Finnish economy 2040

Public finance simulator

How the model calculates

This page has everything the calculation uses: the data, every rule with its formula, all assumptions and the cause-and-effect chains. The figures come straight from the same service the simulator uses.

Overview

The calculation runs year by year from 2025 to 2040. Each year goes through these steps:

  1. Base year: actual figures
  2. The economy
  3. Taxes and other revenue
  4. Spending and its drivers
  5. The government's decisions
  6. Your spending changes
  7. Your tax changes (static effect)
  8. Cause-and-effect chains
  9. Growth effects (optional, disputed)
  10. Supply-side effects (optional, disputed)
  11. Interest
  12. Calibration to the Ministry of Finance forecast
  13. Fiscal balance and debt
  14. Uncertainty (fan charts)

Notation

t = year, t−1 = the year before. Money in millions of euro. GDP = gross domestic product. P65+ = number of people aged 65+ in the population projection.

Data used

The starting data are Statistics Finland statistics, fetched from its API automatically, and Ministry of Finance publications whose figures were transcribed by hand. Every figure's source and release date is on the Sources page.

Base year 2025: revenue, expenditure and taxes

Revenue and expenditure of general government (central government, municipalities, wellbeing services counties, social security funds). The lines add up exactly to Statistics Finland's published totals. The codes are national accounts transactions and tax categories.

Revenue

ItemAmountCodesTable
Taxes on production and imports€39.5bnD2Rntp/15ah
Current taxes on income and wealth€44.7bnD5Rntp/15ah
Social security contributions€34.1bnD61Rntp/15ah
Property income€11.2bnD4RSntp/15ah
Sales and fees€18.1bnP11R + P12R + P131Rntp/15ah
Other current transfers€2.0bnD7RSntp/15ah
Capital transfers (incl. capital taxes)€1.5bnD9RSntp/15ah
Subsidies received€0.0bnD39Rntp/15ah
Total€151.2bn

Expenditure

ItemAmountCodesTable
Social benefits in cash€53.1bnD62Kntp/15ah
Compensation of employees€38.1bnD1Kntp/15ah
Intermediate consumption€32.2bnP2Kntp/15ah
Gross capital formation€12.8bnP5Kntp/15ah
Social transfers in kind (purchased)€9.7bnD632Kntp/15ah
Other current transfers€6.7bnD7KSntp/15ah
Property expenditure (mainly interest)€4.6bnD4KSntp/15ah
Subsidies€3.4bnD3Kntp/15ah
Capital transfers€0.9bnD9KSntp/15ah
Other expenditure€0.3bnD29K + D5K + NPntp/15ah
Total€162.0bn

Taxes by type

ItemAmountCodesTable
Personal income tax€36.2bn110002ntp/15aj
Corporate income tax€7.1bn120001ntp/15aj
Employers' social contributions€21.6bn2200ntp/15aj
Employees' and self-employed social contributions€12.3bn2100ntp/15aj
Value added tax€27.1bn511101ntp/15aj
Energy excise€3.7bn512102ntp/15aj
Alcohol excise€1.4bn512101ntp/15aj
Tobacco excise€1.2bn512105ntp/15aj
Other excises€0.8bn5121 − 512101 − 512102 − 512105ntp/15aj
Real-estate tax€2.3bn410001ntp/15aj
Inheritance and gift tax€1.1bn430001ntp/15aj
Transfer tax€0.7bn440001ntp/15aj
Other taxes€3.8bnSSSntp/15aj
Total€119.2bn

Ministry of Finance forecast

Up to 2030 the development of the economy and the main lines of public finances come from this forecast. Ministry of Finance Economic Survey, autumn 2026 (VM 2026:35). Licence: CC BY-NC-ND 4.0

202520262027202820292030
GDP, bn €281.7292.9304.6316.9329.0340.7
GDP growth, %0.81.61.81.51.51.3
Unemployment rate, %9.710.39.99.58.98.5
Inflation, %0.32.01.92.12.02.0
Earnings growth, %3.13.63.23.43.23.2
Long-term interest rate, %3.03.33.33.23.13.0
Fiscal balance, % of GDP-3.8-4.2-4.5-4.5-4.6-4.4
Debt, % of GDP88.590.191.293.195.397.6

Population projection

Statistics Finland's population projection drives age-related spending (education, pensions, health).

2025203020352040
Aged 0–5283,406279,089284,488290,591
Aged 6–12395,858357,643348,586355,071
Aged 13–18389,011372,181330,847315,501
Aged 19–29724,800762,264780,346738,157
Aged 30–642,530,4082,585,5322,674,1472,809,889
Aged 65–74672,019666,480651,935613,934
Aged 75+674,722766,645829,237875,981

The government's decisions

Cumulative effect of the decisions on public finances, bn €. A positive value strengthens the fiscal balance. The lever column shows which lever scales a line; fixed lines can't be changed because they cancel out within general government. Julkisen talouden suunnitelma 2027–2030, VM 2026:17, taulukko 33. Licence: CC BY-NC-SA 4.0

Item20232024202520262027202820292030Lever
Central government appropriation cuts0.00.52.73.94.84.95.15.2Government's spending cuts
Central government appropriation increases-0.1-0.3-1.0-0.9-1.2-1.8-4.1-4.3Government's spending increases (incl. defence)
Investment programme0.0-0.3-0.7-0.9-0.8-0.8-0.4-0.1Government's spending increases (incl. defence)
Central government tax increases0.01.02.53.13.43.53.63.6Government's tax increases
Central government tax cuts (excl. index adjustments)0.0-0.6-1.3-2.8-4.1-4.1-4.3-4.3Government's tax cuts
Other central government revenue increases0.00.00.00.11.30.20.20.2

fixed

Compensation to municipalities for tax changes0.00.00.40.40.40.40.40.4

fixed

Effects of social contribution changes0.00.50.50.40.40.40.40.3

fixed

Wellbeing counties: revenue increases0.00.00.10.40.70.91.01.1

fixed

Wellbeing counties: revenue cuts0.0-0.1-0.6-0.9-1.3-1.6-1.8-1.8

fixed

Wellbeing counties: expenditure cuts0.00.10.60.91.31.61.81.8

fixed

Wellbeing counties: expenditure increases0.00.0-0.1-0.4-0.7-0.9-1.0-1.1

fixed

Municipalities: revenue measures0.00.0-0.3-0.4-0.4-0.4-0.4-0.4

fixed

Municipalities: expenditure measures0.00.00.20.10.00.00.00.0

fixed

Net effect of state measures on municipal tax revenue0.00.10.50.50.50.50.50.5

fixed

Compensation received for tax changes0.00.0-0.4-0.4-0.4-0.4-0.4-0.4

fixed

Other state measures affecting municipalities0.00.00.30.30.30.30.30.3

fixed

Effects of social contribution changes0.00.30.20.10.10.10.10.1

fixed

Decisions' effect on social security fund expenditure0.00.71.62.21.42.22.32.3Government's social security cuts
Decisions' effect on social security fund revenue0.0-0.3-1.3-1.7-1.8-1.8-1.7-1.7

fixed

Changes in social contributions0.0-2.0-1.2-0.9-1.0-1.0-1.1-1.2

fixed

All sources with release dates

Rules and formulas

Every step with its formulas, in the order it is calculated. The links lead to the assumptions and levers it uses.

1. Base year: actual figures

Revenue(2025) = Σ revenue lines = OTRSExpenditure(2025) = Σ expenditure lines = OTESComponent_f(2025) = (Expenditure − Interest) × COFOG_f / (COFOG total − debt transactions)

The calculation starts from the latest year of Statistics Finland's national accounts. Revenue and expenditure are summed from published transactions (e.g. D2R, D62K) and match the published totals exactly. Spending is split into functions (health, education, social protection…) with the shares of the latest COFOG statistics, which are published a year later. Old-age spending is split into pensions (cash social benefits, D62K) and elderly care services (everything else), and student aid (cash benefits in education) is separated from education.

2. The economy

GDP(t) = GDP(t−1) × Ministry GDP(t) / Ministry GDP(t−1), when t ≤ 2030GDP(t) = GDP(t−1) × (1 + g) × (1 + p), when t > 2030Wages(t) = (1 + inflation) × (1 + g) / (P19–64(t) / P19–64(t−1)) × (1 + r) − 1, when t > 2030

Up to 2030, growth, inflation, wages, unemployment and interest rates come from the Ministry of Finance forecast. After that the model's assumptions apply: g is real growth and p the price increase. Real wages rise with output per working-age person (r is an optional gap to that). Growth responds to your changes only when growth or supply-side effects are on.

