Why a bank is not read like a factory
Six axes. Four change their measure according to the economics of the business.
Most market scores apply one grid to everyone, then adjust by country or sector. The problem is not the sector: it is the DENOMINATOR. A ratio whose denominator does not describe the economics of the business cannot be adjusted — it has to be replaced.
Each substitution below was measured on the full universe before being adopted, and every figure quoted is the measured one — not a round number.
A bank is not scored on ROIC
Return on invested capital puts debt AND equity in the denominator. For an industrial company that debt funds the plant. For a bank, deposits and debt ARE the raw material: the ratio then measures the opposite of what it claims.
What is read insteadReturn on equity, on a scale calibrated on banks.
Measured on the full batch: median ROIC of 1.09% for banks, against 8.32% on the standard grid. Scored on ROIC a bank got 20.8 out of 100; scored on its own ROE, 59.7 — a 38.9-point gap on a signal worth 40% of the axis. The gap is 41.3 against 60.2 for insurers, 45.3 against 59.5 for other balance-sheet financials.
A REIT is read on FFO, not on accounting earnings
Property depreciation crushes a REIT’s accounting earnings — to the point that the payout ratio computed on them exceeded 100% for 95 of them. Those same earnings were already set aside elsewhere in the calculation; the dividend axis kept using them.
What is read insteadFunds from operations, on a sector-specific scale — a REIT’s tax regime REQUIRES it to distribute, so a high payout does not mean there what it means elsewhere.
Across the 277 dividend-paying REITs in the batch: median accounting payout of 0.789, scored 40.1 out of 100; on FFO, 0.711, scored 67.6. The count appearing above 100% falls from 95 to 61. Effect on the dividend axis: 54.1 → 69.7.
A regulated utility is measured on its rate base
The capex that crushes a regulated utility’s free cash flow IS the mechanism of its growth. Scoring it on that cash flow punishes it for what makes it grow — and for a quarter of them the figure was not even interpretable.
What is read insteadRate base growth — the asset on which the regulator allows a return.
Across the 109 regulated utilities in the batch: median annualised free cash flow growth of −0.53%, scored 51.1 out of 100; rate base growth, +6.16%, scored 59.8. And 27 of the 109 showed five-year free cash flow below −100%, hence meaningless. Effect: growth axis 52.8 → 57.1, coverage 0.563 → 0.650.
When the Score stays silent
An axis whose inputs are missing is not scored zero: it is not scored at all. An axis is published from 60% coverage, and a stock only enters a ranking from 70%. Below that the Score exists but does not compare — a score computed on half the data is not worth the same as one where everything is known, and putting them side by side in a sorted list would be the easiest lie to tell.
What the Score is not
- — It is not a forecast. No axis says what the price will do.
- — It is not advice. Polaris is not a registered investment adviser and recommends nothing.
- — It is not a verdict on a company. It is a reading of its financial statements, at one date, with the limits of what has been published.
Model generation: trois-modeles-v19. Every change of scale or weighting increments this label — two scores sharing it are comparable with one another.
The measures quoted are each explained in the glossary.