The State's Ledger
Every year the Portuguese State collects about two-fifths of what the country produces and spends a little more. Assign each euro to the person who pays it and to the person who benefits, and most adults come out ahead, the middle roughly breaks even and the richest tenth pays in about a sixth of its income. How that result is built, what it counts as a transfer and what it assumes, matters as much as the result.
Earned and Unearned included a chart of what each tenth of Portuguese adults gains or loses from all taxes and all public spending together. The poorest tenth gained several times its own income, the middle broke even and the richest tenth lost 18 per cent. The chart gave the result and not the arithmetic: where the money comes from, what is counted as a transfer, and why the poorest tenth’s gain runs to hundreds of per cent. This essay opens the ledger.
It goes in three steps. First the State’s own accounts: what it collects and what it spends. Then the method that assigns each of those euros to people, and the choices in it. Then the check: whether other estimates, built differently, reach the same answer. Along the way it corrects one thing in the earlier chart. The figures labelled 2022 are in fact the structure measured in 2019, carried forward.
Where the money comes from
In 2022 taxes and social contributions brought in 37.9 per cent of GDP, €92 billion, somewhat less than the EU average of 40.7 (Figure 1). The largest single tax is VAT, 9.4 per cent of GDP, well above the EU’s 7.4; with excise duties and the other taxes on products and production, they bring in 15.2 per cent of GDP, two-fifths of the total. Social contributions bring in 12.2 per cent: 6.2 paid by employers, 4.0 by workers and the self-employed, and 2.0 imputed to the State itself as the employer of civil servants whose pensions it pays directly. Personal income tax brings in 6.9 per cent and the tax on company profits 3.3.[1]
Taxes are not the whole of revenue. The State also sells services, receives EU funds and earns property income, which together added about another 5.7 per cent of GDP; and it borrows the difference. In 2022 the deficit was 0.3 per cent of GDP; by 2024 it had become a surplus of 0.6.[2]
Where it goes
The State spent 43.9 per cent of GDP in 2022, against an EU average of 49.1 (Figure 2). The largest item by far is pensions for old age and survivors, 12.6 per cent of GDP, more than the EU average. Health takes 7.1, education 4.4, the economy and transport 4.8, and interest on the public debt 2.0, more than the EU because Portugal’s debt is larger. The biggest gaps with the EU are in the rest of social protection: Portugal spends 1.4 per cent of GDP on sickness and disability against the EU’s 2.7, and 1.3 on unemployment, housing and social exclusion against 3.0.[3]
So the Portuguese State is a little smaller than the European average, and its spending is weighted towards pensions and the two great universal services, health and education, and away from cash support for people of working age. That shape decides most of what follows.
What counts as a transfer
The first choice in any calculation of who gains and who loses is what to do with pensions. They are the largest item of spending, and they could be counted either as a transfer from workers to pensioners or as deferred pay, the return on contributions made during a working life. The two treatments give very different pictures, because pensioners are numerous in the lower half of the income distribution.
The World Inequality Database, whose estimates are behind the chart in the earlier essay, takes the second view. Its “pre-tax income” already includes contributory pensions and unemployment benefits, and deducts only the contributions that pay for them. The net gain it reports is therefore about everything else: all the other taxes, the non-contributory cash benefits, and public services.[4][5]
How much pensions do can be seen from the household survey (Figure 3). Without any pensions or social benefits, 40.7 per cent of people in Portugal would have been below the poverty line on their 2024 incomes. Pensions alone bring that down to 20.8 per cent; all the other benefits together bring it down to 15.4. Pensions do four-fifths of the work. The other benefits reduce poverty by 5.4 points in Portugal against 8.1 in the EU, which is the same gap in working-age support that Figure 2 shows in spending.[6]
From pre-tax to post-tax, one tenth at a time
With pensions inside pre-tax income, the calculation takes two steps.[5][7][8]
The cash step. From pre-tax income it subtracts every tax: income tax, the social contributions that do not finance pensions or unemployment benefits, VAT and excise duties, allocated in proportion to each group’s consumption as measured by the household budget survey, and corporate tax, allocated to the owners of the companies that pay it. It then adds the non-contributory cash benefits, such as child benefit and the minimum-income benefit. The result is post-tax disposable income.
The services step. To that it adds everything else the State spends, so that the total adds up to national income. Health care is given to every adult as an equal amount, on the reasoning that a public health service insures everyone equally. Education and all collective spending, defence, police, courts, roads, administration, are given in proportion to disposable income. The deficit, which is spending not yet paid for by taxes, is spread in the same way. The result is post-tax national income.
