Research brief · Published
Payroll preparation and approval separation coverage
A study design for measuring how often payroll preparation, review, and approval are performed and evidenced by the right roles.

Research finding
Score preparation, independent review, and authorized approval as separate gates rather than one completed status.
Methodology
This brief triangulates the headline measure against official Philippine government, regulatory, development, and labor sources. It translates evidence into an operating control and separates context from recommendations.
| Measure | Interpretation |
|---|---|
| 3 evidence gates applied to each prepared payroll item. | Context signal for planning, not a promise about an individual worker or provider. |
| 4 source records | Primary source links are listed and numbered below for review. |
Key takeaways
- Score preparation, independent review, and authorized approval as separate gates rather than one completed status.
- Record the approver's authority source, scope, and pay group for every sampled item.
- Treat a missing scope or an undocumented substitute as an approval gap.
- Report coverage by pay group and decision type so one easy lane cannot hide another.
Question and scope
This study asks what share of payroll items has evidence that preparation, independent review, and approval were performed by the appropriate roles. The population is every item that changed a pay-affecting value for one employer during a defined period. The unit is one prepared item with its evidence, reviewer, approver, and pay cycle. The question separates the person who assembled the record from the person who accepted it, because a completed screen does not show who decided. The result describes process evidence for this employer, not the legal validity of any payment.
Population and unit
Define the population before sampling. Include new hires, terminations, rate changes, bank changes, deduction elections, timesheet corrections, and off-cycle payments if they affect the pay run. Exclude items that were withdrawn before any payroll effect, but keep a count of them. For each included item, record a unique identifier, pay group, effective period, and current disposition. A population built only from closed items will overstate coverage because abandoned and deferred work disappears. Keep the raw extract so a reviewer can reproduce the frame.
Preparation evidence
Preparation is documented when the record shows the source, the change, the effective date, and the person who assembled it. A source can be a timesheet, an approved form, a signed notice, or a system audit entry. Grade the gate as met only when the evidence links to the specific employee and period. A generic note that the change was requested is not preparation evidence. Preparation may be delegated widely, so this gate should have the highest coverage. Gaps here usually point to intake design rather than reviewer behavior.
Review evidence
Independent review means a second person checked the prepared item against the source before approval. Record the reviewer, the date, and what was checked. When the preparer and reviewer are the same person, mark the gate as not independent even if the item is correct. Independence is a control property, not a judgment about trust. A small team may need a rotational or sampled review rather than a full second pass. Document the chosen model and apply it consistently across pay groups.
Approval authority
Approval is met when a person with current authority for the pay group records a decision on a defined scope. Capture the approver, the authority source, the scope, the pay period, and the decision. A system permission alone does not prove authority. Compare the approver against a current delegation list that includes effective dates. If the list is missing or stale, report the approval gate as unverifiable rather than passing it. An unverifiable gate is a finding, not a neutral result.
Substitution and absence
Separate planned substitutions from emergency ones and score both. A planned substitution should have a scope, dates, and a return-to-owner step. An emergency approval should record why the primary owner was unavailable and who accepted the risk. Count substitutions that lack a documented scope or that extend beyond the stated period. Repeated substitutions may indicate a capacity problem rather than isolated absence. Keep the substitution record linked to the item so coverage is not diluted by a separate roster review.
Timing and cutoff
Timing is part of the evidence because an approval after the file release cannot have governed that run. Record when the item was prepared, reviewed, approved, and released. Compare the approval time with the published freeze point. A late approval that was correctly deferred should count as a met gate for the cycle it actually affected. The study must not reward a fast approval that arrived after the bank file. Report timing separately so a complete but late item is visible.
Sampling and replay
If the population is too large to review in full, use a fixed sampling rule. Record the frame, the selection method, the sample size, and the period. Include deferred and rejected items in the frame; a sample of successful payments only will overstate coverage. Give a second reviewer a de-identified subset and the gate definitions without the first reviewer's notes. Compare the grades and resolve disagreements by inspecting the source. Track disagreement reasons because repeated confusion about one gate usually means the definition is unclear.
