Chapter 2

The machine that never missed

For 26 years HDFC Bank produced the same story every quarter: growth, low bad loans, a premium valuation. The consistency was the product. It was also, critics would later argue, a kind of pressure system.

By the time Aditya Puri stepped down in October 2020, the bank he had run since 1994 was India's most valuable, with a market capitalisation of about US$92 billion, as banking commentator Hemindra Hazari noted in his assessment of the Puri era. Its formula was no secret: lend where risk is measurable, own the salary accounts of corporate India, keep costs lean, never surprise the market.

Puri ran the place as a single instrument. "Nothing moves here without my consent," he would say - a line Hazari quotes while listing the less comfortable features of the era: regulatory non-compliance episodes, mis-selling allegations, intense pressure on junior staff to hit ambitious targets, and what Hazari describes as a lack of transparency with stakeholders about irregularities and sudden senior exits.

"Nothing moves here without my consent."- Aditya Puri, as quoted by banking commentator Hemindra Hazari

The other reading

To be fair to the machine: it worked, honestly, for a long time. The bank avoided the bad-loan crises that gutted Indian banking in the 2010s. Depositors never lost a rupee. Shareholders were rewarded with two decades of compounding. When the RBI later designated Domestic Systemically Important Banks, HDFC Bank was an obvious name on the list.

Hazari's counterpoint was not that the profits were fake, but that the method had a shadow: a culture so tuned to predictability that bad news had to travel very quietly. His parting warning, written days before Puri's exit, was that the successor should ensure HDFC Bank "avoids the fate of Wells Fargo" - the American bank once admired for the same cross-selling machine, before mis-selling scandals destroyed its reputation.

Puri had publicly admired Wells Fargo. Within months of his retirement, the first cracks in the HDFC machine's own cross-selling story were in the newspapers. That is Chapter 5. First, the people who ran the machine.

The four dials

Strip the legend away and the machine's design was elegantly visible. Four dials ran it. First, deposits - above all the salary accounts of corporate India, gathered company by company through relationships that turned every payday into a deposit inflow, the cheapest funding in the industry. Second, distribution - a branch network that functioned less as a place to bank than as a net, trawling every customer interaction for the next product. Third, discipline - a cost culture so lean that the bank's efficiency ratios became a benchmark analysts recited like scripture. Fourth, predictability - the quarterly result that never shocked, the guidance that never wobbled, the bad-loan number that stayed low while the industry's soared. Markets do not pay for profits alone; they pay for profits they can forecast. HDFC Bank sold certainty, and certainty, quarter after quarter, compounded into the steepest valuation premium in Indian banking.

Each dial was defensible. Together they produced something worth a harder look: an organisation whose supreme value was not growth, or service, or even safety - but the uninterrupted continuation of its own story. In such an organisation, the quarterly number is not a report on the business. The business becomes a machine for producing the number. That inversion, invisible in good times, is the key to everything that happened later.

One man, one instrument

Puri ran that machine as a single instrument, and said so. "Nothing moves here without my consent," he would remark - a line Hazari quotes in his 2020 assessment of the Puri legacy, placing it alongside the era's less comfortable features: regulatory non-compliance episodes, mis-selling allegations, intense pressure on junior staff to meet ambitious targets, and what Hazari describes as a pattern of opacity with stakeholders around irregularities and sudden senior-level exits. A chief who insists nothing moves without his consent builds a machine with extraordinary alignment - and a single point of failure. Every escalation path ends at one desk. Every uncomfortable fact must decide, en route to that desk, whether it is worth the journey.

The senior exits Hazari mentions are part of the machine's own folklore. Through the Puri years, a series of deputies seen as potential successors departed the bank - each departure handled, publicly, as routine; each read, privately, as proof that there was room at the top of this bank for exactly one top. The curious reader can line up the dates and draw their own conclusions; the bank offered none. What is documented is the pattern: a long line of strong number-twos, and never a crown prince. When succession finally arrived in 2020, it went to the finance chief - the man who knew the numbers best - and the question of who would hold the culture was left, fatefully, unasked.

The Wells Fargo shadow

Hazari's parting warning, published days before Puri's retirement in October 2020, was the kind of sentence that gets framed if it ages well and forgotten if it doesn't: the successor should ensure HDFC Bank "avoids the fate of Wells Fargo." The comparison was heresy at the time. Wells Fargo, the most admired retail bank in America, had destroyed its own name in 2016 when the world learned that its celebrated cross-selling machine had been manufacturing accounts customers never asked for - millions of them - under sales pressure so extreme that employees described it as a grinder. Puri had publicly admired Wells Fargo's model. The machinery Hazari was pointing at was structurally similar: own the customer relationship, then maximise the products per relationship, with targets that cascade down ten levels of management until they land on a desk where someone's month depends on a signature.

