Accounting Red Flags Checklist: 4 Checks for India

A good cook smells the pot before tasting it. Too much oil, a stale spice, milk on the turn: the nose catches it before the tongue does. No chemistry is needed. The smell comes before the taste, and the smell is where the warning lives.

A company's accounts work the same way. You don't need to be a forensic auditor to catch the first whiff of something wrong; you need a short list of things to sniff for, and you need to know which page of the annual report each one lives on. Not every problem shows up in the filings, but many of the warning signs that matter are printed there, in notes most readers never open. This is a smell test of four checks, each one a question you can answer from documents every listed Indian company has to publish.

Why the smell test matters more than the headline

The number every app shows you, net profit, is the most cooked dish in the kitchen. It sits at the bottom of the income statement, so every choice management made on the way down has already been stirred in. Warren Buffett, in his 2000 letter to shareholders, described chief executives chasing earnings targets they had announced who, "after exhausting all that operating acrobatics would do, they sometimes played a wide variety of accounting games to 'make the numbers'" (Berkshire Hathaway, 2000 letter). The pressure to hit a number the company has already said out loud is where the games begin.

There is a second reason to read the filing rather than the app. Data screens re-label lines to make thousands of companies fit one template, and the re-labelling can hide exactly the thing you are smelling for. An earlier version of this checklist made that mistake itself: it described a large one-off gain at Suzlon Energy as "other income", which is how the summarised data presented it. Suzlon's own filing calls it an exceptional item, and the difference matters for Check 2. So treat the profit figure as the finished plate, and treat any summary table as someone else's description of the plate. Lean over the pot yourself.

Check 1: Does profit turn into cash, or just into receivables?

The first and most important smell is profit rising while cash does not. A real profit eventually shows up as money in the bank. A paper profit shows up as a growing pile of receivables, bills the company has booked as sales but has not collected. Book enough of those and profit climbs while the bank balance stalls. The gap parks itself in receivables and inventory, together the working capital a business needs to keep running.

The cleanest test is cumulative: add up five years of net profit and five years of cash flow from operations, and compare the two totals. A single year can be lumpy; a five-year gap is harder to explain away, though not impossible to fake, as the limits at the end of this piece show. Cumulative cash flow against profit explains why the multi-year sum beats any one year's ratio, and cash is truth covers the reasoning in full.

One fence before you run it: Checks 1 and 3 are built for non-financial companies. At a bank or NBFC, lending is the operating activity, so operating cash flow mostly tracks loan growth and is often negative for a lender that is growing; asset quality is the test there instead.

Next, look in the notes for a trade receivables ageing schedule; it is the smell test made specific. Suzlon's FY2025 annual report carries one in note 11.1 to its consolidated financial statements, on pages 384 and 385. It splits receivables by how long they have been outstanding from the due date of payment, from less than six months to more than three years, and separately into undisputed and disputed amounts (Suzlon Energy annual report 2024-25, note 11.1).

At 31 March 2025, ₹1,731 crore of Suzlon's ₹4,048 crore of gross receivables was not yet due, ₹454 crore was more than a year past due and ₹169 crore of that more than three years. Cut the other way, ₹34 crore was disputed. The allowance for credit impairment and expected credit loss against the whole book was ₹181 crore, and the same note's table for 31 March 2024 put ₹329 crore in the more-than-three-years bucket. Receivables more than a year past due with little provision set against them, or a disputed pile that keeps growing, are revenue the company booked and has not been able to turn into cash.

Some managements put a number on the cash cycle on their earnings calls, a working-capital-days target or a receivables goal. When they do, that number is a commitment you can hold them to, and the Guidance Tracker is where you can see whether later calls kept repeating it, revised it or let it go quiet. None of this is an accusation. It is the signal an owner watches: management itself naming the cash cycle as the number it has to bring down.

Check 2: Is the profit from the business, or from a line that won't repeat?

The second smell is operating profit flat while net profit jumps. Two lines usually carry the jump. Other income is money earned outside the core business, such as interest on cash or gains on investments, and it often recurs in some form. Exceptional items are gains or losses the company itself says are unusual: a debt settlement, a sale of a division, a large write-back. Both deserve the same question, can this happen again, but the exceptional line is where the company has already told you the answer is probably no.

