The Psychology of Compounding: How to Lose Less and Earn More

ELI5 / TLDR

A fund manager with 30 years in Indian markets argues that the way to win at investing is not to be brilliant but to avoid four specific blunders, all of which are entirely within your control. His core obsession is temperament over intellect: most people lose money not because they failed to find the next great stock, but because they couldn’t stand watching a neighbour make more. Buy good companies when they’re temporarily unloved, refuse to overpay, ignore everyone else’s returns, and give it three to five years. The unglamorous goal is to lose less, because losing less leaves more capital to compound into the next cycle.

The Full Story

E.A. Sundaram is the CIO of Bugle Rock PMS and has run money across institutions, mutual funds, and family offices since the mid-1990s. The interviewer (Shrishti Sahu of The India Opportunity) treats him as an icon; the framing is reverent, but the content holds up. He once ran the fund that is now HDFC Flexi Cap, back when it was called Zurich Equity Fund.

Don’t compete on returns

Asked to compress three decades into one sentence, he picks: do not be bothered about how much money someone else has made. He calls this the root cause of nearly every investing mistake.

An investor left to himself or herself is happy with a 15 to 16% rate of return from India stock market. But that is until he or she sees that the neighbor has made 30%.

The danger isn’t envy in the abstract. It’s that the industry rewards being “the best,” and the only way to be the best in a frothy market is to load up on whatever is already dangerously overvalued. His sharpest framing: it is important to be consistently good; it is dangerous to try to be the best, because being the best is a transient position that usually requires taking crazy risk.

Two scars taught him this. In the 1995-97 IPO boom he and his young colleagues bought into junk, made quick money, then got stuck holding illiquid names when the market slumped. Lesson: never buy measurable junk, even free. In 1999, his fund returned 160% during the dot-com boom and it still wasn’t enough, because others had done 250% by going even harder into tech. That’s when it sank in that no return will ever satisfy, because there is no such thing as enough money. Competing on returns alone is a losing game by construction.

The subtraction approach

He splits investing into two styles: succeeding by doing the right things, versus succeeding by avoiding the wrong things. The first requires forecasting hundreds of variables nobody controls (interest rates, election seats, the Ukraine war, commodity cycles, FII flows) against an army of equally-credentialed people with the same Bloomberg terminals reading the same annual reports. He’s blunt that no one has a sustainable edge there.

The second style narrows the entire game to four errors, all 100% within your control:

Don’t buy measurably weak or deteriorating businesses. Don’t buy any company where the management has shown a tendency to go against the interest of minority shareholders. Don’t buy at exorbitant valuations. Don’t be bothered too much about how much money somebody else has made.

The point of avoiding mistakes isn’t to make more; it’s to lose less. Less glamorous, equally important: lose less in a bear market and you carry more capital into the next compounding cycle.

Don’t look at the index

A recurring theme: the market is not a monolith. The index being expensive tells you almost nothing, because at every peak large parts of the market were cheap. In 1999-2000 the “old economy” (FMCG, autos, pharma) was cheap while only tech, telecom, and media raged. In 2007, with the index at a peak, FMCG and pharma and names like Page Industries and Eicher were cheap while real estate and infra boiled. In 2016 small/midcaps roared but Infosys traded at a P/E of 12.5. So opportunities exist in any market environment; you just have to look at companies, not the index. (Unless you’re an index investor, he says, the index movement shouldn’t bother you.)

Buy good companies when they’re in trouble

The capital market exists to reward efficient use of capital. So pick companies that use capital well, and use your own capital well by not overpaying. The catch: a good asset is cheap only when it’s unpopular, and a good company is unpopular only when it’s going through a period of difficulty. He’s happy to buy then, after satisfying himself the difficulty is temporary, not terminal, and the long-term ability to compete is intact.

The opportunity is a gap between reality and perception, when perception is weaker than reality. He never buys in one shot; he averages in over installments, often ending up with an average cost below his first purchase. This requires only a longer time horizon, not superior intelligence or clairvoyance.

His worked example is Indraprastha Gas (IGL), bought in 2012. The regulator (PNGRB) tried to control its selling prices; the stock crashed. The company filed a writ; the Delhi High Court ruled within two months that PNGRB had no power to control prices (the order was ultra vires) and that price control only applies where there’s a restrictive trade practice. The case dragged at the Supreme Court for two years, which was precisely the window to accumulate, because a categorical High Court judgment like that was unlikely to be overturned. When the Supreme Court affirmed, the stock ran. The psychological cover for the two-year wait: he wasn’t holding a weak company.

