Unit Bias and Position Sizing

A unit bias in investing occurs when an investor feels a compulsion to own a round number of shares (100, 500, 1,000) rather than an odd amount, or to allocate capital in round percentages (5%, 10%, 20%), even when the mathematics of diversification and asset allocation demand fractional or irregular amounts. This cognitive quirk distorts portfolio weight and increases risk.

What unit bias looks like in practice

A portfolio manager decides to hold 12 stock positions to achieve adequate diversification, but finds that optimal sizing for one position (given the fund’s total capital, target volatility, and position correlations) is 7.3% rather than the nearest round number, 5% or 10%. The psychological pull toward “10%” is strong—a number that feels complete and is easy to communicate to stakeholders.

Unit bias also appears when investors insist on owning “100 shares” of a stock because 100 is a standard round lot, when the optimal position size for their portfolio is 87 or 93 shares. Before fractional shares became common, this was a genuine market friction; today, it is pure psychology.

The bias extends to asset allocation. An investor with a target of 23% in bonds might round to 20% or 25%, sacrificing their intended risk exposure for the satisfaction of a “clean” number. Portfolio weights of 25%, 25%, 25%, 25% across four asset classes feel orderly and fair, even if the actual optimal split is 24%, 26%, 23%, 27% based on return and volatility forecasts.

In factor investing, unit bias manifests as holding exactly five factors (value, momentum, quality, size, volatility) in equal weight (20% each), when a more rigorous analysis suggests 18%, 22%, 20%, 24%, 16%. The equal-weight solution is psychologically appealing and easy to manage, but mathematically arbitrary.

The cognitive appeal of round numbers

Psychologists and behavioral economists have documented that round numbers feel cognitively “easier” to process and remember. A portfolio with 10 positions, $1,000,000 in value, and 20% in bonds is faster to communicate and mental-model than one with 12 positions, $987,654 in value, and 23.4% in bonds—even if the latter is more accurate or optimal.

Round numbers also create an illusion of control and completeness. An investor who owns exactly 100 shares feels she has “finished” the transaction; buying 87 shares leaves a sense of incompleteness—“should I buy 13 more?” Round positions are easier to liquidate (100 shares, not 87), and they avoid the psychological discomfort of fractional holdings.

Numbers like 50%, 25%, and 10% carry additional force because they are mathematically fundamental—halves, quarters, tenths. A portfolio split 50/50 between stocks and bonds, or 25% allocated to each of four themes, feels “fair” and “symmetrical” even when such allocations conflict with stated risk targets or return forecasts.

How unit bias distorts allocation

The core harm is drift from optimal position sizing. If the mathematics of your portfolio strategy demand that Position A be 7.3% of capital, and unit bias causes you to hold 10% instead, you have overweighted Position A by 3.7 percentage points. That excess capital must come from somewhere—typically an underweighted Position B.

This creates concentrated risk. Positions held at larger-than-optimal size because they land on “nice” numbers become outsized bets. If those positions correlate more with your existing holdings than you anticipated, diversification suffers. The excess capital wasted on rounding is capital no longer available for true diversification.

Unit bias also complicates rebalancing. A portfolio with “round” weights may drift from those weights as markets move, requiring more frequent and larger rebalancing trades. A portfolio with mathematically optimal (non-round) weights typically requires the same rebalancing frequency, but the destination is more precise, reducing overshoot and drift.

Unit bias in fund structure

Mutual funds and ETFs sometimes fall prey to unit bias in their structures. A fund might hold exactly 50 stocks (a round number suggesting diversification) when holdings analysis shows 47 stocks would optimize the Sharpe ratio or minimize tracking error. A target-date fund might allocate 20% to each of five broad asset classes for simplicity, rather than rebalancing quarterly based on Sharpe-optimal weights.

Index funds have a built-in defense: they simply hold all (or a representative sample of all) stocks in the index, following a mathematical rule rather than a psychologically appealing round number. Index funds and passive strategies generally avoid unit bias because the holdings are determined ex-ante by the index methodology, not by committee intuition.

Actively managed funds, which exercise human judgment on position count and weight, are more vulnerable. A manager who believes she should hold 12 to 15 positions might round to exactly 12 because 12 is easier to manage and monitor, even if the current opportunity set supports 13.7.

Overconfidence and unit bias

Unit bias also interacts with overconfidence bias. An investor overconfident in their stock-picking ability might hold exactly 10 “best ideas,” each weighted at 10%, rather than sizing positions based on conviction and risk thresholds. The round weight masks overconcentration; the investor feels diversified with 10 positions, when a more honest assessment suggests 8 positions would be appropriate given actual edge and conviction.

Defense strategies

Modern portfolio technology largely eliminates unit bias. Fractional shares, available through most brokers and index funds, remove the barrier to odd position counts. Automated rebalancing tools ignore roundness and rebalance to precise target weights.

The strongest defense is a written investment policy statement that specifies position-sizing rules mathematically rather than heuristically. A rule like “each position is sized at [formula based on volatility, correlation, and conviction] rounded to 0.1%” produces genuinely random, non-round weights—and thereby forces discipline.

Robo-advisors and quantitative strategies inherit this advantage: allocations are determined by algorithms, not human intuition. A robo-advisor rebalancing a portfolio will produce target weights like 23.7%, 18.4%, 16.9%, 41.0%—odd numbers optimized for some objective, not round numbers optimized for psychological comfort.

For individual investors, the conscious practice is to ask: “Am I holding this position size because the math demands it, or because it’s a round number?” If the answer is the latter, revisit the position size. Over a career of investing, eliminating the drag from unit bias—even if only by 0.5% of returns annually—compounds into meaningful outperformance.

See also