Heresy has a shelf life. Within a year of the warning, HDFC Bank's own cross-selling machinery was in the newspapers - the GPS devices bundled into car loans that customers never knowingly bought (Chapter 5), the regulator's ₹10 crore penalty, the departures at the top of the auto-loan book. The Wells Fargo parallel is not this dossier's invention; it was published, dated and signed, by one of the few banking commentators in the country willing to write it while the myth was still warm. The reader can judge how it aged.

The salary-account moat

Every bank wants deposits. HDFC Bank built a moat around the cheapest kind. Its corporate relationships made it the salary bank of white-collar India: employer after employer routed payroll through the blue grid, and each new company onboarded delivered thousands of individual depositors at once - sticky, low-cost, transaction-heavy balances that funded the loan book at a spread competitors could only envy. The salary account is the quiet engine of the whole legend. It explains the deposit franchise, it explains the cross-sell pipeline (every salary account is a pre-qualified lead for a credit card, a personal loan, an insurance policy), and it explains the market's faith: a bank that owns payday owns the customer. The chapters on people and commissions describe what flowed through that pipeline at full pressure. Here it is enough to note the design: the moat was dug in the 1990s and it has never stopped refilling.

The quarter that never surprised

Ask what, exactly, investors paid the premium for, and the answer is a ritual. For over two decades, HDFC Bank's quarterly results were the least dramatic event in Indian finance: loan growth in the high teens or better, bad-loan ratios a fraction of the industry's, fee income climbing, guidance met. Analysts wrote the preview before the results and rarely had to edit. This consistency was a managerial achievement of the first order - and, Hazari's critique insists, a managerial incentive of the first order too. An institution whose stock price depends on never surprising the market has made a promise that can only be kept two ways: by genuinely flawless execution, or by ensuring that surprises never reach the surface. For most of the Puri era, the evidence supports the first. The question the post-2020 record raises - the RBI's digital-launch ban, the GPS probe, the penalty letters, the chairman's letter - is whether the second began, gradually, to substitute for it. That question is the spine of this book.

Too big, too smooth

The machine's ultimate certification came from the regulator itself. When the RBI designates its Domestic Systemically Important Banks - the institutions whose failure would threaten the system, the formal "too big to fail" list - HDFC Bank sits on it alongside the State Bank of India and ICICI Bank. The designation is an honour with teeth: higher capital requirements, closer supervision, the explicit acknowledgment that the country cannot let this institution stumble. And there is the paradox at the heart of this dossier. A bank too big to fail is also a bank too admired to question - too smooth for the scepticism every other institution gets by default. Regulators penalise it and the penalties are footnotes. Its chairman resigns over values and ethics and the machinery of reassurance starts within hours. The machine that never missed taught everyone - investors, journalists, supervisors - not to look for a miss. What follows is the record of what not-looking cost.

The price of perfection

Perfection has a shadow price, and in organisations it is paid in candour. The management literature is unglamorous on this point: cultures that punish variance eventually stop reporting it. The numbers still look smooth; the turbulence moves underground - into group chats, into resignation letters, into the gap between what the branch knows and what the board is told. HDFC Bank's own later record reads like a case study in the phenomenon. The GPS bundling ran for about four years before it surfaced, and the internal probe that followed cost senior executives their jobs, as Chapter 5 documents. The outages that provoked the RBI's unprecedented digital-launch ban recurred for two years before the regulator acted, per Chapter 6. In each case the machine's public face held steady right up to the moment the record broke. Smooth, then suddenly not.

To be scrupulously fair - and this dossier's credibility depends on fairness - the defenders have a real case. The bank never produced a bad-loan debacle. It never needed a bailout, never froze a depositor's money, never restated its books. India's banking history is littered with institutions that did all three. If the choice is between a machine that runs too hot and a machine that does not run, the first is the better problem to have. The critique here is narrower and, for that reason, harder to dismiss: the very traits that made the machine admirable - the centralisation, the intolerance of variance, the worship of the uninterrupted quarter - are the traits that allowed its failures to gestate unseen. The strengths did not prevent the weaknesses. They concealed them, until they couldn't.

