Branch risk monitor
PAR and a 24 month trend for every branch, filterable by province and district, with overdue migration and concentration across the portfolio.
Risk layer for microfinance institutions
CBSRM reads an authorised CSV export from the institution's existing CBS or MIS and turns it into a month-end view of portfolio risk: PAR, a trend and early-warning alerts against agreed thresholds for every branch, overdue migration and concentration across the portfolio, and stress tests for rate, farm income, monsoon, regional disaster and funding shocks. It is read-only. It posts nothing back and changes no account, so the CBS stays the ledger of record.
The CBS keeps the books. CBSRM reads them and tells you where the risk is.
Direct API connection is pilot work, not built yet.
CBSRM does not replace the Core Banking System. It analyses authorised data from it.
Export, calibrate, review.
The institution exports an anonymised or centre-level aggregated portfolio file from its CBS/MIS in the template format (laghubitta_template.csv).
Thresholds, provisioning rates and scenarios are set together with the institution's risk team against current NRB directives.
Management and board receive the branch risk monitor, alerts, stress results and a board summary each month-end.
Three things, none of them customer-identifying.
From the CBS/MIS, in the template format. Centre-level aggregates are enough. No customer names or IDs are needed.
The handful of figures needed for capital, liquidity and funding stress.
One person who knows the export and can sit in on calibration.
The app processes the file in the browser. Nothing is uploaded.
Five outputs, every month-end.
PAR and a 24 month trend for every branch, filterable by province and district, with overdue migration and concentration across the portfolio.
Rules flag branches that cross agreed thresholds, ranked by severity, each with the value and the threshold that triggered it.
Each scenario recomputes PAR, provisions, capital and the 90 day liquidity gap.
A short month-end summary for management and the board, drawn from the same computed figures as the monitor.
SHA-256 of the input file and the parameter set in the app; for the sample dataset, a hash-chained record of every computed block from the open-source library.
A working prototype. The institution-specific work is still to do.
Try it on your own portfolio export.
Write with the CBS or MIS the institution uses and a contact in risk or MIS. The first step is a short call to walk through the template and the calibration list.
prabhawa@wavervanir.com