Model Portfolio Performance Tracking: Metrics, Benchmarks and Reporting for RAs

Model Portfolio Performance Tracking: The Metrics Every Research Analyst Should Monitor

For a SEBI-registered Research Analyst, performance is what subscribers judge you on. But "we returned X%" says very little on its own. The period, the benchmark, the risk taken and the calculation method all change what that number means. This guide covers the metrics worth tracking, how each is calculated, how to choose a fair benchmark (including NIFTY 50 versus NIFTY 500), and how to report results in a way that holds up to scrutiny.

Return metrics

Absolute return

The simplest measure: (current value − starting value) ÷ starting value. It is easy to understand but ignores time, so it cannot compare portfolios started on different dates.

CAGR

Compound annual growth rate converts a return into a yearly rate: (ending value ÷ starting value)^1/years − 1. It is the standard way to compare performance across periods. Use it only over periods of a year or more. Annualising a few strong months produces a figure that is technically correct and practically misleading.

TWRR versus XIRR

These two answer different questions, and confusing them is a common reporting mistake.

• Time-weighted return (TWRR) removes the effect of money being added or withdrawn. It measures the quality of the portfolio's decisions, independent of when any subscriber joined or how much they put in. This is the right measure for the model portfolio itself and for comparing it with a benchmark.
• XIRR (money-weighted return) accounts for the timing and size of every cash flow. It measures what a particular investor actually earned. Two subscribers following the same model can have different XIRRs because they invested at different times.

Report TWRR for the model. Show XIRR to an individual subscriber when describing their own experience. Do not present a subscriber's XIRR as the model's performance, or the reverse.

Rolling returns

One-year or three-year returns measured from every possible start date. They show how consistent the portfolio has been, rather than how lucky one start date was.

Risk-adjusted metrics

A portfolio that returned 30% with a 45% drawdown is a very different product from one that returned 25% with a 15% drawdown. Risk-adjusted measures make that difference visible.

• Sharpe ratio: (portfolio return − risk-free rate) ÷ standard deviation of returns. Excess return per unit of total volatility. In India, a government bond yield or treasury bill yield is commonly used as the risk-free rate; state which one you use.
• Sortino ratio: like Sharpe, but divides by downside deviation only, so upside volatility is not penalised.
• Information ratio: (portfolio return − benchmark return) ÷ tracking error. It shows how consistently the portfolio has beaten its benchmark, not just by how much.
• Beta: how strongly the portfolio moves with the market. Above 1 means it amplifies market moves; below 1 means it dampens them.

Drawdown metrics

• Maximum drawdown: the largest fall from a peak to a later low. It is the closest thing to "the worst a subscriber could have experienced", and subscribers care about it more than any ratio.
• Drawdown duration: how long it took to get back to the previous peak. A 25% fall recovered in four months and a 25% fall that took two years are not the same experience.
• Calmar ratio: annualised return ÷ maximum drawdown. Return earned per unit of worst-case loss.

Choosing the right benchmark: NIFTY 50 vs NIFTY 500

Most arguments about whether a model portfolio "beat the market" are really arguments about which market. The benchmark decides whether your results look skilful or lucky, so choose it carefully and keep it fixed.

This is also a regulatory point. SEBI's model portfolio guidelines for Research Analysts (circular of 8 January 2025) require each model portfolio to be benchmarked against an appropriate and relevant index, to disclose that benchmark, and to use it consistently. Their examples: an auto-stock portfolio against the Nifty Auto index, a mid-cap portfolio against a midcap index.

What each index represents

The NIFTY 50 holds 50 large, liquid companies weighted by free-float market capitalisation. It is concentrated: a handful of sectors and its largest constituents drive much of its return.

The NIFTY 500 covers the 500 largest companies across large, mid and small caps. Its largest holdings overlap heavily with the NIFTY 50. What sets it apart is the long tail of mid- and small-cap companies the NIFTY 50 never holds.