Uses:Trend real GDP growth 2031–GDP deflator 2031–Consumer price inflation 2031–Real wages compared with productivity 2031–

3. Taxes and other revenue

Tax(t) = Tax(t−1) × [GDP(t) / GDP(t−1)]^εOther revenue(t) = Other revenue(t−1) × GDP(t) / GDP(t−1)

Each tax grows with the economy. The elasticity ε says how much tax revenue grows when GDP grows one percent; a value of 1 keeps the tax ratio constant.

4. Spending and its drivers

Component(t) = Component(t−1) × driver growth(t)

Each spending component grows with its own driver: for example pensions with the number of people aged 65+ and the pension index, health spending with ageing and wages. The drivers' formulas are in their own table below.

5. The government's decisions

Decisions(t) = Σ lines [s × level(t) − level(2025)]s = 1 + (lever / 100 − 1) × phase(t)

The Ministry of Finance table of the government's decisions is added as the change since the base year, because some decisions already show in the 2025 figures. After 2030 the decisions keep their share of GDP. The lever s scales them: 100% = as decided, 0% = reversed from the start year.

6. Your spending changes

Change(t) = p / 100 × Component(t) × phase(t)Part change(t) = [(1 + p × phase(t) / 100) × (1 + q × phase(t) / 100) − 1] × Component(t)phase(t) = 0 when t < start year; otherwise min(1, (t − start year + 1) / phase-in years)

A spending lever p changes a function's spending by a percentage of what it would be on the current path. Health, education and social protection also have part levers q (e.g. hospitals) that change a part on top of the change to the whole function. A cause-and-effect chain triggered by a function's lever is also triggered by its part levers. The change starts in the start year and can phase in gradually.

7. Your tax changes (static effect)

ΔVAT = VAT(t) × general-rate share × Δpp / 25.5ΔCorporate tax = Corporate tax(t) × Δpp / 20ΔHigh earners' income tax = Income tax(t) × high earners' share × p / 100ΔOther tax = Tax(t) × p / 100

The static effect of a tax change is calculated from current revenue assuming behaviour doesn't change. Behavioural effects (consumption, migration, profit shifting) are separate cause-and-effect chains you can switch on or off.

8. Cause-and-effect chains

Effect(t) = s × T(t) × shape(n)

T(t) is the lever's static effect on the fiscal balance (a saving or extra revenue is positive), s is the share and n the years since the effect began. The effect is added to its target line. The shapes' formulas and every effect are described below.

9. Growth effects (optional, disputed)

gap(t) = ρ × gap(t−1) − c(t) × [m_g × ΔT_g(t) + m_t × ΔT_t(t)] / GDP(t), c(t) = max(0, 1 + s × [unemployment(t) − unemployment*])GDP'(t) = GDP(t) × (1 + gap(t)), unemployment'(t) = unemployment(t) − Okun × 100 × gap(t)market rate'(t) = market rate(t) + p × [debt ratio(t−1) − current-path debt ratio(t−1)] / 100

By default the changes don't affect economic growth. With growth effects switched on, your changes' static effect on the balance (T; tightening is positive) moves GDP by a multiplier: one for spending (consumption and investment, m_g) and one for taxes and social benefits (m_t). Only the change in tightening from the year before matters, and the effect fades each year by the factor ρ, so even a permanent change mainly affects GDP in the years it is introduced. Lower GDP lowers tax revenue and raises unemployment, unemployment spending and the debt ratio. In addition, a debt ratio above the current path raises the rate on new borrowing by p basis points per percentage point (and a lower one lowers it). The multipliers can also be made to depend on the cycle (s > 0): they are larger in years when unemployment on the current path is above its long-run level (unemployment*). By default s = 0. The size of the multipliers, whether they depend on the cycle and how long the effect lasts are disputed among economists.

10. Supply-side effects (optional, disputed)

target_c(t) = −(s_c / R(t)) × T_c(t) / GDP(t), level_c(t) = level_c(t−1) + speed_c × [target_c(t) − level_c(t−1)]GDP'(t) = GDP(t) × (1 + gap(t) + level_work(t) + level_capital(t))

Off by default (self-financing share 0%). When set, a lasting tax change moves the level of GDP for good: taxes on work (income taxes and social contributions) through how much people work, and company and capital taxes through investment. The size is given as a self-financing share s: how much of a tax's static revenue comes back in the long run as revenue from a bigger economy (R is revenue as a share of GDP). The effect reaches its full size in about three years for work and about ten for investment (speeds 0.4 and 0.15). Unlike the demand effect it doesn't fade and doesn't change the unemployment rate. If behavioural tax effects are on at the same time, part of the same response is counted twice.

11. Interest

Average rate(t) = Average rate(t−1) × (1 − r) + Long rate(t) × rInterest(t) = Average rate(t) × Debt(t−1)Average rate(2025) = Interest(2025) / Debt(2024)

Each year a share r of the debt matures and is refinanced at the market rate, so the average rate on debt gradually approaches the market rate. As debt grows, interest grows.

12. Calibration to the Ministry of Finance forecast

Calibration(t) = Ministry revenue(t) − model revenue(t), when t ≤ 2030Calibration(t) = Calibration(2030) × GDP(t) / GDP(2030), when t > 2030

So that the current path matches the official forecast, the difference between the model and the forecast is added as a visible calibration line (likewise for spending). Without calibration the model stays within 0.6 percentage points of the forecast deficit and 1.5 points of the debt ratio up to 2030.

13. Fiscal balance and debt

Net lending(t) = Revenue(t) − Expenditure(t)Debt(t) = Debt(t−1) − Net lending(t) + Stock-flow adjustment(t)Debt ratio(t) = Debt(t) / GDP(t) × 100Balance without pension schemes(t) = Net lending(t) − employment pension schemes' surplus(t)

A deficit adds to debt. In addition, debt grows about one percent of GDP a year more than the deficit, mainly because pension funds' surplus is invested rather than used to repay debt. Up to 2030 this stock-flow adjustment comes from the Ministry of Finance forecast. The balance is also shown without the employment pension schemes' surplus, which up to 2030 comes from the Ministry of Finance's fiscal plan and after that from the assumption.

14. Uncertainty (fan charts)

assumption ~ triangular(low, chosen value, high)GDP'(t) = GDP(t) × Π (1 + e_k / 100), e_k ~ N(0, σ_GDP)rate'(t) = rate(t) + Σ u_k, u_k ~ N(0, σ_rate)

The shading in the charts comes from running the calculation 300 times. Each time the uncertain assumptions are drawn from their plausible range (around the chosen value), and GDP growth and the interest rate get random yearly swings that accumulate. The shading shows the 10th–90th and 25th–75th percentiles. The assumptions are drawn independently of each other, and the calculation includes no policy decisions that rising debt would probably prompt, so the range is very wide in the long run. The ranges and the size of the swings are themselves estimates.

Spending drivers

Each spending component grows from year to year with its own driver.

