AIA Engineering Ltd.

Altman Z″ — distress model — Edward Altman, NYU, 1968; the Z″ variant was published in 1995 for emerging markets and non-manufacturers.

A single score built from four balance-sheet ratios that, together, separated companies that later went bankrupt from those that did not. Altman tested it on manufacturers; the Z″ version drops the sales-to-assets term, which made industrial firms look better than service businesses.

X1 · Working capital ÷ total assets

Short-term liquidity. Negative means current liabilities exceed current assets — the company owes more within a year than it holds.

X2 · Retained earnings ÷ total assets

Cumulative profitability. A young or serially loss-making company scores low here regardless of this year.

X3 · EBIT ÷ total assets

Operating productivity of the asset base, before financing and tax.

X4 · Net worth ÷ total liabilities

How far assets can fall before liabilities exceed them.

How to read itAbove 2.6 is the safe zone. Between 1.1 and 2.6 is grey. Below 1.1 is the distress zone. The score is a screen, not a prediction — it tells you which balance sheets deserve a second look.

Where it failsNot meaningful for banks, NBFCs or insurers, whose balance sheets are structurally different. Also unreliable for asset-light businesses, which carry few assets by design, and for holding companies whose value sits in unconsolidated stakes.

Piotroski F-Score — fundamental momentum — Joseph Piotroski, University of Chicago, 2000, in a study of whether accounting signals could improve returns among cheap stocks.

Nine yes-or-no tests across profitability, leverage and operating efficiency. Each pass scores one. It asks a narrow question: is this business getting better or worse on its own terms, year over year?

Profitability (4 tests)

Positive profit, positive operating cash, improving return on assets, and cash exceeding profit. The last is the quality test — profit that outruns cash is the one to question.

Leverage and liquidity (3 tests)

Falling debt, improving current ratio, no new shares issued. Growth funded by dilution scores zero here.

Operating efficiency (2 tests)

Improving margin and improving asset turnover.

How to read it7 or more suggests improving fundamentals; 3 or fewer suggests deterioration. It measures direction, not quality — a weak company improving can score higher than a strong one holding steady.

Where it failsA single year of comparison, so one unusual year distorts it. Says nothing about valuation, competitive position or management. Piotroski designed it to rank already-cheap stocks, not to judge a company in isolation.

Beneish M-Score — earnings manipulation — Messod Beneish, Indiana University, 1999. Best known for the Cornell students who flagged Enron with it a year before the collapse.

Eight ratios comparing this year with last, weighted into one score. It does not detect fraud. It detects the accounting patterns that tend to accompany managed earnings — receivables outrunning sales, margins falling while the business grows, assets shifting into categories that are harder to verify.

DSRI · Days Sales in Receivables Index

Receivables against sales, this year versus last. Above 1 means the company is collecting more slowly — sales may be being recognised before cash is likely.

GMI · Gross Margin Index

Last year's margin divided by this year's. Above 1 means margins deteriorated, which raises the incentive to manage the numbers.

AQI · Asset Quality Index

The share of assets that are neither current nor fixed — intangibles, deferred costs, "other". Above 1 means more of the balance sheet has moved into items whose value rests on judgement.

SGI · Sales Growth Index

Growth itself is not manipulation, but fast-growing companies face more pressure to sustain the trajectory.

DEPI · Depreciation Index

Above 1 means the depreciation rate slowed — assets are being written off more slowly, which flatters profit.

SGAI · Selling, General & Administrative Index

Overheads against sales. A disproportionate rise signals loss of control.

LVGI · Leverage Index

Rising leverage increases the pressure to meet covenants.

TATA · Total Accruals to Total Assets

The heaviest weight in the model. Profit not backed by cash, scaled by assets.

How to read itAbove −1.78 is the threshold at which the model classifies a company as a likely manipulator. That threshold produces false positives — fast growers and companies mid-acquisition often cross it innocently. Treat it as a prompt to read the notes, never as an accusation.

Where it failsNeeds two comparable years. Meaningless after a large acquisition, demerger or accounting-standard change, when the year-over-year ratios compare different businesses. Not applicable to banks or insurers.

DuPont decomposition — Devised inside the DuPont Corporation in the 1920s and still the standard way to read a return on equity.

Splits return on equity into its three sources, so the same headline number can be traced to very different businesses.

