Quantitative Allocation Strategies: Risk Budgeting, Risk Parity, and TAA

You see a 60/40 portfolio and assume balance. Use the inputs from the worked example later in this lesson: 16% equity volatility, 5% bond volatility, and a 0.2 correlation. With those inputs, equity supplies about 92% of the portfolio's volatility.

Mean-variance (moment-based) optimization gives you a precise quantitative framework, but it can magnify estimation errors. The approaches below are alternative allocation methods. You can use them alongside moment-based optimization or in place of it.

Strategic Asset Allocation (SAA). Strategic Asset Allocation (SAA) sets the long-horizon policy mix across asset classes (e.g., 60% equity, 30% fixed income (FI), 10% alternatives). Those weights change slowly. A total portfolio approach (TPA), another allocation model the curriculum names, works differently: rather than filling fixed class weights, it judges each holding by what it adds to the portfolio as a whole. The contrast matters because class labels can mislead.

Core-satellite. Picture an investor sorting its money into two kinds of mandate, with the combined goal of beating a benchmark.

Common mistakes

• Confusing dollar weights with risk weights. Reporting 60% as equity's risk contribution just repeats the dollar weight. Compute PCR from the supplied volatilities and covariances. With 16% equity vol, 5% bond vol, and 0.2 correlation, 60/40 is 92/8 in PCR terms.
• Computing PCR without normalizing by portfolio vol. The weighted contribution $w_i \times \partial\sigma_p/\partial w_i$ is in volatility units; PCR divides it by $\sigma_p$. Reporting a contribution of 0.092 as the PCR in a 10% vol portfolio understates it; the PCR is 0.92. A result off by exactly a factor of $\sigma_p$ signals this error.
• Ignoring the borrowing cost on levered returns. Levered return is L x E[R] − (L − 1) x borrowing rate, not L x E[R]. In Example 1, 1.8x leverage gives 1.8 x 4.19% − 0.8 x 2.5% = 5.54%, not 7.54%. Volatility, by contrast, scales fully to 1.8 x 5.78% = 10.40%.

Bottom line

• Core-satellite pairs a cheap passive core with active satellites seeking alpha; each investor sets the boundary
• Bottom-up ranks deals but can concentrate; top-down needs proactive sourcing; most investors mix them
• Risk budgeting typically hits each bucket's target exactly (limits are ceilings) and needs an objective to pick one portfolio
• Asset contributions $\rho_i \sigma_i w_i$ sum to $\sigma_p$; alternatively, factor contributions $b \rho \sigma_F$ plus the residual sum to $\sigma_p$

Exam shortcut

Use the comparison table to sort the methods by what each one equalizes. Compute PCR from covariances before you judge diversification. The conditional mean is intercept + slope x current value. Levered return is L x E[R] − (L − 1) x borrowing rate, while volatility is L x unlevered volatility.

The full lesson (about 7,090 words, 47 min read) adds 2 worked examples, all 5 common mistakes, a self-check, free in the app.

Learning objectives

• mvo process
• mvo implementation
• mvo multiple risky
• mvo issues
• mvo adjustments
• mvo estimation error
• black-litterman model
• bl model private markets
• constraining bl model
• tpa overview
• tpa defining
• tpa governance
• tpa factor lens
• tpa competition capital
• tpa culture
• transitioning tpa
• tpa implementing
• maintaining tpa
• tpa gp-lp relationship
• core satellite
• top down bottom up
• risk budgeting
• factor risk budgeting
• risk parity
• other quant strategies
• taa
• taa process
• cash commitments illiquidity
• overcommitment model
• takahashi-alexander model
• historical simulation cash flow
• monte carlo-based cash flow
• managing unfunded commitments
• liquid alternatives
• lp direct investment
• co investments
• co investment returns
• secondary market pe
• gp led secondaries
• funds funds
• fof historical returns
• rebal buy hold constant mix
• rebal directional
• rebal cppi
• rebal obpi
• rebal dynamic illiquid
• rebal costs

Browse all free CAIA Level II lessons or jump into free CAIA Level II practice questions.