Figure 4 shows both steps in euros for each tenth. The poorest tenth has a pre-tax income of about €860 a year per adult, at 2025 prices: mostly people with little or no earnings who are not yet or no longer entitled to a contributory pension. It receives slightly more in cash benefits than it pays in taxes, mostly VAT, and then about €4,200 in services. Its post-tax income is about €5,700, a gain of some €4,800 on a base of €860, which is how a percentage of more than 500 arises. From the second tenth upwards the cash step is negative and grows with income; the services step grows too, but more slowly. The two cancel around the eighth tenth. The richest tenth pays about €36,900 per adult more in taxes than it receives in cash, and is credited with about €20,000 in services; its net contribution is about €16,900, 18 per cent of its pre-tax income.[4]
One feature of the chart is a choice rather than a finding. The richest tenth is credited with almost five times as much “service” per adult as the poorest, because education, defence, roads and the rest are assumed to benefit people in proportion to their income. The World Inequality Database calls this the “conservative” assumption, and its own guidelines say the evidence places the true benefit somewhere between an equal amount per person and a share proportional to income.[8] Giving collective spending as an equal amount to everyone would make the richest tenth’s net contribution larger and the poorer half’s gain larger still. In the European study behind these series, doing so “reduces inequality significantly” but changes neither the ranking of countries nor the trends.[5]
What the numbers rest on
Three things should be said plainly. The first is the date. The World Inequality Database’s post-tax series for Portugal is estimated up to 2019; for 2020 to 2022 it applies the 2019 ratio of post-tax to pre-tax income, percentile by percentile, to newer pre-tax figures.[4] The chart in the earlier essay, labelled 2022, therefore shows the redistribution of 2019 applied to the incomes of 2022; tax and benefit changes since 2019, including those of the pandemic, are not in it. That is why the figures here use 2019. The difference is small: the richest tenth’s net contribution is 18.0 per cent in 2019 against 17.8 in the extrapolated 2022.
The second is the data at the bottom. The pre-tax series is built from the EU household survey, adjusted at the top with tax tabulations that for Portugal end in 2019. The latest update of the European series notes that the earlier method “proved inadequate for the bottom percentiles”, sometimes producing negative incomes, and replaced it.[9] Percentages for the poorest tenth, measured against a base of a few hundred euros, should be read as orders of magnitude.
The third is that every allocation is a rule, not an observation. Nobody records how much of the Portuguese navy a given household consumes, or who bears a tax on company profits. The rules used here are standard and published, and the debate about them is real: in the United States, a rival estimate that gives half of government consumption to everyone as an equal amount, “like a big universal basic income” in the words of its critics, produces a noticeably more equal post-tax distribution.[10]
Three estimates, one shape
If the result depended entirely on these choices, other estimates would disagree. Figure 5 sets the World Inequality Database’s numbers, grouped into fifths, beside two estimates built from the OECD’s distributional national accounts for 2017, which rank households rather than adults and treat pensions as a cash benefit rather than as pre-tax income. The first OECD estimate counts only income taxes and contributions against cash benefits and services in kind; it leaves out VAT and collective spending altogether. The second adds an estimate of VAT and excise duties, at Portugal’s average tax rate on consumption.[11]
The three disagree on the size of the gains, and the differences are instructive. Without VAT the OECD accounts show the richest fifth breaking even, because its members receive large pensions, including the higher pensions of former civil servants, which that method counts as a benefit. With VAT the richest fifth pays in about 15 per cent and the fourth fifth about 11, a steeper result than the World Inequality Database’s. But all three have the same shape. The poorest fifth gains roughly its own pre-tax income again or more; the second and middle fifths gain; the crossing point lies in the fourth fifth; and the richest fifth pays in.
Forty years of the ledger
The same calculation for every year since 1980 shows how the balance built up (Figure 6). In 1980 the poorer half of Portuguese adults gained about 15 per cent over their pre-tax income. The gain rose steeply in the 1990s, to 33 per cent in 2000, and peaked at about 39 per cent in 2013, when the crisis had cut pre-tax incomes while the services continued; by 2019 it was 30 per cent. The richest tenth’s net contribution rose from about 9 per cent in 1980 to 21 in 2013 and 18 in 2019. The middle four-tenths have stayed within about two points of breaking even throughout.[4]
The period covers the building of the Portuguese welfare state: the national health service created in 1979, the expansion of schooling, the minimum-income benefit introduced in 1996. That the gains grew as those services grew, and that the middle has stayed at zero while they did, is the ledger’s long-run lesson: the services that Portuguese adults receive are paid for, in net terms, by the top of the distribution, while the middle pays for roughly what it receives.