DriverFormulaSpending components
Nominal GDP (the component keeps its share of GDP)
Spending(t) = Spending(t−1) × GDP(t) / GDP(t−1)
General public services: administration, taxation and foreign affairs, General public services: development aid, General public services: general services (personnel, premises, IT …), General public services: basic research, General public services: other (excl. interest), Defence: military defence, Defence: foreign military aid, Defence: other (civil defence, R&D …), Public order and safety, Economic affairs: business and employment services, Economic affairs: agriculture, forestry and fishing, Economic affairs: transport, Economic affairs: research and development, Economic affairs: other (energy, industry …), Environmental protection, Housing and community amenities, Recreation, culture and religion
Population aged 65+ × pension index × average pension drift
Spending(t) = Spending(t−1) × P65+(t) / P65+(t−1) × [1 + w × inflation(t) + (1 − w) × wages(t)] × (1 + d)
Social protection: old age (pensions)
Unemployment rate × working-age population × wages
Spending(t) = Spending(t−1) × u(t) / u(t−1) × P19–64(t) / P19–64(t−1) × (1 + wages(t))
Social protection: unemployment
Working-age population × wages (earnings-related allowances and disability pensions)
Spending(t) = Spending(t−1) × P19–64(t) / P19–64(t−1) × (1 + wages(t))
Social protection: sickness and disability
Population aged 0–18 × wages (early childhood education is labour-intensive)
Spending(t) = Spending(t−1) × P0–18(t) / P0–18(t−1) × (1 + wages(t))
Social protection: families and children (incl. early childhood education)
Population × consumer prices
Spending(t) = Spending(t−1) × Population(t) / Population(t−1) × (1 + inflation(t))
Social protection: housing benefits, Social protection: social assistance and other social exclusion, Social protection: other (survivors' pensions …)
Age-weighted population × wages × real cost drift
Spending(t) = Spending(t−1) × W(t) / W(t−1) × (1 + wages(t)) × (1 + r), where W = P0–64 + a × P65–74 + b × P75+
Health: hospital services, Health: outpatient care, Health: medicines and equipment, Health: other (public health, R&D …)
Population aged 6–12 × wages
Spending(t) = Spending(t−1) × P6–12(t) / P6–12(t−1) × (1 + wages(t))
Education: pre-primary and grades 1–6
Population aged 13–18 × wages
Spending(t) = Spending(t−1) × P13–18(t) / P13–18(t−1) × (1 + wages(t))
Education: grades 7–9 and upper secondary
Population aged 19–29 × wages
Spending(t) = Spending(t−1) × P19–29(t) / P19–29(t−1) × (1 + wages(t))
Education: universities and polytechnics
Population aged 0–18 × wages
Spending(t) = Spending(t−1) × P0–18(t) / P0–18(t−1) × (1 + wages(t))
Education: other (adult education, support services …)
Population aged 19–29 × consumer prices (the study grant is tied to the national pension index)
Spending(t) = Spending(t−1) × P19–29(t) / P19–29(t−1) × (1 + inflation(t))
Education: student aid
Population aged 75+ × wages (care is labour-intensive)
Spending(t) = Spending(t−1) × P75+(t) / P75+(t−1) × (1 + wages(t))
Social protection: old age (elderly care services)

Assumptions

Assumptions about how the economy develops. You can change them in the simulator's Model assumptions section.

NameDefaultRangeConfidenceBasis

Trend real GDP growth 2031–

Real growth after the forecast period. Defaults to the Ministry of Finance's growth for its last forecast year (2030).

1.3%0.0% … 3.0%
Confidence: Medium
Ministry of Finance, Economic Survey autumn 2026 (VM 2026:35)Ministry of Finance, background analyses for the macroeconomic forecast 2025 (VM 2025:57): long-term potential growth settles at 1.3% (in Finnish)

Matches the Ministry of Finance's estimate of long-term potential growth (1.3%). In 2020–2024 potential output grew only 0.8% a year.

GDP deflator 2031–

Annual change in the price of output after the forecast period.

2.0%0.0% … 5.0%
Confidence: Medium

Consumer price inflation 2031–

Drives price-indexed benefits and pensions after the forecast period.

2.0%0.0% … 5.0%
Confidence: Medium

Real wages compared with productivity 2031–

After 2030, real wages rise at the same pace as output per working-age person (trend GDP growth minus growth of the working-age population, about 0.7% a year). With this you can assume wages rise faster (+) or slower (−) than that. Wages drive wage-intensive spending (health, education, elderly care) and part of the pension index.

0.0 pp−1.5 pp … 1.5 pp
Confidence: Medium
Ministry of Finance, background analyses for the macroeconomic forecast 2025 (VM 2025:57): long-term potential growth settles at 1.3% (in Finnish)

In the long run real wages follow productivity. In Statistics Finland's population projection the working-age population grows by about 0.6% a year in the 2030s thanks to immigration, so of 1.3% GDP growth about 0.7% is output per working-age person. With wages tied to this, changing the growth assumption also changes wages, not only tax revenue. (The Ministry of Finance's sustainability calculation has productivity growing 1.3%, but with hardly any growth in labour input.)

Unemployment rate 2031–

Drives unemployment spending after the forecast period. Defaults to the Ministry's forecast for 2030.

8.5%3.0% … 15.0%
Confidence: Medium

Long-term interest rate 2031–

10-year government bond yield at which new and maturing debt is financed. Defaults to the Ministry's forecast for 2030.

3.0%0.0% … 8.0%
Confidence: Medium

Share of debt refinanced each year

How fast the average interest rate on debt moves to the market rate: the share of debt whose interest rate is reset each year.

19%5% … 40%
Confidence: Medium
State Treasury, Debt Management Annual Review 2025 (average refixing time 5.39 yrs, maturity 7.84 yrs; in Finnish)

Derived: 1 / the average interest-rate refixing time of central government debt, 5.39 years (end of 2025) ≈ 19%. This is central government; local government and wellbeing county debt is shorter, so the general government share may be slightly higher.

Stock-flow adjustment 2031–

Debt grows faster than the deficit each year, mainly because pension-fund surpluses are invested rather than used to repay debt. Default is the 2030 value implied by the Ministry's forecast.

1.1% of GDP−1.0% of GDP … 3.0% of GDP
Confidence: Medium
Ministry of Finance, Economic Survey autumn 2026 (VM 2026:35)Eurostat: stock-flow adjustment between deficit and debt

Derived: change in debt − deficit in the Ministry's 2030 forecast (18.9 − 15.1 bn €).

Tax revenue elasticity to GDP

How much tax revenue grows when nominal GDP grows 1%. A value of 1 keeps the tax ratio constant.

1.00 ×0.50 × … 1.50 ×
Confidence: Medium

No source – estimate

Share of VAT revenue at the general rate

Used to compute the static effect of a change in the general VAT rate.

73%30% … 90%
Confidence: Medium
Government bill HE 61/2024: general VAT rate 24 → 25.5%, static effect about €1,167m a year (in Finnish)

Derived: HE 61/2024's static effect of €1,167m / 1.5 pp × 25.5% ≈ €19.8bn at the general rate, about 73% of 2025 VAT revenue (€27.1bn). The bill's tax base is a forecast, and the reduced rates changed in 2025.

High earners' share of income tax

How much of earned and capital income tax the top income decile (earning roughly over €6,000 a month) pays. Splits the income tax levers into high earners and low and middle incomes.

43%20% … 70%
Confidence: High
Tax Administration, personal taxes and payments by income class, tax year 2024 (income taxes €38.1bn in total; in Finnish)Yle 2019: the top income decile paid about €13.8bn of about €30bn in income taxes (Tax Administration data, tax year 2018; in Finnish)

In tax year 2024 the top tenth of taxpayers (income over about €72,000 a year) paid 41% of income taxes (€38.1bn). Counted over all income recipients, as in the tax-year 2018 figure of 13.8 / 30 bn €, the share is about 44%. Everyone earning over €60,000 (15% of taxpayers) paid 51%.

Capital income tax's share of income tax

How much of personal income tax is tax on capital income (dividends, capital gains, rents, interest). The capital income tax lever changes this part.

10%5% … 20%
Confidence: High
Tax Administration, personal taxes and payments by income class, tax year 2024 (income taxes €38.1bn in total; in Finnish)

In tax year 2024 capital income tax was €3.58bn, 9.4% of income taxes (€38.1bn); in 2019–2024 the share varied between 8.9% and 12.4%. It moves with the stock market, because capital gains are realised in good years.

Price weight of the pension index

Pensions in payment are indexed with prices weighted 80% and wages 20%.

80%0% … 100%
Confidence: High

Average pension growth above the index

The average pension rises faster than the pension index, because new wage-based pensions replace older, smaller ones (cohort effect).

0.3%0.0% … 2.5%
Confidence: Medium
Finnish Centre for Pensions: average total pension of old-age pensioners 2012–2025 and changes in the earnings-related pension index (in Finnish)2024 Ageing Report, country fiche for Finland: benefit ratio (average pension / average wage) 51% (2022), 50% (2030), 47% (2040)

History: the average pension of old-age pensioners grew on average 0.8% a year faster than the index in 2013–2025, but only about 0.5% in 2023–2025. Future: in the Ageing Report the average pension relative to the average wage falls from 50% to 47% in 2030–2040 (life expectancy coefficient), which means about 0.8 × real wage growth − 0.6% a year: about 0% with real wages growing 0.7% (this calculation's assumption) and 0.5% at 1.4% (the report's assumption). The default 0.3% lies between these and history. If you change real wage growth, move this in the same direction by about 0.8 times as much.

Health spending: relative weight of ages 65–74

Average health and care spending on a person aged 65–74, as a multiple of a person under 65.

