Mortgage math breaks on a seventy-two-hour loan.

Put a three-day digital working-capital loan in front of a traditional economist, a parliamentary committee or an overseas ESG analyst and watch the first calculation they perform on the back of an envelope. They take a 6 or 10 percent transaction fee on a 72-hour facility, multiply it by 120 three-day cycles in a year, and announce in horror that the borrower is paying an annual percentage rate of 720 or 1,200 percent. Judged by the yardstick of a 25-year salaried mortgage in London or a 12-month corporate overdraft in Nairobi, a 10 percent fee for three days of liquidity looks like extortion, and the resulting headline writes itself.

Step out of the seminar room and stand beside a vegetable wholesaler in Wakulima Market or a boda boda rider in Kisumu at 5.30 am. Neither trader is borrowing for 365 consecutive days to buy a house. Both are financing a 48-to-72-hour inventory or fuel cycle that yields a 40 to 100 percent gross margin before the weekend. Applying annualized mortgage mathematics to ultra-short-term informal working capital is like taking the nightly rate of a Nairobi hotel room, multiplying it by 365 and declaring that travelers pay KSh9 million a year for shelter. The error matters beyond rhetoric, because policy built on the wrong metric strangles the productive credit while leaving the real abuses untouched.

Walking through the seventy-two-hour return on working-capital velocity.

To see why a rational informal entrepreneur willingly pays a flat short-term fee, and when that fee crosses from productive fuel into destructive debt, run the shilling cash book of a three-day trade cycle.

Suppose a market trader has zero liquid cash on Thursday morning, yet knows that KSh2,000 of fresh tomatoes or petrol will generate KSh3,600 in gross receipts across Thursday, Friday, and Saturday trade. She takes a KSh2,000 three-day facility priced at a flat 8 per cent fee of KSh160, repays KSh2,160 on Saturday evening, and walks home with KSh1,440 in net cash profit she could not have earned without the bridge. Measured against her gross profit of KSh1,600, the financing consumed 10 percent of the upside while leaving her family KSh1,440 richer. Annualising that KSh160 into a three-digit percentage tells the credit committee nothing about whether the loan made the borrower richer or poorer. The metric that governs informal business survival is the return on working-capital velocity: the net shilling cash retained by the trader, after cost of goods, operating expenses and the exact shilling credit fee, across a single completed stock turn. Lend against the turn, measure the turn, and the product either feeds the business or it does not.

Where the short-term model fails: consumption leakage and multi-app stacking.

Respecting the mathematics of high-velocity working capital does not hand high-cost digital credit a free pass, because short-term liquidity turns toxic the moment either of two boundary conditions breaks.

The first is purpose. When the KSh2,000 facility is used for non-productive consumption or to plug a chronic household income deficit in an economy where 83.8 percent of workers are informal and fewer than one in five adults is financially healthy, there is no Saturday inventory profit to repay the KSh2,160; there is only a deeper hole. The second is stacking. If the borrower rolls the three-day loan over fifty times a year across ten competing apps, the hypothetical annualized nightmare becomes her literal reality, cycle by cycle.

Two 2026 guardrails change that equation for Kenyan credit. The industry's unified single-portal credit information sharing hub now gives underwriters a real-time view across banks, SACCOs and credit providers, so a borrower cannot quietly stack dozens of concurrent short-term loans. And the Central Bank of Kenya's Non-Deposit-Taking Credit Providers Regulations, 2026, gazetted on 29 September 2026 under Legal Notice No. 191, codify a strict in duplum cap ensuring interest on a non-performing loan can never exceed the principal, together with the statutory right to repay early, in full or in part, at any time without penalty, immediately extinguishing interest for the remaining period. The rules, in other words, police the two genuine failure modes, leakage and stacking, instead of outlawing the velocity that makes the product work.

Replacing armchair algorithms with field verification.

Pure smartphone-scraping models cannot tell whether a borrower is buying KSh5,000 of fast-moving hardware stock or betting on weekend football, just as chatbots cannot verify a market stall's inventory and automated rollover scripts only deepen distress. The lenders sustaining the strongest repayment rates in the informal sector pair algorithmic scoring with physical field discipline: tablet-equipped officers who verify the trader's stall, stock turn and supplier receipts before unlocking a working-capital line, backed by hard system locks that block automatic rollovers after two consecutive extensions. The algorithm prices the loan; the officer confirms the business exists; the system refuses to let either of them pretend a distressed borrower is a repeat customer.

Before submitting updated pricing parameters to the Central Bank of Kenya under the new rules ahead of the 29 March 2027 deadline, every chief credit officer and chief product officer should re-anchor product suitability around verified inventory turnover cycles: build pro-rata daily fee rebates into the core ledger for borrowers who settle early, cap consecutive renewals at two cycles before a mandatory cooling-off and cash-flow review, and measure portfolio health by the net shilling margin left in the merchant's till after repayment. When your credit committee reviews short-term SME products in 2027, will it be measuring whether the borrower's working capital grew, or just whether she borrowed again to pay you back?