Why buying every cheap PSU on MTF is a trap: the factor-investing verdict

The pitch is seductive because it sounds like arithmetic. Public-sector undertakings — the government-owned oil marketers, banks, miners, power and coal companies — screen absurdly cheap. Price-to-earnings under 6. Price-to-book under 1. Many pay fat dividends. So the plan writes itself: buy every PSU trading at PE < 6 and PB < 1, do it on Margin Trading Facility (MTF) so your capital works twice as hard, and just hold. Cheap stocks bought with borrowed money, held through the cycle. What could go wrong?

Almost everything, and not for the reasons most people fear. This is not a "be careful, markets are risky" hand-wave. There is a specific, well-documented body of factor-investing evidence that speaks directly to this exact bet — a leveraged basket of the cheapest available stocks — and its verdict is close to brutal. The single most famous case of a value investor successfully using leverage is Warren Buffett, and when academics reverse-engineered how he did it, they found he did nearly the opposite of what the cheap-PSU-on-MTF plan does. The plan keeps the one leg of Buffett's formula that is easy to copy and discards the three that actually made it work.

This post answers one question honestly: why shouldn't I just buy every PE < 6 / PB < 1 PSU stock available on MTF and hold? The answer is not "cheap stocks are bad" — deep value is a real, studied premium. The answer is that cheapness alone, leveraged, in one correlated corner of the market is a near-perfect recipe for the tail risk that leverage punishes hardest. Let's walk the evidence.

Why the instinct feels right — and why that's the trap

The plan feels right because every individual step is defensible in isolation:

Value works. Buying cheap earnings and cheap assets is one of the best-documented return premia in finance. Over long horizons, cheap has beaten expensive.
Leverage amplifies returns. If value earns a premium, borrowing to hold more value should earn more of that premium. That's just multiplication.
PSUs are genuinely cheap. The low multiples are real. The dividends are real. You are not imagining the discount.

Each premise is true. The conclusion is still wrong, because the plan stacks them in a way that quietly assumes three things that are false: that the cheap stocks are safe, that they are high quality, and that your borrowing is patient. Strip those assumptions out and the multiplication runs in reverse.

The instinct to "just buy the cheapest thing and lever it" is itself a known failure mode in the literature — it is what happens when you take a factor premium and apply it naively, without the companion factors that make it survivable. The academics have a name for the companion factors: quality and safety. Leave them out and you are not running the value strategy that won; you are running the caricature of it that loses.

The Buffett's Alpha inversion: the one case that proves the point

In 2013, Andrea Frazzini, David Kabiller, and Lasse Pedersen of AQR published Buffett's Alpha (SSRN 2312432; later in the Financial Analysts Journal, 2018). They asked a precise question: Buffett's returns are extraordinary — is it skill, luck, or a replicable formula? Their answer is the most useful thing ever written for anyone thinking about leveraging value.

They found Buffett's edge decomposes into a handful of factors, and it worked because of four legs working together:

Cheap. He bought at attractive valuations — the value leg.
Safe. He bought low-beta, low-volatility stocks — companies that don't swing violently with the market.
High quality. He bought profitable, stable, growing, well-run businesses — the QMJ ("Quality Minus Junk") leg.
Patient, cheap, non-callable funding. He leveraged roughly 1.6-to-1 — modest — and, critically, financed it with insurance float: money that cost him below the T-bill rate (around 1.7%) and, decisively, could not be called away. No margin clerk could ever force him to sell at the bottom.

Now line up the cheap-PSU-on-MTF bet against those four legs. It keeps leg 1 and violates the other three — one by one, in the worst possible direction.

