Cross-border shadow payroll reconciliation study

Methodology

This brief triangulates the headline measure against official Philippine government, regulatory, development, and labor sources. It translates the evidence into an operating control and separates context from recommendations.

Key takeaways

• Freeze populations and definitions before reviewing outcomes.
• Separate facts, owner decisions, analysis, inference, and uncertainty.
• Use protected references and stop at authority or interpretation questions.

Construct paired populations from the assignment roster

Start with the authorized assignment roster, then freeze home and host extracts using protected tokens, entities, jurisdictions, earning and payroll periods, element codes, currencies, source references, and extraction times.

This keeps a person missing from one payroll visible.

Record split pay, multiple entities, off-cycle results, and third-party feeds.

Do not join on names.

Define mid-period starts and ends and late-record roll-forward.

Maintain a versioned one-to-many element map with owner and effective date rather than forcing systems into identical codes.

Exclusions such as cancelled assignments or travelers outside scope need individual reasons.

Balanced totals cannot compensate for a missing person or wrong element mapping.

Separate currency, timing, and treatment

Retain source currency and amount, approved rate source, rate date, value, precision, and converted amount.

Never infer method from rounded output.

Map earning period, payment date, reporting period, cutoff, and filing period separately.

A timing difference needs a referenced future or prior record expected to resolve it.

Treatment differences need a qualified owner reference and cannot be inferred from similar labels.

Benefits, equity, bonuses, allowances, hypothetical tax, and employer-paid items may follow different feeds.

Explain whether each variance comes from currency mechanics, timing, source completeness, mapping, or approved treatment.

Net variance obscures the action and authority required.

Reconcile in both directions

From the roster, confirm expected home and host presence.

From each payroll, confirm the assignment and mapping justifying inclusion.

Trace home elements to expected host records and host records to an authorized source or local treatment.

Aggregate only after eligibility and mapping are established.

Use tolerances only for declared rounding or conversion, never for missing elements.

Sample matches because stale copied amounts can balance.

Follow corrections into later outputs without erasing original states.

Cover different currencies, assignment phases, and element families in independent review.

Unresolved source conflict goes to global owners; support must not select whichever record makes totals align.

Bound the cross-border support lane

Report paired and unpaired assignments, mapped and unmapped elements, mechanisms, unresolved age, owner deadlines, and correction confirmation in source currencies and approved reporting currency.

Do not net overstatements and understatements or claim filing accuracy.

Global mobility, tax, payroll, finance, HR, and legal owners define applicability, exchange method, source authority, treatment, and correction.

Support can maintain controlled extracts and comparisons.

It cannot determine residence, taxable compensation, social insurance, permanent-establishment questions, employment status, or entitlement.

Minimize data and never place returns, passports, bank details, health records, or assignment documents in the analysis workbook.

Explain one difference without flattening home and host records

Take one protected assignee and one compensation element that differs between the home and host extracts.

Preserve each source amount, currency, earning period, payment date, reporting window, element code, mapping version, exchange-rate method, and owner reference.

Show the reconciliation first in source currencies, then under the approved reporting conversion.

Identify whether the difference is timing, currency, mapping, source completeness, approved treatment, or unresolved.

Do not designate either payroll as universally correct.

If a later record is expected to resolve a timing item, name that record and verify it in the next window rather than closing the exception immediately.

Have the global owner confirm any treatment conclusion.

This worked trace demonstrates whether the support lane can maintain paired evidence and carry uncertainty forward without interpreting tax, residence, employment, social insurance, or filing obligations.

It also exposes when a balanced consolidated number conceals an unmatched person or element.

Report the unresolved item in both source currencies and keep its next expected evidence date visible to the owner.

Retest the pair in the following window.

Pair populations before comparing compensation

Freeze the authorized assignment roster and independent home and host payroll extracts for the same declared window.

Use protected stable tokens, then classify every assignee as present in both, roster-only, home-only, host-only, or otherwise unresolved.

Record entity, jurisdiction, assignment phase, payroll and earning periods, extraction time, and source version.

Investigate population differences before monetary reconciliation.

Do not join on names or allow a balanced consolidated headcount to hide a missing assignee.

Document travelers, cancelled assignments, split arrangements, mid-period starts and ends, and third-party feeds according to the approved scope.

This step shows whether later element comparisons use a defensible denominator and identifies data-sharing or source-access limitations without requiring unnecessary identity documents in the analysis file.

Analyze elements in source currency and approved mapping

For paired assignees, retain each element code, owner-approved mapping, earning period, payment date, source currency and amount, approved exchange-rate source and date, converted value, and reporting window.

Classify differences as timing, currency method, mapping, missing record, duplicate, approved treatment, or unresolved.

Never infer treatment from labels that appear similar across countries.

A future-period timing item stays open until the named record appears.

Compare line and cohort results before totals, and do not net overstatements against understatements.

Reviewers must cite the mapping version and owner decision for every nonliteral match.

This makes the reconciliation reproducible while keeping tax, residence, employment, social insurance, and filing interpretations outside the support analyst's role.

Report uncertainty and correction propagation

Publish the paired-population denominator, state counts, unmapped elements, timing items, currency-method differences, unresolved age, owner deadlines, and confirmed later records.

Show amounts in source currencies and only then under the approved reporting conversion.

Preserve initial and corrected states so a later match does not erase the original handoff gap.

A second reviewer should repeat all unmatched cases and a declared sample of matches across jurisdictions and element families.

Explain source limitations and small populations.

The buyer may use the study to decide whether outsourced reporting support can maintain evidence and route exceptions.

It must not present the findings as tax compliance, filing accuracy, employee entitlement, provider performance, or proof that either home or host payroll is universally authoritative.

Reproduce the comparison after maps or rates change

Retain source extract hashes, protected token rules, roster version, element-map version, exchange-rate table, conversion logic, reporting-window definitions, reviewer notes, and exception dispositions.

When an owner changes an element mapping or currency method, keep the prior version and rerun only the affected comparisons under the newly approved rule.

Do not rewrite the original observation.

A later study should distinguish genuinely corrected records from results that differ because the method changed.

Document local records the reviewer could not access and avoid assuming the accessible payroll was authoritative.

This reproducibility record lets an authorized team repeat the analysis without copying passports, returns, bank details, or full assignment files into the research workspace.

FAQs