Synthetic Form 941 Schedule D - Report of Discrepancies Caused by Acquisitions, Statutory Mergers, or Consolidations Data
Synthetic training data — no real PII, fully coherent identities
Generate synthetic Form 941 Schedule D discrepancy reports — the schedule that explains why the wages an employer reported to the IRS do not match what it reported to the Social Security Administration after a merger, acquisition, or consolidation. Every wage line appears as an IRS amount, an SSA amount, and a signed difference.
64
Fields per document
2
Pages
Tax
Category
What this document is
Schedule D (Form 941) is filed when an employer's totals on its Form 941 filings disagree with the totals on the Forms W-2 it sent to the Social Security Administration, and that disagreement is explained by a corporate restructuring. The employer identifies itself as the predecessor or successor, or as the acquiring or surviving corporation, states which restructuring path applies, and then reconciles each wage and tax line three ways: what the IRS shows, what the SSA shows, and the difference between them.
Why generate synthetically
Three-column reconciliation documents are structurally different from ordinary returns and structurally different from amendments, because the third column is a signed difference that can be negative. Models trained only on returns have never seen a negative currency value in a tax-form context and routinely drop the sign, which converts a small overstatement into a large one. Real Schedule Ds are rare, restructuring-specific, and always name identifiable companies, so no meaningful labeled corpus exists.
What makes synthetic data useful
The IRS column and the SSA column are generated as two plausible views of the same underlying payroll, not as independent draws, and the difference column is derived from them. That means the arithmetic a reviewer checks — IRS minus SSA equals difference — actually holds on every document, so you can train and evaluate arithmetic consistency validation instead of just field extraction. The restructuring role checkboxes are drawn as coherent combinations rather than sampled independently, so you never get a document that claims to be both the predecessor and the successor.
Training challenges
The three columns are typed identically and spaced tightly, so column assignment is the whole game: read the SSA figure into the IRS slot and the extraction is wrong even though the OCR was right. The difference column carries genuine negatives — the median Social Security wage difference is −$70.34, with values running from −$36,771 to +$34,633 — and sign handling on small negative amounts adjacent to large positive ones is where most currency extractors fail. Several line pairs exist in two variants, a primary set and a secondary set for employers reporting on multiple returns, and the secondary set is legitimately zero in most documents, so a model can score deceptively well by predicting zero everywhere until it meets the minority of documents where those lines matter. Finally, eight small checkboxes encode four mutually constrained binary choices whose semantics change how the whole page should be interpreted.
Generate synthetic Form 941 Schedule D - Report of Discrepancies Caused by Acquisitions, Statutory Mergers, or Consolidations data
Start with 500 free credits. No credit card required.
Generate NowWho uses this data
M&A payroll integration teams reconciling predecessor and successor wage reporting, corporate tax and audit groups building extraction over restructuring documentation, and wage-reporting reconciliation platforms that have to detect IRS-versus-SSA mismatches automatically. It is also the sharpest available test for any extraction stack that claims to handle signed amounts and multi-column financial layouts.
Document complexity profile
64 fields across 2 pages: 30 currency amounts arranged as IRS / SSA / difference triples, 25 text fields, 8 checkbox targets, and 1 date. 94 annotation relations — half again the field count — because a single printed line label governs all three columns of its row. The binding graph is moderate: 16 conditional bindings and 24 function calls at a maximum expression depth of 2, driving the restructuring-role logic and the derivation of every difference amount from its two source columns.
