Portfolio construction is the process of choosing the holdings and weights that turn an investment objective into an implementable portfolio. It brings together the eligible assets, return forecasts, risk estimates, costs and constraints. A promising investment idea can still produce an unsuitable portfolio if it receives too much capital, duplicates existing exposures or requires costly trading.
Construction answers “How much should we hold, together?” Security analysis may rank an asset highly; that ranking alone does not specify its position size. Nor does a list of last year's winners supply a forward-looking risk model or a feasible investment policy.
The decisions behind a weight
Start with the job the portfolio must do. A retirement account, a fund seeking to beat an equity benchmark and a portfolio funding a near-term payment face different objectives. The time horizon, cash needs and capacity to absorb losses affect which holdings are appropriate. A written investment policy statement records these choices; the CFA Institute's planning summary explains how objectives and constraints guide construction.
For a quantitative strategy, the inputs then need to describe the same portfolio and horizon:
| Input | What must be decided |
|---|---|
| Investable universe | Which securities, funds or instruments can actually be held? |
| Return forecasts | What outcomes are expected, over what horizon and in which currency? |
| Risk model | How can positions lose together, including common factors and stressed conditions? |
| Costs and current holdings | What trades are needed, and what spreads, fees or market impact might they incur? |
| Constraints | Which position, leverage, liquidity, issuer or other exposure limits must hold? |
| Construction rule | How are these inputs converted into target weights? |
Asset allocation chooses the broad asset-class mix. Construction also covers the instruments and sizes used to express that mix, and applies within a single asset class. Buying several funds does not guarantee distinct exposures if their underlying holdings overlap.
A rule can assign equal weights, track a benchmark or size positions using a specified risk measure. An optimiser can instead solve an explicit objective subject to constraints. Both are construction methods. Neither escapes the need to choose assumptions and check the resulting exposures. As the CFA Institute's active-equity summary emphasises, costs, risk and the investment mandate affect the portfolio alongside return predictions.
Assess a target before calling it a solution
Consider two hypothetical securities, A and B, issued by different issuers. Their stipulated one-year nominal simple total returns in USD have these moments:
| Security | Expected return | Standard deviation | Illustrative starting weight |
|---|---|---|---|
| A | 8% | 20% | 60% |
| B | 4% | 10% | 40% |
The correlation of their returns is . The account is fully invested and long-only; there are no external cash flows or interim trades. Ignore transaction costs, fees and taxes. These inputs are synthetic assumptions, not market estimates or an allocation recommendation. The 60/40 target is supplied for assessment; no objective has been specified precisely enough to establish that it is optimal.
For starting weights and , the portfolio return is
Its expected return is therefore
Its variance also depends on how the returns move together. With standard deviations and correlation , the covariance is . Thus,
The variance uses squared decimal-return units. Taking its square root gives one-year portfolio volatility:
The diagram rounds this to 13.5647%. Weighting the two standalone standard deviations instead gives , which is not this portfolio's volatility. The covariance term matters. None of these moment calculations requires normal returns; moments alone do not specify a probability of loss or a worst-case loss.
Now impose a hypothetical maximum issuer weight of 60%. A is exactly at the limit and B is below it, so the target passes that constraint and its weights sum to 100%. Holding only A would have a higher assumed expected return, but would violate the cap. The cap is an invented policy for this example, not a regulatory rule. Passing it establishes feasibility under that rule; it does not establish that the portfolio suits an investor or maximises an objective.
Original synthetic construction map. The target illustrates assessment, not an optimisation result. Moments share one year and USD simple total-return conventions; costs and taxes are zero.
The target must survive implementation
Mean–variance optimisation is one formal way to trade expected return against risk. It cannot certify the forecasts it receives. Changing a forecast or covariance estimate can change the preferred weights substantially, so sensitivity checks and stress scenarios are part of assessing the output.
The current holdings also matter. Two accounts can share the same desired target yet require different trades to reach it. Transaction costs can make a small forecast improvement uneconomic; an illiquid instrument may be impossible to buy at the assumed size. A real construction process can therefore retain a slightly different portfolio rather than trade immediately to an idealised target.
After trading, prices and circumstances change. Monitoring checks actual exposures against the mandate; rebalancing moves holdings toward a chosen target when the policy calls for it. Revising that target after a change in objectives is a separate decision. Construction is consequently a repeated decision process, not a one-time optimiser run.
For the mathematical next step, Mean-Variance Optimisation develops the portfolio moment formulas and allocation problems with different funding constraints. Its full lesson requires membership; a public preview is available.
Sources
- Qian, E. E., Hua, R. H. and Sorensen, E. H. Quantitative Equity Portfolio Management: Modern Techniques and Applications. First edition, Chapman & Hall/CRC, 2007. Section 1.3, pp. 5–7, “Quantitative Investment Process”; Chapter 2, pp. 23–24, “Portfolio Theory.”
- CFA Institute. Basics of Portfolio Planning and Construction, 2026 curriculum. Public introduction and summary.
- CFA Institute. Active Equity Investing: Portfolio Construction, 2026 curriculum. Public introduction and summary.