Leg of the formula	What Buffett actually did (the alpha)	What "cheap PSUs on MTF" does
Cheap	Bought cheap valuations	Buys cheap valuations — the only leg it keeps
Safe (low beta)	Bought low-beta, low-volatility stocks	Buys high-beta cyclicals (OMCs, PSU banks, metals, power) — the opposite of safe
High quality	Bought profitable, growing, well-governed firms	Buys PSU value-traps — political interference, capital misallocation, forced dividends/OFS — the opposite of quality
Patient, cheap, non-callable funding	~1.6x leverage on ~1.7%, non-callable insurance float	Callable ~12-15% margin debt — expensive, and a margin call can force-sell you at the bottom

Read that table twice. The cheap-PSU-on-MTF plan is not a scaled-down Buffett. It is close to the mathematical inverse of Buffett on three of the four legs that generated the alpha. It keeps the one leg (cheap) that is necessary but nowhere near sufficient, and inverts safety, quality, and funding — the three legs that actually did the work.

And here is the number that should end the fantasy of "extraordinary leverage-worthy returns": even with all four legs firing, over decades, Buffett's lifetime Sharpe ratio was about 0.76. That is excellent — the best long-run track record on record — but it is not magic. If the greatest value-plus-leverage investor in history, doing everything right, earned a Sharpe near 0.76, then the idea that a naive cheapest-basket on callable leverage will produce extraordinary risk-adjusted returns has no support in the evidence. You would be attempting something harder than what Buffett did, with worse tools, and expecting a better result.

Cheapness alone is incomplete: what the factor literature actually says

The deeper problem is that "buy the cheapest stocks" is a misreading of the value research. The literature is emphatic that cheapness is a component, not the whole engine, and that leveraging it naked is exactly the mistake.

Betting Against Beta (Frazzini & Pedersen, 2014). This is the paper that matters most for the leverage part of the plan. Its central finding: high-beta stocks deliver poor risk-adjusted returns, and low-beta stocks deliver better ones. The profitable strategy is to lever up low-beta assets and short high-beta ones. The cheap-PSU-on-MTF bet does the reverse — it levers up high-beta cyclicals. That is not a neutral choice; it is documented, by name, as the opposite of what works. You are buying the leg of the market that pays you least per unit of risk, and then adding risk to it.

Quality Minus Junk (Asness, Frazzini, Pedersen). The QMJ factor shows that, holding price constant, high-quality firms — profitable, growing, safe, well-run — systematically outperform junky ones. Investors underpay for quality. The implication for a value screener is direct: among cheap stocks, the quality ones are the mispricings and the junky ones are the traps. A raw PE < 6 / PB < 1 screen makes no distinction. It buys the quality and the junk indiscriminately — and PSU cyclicals are disproportionately on the junk side of that line.

The Devil in HML's Details (Asness & Frazzini, JPM 2013). This one is subtle but load-bearing. It shows the value factor itself is measured and used far better when combined with other signals — that value works best jointly with momentum and with fresh price information, not as a static "cheapest wins" sort. The takeaway: even the value premium you are trying to harvest is not captured by a naive low-multiple screen. The premium lives in a disciplined, multi-factor construction, not in "buy everything under PE 6."

Put together, the message from the people who literally built the value and leverage factors is: the load-bearing factors are quality and safety, not cheapness alone — and value is best used jointly with other signals. A naive low-PE/PB PSU basket omits every filter the research says is essential and keeps only the one that isn't sufficient.

The Piotroski quality guard — built for exactly this problem

If you want proof that quality-guarding a cheap screen is not my opinion but the established fix, look at Joseph Piotroski's 2000 paper in the Journal of Accounting Research. Piotroski studied a universe of cheap (high book-to-market) stocks — the exact universe the PSU plan targets — and asked: can simple accounting signals separate the winners from the value-traps within the cheap group? His answer was the 9-point F-score, and it works. Applying it to a cheap universe meaningfully raised returns and shifted the whole distribution up, precisely by weeding out the deteriorating names.

What makes the F-score damning for the raw PSU plan is which nine things it checks. Its criteria target, almost line for line, the specific ways PSUs disappoint shareholders:

Decreasing long-term leverage (debt). Rewards firms cutting debt. Many PSUs — power, infrastructure, some banks — carry heavy or rising leverage.
No new share issuance. Penalises dilution. PSUs are serially diluted through Offer-for-Sale (OFS) as the government sells down its stake to raise money — the F-score flags exactly this.
Improving return on assets (ROA). Rewards rising profitability on the asset base. PSU capital misallocation — projects chosen for political rather than return reasons — shows up as falling or stagnant ROA.
Improving gross margin and asset turnover. Rewards operating efficiency gains. Overstaffed, mandate-driven PSUs frequently show the opposite.