Key stats from our synthetic corpus
Quantitative characteristics of the Form 941 Schedule D - Report of Discrepancies Caused by Acquisitions, Statutory Mergers, or Consolidations documents our generator produces.
| Metric | Value | Detail |
|---|---|---|
| Median Social Security wage discrepancy | −$70.34 | Across 641 eligible synthetic identities, the median difference between the IRS and SSA Social Security wage figures is −$70.34, with a p25–p75 range of −$3,766.69 to +$3,499.58 and extremes of −$36,771 and +$34,633. Small signed differences beside five-figure gross amounts are the hardest currency-extraction case on the form. |
| Median IRS Social Security wages | $34,264.10 | The IRS column reports a median of $34,264.10 in Social Security wages (p25 $20,313.71, p75 $55,679.07) against an SSA median of $34,762.20 — two independently plausible views of the same payroll, which is what makes the reconciliation non-trivial. |
| Merger versus acquisition split | 50 / 50 | 50.1% of synthetic filings check the statutory-merger-or-consolidation path and 49.9% check the acquisition-or-alternate-procedure path. The two paths change which role checkboxes are valid below them, so an even split is required to train the conditional structure. |
| Original versus corrected filings | 87% / 13% | 86.6% of documents are original Schedule D filings and 13.4% are corrected ones. The original/corrected pair sits in the header and is easy for a classifier to miss, but it determines whether the figures below supersede a prior submission. |
| Population coverage | 64% | 641 of 1,000 simulated identities are employed with an employer and therefore eligible to appear on a Schedule D, so a 1,000-identity run yields roughly 640 documents. |
How this document co-occurs with others
Rates at which identities in our corpus that produce a Form 941 Schedule D - Report of Discrepancies Caused by Acquisitions, Statutory Mergers, or Consolidations also produce other documents.
| Correlation | Rate | Detail |
|---|---|---|
| The quarterly returns being reconciled | 100% | Every synthetic Schedule D filer also files Form 941. The quarterly returns supply the IRS-side totals the schedule disputes, making the pair a genuine cross-document reconciliation exercise. |
| The SSA-side wage statements | 100% | The SSA column on Schedule D exists because of the W-2s the employer transmitted. Generating W-2s alongside gives you both sides of the discrepancy in one dataset. |
| Correction pathway | 100% | Employers who discover a reporting discrepancy often file Form 941-X to correct it. The two forms use different multi-column geometries for the same conceptual task, which makes them a strong generalization test. |
All stats above are corpus-derived: they were computed on a local synthetic corpus of 1,000 generated identities produced by SymageDocs' World Simulation Engine. No real employer, SSA, or payroll data was used. `make corpus-stats` regenerates the shipped corpus_stats.json for the site's analytics form set, which does not include this edition; reproduce these figures with `python -m symagedocs.corpus_analysis --corpus-size 1000 --form-id irs_f941_schedule_d_standard_2024`.
Frequently asked questions
- What data format do synthetic Form 941 Schedule D documents include?
- Each generated identity produces a rendered PDF plus a structured JSON annotation file with bounding boxes, field types, and ground-truth values for all 64 fields across both pages, with the IRS, SSA, and difference amounts annotated as three distinct fields per line rather than one merged row. COCO, YOLO, FUNSD, and BIO/NER exports come from the same job.
- Which wage lines are reconciled?
- Social Security wages, Social Security tips, Medicare wages and tips, federal income tax withheld, and advance EIC payments — each as an IRS amount, an SSA amount, and a difference, plus a parallel set of the same five lines for employers whose totals span multiple returns. Employer identification, address, phone, and the restructuring date round out the page.
- How does labeling work for the three-column layout?
- Field ids are prefixed irs_, ssa_, and diff_, so irs_ss_wages, ssa_ss_wages, and diff_ss_wages are three separately annotated regions sharing one printed line label. That is why the form carries more annotation relations than fields — the relation graph records that one label governs three values, which is exactly the structure a column-aware model needs to learn.
- Can I use this data commercially?
- Yes. Every employer, restructuring event, and wage figure is synthetic, contains no real company or SSA data, and is licensed for commercial use including model training, benchmarking, and redistribution inside your own products.
- Does it pair with the underlying Form 941?
- Yes. Generate Form 941 and Schedule D together and the quarterly returns provide the IRS-side figures the schedule reconciles against, giving you a realistic multi-document discrepancy investigation rather than an isolated page.