Rejected transfers mean food bought on credit and medicine postponed. This article proposes five numbers every agency should publish, a single complaint channel and donor financing tied to service quality, so that digital welfare payments are judged from the recipient's side.

Judge digital welfare payments by how many fail

Analysis

Digital welfare payments are not delivered when a ministry sends an instruction to a bank. They are delivered when the right person can use the money. Between those two moments, a transfer can be rejected, returned, delayed by an identity mismatch or left inaccessible at the point of withdrawal. Yet governments and donors still tend to celebrate the number of accounts opened, beneficiaries registered or payments initiated. Those are measures of administrative activity, not reliable service.

South Asian governments are right to digitise social assistance. Electronic payments can reduce handling costs, shorten queues and make it harder for cash to disappear before reaching a household. A 2026 World Bank stocktake examined 383 social-assistance programmes in 123 countries and found a wide range of delivery arrangements, including systems that still combine digital and manual steps. That variety is a warning against treating the word “digital” as a guarantee of completion.

What digital welfare payments should report

The missing public measure is a payment-failure rate. For every benefit cycle, an agency should publish five numbers. They are payments due, payments attempted, payments completed by the deadline, payments rejected or returned and failed payments corrected within a stated period. The figures should be broken down by district, payment provider and failure category without exposing anyone’s identity.

This is not a technical detail for a back-office report. A rejected transfer can mean food bought on credit, medicine postponed or a fare unavailable on the day it is needed. A national success rate can also conceal places where residents travel farther to reach an agent, face poorer connectivity or encounter more frequent biometric and identity errors. Publishing local results would show whether digitisation is redistributing inconvenience onto people with the least ability to absorb it.

A useful dashboard must separate different failures. An eligibility decision is not the same as a bank rejection. A duplicate record is not the same as an agent running out of cash. A payment that reaches an account but cannot be withdrawn should not be recorded as equivalent to money used by the recipient. Clear categories allow the responsible institution to fix the problem instead of sending a person from a local office to a bank and back again.

Nepal and unfinished last mile

Nepal illustrates both the promise and the unfinished work. The World Bank’s account of Nepal’s digitised civil registration and cash transfers says social-security allowances moved into bank accounts for millions of beneficiaries, improving efficiency and reducing leakage risks. It also identifies continuing challenges in connectivity, staffing and access to banks. The sensible response is not to retreat to cash. It is to measure the last mile with the same seriousness used to measure enrolment.

Complaints, donors and accountability

The same distinction should shape complaints. One free entry point, available in person, by telephone and through a simple digital channel, should accept every payment complaint. The agency can route the case internally, but it should not require the recipient to diagnose which institution failed. The World Bank’s guidance on government-to-person payments stresses clear and responsive grievance redress, appeal mechanisms for errors and payment-reconciliation information for programme administrators. Those safeguards should produce public service standards, namely time to acknowledge, time to resolve and the share of cases reopened.

Donors have leverage here. A project milestone should not be satisfied merely because a database was built or a target number of accounts was created. Part of the financing should depend on measurable service quality, including timely completion, correction of rejected payments, accessible complaint channels and declining disparities between districts. The World Bank’s own delivery-systems overview recognises multi-channel grievance redress and results linked to coverage and service quality. Payment reliability belongs in that definition of quality.

Publishing failures without publishing people

Officials may worry that publishing failures will invite criticism or create misleading league tables among banks. That risk can be managed with a standard method, explanatory notes and independent sample audits. Secrecy creates a worse incentive, because a system can look successful as long as it records the outgoing instruction and does not count the person still waiting.

Providers should also be allowed to correct records before the next public release, but not to erase the original failure. A repaired payment is good service recovery; it is not proof that the failure never happened. Showing both the initial error rate and the correction time would reward providers that respond quickly while preserving the evidence needed to prevent recurrence.

None of this requires publishing names, account numbers or benefit amounts. Aggregated operational data can protect privacy while making performance visible. Nor should a dashboard replace direct assistance. Its purpose is to reveal patterns that individual complaints cannot, such as a district where rejections spike, a provider whose corrections take weeks or a payment category repeatedly trapped by the same data mismatch.

Judging digital welfare payments from recipient’s side

Digital social protection should be judged from the recipient’s side of the transaction. Registration matters. Fraud controls matter. But the decisive question is more ordinary. Did the money arrive in usable form and on time, and was a failure fixed without making the person navigate the machinery of the state? Until governments and donors publish that answer, claims of successful digitisation remain incomplete.

Swaansh Mahat is an independent writer and English-language educator based in Kathmandu, Nepal. He writes about education, public services and everyday institutional life.