Investing is a journey through uncharted territory. If it is a journey through uncharted territory, it is better to have strong companions.

The strong companions: quality of business, quality of balance sheet, management’s friendliness to minority holders, and the entry price. Pay attention to those and you have less blood pressure. He’s wrong two or three times out of ten and is fine with that, because every style has that hit rate.

What “high quality” actually means (measurable)

He insists the parameters be verifiable, so a client can check that what he claims matches what he does:

Track record of at least 15 years. He is not an early-stage investor. A long track record means the company has survived at least one full cycle, and a difficult period teaches management what no MBA school can.
Minimum size: 500 crore of revenue (for non-financials).
Consistently high return on capital employed, and consistent free cash flow generation. Free cash flow matters more than reported net profit because profit can be massaged with accounting; free cash flow can’t. It also proves the company lives within its means. He admires companies (and families) that do, and avoids businesses that constantly need external capital to fund their own growth.
Financials get their own set: ROA, ROE, cost-to-income, NPA levels, a longer track record.

Running the whole database through these filters leaves roughly 210 companies out of ~5,300 listed names. The portfolio of ~25 names is built only from those 210. But he keeps a small, deliberate allowance for exceptions, strong companies in cyclical businesses that fail the ROC/FCF screens, like Larsen & Toubro (bought ~3 years ago) or State Bank of India during COVID. Discipline without dogma; each exception made with eyes open and a hard cap on total exceptions.

Past track record is necessary but not sufficient. He also wants confidence the competitive position will stay intact, the ability to grow remains (he doesn’t forecast earnings to the millimetre, just needs revenue and profit to be substantially higher in 3-5 years), and clean management, judged by capital allocation, consistent dividends, transparency in bad times as well as good, and no routing of profitable business through unlisted entities.

On valuation, he corrects a common misreading: value investing is not “never pay above 9x P/E.” Buying junk cheap is not value investing. Value investing is buying quality and not overpaying. Given a choice, he’ll take a good business at a reasonable price over a mediocre business at a bargain price every time.

The SBI trade

He walks through State Bank of India around 2020 in detail, because the numbers were visible to everyone yet the stock was hated. Between 2010 and 2020 SBI’s market share was rising across housing, auto, personal, and corporate loans; NPAs fell from ~5.3% in 2018 to ~2.4% by 2020; it led HDFC and ICICI in digital banking, had 20,000 branches and 50,000 ATMs, one in four Indians as customers, every government department and the armed forces banking with it, plus successful mutual fund and insurance subsidiaries. The stock traded at 0.7x book, a multi-year low. The market’s argument was that public-sector banks were losing to private, which was true in aggregate but not for SBI. “You can’t paint everything with the same brush.” Perception weaker than reality, again. It took 3-4 years to play out.

Selling, sizing, cash

Four reasons to sell, and only four: (1) the original assumptions become invalid (rare; he sells regardless of profit or loss); (2) the stock has gone well beyond what he thinks it’s worth (the most common reason); (3) a clearly superior opportunity appears and he switches from A to B; (4) the client wants a redemption. His patience rule: he’ll hold for three years even with no return, and only replaces a name if both the stock is flat for three years and the fundamentals haven’t improved. If fundamentals improve, he holds (and buys more) regardless of price action.

Position-sizing discipline: minimum 2.5-3% per name (a quarter-percent position can’t move the needle even if it quadruples), maximum 10% per stock, maximum 30% per sector, at least four non-correlated sectors at all times, and enough liquidity to exit the entire position within five to six trading sessions. The portfolio is deliberately a mix of fast growers, blue-chip stalwarts, secular compounders, and cyclicals, not just “best ideas,” because construction is about balance.

Cash is a residual, never a target. It’s whatever’s left after opportunities are taken, usually 6-7%, and he was nearly fully invested during COVID. Setting a fixed cash level makes no sense to him precisely because opportunities don’t track the index.

Temperament, the architect metaphor, and enough

His central complaint about the industry: the “intellect” part of investing (economics, valuation, competitive analysis) gets all the attention, while building a sound temperament gets none, even though it’s equally important and is the part within your control. You can control what you buy, how much, and at what price. That’s it.

