How to Screen Indian Stocks for Quality: A Four-Test Framework Worked on the Nifty 200

There are two ways to look for good companies in the Indian stock market. You can wait for a name to come up in the news or a WhatsApp group and then try to decide whether it is any good. Or you can decide in advance what a good business looks like, write that down as a few rules, and test every company against them.

The second approach is called a stock screen, and this post teaches you how to build one. We wrote four tests, a valuation check and a short list of red flags, then ran every company in the Nifty 200 (India's 200 largest listed companies, explained in our guide to the Nifty 50) through them using numbers collected from screener.in on 25 October 2026.

This is a lesson in method, not a list of stocks to buy. We use real companies throughout, but only to show you how each test behaves when a company passes or fails it. Several of the most instructive examples are companies that fail.

By the end you will be able to run the same tests yourself on any company, in about fifteen minutes, with a free screener.in account and the checker further down this page.

Why screen at all, and why now

The Nifty 200 is a good place to practise because it is big enough to show patterns and small enough that every name is a real, well-covered business.

It is also a useful moment. When we collected the data, 132 of the 200 companies (66%) were trading 20% or more below their 52-week high, and 60 were down 30% or more. The median company was 24% below its high.

When the whole market falls together, good businesses get marked down alongside weak ones. Prices move fast and the quality of a business moves slowly, so a broad decline is the time a screen earns its keep. The hard part is telling a company that is temporarily cheap from one that is cheap because something inside it is breaking. That is the problem a screen is designed to help with, and it only helps if the rules were written before you looked at any prices.

The four tests

Every non-financial company had to clear four tests, all using consolidated figures (the parent company plus its subsidiaries, which is the version that reflects the whole business) over the last five years.

Each test closes a gap that the others leave open. Here is why each one exists.

Quality, measured by ROCE

Return on capital employed tells you how much operating profit a company earns for every rupee tied up in the business, counting both shareholders' money and borrowed money. A company with a ROCE of 25% earns ₹25 of operating profit a year on every ₹100 invested in factories, stores, inventory and everything else.

We use the median of five years rather than the average. The median is the middle value when you line the five years up in order, so one freak year cannot drag it up or down. A company with ROCE of 22%, 18%, 12%, 24% and 26% has a median of 22%. The second half of the rule, no year below 10%, catches companies whose returns swing wildly with the economic cycle.

Why 15%? Roughly, it is a return comfortably above what the capital costs. A business that earns less than its cost of capital destroys value every time it grows. A business that earns well above it, year after year, almost always has some kind of economic moat protecting it.

Cash, measured by conversion

Profit is an accounting opinion. Cash is a fact. A company can book a sale, record the profit and never collect the money, and the profit and loss statement will look perfect until the day the receivable is written off.

The test adds up five years of cash from operating activities (the cash the business generated from its day-to-day operations) and divides it by five years of reported net profit. Below 80% means a fifth or more of the reported profit never arrived as cash over five whole years. Five years matters because any single year can be distorted by a large order shipped in March.

Balance sheet, measured by debt and interest cover

Two numbers. Debt to equity compares what the company has borrowed with what shareholders have put in and left in. Below 0.5 means borrowings are under half of shareholder funds. Interest coverage is operating profit divided by interest cost. A coverage of 5 means the company could see operating profit fall by 80% and still pay its lenders.

Companies with heavy debt can do very well in good years. The test is there for the bad year, which is exactly the kind of year a falling market might be pricing in.

Growth, measured by sales

The last test asks whether customers want more of what the company sells. Sales growth of 10% a year over either five or three years gives a company credit for a recent acceleration, while the "not shrinking now" rule screens out companies whose good years are behind them.

We use sales rather than profit because profit can be grown for a while by cutting costs. Sales cannot be faked for long.

Banks and lenders need their own rules

Apply the debt test to a bank and every bank fails, because borrowing (taking deposits) is the bank's raw material. Cash conversion is also meaningless for a lender, since lending money out shows up as negative operating cash flow when the business is growing. So the 48 financial companies in the Nifty 200 got a different set of four tests:

Return on equity (profit divided by shareholders' funds) replaces ROCE because for a lender, shareholders' money is the capital that absorbs losses. The stability test catches lenders whose returns collapsed in a bad credit year, which is the main way lenders destroy shareholder wealth. Our HDFC Bank analysis walks through the other numbers you need for banks, such as asset quality and the deposit base, and our post on the insurance business model explains why insurers need yet another lens.

