Portfolio Management for the CMT Exam: From Signal to Weight

The portfolio question differs from the chart question

A security can look attractive while being unsuitable for the current portfolio. Existing exposures, mandate limits, liquidity, and taxes affect implementation. The 2026 CMT Program Guide asks Level II candidates to establish weights using relative strength, trend, and momentum. Level III emphasizes integrated portfolio choices, sector rotation, model suitability, and risk.

Technical analysis may aid selection, timing, risk control, or tactical allocation. A manager can retain a long-term policy benchmark yet adjust a limited tactical sleeve. State which decision is being made. An authorized 10 percentage point band is different from permission to rewrite strategic policy.

Translate signals to weights

A transparent process defines an investable universe, benchmark, review date, lookback, and eligibility tests. Rank candidates by relative strength, then require an absolute trend or liquidity filter. Allocate a known sleeve subject to name, sector, and asset-class caps. State how ties, missing prices, and corporate actions are handled.

A rising relative ratio when both asset and benchmark fall indicates outperformance, not necessarily an attractive long trade. An absolute filter can address this. On the other hand, selecting several leading stocks may create one concentrated factor bet. Count common economic exposure, not merely number of tickers.

Worked investment-committee case

A 1,000,000 portfolio has weights of 60% equities, 35% bonds, and 5% cash. Its mandate allows equity between 50% and 70%, bonds between 25% and 45%, and cash of at least 5%. Sector A has a rising absolute trend and relative line. Sector B is falling and underperforming. The committee proposes shifting five percentage points from bonds to Sector A. That is 0.05 × 1,000,000 = 50,000. Asset-class weights would become 65/30/5, within those bands.

But existing Sector A exposure is 120,000 and the sector cap is 15%, or 150,000. Only 30,000 additional A fits. If the committee implements only that amount, final class weights are 63% equities, 32% bonds, 5% cash. The remaining 20,000 stays in bonds. If another sector qualifies, it can be considered separately; do not treat a strong A signal as permission to breach the cap.

A complete instruction says: add up to 30,000 to A at the scheduled rebalance if its relative ratio remains above a reference level and its absolute trend is positive; reassess if either fails; trim if the 15% cap is crossed after market moves. Set the reference level before the next observation. If estimated turnover costs exceed expected benefit, hold rather than trade.

Measure results on the right basis

Total return includes price and distributions. If the portfolio grows from 1,000,000 to 1,060,000 with no cash flows, return is 6%. A midyear contribution would make that simple calculation inappropriate for manager evaluation; use a method that accounts for flows, such as time-weighted subperiod returns.

Compare the result with the preselected policy benchmark, not a benchmark chosen after performance is known. Drawdown, volatility, Sharpe, information ratio, and downside measures answer different questions. State the risk-free rate and benchmark when relevant. A high return can still violate liquidity needs or concentration limits.

Backtests need constituent lists known at each date, tradable prices, costs, delistings, and out-of-sample review. Survivorship bias, look-ahead bias, and repeated parameter searches can overstate a sector-rotation result. An appealing historical chart is evidence to investigate, not a live-performance promise.

Review and invalidation

At each scheduled review, check mandate bands, signal status, concentration, liquidity, and costs. Use an event review after a large risk change. A tactical signal flip may reduce the overlay without liquidating a strategic position. Decide that distinction in advance. Requiring repeated observations can reduce turnover but delay exits; assess that tradeoff for the mandate.

For a Level III written response, connect signal to a specific weight, then explain the risk constraint that governs it and the condition that would alter the recommendation. "Buy the strongest sector" omits the portfolio decision.

Separate skill from the market's direction

A portfolio can gain because its benchmark rose, because tactical choices added value, or because it took more risk. Compare the portfolio with its authorized benchmark over matching dates and cash-flow conventions. If a 60/40 policy benchmark gains 7% and a tactical portfolio gains 8%, the one-point excess return is a starting observation. It is not proof of skill: higher equity exposure, sector concentration, currency moves, or luck may explain it. Attribution should identify those exposures and the cost of changing them.

For the committee case, the decision to add Sector A should be evaluated against the counterfactual of keeping the 30,000 in bonds, not against a later handpicked sector. Record the chosen benchmark and rationale before trading. If A outperforms but the added concentration causes an unacceptable drawdown, the portfolio decision may still be poor for that mandate.

The next review should ask whether the original signal still holds, whether weight drift breached limits, and whether the thesis remains useful after costs. A rebalancing rule is not a promise to buy every dip. It is a defined process for returning risk to an acceptable range while preserving the client's strategic objectives.

If a tactical overweight is right but the portfolio underperforms, investigate attribution rather than declaring the signal wrong. Other active positions, cash drag, and rebalancing costs may dominate the sector's contribution. Conversely, a wrong signal can be masked by a rising benchmark. A useful review separates signal accuracy, sizing, and execution. That separation tells the committee whether to modify the model, the risk cap, or the implementation process.

Common traps and self-check

Avoid equating one strong security with a diversified portfolio, changing benchmarks to flatter performance, reporting price return as total return, and using a backtest without costs. A position can obey asset-class bands yet breach a sector cap, as the example shows.

Self-check: A 500,000 portfolio is 55% equity, 40% bonds, 5% cash; equity is capped at 60%. How much may move from bonds to equity before other limits? Answer: Five percentage points, or 25,000, producing 60/35/5. Sector or name caps may lower the allowable trade.

Next, write a one-page committee memo with current and proposed dollar weights, risk check, transaction cost, and invalidation. Compare it with official objectives and the portfolio practice topic.