That is not a coincidence. The F-score's failure modes are the PSU failure modes. This is the mechanistic reason a quality-guarded cheap screen works and a raw cheapest-first screen doesn't: the guard is specifically constructed to detect the deterioration that makes a cheap stock a trap, and PSUs are unusually prone to that deterioration. Screen PE < 6 / PB < 1 with no F-score filter and you are buying the exact stocks Piotroski's method exists to help you avoid.

The structural PSU discount: why cheap ≠ mispriced

Here is the argument that most retail buyers never confront. The plan assumes the low multiples are a mispricing — that the market is wrong and will eventually re-rate the stocks up. But there is a strong, reasoned case that the PSU discount is structural — that the market is pricing a set of permanent disadvantages correctly, and the discount will not close because the reasons for it do not go away.

Consider what a PSU actually is, from a minority shareholder's seat:

Government majority ownership and political interference. The controlling shareholder is the state, whose objectives — employment, inflation control, strategic priorities, votes — are not "maximise the minority shareholder's return." Decisions can and do run against your interest.
Capital misallocation. Investment is often directed by policy rather than by return on capital. Loss-making mandates, non-commercial projects, and cross-subsidies are features, not bugs.
Forced dividends and OFS to plug the fiscal deficit. When the government needs money, PSUs are a source — via special dividends that drain cash the business might have reinvested, and via OFS that dilutes or overhangs the stock. The controlling shareholder can extract value from the company to fund the exchequer.
Low free float and disinvestment overhang. A small tradable float and the perpetual possibility of the government selling more shares caps the upside — every rally invites another OFS.
Energy-transition and ESG risk. The oil, coal, and thermal-power PSUs that dominate the cheap screen sit directly in the path of the energy transition. Some of their "cheapness" is the market discounting assets that may be structurally challenged over the coming decades.

None of these are temporary. A rational market that knows PSUs will be periodically milked for dividends, diluted through OFS, run partly for non-shareholder objectives, and exposed to transition risk should assign them a lower multiple than an otherwise-identical private firm. In that world, a low PE/PB is not an opportunity — it is the fair price of a permanently disadvantaged claim on earnings. "Cheap" and "mispriced" are different words, and the whole leveraged bet rests on assuming they are the same.

That does not mean every PSU is uninvestable — some are well-run and genuinely cheap. It means you cannot infer mispricing from the multiple alone, which is exactly what the buy-everything-under-PE-6 plan does.

Concentration and correlation: the basket is one bet, not many

Buyers imagine a PE < 6 / PB < 1 PSU basket as diversification — twenty stocks, surely that's spread out. It is not. Run that screen and watch what it returns: it clusters, hard, into a handful of correlated cyclical sectors:

Oil marketing companies (the BPCL / IOCL / HPCL cluster)
PSU banks
Metals and mining
Power and coal

These are not independent bets. They are macro-linked and they move together. A spike in crude hammers the OMCs' marketing margins at the same time. A rate shock or a credit-cycle turn hits the PSU banks together. A global growth scare knocks metals, mining, and coal in unison. The screen feels like twenty positions; economically it is closer to two or three correlated macro bets wearing twenty tickers.

That is fatal for the diversification story. Diversification only reduces risk when holdings are imperfectly correlated. A basket that all rises and falls on the same two or three macro variables gives you the illusion of spread with the reality of concentration. And concentration is precisely the thing you cannot afford when you have added leverage on top — because when the correlated basket falls, it all falls at once, and it takes your margin with it.

Leverage amplifies the tail: the margin-call math

This is where the three inversions compound into a single catastrophic mechanism. High-beta + low-quality + callable-funding + concentration don't just add up — they multiply, and the product is forced liquidation at the bottom.