2.0 ×1.0 × … 5.0 ×
Confidence: Medium
Kapiainen & Eskelinen: Health expenditure by age and sex 2011 (THL report 17/2014; in Finnish)

According to THL, health spending in 2011 was €4,867 per person aged 65–74 and about €2,220 per person under 65 (2.2 times). With elderly institutional care and service housing removed (in this calculation they are in social protection, old age), the ratio is about 2.0. The data is from 2011; the ratios change slowly.

Health spending: relative weight of ages 75+

Average health and care spending on a person aged 75+, as a multiple of a person under 65.

3.6 ×1.0 × … 8.0 ×
Confidence: Medium
Kapiainen & Eskelinen: Health expenditure by age and sex 2011 (THL report 17/2014; in Finnish)

According to THL, health and care spending in 2011 was about €12,800 per person aged 75+ (5.8 times a person under 65). With elderly institutional care, service housing, dementia units and informal-care support (€2.2bn, in this calculation in social protection, old age) removed, the ratio is about 3.6. One weight for the whole 75+ group slightly understates growth, because the share of people over 85 in it is rising and their spending is about twice that of 75–84-year-olds.

Real cost drift in health care

Annual cost growth beyond ageing and wages (e.g. new treatments).

0.1%−1.0% … 2.0%
Confidence: Medium
2024 Ageing Report: income elasticity of health spending 1.1 (baseline) and 1.5 (risk scenario), converging to 1 by 2070

In this calculation health spending already rises with wages. In the EU Ageing Report baseline, spending per person grows 1.1 times as fast as income (falling to 1 by 2070), about 0.1% a year extra over this period. In its risk scenario (elasticity 1.5, mainly new treatments) the extra growth is about 0.4% a year. In the international literature health spending has grown 1–2% a year faster than GDP per person, but part of that is wage growth, which this calculation already includes.

Employment pension schemes' surplus 2031–

Employment pension schemes collect contributions into funds for future pensions, so they run a surplus. It can't be used for other spending, so the calculation also shows the balance without it.

1.1% of GDP0.0% of GDP … 2.0% of GDP
Confidence: Medium
Ministry of Finance, General Government Fiscal Plan 2027–2030 (VM 2026:17), annex 3, tables 39–43: net lending by subsector (in Finnish)

Up to 2030 the calculation uses the Ministry of Finance's fiscal plan forecast (€2.8bn in 2026 → €3.8bn in 2030, about 1.1% of GDP); after that the surplus is held at its 2030 share. In reality the surplus depends on investment income and pension spending; the share is an estimate.

Spending multiplier

How many euros GDP changes when government consumption and investment change by one euro (in the first year).

0.6 ×0.0 × … 1.5 ×
Confidence: Medium

Typically about 0.3–1.0 in the literature; larger in a downturn and when monetary policy doesn't offset it (Finland in the euro area). The size is disputed.

Tax and transfer multiplier

How many euros GDP changes when taxes or social benefits change by one euro. Smaller than the spending multiplier because part of the change goes into saving.

0.3 ×0.0 × … 1.5 ×
Confidence: Medium

Typically about 0.2–0.6 in the literature; larger for changes aimed at low and middle incomes. Estimate.

Cycle sensitivity of the multipliers

How much the multipliers grow for each percentage point by which unemployment (on the current path) is above its long-run level, and shrink when it is below. 0 = the same multipliers every year. E.g. 0.25: with unemployment 2 percentage points above its long-run level, the multipliers are 1.5 times as large.

0.00 ×0.00 × … 0.50 ×
Confidence: Speculative
Auerbach & Gorodnichenko (2012): measuring the output responses to fiscal policy (recession vs expansion)Ramey & Zubairy (2018): government spending multipliers in good times and in bad – evidence from US historical data

Disputed, so 0 by default. According to Auerbach and Gorodnichenko (2012) multipliers are much larger in a recession (about 1.5–2) than in an expansion (near 0), but Ramey and Zubairy (2018) find little difference in long US data. The long-run level is the 'Unemployment rate 2031–' assumption, at which the Ministry of Finance forecast has unemployment back at its structural level by 2030.

Persistence of the growth effect

Share of the GDP effect left the next year. 0.5 = half fades each year. Even a permanent change's effect on GDP fades in the end as the economy returns to its potential.

0.5 ×0.0 × … 0.9 ×
Confidence: Low

Estimate. Some researchers argue that long downturns leave a permanent mark (hysteresis), in which case the effect doesn't fade completely.

Okun's coefficient

How many percentage points the unemployment rate rises when GDP falls one per cent below the current path. Drives unemployment spending.

0.50 ×0.00 × … 1.00 ×
Confidence: Medium

A standard rule of thumb; the real value varies between countries and over time. Estimate, to be refined from Finnish data.

Interest effect of debt

How many basis points the interest rate on new borrowing rises when the debt ratio is one percentage point above the current path. Works the other way too.

3.0 bp0.0 bp … 10.0 bp
Confidence: Low

Typically about 1–5 basis points per percentage point of debt in the literature; much more in crises and when confidence is weak. In the euro area the rate level also depends on the ECB. Estimate.

Self-financing share of taxes on work

How much of a change in taxes on work (income taxes and social contributions) comes back in the long run because people work more or less. E.g. 20%: when taxes are cut by €1bn, the economy grows enough for about €0.2bn to come back as tax revenue. 0 = no effect (the default).

0%0% … 100%
Confidence: Low
Kiema & Koistinen (Labore, 16 May 2025): the growth effects of tax changes are overstated (in Finnish)FIT – Finnish Centre of Excellence in Tax Systems Research: statement on fiscal consolidation measures (1 April 2026; in Finnish)Saez, Slemrod & Giertz (2012): the elasticity of taxable income – a review

Disputed. Research finds that moderate income tax changes across the whole income scale hardly change how much people work (FIT 2026); with usual elasticities of taxable income the self-financing share works out at about 10–30%. The Ministry of Finance has assumed that a cut in the top marginal tax rate pays for itself fully (100%), which some researchers consider over-optimistic (Labore 2025). Suggested with the switch: 20%.

Self-financing share of company and capital taxes

How much of a change in corporate and capital income taxes comes back in the long run because firms invest more or less. The effect builds slowly, over about ten years. 0 = no effect (the default).

0%0% … 100%
Confidence: Low
VATT: the corporate tax cut did not increase investment but boosted small firms' activity (Harju, Koivisto & Matikka)Kiema & Koistinen (Labore, 16 May 2025): the growth effects of tax changes are overstated (in Finnish)

Disputed. According to VATT research, the 2014 corporate tax cut (24.5 → 20%) did not increase small firms' investment in the first years, although their turnover grew. International studies rank company taxes among the most harmful to growth, but estimates of the size vary a lot. Suggested with the switch: 10%.

Year-to-year variation in GDP growth

In the fan charts, each year's GDP growth gets a random deviation with this standard deviation. The deviations accumulate in the level of GDP.

1.5 pp0.0 pp … 4.0 pp
Confidence: Low

Estimate: Finland's GDP growth has varied with a standard deviation of about 2–3 percentage points since 2000 (incl. the 2009 and 2020 collapses); forecast errors are smaller than that.

Year-to-year variation in the interest rate

In the fan charts, the long-term interest rate moves randomly each year with this standard deviation, and the moves accumulate.

0.4 pp0.0 pp … 1.5 pp
Confidence: Low

Estimate; annual changes in the 10-year government bond yield have typically been a few tenths of a point, but in 2022 the yield rose by over 2 points.

Levers

Levers are policy choices you can change in the simulator.

NameDefaultRangeConfidenceBasis

Year changes start

The year in which changes to spending, taxes and the government's decisions take effect.

20272026 … 2035
Confidence: High

No source – estimate

Phase-in period

Number of years over which a change reaches its full size.

1 yrs1 yrs … 5 yrs
Confidence: High

No source – estimate

General public services (excl. interest)

Change to the whole function from the current path, %. The detailed sliders below change individual parts on top of this change.

0%−50% … 50%
Confidence: High

No source – estimate

General public services (excl. interest): Administration, taxation and foreign affairs

Change to this part from the current path, %, on top of the change to the whole function.

0%−50% … 50%
Confidence: High

No source – estimate

General public services (excl. interest): Development aid

Change to this part from the current path, %, on top of the change to the whole function.

0%−100% … 100%
Confidence: High

No source – estimate

General public services (excl. interest): General services (personnel, premises, IT …)

Change to this part from the current path, %, on top of the change to the whole function.

0%−50% … 50%
Confidence: High

No source – estimate

General public services (excl. interest): Basic research

Change to this part from the current path, %, on top of the change to the whole function.