Compare a stock with its own past, not with the market

After the four tests comes a valuation check, and this is the part most beginners get wrong.

The usual approach is to compare a stock's P/E ratioP/E Ratio (Price to Earnings)Tells you how many years of current earnings you're paying for when you buy a stock. A P/E of 25 means you pay 25 years of today's profits upfront. High P/E = market expects strong growth; low P/E = slow growth or high risk.See all terms in the glossary with the market's and call it cheap if the number is lower. Our guide to P/E ratios explains how that works. The problem in India is that the best businesses have traded above the market for decades. Wait for them to look cheap against the Nifty and you may wait for ten years.

So the screen compares each company with its own 10-year median valuation. For most companies the measure is EV/EBITDA: enterprise value (the market value of the shares plus debt minus cash) divided by EBITDAEBITDAEarnings Before Interest, Taxes, Depreciation and Amortisation. A measure of a company's core operating profitability, stripping out financing and accounting effects. Useful for comparing companies across industries.See all terms in the glossary (operating profit before depreciation and amortisation). It is better than P/E for comparing across years because it is not affected by changes in debt or tax rates. For lenders the measure is price to book valueBook ValueA company's assets minus its liabilities, as recorded on the balance sheet — its net worth on paper. For most of Berkshire's history, growth in book value per share was Buffett's headline measure of value creation.See all terms in the glossary.

The calculation is a single division:

Valuation ratio = today's EV/EBITDA ÷ the company's 10-year median EV/EBITDA

A ratio of 0.75 means the market is paying three-quarters of what it has usually paid for this business. A ratio of 2.0 means it is paying double. The screen treats anything above 1.15 as stretched.

Across the Nifty 200, 130 companies had a comparable ten-year valuation history. Of those, 79 (61%) were below their own median and 39 were more than 1.15 times above it. The median company sat at 0.91. Where the history was too short or measured on a different basis from today's number, we left the comparison blank rather than guess.

A low ratio is not a bargain signal. It means the market expects the future to look worse than the past. Sometimes the market is overreacting. Sometimes it has spotted something you have not. Either way, the ratio tells you what question to ask next, and our valuation guide covers how to think about the answer.

Red flags that override the numbers

Some risks do not show up in five years of ratios. The screen marks four hard red flags, and a company with any of them cannot be called a clean pass no matter how good its numbers are:

1. Government ownership. When the government is the majority owner, policy (dividends demanded to fill the budget, prices held down for consumers, a stake sold at a discount) can take a share of profits before minority shareholders see them. 37 of the 200 carried this flag.
2. Commodity or gold-price exposure. If profits follow a price the company does not set, five good years may just mean five years of a high price. 26 companies were flagged for commodities and 3 for gold.
3. Promoters pledging or selling. Promoters (the founding family or controlling group) know the business best. When they borrow against their shares or cut their stake, it is worth asking why. 18 companies showed promoter selling over the year and 4 had pledges noted.
4. Profit down 10% or more in a year. A sharp drop says the past five years may not describe the next one.

The screen also tracks soft flags that are shown but do not change the result. Two are worth knowing about because they are easy to miss:

• Capitalised interest (26 companies). When a company borrows to build an asset, accounting rules let it add the interest to the cost of the asset instead of charging it as an expense. Reported profit looks higher today, and the cost shows up later as depreciation.
• An unusually low tax rate (21 companies). It can be legitimate (a tax holiday, losses carried forward), but profits taxed at 10% today may be taxed at 25% tomorrow.

What the screen found across all 200 companies

Here is how the 200 companies did, counted from the screen results. A company can miss more than one test.

Across all 200 companies, 61 passed all four, 69 passed three, 43 passed two, 24 passed one and 3 passed none.

The headline is that quality is the hardest test. Even among India's 200 largest companies, 43% of non-financial businesses do not consistently earn a 15% return on the capital they use. Being big and well known says very little about being a good business. The second-biggest filter is debt: more than a third of the large non-financial companies borrow more than half of their equity or struggle to cover interest.

Almost no government-owned companies or commodity producers made it through cleanly, and that is by design. The framework is hunting for businesses that control their own economics.