What the next custodian inherited

Puri's exit, on 26 October 2020, was choreographed to reassure: a planned retirement, a respected insider successor, a market capitalisation near US$92 billion, as Hazari noted at the time. The machine, apparently, would run itself. But a machine built around one man's vigilance does not run itself; it runs on momentum while the new driver learns the controls. The next chapters trace the momentum's decay in real time: the digital crashes that brought the regulator to the door within six weeks of the handover, the forced-selling scandal breaking within months, the attrition spike, and eventually the letters - a chairman's resignation, an FIR, a board fining its own CEO - that momentum could not absorb. The machine did not fail all at once. Machines like this never do. They miss a little, and then, one day, the miss is the story.

The numbers behind the legend

Measure the machine at the moment of handover. Market capitalisation around US$92 billion, per Hazari's October 2020 tally - at points in that era, the most valuable bank in the emerging markets outside China. A branch network that had grown from one Sandoz House office to thousands, an ATM estate in the tens of thousands, a workforce past one lakh and hiring at industrial pace. And beneath it all, the metric the street cared about most: an unbroken multi-decade run of profitable quarters through every macro shock India could supply - the 1997 Asian crisis, the 2008 global crisis, demonetisation in 2016, the IL&FS panic, the pandemic. Each shock wrecked somebody. None of them wrecked the blue grid. That is not a small thing, and nothing in this dossier pretends otherwise.

But note what a record like that does to the people paid to extend it. Every new executive inherits not a business but a streak. Streaks create their own gravity: no one wants to be the quarter it ends, the desk where the number slipped, the officer whose name is on the first red cell. Multiply that psychology across two lakh employees and you have an institution with a tremendous, unspoken incentive to make problems small before they become official. The regulators' files in the later chapters - the show-cause notices, the penalties, the bans - are what happens when small problems refuse to stay small.

The consent bottleneck

There is one more structural feature worth naming, because the succession chapters depend on it. A bank where "nothing moves without my consent" is a bank that never has to develop institutional consent - the committees, the challenge functions, the documented disagreements through which lesser organisations grope toward decisions. Puri's bank outsourced those to Puri. It worked because he was, by every account including his critics', exceptional. The risk it stored up is the risk of all bottlenecks: remove the person, and the organisation must suddenly operate muscles it never built. Watch, in Chapter 8 and Chapter 12, how quickly the post-Puri bank started exhibiting exactly the failures a strong challenge function exists to catch - an IT estate allowed to decay until the regulator intervened, a sales culture whose excesses reached the newspapers, a board that ended up fining its own chief executive. The machine's greatest strength was its single instrument. It turned out to be its greatest vulnerability too.

The religion of the number

Wells Fargo deserves one more paragraph, because the parallel is mechanical, not rhetorical. The American bank's crime was not greed in the abstract; it was a specific device - "eight is great," the target of eight products per household - wired into pay, promotion and punishment until employees opened accounts customers never requested to keep their jobs. The device is the important part. Any bank that worships products-per-customer and pays on the count will eventually manufacture the count. HDFC Bank never faced a Wells-scale accounting of its own cross-selling; the GPS affair documented in Chapter 5 was a single product line, caught by an internal probe and punished by the regulator with a ₹10 crore fine. But the structural ingredients Hazari listed in 2020 - the ambitious targets, the pressure on junior staff, the opacity about irregularities - are the same ingredients, arranged the same way. The reason to study the machine is not to convict it of Wells Fargo's sins. It is to notice that the machine was built to the same blueprint, and to ask who, in the years after its architect left, was checking what the blueprint produced.

That is the chapter's ledger. On one side: twenty-six years of compounding, a deposit franchise without peer, a risk record that survived every crisis, an institution so reliable the country formally cannot let it fail. On the other: a culture of consent centralised in one man, a streak psychology that punishes candour, and a cross-selling engine one step from the blueprint that broke American banking's most admired name. The legend and the liability were never opposites. They were the same machine, photographed from two sides - and the rest of this book is the second photograph.

One closing image, and then the people. In the bank's own mythology the machine's emblem is not the branch but the grid itself - four squares, orderly, identical, each in its place. It was meant to signal solidity. It also, read with a colder eye, describes the architecture perfectly: everything in its cell, every number in its box, every quarter aligned with the last. Order is a genuine achievement. It is also, sometimes, a way of making sure nothing unexpected is ever in frame. The next chapter is about the people who lived inside the grid - and the year a third of them climbed out.