Everyone tells you to diversify. Almost no one tells you there's a point at which diversification stops protecting you and starts guaranteeing that you can't beat the thing you're trying to beat.

The standard advice is to hold 20, 30, 50, or more stocks across sectors and geographies. By this logic, more is better — more holdings means less risk. The problem is that this is approximately true up to a point, and then it becomes an excuse to own businesses you've never seriously thought about, correlated assets dressed up as diversification, and index-like exposure that costs more than an index.

This piece is a case for deliberate concentration — not recklessness. After reading it you'll understand exactly how much diversification the evidence supports, why "over-diversification" is a real phenomenon that most retail investors are living in, and how to construct a portfolio that balances genuine protection with the analytical capacity to actually monitor what you own.

TL;DR

• Research consistently shows that most idiosyncratic (company-specific) risk is eliminated by 15–20 carefully chosen holdings. Beyond that, you're not reducing risk — you're eliminating upside.
• The biggest single-stock risk (fraud, sector collapse, permanent impairment) is not solved by adding a 30th holding; it's solved by quality filtering before you buy.
• In a real market downturn, correlation spikes. Your 40-stock "diversified" portfolio may move nearly in lockstep with the index anyway.
• If you can't monitor 40 positions meaningfully, you don't have a diversified portfolio — you have an unmanaged one. That's a different kind of risk.
• The honest alternative: fewer, better-understood positions; genuine diversification across uncorrelated asset classes; or a low-cost index if you can't or won't research individual businesses.

1. What diversification actually does — and the steelman

The logic behind diversification is solid at its core. Every company carries two types of risk:

• Systematic (market) risk: the risk that moves with the broader economy — recessions, interest rate shifts, geopolitical events. You cannot eliminate this by adding more stocks. It affects everything.
• Idiosyncratic (company-specific) risk: the risk unique to one business — fraud, management failure, product obsolescence, competitive disruption. This can be reduced by holding more companies.

The steelman: if you hold only one stock and the company commits accounting fraud, you lose everything. If you hold 20 stocks and one does, you lose 5%. Diversification genuinely protects against concentrated idiosyncratic catastrophe. No serious argument against diversification disputes this.

The question is where the marginal benefit ends.

2. The diminishing returns: the empirical case

The foundational work on portfolio risk reduction comes from studies of equity portfolios in major markets. The consistent finding across markets and time periods:

The reduction going from 1 to 10 holdings is enormous. From 10 to 20, meaningful. From 20 to 50, small. From 50 to 100, negligible. The asymptote is the market itself. By the time you hold 30–40 representative stocks, you've captured perhaps 90–92% of the diversification benefit an index provides — at the cost of index-like performance without the index's low fees.

This is the mathematical case for "15–20 well-chosen stocks eliminate most idiosyncratic risk." It is not a fringe view. Warren Buffett, Charlie Munger, and the academic literature largely agree on the number, even if they disagree on what else to do with it.

3. The correlation problem: diversification fails when you need it most

Here is the inconvenient empirical fact about diversification in bear markets: correlation rises.

Under normal market conditions, individual stocks in different sectors move somewhat independently. A pharmaceutical stock and an energy company have low correlation — their businesses don't move together for most of the year. This is the environment in which holding 40 stocks across sectors looks like genuine diversification.

In a real downturn, this disappears. During the 2008–09 financial crisis, the 2020 COVID crash, and the 2022 rate-shock selloff, correlations across equities spiked sharply. Global markets fell together. Sectors that "had nothing to do with each other" in calm markets declined in near lockstep during the actual stress event. Investors who held 40 domestic stocks discovered their "diversified" portfolio fell almost as much as the index.

This does not mean diversification is useless. It means equity diversification within one asset class does not protect you against macro risk. More stocks does not solve the problem that all stocks are correlated to growth expectations and risk appetite.

Real diversification against macro risk requires diversification across asset classes — equities, fixed income, real assets, cash, and potentially alternative strategies with genuinely different payoff structures. Holding 50 stocks instead of 20 does almost nothing here.

4. The analytical capacity problem

Here is the argument most portfolio theory ignores entirely: you cannot meaningfully monitor 40 businesses.

