---
name: revacc-contribution-framing
description: Use when results exist but the "so what for accounting" is thin for a Review of Accounting Studies (RAST) manuscript — turning estimates or propositions into one sharp, defensible accounting contribution. Sharpens the claim; it does not position against the literature (revacc-literature-positioning) or polish the prose (revacc-writing-style).
---

# Contribution Framing (revacc-contribution-framing)

## When to trigger

- The results are in but you cannot state the contribution in one sentence
- The intro lists what you *did* rather than what the field *learns*
- A referee says "interesting, but what is the takeaway for accounting?"
- An analytical paper has proven propositions but no crisp accounting implication
- The contribution is overclaimed relative to what the design or model actually shows

## State the contribution as what accounting learns

A RAST contribution is not "we run X regression" or "we prove Proposition 1." It is a sentence of the form: **"We show that [accounting construct/mechanism] [does what] through [friction/channel], which means [implication for reporting, standard-setting, investors, auditors, or theory]."** RAST's fast first-round-decision culture means the editor reads the contribution claim hard on the first pass; an underframed or overclaimed contribution is a leading reject reason.

Calibrate the claim to the evidence:

- **Empirical archival.** The contribution is a *credibly identified* accounting effect plus the channel evidence. Do not promise causality your design cannot deliver; if the cleanest claim is an association under stated assumptions, say so and make the association informative.
- **Analytical.** The contribution is the *insight* the model delivers — an optimal-disclosure result, a rationalized puzzle, a comparative static with a reporting reading — not the algebra. Translate the proposition into the accounting decision it informs.
- **Experimental.** The contribution is the *channel* the manipulation isolates that archival data cannot, and what it implies for how the accounting information is processed or produced.

## The four contribution registers RAST rewards

| Register | What it offers | Watch out for |
|----------|----------------|---------------|
| **Empirical regularity** | A new, credibly identified fact about reporting/markets | Over-claiming causality; thin channel |
| **Theory ↔ evidence bridge** | A model that explains data, or a test of a disclosure-theory prediction | Loose fit between the model and the test |
| **Construct / measurement** | A better measure of an accounting construct, validated | A measure with no question behind it |
| **Mechanism / channel** | Why an established effect occurs | A channel the design cannot distinguish |

The **theory↔evidence bridge** is RAST's sweet spot — lean on it when the paper supports it.

## Sharpen and bound the claim

1. Write the one-sentence contribution; cut every clause not supported by a table or a proof.
2. State the **scope condition** — the setting, sample, or assumptions under which it holds.
3. Name what the paper does *not* show, pre-empting the obvious referee objection.
4. Connect the contribution to a decision: what should a standard-setter, investor, auditor, or theorist do differently?

## Checklist

- [ ] The contribution is one sentence: construct → effect → channel → implication
- [ ] The claim is calibrated to the evidence (no causality beyond the design; no insight beyond the proof)
- [ ] The scope condition is stated explicitly
- [ ] What the paper does *not* show is named
- [ ] The contribution names a decision or debate it changes
- [ ] The theory↔evidence bridge is used where the paper supports it

## Anti-patterns

- **Activity for contribution:** listing what you did instead of what is learned.
- **Causality inflation:** claiming a causal effect a DiD/IV cannot support.
- **Insight inflation:** selling algebra as if the equilibrium itself were the contribution.
- **Unbounded claim:** no scope condition, so a referee finds the boundary for you.
- **Orphan result:** a finding connected to no decision, debate, or theory.

## Output format

```text
【One-sentence contribution】[construct] → [effect] via [channel] ⇒ [implication]
【Register】regularity / theory-evidence bridge / measurement / mechanism
【Calibration】claim ≤ evidence? (causal / associational / model-insight)
【Scope condition】where it holds
【Does NOT show】the honest limit
【Decision changed】standard-setter / investor / auditor / theorist
【Next skill】revacc-tables-figures
```