Portfolio management in lending: risk, funding and profit
You have the risk pack open when the monthly portfolio review turns to what is changing around the book. Treasury says the funding line comes up for renewal next quarter. The product team says the first-loan cohort from the new paid channel is repaying on time. Behind them, the vintages, roll rates and cure rates are within range, and the provision is on the page. It is a routine meeting, but the discussion has moved from recorded portfolio performance to a funding decision and a new customer channel, without yet putting either into the same account of profit.
Use risk measures to decide where to look next
Portfolio management in lending starts with the outstanding book, and the risk pack tells you where losses may emerge. Vintage curves separate origination quality from loan age, so a change in underwriting or channel can be read against the right cohort.
Roll rates show whether accounts are moving into worse delinquency buckets. Cure and recovery rates show what returns to current and what comes back after default. Those movements affect expected loss, then provisions, earnings and capital.
| measure | what it shows | what it changes |
|---|---|---|
| vintage curves | origination quality apart from age | underwriting and channel policy |
| roll rates | movement between delinquency buckets | underwriting, servicing, collections |
| cure rates | delinquent accounts returning to current | collections intensity, restructuring |
| recovery rates | value back from defaulted accounts | collection strategy, charge-off view |
| provisions and expected loss | risk turned into reserves | earnings and capital |
| cash timing | when the money arrives | liquidity and funding |
| collections staffing | capacity by delinquency stage | headcount by bucket |
| portfolio score | quality of the whole pool | securitization and investor reporting |
| stress test | the book under rate, funding and downturn shocks | limits and capital |
| concentration | risk clustered in one segment, region, industry or channel | limits and pricing |
Use the measures to decide where underwriting, servicing or collections need attention, and what the change could mean for cash and capital. Keep collections capacity aligned with expected delinquency inflows, because understaffing can let accounts age even when underwriting holds. Review concentration too. A loss pattern clustered in one channel, segment, region or industry can make the rest of the book look safer than it is.
A pack within range cannot answer either commercial question. Use the cohort calculation to test whether the new channel earns back its acquisition cost and whether the funding renewal leaves enough margin on loans already on the book.
Calculate what the customer contributes after acquisition
Assess the paid-channel cohort by its expected profit over the customer relationship, alongside its first-payment performance. The generated example starts with a first loan that earns 255 in interest.
Funding costs 51, expected loss costs 75, and servicing costs 40 in that example. Those costs leave a margin of 89 before acquisition cost. An acquisition cost of 90 makes the first loan a loss of 1.
The example assumes the second product contributes an expected margin of 196 after the first loan's 12 months, with no discounting and no new acquisition cost. At a 40% cross-sell probability, the total becomes 77 per acquired customer. For a customer acquired for 150, break-even requires a 31% cross-sell probability.
| cross-sell probability | acquired for 90 | acquired for 150 |
|---|---|---|
| 0% | -1 | -61 |
| 10% | 19 | -41 |
| 20% | 38 | -22 |
| 30% | 58 | -2 |
| 40% | 77 | 17 |
Treat these figures as assumptions in a generated example. The calculation is interest and fee income, less funding, expected loss, servicing and acquisition cost, then the expected margin from a later product. Reaching an existing customer costs less than acquiring a new one, so that difference belongs in the price of a second product. Account for its funding and risk costs too, as in risk-based pricing for an existing customer.
The cross-sell calculation answers what the paid-channel cohort is worth, and the funding-renewal example below shows what renewal does to loans already on the book. On your own book, price the cohort using its funding cost, then test what a renewal would change for the loans and customers still outstanding.
Give the next product a real path
Build a cross-sell plan by tracking repeat borrowing and second-product take-up by cohort and channel, then check whether the customers you acquire actually reach the next offer.
Set out who qualifies after the first loan, which product they can take and when the offer is made. Compare take-up and customer profit across channels. A channel with a higher acquisition cost works when enough customers qualify and take the next product. Count the expected margin when a clear offer reaches them.
Keep a channel or first product in the mix when its cohort's observed take-up reaches the break-even cross-sell probability for its acquisition cost.
Treat repayment service as part of cash timing
Make repayment straightforward and send a reminder before the due date. On-time payments limit your own cashflow problems by bringing cash in closer to when the book expects it.
In a randomized field experiment with a micro lender in Uganda, text reminders before payment raised the probability of paying on time and reduced average days late. The Cadena and Schoar study found that every treatment raised the probability of paying on time by 7 to 9% and cut days late by 2 a month. A text reminder did as much as a 25% interest rate reduction.
