Monetary correction: what the official table leaves out

Why the acquisition value is uprated

A capital gain is, at its simplest, the difference between the sale price and the purchase price. The trouble is that those two amounts are measured in different currencies: the euro of 1998 is not the euro of 2025. If the state taxed the nominal difference, it would be taxing inflation, money the seller never earned.

Monetary correction fixes that on the cost side. The acquisition value is multiplied by a currency-devaluation coefficient for the year the asset was bought, and it is that corrected figure that is subtracted from the sale price ². The arithmetic is deliberately simple:

corrected acquisition value = acquisition value (in euros) × coefficient

Someone who bought a flat in 1998 for €50,000 sees the acquisition cost rise to €86,000, because the 1998 coefficient is 1.72. That is €36,000 that stops being taxed as property capital gains, not because the law is granting a discount, but because those €36,000 were never profit in the first place.

Where the coefficient comes from

The coefficient is not an inflation estimate each taxpayer makes. It is a published number, and published on a very regular schedule.

Every November, the government member responsible for finance approves by Portaria the table of coefficients to be applied to assets and rights disposed of during that year. The Portaria is issued simultaneously under Art. 50 of the IRS Code and Art. 47 of the IRC Code, so the table is literally the same one for individuals and for companies ¹.

The table in force is Portaria n.º 382/2025/1 of 11 November, published in the Diário da República, 1st series, no. 218, and it applies to disposals made during 2025. It has 76 rows covering 122 years of acquisitions:

• the first row is open-ended ("Até 1903", "until 1903") and carries the highest coefficient, 5,585.78;
• the oldest rows group several years together ("1904 a 1910", for instance);
• from 1915 onward the table is almost entirely year by year;
• the last row is 2024, with a coefficient of 1.00.

One detail worth knowing before anyone tries to "tidy up" the series: the table does not always descend. The coefficient for 2009 is 1.30, higher than both 2008 and 2010 at 1.28. It is not a transcription error; it reflects the fall in Portuguese consumer prices during 2009, so money from that year needs a slightly larger correction than money from the year before. It is the only inversion in the whole table, and it is enough that any interpolation or smoothing of the series returns a figure that is not the published one.

The rule the table does not contain: the holding period

The table gives you a coefficient per year of acquisition and nothing else. But the coefficient applies only if the asset was held long enough, and this is where most pages that republish the table go wrong, because the condition is not in the table, it is in the law.

And the two laws do not word it the same way.

Under personal income tax, Art. 50(1) of the IRS Code corrects the acquisition value "sempre que tenham decorrido mais de 24 meses entre a data da aquisição e a data da alienação", whenever more than 24 months have elapsed between acquisition and disposal ².

Under corporate tax, Art. 47(1) of the IRC Code uprates the acquisition value "sempre que, à data da realização, tenham decorrido pelo menos dois anos desde a data da aquisição", whenever at least two years have elapsed since the acquisition, measured at the date of realisation ³.

At first glance these are the same thing said twice. They are not. "More than 24 months" excludes the very day on which 24 months are completed; "at least two years" includes it. In practice:

An asset bought on 15 March 2023 and sold on 15 March 2025 is corrected under corporate tax and is not corrected under personal income tax. On an acquisition value of €200,000, the 2023 coefficient (1.02) is worth €4,000 of extra cost, which the company can deduct and the individual cannot. One day later, both would qualify.

Two things follow. First, the holding period is counted between dates, not between calendar years: looking at the table and seeing that you bought in 2023 and sold in 2025 tells you nothing about whether you qualify. Second, the same Portaria, applied to the same asset on the same dates, can give two different answers depending on which tax is at stake.

The other rule the table does not contain: the currency

The coefficients on the oldest rows are enormous. "Até 1903" is 5,585.78; 1950 is 101.78; 1980 is 12.68. They are large because they bridge the escudo era, but they still apply to a value in euros.

This is the source of the commonest error in these calculations. A 1995 deed reads 12,000,000$00. The number that enters the formula is not 12,000,000: it is that amount converted from escudos into euros at the fixed, irrevocable rate of 200.482 escudos per euro, i.e. €59,855.75. Only then is it multiplied by the 1995 coefficient (1.84), giving €110,134.58.

Anyone applying the coefficient straight to the 12,000,000 would get an acquisition cost of €22,080,000, about 200 times the right figure. On a real sale that wipes out the gain entirely and produces a return that is wrong in the taxpayer's favour, which is precisely the sort of discrepancy the tax authority cross-checks easily, because it holds the deed.

If the figure is in contos, there is one more step: one conto is 1,000 escudos, so 12,000 contos is 12,000,000$00 and only then €59,855.75.

Which assets it reaches

Art. 50(1) of the IRS Code is explicit about its own reach: it corrects the acquisition value "de direitos reais sobre os bens referidos na alínea a) do n.º 1 do artigo 10.º" (real estate), "bem como de partes sociais no caso da alínea b) do referido número" ².

Paragraph (b) of Art. 10(1) covers the "alienação onerosa de partes sociais e de outros valores mobiliários": the disposal of shares and of other securities ⁴. The cross-reference in Art. 50 picks only the first half of that paragraph. So:

Under corporate tax the line is drawn from the other side and lands in the same place: Art. 47(2) states that monetary correction "não é aplicável aos instrumentos financeiros, salvo quanto às partes de capital": it does not apply to financial instruments, except to equity holdings ³.

The implication is worth spelling out, because it surprises almost everyone: someone selling shares held for more than two years is entitled to uprate what they paid for them, exactly as someone selling a house is. Someone selling an ETF held for twenty years is not.

Where the correction sits in the calculation

Monetary correction is the first step, not the last. On a resident's property, the full sequence is:

1. convert the acquisition value into euros, if needed;
2. multiply it by the coefficient for the year of acquisition (if the minimum holding period is met);
3. subtract from the sale proceeds that corrected value, the acquisition and sale costs and the improvement costs;
4. of any positive balance, take only 50% (the residents' rule);
5. add that half to the rest of your income and apply whatever progressive rate falls on it.

One consequence is worth keeping in mind: a euro of correction is not a euro of tax saved. On a resident's property it is fifty cents of taxable base, and that is then multiplied by the marginal rate. On an inherited property, the correction applies to the taxable patrimonial value at the date of death, and the year of acquisition is the year of the death, not the year the deceased bought the house.

What to do about a 2026 sale

If you sold in 2026, the table that governs your case has not been published yet: the Portaria for 2026 disposals will come out in November 2026, with the same regularity as its predecessors.

In practice this is almost never a blocker. A 2026 sale is reported on the IRS return filed in 2027, and the right table will be published well before the deadline. The return filed in 2026 concerns 2025 sales and uses Portaria n.º 382/2025/1: the one in force, and the one our calculator applies.

If you need an estimate before then, the sensible practice is to use the most recent published table and treat the result as an approximation: the coefficients for older years move very little from one Portaria to the next, and the coefficient for recent years is almost always 1.00 or close to it. What you should not do is present that figure as final on a return.