Worked examples, one lesson per test

The numbers below come from the same screener.in snapshot. Each company is here because it teaches something about how a test behaves, not because of what it might do next.

Reliance and Nestle, the two ends of ROCE

Reliance Industries is India's most valuable company and generated operating cash flow of 190% of its reported profit, but its median ROCE was 10%, so it missed the quality test. Refining, petrochemicals, telecom and retail all need enormous amounts of capital, and the return on all of that capital has been modest. The cash test and the quality test are asking different questions, and here they give different answers.

At the other end, Nestle India posted a median ROCE of 138%. That does not mean it is ten times better than a company earning 14%. A business that has paid out most of its earnings as dividends for years carries very little capital on its balance sheet, so the ratio balloons. When you see a ROCE above 50%, check the capital base before you get excited. The return is real, but it tells you more about how little the business needs to reinvest than about how fast it can grow.

ITC, a miss by one percentage point

ITC had a median ROCE of 37% and almost no debt, yet it passed only two of the four tests. Over five years, operating cash covered 79% of reported profit, against a bar of 80%. Sales in the latest twelve months were also down 3%.

A mechanical screen cannot tell the difference between 79% because of a one-off and 79% because customers are not paying. A human can. ITC's figures include the effects of the hotels demerger, which most likely distorted both profit and cash for a period. This is why a screen is a starting point. The right response to a near miss is to open the cash flow statement and find the line that explains the gap. Our ITC analysis goes through the business segment by segment.

Titan, when the funding model fools the test

Titan Company grew sales 32% a year over five years and earned a median ROCE of 21%, but its operating cash covered just 43% of profit and its debt was 1.95 times equity. On the screen, it looks like a company in trouble.

Much of that comes from how jewellers fund themselves. A jewellery retailer holds a large stock of gold, and gold bought on credit or leased from banks shows up as borrowing, while the growing inventory soaks up operating cash flow. The screen also flagged capitalised interest and gold-price exposure. Whether you are comfortable with that model is a judgement call, but you can only make it once you understand why the numbers look the way they do. Our post on gold, the dollar and the rupee explains why the gold price ripples through Indian companies like this.

Hindustan Unilever and Britannia, quality without growth

Both are among the best-run consumer companies in India. Britannia's median ROCE was 49% and Hindustan Unilever's 27%, both with almost no debt and cash conversion above 100%. Both missed only on growth: Hindustan Unilever's sales grew 7% a year over five years and 2% over three, and Britannia's 8% and 6%.

Both also traded below their own history (0.69 and 0.77 times). The screen is pointing out a trade-off, not a flaw. With companies like these you are paying for steadiness, and the price has to reflect slower growth.

Hindustan Aeronautics, a good business at a stretched price

HAL passed on quality, cash and balance sheet, with a median ROCE of 32% and cash flow at 144% of profit. It missed on growth (8% a year) and carried the government-ownership flag. The more striking number was valuation: 2.27 times its own 10-year median, after the rally in defence stocks. The market is paying more than double what it usually has for this business, which means it expects the next decade to be far better than the last. It may be right, but at that price the framework has no margin of safety to work with.

Muthoot Finance and LIC, when a flag outweighs the numbers

Muthoot Finance passed all four lender tests, with five-year ROE of 23% and profit up 85% over twelve months. It still cannot be a clean pass, because a gold-loan business lends against gold jewellery, so the gold price shapes both how much it can lend and what happens when borrowers default. When we collected the data, gold was well below its January peak, and the stock was 43% below its own high.

LIC showed a five-year ROE of 52% and traded at 0.64 times its own history, but revenue grew 7% a year, it is government-owned, and the government cut its stake by 6.5 percentage points during the year. Three separate warnings sit behind a very attractive-looking ratio.

Run the tests yourself

Pick any company you are curious about, open its page on screener.in (switch to the consolidated view if there is one), and copy the numbers into the checker. Every field tells you where on the page to find it. The checker starts with an imaginary company that misses the cash test, so you can see what a miss looks like.

The two numbers people find hardest to locate are the five-year totals for the cash test and the ten-year valuation median. For the cash test, add up the last five values in the "Cash from Operating Activity" row of the cash flow table and the "Net Profit" row of the profit and loss table. For the valuation median, screener's chart section can plot EV/EBITDA over ten years; read the middle value off it, or write down the ten year-end values and take the middle one.

