Authorised but Not Autonomous: Loss Allocation in India’s Coerced Push Payment Frauds

India allocates loss from electronic banking fraud through a Reserve Bank circular of 6 July 2017, which distributes liability across three tiers, each presupposing that the disputed transaction was unauthorised. The fastest-growing category of Indian payment fraud does not fit that premise. In frauds described as "digital arrest", the victim is persuaded by impersonators of the police or the judiciary that criminal process has been initiated against her, and transfers the money herself using her own credentials and authentication. The transaction is authorised in every technical sense, the framework therefore does not reach it, and the victim bears the whole loss by default rather than by any considered allocation. This article asks how loss should be distributed when a payment instruction is genuine but the consent behind it was procured by fraud. It argues that the circular asks who authenticated the transfer when it should ask whether consent was free, and that contract law cannot close the gap, the Indian Contract Act locating vitiated consent between victim and fraudster rather than between victim and bank. The study is doctrinal and functionally comparative, drawing on Reserve Bank instruments, decisions of the Delhi and Allahabad High Courts, orders of the Supreme Court in its suo motu digital arrest proceedings, and data placed before Parliament by the Ministry of Home Affairs. It finds that India has built substantial capacity to intercept fraudulent funds in transit while restoring close to two per cent of reported losses, and that High Courts have had to improvise the distinction between coercion and self-dealing case by case, the regulator having left the question open. It recommends a third liability category for fraudulently induced authorised transactions, with loss shared between remitting and beneficiary institutions.