Why a mismatch distorts your scorecard

• Benchmark a multi-cap model against the NIFTY 50 and a broad small-cap rally will make you look skilful when you are simply holding a part of the market the index does not.
• Benchmark a large-cap model against the NIFTY 500 and a small-cap sell-off will flatter "defensive" calls that were never exposed to that risk.

Match the benchmark to the universe

• Large-cap model (recommending from roughly the top 50 to 100 names): NIFTY 50, or NIFTY 100 if your universe is a little wider.
• Multi-cap or flexi-cap model: NIFTY 500, because it contains the same cap range your picks come from.
• Mid- or small-cap model: the NIFTY Midcap 150 or NIFTY Smallcap 250, with the NIFTY 500 as a secondary reference if you want one.
• Thematic model: the relevant sector index.

Use the total return index

If your portfolio returns include dividends, compare against the total return index (TRI) variant of the benchmark, not the price index. Comparing a dividend-inclusive portfolio with a price-only index creates outperformance that is not real.

Tracking error, liquidity and rebalancing

A concentrated large-cap model will show low tracking error against the NIFTY 50 and high tracking error against the NIFTY 500, simply because it leaves out the mid- and small-cap tail. That is a structural mismatch, not skill. If you anchor to the NIFTY 500, build a liquidity filter into stock selection, because some of its smaller names cannot absorb many subscribers buying at once. The broader index also changes membership more often at the mid/small boundary, so set a rule for handling those changes.

A simple decision rule

1. Write down the actual cap range and liquidity floor you recommend from.
2. Choose the narrowest index that fully contains it.
3. Check how your model reads against it in both a large-cap-led year and a broad-market year. If your outperformance only appears in one, it may be a cap or sector tilt rather than stock selection.
4. Disclose the benchmark and keep it. Switching because another index now looks easier to beat misleads subscribers.

Portfolio characteristics to monitor

• Concentration: share of value in the top five and top ten holdings.
• Sector weights against your stated limits and against the benchmark.
• Turnover: how often holdings change, which drives costs and tax for subscribers.
• Hit rate: the share of closed recommendations that made money, shown alongside average gain and average loss so one does not mislead without the other.
• Average holding period, to check that what you do matches what you say your approach is.

Attribution: where did the return come from?

Each holding's contribution is its weight multiplied by its return. A 10% position that rose 20% added about 2 percentage points. Go one level further and separate sector allocation (being over- or underweight sectors that did well or badly) from stock selection (picking better or worse stocks within each sector). Include stocks exited during the period. Analysing only current holdings hides the losses that were cut.

Reporting performance to subscribers

How often

Many research analysts send a short monthly update and a fuller quarterly report with attribution and outlook. Whatever you choose, stay on schedule in poor periods as well as good ones.

What to include

• Returns over several periods (one month, three months, one year, three years, since inception), next to the benchmark over the same periods.
• Maximum drawdown alongside returns, never returns alone.
• Whether figures are gross or net of costs and fees.
• The benchmark used, and whether it is the TRI.
• A clear statement that past performance does not guarantee future results, and your SEBI registration number.

Disclosure and advertising rules for research analysts change, so check the current SEBI circulars before publishing performance in marketing material.

Verified performance and PaRRVA

SEBI's Past Risk and Return Verification Agency (PaRRVA) verifies past performance claims made by research analysts and investment advisers. If you want to communicate past performance publicly, it needs to come from recommendations recorded at the time they were made, not reconstructed later. Read what PaRRVA is and how it works, or see how AlphaQuark supports PaRRVA performance verification.

Automating performance tracking

Calculating NAV, TWRR, drawdowns and attribution by hand across several model portfolios is slow and error-prone, and one spreadsheet mistake can end up in a client report. AlphaQuark's platform for SEBI research analysts records every rebalance as a dated advice record and tracks model performance against the benchmark, so your figures come from the same records you would submit for verification.

New to running a model portfolio? Start with our guide to building your first model portfolio.