The Economics of a Lending Business

If you strip away the positioning and strategic narratives, every lending business is simply a continuous tug-of-war between three levers: Yield, Operating Expenses (Opex), and Non-Performing Assets (NPAs). In this opening module, we deconstruct the fundamental math that governs credit platforms — the ROA tree that dictates profitability, the "Valley of Death" that kills early-stage fintechs, how leverage shapes capital structure, and how subvention hacks the traditional rules of underwriting.

Principle 1.1: The ROA Tree

How a Lending Business Actually Makes Money

Every lending business earns money through this waterfall, expressed as a percentage of the loan book (AUM). The ROA tree is the single most important diagnostic tool for understanding a credit business. While the abstracted formula is universal, the second-order variables interact in wildly different ways depending on the strategy.

When you look at the public P&Ls of the industry giants, you see exactly how they chose to play this game:

What is the real story behind the numbers?

• The Volatility Play (SBI Cards): Credit cards are the highest-yielding asset class, but they are completely unsecured. When inflation bites, the borrower stops paying the credit card first. Notice the brutal 7.7% credit cost and 2% collection cost. Their ROA looks lower than you'd expect from a 23% yield because the loss and operational intensity are massive.
• The B2B2C Masterclass (Synchrony): Notice their Customer Acquisition Cost (CAC) is only 1%. Because merchants like Amazon and Walmart acquire the customer at the checkout counter, Synchrony sidesteps the bloodbath of consumer marketing.
• The Risk Pricers (Shriram) vs. The Balance Sheet Optimizers (Chola): Shriram lends to first-time truck buyers with thin files. They know their NPAs will be high (5.5%) and that repossessing trucks is expensive (2.5% collection cost). They don't try to artificially lower their NPAs; they just price the risk perfectly at 19.5% gross yield. Chola, on the other hand, serves the premium buyer in the same category. Their credit cost (1.9%) is less than half of Shriram's. By maintaining high asset quality in a secured book, Chola enjoys one of the lowest Costs of Funds in the market, allowing them to maintain a higher ROA despite charging a yield that is 200 bps lower.

Principle 1.2: Minimum Viable Scale and the "Lumpy Opex" Trap

In a traditional SaaS business, you see the famous hockey stick: upfront development costs followed by Opex that scales in a smooth, predictable curve alongside revenue. Lending is very different. In a credit business, Opex doesn't look like a curve — it looks like a staircase.

This is the "Lumpy Opex" trap, and it creates a "Valley of Death" for early-stage fintechs. Look at the actual timeline of building out a credit business as an LSP:

• Month 1 (The Build): You hire a Chief Risk Officer and a Tech Lead, and set up your back end for the lending infrastructure. Your central overhead instantly spikes to ₹20 Lakhs/month. Your AUM is ₹0.
• Month 3–6 (The Illusion of Scaling): You start disbursing. You set up a basic underwriting and repayment team. Fixed costs hit ₹30 Lakhs/month, but you've maybe disbursed ₹5 Crores. Your fixed Opex-to-AUM sits at a brutal 6%.
• Month 12–15 (The False Summit): AUM scales to ₹25 Crores. Fixed Opex temporarily stabilizes. Suddenly, Opex-to-AUM drops to a highly profitable 1.2%. You think you've made it.
• Month 18–24 (The System Break): At ₹100 Crores AUM, the early infrastructure breaks under the weight of volume. Compliance nuances pile on. You're forced to hire a 20-person team spanning fin-ops, collections, and credit policy, reinforce your tech stack, and migrate to an enterprise LMS. Fixed costs instantly gap up to ₹1 Crore/month. Profitability vanishes overnight until AUM can catch up to this new "lump."

Lending businesses only become profitable when the AUM grows large enough to absorb the last lump of fixed costs before the next lump is required.

Principle 1.3: Capital Structure is the Science of Leverage

LSP vs. NBFC

Capital structure is not just about who owns the customer; it is a profound strategic trade-off. To see how leverage dictates outcomes, look at the correlation between Return on Assets (ROA), Debt-to-Equity (D/E), and the resulting Return on Equity (ROE):

The Physics of Leverage

Notice how Five Star Business Finance and Chola generate nearly identical ROEs (~17.6% vs. ~18.2%), but do it using vastly different physics. Five Star relies on a massive ROA (8.4%) and runs low leverage (1.5x D/E). Chola runs a tighter ROA (2.7%) but cranks their leverage up to 6.6x to achieve the same equity return.

Regulators and investors are hyper-sensitive to Capital Adequacy and Leverage Ratios for a reason. High Debt-to-Equity ratios create phenomenal ROE during bull markets, but leverage amplifies NPAs. If a highly leveraged book experiences even a minor NPA spike, the underlying equity is wiped out.

Principle 1.4: Product Economics Are Strictly Non-Fungible

Not all loans are equal. A VC or an operator must evaluate every product line separately and never accept "blended" metrics. Product archetypes are governed by the interplay of Ticket Size × Tenure × Yield. These variables create distinct operational machinery.

Industry Examples

• Kisht (Digital Unsecured): Small-ticket unsecured loans (ATS ~₹25k), very short tenure, hyper-high yield, and massive tech intensity. Focuses on mass-market digital acquisition, requiring heavy AI/ML underwriting to keep credit costs manageable at scale.
• HDFC Bank Auto Loans (8.15–8.35% IRR): Low yield but massive scale (26% market share), low credit cost because the car is collateral. ROA ~1.9% but enormous AUM.
• SBI Cards (30–40% IRR on revolving balances): Highest yield product in Indian banking, but credit cost 6–7% and opex 4–5%. The ROA (3.9%) looks lower than you'd expect from 36% yields because the loss and opex are brutal.
• Shriram Finance (Used Commercial Vehicles): Higher yields (14–16%) in underserved segments, longer tenures (5–7 years), higher operational intensity (field collections) — but commanding economics because nobody else wants to serve this segment.
• Used Car Financing (India): USD 8.85B (FY24) → projected USD 26.59B (FY33) at 13% CAGR. Financing penetration moved from 75% to 80%. Different product economics than new car loans entirely.