0%−50% … 50%
Confidence: High

No source – estimate

General public services (excl. interest): Other general public services

Change to this part from the current path, %, on top of the change to the whole function.

0%−50% … 50%
Confidence: High

No source – estimate

Defence

Change to the whole function from the current path, %; the detailed sliders change parts on top of it. The slider goes up to +100% so that large increases such as NATO's 3.5%-of-GDP target can be tried; the current path already includes the increases decided so far.

0%−50% … 100%
Confidence: High

No source – estimate

Defence: Military defence

Change to this part from the current path, %, on top of the change to the whole function.

0%−50% … 100%
Confidence: High

No source – estimate

Defence: Foreign military aid

Change to this part from the current path, %, on top of the change to the whole function.

0%−100% … 100%
Confidence: High

No source – estimate

Defence: Other (civil defence, R&D …)

Change to this part from the current path, %, on top of the change to the whole function.

0%−50% … 50%
Confidence: High

No source – estimate

Public order and safety

Change from the current path, %.

0%−50% … 50%
Confidence: High

No source – estimate

Economic affairs and transport

Change to the whole function from the current path, %. The detailed sliders below change individual parts on top of this change.

0%−50% … 50%
Confidence: High

No source – estimate

Economic affairs and transport: Business and employment services

Change to this part from the current path, %, on top of the change to the whole function.

0%−50% … 50%
Confidence: High

No source – estimate

Economic affairs and transport: Agriculture, forestry and fishing

Change to this part from the current path, %, on top of the change to the whole function.

0%−50% … 50%
Confidence: High

No source – estimate

Economic affairs and transport: Transport

Change to this part from the current path, %, on top of the change to the whole function.

0%−50% … 50%
Confidence: High

No source – estimate

Economic affairs and transport: Research and development

Change to this part from the current path, %, on top of the change to the whole function.

0%−50% … 50%
Confidence: High

No source – estimate

Economic affairs and transport: Other (energy, industry …)

Change to this part from the current path, %, on top of the change to the whole function.

0%−50% … 50%
Confidence: High

No source – estimate

Environmental protection

Change from the current path, %.

0%−50% … 50%
Confidence: High

No source – estimate

Housing and community amenities

Change from the current path, %.

0%−50% … 50%
Confidence: High

No source – estimate

Health

Change to the whole function from the current path, %. The detailed sliders below change individual parts on top of this change.

0%−50% … 50%
Confidence: High

No source – estimate

Health: Hospital services

Change to this part from the current path, %, on top of the change to the whole function.

0%−50% … 50%
Confidence: High

No source – estimate

Health: Outpatient care

Change to this part from the current path, %, on top of the change to the whole function.

0%−50% … 50%
Confidence: High

No source – estimate

Health: Medicines and equipment

Change to this part from the current path, %, on top of the change to the whole function.

0%−50% … 50%
Confidence: High

No source – estimate

Health: Other health (public health, R&D …)

Change to this part from the current path, %, on top of the change to the whole function.

0%−50% … 50%
Confidence: High

No source – estimate

Recreation, culture and religion

Change from the current path, %.

0%−50% … 50%
Confidence: High

No source – estimate

Education

Change to the whole function from the current path, %. The detailed sliders below change individual parts on top of this change.

0%−50% … 50%
Confidence: High

On the current path education spending falls from about 5.9% to 5.1% of GDP in 2030–2040, because there are fewer children. In the EU Ageing Report the fall is about half of this. In reality it may be slower, because schools' fixed costs don't shrink with pupil numbers.

Education: Pre-primary and grades 1–6

Change to this part from the current path, %, on top of the change to the whole function.

0%−50% … 50%
Confidence: High

No source – estimate

Education: Grades 7–9 and upper secondary

Change to this part from the current path, %, on top of the change to the whole function.

0%−50% … 50%
Confidence: High

No source – estimate

Education: Universities and polytechnics

Change to this part from the current path, %, on top of the change to the whole function.

0%−50% … 50%
Confidence: High

No source – estimate

Education: Student aid

Change to this part from the current path, %, on top of the change to the whole function.

0%−50% … 50%
Confidence: High

No source – estimate

Education: Other education (adult education, support services …)

Change to this part from the current path, %, on top of the change to the whole function.

0%−50% … 50%
Confidence: High

No source – estimate

Social protection (incl. pensions)

Change to the whole function from the current path, %. The detailed sliders below change individual parts on top of this change.

0%−50% … 50%
Confidence: High

No source – estimate

Social protection (incl. pensions): Old age (pensions)

Change to this part from the current path, %, on top of the change to the whole function.

0%−50% … 50%
Confidence: High

No source – estimate

Social protection (incl. pensions): Old age (elderly care services)

Change to this part from the current path, %, on top of the change to the whole function.

0%−50% … 50%
Confidence: High

No source – estimate

Social protection (incl. pensions): Sickness and disability

Change to this part from the current path, %, on top of the change to the whole function.

0%−50% … 50%
Confidence: High

No source – estimate

Social protection (incl. pensions): Families and children (incl. early childhood education)

Change to this part from the current path, %, on top of the change to the whole function.

0%−50% … 50%
Confidence: High

No source – estimate

Social protection (incl. pensions): Unemployment

Change to this part from the current path, %, on top of the change to the whole function.

0%−50% … 50%
Confidence: High

No source – estimate

Social protection (incl. pensions): Housing benefits

Change to this part from the current path, %, on top of the change to the whole function.

0%−50% … 50%
Confidence: High

No source – estimate

Social protection (incl. pensions): Social assistance and other social exclusion

Change to this part from the current path, %, on top of the change to the whole function.

0%−50% … 50%
Confidence: High

No source – estimate

Social protection (incl. pensions): Other social protection (survivors' pensions …)

Change to this part from the current path, %, on top of the change to the whole function.

0%−50% … 50%
Confidence: High

No source – estimate

General VAT rate (now 25.5%)

Change in percentage points. Static effect: VAT revenue × general-rate share × change / 25.5.

0.0 pp−5.0 pp … 5.0 pp
Confidence: Medium

No source – estimate

Corporate tax rate (18% from 2027)

Change in percentage points. Static effect: corporate tax revenue at the 20% rate / 20 × change.

0.0 pp−8.0 pp … 8.0 pp
Confidence: Medium

No source – estimate

Income taxes of high earners (top decile)

Income taxes of the top income decile change proportionally, e.g. +5% means 50% → 52.5%. Static effect: income tax × high earners' share × change.

0%−30% … 30%
Confidence: Medium

No source – estimate

Income taxes of low and middle incomes

Income taxes of everyone outside the top decile change proportionally. Static effect: income tax × (1 − high earners' share) × change.

0%−30% … 30%
Confidence: Medium

No source – estimate

Capital income taxes

Tax on capital income (dividends, capital gains, rents, interest) changes proportionally, e.g. +10% means 30% → 33%. Static effect: income tax × capital income tax's share × change.

0%−30% … 30%
Confidence: Medium

No source – estimate

All earned and capital income taxes

All income tax rates change proportionally, e.g. +5% means 30% → 31.5%.

0%−30% … 30%
Confidence: Medium

No source – estimate

Employers' social contributions

Change in contributions from the current path, %.

0%−30% … 30%
Confidence: Medium

No source – estimate

Employees' social contributions

Change in contributions from the current path, %.

0%−30% … 30%
Confidence: Medium

No source – estimate

Energy excise

Change in the tax from the current path, %.

0%−50% … 50%
Confidence: Medium

No source – estimate

Alcohol excise

Change in the tax from the current path, %.

0%−50% … 50%
Confidence: Medium

No source – estimate

Tobacco excise

Change in the tax from the current path, %.

0%−50% … 50%
Confidence: Medium

No source – estimate

Real-estate tax

Change in the tax from the current path, %. The slider goes up to +100% because the tax is small and substantial increases have been proposed in tax-structure reform discussions.

0%−50% … 100%
Confidence: Medium

No source – estimate

Government's spending cuts

Decisions cutting central government appropriations (5.2 bn € in 2030). 100% = as decided, 0% = cuts reversed.

100%0% … 200%
Confidence: High

Government's spending increases (incl. defence)

Decisions increasing central government appropriations, and the investment programme (4.4 bn € in 2030).

100%0% … 200%
Confidence: High

Government's tax increases

Decisions raising central government taxes, e.g. the VAT increase (3.6 bn € in 2030). Excludes one-off transfers such as the extra 2027 transfer from the State Pension Fund, which is neutral for general government.

