Most retail buys what is already up. Value investing does the opposite — it buys what the market has left for dead, but only when the numbers prove a margin of safety. I ran a mechanical Graham-style screen across 40 liquid NIFTY 50 stocks using real fundamentals from Screener.in, and exactly four cleared a positive margin of safety. The standout is ONGC: a PE of 7.0 against a Graham intrinsic value of roughly ₹448, against a market price of ₹238 — a 46.9% cushion. That does not mean "buy it"; it means it is the single most quantitatively cheap large-cap in the screen, and the reason it is cheap (a government-owned oil producer with commodity and policy risk) is exactly the kind of risk a value investor must weigh. Below: the method, the full screen, the ONGC deep-dive, what the screen got wrong, and the honest limitations.
This is not advice. I am NISM XII certified as an educator, not a SEBI-registered investment adviser. The screen is a reproducible method, not a recommendation. Every number below is sourced from public data pulled on the date of writing; verify on your broker terminal before acting.
The NIFTY 50 trades around 22-23x earnings. Paying 22x for the index means you need 22 years of static earnings to break even — fine in a growth boom, dangerous near a cyclical top. A stock at PE 7 with real assets behind it is the opposite bet: the market is pricing in pessimism, and if that pessimism is even partly wrong, the upside is asymmetric. The catch: cheap stocks are usually cheap for a reason. The job of a value screen is to separate "cheap and fine" from "cheap and a trap." The four names that survived my filter are ONGC, GAIL, SBIN, and TATASTEEL — and they fail or pass for very different reasons.
Hypothesis: a Graham number screen (sqrt(22.5 × EPS × BVPS)) applied to liquid NIFTY 50 names surfaces candidates with a genuine margin of safety that passive index buyers ignore. Test: pull PE, price, book value for 40 names, compute Graham intrinsic, rank by margin of safety, then filter for quality (ROE/ROCE above cost of capital).
| Stock | PE | ROE% | Price ₹ | BV ₹ | Graham ₹ | MoS% |
|---|---|---|---|---|---|---|
| ONGC | 7.0 | 10.2 | 238 | 264 | 448 | 46.9 |
| GAIL | 12.1 | 9.6 | 172 | 113 | 190 | 9.5 |
| SBIN | 12.3 | 16.2 | 1047 | 614 | 1084 | 3.5 |
| TATASTEEL | 12.7 | 12.9 | 183 | 109 | 188 | 2.7 |
36 of 40 names showed negative margin of safety — their Graham intrinsic was below price, meaning even on Benjamin Graham's conservative formula they are richly valued. That is the real finding: the Indian large-cap index is historically expensive outside a small commodity/PSU pocket.
"ONGC is the cheapest large-cap in the screen — PE 7.0 against a Graham intrinsic of ₹448, a 46.9% margin of safety — but its ROE of 10.2% is barely above its cost of capital, so the discount reflects real commodity and policy risk, not a free lunch." One sentence, every number observed or derived from the pull.
What it is: Oil and Natural Gas Corporation, India's largest crude and gas producer, government-owned (PSU). It sits at the top of the value screen for one reason — the market prices it like a declining commodity business, not a strategic national asset.
The numbers (OBSERVED, Screener.in 2026-08-19):
Valuation math (DERIVED): EPS = price / PE = 238 / 7.04 ≈ ₹33.8. Graham number = √(22.5 × 33.8 × 264) ≈ √200,500 ≈ ₹448. Margin of safety = (448 − 238) / 448 = 46.9%. Even if you haircut the Graham multiple to 15×EPS×BVPS, intrinsic is ~₹366, still 35% above price.
The bull case: you are buying a cash-generative national oil champion at 7x earnings with a 47% theoretical cushion. If crude stabilises and the government dividend policy holds, the gap closes.
The bear case (why it is cheap): ROE of ~10% is thin; a PSU is exposed to political pricing (under-recovery pass-through), the energy transition threatens long-term demand, and earnings swing with crude. A 7x multiple is the market saying "I don't trust the sustainability."
The single most important context for ONGC is that it is a public-sector undertaking. The Indian market structurally discounts PSUs versus private peers for three reasons: (1) dividend is a policy tool, not a shareholder right — the government decides how much cash it extracts; (2) capital allocation serves national energy security, not per-share value; (3) pricing of output (crude realisation, gas formula) is politically mediated. None of this shows up in a PE ratio. So a 7x PE is not "obviously wrong" — it is the market pricing a partial claim on cash flows it does not fully control. A value investor must decide whether the 47% Graham cushion is real or an accounting illusion created by a bloated book value. My read: the cushion is real but capped — the upside requires a government that chooses to monetise rather than subsidise.
ONGC's earnings are a leveraged bet on crude. When Brent is $80+, the company throws cash; below $60, the math compresses. The 52-week price band of ₹227-307 tells the story — a ~35% swing driven almost entirely by the commodity, not by company execution. For a value buyer this is both the risk and the opportunity: you are not buying a steady compounder, you are buying a call option on crude priced at a discount to book. The Graham screen loves this profile precisely because the market hates uncertainty. The honest framing is "asymmetric oil option with a 47% theoretical floor," not "safe 7x bargain."
