How to Build a Model Portfolio: Step-by-Step Guide for Research Analysts

How to Build a Model Portfolio: A Step-by-Step Guide for SEBI Research Analysts

For a SEBI-registered Research Analyst, the model portfolio is the product. It sets out your investment approach, it is what subscribers pay for, and over time it becomes your track record. This guide walks through building your first model portfolio: the philosophy, the stock universe, selection, weights, benchmark, risk rules, documentation, launch, rebalancing and the quarterly review that keeps it honest.

If you are new to the idea, start with what a model portfolio is. For the regulatory side, read our guide to SEBI rules for model portfolios alongside this one.

Step 1: Define your investment philosophy

Before picking a single stock, write down why your portfolio should do better than simply buying the index. That answer is your philosophy, and every later decision should trace back to it. Common approaches for Indian equity portfolios:

• Value: good businesses bought below a reasonable estimate of intrinsic value, using measures such as price-to-earnings, price-to-book and free cash flow yield.
• Growth: companies growing revenue and earnings faster than the market.
• Growth at a reasonable price: growth screens combined with valuation limits, for example a PEG ratio ceiling.
• Quality: high and stable return on equity, low debt, consistent cash generation and clean governance.
• Momentum or factor-based: stocks ranked by measurable signals such as relative strength, rebalanced on a fixed schedule.

Pick the approach that matches your actual analytical strengths. A portfolio described as "conservative quality" but filled with speculative small caps will lose subscriber trust the first time the market falls, and the mismatch will be obvious to anyone reviewing your records.

Step 2: Choose your investment universe

The universe is the set of stocks you are willing to recommend from. It should follow from your philosophy and from how many companies you can genuinely research.

• Large-cap portfolio: the NIFTY 50, NIFTY 100 or NIFTY 200.
• Multi-cap portfolio: the NIFTY 500 or BSE 500.
• Mid- or small-cap portfolio: the NIFTY Midcap 150, NIFTY Smallcap 250 or BSE SmallCap index.
• Thematic portfolio: the relevant sector index.

A defined universe of a few hundred stocks keeps research manageable. Add a liquidity floor at this stage, not later. If many subscribers will act on the same recommendation, a thinly traded stock gives them poor fills and can move the price against them. A practical check is to compare the total amount your subscribers might invest in a stock with its average daily traded value.

Step 3: Build a repeatable stock selection process

A selection process that you can describe and repeat is what separates research from tips. A typical process has four stages:

1. Screen. Filter the universe to a shortlist using numbers that match your philosophy. A quality-growth screen might require sustained return on equity, revenue growth, low debt and positive free cash flow over several years.
2. Analyse. Read annual reports and results for each shortlisted company. Understand the business model, its competitive position, management's record on capital allocation, and the industry's direction.
3. Value. Decide whether the current price leaves room for error, using discounted cash flow for predictable businesses and relative valuation against history and peers.
4. Record the thesis. For every stock, write what you expect to happen, over what period, and what would prove you wrong. This becomes your exit trigger and your evidence of research.

If you run a rules-based or factor portfolio, test the rules on historical data before launch. Our guide to backtesting a stock strategy in India covers the biases to avoid.

Step 4: Construct the portfolio and size positions

Most concentrated equity model portfolios hold somewhere between 12 and 20 stocks. Fewer than that and one stock can dominate results; many more and each position becomes too small to matter, and harder to follow.

Position sizing methods

• Equal weight. Every stock gets the same share, roughly 6.7% each in a 15-stock portfolio. Simple to explain and recommended for a first portfolio, when your conviction levels are not yet well calibrated.
• Conviction weighted. Higher weights for your strongest ideas, lower weights for positions where the thesis is still developing. Set a cap on any single position at the time of recommendation.
• Risk weighted. Weights set so that each stock contributes roughly similar volatility. In a simple version, take the inverse of each stock's recent volatility and scale the results to 100%. Lower-volatility stocks get larger weights.

Diversification limits

• A maximum weight for any single stock, both at entry and after it has risen.
• A maximum weight for any single sector, and awareness of hidden overlaps (banks and NBFCs are both financials).
• Market-cap limits if the portfolio mixes large, mid and small caps.
• A small cash buffer, if your approach allows one, and a rule for when it is deployed.

Step 5: Choose an honest benchmark

Pick the narrowest index that fully contains your universe: the NIFTY 50 or NIFTY 100 for large caps, the NIFTY 500 for multi-cap, the relevant midcap or smallcap index for those portfolios. Compare against the total return version of the index if your returns include dividends. Fix the benchmark at launch and keep it. Switching benchmarks because another one now looks easier to beat misleads subscribers. Our performance tracking guide covers benchmark choice in more detail.

Step 6: Write a risk management policy

Every model portfolio needs written risk rules before launch. Decisions made during a sharp fall are usually emotional; decisions written in advance are not. Your policy should cover:

• Position limits: maximum weight at entry and the level at which a position that has grown is trimmed.
• Sector and market-cap limits.
• Exit rules: price-based stops (fixed or trailing, set wider for more volatile stocks), fundamental stops (for example, repeated earnings misses, governance concerns, or a broken thesis), or both.
• Liquidity requirement for any stock you recommend.
• Drawdown response: what you will do, and how you will communicate, as the portfolio falls by increasing amounts.
• Monitoring frequency for portfolio-level measures such as beta, volatility and maximum drawdown.