100%0% … 200%
Confidence: High

Government's tax cuts

Decisions cutting central government taxes, e.g. earned-income and corporate tax cuts (4.3 bn € in 2030).

100%0% … 200%
Confidence: High

Government's social security cuts

Effect of decisions on benefit spending of social security funds (2.3 bn € in 2030).

100%0% … 200%
Confidence: High

Cause-and-effect chains

A cause-and-effect chain is always calculated the same way: effect = share × the lever's static effect × shape. T(t) is the lever's effect on the fiscal balance (a saving or extra revenue is positive), s is the share and n the years since the effect began.

Effect shapes

Proportional

Each year the target changes by a fixed share of the lever's effect.

E(t) = s × T(t)

Delayed

The effect starts after a delay and builds up to its full size over several years.

E(t) = s × T(t) × min(1, max(0, (n − delay + 1) / build-up))

Growing

The effect grows every year the change stays in place, up to a cap.

E(t) = T(t) × sign(s) × min(|s| × (1 + growth)ⁿ, maximum share)

Threshold

A small change causes no effect; the part beyond a threshold does.

E(t) = s × T(t) × max(0, 1 − threshold / |lever change|)

Fading

The effect is largest in the first year of the change and gets smaller every year after that.

E(t) = s × T(t) × (1 − fading)ⁿ

The model's effects

Deferred care raises health costs

Confidence: Low
On by default
Shape: Growing

When health care is cut, queues grow and conditions worsen. Part of the saving comes back later as costlier treatment, and that part grows as long as the cut continues.

Of every 100 euros that a change to “Health” saves or brings in, 15 € comes back as extra spending on the line “Health: hospital services” in the first year, growing 10% a year, up to 60%. (The other way round if the change adds spending or cuts taxes.)

E(t) = T(t) × sign(s) × min(|s| × (1 + growth)ⁿ, maximum share)

Longer queues raise sickness absence

Confidence: Low
On by default
Shape: Delayed

When people wait longer for treatment, sick leave gets longer and Kela's sickness allowance spending grows.

Of every 100 euros that a change to “Health” saves or brings in, 5 € comes back as extra spending on the line “Social protection: sickness and disability”, starting after 1 yr and building up to its full size over 2 yrs. (The other way round if the change adds spending or cuts taxes.)

E(t) = s × T(t) × min(1, max(0, (n − delay + 1) / build-up))

Part of benefit cuts shifts to last-resort assistance

Confidence: Medium
On by default
Shape: Proportional

When social security is cut, some people apply for last-resort social assistance, so the real saving is smaller than the cut.

Of every 100 euros that a change to “Social protection (incl. pensions)” saves or brings in, 15 € comes back as extra spending on the line “Social protection: social assistance and other social exclusion” every year. (The other way round if the change adds spending or cuts taxes.)

E(t) = s × T(t)

Sources: Social Affairs and Health Committee report StVM 14/2023 (bill HE 74/2023): housing allowance cuts lower housing allowance spending by €385m and raise social assistance spending by €77.2m a year (in Finnish)


Maintenance backlog grows

Confidence: Low
On by default
Shape: Growing

When maintenance of roads and railways is cut, repairs get more expensive later and part of the saving returns as a growing cost.

Of every 100 euros that a change to “Transport” saves or brings in, 10 € comes back as extra spending on the line “Economic affairs: transport” in the first year, growing 8% a year, up to 50%. (The other way round if the change adds spending or cuts taxes.)

E(t) = T(t) × sign(s) × min(|s| × (1 + growth)ⁿ, maximum share)

Education cuts weaken future tax revenue

⚠ Disputed
Confidence: Speculative
Off by default
Shape: Delayed

When education is cut, skills and productivity weaken after several years, which lowers future wages and tax revenue.

Of every 100 euros that a change to “Education” saves or brings in, 30 € is lost as revenue on the line “Personal income tax”, starting after 8 yrs and building up to its full size over 10 yrs. (The other way round if the change adds spending or cuts taxes.)

E(t) = s × T(t) × min(1, max(0, (n − delay + 1) / build-up))

Income tax changes work and reported income

⚠ Disputed
Confidence: Medium
Off by default
Shape: Proportional

When income tax is raised, some people work less or shift income to lower-taxed forms, so revenue grows less than the static estimate. A cut works the other way.

Of every 100 euros that a change to “All earned and capital income taxes” saves or brings in, 25 € is lost as revenue on the line “Personal income tax” every year. (The other way round if the change adds spending or cuts taxes.)

E(t) = s × T(t)

Sources: Saez, Slemrod & Giertz (2012): the elasticity of taxable income – a review


Corporate tax moves investment, profits and activity abroad

⚠ Disputed
Confidence: Low
Off by default
Shape: Delayed

When corporate tax is raised, firms invest less in Finland, shift profits to lower-tax countries and some move activity abroad over a few years, so part of the extra revenue disappears. With a cut, part of the loss comes back.

Of every 100 euros that a change to “Corporate tax rate (18% from 2027)” saves or brings in, 30 € is lost as revenue on the line “Corporate income tax”, starting after 0 yrs and building up to its full size over 3 yrs. (The other way round if the change adds spending or cuts taxes.)

E(t) = s × T(t) × min(1, max(0, (n − delay + 1) / build-up))

Higher taxes on high earners drive them to move or shift income

⚠ Disputed
Confidence: Low
Off by default
Shape: Delayed

When taxes on high earners are raised, some of them move abroad, don't return, or shift income to lower-taxed forms, so Finland collects less than the static estimate. A cut works the other way.

Of every 100 euros that a change to “Income taxes of high earners (top decile)” saves or brings in, 35 € is lost as revenue on the line “Personal income tax”, starting after 1 yr and building up to its full size over 4 yrs. (The other way round if the change adds spending or cuts taxes.)

E(t) = s × T(t) × min(1, max(0, (n − delay + 1) / build-up))

Sources: Kleven, Landais, Saez & Schultz (2014): migration and wage effects of taxing top earners, evidence from Denmark; Saez, Slemrod & Giertz (2012): the elasticity of taxable income – a review


Higher taxes on low and middle incomes cut consumption and VAT revenue

Confidence: Medium
On by default
Shape: Proportional

People on low and middle incomes spend most of their income. When their taxes rise, consumption falls by nearly the same amount, and the VAT on that spending is lost. A cut raises consumption.

Of every 100 euros that a change to “Income taxes of low and middle incomes” saves or brings in, 15 € is lost as revenue on the line “Value added tax” every year. (The other way round if the change adds spending or cuts taxes.)

E(t) = s × T(t)

Lower consumption weakens company profits

Confidence: Low
On by default
Shape: Proportional

When people on low and middle incomes spend less, company sales and profits fall, which lowers corporate tax revenue. Wider effects on jobs and growth will be added later with multiplier effects.

Of every 100 euros that a change to “Income taxes of low and middle incomes” saves or brings in, 3 € is lost as revenue on the line “Corporate income tax” every year. (The other way round if the change adds spending or cuts taxes.)

E(t) = s × T(t)

Higher income taxes cut consumption and VAT revenue

Confidence: Medium
On by default
Shape: Proportional

When everyone's income taxes rise, consumption falls, especially among low and middle incomes, and VAT revenue is lost. The effect is smaller than for a rise on low and middle incomes alone, because high earners save more of their income.

Of every 100 euros that a change to “All earned and capital income taxes” saves or brings in, 8 € is lost as revenue on the line “Value added tax” every year. (The other way round if the change adds spending or cuts taxes.)

E(t) = s × T(t)

A VAT increase reduces consumption

Confidence: Low
On by default
Shape: Proportional

When VAT is raised, prices rise and consumption falls somewhat, so revenue grows less than the static estimate.

Of every 100 euros that a change to “General VAT rate (now 25.5%)” saves or brings in, 10 € is lost as revenue on the line “Value added tax” every year. (The other way round if the change adds spending or cuts taxes.)

E(t) = s × T(t)

Alcohol tax increases shift purchases to Estonia

Confidence: Medium
On by default
Shape: Threshold

A small increase raises revenue almost fully, but with a larger increase more people buy their alcohol in Estonia, and part of the extra revenue is lost.

Of every 100 euros that a change to “Alcohol excise” saves or brings in, 50 € is lost as revenue on the line “Alcohol excise” for the part of the lever change beyond 10%. (The other way round if the change adds spending or cuts taxes.)

