Really nice work on this. The model is well built and you’re watching the right thing: cost to income. That’s the real story here. As the branches added last year fill up with customers, costs stay flat while income keeps rising, and profit jumps. I agree FY27 will be a big step up from last year’s 332 crore. So no argument with the direction.
I’d only push back on how optimistic the two cases are, and on one thing that’s missing.
1. The rainy-day cushion is the catch. Your base case has credit cost falling to about 2.2 percent and the bull case holds it at 1.95 percent. But that 1.95 percent is Satin’s number before the extra cushion they set aside. On the call, management guided credit cost of 3 to 3.5 percent including that cushion, and said clearly they will keep building it every good quarter. In other words, they are choosing to report lower profit on purpose to store up protection. So the model is really showing what the company could earn, not what management will actually report. To hit 701 crore, you basically have to assume they stop building the cushion, which is the opposite of what they just told everyone they will do.
2. The bull case needs too many good things at once. To get to 845 crore you need a rating upgrade, no cushion building, almost no Assam damage, full cost savings, and better margins, all landing in nine months. That would put Satin at roughly 6 percent ROA and 30 percent plus ROE, which is CreditAccess Grameen territory, and they took years to get there. Management themselves said the 4.7 percent ROA from last quarter will not repeat and guided 3.5 to 4 percent. Each of those good things is possible on its own. All of them together in one year is a long shot, so I’d treat 845 as the best-case dream, not the bull base.
3. There’s no bear case, and it’s the one that matters most. Both cases assume the good times keep rolling and treat Assam as the only risk. But the real question for this stock is the quality of the low bad-loan number. Bad loans fell to 2.2 percent partly because they wrote off 127 crore, which is more than the 49 crore of fresh bad loans, so some of the “improvement” is just loans leaving the books. On top of that, the loan book grew 27 percent while the number of borrowers barely moved, which means bigger loans to the same people, and that is exactly the setup that goes wrong in microfinance. Management is building those cushions because they see this risk too. A proper bear case would push credit cost up toward 3.5 to 4 percent and pull profit down to maybe 430 to 480 crore.
My rough numbers: I’d put the realistic base around 560 to 650 crore reported (still up 70 to 95 percent on last year), with your 701 as the upside if they ease off building the cushion, 845 as the dream, and a missing bear around 430 to 480.
Bottom line: the model does a great job raising the ceiling, but it under-draws the floor, and it quietly assumes management stops doing the one thing they say defines them. The good news is the takeaway is the same either way. At about 0.9 times book value, even my more cautious 560 to 650 crore puts this on a single-digit P/E. On any of these cases the stock looks cheap, if the low bad-loan number is real. And that is the one thing none of the three cases can prove yet, because the company still won’t show the write-off split, the early-stress loans, or its group return on equity.
Discl- Holding