Why drawdowns matter so much

Losses and recoveries are not symmetrical. A 20% fall needs a 25% gain to recover. A 30% fall needs about 43%. A 50% fall needs 100%. Keeping drawdowns contained keeps recovery times realistic, and it keeps subscribers with you long enough to benefit from the recovery.

Risks to plan for

• Market risk, which diversification cannot remove.
• Concentration risk in one stock or one sector.
• Liquidity risk, mainly in small and micro caps.
• Sector risk from regulation, commodity prices or technology change.
• Governance risk. No analysis fully protects against fraud, but screening promoter pledging, related-party transactions and auditor changes reduces it.

Step 7: Document the methodology

Before you publish, put your approach in writing: philosophy and objective, universe and liquidity floor, selection criteria, sizing rules, sector limits, rebalancing triggers, risk policy and benchmark. This document helps in three ways. It shows the basis for your recommendations if your records are ever reviewed. It gives subscribers something to read before they join, which reduces anxious questions when markets fall. And it forces you to make the process repeatable.

Keep a dated record of every recommendation and its rationale from the first day. SEBI's performance verification framework, PaRRVA, relies on recommendations recorded at the time they were made. See our guide on what PaRRVA is and how to prepare.

What SEBI's model portfolio guidelines require

SEBI's circular of 8 January 2025, Guidelines for Research Analysts (SEBI/HO/MIRSD/MIRSD-PoD-1/P/CIR/2025/004), includes an annexure of guidelines for model portfolios recommended by RAs. A model portfolio is defined there as a basket of securities for which a research report recommends weightages. The main requirements:

• Research report with a factsheet. The model portfolio is issued through a research report that covers every constituent, with a factsheet, disclosures, rationale, methodology, launch date, update date and portfolio type. Your view on a stock inside the portfolio must not contradict your view on that stock on its own.
• Methodology and universe. The report must describe the selection framework and label the underlying universe (large cap, mid cap, multi cap), theme or sector.
• True to label. The name must state the type of portfolio, with a one-line description of its theme or objective.
• Investment horizon stated, so investors can match it to their own.
• Review frequency defined in advance. Rebalancing must stay within the stated framework and be communicated to clients with the rationale.
• Risk disclosure in the report.
• Benchmark. Performance validated by the SEBI-specified agency, shown over different periods, against an appropriate index that is disclosed and used consistently.
• Audit. The annual compliance audit under the RA Regulations also covers these obligations.

The steps in this guide map onto these requirements. Check the current circular before launch in case anything has changed.

Step 8: Set rebalancing rules

Decide before launch what will trigger a change:

• Calendar: a fixed review, usually monthly or quarterly.
• Threshold: when a position or sector drifts beyond its limits.
• Event: results, corporate actions or news that affects a thesis.
• Hybrid: a quarterly review with event-driven exceptions. This suits most first portfolios. It gives subscribers a regular rhythm without the costs and tax drag of frequent trading.

Step 9: Run a quarterly review

A quarterly review is where you check whether the portfolio is still doing what you said it would. A structured review has four parts.

Prepare the data

• Holdings and weights at the start and end of the quarter, and every change made during it.
• Benchmark data for the same period.
• Dividends and corporate actions affecting your stocks.
• Quarterly results and management commentary for every holding.

Measure

Look at absolute return, return against the benchmark, since-inception CAGR, maximum drawdown, beta and risk-adjusted ratios. Include stocks you exited during the quarter. Leaving them out makes the review look better than the quarter really was.

Attribute

Work out where returns came from: which stocks added and which detracted (weight multiplied by return), and how much came from sector allocation compared with stock selection. Then check every holding against its written thesis.

Decide: rebalance or hold

Act when a position or sector has drifted beyond your limits, a thesis has broken, risk measures are outside your stated range, or a clearly better opportunity needs capital.

Hold when a stock has underperformed for a quarter or two but the thesis is intact, when the whole market has fallen and your portfolio has simply fallen with it, or when the tax cost of selling outweighs the benefit for subscribers.

Then tell subscribers what you found, including in poor quarters. Lead with the numbers, explain without blaming the market, and say specifically what you changed and why. Our guide to client communication for research analysts covers how to deliver these updates.

Step 10: Launch on a proper platform

Sending stock names in a chat group leaves no clean record, no consistent execution for subscribers and no performance history you can verify later. A model portfolio platform gives subscribers a dashboard, sends rebalancing updates, lets them execute through their own broker, tracks performance against the benchmark, and keeps a time-stamped record of every change. AlphaQuark's platform for SEBI research analysts covers model portfolio publishing, rebalancing across client brokers, onboarding and PaRRVA-ready records.

Common first-portfolio mistakes

• Too many stocks, which dilutes your best ideas and makes the portfolio hard to follow.
• Recency bias: filling the portfolio with last year's winners.
• Ignoring liquidity, so subscribers cannot get fair fills.
• No exit rules written before they are needed.
• The wrong benchmark, which makes performance look better or worse than it is.
• Copying another analyst's holdings, which gives subscribers nothing they could not get elsewhere.
• Skipping or delaying reviews after a weak quarter, which damages trust more than the weak quarter itself.