E(t) = s × T(t) × max(0, 1 − threshold / |lever change|)

Sources: THL blog: Cheap alcohol would collapse alcohol tax revenue (revenue, traveller imports and the 2004 cut; in Finnish)


Alcohol tax increases lower harm and health costs

Confidence: Speculative
Off by default
Shape: Delayed

Pricier alcohol reduces consumption and alcohol-related harm, which lowers health care spending after a few years.

Of every 100 euros that a change to “Alcohol excise” saves or brings in, 20 € is saved on top on the line “Health: hospital services”, starting after 1 yr and building up to its full size over 3 yrs. (The other way round if the change adds spending or cuts taxes.)

E(t) = s × T(t) × min(1, max(0, (n − delay + 1) / build-up))

Sources: THL blog: Cheap alcohol would collapse alcohol tax revenue (revenue, traveller imports and the 2004 cut; in Finnish)


Benefit cuts raise employment

⚠ Disputed
Confidence: Low
Off by default
Shape: Delayed

The government estimates that social security cuts encourage work, which raises tax revenue after a few years. If the cuts are reversed, this effect is lost.

Of every 100 euros that a change to “Government's social security cuts” saves or brings in, 30 € comes in as extra revenue on the line “Personal income tax”, starting after 1 yr and building up to its full size over 3 yrs. (The other way round if the change adds spending or cuts taxes.)

E(t) = s × T(t) × min(1, max(0, (n − delay + 1) / build-up))

Outpatient care cuts raise disability pensions

Confidence: Low
On by default
Shape: Delayed

When outpatient care – for example primary care and mental-health services – is cut, some illnesses drag on and lead to a disability pension a few years later. The increase shows in sickness and disability spending.

Of every 100 euros that a change to “Outpatient care” saves or brings in, 10 € comes back as extra spending on the line “Social protection: sickness and disability”, starting after 3 yrs and building up to its full size over 5 yrs. (The other way round if the change adds spending or cuts taxes.)

E(t) = s × T(t) × min(1, max(0, (n − delay + 1) / build-up))

Poverty raises the need for child welfare and family support

Confidence: Speculative
Off by default
Shape: Delayed

When social security is cut, poverty deepens among the lowest-income families. Research links child poverty to the need for child welfare and later problems, so part of the saving comes back over the years as spending on family and children's services.

Of every 100 euros that a change to “Social protection (incl. pensions)” saves or brings in, 5 € comes back as extra spending on the line “Social protection: families and children (incl. early childhood education)”, starting after 2 yrs and building up to its full size over 8 yrs. (The other way round if the change adds spending or cuts taxes.)

E(t) = s × T(t) × min(1, max(0, (n − delay + 1) / build-up))

Basic research cuts weaken future productivity

⚠ Disputed
Confidence: Speculative
Off by default
Shape: Delayed

When basic research is cut, less new knowledge and innovation emerges, which slows productivity growth and lowers wages and tax revenue after a long lag. Finland has a statutory target of raising R&D spending to 4% of GDP by 2030.

Of every 100 euros that a change to “Basic research” saves or brings in, 30 € is lost as revenue on the line “Personal income tax”, starting after 7 yrs and building up to its full size over 10 yrs. (The other way round if the change adds spending or cuts taxes.)

E(t) = s × T(t) × min(1, max(0, (n − delay + 1) / build-up))

Cuts to business R&D support weaken future tax revenue

⚠ Disputed
Confidence: Speculative
Off by default
Shape: Delayed

When research and development support for business (for example Business Finland funding) is cut, firms do less product development, which slows productivity and lowers later profits, wages and tax revenue.

Of every 100 euros that a change to “Research and development” saves or brings in, 30 € is lost as revenue on the line “Personal income tax”, starting after 5 yrs and building up to its full size over 8 yrs. (The other way round if the change adds spending or cuts taxes.)

E(t) = s × T(t) × min(1, max(0, (n − delay + 1) / build-up))

Elderly care cuts block hospital beds

Confidence: Low
On by default
Shape: Delayed

When elderly care (service housing, home care) is cut, some older people have to wait in hospital for a care place. A hospital day costs about twice a day in service housing, so part of the saving comes back as health spending within the same year.

Of every 100 euros that a change to “Old age (elderly care services)” saves or brings in, 30 € comes back as extra spending on the line “Health: hospital services”, starting after 0 yrs and building up to its full size over 1 yr. (The other way round if the change adds spending or cuts taxes.)

E(t) = s × T(t) × min(1, max(0, (n − delay + 1) / build-up))

Sources: Mäklin & Kokko (eds.): Unit costs of health and social care in Finland in 2017 (THL working paper 21/2020): a health-centre ward day €308 (short-term €359), a day of 24-hour service housing €124–174 (in Finnish)


Student aid cuts make students work more

Confidence: Low
On by default
Shape: Proportional

When student aid is cut, some students make up for it by working more. Their wages are taxed, so part of the saving comes back as tax revenue straight away. (Some take a student loan instead, and some spend less.)

Of every 100 euros that a change to “Student aid” saves or brings in, 10 € comes in as extra revenue on the line “Personal income tax” every year. (The other way round if the change adds spending or cuts taxes.)

E(t) = s × T(t)

Sources: Reforming student financial aid, rapporteur's report (Ministry of Education and Culture; in Finnish)


Student aid cuts slow down graduation

⚠ Disputed
Confidence: Speculative
Off by default
Shape: Delayed

When students work more alongside their studies, some graduate later and some drop out. Skilled workers enter jobs matching their education later, which lowers tax revenue after several years.

Of every 100 euros that a change to “Student aid” saves or brings in, 20 € is lost as revenue on the line “Personal income tax”, starting after 3 yrs and building up to its full size over 8 yrs. (The other way round if the change adds spending or cuts taxes.)

E(t) = s × T(t) × min(1, max(0, (n − delay + 1) / build-up))

Sources: Reforming student financial aid, rapporteur's report (Ministry of Education and Culture; in Finnish)


Higher capital income tax delays selling shares and property

⚠ Disputed
Confidence: Low
Off by default
Shape: Proportional

Capital gains tax is only paid when an asset is sold. When the tax is raised, owners put off sales, so fewer gains are realised and the extra revenue falls short of the static estimate. A cut works the other way.

Of every 100 euros that a change to “Capital income taxes” saves or brings in, 30 € is lost as revenue on the line “Personal income tax” every year. (The other way round if the change adds spending or cuts taxes.)

E(t) = s × T(t)

Sources: Dowd, McClelland & Muthitacharoen (2015): persistent tax elasticity of capital gains realisations −0.72 (United States)


Health staff cuts raise unemployment spending

Confidence: Low
On by default
Shape: Fading

When health care spending is cut, part of the saving comes from staff: some employees are dismissed or their fixed-term contracts are not renewed. They receive unemployment benefit until they find new work – most within a year – so the effect is largest in the first year and then fades. A change to medicines and equipment doesn't trigger it, as no own staff are involved. Extra spending works the other way, but probably less strongly, as new staff are often hired from other jobs.

Of every 100 euros that a change to “Health” saves or brings in, 2 € comes back as extra spending on the line “Social protection: unemployment” in the first year, getting 55% smaller every year after that. (The other way round if the change adds spending or cuts taxes.)

E(t) = s × T(t) × (1 − fading)ⁿ

Sources: Statistics Finland: general government expenditure by function 2024 (table 12a6): compensation of employees is 44% of health spending (excl. medicines), 51% of education (excl. student aid) and 43% of elderly care; OECD (2016): Back to Work: Finland – nearly five in six displaced workers are re-employed within a year; those over 55 and the less educated more slowly (Ministry summary); Yle 23 Sep 2026: Pohjois-Savo wellbeing services county cuts 280 person-years – 50 dismissed, 14 fixed-term contracts not renewed, 88 retire (in Finnish); Yle 1 Sep 2026: at Pohjois-Pohjanmaa wellbeing services county 511 employment relationships end, 302 of them dismissals; saving €25.4m (in Finnish); Verho (2020), Economica: workers displaced in the 1990s recession were almost 50% more often unemployed five years later than a comparison group (VATT summary); Algan, Cahuc & Zylberberg (2002): public jobs crowd out private ones – in the long run a change in public employment hardly changes total employment; Faggio & Overman (2014): public sector jobs have no identifiable effect on total local private employment (England)


Health staff cuts lower tax and contribution revenue

Confidence: Low
On by default
Shape: Fading

When health care spending is cut and some employees become unemployed, they pay no income tax or social contributions on a wage, and no employer pension or other social contributions are paid for them, until they find new work. The effect is largest in the first year and then fades. A change to medicines and equipment doesn't trigger it, as no own staff are involved. Extra spending works the other way, but probably less strongly.

