<aside> 🌐

Parent: Sequential Provenance Comparison (SPC) | Public Page

Language: English ｜ 会計整合的な縮約は政策推論を支えられるか｜公開論文ページ

Published: October 4, 2026

</aside>

<aside> 🔎

Balanced accounts do not establish that a policy path can be executed.

SFC models maintain consistency between transaction flows and changes in assets and liabilities. Yet reducing detailed institutional states to a smaller set of variables can preserve those numerical relationships while losing the information that determines whether the resulting payments are permitted.

The question is not whether every detail of reality must be restored. It is whether the reduction discards information governing the admissibility and updates of operations required by the model’s own calculation of policy effects.

Deposit-use eligibility is a minimal witness, not the limit of the problem. The paper’s central proposal is to audit the institutional feasibility of policy paths independently of their accounting consistency.

</aside>

Can Accounting-Consistent Reduction Support Policy Inference? Preserving Institutional Feasibility in SFC Models: A Minimal Counterexample

会計整合的な縮約は政策推論を支えられるか――SFCモデルにおける制度的実行可能性の保存と最小反例

<aside> 📄

Author: Nyun ｜ Version: 1.0 ｜ Status: Working Paper / pre-peer-review preprint

Language versions: English (15 pages) and Japanese (13 pages), representing the same paper.

Zenodo: public record ｜ DOI: 10.5281/zenodo.23127811

</aside>

The Main Points in 30 Seconds

• Starting distinction: numerical accounting consistency and institutional permission are separate requirements.
• Minimal counterexample: two loans of 20 from a common initial state produce the same aggregate deposits and borrowing, but different use conditions permit the same next payment in one state and prohibit it in the other.
• A further problem: a reduction can reproduce current admissibility exactly yet lose information governing a review update, causing subsequent paths to diverge.
• Connection to policy predictions: a closed one-period model yields realized expenditure of 20 or zero after quantitatively equal credit formation. A total-deposit-based prediction overstates expenditure by 20 under the restricted contract.
• Repair: retain the required eligibility information and update rules, explicitly limit the domain of application, or report uncertainty.
• Object of criticism: the reduction adopted for policy inference, not SFC accounting principles.

What Is at Stake in Reduction?

Reduction summarizes detailed states using variables selected for analysis. Aggregating balances, consolidating agents, classifying instruments, and combining transactions into period flows are examples. Reduction is necessary for tractability; the issue is whether discarded distinctions are irrelevant to the analytical purpose.

Matching assets and liabilities does not necessarily reveal whether funds are usable for a target payment or what information is needed to authorize release. A calculation proceeding from policy intervention to credit formation, payments, expenditure, income, and prices presupposes that the required operations can occur.

If omitted information changes those execution conditions, internally balanced accounts cannot by themselves justify the policy path.

Identical Accounting, Different Permission for the Next Payment

The first construction contains a household, a firm, and a bank. One institution permits both a contract forming deposits usable for the target payment and a contract forming deposits unusable until release. Loan amounts, interest rates, repayment terms, and other execution conditions are identical.

Both states are reachable from a common zero-balance initial state, with identical numerical formation flows. Restricted deposits remain the firm’s assets and the bank’s liabilities; no alternative ownership is introduced.