← Q&A archive
02-strategic-allocation May 2, 2026

Walk me through the efficient frontier

The Efficient Frontier

What it is

Every combination of assets you can hold — Nifty 50, fixed income, gold, international equity, cash — plots as a point in risk-return space (x-axis: Volatility — definition">volatility, y-axis: Expected Return — definition">expected return). There are theoretically infinite such combinations.

The Efficient Frontier — definition">efficient frontier is the upper-left boundary of that cloud. [8] Any portfolio on it either:

Portfolios inside the cloud are wasteful — you're taking more risk than necessary, or earning less return than possible. [8]


The maths behind it

For each target return $\mu^*$, you solve:

$$\min_{\mathbf{w}} \ \mathbf{w}' \Sigma \mathbf{w}$$
$$\text{subject to} \quad \mathbf{w}' \boldsymbol{\mu} = \mu^*, \quad \mathbf{w}' \mathbf{1} = 1$$

Sweep $\mu^*$ upward from the lowest achievable return and you trace out the full frontier. [2]


Two landmark points

Point What it is
Minimum Variance Portfolio (MVP) Leftmost point — lowest possible volatility. Often has meaningful gold and bond exposure even for growth investors.
Tangency Portfolio The point with the highest Sharpe Ratio — definition">Sharpe ratio $\left(S = \frac{E[R_p] - R_f}{\sigma_p}\right)$. This is where the CAL — Capital Allocation Line — definition">Capital Allocation Line (CAL) from the Risk-Free Rate — definition">risk-free rate just touches the frontier.

[2][1]


Adding the risk-free asset

Once you introduce a risk-free asset (think: RBI repo rate, ~6.5%), you can mix it with any frontier portfolio. The best such mix is the Capital Allocation Line (CAL):

$$E[R_C] = R_f + \frac{E[R_T] - R_f}{\sigma_T} \cdot \sigma_C$$

The steeper this line, the better — and it's steepest when it touches the tangency portfolio. Every rational investor then holds some blend of the risk-free asset and this tangency portfolio. Your risk aversion determines the blend, not which risky portfolio you hold. [7]


An illustrative Indian example

Asset Expected Return Volatility
Nifty 50 12% 20%
Indian Fixed Income 7.5% 5%
Gold (INR) 9% 16%
International Equity 11% 18%
Cash (liquid) 6.5% 0.5%

With these inputs and the correlations between asset classes (e.g., Nifty/fixed income ≈ −0.10, Nifty/gold ≈ 0.05), the tangency portfolio lands roughly at 55–65% equity, 15–20% fixed income, 10–15% gold. [2]

Note that gold earns a relatively high weight not because its standalone return is the best, but because its low Correlation — definition">correlation with equity reduces overall portfolio risk — exactly what the maths rewards. [1]


One critical caveat

Don't over-optimise. When you feed sample return estimates into the optimiser, it amplifies estimation errors. DeMiguel et al. (2009) showed that a simple equal-weight (1/N) portfolio beat 14 sophisticated optimisation strategies out-of-sample across 7 real datasets. [4]

The practical lesson: use the frontier to understand why diversification works and to set rough allocation bands — not to chase a precise 39.7% vs. 40.0% equity weight. [4]


Apply this → Portfolio Builder — enter your own expected returns and volatilities, select asset classes, and see your current portfolio plotted against the efficient frontier in real time.

Sources cited

nism 14.6 The concept of Efficient Frontier
lecture Strategic Asset Allocation