The Piotroski F-score: which financial-statement signals add up to the score

The F-score is a financial-statement checklist that awards a point for every condition met, and the total tells you not how much a stock is worth but whether the business behind it is improving. Joseph Piotroski put it together in a paper published in 2000: he took companies that were cheap on book-to-market and showed that, within this cheap group, the financial statements make it possible to separate businesses that are recovering from those that are deteriorating. The model is deliberately simple: no weights, no regression, just a set of binary yes/no tests, each of which relies on figures the company itself publishes.

What exactly is tested

The tests fall into groups by meaning, and the order within each group does not matter — every test carries the same weight.

Profitability and earnings quality:

• return on assets is positive in the reporting year;
• operating cash flow is positive;
• return on assets is higher than a year earlier;
• operating cash flow exceeds net income.

The last item is the most underrated. It compares cash with accounting profit and catches accruals: if earnings are rising but cash flow is not following them, the profit is being generated by revenue recognition, revaluations and the capitalisation of expenses rather than by sales. For the Russian market this is the key test, because revaluations of foreign-currency assets and liabilities can conjure up profit that never existed.

Financial strength and sources of funding:

• long-term debt as a share of assets has decreased;
• the current ratio has increased;
• the company has issued no new shares during the period.

The ban on share issuance looks odd until you recall the logic: a company in good shape funds itself from operating cash flow, while the one that goes to shareholders for money is the one whose cash flow falls short. A secondary offering here is not a judgement call but an observable fact, and it costs a point.

Operating efficiency:

• gross margin has risen year on year;
• asset turnover has risen year on year.

How to read the total

The scale works like this: its top end means every condition is met, its bottom end means none is. A meaningful conclusion is drawn not from the middle of the scale but from its extremes. Piotroski compared the performance of the top and bottom groups using data from 1976 to 1996 and found a persistent gap in returns between them. The middle of the scale is almost useless as a diagnostic: that is where most companies sit, and their score is made up of a random assortment of tests passed.

It matters what the score does not do. It does not tell you a stock is cheap — the valuation has to come from elsewhere, for example by comparing the result with 3,92 and the rest of the sector. It does not predict default: for the probability of bankruptcy you turn to a different construct, one that looks at balance-sheet structure and works the other way round — our analysis of the Altman model shows where the two tools diverge. Nor is it a return forecast for any single stock: the original result was obtained on a broad basket over a long horizon, not by picking among a handful of names.

Where the model breaks down on Russian stocks

Gross margin and asset turnover have no economic meaning for banks and insurers: they have no cost of sales in the usual sense, and their assets are a loan book. An automated calculation for such issuers will produce a score with nothing to interpret.

The second problem is comparability. The tests are built as comparisons with the previous period, so a change in accounting standards, in the scope of consolidation or the transfer of a business between parts of a group breaks the comparison without breaking the formula: the score will still be calculated, but it will measure the reorganisation rather than operating performance.

The third is data availability. The period when some Russian issuers cut back disclosure or published incomplete financial statements leaves gaps precisely where the model needs cash flow and the prior-year base. The site does not provide a ready-made aggregate F-score as a separate directive — no such metric exists in the database, so the score has to be assembled by hand from the published statements. The source material is in issuers' financial statements, and the page for a specific stock — for example, SBER — gives a starting point for checking the scope of consolidation and share issues.

A practical workflow

Step 1 — assemble a group of stocks that are cheap on book value from the list of stocks, without filtering it for quality. Step 2 — drop the financial sector and every company whose year is not comparable with the previous one. Step 3 — run the tests on annual data, not quarterly: quarterly seasonality distorts margins and turnover. Step 4 — work only with the extremes of the scale and hold a basket rather than a single name.

It is also worth checking the dividend history separately: a company that is borrowing and paying out at the same time easily scores on liquidity and loses the point on debt — the payout calendar helps you see what is actually sustaining the cash flow. If the definitions of individual ratios differ from yours, check them against the glossary.