Why India Needs Rule-Based Portfolio Governance

India needs rule-based portfolio governance because a professional investment process must do more than find good stocks. It must apply the same mandate across a broad market, preserve what was known at the time, document why judgement overrode a rule, connect every holding to measurable watchpoints and reproduce the decision months later. Rules make that process consistent and reviewable. They do not automate conviction.

That distinction matters. “Rule-based” is often heard as “the computer picks the stocks.” That is only one possible design, and usually not the most useful one for a fundamental PMS, AIF or family office.

A better interpretation is simpler: the firm writes down how research becomes capital, then makes the record difficult to rewrite after the outcome is known.

The real governance problem

Most investment teams already have a process. The trouble is that different parts of it live in different places:

• the universe and ratios sit in a screener;
• the scorecard sits in one analyst’s spreadsheet;
• the thesis sits in a note or presentation;
• the committee discussion sits in memory;
• the portfolio sits in an accounting system;
• the alerts sit in email or chat;
• the post-mortem, if it happens, starts from reconstructed history.

Each part may be reasonable on its own. The failure appears at the joins.

Was the ratio in the committee memo the one that existed on the approval date, or today’s restated value? Did the portfolio buy a company because it passed the scorecard, because an analyst overrode it, or because the scorecard itself changed? When management later missed guidance, who was supposed to notice? When a client asks why the holding existed, can the firm show the rule, evidence and version that produced the decision?

Portfolio governance is the system that answers those questions before they become archaeology.

Why the Indian market makes this harder

One ratio does not mean one thing across the market

A bank, NBFC, insurer, holding company and manufacturer do not share the same economic or accounting shape. Return on capital is central for many industrial businesses and often inappropriate for a bank. Debt is operating raw material for a lender and financial risk for many non-financial companies.

A universal scorecard that forces every company through the same factors creates false precision. A governed process specifies which measures apply to which business type, how peer groups are chosen, and what happens when a factor is not meaningful.

The date attached to a number is not enough

A quarter ended on 31 March was not knowable on 31 March. Results arrived later. A prior period may then be restated or reclassified in a subsequent filing.

If a backtest reads the latest cleaned version of an old number and stamps it against the period end, it learns from the future. If a committee record later displays the latest value instead of the value it actually saw, it rewrites the past.

That is why point-in-time data is a governance requirement, not only a quant nicety. Read why point-in-time data matters for the mechanics.

Corporate actions and identifier changes are common enough to be structural

Splits, bonuses, rights issues, mergers, demergers, renamed entities and changing index membership all affect historical tests and portfolio records. A rule cannot be reproduced if the system cannot identify the same economic security across those changes.

Liquidity differs sharply across the listed universe

A screen may find an attractive company that a small portfolio can buy and a larger one cannot. A governed process therefore separates “passes the research rule” from “fits the mandate at this capital base.” Liquidity, capacity and position limits belong in the specification, not in a last-minute judgement after the backtest looks good.

Quantitative and qualitative evidence arrive on different clocks

Prices move daily. Ownership is periodic. Financial statements arrive quarterly. Annual reports contain details that may not exist anywhere else. Management guidance changes in calls and presentations. Regulatory or governance events arrive when they arrive.

A portfolio is not governed if the quarterly scorecard is disciplined but the evidence that could invalidate it has no owner, trigger or review rule.

Rule-based does not mean judgement-free

Every systematic process contains judgement. Someone chooses the universe, factor definition, window, threshold, weight, benchmark and review frequency. The difference is that the judgement is made explicitly and can be challenged.

The same is true for a fundamental process. A view on management quality cannot be reduced honestly to one universal number. It can still be governed:

• which evidence must be read;
• which promises are tracked;
• what counts as a missed or changed commitment;
• who may record a qualitative veto;
• what reason code accompanies the veto;
• when the decision must be reviewed.

The rule governs how judgement enters, not what the judgement must be.

The eight objects in a governed portfolio process

1. Mandate and eligible universe

Write down what the strategy is allowed to own: exchanges, market-cap or liquidity bounds, security types, listing history, sector restrictions, promoter pledge rules and any client-specific exclusions.

The universe must be reconstructable on a past date. “Today’s Nifty 500” is not the Nifty 500 that existed five years ago.

2. Hard gates

Gates remove companies before ranking. Examples include insufficient history, missing financial statements, unacceptable liquidity, leverage beyond the mandate or a governance condition the firm has decided it will not underwrite.

A missing observation should produce “not assessable” or “ineligible,” according to the written policy. It should not silently become zero.

3. Versioned scorecard

The scorecard defines the factors, formulas, weights, peer groups and treatment of outliers. A useful score is inspectable. “Quality: 82” is not a definition.

When the model changes, the version changes. The old run remains available so a past decision does not inherit today’s improved methodology.

