Wondering which are some standard constraints in portfolio optimization in practice?

For example, assuming we want to maximize expected returns subject to a risk constraint, typically we may have

-constraints on gross exposure -constraints on net exposure -constraints on individual asset gross exposure etc.

Is there a good document/ book that has standard constraints that are used in practice?


Constraints obviously also depend on the problem formulation (for example: weights sum to 1 will translate to active weights sum to 0 if we optimize relative positions)

Apart from the most important ones the unit investment constraints (weights sum to 1 or money invested sums to NAV), sometimes short-selling constraints, very often you see constraints on the number of securities held or traded.

Depending on the problem, also constraints on portfolio characteristics, such as FX exposure, Non-rated or asset class exposures.

Unfortunately there is no standard literature on this. The reason for this being: Every optimization problem is different and the constraints come from the practical application itself.

Disclosure/Addendum: One thing is considering constraints, the other one is translating them into a format the optimizer understands. More often than not, this is harder than you think.


I have worked with:

Portfolio stability constraints

  • Long-only constraints (to avoid unrealistic short positions)
  • Maximum and minimum allocation to asset classes

Liquidity and illiquidity constraints

  • Minimum posititon in liquid instruments (money market, govies, etc.)
  • Maximum hair-cut due to forced selling

Concentration constraints

  • Asset within asset class concentration constraint (eaxh asset should not be 100% of the exposure in the particular asset class)

Portfolio factor beta exposure

  • Minimum and maximum limits to the relative beta exposure of the portfolio aganst a given factor (e.g. market factor)

There will certaintly be more possible constraints (see e.g. the implementation here).


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.