Walk the sequence. You hold a correlated cyclical PSU basket on ~2x MTF. A macro shock hits — crude spikes, or a rate shock, or a global growth scare — and because the basket is correlated and high-beta, it doesn't drop 10% politely over a year. It gaps down 30-40% together, fast.

Now the leverage turns on you:

Your equity in the position collapses (leverage multiplies the drawdown on your own capital).
You breach the maintenance margin. The broker issues a margin call.
If you can't top up in cash, the broker squares off — sells your positions at the market's price, at the bottom, at the worst possible moment.
What was a paper loss you might have ridden out becomes a realised wipeout. The forced sale converts a temporary drawdown into a permanent loss.
And through the entire drawdown, ~12-15% annual interest has been accruing daily on the borrowed half — you paid to hold the thing all the way down.

Basket move (correlated, fast)	On cash (unlevered)	On ~2x MTF, after carry
-20%	-20%, you wait it out	-40%+ on your capital; margin under stress
-35%	-35%, still your choice to hold	Margin call → forced square-off at the bottom; loss realised
Then recovery	You're still holding, you recover	You're out. You ate the drawdown, missed the recovery, and paid interest through both

The cruelty is in the last row. In cash, a cyclical crash is survivable — you hold, the cycle turns, you recover. On leverage in a correlated basket, the crash removes you from the position before the recovery, at the price you least wanted, while charging you interest for the privilege. Leverage doesn't just amplify volatility; against a callable lender it amplifies the tail — and the tail is exactly where correlated high-beta cyclicals live. This is, plainly, the single most dangerous combination available to a retail investor: leverage plus high-beta cyclicals funded by callable debt. It is the precise inverse of Buffett's non-callable, low-beta, modest-leverage structure — and it fails in precisely the way his was engineered to never fail.

What would actually make it defensible

None of this means deep value is worthless or that PSUs are untouchable. It means the naive leveraged cheapest-basket is the reckless version. There is a disciplined version, and it is defined entirely by putting back the legs the naive plan strips out. Here is the honest line between the two:

Dimension	Reckless version (the trap)	Defensible version (disciplined deep value)
Leverage	~2x callable MTF at 12-15%	Unleveraged, or low leverage you can service and survive
Selection	Every stock under PE 6 / PB 1	Quality-guarded: ROE, low promoter pledge, low debt, Piotroski ≥ 6
Diversification	All PSU cyclicals (OMCs, banks, metals, power)	Spread across uncorrelated sectors, not one macro bet
Position sizing	Sized to maximise exposure	Sized to survive a 40% drawdown with no margin call
Holding	"Hold" — but forced out on a margin call	Genuinely held through the cycle, on your terms
Funding character	Callable, expensive	As close to patient and non-callable as you can get (ideally cash)

The disciplined version is recognisably the same factor bet the research supports: cheap and quality and safe, diversified, held patiently, with leverage that can't force-sell you. That's Buffett's four legs, restored. The reckless version keeps one leg and inverts three. The gap between them is the entire difference between harvesting the value premium and being destroyed by the tail.

If you want a single filter that captures the spirit of the fix: the quality guard. A cheap screen without ROE, debt, pledge, and Piotroski checks is not a value strategy — it's a trap-collection machine.

The verdict

So, why shouldn't you just buy every PE < 6 / PB < 1 PSU on MTF and hold?

Because the one documented case of value-plus-leverage that worked — Buffett — won on four legs, and your plan inverts three of them. Because the factor literature says the load-bearing factors are quality and safety, not cheapness alone, and your screen omits both. Because Piotroski's F-score exists specifically to separate value from value-traps within the cheap universe, and your screen skips it — while PSUs are unusually prone to the exact deteriorations it detects. Because the PSU discount is plausibly structural, not a mispricing, so "cheap" may be the correct price of a permanently disadvantaged claim. Because the basket is not diversified — it's two or three correlated macro bets in disguise. And because leverage against a callable lender turns the correlated tail into a forced square-off at the bottom, converting a survivable drawdown into a permanent wipeout while charging you 12-15% the whole way down.

