Jfm Rebuttal

Rebuttal Strategy (jfm-rebuttal)

When to trigger

• A JFM decision letter (R&R or reject-and-resubmit) has arrived and a response letter is due
• The referee asks for new microstructure analysis (alternative measures, a placebo, a subsample) and you must plan it
• You disagree with a comment and need to push back without antagonizing the referee
• Comments conflict across referees and the editor must be asked to adjudicate
• The revision risks ballooning and you need to keep it inside the editor's real decision rule

Structuring the JFM response

The response letter is read by busy specialists; make compliance visible and verifiable. The governing principle: do the work, then make it effortless to see you did it.

1. Open with a short summary of the main changes and the headline result's stability ("the X-bps effect is unchanged after the requested checks").
2. Reproduce every comment verbatim, numbered by referee, each followed by a concise response and a pointer to the exact revised location ("new Table 5, p. 14; Internet Appendix Table IA.3").
3. Lead with the fatal-if-true comments the editor flagged (from jfm-referee-strategy), then the rest in order.
4. For new analysis, state what you ran, the result, and whether the headline survives — show the number, not just "we addressed this."
5. For push-backs, concede the framing, then give the microstructure evidence. "The referee is right to worry about endogeneity; the tick-size shock is exogenous to firm fundamentals because …" beats a flat refusal.
6. For conflicting referees, name the conflict and ask the editor for guidance rather than silently siding with one.

Microstructure-specific moves that win R&Rs

• Measurement-artifact comment → add the alternative liquidity/impact construct and show the effect holds; this is the single most common JFM ask.
• Endogeneity comment → strengthen the design (find/justify the exogenous market-structure variation) rather than adding controls.
• Inference comment → re-report with two-way clustering / NW and state it; usually cheap and fully closes the point.
• "Not microstructure enough" → sharpen the mechanism framing and add the heterogeneity exhibit showing the effect where the trading-process theory predicts.
• Generalizability → add an asset class / venue / period, or scope the claim honestly; do not over-promise.

Documenting result stability across the revision

Because most JFM revisions add measurement and inference checks, the headline number often shifts slightly. Get ahead of this: maintain a small table (for yourself, and summarized in the letter) showing the headline coefficient under the original and each newly requested specification, so the reader sees the result is stable rather than wondering why a number changed between rounds. If a check does move the estimate materially, explain why in microstructure terms rather than burying it — a transparent "the effect attenuates to X bps once we use realized spreads, consistent with [reason]" is far stronger than a silent change a referee notices.

Worked response entry (illustrative)

Referee 2, Comment 3: "The spread effect may simply reflect the alternative liquidity measure chosen."

Response: We thank the referee for raising this. We now re-estimate the main result using quoted, effective, and realized spreads and the Amihud measure (new Table 4, columns 2-5; full battery in Internet Appendix Table IA.3). The effect ranges from 9 to 14 bps across all four constructs and remains significant at the 1% level, so it is not an artifact of any single measure. We added a sentence to Section 4.2 (p. 15) noting this.

This entry reproduces the comment, names the new analysis, gives the actual numbers and their stability, and points to exact locations. It does not say "we addressed this" — it shows it.

Sequencing and tone

Order the letter to mirror the editor's priorities: fatal/editor-flagged comments first, then referee-by-referee. Open each referee's section by thanking them and summarizing the changes their report produced. For a genuine disagreement, never refuse flatly — concede the concern's legitimacy, then present the microstructure evidence that resolves it; if it cannot be resolved, scope the claim honestly. When two referees pull in opposite directions, state both positions and ask the editor for direction rather than silently complying with one and ignoring the other. The revised manuscript and Internet Appendix must say exactly what the letter claims they say — an editor will spot-check.