Evidentiary Asymmetry and a Warning from the Horizon Case: Analysing the Bankers’ Books Evidence Bill 2026

[Sai and Harinandana are students at Tamil Nadu National Law University and Indian Institute of Management, Rohtak, respectively.]

On 10 August 2026, the Rajya Sabha passed the Bankers’ Books Evidence Bill 2026 (Bill), which seeks to repeal and replace the 135-year-old Bankers' Books Evidence Act 1891. This long overdue Bill has outlined a new procedure for the certification of banking-related documents. It also touches on other realms such as the applicability to other financial sector entities and arbitration proceedings, thus having far-reaching commercial ramifications. 

In this article, we argue that the new procedure with respect to the conditions for certified copies under Sections 3, 5, 6 and 7 of the Bill holds the door wide open for evidentiary asymmetry in digital records. We shed light on its interplay with the Bharatiya Sakshya Adhiniyam 2023 (BSA), which appears to be a dilution of the evidentiary standard.

A Lack of Expertise? Analysing the Single-Certificate Standard

The most crucial shift in the Bill occurs through the, arguably simpler, scheme envisaged in Sections 3, 5, 6 and 7, which adopt a novel single-certificate architecture for the certification of copies of bankers’ books. This is in stark contrast to the high standard in place under the BSA; and we argue that this new design paves the way for evidentiary asymmetry.

Under Section 63(4) of the BSA, the certificate accompanying electronic evidence must be signed by the person in charge and an expert. The Schedule gives effect to this through two components: Part A requires certification by the custodian of the computer or system, including particulars of the system and the calculated cryptographic hash, and Part B requires independent certification by an electronic evidence expert concerning the integrity of the system and absence of tampering.

On the other hand, Section 3(3) of the Bill provides that the certificate shall be signed or authenticated by the “branch head or the office head or any other officer of the bank duly authorised” by the bank, with a standard good faith proviso. While the BSA expressly places an expert within the certification chain, Section 3(3) of the Bill does not reproduce an equivalent requirement. This is to be read with Section 6, which begins with a crucial non-obstante clause for any other law for the time being in force, with respect to the admissibility of electronic or digital records. It is to be read subject to Section 7. 

This non-obstante clause establishes a specialised evidentiary route for bankers’ books. The concern is not that Section 6 makes every banker’s record automatically conclusive. It does not. The concern is that the Bill creates a statutory route in which the bank can place electronically generated records before the adjudicatory forum through a certification architecture that does not replicate the BSA’s express requirement of certification by the custodian and an expert.

In Anvar PV v. PK Basheer, the Honourable Supreme Court treated compliance with statutory requirements with respect to electronic evidence as integral to its admissibility. This position was later reaffirmed by the Three-Judge Bench in Arjun Panditrao Khotkar v. Kailash Kushanrao Gorantyal. The court described the certificate under Section 65B(4) of the Indian Evidence Act 1872 as a condition precedent to the admissibility of electronic evidence. 

Hence, the question arises: if independent certification is required to ensure the integrity of ordinary electronic evidence, why is the same expert safeguard not required when the record is generated and produced by the financial institution itself? This marks our first point of contention.

A Warning from the Horizon Case

A good reference point for analysing the nuances of electronic evidence in banking is the position in the United Kingdom. The Post Office Horizon scandal illustrates the institutional risk in placing excessive confidence in computer-generated financial records. Between 1999 and 2015, the UK Post Office prosecuted more than 700 sub-postmasters on the basis of the accounting shortfalls generated by Fujitsu’s Horizon system. In Hamilton v. Post Office Limited (Horizon case), the Court of Appeal addressed the reliability problems associated with Horizon and the evidentiary difficulties faced by individuals attempting to challenge records generated by a system whose underlying operation was not within their access or control.

These cases relied on the common law presumption that a computer was functioning properly unless there was evidence to the contrary. Singapore follows a similar approach through the statutory presumptions under Section 116A of the Singapore Evidence Act 1893. The BSA, however, takes a different route by requiring dual certification, whereas the Bill provides for a single-certification model for bankers’ books. 

