How can I see what stocks a mutual fund actually holds, and how do I judge it against the right benchmark?

What is disclosed, and how often

SEBI requires every AMC to publish each scheme's complete portfolio every month — every security, its quantity, its market value and its percentage of net assets. This is the most under-used data in Indian fund research, because it is published as a spreadsheet per fund house and almost nobody reads it.

What we hold, counted on 3 October 2026: 320,336 disclosed positions across 2,399 schemes from 49 fund houses, with the newest portfolios dated 31 August 2026.

Two honest limits. Disclosure is monthly and in arrears, so a holdings view describes the portfolio as filed at month-end, not as it stands today — a manager may have sold out of a position weeks ago. And a fund house that has not filed is simply absent; those funds show as missing rather than as zero, because an empty chart reads as "holds nothing", which would be false.

What the holdings tell you that returns cannot

• Concentration. The top-10 weight is the fastest read on how much the fund is betting. A fund with 55% in ten names is a different animal from one with 22%, whatever their returns look like.
• Whether the label is true. A "flexi cap" fund holding 85% large caps is a large-cap fund with a wider mandate it is not using. The market-cap split from the disclosed portfolio settles it.
• Sector bets. Strong returns often come from one sector call. The sector split shows whether you are buying a stock-picker or a bet on financials.
• Overlap with what you already own. The single most useful use of holdings data, and the reason it is worth having at all.

Because every weight is a percentage of net assets, the market-cap split is shown as a share of what has been classified, with the coverage stated. A bar that is two-thirds full because a third of the portfolio could not be classified, drawn as though it were the whole picture, is a lie by omission.

Matching is by ISIN, not by name

"Bajaj Finance Ltd." and "Bajaj Finance Limited" are the same company and a name match misses it, which quietly understates overlap and concentration everywhere. Every security is matched on its ISIN. Where a name must be shown, the scheme's plan and option are appended if they would otherwise be ambiguous, because two plans of one scheme sharing a name is how a table ends up claiming a fund overlaps itself.

The right benchmark, and why TRI matters

A fund's benchmark is set by its SEBI category, not by its name:

All 32 indices here are Total Return indices, with daily values from 2 January 2012. A price index ignores the dividends the companies paid; the fund collects them. Measure a fund against a price index and it wins by roughly the market's dividend yield every year without the manager doing anything.

The quieter version of the same mistake is benchmarking against an index fund instead of the index — then you are partly measuring the tracker's expense ratio and cash drag, and every fund in the category looks slightly better than it is.

Does it beat its benchmark consistently?

"Consistently" needs more than one window. The useful reads are alpha over several windows rather than one, and up/down capture, which splits the answer into rising and falling markets. A fund can beat its benchmark overall purely by losing less in falls — worth knowing, and invisible in a single alpha number.

Tracking error and tracking difference

For an index fund these are the only measures that matter, because it is not trying to beat anything. Tracking difference is how far its return fell short of the index over a period — mostly the expense ratio. Tracking error is how much its returns wobble around the index, annualised; it measures how reliably it tracks. A fund can have a small tracking difference and a large tracking error, which means it got to roughly the right place by a scenic route.