Net margin

What the company keeps from each rupee of sales. High margin points to pricing power or a genuine cost advantage.

Asset turnover

Sales generated per rupee of assets. High turnover points to efficiency — a retailer earns this way, a utility never will.

Leverage (equity multiplier)

Assets divided by equity. This multiplies whatever the first two produce, in both directions.

How to read itA 20% ROE built on margin and turnover is a different proposition from a 20% ROE built on 3× leverage. The first survives a downturn; the second amplifies it.

Where it failsA single year. Negative equity makes it meaningless. Leverage is structural for lenders, so the third term carries no signal there.

Montier C-Score — the cooking score — James Montier, then at Dresdner Kleinwort, 2008, as a deliberately simpler companion to Beneish.

Six binary flags, each a condition that appears when earnings are being flattered. Montier's argument was that you do not need a weighted regression — the conditions tend to cluster, and counting them is enough.

Profit and cash diverging

Net income growing faster than operating cash flow. The single most reliable warning in accounting.

Receivables growing faster than sales

Revenue recognised ahead of collection.

Inventory building against sales

Production outrunning demand, with the write-down still to come.

Other assets rising as a share of the balance sheet

Value migrating into items that are difficult to verify.

Depreciation slowing against fixed assets

Useful lives extended, which lifts reported profit without any operating change.

Assets growing unusually fast

Above roughly 25% a year, often through acquisition, which resets the comparative base and obscures the underlying trend.

How to read it4 or more flags warrants a careful read of the notes. 0 or 1 is unremarkable. Each flag is a question — several together are a pattern.

Where it failsA company mid-expansion trips several flags legitimately: inventory builds ahead of a launch, assets grow with a new plant. Read alongside what the business is actually doing.

Return on invested capital, and incremental ROIC — Standard in corporate finance; the incremental form was popularised by Michael Mauboussin as the test of whether growth creates value.

ROIC measures what the business earns on all the capital it employs — equity plus debt, less cash. Incremental ROIC asks a sharper question: what has it earned on the money put in since a chosen year?

NOPAT

Operating profit after a notional tax charge, so the figure is independent of how the company is financed. We use 25%.

Invested capital

Equity plus borrowings less cash — the money actually at work.

Incremental ROIC

Change in NOPAT divided by change in invested capital. If it sits below the cost of capital, growth is destroying value however fast revenue rises.

How to read itROIC comfortably above the cost of capital — call it 11–13% in India — means growth compounds. Below it, growth consumes. Incremental below headline means recent investment is earning less than the legacy business.

Where it failsDistorted in the year of a large acquisition. Understated for companies mid-build, where capital is deployed but capacity has not yet been commissioned — an incubator will look poor until it does not.

The earnings quality ladder — Not a named model — the standard sequence an analyst walks when testing whether reported profit is real.

Three ratios read in order, each stricter than the last.

Cash ÷ EBITDA

Does operating profit arrive as cash? Below 0.8 points to working capital absorbing it.

Cash ÷ profit

Does bottom-line profit arrive as cash? Below 1.0 persistently is the classic warning.

Free cash ÷ profit

Does anything survive capex? This is where capital-hungry businesses reveal themselves — a company can pass the first two and still never generate spendable cash.

How to read itRead downward. Each rung failing where the one above passed tells you exactly where the cash is going.

Where it failsA single year of heavy capex depresses the third rung legitimately. Judge it across a cycle.

Reverse DCF — the growth already in the price — A standard inversion of discounted cash flow, used to avoid the forecasting problem entirely.

Instead of forecasting cash flows and deriving a value, it takes today's market value as given and solves for the growth rate that would justify it. The output is not a view — it is the arithmetic of what the market is currently assuming.

Discount rate

The return required for the risk taken. We use 11.5%, roughly the long-run cost of equity in India.

Terminal growth

Growth beyond the explicit ten years. We use 4%, near long-run nominal GDP.

The output

The free-cash-flow growth rate, every year for a decade, that makes the discounted total equal today's market value.

How to read itCompare it with what the business has actually delivered. A price implying 30% a year against a decade of 15% is a demanding assumption; the reverse is a modest one.

Where it failsUseless when free cash flow is negative or unusually depressed, which is common mid-capex. Highly sensitive to the discount rate — a point either way moves the answer materially.