Suzlon's FY23 is the clearest example, in its own filing. The press release it filed with the exchanges on 30 May 2023 reports FY23 revenue of ₹5,947 crore and EBITDA of ₹832 crore, and headlines "PAT before exceptional items of INR 167 crores" (Suzlon Q4 and FY23 press release and investor presentation, BSE filing). The investor presentation's results table, in the same filing, shows profit after tax of ₹2,887 crore, including an exceptional gain of ₹2,721 crore. About 94% of the headline profit (₹2,721 crore of ₹2,887 crore) came from the exceptional line.

Suzlon's own note to its results breaks that line down: ₹2,397 crore was the "difference on de-recognition as a result of conversion in full of OCD and CCPS". In May 2022 its optionally convertible debentures were converted into 57.14 crore new Suzlon shares, and a subsidiary's compulsorily convertible preference shares into that subsidiary's shares (Suzlon allotment on conversion of OCDs, BSE filing, 24 May 2022). The instruments came off the balance sheet and the accounting difference was booked as a gain, with no cash received and new shares issued in their place. The rest came from de-recognising assets and liabilities, selling property and plant, and selling stakes in associates, a joint venture and a subsidiary (Suzlon results for the quarter to 30 September 2023, note 5 with the audited FY23 column, BSE filing). None of it came from selling or servicing turbines.

Nothing was hidden. Suzlon's chief financial officer led with the smaller number: "Our consolidated FY23 PAT before exceptional items of Rs. 167 crores is the highest reported by the company in 6 years." Anyone who anchored on ₹2,887 crore, or on a summary table that folded the exceptional gain into other income, read the year's earning power as roughly seventeen times what the business produced.

The next year showed it. On the FY24 results call, the CFO reported EBITDA of ₹1,029 crore and PAT before exceptional items, management's own measure, of ₹714 crore, more than four times the FY23 figure of ₹167 crore (Suzlon Q4 and FY24 earnings call transcript, filed 31 May 2024; FY23 comparison from Suzlon audited FY24 results and investor presentation, BSE filing, 24 May 2024). Reported net profit, after a ₹54 crore exceptional loss and with no ₹2,721 crore gain to repeat, was ₹660 crore, a fall of more than three-quarters from the FY23 headline (Suzlon audited FY24 results and investor presentation, BSE filing, 24 May 2024).

The business did not improve fourfold, though. EBITDA rose about 24%, and the CFO credited the rest partly to "lower interest costs as a result of repayment of the entire debt", with net finance cost falling from ₹401 crore to ₹126 crore. Suzlon funded that repayment largely by issuing new shares, raising ₹2,000 crore by selling 114 crore new shares to institutions in August 2023 and repaying ₹1,789 crore of long-term borrowings in the same half-year (Suzlon results for the quarter to 30 September 2023, cash flow statement, BSE filing), so the bigger profit is shared across more shares.

The check does not predict the share price, so a fall in the price proves nothing either way. What it settles is which number you are paying for: a price-to-earnings ratio worked out on the FY23 headline makes the stock look far cheaper than one worked out on profit before exceptional items.

Check 3: Are real costs being parked on the balance sheet?

The third smell is subtler: costs that should reduce this year's profit being carried as an asset instead. When a company spends on something that lasts years, a plant or a line, it is fair to spread that cost over time; until the asset is ready, the spending sits in capital work-in-progress (CWIP), or in intangible assets under development for software and product spend. Some of this is legitimate, including some interest on money borrowed to build the asset. The trick is when interest and overheads keep being capitalised after work has stalled, or after the asset is already in use. Profit looks higher because a real expense is sitting in CWIP.

The CWIP note in the annual report makes this checkable. Suzlon's FY2025 report carries a CWIP ageing schedule in note 5.2 to its consolidated statements, on page 380: the balance split into less than one year, one to two years, two to three years and more than three years, and into projects in progress and projects temporarily suspended. Under the table the company states whether any project's completion is overdue or has exceeded its cost compared with its original plan. For FY2025 it says none had, and that the ₹89 crore in CWIP at the year-end is expected to be completed within one to two years (Suzlon Energy annual report 2024-25, note 5.2).

