The forensic mindset

There is a way of reading a company's accounts that people mistake for cynicism, and it is nothing of the kind. It does not begin by assuming the numbers are false; it begins by assuming nothing at all, and then it does one disciplined thing over and over - it ties every number the company reports to another number that should move with it, and asks the company to explain any place where the two have come apart. Profit should turn into cash. Sales should show up as cash collected or as a receivable owed. A plant of a given capacity can make only so much output. A company earning book profit should, over time, pay tax on it. Each of those is a tie, and the forensic mindset is simply the habit of never letting one side of a tie stand alone. illustrative

The distinction from paranoia matters, because paranoia is useless. A paranoid reader assumes fraud, hunts only where he already suspects it, and finds whatever he set out to find. A forensic reader assumes neither honesty nor dishonesty; he reconciles. He runs the same ties on the dull, clean company as on the exciting one, because it is only by knowing what reconciled accounts look like that he can notice the year one line has drifted from the number it should agree with. The stance is not an accusation aimed at a suspect. It is a standing question - if this company wanted to look better than it is, how would it do it, and does the reconciliation rule that out? - applied to every report that crosses the desk.

The idea underneath the ties is articulation: the three statements are not independent but three views of one underlying reality, so a number in one must be consistent with numbers in the others. Because a business runs on the accrual basis - recording revenue when earned, not when cash arrives - profit and cash can legitimately diverge for a while; but they cannot diverge forever, and the shape and location of the divergence is exactly what the forensic reader reads. This module installs the posture; the rest of Part Four is the catalogue of specific ways each tie gets broken on purpose. The revenue games, the expense games, the balance-sheet games, the cash-flow games are all, at bottom, ways of making one side of a tie say something the other side would contradict if you checked. The defence against every one of them is the same word: reconcile.

The method is a small set of ties, run every time, in the same order.

Profit to cash. The first and most powerful reconciliation is reported profit against operating cash flow. Over a single year they can diverge for good reasons - a growing company funds working capital, a seasonal business ends mid-cycle - but over two or three years profit that is real turns into cash. When profit rises and operating cash does not follow, the forensic reader concludes nothing yet; he asks where the profit went, and the cash flow statement answers: into receivables, into inventory, into something. The gap is not the verdict. The gap is the question, and the account it landed in is the next thing to read.

Sales to receivables and collections. A sale recognised when earned becomes either cash or a trade receivable that is later collected. When reported sales grow much faster than the business collects, the difference piles up as receivables - and a receivable growing faster than sales, year after year, is revenue booked but not turned into money. It may be a legitimate change in terms or a large late-year order; it may be channel stuffing, pushing product to distributors to book the sale, dressed as demand. The reconciliation does not tell you which. It tells you the sales number and the collection number disagree, and that the company owes you an explanation.

Capacity to output, and book profit to tax. Some ties are physical: a plant of a stated capacity can make only so many tonnes, a hotel has only so many rooms, an airline only so many seats, so revenue implying more output than the assets could produce is revenue that is not what it claims. Others are fiscal: a company reporting healthy book profit should, across years, pay cash tax roughly consistent with it, because the tax authorities reconcile too. A large and persistent gap between book profit and tax paid is either a genuine timing or exemption difference - or a sign that the profit shown to shareholders is larger than the profit shown to the taxman. Either way, the tie names the thing to ask about.

These four are not an exhaustive list; they are the template. The move underneath all of them is identical - take the number the company wants you to read, find the number that must agree with it, and treat any drift between them as a claim to be substantiated. You run the ties on every company, clean or suspect, because the value of the stance is comparative: only by reconciling the honest accounts do you build the sense of what agreement looks like, so that the year a tie breaks, you notice.

The stance is universal, but the tie that carries the most weight moves from sector to sector - the same instruction, "reconcile," points at a different account depending on the business, and that is what makes a single forensic posture work everywhere.

The inversion is that the same instruction interrogates a completely different line in each business, and the line that matters most in one is invisible in another. Tell a reader to reconcile a bank and the whole weight falls on the provision and the stock of bad loans, because that is where a lender manufactures or destroys profit - yet a manufacturer has no such line, and its decisive tie is sales against physical output and inventory. An IT firm's danger lives in unbilled revenue, a developer's in the gap between recognised revenue and cash collected, and an FMCG company's almost nowhere. A reader who learns one sector's reconciliation as the reconciliation will apply the wrong tie everywhere else - challenging a manufacturer's provisions while ignoring its inventory build, or trusting a bank because its receivables look clean while its under-provisioning goes unchecked. The stance survives the move across sectors precisely because it tells you to find the number that should agree, and lets the business tell you which number that is.

