Understanding Provision for Doubtful Accounts: More Than Just a Buffer
Let’s clear something up: PDA isn’t money the bank admits is lost. Not yet. It’s a precaution. A promise to itself that if a loan goes south, the hit won’t come out of nowhere. Banks lend money—millions, sometimes billions—knowing full well not all of it will return. So they don’t wait for defaults to happen. They predict them. And that prediction? That’s the PDA. Think of it like car insurance. You pay monthly, hoping you’ll never need it. But if you crash, you’re covered. Same idea.
Now, here’s the thing—banks don’t guess randomly. They use models, historical data, economic forecasts. A loan to a shaky startup in a recession? Higher PDA. A mortgage from someone with a 780 credit score and steady job? Lower. The system isn’t perfect, but it’s structured. Regulatory bodies like the RBI in India or the FDIC in the U.S. set guidelines. Under Ind-AS 109 (India’s version of IFRS 9), banks must now use expected credit loss models—meaning they can’t wait for signs of trouble. They must anticipate it from day one. And that changes everything. It forces banks to be forward-looking, not reactive.
How Provisions Are Calculated: The Mechanics Behind the Numbers
Each loan is categorized by risk. Tier 1—low risk. Tier 3—high risk. A construction company with delayed projects? Tier 2 or 3. A government employee with a home loan? Tier 1. Then, banks apply a percentage. Maybe 2% for Tier 1, 15% for Tier 3. Multiply that by the loan amount. That’s the PDA. For a ₹50 lakh commercial loan in Tier 3, that’s ₹7.5 lakh set aside. Multiply that across thousands of loans and the numbers explode. In 2023, HDFC Bank booked ₹12,800 crore in provisions. ICICI? ₹9,400 crore. These aren’t minor expenses—they’re major profit dampeners.
Models use variables: unemployment trends, sector performance, interest rates. If steel prices crash, loans to steel manufacturers get marked up. It’s dynamic. But—and this is critical—different banks can arrive at different figures using the same data. Why? Because judgment calls exist. How much weight do you give to a company’s new CEO? Is inflation temporary or structural? These nuances shape the PDA. And that’s why auditors scrutinize them so closely.
Why PDA Matters: The Ripple Effect on Profit and Trust
Here’s a fact people don’t think about enough: PDA directly hits the bottom line. Every rupee or dollar set aside is one less for dividends, bonuses, expansion. When ICICI Bank increased its PDA by 30% in Q2 2022, its net profit fell by 14%. Investors reacted. The share price dipped. And that’s exactly where the pressure builds. Management wants profits up. Regulators want prudence. Shareholders want returns. Banks are caught in the middle.
But there’s a bigger picture. Under-provisioning—setting aside too little—can backfire spectacularly. Remember Yes Bank? In 2018, it was booking minimal provisions on bad loans. Then reality hit. The RBI stepped in. By 2020, it needed a ₹10,000 crore bailout. Confidence evaporated. So yes, PDA protects the bank, but it also protects depositors. We trust banks with our money because we assume they’re prepared for failure. Without strong PDA practices? That trust frays.
To give a sense of scale: India’s banking sector held over ₹7.2 lakh crore in NPAs in 2023. If provisions covered only half, the shortfall would be catastrophic. PDA is the dam holding back the flood.
PDA vs. NPA: What’s the Real Difference?
People mix these up constantly. NPA stands for Non-Performing Asset—loans that have stopped paying interest or principal for 90 days. PDA is the money banks set aside before that point. NPA is the corpse. PDA is the funeral fund. One is backward-looking. The other is predictive. And that distinction matters.
For example: A textile exporter misses a payment in January. Not yet an NPA. But if the bank sees global cotton prices crashing and orders drying up, it might increase the PDA in December—before the default even occurs. That’s proactive risk management. But if the bank ignores the warning signs? Come April, that loan becomes an NPA. The PDA was too low. The loss hits harder.
And here’s a twist: a loan can have a high PDA without becoming an NPA. Maybe the borrower restructures the debt, finds new investors. The bank recovers most of the amount. The PDA was precautionary, not prophetic. Yet it still reduced reported profits. That’s the irony—being cautious can make you look less profitable, even if you’re wiser.
How Economic Cycles Influence PDA Decisions
Recessions spike PDA. Pandemic? Skyrocketing. 2008 crisis? Banks went into overdrive. During India’s 2016-2018 bad loan crisis, PDA levels averaged 1.8% of total advances. In normal years, it’s closer to 0.7%. That’s a 157% increase. Why? Because risk perception shifts. Lenders start seeing ghosts in every balance sheet.
But the reverse happens too. In boom times, banks get complacent. Startups get easy credit. Real estate loans balloon. PDA ratios dip. We saw this in 2021—post-vaccine optimism, low provisions, high lending. Then inflation hit. Rates rose. Defaults followed. And banks scrambled to catch up. The problem is, PDA adjustments lag reality. By the time banks admit risk is rising, it’s often already here.
Which explains why some experts argue for counter-cyclical provisioning—building reserves during good times to cushion bad ones. But regulators hesitate. It would penalize banks for doing well. Hence, we’re far from a global standard. The EU experimented with it. India rejected it in 2019. The issue remains unresolved.
Frequently Asked Questions
Is PDA the same as a loan loss reserve?
Yes—almost. In U.S. banking, it’s often called the Allowance for Loan and Lease Losses (ALLL). Same concept: funds set aside for expected defaults. But U.S. banks use CECL (Current Expected Credit Loss) models, which are more aggressive than India’s ECL Stage 1-3 system. Under CECL, banks estimate losses over the entire loan life at origination. That means a 30-year mortgage gets a provision on day one. It’s more conservative, but critics say it overstates risk. Data is still lacking on long-term impact.
Can banks reverse PDA if loans recover?
They can—and do. If a borrower repays, the bank reverses part of the provision. That reversal boosts profits. In 2022, State Bank of India reversed ₹2,100 crore in provisions after recovering dues from Bhushan Power & Steel. That wasn’t new revenue—it was old money returning. But it still lifted net income by 8%. That said, reversals are rare. Most defaults don’t end happily.
Do all types of banks use PDA?
Yes. Commercial banks, co-operative banks, even NBFCs (Non-Banking Financial Companies) must maintain provisions. But enforcement varies. Urban co-ops? Often under-provisioned. Large private banks? Tightly monitored. The difference in compliance explains why some collapsed during the 2019 PMC Bank crisis while others sailed through.
The Bottom Line: PDA Is a Discipline, Not a Choice
Let me be blunt: PDA isn’t optional. It’s survival. You can’t run a bank without it. But how strictly it’s applied—that’s where judgment, culture, and regulation collide. Some banks (like Kotak Mahindra) take a hardline, conservative approach. Others (like certain regional lenders) cut corners until auditors force corrections. I find this overrated: the idea that technology alone can fix provisioning. AI models help, yes. But they can’t replace skepticism, experience, or regulatory teeth.
We’re not dealing with abstract numbers. A poorly managed PDA system can wipe out a bank. It can trigger runs. It can destabilize economies. The 2008 global crisis wasn’t just about subprime mortgages—it was about banks not setting aside enough for what they knew, deep down, was a ticking time bomb.
So what’s the takeaway? PDA stands for Provision for Doubtful Accounts—but what it really measures is a bank’s honesty about risk. Its courage to face the worst-case scenario. And its respect for the people who trust it with their money. That’s not just accounting. That’s ethics. And that, suffice to say, is something no algorithm can fully capture. Honestly, it is unclear whether we’ll ever build a system that balances caution and confidence perfectly. But until then, PDA remains our best imperfect shield.