@inproceedings{tumpati-etal-2026-finbalance,
title = "{F}in{B}alance: A Multi-Document Accounting Reconciliation Benchmark",
author = "Tumpati, Sasank and
Agarwal, Devansh and
Kedia, Ayush and
Neekhra, Arjun and
Mandal, Murari and
Garg, Krishna and
Sinha, Yash and
Gupta, Suman and
Kumar, Dhruv",
editor = "Lee, Hung-Yi and
Dev, Sunipa and
Li, Jessy and
Elliott, Desmond",
booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {EMNLP} 2026",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.findings-emnlp.675/",
pages = "14168--14191",
ISBN = "979-8-89176-462-0",
abstract = "Existing financial-NLP benchmarks mostly evaluate prepared artifacts such as filings, tables, or extracted values. Real accounting begins earlier: source documents must be reconciled into cited journal entries, aggregated into a balance sheet, and checked for contradictions. We introduce FINBALANCE, a multi-document accounting reconciliation benchmark built from source-document bundles across eight industries, three period types, and five difficulty levels. Human-authored business scenarios, accounting policies, tax/FX treatments, document schemas, distractors, and inconsistency templates are composed by a deterministic generator whose ledger produces journal entries, balance sheets, and 23 inconsistency-code labels. On a 710-record evaluation split, six contemporary large language models (LLMs) reach at most 46{\%} exact final-balance-sheet accuracy. For four models, balance sheets reconstructed from their journal entries score 26{--}41 percentage points higher than their own reported balance sheets. Models often recover numerically plausible entries but fail to bind them to supporting documents and aggregate them consistently. Citation-pressure prompting barely changes document-linking errors, while ledger-feedback ablations substantially improve reported balance sheets and expose inconsistency-detection trade-offs. Expert finance reviewers validate the benchmark design and labels."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="tumpati-etal-2026-finbalance">
<titleInfo>
<title>FinBalance: A Multi-Document Accounting Reconciliation Benchmark</title>
</titleInfo>
<name type="personal">
<namePart type="given">Sasank</namePart>
<namePart type="family">Tumpati</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Devansh</namePart>
<namePart type="family">Agarwal</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ayush</namePart>
<namePart type="family">Kedia</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Arjun</namePart>
<namePart type="family">Neekhra</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Murari</namePart>
<namePart type="family">Mandal</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Krishna</namePart>
<namePart type="family">Garg</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yash</namePart>
<namePart type="family">Sinha</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Suman</namePart>
<namePart type="family">Gupta</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Dhruv</namePart>
<namePart type="family">Kumar</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-10</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: EMNLP 2026</title>
</titleInfo>
<name type="personal">
<namePart type="given">Hung-Yi</namePart>
<namePart type="family">Lee</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sunipa</namePart>
<namePart type="family">Dev</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jessy</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Desmond</namePart>
<namePart type="family">Elliott</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Budapest, Hungary</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-462-0</identifier>
</relatedItem>
<abstract>Existing financial-NLP benchmarks mostly evaluate prepared artifacts such as filings, tables, or extracted values. Real accounting begins earlier: source documents must be reconciled into cited journal entries, aggregated into a balance sheet, and checked for contradictions. We introduce FINBALANCE, a multi-document accounting reconciliation benchmark built from source-document bundles across eight industries, three period types, and five difficulty levels. Human-authored business scenarios, accounting policies, tax/FX treatments, document schemas, distractors, and inconsistency templates are composed by a deterministic generator whose ledger produces journal entries, balance sheets, and 23 inconsistency-code labels. On a 710-record evaluation split, six contemporary large language models (LLMs) reach at most 46% exact final-balance-sheet accuracy. For four models, balance sheets reconstructed from their journal entries score 26–41 percentage points higher than their own reported balance sheets. Models often recover numerically plausible entries but fail to bind them to supporting documents and aggregate them consistently. Citation-pressure prompting barely changes document-linking errors, while ledger-feedback ablations substantially improve reported balance sheets and expose inconsistency-detection trade-offs. Expert finance reviewers validate the benchmark design and labels.</abstract>
<identifier type="citekey">tumpati-etal-2026-finbalance</identifier>
<location>
<url>https://aclanthology.org/2026.findings-emnlp.675/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>14168</start>
<end>14191</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T FinBalance: A Multi-Document Accounting Reconciliation Benchmark
%A Tumpati, Sasank
%A Agarwal, Devansh
%A Kedia, Ayush
%A Neekhra, Arjun
%A Mandal, Murari
%A Garg, Krishna
%A Sinha, Yash
%A Gupta, Suman
%A Kumar, Dhruv
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Findings of the Association for Computational Linguistics: EMNLP 2026
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-462-0
%F tumpati-etal-2026-finbalance
%X Existing financial-NLP benchmarks mostly evaluate prepared artifacts such as filings, tables, or extracted values. Real accounting begins earlier: source documents must be reconciled into cited journal entries, aggregated into a balance sheet, and checked for contradictions. We introduce FINBALANCE, a multi-document accounting reconciliation benchmark built from source-document bundles across eight industries, three period types, and five difficulty levels. Human-authored business scenarios, accounting policies, tax/FX treatments, document schemas, distractors, and inconsistency templates are composed by a deterministic generator whose ledger produces journal entries, balance sheets, and 23 inconsistency-code labels. On a 710-record evaluation split, six contemporary large language models (LLMs) reach at most 46% exact final-balance-sheet accuracy. For four models, balance sheets reconstructed from their journal entries score 26–41 percentage points higher than their own reported balance sheets. Models often recover numerically plausible entries but fail to bind them to supporting documents and aggregate them consistently. Citation-pressure prompting barely changes document-linking errors, while ledger-feedback ablations substantially improve reported balance sheets and expose inconsistency-detection trade-offs. Expert finance reviewers validate the benchmark design and labels.
%U https://aclanthology.org/2026.findings-emnlp.675/
%P 14168-14191
Markdown (Informal)
[FinBalance: A Multi-Document Accounting Reconciliation Benchmark](https://aclanthology.org/2026.findings-emnlp.675/) (Tumpati et al., Findings 2026)
ACL
- Sasank Tumpati, Devansh Agarwal, Ayush Kedia, Arjun Neekhra, Murari Mandal, Krishna Garg, Yash Sinha, Suman Gupta, and Dhruv Kumar. 2026. FinBalance: A Multi-Document Accounting Reconciliation Benchmark. In Findings of the Association for Computational Linguistics: EMNLP 2026, pages 14168–14191, Budapest, Hungary. Association for Computational Linguistics.