From shortlist to research, five questions to answer next

A screen gives you a short list. What you do with that list matters more than the list itself. For each company that comes through, work through five questions, in this order:

1. Why is the price where it is? If the stock is below its own history, write down in one sentence what the market is worried about. If you cannot, read the last two earnings call transcripts until you can. Across the screen, the reasons ranged from fears that AI will shrink pricing for IT services companies (TCS, Infosys and their peers traded at 0.61 to 0.66 times their own history) to leadership change and post-merger margin pressure at HDFC Bank (0.49 times its own price to book) and a large new competitor entering wires and cables.
2. Is the worry temporary or structural? A temporary problem is a bad quarter, a one-off cost or a cyclical slowdown. A structural one changes what the business can earn for the next ten years. The AI question for Indian IT companies is a live example of an argument that is not yet settled; our TCS analysis lays out both sides.
3. What would have to be true for the past five years to repeat? Write it down. "Sales keep growing at 12%" is a claim you can check against the order book, the industry's growth rate and the company's market share.
4. Who runs it, and what do they do with the cash? No formula captures promoter integrity, related-party dealings or a habit of expensive acquisitions. This is what the annual report is for. Our guide on how to read annual reports shows you where to look.
5. What is the moat? A company that passed the quality test has earned high returns for five years. Something is stopping competitors from taking those returns away. Name it, and ask whether it is getting stronger or weaker.

Only after those five do you have something like a view. Most companies that pass a screen will not survive all five questions, and that is the point. You want to spend your reading time on ten companies, not two hundred.

What a screen cannot tell you

Every screen has blind spots, and you should know ours.

It looks backward. Five years of good numbers say nothing certain about the next five. That is the entire debate around Indian IT services right now: the businesses pass today, and the question is whether the next five years will look like the last five.

It cannot judge people. Capital allocation, governance and how management treats minority shareholders all live in the annual report and the notes, not in ratios.

The data has gaps. Promoter pledging was picked up only from screener's own notes. Recent listings have short valuation histories. Some companies report on a different basis in different years.

The thresholds are arbitrary. Why 15% ROCE and not 14%? There is no magic in the number. ITC missed by a single percentage point on cash. Treat near misses as companies worth a look, and treat a narrow pass with the same suspicion.

It says nothing about timing. A good business at a fair price can keep falling for months if interest rates rise or foreign investors keep selling. People who invest in stages over several months do so because nobody can call the bottom.

Practice exercises

Exercise 1: one company, four tests. Choose a company you already know as a customer, such as a paint maker, a bank you use or a consumer brand in your kitchen. Run it through the checker above. Which test does it fail, if any? Find the line in the annual report that explains the result.

Exercise 2: a near miss. Find a company that passes three of the four tests. Write one paragraph on whether the missed test reflects something temporary or something permanent, using the company's last two annual reports.

Exercise 3: same sector, different results. Pick two companies in the same industry. Run both. If one passes and the other fails, work out what the passing company does differently, and connect it to the moat types in our economic moats guide.

Key takeaways

• A stock screen turns "is this a good business?" into a handful of rules you can apply to every company the same way, before you look at prices.
• The four tests (quality, cash, balance sheet, growth) work as a set, because each one blocks a way the others can be fooled. Lenders need their own version built on ROE.
• In the Nifty 200 snapshot, quality was the hardest test: 43% of large non-financial companies did not earn a median ROCE of 15%. Being big is not the same as being good.
• Compare a company's valuation with its own 10-year history, not with the market. A low ratio means the market expects a worse future, and your job is to find out why.
• Hard red flags (government ownership, commodity exposure, promoter selling, a sharp profit drop) can outweigh even excellent numbers.
• Near misses like ITC and odd results like Titan teach more than clean passes. Always find the line in the accounts that explains a result.
• A screen produces a research list, not a decision. The work that matters starts afterwards, with the annual report and the earnings calls.

This article explains a method for analysing companies and is for educational purposes only. It is not investment advice, a research report or a recommendation to buy, sell or hold any security, and the author is not registered with SEBI as an investment adviser or research analyst. Figures come from screener.in as collected on 25 October 2026 and may contain errors or have changed since. Investing in equities involves risk, including loss of capital. Do your own research or consult a SEBI-registered adviser before making any investment decision.