Take a composite mid-cap that reported a strong year. Revenue up 38%, profit up 44%, the management commentary all momentum and market share. On the headline it is a fine result. The forensic reader does not stop at the headline; he runs the ties. illustrative

First, profit to cash: reported profit was ₹520 crore, but operating cash was ₹180 crore - a third of it - where the year before the two had been close. That is the gap, and the cash flow statement shows where the profit went. Second, sales to receivables: revenue rose 38% while trade receivables rose 96%, from ₹410 crore to ₹800 crore, and receivable days stretched from 74 to well over a hundred. So most of the profit that did not become cash is sitting in a receivable that grew almost three times as fast as sales - growth recognised but not collected. Third, book profit to tax: cash tax barely moved despite the 44% profit jump, another line that should have followed the profit and did not. illustrative

None of these three ties, alone, proves anything wrong. A large late-year order, a deliberate extension of credit terms to win share, a genuine tax timing difference - each could innocently explain one gap. But the forensic mindset reads the ties together, and three breaking in the same direction at once - profit outrunning cash, sales outrunning collection, profit outrunning tax - is a single coherent pattern: growth booked into the P&L faster than it became real. That is now the thing the company must explain, and the next four modules are the catalogue of explanations, honest and dishonest, that produce exactly this signature. The habit to build is to run the ties before reading the management's story, not after, so the story has to fit the reconciliation rather than the reconciliation being bent to fit the story.

When a company posts strong profit growth that has not turned into cash, the forensic analyst does not challenge the profit directly - he asks the reconciliation into the open: "Profit grew 44% but operating cash was about a third of it and receivables grew 96% - can you walk us through why cash and collections did not follow, and how much of the receivable is over ninety days?" The question does not accuse; it names the broken tie and asks the company to close it.

A good answer accepts the gap and decomposes it into checkable pieces:

"Fair question, and the gap is real this year. About ₹240 crore of the receivable is one large infrastructure client we moved to 120-day terms to win the contract; it's current and already partly collected this quarter. Another chunk is a late-March order we'd flag as genuinely one-off. Strip those two and receivable days are up only modestly, in line with the mix shift. Cash tax is low because of accelerated depreciation on the new plant, which reverses over the next three years. We'd expect operating cash to converge back toward profit next year." illustrative

It names the client, the ageing, the cause, and a future number you can hold it to. An evasive answer reassures without reconciling:

"We're very comfortable with the quality of our receivables and our collection track record has always been strong. Cash flows can be lumpy quarter to quarter and we manage for the full year. Our auditors are comfortable and our balance sheet has never been healthier. We remain confident in the demand environment."

"Comfortable with the quality" and "cash can be lumpy" are assertions, not the decomposition the question asked for; it names no client, no ageing, no cause, and reaches for the auditor and "momentum" in place of the numbers that should agree. The follow-up that decides it - "how much of the receivable is over ninety days and over a year, and what did you actually collect in cash against last year's sales?" - forces the ageing into the open, where a swelling, stale receivable cannot hide behind "we manage for the year." A company whose receivable would survive ageing usually volunteers the ageing; the silence is the tell.

Reconciliation finds the gap; it does not, by itself, name the cause. A profit that has outrun cash into a swelling receivable is consistent with aggressive revenue recognition, with channel stuffing, with a genuine shift to longer terms to win a large customer, and with a single big order booked in the last week of the year. The stance is precise about locating the question and silent about the answer - which comes from the notes, the concall, the segment detail and the history, the work of the modules that follow. Treating a broken tie as proof of fraud is the paranoid error, and it is wrong because it stops the investigation at the exact point it should begin. Nor can reconciliation certify that accounts which tie out are honest: some manipulations are built to keep the visible ties closed, borrowing to convert fake profit into real-looking cash or routing a fabricated sale through a related party that also pays. Reconciliation raises the cost of deception and catches the large majority of it - keeping every tie closed while inflating results is genuinely hard - but a clean set of ties lowers suspicion, it does not eliminate it. This is why the forensic reader adds ties rather than resting on a few: the more independent numbers a lie must keep consistent, the likelier one of them betrays it.

Against those limits, three ways readers slip. The first is reading numbers one at a time - a profit figure read alone is a claim, convincing because it is specific and audited and printed in bold, and the forensic error is never placing it beside the number that would have contradicted it. The defence is mechanical and unglamorous: never let a key number stand without the number it must agree with. The second is mistaking the stance for paranoia and then abandoning it - a reader who thinks forensic reading means assuming fraud tries it, finds most companies honest and most gaps innocent, decides the whole exercise is cynical noise, and stops. The forensic reader is not disappointed to find the ties close: that is the normal, valuable result, because it is the library of closed ties that lets him recognise the broken one. Dropping the stance because it usually finds nothing is like dropping a smoke alarm because it usually does not go off. The third is running the wrong tie for the sector - challenging an IT firm's inventory, which barely exists, while ignoring its unbilled revenue, or trusting a bank because its trade receivables look clean while its provision, the line that actually carries a lender's profit, goes unexamined. The remedy is to carry the stance, not the checklist: in every new business, ask which number the reported profit most depends on and least easily corroborates, and make that the tie you run first.