the card had not declined despite the financial panic. This ground-level research gave him conviction to invest 40% of the Buffett Partnership's assets in a single stock. That position produced extraordinary returns. An investor holding 50 or 100 positions could never have conducted that level of due diligence on each holding, and would never have had the conviction to allocate 40% to a single idea.
Concentration forces discipline. When an investor can only hold 10-15 positions, each new purchase must displace an existing holding or represent a genuinely compelling opportunity. This creates a natural quality filter that prevents the portfolio from filling up with mediocre ideas. Munger described this as the "punch card" approach: imagine having a lifetime card with only 20 punch holes, one for each investment. The constraint forces the investor to be extremely selective.
The best ideas drive returns. Empirical evidence from mutual fund research shows that fund managers' highest-conviction positions (their largest overweights relative to the benchmark) consistently outperform their lower-conviction positions. A 2005 study by Cohen, Polk, and Silli found that the "Best Ideas" of active managers outperformed the market by approximately 1.6 to 4.0 percentage points annually, while their lower-conviction positions performed roughly in line with or below the benchmark. Diversification dilutes the impact of the best ideas with the mediocrity of the rest.
Fewer decisions, fewer mistakes. Each investment decision is an opportunity to be wrong. A portfolio with 100 stocks requires 100 buy decisions, periodic rebalancing decisions, and eventually 100 sell decisions. A portfolio with 10 stocks requires a fraction of the decisions. Fewer decisions mean fewer chances to make behavioral errors like panic selling, anchoring to a purchase price, or averaging down in a deteriorating situation.
The Case for Diversification
The diversified value investors, Graham and Schloss in particular, make equally compelling arguments.
Analytical uncertainty is irreducible. No matter how thoroughly an investor researches a company, significant uncertainties remain. Management may be dishonest. Accounting may be misleading. Competitive threats may emerge unexpectedly. Regulatory changes may impair the business model. These are not failures of analysis; they are structural features of investing in complex businesses in an unpredictable world. Diversification protects against the inevitable occasions when careful analysis reaches the wrong conclusion.
Graham's experience with his net-net portfolio demonstrated this principle. Many of his holdings were in troubled businesses where the outcome was genuinely uncertain. Some recovered. Some did not. The portfolio-level result was excellent because the winners more than compensated for the losers, but no amount of research could have predicted which specific stocks would fall into which category. Diversification was not a concession to ignorance. It was a rational response to genuine uncertainty.
Schloss's 45-year record is evidence. Schloss held 100+ positions, did minimal deep research on each one, and compounded at approximately 16% annually for nearly half a century. His approach worked not despite the broad diversification but because of it. Wide diversification in cheap stocks produced a reliable statistical edge that was immune to the failure of any individual position. Schloss did not need to be right about any particular stock. He needed the portfolio-level statistics to work, and they did, consistently, for decades.
Concentration risk is real and severe. Bill Ackman's Pershing Square held a concentrated position in Valeant Pharmaceuticals that ultimately lost approximately $4 billion. Ackman had done extensive research, had high conviction, and was one of the most successful investors of his generation. The loss happened anyway. Concentrated portfolios are vulnerable to company-specific events (fraud, regulatory action, competitive disruption) that no amount of research can fully anticipate. A diversified portfolio absorbs such shocks; a concentrated portfolio is devastated by them.
Position sizing errors compound. Even if an investor's stock selection is excellent, errors in position sizing can destroy returns. An investor who allocates 30% to their "best idea" and 5% to a stock that triples has made a sizing error that no amount of analytical skill can correct. In a diversified portfolio, sizing errors are diluted. In a concentrated portfolio, they are magnified.
What the Track Records Actually Show
Examining the track records of successful value investors reveals a more nuanced picture than either camp suggests.
Concentrated investors (fewer than 20 positions):
- Buffett (Berkshire Hathaway, 1965-2024): ~20% CAGR
- Munger (Wesco Financial, 1977-2004): ~17% CAGR
- Greenblatt (Gotham Capital, 1985-1994): ~50% CAGR (very small AUM)
- Klarman (Baupost, 1983-2024): ~15% CAGR
Diversified investors (30-100+ positions):
- Schloss (WJS Partners, 1955-2002): ~16% CAGR
- Graham (Graham-Newman, 1936-1956): ~15% CAGR
- Tweedy Browne (various funds, 1968-2024): ~15% CAGR
- Lynch (Magellan Fund, 1977-1990): ~29% CAGR (1,000+ positions at peak)
Both columns contain exceptional results. The concentration question does not separate the winners from the losers. What separates them is stock selection skill, the ability to identify undervalued securities consistently over long periods. Concentration amplifies the skill signal, producing higher highs (Greenblatt's 50% CAGR) and lower lows (Ackman's Valeant disaster). Diversification dampens the signal, producing more consistent but less extreme results.