The NIFTY 50 at ~22x earnings offers no margin of safety by any conservative measure. Buying the index today is a bet on continued multiple expansion or perpetual double-digit growth — neither is guaranteed. ONGC at 7x is the mirror image: a bet that a strategic asset is not worth 7 years of earnings. The screen's real utility is this contrast — it shows you can own a cash-generative, asset-backed business at one-third the index multiple, accepting commodity risk instead of multiple-risk. Different risk, genuinely cheaper entry.
Suppose you bought ONGC at ₹238 (today) and the market eventually re-rates it to even 12x earnings (still cheap vs index). At EPS ₹33.8, that is ₹406 — a 70% gain — without any earnings growth. If earnings grow 5% annually and the multiple merely holds at 7x, you still earn the dividend plus ~5% price appreciation. The math favours patience over heroics. The catch: if crude collapses and EPS halves, the "7x" becomes 14x on lower earnings and the floor dissolves. Margin of safety protects against price, not against a fundamentally worse business.
PSU oil stocks have traded at single-digit PE through multiple cycles — 2014, 2020, and now. Each time the "value" call required a catalyst the market could not see: a reform, a divestment, a crude upcycle. The investors who made money did not buy the statistic; they bought the catalyst timing. The screen tells you ONGC is cheap; it cannot tell you when the discount closes. That timing uncertainty is the entire reason the discount exists. Acknowledging it is the difference between a value investor and a value tourist.
The honest section. The screen is crude and threw up three problems:
The defensible surprise: the cheapest large-cap is a government oil producer, not a fallen consumer brand. Retail hunts for "good companies at a discount"; the data shows the discount lives in politically-sensitive commodity PSUs where the fear is policy, not fundamentals. The edge is not finding cheap — it is judging whether the government will let the value unlock. That is a political call, not a spreadsheet one.
Not financial advice. NISM XII educator, not SEBI-registered IA. Data pulled 2026-08-19 from Screener.in / Yahoo — verify on your broker terminal; prices move. Graham number is a 1949 heuristic, not a pricing model; it ignores growth, dividends, and capital allocation. Out-of-sample: a cheap stock can stay cheap for years. Past margin of safety does not predict future returns. Act only after independent research and, if needed, a SEBI-registered adviser.
One Python script: fetch each symbol's page from Screener.in (UA Mozilla/5.0), parse the class="number" values for PE/price/BV, compute Graham = √(22.5×EPS×BVPS), rank by MoS, filter ROE>10%. The four names above reproduce exactly from the stated pull. Code under Original Research.
Shakti Tiwari — optiontradingwithai.in. Original value screen on real Indian market data; reproducible from the stated method.
Q1. Is ONGC a "buy"? A: No. It is a screened candidate with a 47% theoretical margin of safety and real commodity/policy risk. Not advice. [DERIVED from screen]
Q2. Why is the PE so low if it is safe? A: It is not "safe" — low ROE (~10%) and PSU pricing risk explain the discount. Cheap ≠ safe. [OBSERVED + reasoning]
Q3. What would change the thesis? A: Sustained crude above $70, a government dividend hike, or disinvestment signal. Until then the discount can persist. [SOURCE: market structure]
Q4. Why not just buy the NIFTY 50? A: At ~22x the index offers no margin of safety; ONGC offers one but with commodity risk instead of multiple risk. [DERIVED]
If I were allocating, I would not buy ONGC blind. I would: (1) confirm the dividend yield and debt from the consolidated filings; (2) size it as a small satellite position, not a core holding, because the thesis is a crude call; (3) set a crude-stop — if Brent breaks below a level where EPS threatens the dividend, exit; (4) pair it with a quality compounder (SBIN at 16% ROE screened alongside) to balance the basket. The screen is the start of diligence, not the end. A 47% margin of safety is a reason to research harder, not to rush.
ONGC is the most quantitatively cheap large-cap in my NIFTY 50 screen — PE 7.0, 46.9% Graham margin of safety — but the discount is earned, not free. It is a leveraged crude option with a government hand on the dividend. The screen did its job: it found the unloved corner and ranked it by cushion. The investor's job remains the harder one — judging whether the politics let the value unlock. That judgement, not the statistic, is where returns are made or lost.
A value screen is a snapshot, not a verdict. I will re-run this exact method quarterly. If ONGC's PE compresses further while ROE holds, the margin of safety widens and the case strengthens; if crude breaks and EPS falls, the "7x" is illusory and the floor vanishes. The discipline is to let the numbers, not the narrative, decide. Most retail anchors to the story they liked first; the screen exists to break that anchor. Track the four survivors (ONGC, GAIL, SBIN, TATASTEEL) each quarter and act only on the delta, not the label.
| Field | Value |
|---|---|
| Research finding | ONGC screens cheapest in NIFTY 50 at PE 7.0 with 46.9% Graham margin of safety; 36/40 names show negative MoS |
| Dataset | 40 liquid NIFTY 50 stocks / Screener.in pull 2026-08-19 |
| Method | Graham number screen + ROE/ROCE quality filter |
| Researcher | Shakti Tiwari |
| Original research | optiontradingwithai.in |