The plan keeps the seductive, easy leg — cheap — and discards the three legs that require discipline and cost you upside. That is not a scaled-down version of what works. It is close to its inverse.

The defensible move is the boring one: quality-guard the screen, diversify across uncorrelated sectors, size to survive, and either drop the leverage or keep it low and non-forced. Do that, and you are running deep value. Skip it and lever the cheapest PSUs, and you are running the caricature the research was written to warn you about.

How this maps to a screener

The practical fix — the quality guard — is exactly what a good deep-value preset builds in. On nifty.oriz.in, the "Deep Value" preset pairs the deep-value valuation cut-offs (in the region of PE < 6 / PB < 1.5) with the quality guards this post insists on: ROE, a Piotroski F-score gate, and a debt-to-equity guard, rolled into a composite score. That is the difference between "everything under PE 6" and "the cheap names that survive the quality check." It surfaces a ranked list you can then diversify across sectors yourself — turning a raw cheapest-first sort into the disciplined version above. The screener can enforce the guard; it can't enforce your sector discipline or your leverage restraint. Those are on you.

Honest limitations

Two caveats, stated plainly, because the whole point of this post is to be honest rather than to sell you a story.

First, the evidence here is factor theory and mechanism, not an India-PSU backtest. Everything cited above — Buffett's Alpha, Betting Against Beta, Quality Minus Junk, The Devil in HML's Details, Piotroski — is well-documented, peer-reviewed factor-investing research, and the leverage/quality/safety mechanism it describes is general and transferable. But this analysis did not independently verify India-specific PSU index return series. If you have heard about a PSU "lost decade" through roughly 2010-2020, or the sharp PSU rally of 2023-24 and its subsequent correction, treat those as widely-reported market context to verify yourself, not as cited datapoints from this post. I am deliberately not quoting specific Nifty PSE, Nifty CPSE, or BSE PSU index return figures, because I have not verified them here. Pull the actual index series before you rely on any historical PSU return number — including any you may remember.

Second, this is general information, not a recommendation about any stock or strategy. The factor evidence is robust; its application to your situation, your risk tolerance, and today's specific PSU valuations is not something a blog post can decide for you.

General education, not investment advice. Leverage can cause losses exceeding your margin; MTF interest accrues daily and a margin call can force a sale at the worst possible time. Factor premia are long-horizon and can underperform for years. The India-specific PSU return history referenced as context is not independently verified here — confirm it yourself. Consult a SEBI-registered investment adviser and a tax professional before acting.

Sources

Frazzini, Kabiller & Pedersen — "Buffett's Alpha" (SSRN 2312432; Financial Analysts Journal, 2018). The four-leg decomposition: cheap + safe (low-beta) + high-quality, ~1.6x leverage on non-callable insurance float at ~1.7%; lifetime Sharpe ≈ 0.76.
Frazzini & Pedersen — "Betting Against Beta" (2014). High-beta stocks deliver poor risk-adjusted returns; the edge is levering low-beta, not high-beta.
Asness, Frazzini & Pedersen — "Quality Minus Junk" (QMJ). Investors underpay for quality; among cheap stocks, quality separates mispricings from traps.
Asness & Frazzini — "The Devil in HML's Details" (Journal of Portfolio Management, 2013). The value factor is captured by disciplined, multi-signal construction, best used jointly with momentum — not a naive cheapest sort.
Piotroski — "Value Investing: The Use of Historical Financial Statement Information to Separate Winners from Losers" (Journal of Accounting Research, 2000). The 9-point F-score, built to separate strong value from value-traps within the cheap universe; criteria target leverage, share issuance, and ROA/margin/turnover trends.
nifty.oriz.in — "Deep Value" preset with ROE + Piotroski + debt-to-equity quality guard.
Companion post: How to pick value stocks to buy on MTF in India (2026) — the quality-value screen and the annual-churn LTCG method.