the-playbook.in

In the previous article, “ How Probability of Default Is Actually Calculated in the Real World ” we did one thing clearly: We showed that PD is nothing but loans migrating into the 90+ DPD bucket over a defined time window. PD answers: Which loans enter default? Once that happens, PD’s job is over. From that exact moment, a new question takes over: Out of the money that was outstanding at default, how much will we finally lose — and when? That question is LGD . This guide explains LGD the same way the PD guide explained PD: Using tables Using behaviour Using time Using cash No theory first. Reality first. Step 0: Fix the universe (this is non-negotiable) LGD is never computed on: The full portfolio Performing loans Stage 1 accounts LGD universe is only loans that have already defaulted . Meaning: Loans that crossed 90+ DPD Loans that entered Stage 3 So your LGD input table must look like this: Table 1: Default entry universe (output of P...

If you have ever tried to understand Probability of Default (PD) , chances are you’ve been told things like: “PD comes from complex statistical models” “You need logistic regression” “You need advanced mathematics” “This is only for quants” That framing scares people away. The truth is simpler. In the real lending world, PD starts with a very basic question : Out of all loans that are performing today, how many of them will become 90+ days past due within the next 12 months? That’s it. Everything else comes later. Let’s walk through this step by step, using plain data and plain logic. First, let’s fix the definition in your head In most retail lending setups: Default = loan reaching 90+ DPD 12-month PD = probability that a loan which is currently below 90 DPD will migrate to 90+ DPD within the next 12 months PD is not : Recovery Loss Write-off Provision PD is only about migration to default status . Once this clicks, everything else becomes ...

 This is not a textbook explanation. This is how I actually think about portfolios when someone asks: “What will my book look like after 6 months?” “Why are my NPAs rising even though disbursements are stable?” “What happens if my collection efficiency drops slightly?” Let’s start from first principles and build our way up. The easy part. Total AUM forecasting If all you care about is overall AUM , life is simple. You take: Opening AUM Add fresh disbursements Subtract normal amortisation Subtract prepayments And you get closing AUM. For example: Particulars Amount (₹ Cr) Opening AUM 100 Disbursements +20 Normal amortisation −8 Prepayments −2 Closing AUM 110 This works fine if you don’t care where the AUM is sitting . But lending is never that simple. Why bucket-wise AUM actually matters Most real decisions depend on where the AUM sits, not just how much. Bucket-wise AUM is required for: ECL provisioning Understanding delinquency build-up Estimati...

There’s a thought that has been sitting with me for a while, and the more portfolios I look at, the harder it is to ignore. Keeping portfolio delinquency low is not just about credit quality. It’s also about keeping the portfolio young. By young, I don’t mean new as in inexperienced teams or reckless growth. I mean average age of the loan book , measured in months on book. Let me explain where this comes from. Delinquency doesn’t show up evenly over time Almost every lending product behaves the same way if you track it long enough. Early months are usually clean. Then delinquencies start appearing. There’s a phase where defaults peak. Eventually, things stabilize or run off. This is not theory. You see it in microfinance, MSME, LAP, unsecured personal loans. The timeline differs, but the shape doesn’t. So when I see a portfolio with very low delinquency, one of my first silent questions is not “how great is underwriting?” but: How old is this book? Why a young book automati...

Somewhere between the tenth and twentieth portfolio I reviewed, I realised I was no longer surprised by failures. The products kept changing. Home loans, MSME, EV, solar, supply chain. Secured, unsecured. But the reasons things went wrong started repeating themselves. Sometimes slowly, sometimes all at once.

I didn’t start my career in risk modelling. I’m an accountant by training. My early years were spent in accounting and finance roles, working on things like OPEX and CAPEX planning, budgets, and variance tracking. Numbers, yes. But not credit models. Statistically, my toolkit was basic. Mean, median, standard deviation. Some idea of correlation, regression, and probability from textbooks. Nothing fancy. No machine learning. No advanced econometrics. So when I say this, I mean it honestly: If I could understand ECL and work with it comfortably, almost anyone in finance can. ECL looks complex mainly because of how it’s presented, not because of what it’s trying to do. At its core, ECL is built around one very practical question: If this borrower fails to pay in the future, how much will I realistically lose? Everything else exists only to answer this question in a structured, regulator-acceptable way. Before we talk about ECL, let’s talk about provisioning When I was in core fi...