Benchmark-based investing is commonly implemented by requiring a portfolio to track a reference

index with fixed, often unit, intensity. This paper develops a closed-form theory of endogenous

benchmark orientation in which benchmark exposure is a decision variable. The model minimizes

tracking-error variance subject to a target active-return bias and a budget constraint, allowing the optimal

benchmark intensity to be positive, zero, or negative. The optimal intensity depends explicitly

on asset expected returns, asset covariances, benchmark mean and variance, and asset–benchmark covariances,

yielding analytical conditions under which a portfolio should track the benchmark, ignore it,

or move against it. The theory provides managerial diagnostics for benchmark-mandate design. The

loss from imposing any fixed benchmark intensity is quadratic in its distance from the optimal intensity,

which gives closed-form measures of the value of benchmark flexibility and the cost of full-intensity

tracking. The model also implies a binary mandate rule: if the optimal intensity is below one-half, a

benchmark-neutral mean–variance mandate dominates full-intensity tracking. Extensions characterize

constrained intensity mandates and transaction-cost-aware rebalancing. A numerical illustration and an

out-of-sample case study show how the approach can support data-driven portfolio implementation and

governance decisions involving benchmark exposure.

Keywords: portfolio optimization; benchmark orientation; index tracking; tracking-error variance; activereturn

bias; mandate design;