Data quality
Before scoring, check that each item has one identifier and one employee link for the period. Look for duplicate items across channels, impossible timestamps, and approvals attached to the wrong pay group. Record every cleaning decision and keep the raw extract separate from the working copy. A high coverage rate on a duplicate-heavy extract is misleading. The report should state how many records were excluded and why. Data quality is part of the finding, not an administrative footnote.
Limitations
This design measures evidence coverage, not whether the underlying pay was correct or lawful. It depends on the quality of system logs and on the delegation list being current. It cannot prove that a reviewer actually read every source, only that a review was recorded. A missing record may reflect a real gap or a process that documents elsewhere. State those limits before interpreting the result. Use the finding to direct review, not to assign fault to an individual.
Access and tool boundaries
Record which system each role can use and which fields they can change. A preparation role should read the payroll record and write only to the queue or a draft, not to the pay value itself. Compare the permission with the task evidence for the same period. If a support worker can edit a pay-affecting field without a separate approval step, the control has a technical gap even when the paperwork looks complete. Report that gap beside the approval coverage result.
Communicating an approval gap
When an approval gap is found, tell the owner before the next pay run. State the item, the missing element, and the risk in plain language. Do not send a general warning that a few approvals were incomplete. The owner needs the specific records to decide whether a correction, a disclosure, or a process change is required. Record the response and the date. A finding that is never discussed cannot change the next cycle.
Exception concentration
Group approval gaps by preparer, approver, pay group, and decision type. A single approver with several scope-free decisions may need a narrower delegation or more support. A pay group with repeated preparation gaps may have an intake form that does not require the right fields. Concentration is more actionable than a site-wide rate because it points to one owner or one workflow. Report the largest groups and the evidence behind them.
Record retention
Keep the evidence long enough to support a later review, and no longer than the employer's policy allows. Store the extract, the gate definitions, the sample list, and the reviewer notes together. Remove sensitive identifiers that are not needed for the analysis. A later reviewer should be able to rerun the sample without asking for the original files scattered across inboxes. Record the retention decision with the study.
Cross-period comparison
Compare two periods only when the population rule, gate definitions, and delegation list are similar. Record any change in pay groups, systems, or decision types before the comparison. A new intake form can raise preparation coverage simply by requiring fields that were always present but undocumented. A leadership change can alter approval behavior without changing the process. State the difference and interpret the trend cautiously. The study should show what changed, not just that a number moved.
Decision link
Name the decision the result would change. If approval coverage is low for one pay group, the owner may need to narrow who can approve or add a second check. If preparation gaps cluster around a channel, the intake form may need a required field. If substitutions lack scope, the delegation process needs repair before the next close. The study should end with one repair, one owner, and a retest date. A result without a decision is only a chart.
Retest after repair
Retest the same gates after the repair with a comparable period. Keep the definitions and population rule stable so the comparison is meaningful. Report the before and after result by gate and by pay group. If one gate improved while another declined, say so instead of reporting a blended number. A retest that changes several definitions at once cannot show which change helped. Record the sample, the reviewer, and the date before deciding whether to keep the repair.
Sources
Prepare a reviewable approval trail
A Philippines-based payroll preparation specialist can assemble the source records, mark the evidence gates, and prepare the owner decision summary. Your authorized payroll owner still approves pay-affecting changes and decides what the result means.
Review payroll preparation supportFAQs
Does this study prove that every approved item was paid correctly?
No. It measures whether preparation, review, and approval evidence exists. Payment accuracy requires a separate reconciliation against the payroll register and source records.
Why separate review from approval?
A reviewer can check the evidence without having authority to accept the pay-affecting decision. Combining the gates hides whether either step actually happened.
What should a small team do first?
Write the delegation list and the stop rules. Then sample one pay group to see which gate has the weakest evidence before adding more controls.
For adjacent operating context, see Payroll Preparation and the payroll operations guide library.