Capital work-in-progress as a red flag walks through a much larger one, NHPC's FY25 consolidated schedule, line by line. The flag is money that sits in the "more than three years" bucket while the completion schedule keeps moving, or CWIP growing as a share of total assets while operating profit and cash go nowhere. Growth companies legitimately carry rising CWIP; the question is whether it ever finishes and starts earning.

Check 4: Who is leaving the kitchen, the auditor or the CFO?

The last smell is not in a number at all. It is in the people. A statutory auditor or chief financial officer who leaves abruptly is the cook walking out mid-meal. They have seen the pot up close. Indian rules make the auditor's exit unusually readable. The SEBI listing regulations require a listed company to disclose the detailed reasons for an auditor's resignation to the exchanges "as soon as possible but not later than twenty four hours of receipt of such reasons from the auditor" (SEBI listing regulations, Schedule III, Part A, Para A, clause 7A, as amended up to 14 July 2026). SEBI's circular of 18 October 2019 (CIR/CFD/CMD1/114/2019) went further (SEBI circular CIR/CFD/CMD1/114/2019), and its rules now sit in SEBI's master circular for listed companies (SEBI master circular on the listing regulations, last updated 30 January 2026, Section V-D, paragraphs 6.1, 9.1 and 9.3).

The circular's additions are practical. A resigning auditor cannot walk away from work in progress: one who leaves within 45 days of a quarter's end first signs that quarter's limited review or audit report, one who leaves later also signs the next quarter's, and one who has signed the first three quarters' reports signs both the last quarter's limited review or audit report and the annual audit report. The auditor's information then follows a set format, and the company must disclose its audit committee's views to the exchanges within twenty-four hours of the meeting that considers them.

That format is the useful part for a reader, because it tells you what to look for in the filing. Annexure A of the 2019 circular (Annexure 21 of the master circular) asks the auditor for the detailed reasons for resignation, any concerns raised and the efforts made before leaving, and, where information was not provided, whether that was "due to a management-imposed limitation or circumstances beyond the control of the management", whether the missing information "would have significant impact on the financial statements", and whether the problem existed in earlier reported results. An auditor who answers those with "pre-occupation" and nothing else has told you something by not telling you.

How to read what the auditor actually signed is covered in the auditor's report; the pattern of auditor and CFO changes over five years is the thing to tabulate. One change is life. A revolving door is a smell.

The checklist, on one page

The smell test isn't a verdict; it's a reason to look closer

One honest caveat, because a checklist can be misused. A single red flag is not proof of fraud, and Suzlon is named here to teach the checks, not as a buy or sell call. Other income can be legitimate. CWIP can be real building. An auditor can reach the end of a term. A cook's nose doesn't condemn the dish; it says to taste carefully before serving it.

The other caveat runs the opposite way. These checks are better at catching aggressive accounting than determined fraud, and a company that fakes its bank balances can pass Check 1. In his email to SEBI of 7 January 2009, Satyam Computer Services' chairman admitted that its balance sheet carried "inflated (non-existent) cash and bank balances" of ₹5,040 crore as of 30 September 2008, ₹376 crore of accrued interest that did not exist, and liabilities understated by ₹1,230 crore (SEBI order in the matter of Satyam Computer Services, 2014).

Because the interest was faked too, a reader who checked that the pile earned a normal return would have been reassured. Two questions get closer, though neither is proof: is the interest on a large cash pile actually received, or does it pile up as interest accrued and not yet received, and why is a cash-rich company still borrowing? A fraud that complete can pass every check in this piece.

That is the right use. A flag doesn't end the analysis, it starts it: it tells you which note in the annual report to read next and which question to ask on the next concall. If you want the order to read the whole document in, how to read an annual report like an analyst sets it out, and related party transactions and contingent liabilities are the next two pots to lean over. The investors who get hurt aren't the ones who smelled something and looked closer. They are the ones who saw a fat headline and tasted without smelling. What you do after you smell something is where the real work, and the margin of safety, begins.

Frequently asked questions