From 2027, Banks Will Score Your Probability of Default. Here's What That Means.
RBI's Expected Credit Loss (ECL) rules are the biggest change to Indian lending in a decade — and almost nobody is explaining them to the people they affect most. This guide does, in plain language.
Section 1
What changed?
Today, Indian banks treat a loan as a problem only after something goes wrong — you miss payments, the loan turns into an NPA, and only then does the bank set money aside for the loss. This is called the incurred loss model. It looks backwards.
From April 2027, RBI replaces it with Expected Credit Loss. Banks must now estimate — at the moment they give you a loan, and continuously for its whole life — the mathematical probability that you will default. They must set aside capital based on that probability even if you have never missed a single EMI. This brings India in line with how banks already work globally under IFRS 9.
The simplest way to think about it
It works like vehicle insurance. A careful driver and a risky driver both get a policy — but at different premiums, because the insurer prices the expected cost of each driver. From 2027, loans work the same way: the bank prices the expected cost of you.
Section 2
The three numbers that decide everything
Every bank's ECL model rests on three quantities. Expected loss = PD × LGD × EAD.
PD
Probability of Default
How likely are you to stop paying? Built from your income type and stability, credit history, debt load, and even conditions in your sector.
Example: A government employee with steady salary might be scored at a fraction of the PD of a freelancer whose income halves in some months — even at the same CIBIL score.
LGD
Loss Given Default
If you did default, how much would the bank actually lose? Collateral slashes this number — that's why secured loans are cheaper.
Example: A home loan is backed by the house, so the bank might recover 80% of its money. A personal loan has nothing behind it — the loss is much larger.
EAD
Exposure at Default
How much money would be on the table at the moment of default? Early in a loan it's almost the full amount; it shrinks as you repay.
Example: Defaulting in year 2 of a 20-year home loan exposes the bank to far more than defaulting in year 18 — which is why long tenures carry more lifetime risk.
Section 3
Every loan now lives in one of three stages
This is the part that matters most for existing borrowers — because the move from Stage 1 to Stage 2 does not require a missed payment.
Performing
When: The loan is healthy and risk hasn't changed meaningfully since approval.
What it means for you: Business as usual. The bank provisions only for the next 12 months of expected loss.
Risk has increased
When: Your measured risk rises significantly — income dropped, your sector weakened, your credit profile slipped. No missed EMI required.
What it means for you: The bank must provision for losses over the loan's whole lifetime — far more expensive. That cost can come back to you as rate resets or tougher terms on top-ups and renewals.
Defaulted
When: Payments are overdue past the NPA threshold (typically 90 days).
What it means for you: Full provisioning, recovery proceedings, and severe long-term damage to your credit history.
Section 4
Who is most affected?
Gig workers and freelancers
Income volatility feeds directly into PD models. Two people earning ₹50,000 a month on average will be scored very differently if one earns it steadily and the other swings between ₹20,000 and ₹90,000. Expect more Stage 2 reclassifications in this group than any other.
The self-employed and small business owners
Sector conditions now matter. If your industry hits a rough patch, your loan can be flagged even while your own business is fine — the model works on probabilities across groups, not just on you.
Anyone stretched thin
A high EMI-to-income ratio (FOIR above ~50%) plus a thin savings buffer is the classic high-PD profile, whatever your job. One income shock away from missed payments is exactly what the models are built to detect.
And one more subtle effect: two borrowers with the same CIBIL score can now get different rates, because CIBIL is only one input into PD. Your income type, stability and buffer are the rest — which means they are now levers you can actually use.
Section 5
Five things that genuinely lower your risk score
Build a 3–6 month buffer
Demonstrated liquidity is the single strongest defence against Stage 2. Park it in a savings account or FD — visible to a bank, instantly usable by you.
Make your income legible
Route everything through one bank account. Twelve months of clean statements turns "volatile freelancer" into "documented earner with a track record" in the model's eyes.
Keep FOIR under 40%
If EMIs eat more than half your income, you are in the high-PD zone. Close the smallest loan first — it is the fastest visible improvement.
Add a salaried co-applicant
A co-applicant with steady income directly cuts the household PD — often worth more than a 30-point CIBIL improvement.
Prefer shorter tenures when you can
Lifetime PD grows with time. A shorter loan means less time for things to go wrong, which models reward with better pricing.
See where you stand
Banks will compute this number about you. Compute it about yourself first.
Set up your free financial profile — income, EMIs, savings — and get your indicative risk band with the exact factors driving it. Everything stays on your device.
Sources: RBI discussion paper and directions on ECL-based provisioning; KPMG India, “Expected Credit Loss” (November 2025). This guide is educational — banks' actual models are proprietary and bank-specific.