Track on-time payment and days late by cohort and channel. That connects the repayment experience to the cash timing in your own book and shows where service or payment arrangements need attention.
Test the funding renewal against the loans already on the book
In the generated example, a 12-month funding line reprices 4 percentage points higher while fixed-rate loans run for 36 months. Interest on the old loans stays the same, funding cost doubles, and year-two profit turns to a loss.
| funding | year 1 | year 2 | year 3 | ROE, years 2–3 |
|---|---|---|---|---|
| short, rates flat | 172,000 | 110,000 | 41,000 | 16.5% |
| short, rates up 4 points | 172,000 | -86,000 | -32,000 | -12.8% |
| matched, rates up 4 points | 134,000 | 86,000 | 32,000 | 12.8% |
Interest income in year two stays at 668,000, but the funding cost rises from 196,000 to 392,000. The margin falls, and return on equity drops below the example’s 12% cost of equity. Matched funding costs more in a calm year, but in the rate-rise case it keeps return on equity at 12.8%, above that hurdle.
For the renewal, identify the loans that will still be outstanding when the line reprices, and recalculate their funding cost with the higher rate. Carry that change into profit and return on equity, then test tighter funding or harder refinancing against the rate pressure the Federal Reserve describes in its review of Silicon Valley Bank’s short-funded, long-dated securities portfolio.
Choose the product mix against the cost of equity
Keep a product in the mix only if its return on equity clears the cost of equity under the funding conditions the book may face. Compare each product’s return on equity with the cost of equity to decide whether its return justifies the capital.
In the generated example, matched funding gives up return when rates stay flat. When rates rise, it keeps return on equity above the cost of equity, while the short-funded book falls below it. That is the choice: accept a lower return in the calm case, or take more exposure to a repricing the old loans cannot pass through to customers.
A realistic next-product path adds margin to a first loan.
Run the next portfolio review across risk and profit
At the next monthly review, bring the paid-channel cohort’s contribution after acquisition and the renewal’s effect on profit on loans still outstanding, alongside the risk pack, so the team can judge them in one decision.
- Review risk by cohort and segment each month. Read vintages, roll rates, cure rates, recoveries, provisions, cash timing, collections staffing and concentration together. Use each measure to identify where underwriting, servicing, collections or capital needs attention.
- Calculate profit per acquired customer. Add interest and fee income, then subtract funding, expected loss and servicing costs to get the first product’s margin. Subtract acquisition cost, then add the expected margin of a later product multiplied by the observed probability that the customer takes it.
- Test acquisition cost against repeat value. Track repeat-loan or second-product take-up by cohort and channel, and compare the customer contribution with what it cost to acquire that cohort.
- Define the next-product path. Record who qualifies after the first loan, which offer they receive and when it is made. Check take-up against the customers who were eligible.
- Make repayment easy to complete. Send a reminder before each due date, then track on-time payment and days late by cohort.
- Recalculate the book at renewal. Identify which loans remain outstanding, apply the higher funding cost to them and carry the change into profit and return on equity. Test tighter funding and harder refinancing as well.
- Set the product mix against the capital hurdle. Compare each product’s return on equity with the cost of equity under the renewal and stress cases, then shape the mix around the products that clear it.
Questions about portfolio management in lending
What is portfolio management in lending?
It is oversight of the outstanding book across credit risk, cash, funding and profitability over the customer relationship. Review customer economics alongside risk performance.
Which metrics should a lender track in portfolio management?
Track vintage curves, roll rates, cure rates, recovery rates, provisions and expected loss, cash timing, collections staffing, stress-test results and concentration. Review them monthly by cohort and segment, and add a whole-pool score where securitization requires one.
How should a lender match funding to the loans in its book?
Use funding with a term that fits the loans it supports, and test the effect of repricing short-term funding when existing loans keep their rates. Recalculate profit and return on equity for the loans still outstanding.
Why can a lender lose money on a first loan and still grow profitably?
A later product can add margin without a new acquisition cost if enough customers take it and there is a real path to the offer. Compare expected customer contribution with acquisition cost by cohort and channel.
How does ease of repayment affect a loan portfolio?
It supports on-time payment and helps cash arrive when expected. Make repayment straightforward, send reminders before due dates and monitor payment timing by cohort.
How often should a consumer lender review portfolio performance?
Review it monthly by cohort and segment. Use the review to catch changes in risk, cash timing and customer economics.
Which scenarios should a lender include in a portfolio stress test?
Test rate increases, tighter funding or harder refinancing, a downturn, falling collateral values and a sector shock. Measure the effects on delinquency, losses, recoveries, provisions, cash flow and capital.