4. Qualitative research and exception policy

The quantitative process produces a shortlist, not a verdict. The qualitative layer asks what the numbers do not capture: business model, competitive structure, capital allocation, management promises, accounting choices, regulatory exposure and disconfirming evidence.

A human veto is legitimate. An invisible veto is not. Record who made it, the evidence, the reason and the date. Then measure whether vetoes added value or merely reflected discomfort.

5. Position sizing and concentration limits

Approval and size are separate decisions. The rule should specify starting weight, maximum weight, sector and factor concentration, liquidity capacity, escalation conditions and what happens after a large price move.

For a family office, the exposure map should include direct stocks and the companies held indirectly through funds. Ten line items are not ten independent risks if they own the same underlying businesses.

6. Decision record

Each live position should retain the mandate version, scorecard version, input date, quantitative output, qualitative evidence, committee decision, exceptions, approved size and expected review date.

This does not need to become a bureaucratic essay. A compact record is enough if another person can reproduce why the position entered the book.

The SEBI regulations index is the authoritative place to consult for the current Portfolio Managers Regulations and formal obligations. This article describes operating discipline, not legal compliance, and a software workflow does not replace a firm’s regulatory duties.

7. Monitoring rules

The thesis should create the watchlist, not the other way around. Convert its load-bearing assumptions into observable conditions:

• a financial threshold is crossed;
• management changes or withdraws guidance;
• cash conversion weakens beyond the agreed range;
• promoter pledge or institutional ownership changes materially;
• a filing introduces a new related-party, contingent or regulatory risk;
• portfolio exposure breaches a limit after market movement.

An alert should say which rule fired, which holding is affected and which evidence changed. A stream of generic news is not governance.

8. Review, rebalance and feedback

Every rule needs a review clock and an exception process. The team should know when the portfolio is recomputed, when a thesis is formally re-underwritten, who can defer action and how that deferral is recorded.

After the outcome, compare expectation with reality. Which factor failed? Which management promise was missed? Were analyst vetoes useful? Did the process behave differently from its written description? Governance becomes an advantage only when it learns.

Where Excel fits

Excel remains one of the best tools for understanding and challenging a calculation. Analysts can trace formulas, change an assumption, build a bridge and explain a result to a committee.

Its weakness is not arithmetic. Its weakness is control at scale. Files are copied, formulas drift, market-wide rankings become fragile, historical inputs are overwritten and nobody can always tell which version made the decision.

The useful architecture is therefore not “replace Excel.” It is:

1. keep source data, time, rules and versions in a controlled research system;
2. compute market-wide screens and rankings consistently;
3. export the input values and live formulas;
4. let the analyst reproduce, challenge and extend the work in Excel;
5. preserve the approved version back in the decision record.

Verification is stronger when the platform does not ask the user to trust the platform.

How Altys approaches the problem

Altys is designed around this rules-to-record loop for Indian investment teams. The same point-in-time research layer can support market screens, factor scorecards, historical tests, company diligence, committee output and monitoring. Quantitative calculations are produced by explicit formulas; qualitative outputs remain connected to source evidence; and important outputs can be exported to Excel for independent inspection.

The platform does not decide what a firm should believe. A PMS, AIF or family office brings its own mandate, factors, weights, questions and judgement. Altys makes that process repeatable and reviewable.

That is also the boundary of its role. Altys is not a broker, does not execute trades, does not provide stock tips and is not a substitute for an investment committee.

A 30-day starting plan

Do not begin by translating the whole firm into software. Pick one real strategy and one real review cycle.

Week 1: write the current process. Capture the universe, gates, ranking logic, sizing rules, exceptions and monitoring habits exactly as they exist, including the informal parts.

Week 2: reproduce today’s portfolio. Run the written process and explain every mismatch. A mismatch may reveal an undocumented judgement, a data definition problem or a rule nobody follows.

Week 3: replay history honestly. Test one sealed version using point-in-time inputs, historical membership, implementable timing, costs and liquidity assumptions.

Week 4: arm monitoring. Turn each live thesis into a small set of evidence-linked watchpoints, assign owners and replay recent history to remove noisy conditions before alerts go live.

The goal is not a prettier dashboard. It is a process the firm can explain, inspect and improve.

The principle to keep

India does not need more black-box scores or more confident AI recommendations. It needs infrastructure that helps investment teams state their process clearly, apply it consistently, preserve what happened and learn when reality disagrees.

A rule is not valuable because it is always right. It is valuable because it makes the decision visible enough to test.

Related reading

• Best quant investing tools in India
• Best qualitative stock research tools in India
• A rule-based investment committee framework
• Systematic versus discretionary investing
• Portfolio monitoring versus portfolio tracking

This article is educational and describes research-process design, not legal or investment advice. Altys Labs is not a SEBI-registered Research Analyst or Investment Adviser, and nothing here is a recommendation to buy, sell or hold a security.