We don't ask you to trust our maths

The problem with "trust us"

You cannot tell whether a payroll engine is right by looking at it. The screen shows a number. It looks like a number. If it's wrong by $16 a fortnight, nothing on screen tells you — you find out when an employee queries their payslip, or when a return doesn't reconcile at year end, and by then it has been wrong for months across every employee you process.

That's an awkward problem for a vendor, because the honest answer to "is it accurate?" can't be the vendor's own assurance. Anyone can say it. So Brawta doesn't.

What we do instead

TAJ publishes worked examples — scenarios where they print both the inputs and the answers. Those examples are turned directly into automated tests. Brawta is given TAJ's inputs and required to produce TAJ's printed figures, to the cent. If it doesn't, the build fails and the release does not go out.

The point is who is being trusted. Not the developer, not the accountant who wrote the spec, not a reviewer signing off because they know us. The authority is the regulator's own published document, and the check runs automatically rather than when someone remembers to do it.

Tested against TAJ Technical Advisory 04/2025

"Increase in Income Tax Threshold effective April 1, 2025 – December 31, 2028" — cumulative PAYE worked scenarios.

Tested against the TAJ Employer's Guide

Employer's Guide to Income Tax, Statutory Deductions & Contributions (13 Dec 2019 edition).

The $16.80 that most engines get wrong

Here is a concrete example of what this method catches — and it is the reason we think it's worth doing.

In a year where the income tax threshold rises on 1 April, the increase lands part-way through a fortnight. The obvious approach is to take the annual threshold and divide by 26. TAJ does not do that. They publish a specific per-fortnight figure that accounts for the split year:

Weekly and monthly payrolls divide evenly at the April boundary, so they're unaffected. Fortnightly does not — and fortnightly is extremely common in Jamaica. Nineteen fortnights, fifty employees, and the "obvious" method is out by roughly $16,000 for the year, in a direction that under-taxes and leaves the employer short at reconciliation.

What this does and doesn't prove

Being straight about the limits, because a claim you have to over-read isn't worth making:

• It proves these cases. Where TAJ published an answer and we reproduce it, that line is validated by an authority that isn't us. It does not prove every possible payroll situation is handled correctly.
• TAJ has not endorsed Brawta. They haven't reviewed it, certified it, or approved it — and no payroll software in Jamaica is TAJ-certified. We use their published material as a measuring stick. That's all, and it's deliberately all we claim.
• It isn't a substitute for your judgment. Brawta computes; you remain responsible for what you file. Severance, unusual terminations and disputed cases still want a professional's eye.
• The rest of the engine is covered too — over 790 automated checks run on every change, covering NIS ceilings, Education Tax, NHT, HEART, week 53, mid-period rate changes, terminations and the statutory return forms. The TAJ-anchored ones above are simply the checks where the answer isn't ours to decide.

Rates that don't go stale

Correct-at-release isn't much use if the figures drift. When TAJ changes a threshold or a contribution rate, Brawta picks up the new figures on its own — signed by us and verified before they're applied — so a correct payslip doesn't depend on somebody remembering to install an update. You can turn that off, and you can always enter a rate yourself.