Skip to main content
    Back to the glossary

    optimizer

    Target Risk

    An optimization mode that finds the highest-return portfolio whose volatility is at most a chosen risk budget.

    Target risk is an optimization mode that finds the highest expected return available without letting portfolio volatility exceed a ceiling you choose. You state how much fluctuation you can live with, and the optimizer looks for the best expected return inside that budget. It suits investors who know their risk tolerance better than their return needs.

    The optimizer maximizes the weighted sum of expected returns subject to portfolio variance staying at or below the target volatility squared, along with the usual rules (weights add to 100%, no shorts, caps). Variance appears because it is what the covariance matrix produces directly, so squaring the ceiling turns a 12% volatility limit into a 0.0144 variance limit. The solution is the point on the efficient frontier at that risk level.

    A higher ceiling lets the optimizer reach for more expected return, with smaller gains for each extra point as you move up the frontier. A ceiling below the minimum-variance portfolio's volatility cannot be met with the chosen holdings, and a ceiling above the most volatile holding's volatility never binds. The target applies to volatility estimated from history; realized volatility over the next year can be well above it, especially in a crisis when correlations jump.

    Example with illustrative numbers: you set a 12% volatility ceiling, and the optimizer finds a mix expecting 8.2% with exactly 12% volatility. Loosening the ceiling to 15% might lift expected return to 9.4%, while tightening it to 9% might drop it to 6.5%. The steps between those points show what each extra point of risk buys.

    On GNG, Target Risk is an engine in the Portfolio Optimizer's goal-targeting group. You enter the ceiling in the Target volatility (hard ceiling) field, and the engine maximizes expected return subject to it, with your other limits built into the same problem. With no ceiling entered, it uses the median volatility of the names in your universe, and a ceiling above the most volatile name is pulled back to that level. If the solver fails, GNG falls back to inverse-volatility weights and flags the run. The final single-name cap and minimum-position steps, and in rebalance or refine runs the turnover check, can also move the weights after the solve, so the final volatility can end up above your ceiling; check it in the results.

    Target Return is the mirror-image mode, and Volatility explains the quantity being capped.

    More optimizer terms

    Get full access to GNG Research

    Create a free account to access portfolio tracking, advanced tools, and more.