A serious investor who owns a stock should know:

• The business model and unit economics
• The primary competitive advantage (if any) and how durable it is
• The key risks to the thesis (regulatory, competitive, management)
• The valuation framework and what would make them sell

That is not a 20-minute exercise. Reading an annual report with genuine comprehension, understanding the competitive dynamics, and forming a defensible view on valuation takes serious time. Most investors who hold 40–50 stocks have not done this for most of what they own. They bought on a tip, a screen, a headline, or a vague sector thesis.

The result: they don't know when a thesis has broken. They hold a business through a structural deterioration they didn't see coming because they never understood what they were watching for. They mistake inaction for conviction.

If you own more companies than you can seriously follow, you haven't diversified — you've passively accumulated positions you will exit at the wrong time, because you have no framework for when to exit.

5. What actually protects you

Three things provide genuine protection that more holdings cannot replicate:

Quality filtering before purchase. The most destructive outcomes in equity portfolios — fraud, permanent capital impairment, zero — are concentrated in low-quality businesses with poor governance, high leverage, or implausible economics. Avoiding these isn't diversification; it's judgment applied before the investment is made. No amount of diversification fully mitigates holding a business that commits fraud, but rigorous quality filters prevent most such positions from entering the portfolio in the first place.

Asset class diversification. Holding equities, fixed income, and real assets creates genuine diversification because their risk factors are structurally different. This is where the risk reduction work happens that equity diversification claims to do but can't deliver in a real crisis. A portfolio of 15 quality equities plus a meaningful fixed-income allocation is more genuinely diversified against macro risk than a portfolio of 50 equities with no other asset classes.

A written, pre-committed sell thesis. The most common portfolio risk isn't concentration — it's holding through deterioration because the investor never defined what would make them wrong. A position sizing limit and a written "what would make me sell" for each holding does more for risk management than adding 10 more positions.

6. The honest case for indexing

If you are not going to do the analytical work, index. A low-cost broad market index fund — with expense ratios of 0.05–0.20% — gives you true diversification, systematic rebalancing, and no decisions to execute incorrectly. This is not settling for mediocrity; it is choosing an approach that matches your actual capacity and beats most active managers over long periods.

The portfolio with 40 undifferentiated stocks is neither the active investor's portfolio nor the index. It is the worst of both: the costs and decisions of active management with the returns (and attention) of a passive strategy. It is also the most common portfolio of retail investors who read about diversification without reading the fine print.

Common mistakes

Confusing the number of holdings with the quality of diversification. Thirty stocks in one country, one sector, and one business model are more concentrated than 10 stocks across three genuinely uncorrelated sectors in different markets. Count less; diversification type more.

Treating a diversified equity portfolio as protection against systemic risk. In a 2008-style crisis, the number of stocks you hold matters far less than your allocation to asset classes with different risk factors. Ten extra equity positions provide almost no protection when equity risk premiums are repricing across the board.

Buying "just to diversify." A new holding purchased because you "need more names" rather than because you have a genuine thesis is not risk reduction. It is noise reduction — diluting your best ideas with lower-conviction positions. The mathematical risk reduction of going from 25 to 26 holdings is approximately zero.

Never pruning. A portfolio that grows via addition but never via subtraction becomes an unmanaged index of past decisions. Holdings where the original thesis has been wrong, slow, or superseded by new information should be evaluated against the next-best use of that capital — which often turns out to be adding to an existing high-conviction position.

Anchoring to "I own X stocks" as a risk signal. Concentration risk is not the number; it is the maximum single-position loss that would permanently impair your financial goals. Five percent position sizing with quality filtering and a sell thesis may be more controlled risk than fifteen percent average sizing with 40 positions you haven't reviewed in two years.

Summary and next step

The evidence on diversification supports holding 15–20 well-researched positions to eliminate most idiosyncratic company risk. Beyond that, the marginal benefit to risk reduction is small, the marginal cost to analytical capacity is large, and the result often looks more like an expensive, unmanaged index than a considered portfolio.

More holdings solve a narrow problem. The problems that actually destroy portfolios — holding bad businesses, selling at the bottom, miscalibrating risk capacity — are not solved by adding a 31st stock. They are solved by the work you do before and after each position.

Related reading: [How to Size a Position](/content/published/investing-research/how-to-size-a-stock-position.md) — the framework for deciding how much of your portfolio any single holding deserves, given conviction and downside asymmetry.