Research brief · Published
Payroll records retention schedule adherence
A bounded study design for testing whether payroll record classes are kept, retrieved, and destroyed according to the employer's own schedule.

Research finding
Adherence is measured against the employer's written schedule, not an assumed legal period.
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 |
|---|---|
| 4 adherence measures applied to each record class. | 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
- Adherence is measured against the employer's written schedule, not an assumed legal period.
- Retrievability is tested separately from whether a record still exists.
- Destruction evidence needs a date, an approver, and a hold check.
- Holds suspend the schedule and must be released by an owner.
Question and scope
This study asks whether payroll record classes are retained, retrieved, and destroyed according to the employer's written retention schedule. The population is the record classes named in that schedule, not every file the team has ever created. The unit is one record class within one review period, with measures for presence, retrievability, destruction evidence, and hold handling. The study does not decide the correct legal retention period and does not evaluate whether the schedule is optimal. It tests whether the organization follows the schedule it wrote. That distinction matters because a modest schedule applied consistently is more defensible than an ambitious one nobody can follow.
Record classes and triggers
List each record class with its purpose, owner, location, retention trigger, and review date. A trigger is the event that starts the retention clock, such as the end of the pay period, the filing date, or the close of a case. Record classes with no clear trigger as undocumented rather than assigning one during collection. Check whether the class matches how the team actually works, because a schedule written for a paper file room may not describe a payroll system, a shared drive, and a vendor portal. Keep the schedule version beside the sample so a later reviewer can see which rules applied.
Sampling at review time
Sample record classes at the review date rather than at creation, because the question is whether the schedule holds over time. Include at least one class that should still be present, one that is near its review date, and one that is eligible for destruction. Record the location checked and the person who produced the record. If a class exists in several systems, state which system was treated as authoritative and why. A sample drawn only from the easiest system can overstate adherence, so the report should name the locations that were not checked.
Retrievability as a separate measure
A record can exist and still be unusable. Measure retrievability by asking a reviewer who did not create the file to locate a specific record within a stated time. Record the time taken, the access path, and whether the reviewer needed help. A pass means the record was found and opened without relying on the original owner's memory. Keep this measure separate from presence, because a complete archive with a fragile access path will fail when the person who knows the path is unavailable. The study reports both results rather than merging them into one readiness score.
Destruction evidence
For each destruction event, record the record class, the date, the approver, the hold check, and the list of removed items. A destruction without an approver or a hold check is unverified even if the deletion was appropriate. Check whether the event appears in the same calendar the team uses for review, because a separate log is easy to abandon. Do not accept a folder's empty state as evidence that the schedule was followed; an empty folder can also mean a lost or moved record. The study grades the evidence, not the intent behind the deletion.
Independent replay
Give a second reviewer the schedule, the class definitions, and a fixed subset of records. Ask that person to grade presence, retrievability, and destruction evidence without seeing the first reviewer's notes. Compare the grades and the evidence cited, then resolve disagreements by inspecting the record or the log. Disagreements about the schedule go to the records owner, while disagreements about a location are settled by the system. The replay tests whether the measures are clear enough for another person to apply. Record the time spent, because a slow replay may reveal an access problem rather than a method problem.
Confounders and holds
A record may be absent for reasons unrelated to the schedule. A legal hold, an audit, a migration, or a vendor change can move or preserve records outside the normal path. Record every hold that applies during the review period, the classes it covers, and the date it was released. Do not count a held record as a schedule failure, and do not count a release as evidence that destruction occurred. Migration is another confounder, because a moved record may be present but no longer retrievable through the old path. The report should name these conditions before interpreting the adherence rate.
Privacy and access
The study uses record identifiers and keeps any identity crosswalk in the approved payroll system. Limit the sample to the fields needed to test presence and retrievability, and redact sensitive content from the working file. Record unavailable fields rather than reconstructing them. Temporary access should end when the replay closes, with each grant and removal dated. Privacy handling protects the records but does not validate the schedule. A well-protected sample can still miss a record class, so access control and coverage remain separate findings. The report should state which locations could not be checked and why.
Limits and next step
The result depends on the record classes, the review date, the access paths, and the reviewer's judgment. Repeat the analysis after one review cycle before treating the pattern as stable. If the result changes, report both readings and name the decision that depends on the definition. Use the evidence to improve one control, such as requiring a hold check before any destruction. Then schedule a retest rather than declaring the schedule solved. Name the decisions that would change if the finding is wrong, such as the retention period for a class or the access granted to a support role.
Sources
FAQs
Does this study set a legal retention period?
No. It tests adherence to the employer's written schedule. The employer sets the period with advice from payroll, legal, and tax resources.
Why measure retrievability separately?
A record can exist but be unusable if access depends on one person or an old path. Retrievability is a distinct operational test.
How should a legal hold affect the result?
A hold suspends destruction for the affected classes. The study records the hold and excludes those records from a destruction failure.
For adjacent operating context, see Payroll Preparation and the payroll operations guide library.