What the ledger shows
Read with its assumptions, the ledger says four things. The Portuguese State collects less than the European average, leans more on VAT and other taxes on products, and spends more of what it collects on pensions and less on working-age support. Counting pensions as deferred pay, the poorer half of adults receives much more in public services and benefits than it pays in taxes, mostly because health and education are worth about the same to everyone and are a large share of a small income. The middle four-tenths pay about what they receive. The richest tenth pays in, net, about a sixth of its pre-tax income.
This does not settle the complaint in the earlier essay, that the middle pays for its own services almost entirely out of wages, at tax rates that rise quickly, while income from capital pays a flat rate. The ledger is about how much each group pays in total, not about which income it is taken from. Both are true at once: the middle breaks even, and it breaks even on its wages.
This piece was written collaboratively with Claude Opus 5.5 (Anthropic): human specification, editorial direction and critical review; machine data research, analysis and drafting. The figures and derived numbers are computed by scripts/state_ledger.py from Eurostat’s government finance statistics (tax receipts, expenditure by function and main aggregates, downloaded 28 September 2026), EU-SILC poverty rates before and after transfers, the World Inequality Database bulk file for Portugal (average pre-tax income, post-tax disposable income and post-tax national income by percentile, adults aged 20 and over, equal split within couples, at 2025 prices), and the OECD’s experimental distributional accounts for 2016–2017. The description of the World Inequality Database method is taken from its published methodology papers and the metadata of the Portuguese series; where these disagree, the text follows the metadata. The OECD estimate with VAT is an illustrative sensitivity, not a published figure. Results are in docs/state-ledger-results.json and the figures in docs/state-ledger-figures.html.
The cover photograph is the Palácio de São Bento, seat of the Assembleia da República, by Alvesgaspar, April 2009; CC BY-SA 3.0, via Wikimedia Commons, cropped.
Authored by: Luis Matos Ferreira — Physicist, Developer, Writer
- Earned and Unearned — the split of income between work and capital, and how each is taxed.
- The Sustainability Story — the Portuguese pension system and its forecasts.
- The Two-Thirds Country — Portugal’s productivity, hours and wages.
- The Duty to Work — how the centrality of work was instilled.
- Eurostat, Main national accounts tax aggregates (gov_10a_taxag), general government, Portugal and EU-27, 2022 and 2024, % of GDP and EUR million.
- Eurostat, Government revenue, expenditure and main aggregates (gov_10a_main), Portugal, 2022 and 2024.
- Eurostat, General government expenditure by function (COFOG) (gov_10a_exp), Portugal and EU-27, 2022 and 2024.
- World Inequality Database, Portugal: pre-tax national income, post-tax disposable income and post-tax national income by percentile (aptinc, acainc, adiinc; adults 20+, equal split), 1980–2024, bulk data of September 2026, with variable metadata.
- Blanchet, Chancel & Gethin, “Why Is Europe More Equal than the United States?”, American Economic Journal: Applied Economics 14(4), 480 (2022); World Inequality Lab working paper 2020/19, version of 9 November 2021, pp. 9–16, 28–29 and Table I.
- Eurostat, EU-SILC 2025: at-risk-of-poverty rate after social transfers (ilc_li02), before social transfers, pensions included in social transfers (ilc_li09), and pensions excluded from social transfers (ilc_li10).
- Fisher-Post & Gethin, “Government Redistribution and Development: Global Estimates of Tax and Transfer Progressivity, 1980–2019”, World Inequality Lab working paper 2023/17, pp. 14–15.
- Chancel, Flores, Moshrif, Nievas & Piketty (eds.), Distributional National Accounts Guidelines, World Inequality Lab, 2025, sections 2.2.3–2.2.5.
- Andreescu & Sodano, “2025 Regional DINA Update for Europe”, World Inequality Lab technical note 2025/05, December 2025, pp. 3–5 and Table 1.
- Piketty, Saez & Zucman, “Income Inequality in the United States: A Comment”, World Inequality Lab technical note 2024/04, pp. 13–14 and Table 2.
- OECD Expert Group on Disparities in National Accounts, household income by quintile, Portugal, 2016–2017; European Commission, DG TAXUD, Data on Taxation Trends, implicit tax rate on consumption, 2017.
Comentários
Enviar um comentário