Notably, Lynch's record is the most intriguing anomaly. He held over 1,000 positions, which conventional wisdom would suggest leads to index-hugging mediocrity. Instead, he delivered 29% annually for 13 years, one of the greatest records in mutual fund history. Lynch's approach was to make many small bets, sell losers quickly, and add to winners aggressively. His diversification was not a compromise; it was a research advantage. By talking to hundreds of companies, visiting factories, and shopping at stores, he developed a broad mosaic of economic intelligence that informed every individual decision.
A Framework for Deciding
Rather than adopting a dogmatic position, value investors should match their portfolio structure to their circumstances.
Strategy determines structure. A deep value approach (buying statistically cheap stocks with modest research per position) naturally requires wider diversification, typically 30-100 positions. The edge is statistical, and diversification ensures the law of large numbers works in the investor's favor. A quality value approach (buying excellent businesses after deep research) naturally supports greater concentration, typically 10-25 positions. The edge is analytical, and concentration maximizes the payoff from superior research.
Skill level determines appropriate concentration. An investor who has been analyzing businesses for 20 years, has a track record of correct assessments, and thoroughly understands a specific industry can rationally hold a concentrated portfolio. An investor who is newer to value investing, has not yet tested their analytical framework across multiple market cycles, or is investing outside their area of expertise should diversify more broadly. Overestimating one's own skill is the most common mistake in the concentration-versus-diversification decision. The research on overconfidence bias strongly suggests that most investors overrate their analytical edge, which argues for more diversification than feels "necessary."
Capital size matters. An investor with $100,000 can build a concentrated portfolio of 10-15 stocks without liquidity issues. An investor with $10 million may face liquidity constraints in small caps, requiring either broader diversification or a shift toward larger companies. An investor managing $1 billion has a very different optimization problem than one managing $500,000. The capacity constraint is practical, not theoretical, and it should influence portfolio construction.
Correlation awareness matters more than position count. Twenty stocks in the same sector, with similar business models and overlapping customer bases, provide far less diversification than ten stocks across unrelated industries. An investor holding 10 positions, one each in banking, technology, consumer staples, healthcare, energy, real estate, telecommunications, industrials, materials, and utilities, has more genuine diversification than an investor holding 30 positions all in regional banking. Position count is a crude measure. Correlation structure is what matters.
Position Sizing Within the Framework
Regardless of overall portfolio concentration, position sizing within the portfolio deserves careful thought.
Equal weighting is the simplest approach and has much to recommend it. An equally weighted 20-stock portfolio allocates 5% to each position. This prevents any single mistake from being catastrophic and forces the investor to maintain discipline when adding or removing positions. Schloss used approximately equal weighting across his 100+ positions, and the simplicity of the approach contributed to his long-term consistency.
Conviction weighting allocates more capital to the investor's highest-conviction ideas. Buffett might allocate 20-40% to his single best idea and 5-10% to lower-conviction positions. This approach maximizes returns if the investor's conviction is well-calibrated, meaning they are actually better at predicting outcomes for their high-conviction positions. The evidence from mutual fund research suggests that managers' highest-conviction positions do outperform, supporting conviction weighting, but the variance of outcomes is also higher.
A practical middle ground is to set a maximum position size (10% for moderate concentration, 15-20% for high concentration) and a minimum position size (2-3%), with new positions starting small and growing as conviction builds. This approach limits initial risk while allowing successful positions to become larger holdings. It also provides a natural framework for adding to positions as prices decline and the margin of safety widens, a core value investing practice.
The Honest Assessment
The concentration-versus-diversification debate produces strong opinions because it touches on investors' self-image. Concentrated investors often see themselves as bold, conviction-driven, and intellectually superior. Diversified investors often see themselves as humble, disciplined, and risk-aware. Both self-images can be accurate or self-serving, and neither guarantees good results.
The honest assessment starts with recognizing that both approaches have produced extraordinary long-term records, and that the choice between them is secondary to the quality of the underlying stock selection. An investor who consistently identifies undervalued stocks will do well with 10 positions or 100. An investor who consistently overpays for stocks will do poorly regardless of portfolio construction.
The more practical question is whether the investor's portfolio structure matches their strategy, skill, and temperament. A deep value investor who holds only 5 positions is taking concentration risk that their strategy does not require or reward. A quality value investor who holds 100 positions is diluting their best ideas with their mediocre ones. Matching the structure to the approach, and honestly assessing one's own skill level, produces better outcomes than dogmatically committing to either extreme.
Put these principles into practice. Track fundamentals, build portfolios, and analyze stocks with AI-powered insights.
Start Free on GridOasis →