We agree that a comparison to the Horizon case should be limited; the Bill does not create an irrebuttable presumption that banking systems are infallible. The case, however, serves as a warning sign. The takeaway for policymakers should be that where the bank controls the system, produces the record and certifies its integrity, parties on the other side should be able to test the reliability of that record. This is where the purpose of a high evidentiary standard such as expert certification and hash values, inter alia, comes in.

Evidentiary Asymmetry in a Wider Context

As per Section 5 of the Bill, a certified copy of an entry in a banker’s book is “received as prima facie evidence” of the existence of that entry or information. Where the record is generated through a distributed system and certified by a bank officer without the independent expert certification contemplated under Section 63(4) of the BSA, a customer, employee or any other litigant may have little means to challenge its reliability, particularly where the underlying system is not accessible to them. This creates an issue of evidentiary asymmetry, much like what arose in the Horizon case.

This is not just in the context of banks. During the debate in the Rajya Sabha, the finance minister hailed Section 4 of the Bill for giving the government the power to extend the Act to any other entity or class of entities operating in the financial sector, to ensure that a uniform evidentiary framework is available across the financial sector. 

This marks our second point of contention. While uniformity is much desired in India’s vast legal landscape, one must critically assess the implications of extending this single-certificate standard to other financial entities. With the emerging fintech sector in India, the ambit of who could be included under this Bill stands wide. Section 4 also allows the Central Government to prescribe “conditions, exceptions or modifications” by notification. The scope of this delegated power requires closer legislative scrutiny. 

Thirdly, these concerns also extend into the realm of arbitration. Section 2(1)(f)(ii) of the Bill brings arbitrations within the definition of “legal proceeding” while the BSA explicitly excludes arbitrations. This inclusion is not without commercial justification. Banking facility agreements, consortium lending arrangements, derivatives and commercial recovery disputes frequently contain arbitration clauses. 

However, Section 5 of the Bill, which gives prima facie status to certified copies of any entry in a banker’s book, raises a concern. This section provides a procedural head start for parties producing certified copies. Parties on the other side must now be vigilant, as such prima facie status shifts the burden of proof onto the respondent. While the Arbitral Tribunal retains the discretion to determine the weight of evidence under Section 19(2) of the Arbitration and Conciliation Act, 1996, such interplay with Section 5 of the Bill may have significant ramifications in practice. Tribunals, particularly ad hoc ones, may be less inclined to look beyond a statutorily privileged document.

Conclusion: Modernisation without Dilution

Does the need for simpler evidentiary rules for digital bankers’ books stand before the significance of a foolproof evidentiary standard for electronic records? The issue at hand is not whether such digital banking records should be recognised, but rather if there are adequate safeguards to accompany their admissibility as evidence. That is, if there are mechanisms to prevent evidentiary asymmetry.

Modern banking infrastructure may involve core banking systems, payment gateways, APIs, cloud servers, edge infrastructure and offsite disaster-recovery systems. A failure may occur at one of these points. Any synchronisation delay or a mismatch between distributed ledgers is, therefore, capable of affecting the evidentiary reliability of a particular transaction. The key concern is whether the integrity of such records is being independently verified when the bank itself generates, maintains and certifies them.

In our view, policymakers need to focus on two issues as the new framework comes into force. First, the evidentiary standard under Section 3(3) of the Bill needs additional safeguards. We propose that for electronic copies, the digital signature should be made mandatory to ensure cryptographic verifiability by way of hash values and to prevent tampering of electronic copies. An electronic signature does not offer the same level of security. Mandating a digital signature may be a more financially and commercially practical way of introducing verifiability. Second, Section 4 should set clearer statutory conditions and provide for legislative scrutiny before the regime is extended to other entities or classes in the financial sector.

The ultimate aim is to ensure that when banking records become more technically complex, the law does not respond by making their certification less independently verifiable. The Bill modernises the form of bankers’ books. Its certification architecture must also be modernised without diluting the evidentiary discipline that gives these records their legal weight. This will prevent any occurrence of evidentiary asymmetry and ensure fairer dispute resolution.