IRC and IRS Revenue Forecast
Portugal collects corporate income tax (IRC) and personal income tax (IRS) through several channels: withholding, payments on account, self-assessment, collection notices, instalment and enforced payments. Refunds, transfers to the autonomous regions and interest come off the top. This tool estimates the year's outturn, the revenue collected over the whole year, from the months already known, item by item, and forecasts the next year.
How the estimate is built
Each item's missing months are filled by one of six rules, applied to the same month of the previous year:
- R1, same as previous year:
x̂ = xt−1. - R2, growth:
x̂ = xt−1 × (1 + g). - R3, mean change:
x̂ = xt−1 × (1 + v̄), where v̄ is the mean of the year-on-year changes observed so far, or the change in their total. - R4, target: an estimate M from outside the model covers some months, such as the year's
payments on account or the months of the IRS campaign. The key months still to come get what is left,
M × (1 − q)less the months already fixed, split in the previous year's proportions; q is a shortfall. - R5, zero, for an item that no longer exists.
- R6, average: some months are set to the average of chosen actual months of the same year, for when the previous year's months are not a good guide.
Net revenue is gross revenue less charges (refunds, transfers to the autonomous regions, interest, and the municipal surcharge for IRC or LLR for IRS). The deviation compares it with the State Budget. The forecast for the next year multiplies each item's total by a factor, such as GDP or wage growth.
How to use it
The calculator opens on an illustrative example with made-up figures. Switch between IRC and IRS, and change any item's rule, growth rate, target or months; the results update at once. To use real data, load your files:
- the Excel workbook of the outturn estimate (IRC: sheet Cálculos Auxiliares; IRS: sheet RECEITA_CENÁRIOS). Only typed numbers count as actual revenue; formula cells are the workbook's own estimates and are ignored;
- or a CSV with columns
year,itemand1to12; - and, optionally, the assumptions file (
.toml) with the rules, targets and budget.
Files are read in your browser and never sent anywhere. With the same files, the results match the Python tools behind this model to the cent.
What it leaves out
- Macroeconomic feedback: the rules extrapolate the previous year's monthly pattern, with the growth rates you set.
- The targets for R4 come from outside the model, from the payments-on-account liability or the IRS campaign statistics.
- IRC refunds do not include REAID restitutions.