On geopolitics and AI disruption, he refuses to forecast and instead builds resilience. The metaphor: a portfolio manager is like an architect building a house by the sea. The job isn’t to predict the next tsunami; it’s to build a house that survives one. He does ring-fence the portfolio away from sectors most exposed to AI (low-level coding, mundane jobs) and away from highly capital-intensive businesses, cutting-edge fast-changing technology, and businesses with violently fickle demand. Governance is non-negotiable.

On India specifically: bullish on consumer spending (both FMCG and discretionary) for the next couple of decades, helped by GST reforms, the infrastructure multiplier, and PLI scheme success in areas like mobile manufacturing. His main worry is the slow pace of job creation in a country where more than half the population is under 30.

The closing register turns personal. His late father spent his last decade only in bank fixed deposits at 7-8%, distrusting his own son’s investing, and was happy. The point: he knew what he wanted, and that’s what matters. The big takeaway he’d choose if forced to one: being consistently good beats trying to be the best, because the best is never a sustainable position. And the rule he’d pass to his children: spending less is as important as making more, in proportion to what you already have. A successful life, he says, is health, family, friends, enough money for your needs, and giving the surplus to those less fortunate.

Key Takeaways

The four permanent-damage errors, all within your control: weak/deteriorating businesses, managements hostile to minority shareholders, exorbitant valuations, and caring how much money others have made.
“It is important to be consistently good. It is not important to be the best.” The best position is transient and usually demands reckless risk.
The goal of avoiding mistakes is to lose less, not earn more, because losing less leaves more capital to compound into the next cycle.
The index is not the market. At every historic peak, large swaths of the market were cheap (old economy in 2000, FMCG/pharma in 2007, large caps in 2021).
A good asset is cheap only when unpopular, and a good company is unpopular only during a temporary, non-terminal difficulty. Buy then.
The edge is a gap where perception is weaker than reality. Establish through research that the long-term competitive position is intact.
Quality, measured: 15-year minimum track record, 500 crore minimum revenue, consistently high ROCE, and consistent free cash flow (which can’t be faked the way net profit can).
Free cash flow proves the figures are believable and the company lives within its means. Avoid businesses that constantly need outside capital to grow.
His filter leaves ~210 of ~5,300 listed Indian companies; the 25-stock portfolio is drawn only from those, with a small capped allowance for strong cyclicals (L&T, COVID-era SBI).
Value investing is buying quality without overpaying, not buying junk at a low P/E. Prefer a good business at a fair price over a mediocre one at a bargain.
SBI at 0.7x book (~2020): rising share, falling NPAs, digital lead, vast deposit base, hated only because “PSU banks are losing” was true in aggregate but false for SBI.
Four (and only four) reasons to sell: thesis broke, price far exceeds worth, a clearly superior swap appears, or client redemption.
Patience rule: hold three years even at zero return; replace only if flat for three years AND fundamentals haven’t improved.
Sizing discipline: 2.5-3% min, 10% max per stock, 30% max per sector, four-plus non-correlated sectors, full exit possible within 5-6 trading sessions.
Cash is a residual (usually 6-7%), never a target, because opportunities don’t track the index.
The architect metaphor: don’t predict the tsunami; build a house that survives one. Resilience over forecasting for macro, geopolitics, and AI.
A long track record matters because surviving a full cycle teaches management what no business school can; a difficult period is the best teacher.

Claude’s Take

This is the genuine article: a practitioner who’s internally consistent across two hours and never once reaches for a tip or a forecast. The signal-to-noise is high because his entire philosophy is built to suppress noise, so the interview ends up modelling its own thesis. The four-error frame and the “consistently good beats best” line are worth keeping; the SBI and IGL examples are concrete enough to verify and show the method has teeth, not just aphorisms.

Two honest caveats. First, survivorship and framing: this is a reverent interview, not an audit. He cites his wins (SBI, IGL, calling the IT/infra/NBFC tops) and mentions a 2-3 out of 10 miss rate without naming a single specific loss, so you’re taking the track record on faith. Second, a fair amount of it is the standard quality-value-temperament canon (circle of competence, Buffett’s two rules, “be greedy when others are fearful”) that any reader with a finance background has met before, hence the lighter fermentation here. What lifts it above a truisms reel is the operational specificity: the actual screens, the 210-name universe, the sizing limits, the five-to-six-session liquidity rule, the three-year patience condition. That’s the part you can’t get from a quote graphic.

Score 8. Docked from higher because the back third softens into life-philosophy and the format never pushes back on him, but the framework density and the candor about temperament being the real game make it well above the average finance interview.

The Psychology of Compounding: How to Lose Less and Earn More | Ft. E.A. Sundaram