Business growth and shareholder result differ
Revenue can rise while earnings per share do not. Extra costs, low margins in a new activity or new share issuance can change the result. Check expansion through costs, investment and share count, not just the top line. Distinguish achieved growth from expectation: continuation requires customers, capacity and resources. A price may already assume success, so a good report can still be below expectations.
Valuation links value to future flows and uncertainty.[11](#source-valuation-teaching) The question is not only whether growth occurs, but what cash result it creates and what is paid for it. Good business and attractive price are separate claims.
What a low multiple can hide
A low P/E may look attractive while its denominator is temporarily high. For a cyclical business, test whether profit reflects unusually favourable conditions. In a fictional example, Company A is worth 200 million roubles and earns 10 million, so P/E is 20; B is worth 80 million and earns 8 million, so P/E is 10. The formula does not decide which is better.[15](#source-valuation-pe)
Productive expansion may explain A’s higher multiple; persistent loss-making investment weakens that explanation. B’s low valuation may reflect lost customers or obligations not captured by the ratio. Formulate testable hypotheses, then seek data.
Reinvestment must create a result
In an educational model, a company earns 10 million and invests 70%, or 7 million, in new projects. At an assumed 15% return on new capital, additional annual result is 1.05 million, or 10.5% of the original 10 million. This is not a forecast: delays, costs, accounting profit and cash flow differ. Ask what is actually reinvested, when it works, whether returns persist and whether debt or new shares are required.
More reinvestment is not automatically better. Weak projects can increase turnover without owner value, while a stable business need not grow rapidly. Analyse the mechanism, not one rewarded metric.
Check without choosing a favourite style
Ask both categories the same questions: repeatability of profit, what is in the price, how growth is financed and what would disprove expectations. Compare successful continuation, slower development and deterioration without inventing probabilities. Show dependence of value on conditions and reasons for peer differences.
From broad list to testable short list
The following is an original educational filter, not a reconstruction of Graham, Lynch or Fisher. Historical thresholds do not transfer automatically. A primary report can verify activities, risks and financial data, not fair price.[1](#source-rm-read-10k)
Define scope: non-financial firms with comparable reporting, positive repeatable operating results and a clear source of money. Exclude missing data, unclear perimeter, inconsistent share classes and unknown material obligations, recording each reason before looking at price. For growth firms test customer retention, unit economics, support costs, working capital and dilution. For cyclical firms restore normal utilization and margins. For apparently cheap firms test earnings quality, asset realizability and reason for the discount.
Fictional Firm R sells 100, earns 15 repeatable profit and forecasts 20% revenue growth. At unchanged margin profit becomes 18, but 30 of new investment and 10 of working capital require financing; profit growth of 3 alone does not prove cash attractiveness. Firm C reports 12 including one-off profit of 5; at business price 84, reported P/E is 7, but on repeatable 7 it is 12. All numbers are fictional.
Entry conditions, margin of safety and review
Instead of ordering a purchase, write confirmed economics, scenario range, price, liquidity risk, unresolved questions and the fact that would disprove the hypothesis. Research inclusion, valuation and trade decision are three stages. In a fictional range of 80–120 at price 70, discount to the lower edge is (80−70)/80=12.5%, versus 41.67% to the upper edge. If the lower edge is revised to 60, price 70 is already above it. Margin of safety depends on assumptions and is not insurance.
Set data and cost conditions before entry, and reject when a key obligation is unknown regardless of rating. After deciding, review the named metric and deadline, not just the quote. Active and passive processes differ by procedure, not promised result.
A screen threshold is a search rule, not a safety certificate
Minimum size, current ratio, dividend history and P/E×P/B can define a candidate screen, not universal safety thresholds. Size depends on currency and date; current assets may be illiquid inventory; past payments do not compel repetition. P/E changes meaning with one-off or negative profit, and P/B with asset-recognition history. Different industries have different obligations. Preserve formula, universe, report-availability date, exclusions and costs, then test another period and nearby thresholds. One historical screen does not establish causality or future return.
Growth plus dividend yield divided by P/E
A rough heuristic is H=(g+y)/(P/E), with g forecast annual EPS growth and y dividend yield. If g=8%, y=2%, P/E=10 and percentages are entered as 8 and 2, H=1; entered as 0.08 and 0.02, H=0.01. The economics are the same but the number differs by 100; a threshold without units is meaningless. H is not expected total return or a probability.
EPS growth is not share-price growth. In a fictional one-period example, EPS rises from 10 to 10.8, P/E falls from 10 to 8, price from 100 to 86.4, and dividend 2 gives (86.4+2−100)/100=−11.6%, despite positive H. Keep EPS, growth horizon and payout base in one model.
Company research is not a market-turn forecast
Revenue, margin, return on capital and dividend are different measures. Check each link separately: profit through costs, cash through investment and working capital, and per-share payment through rights and share count. No dividend today does not prove no future value; a generous payment does not prove durability. Entry price determines which future flows the owner pays for.
A value range can be stated without claiming the whole market will turn tomorrow. Check personal payment deadlines and financing for waiting. Registration country is not a certificate of independent risk: linked markets, revenue currency, settlement access and regime changes require a fresh exposure map. Country diversification and careful company selection answer different questions; neither guarantees no loss.