Of every 100 euros that a change to “Health” saves or brings in, 2 € is lost as revenue on the line “Personal income tax” in the first year, getting 55% smaller every year after that. (The other way round if the change adds spending or cuts taxes.)

E(t) = s × T(t) × (1 − fading)ⁿ

Sources: Statistics Finland: general government expenditure by function 2024 (table 12a6): compensation of employees is 44% of health spending (excl. medicines), 51% of education (excl. student aid) and 43% of elderly care; OECD (2016): Back to Work: Finland – nearly five in six displaced workers are re-employed within a year; those over 55 and the less educated more slowly (Ministry summary); Yle 23 Sep 2026: Pohjois-Savo wellbeing services county cuts 280 person-years – 50 dismissed, 14 fixed-term contracts not renewed, 88 retire (in Finnish); Yle 1 Sep 2026: at Pohjois-Pohjanmaa wellbeing services county 511 employment relationships end, 302 of them dismissals; saving €25.4m (in Finnish); Verho (2020), Economica: workers displaced in the 1990s recession were almost 50% more often unemployed five years later than a comparison group (VATT summary); Algan, Cahuc & Zylberberg (2002): public jobs crowd out private ones – in the long run a change in public employment hardly changes total employment; Faggio & Overman (2014): public sector jobs have no identifiable effect on total local private employment (England)


Education staff cuts raise unemployment spending

Confidence: Low
On by default
Shape: Fading

When education spending is cut, part of the saving comes from staff: some employees are dismissed or their fixed-term contracts are not renewed. They receive unemployment benefit until they find new work – most within a year – so the effect is largest in the first year and then fades. A change to student aid doesn't trigger it, as it is a cash benefit. Extra spending works the other way, but probably less strongly, as new staff are often hired from other jobs.

Of every 100 euros that a change to “Education” saves or brings in, 3 € comes back as extra spending on the line “Social protection: unemployment” in the first year, getting 55% smaller every year after that. (The other way round if the change adds spending or cuts taxes.)

E(t) = s × T(t) × (1 − fading)ⁿ

Sources: Statistics Finland: general government expenditure by function 2024 (table 12a6): compensation of employees is 44% of health spending (excl. medicines), 51% of education (excl. student aid) and 43% of elderly care; OECD (2016): Back to Work: Finland – nearly five in six displaced workers are re-employed within a year; those over 55 and the less educated more slowly (Ministry summary); Yle 23 Sep 2026: Pohjois-Savo wellbeing services county cuts 280 person-years – 50 dismissed, 14 fixed-term contracts not renewed, 88 retire (in Finnish); Yle 1 Sep 2026: at Pohjois-Pohjanmaa wellbeing services county 511 employment relationships end, 302 of them dismissals; saving €25.4m (in Finnish); Verho (2020), Economica: workers displaced in the 1990s recession were almost 50% more often unemployed five years later than a comparison group (VATT summary); Algan, Cahuc & Zylberberg (2002): public jobs crowd out private ones – in the long run a change in public employment hardly changes total employment; Faggio & Overman (2014): public sector jobs have no identifiable effect on total local private employment (England)


Education staff cuts lower tax and contribution revenue

Confidence: Low
On by default
Shape: Fading

When education spending is cut and some employees become unemployed, they pay no income tax or social contributions on a wage, and no employer pension or other social contributions are paid for them, until they find new work. The effect is largest in the first year and then fades. A change to student aid doesn't trigger it, as it is a cash benefit. Extra spending works the other way, but probably less strongly.

Of every 100 euros that a change to “Education” saves or brings in, 2 € is lost as revenue on the line “Personal income tax” in the first year, getting 55% smaller every year after that. (The other way round if the change adds spending or cuts taxes.)

E(t) = s × T(t) × (1 − fading)ⁿ

Sources: Statistics Finland: general government expenditure by function 2024 (table 12a6): compensation of employees is 44% of health spending (excl. medicines), 51% of education (excl. student aid) and 43% of elderly care; OECD (2016): Back to Work: Finland – nearly five in six displaced workers are re-employed within a year; those over 55 and the less educated more slowly (Ministry summary); Yle 23 Sep 2026: Pohjois-Savo wellbeing services county cuts 280 person-years – 50 dismissed, 14 fixed-term contracts not renewed, 88 retire (in Finnish); Yle 1 Sep 2026: at Pohjois-Pohjanmaa wellbeing services county 511 employment relationships end, 302 of them dismissals; saving €25.4m (in Finnish); Verho (2020), Economica: workers displaced in the 1990s recession were almost 50% more often unemployed five years later than a comparison group (VATT summary); Algan, Cahuc & Zylberberg (2002): public jobs crowd out private ones – in the long run a change in public employment hardly changes total employment; Faggio & Overman (2014): public sector jobs have no identifiable effect on total local private employment (England)


Elderly care staff cuts raise unemployment spending

Confidence: Low
On by default
Shape: Fading

When elderly care spending is cut, part of the saving comes from staff: some employees are dismissed or their fixed-term contracts are not renewed. They receive unemployment benefit until they find new work – most within a year – so the effect is largest in the first year and then fades. Extra spending works the other way, but probably less strongly, as new staff are often hired from other jobs.

Of every 100 euros that a change to “Old age (elderly care services)” saves or brings in, 2 € comes back as extra spending on the line “Social protection: unemployment” in the first year, getting 55% smaller every year after that. (The other way round if the change adds spending or cuts taxes.)

E(t) = s × T(t) × (1 − fading)ⁿ

Sources: Statistics Finland: general government expenditure by function 2024 (table 12a6): compensation of employees is 44% of health spending (excl. medicines), 51% of education (excl. student aid) and 43% of elderly care; OECD (2016): Back to Work: Finland – nearly five in six displaced workers are re-employed within a year; those over 55 and the less educated more slowly (Ministry summary); Yle 23 Sep 2026: Pohjois-Savo wellbeing services county cuts 280 person-years – 50 dismissed, 14 fixed-term contracts not renewed, 88 retire (in Finnish); Yle 1 Sep 2026: at Pohjois-Pohjanmaa wellbeing services county 511 employment relationships end, 302 of them dismissals; saving €25.4m (in Finnish); Verho (2020), Economica: workers displaced in the 1990s recession were almost 50% more often unemployed five years later than a comparison group (VATT summary); Algan, Cahuc & Zylberberg (2002): public jobs crowd out private ones – in the long run a change in public employment hardly changes total employment; Faggio & Overman (2014): public sector jobs have no identifiable effect on total local private employment (England)


Elderly care staff cuts lower tax and contribution revenue

Confidence: Low
On by default
Shape: Fading

When elderly care spending is cut and some employees become unemployed, they pay no income tax or social contributions on a wage, and no employer pension or other social contributions are paid for them, until they find new work. The effect is largest in the first year and then fades. Extra spending works the other way, but probably less strongly.

Of every 100 euros that a change to “Old age (elderly care services)” saves or brings in, 2 € is lost as revenue on the line “Personal income tax” in the first year, getting 55% smaller every year after that. (The other way round if the change adds spending or cuts taxes.)

E(t) = s × T(t) × (1 − fading)ⁿ

Sources: Statistics Finland: general government expenditure by function 2024 (table 12a6): compensation of employees is 44% of health spending (excl. medicines), 51% of education (excl. student aid) and 43% of elderly care; OECD (2016): Back to Work: Finland – nearly five in six displaced workers are re-employed within a year; those over 55 and the less educated more slowly (Ministry summary); Yle 23 Sep 2026: Pohjois-Savo wellbeing services county cuts 280 person-years – 50 dismissed, 14 fixed-term contracts not renewed, 88 retire (in Finnish); Yle 1 Sep 2026: at Pohjois-Pohjanmaa wellbeing services county 511 employment relationships end, 302 of them dismissals; saving €25.4m (in Finnish); Verho (2020), Economica: workers displaced in the 1990s recession were almost 50% more often unemployed five years later than a comparison group (VATT summary); Algan, Cahuc & Zylberberg (2002): public jobs crowd out private ones – in the long run a change in public employment hardly changes total employment; Faggio & Overman (2014): public sector jobs have no identifiable effect on total local private employment (England)

Limitations

The model is a simplification. The main limitations:

  • Growth and supply-side effects are optional and off by default, because economists disagree about their size. Indirect cause-and-effect effects (for example the cost of longer care queues) are included as assumptions you can switch off, adjust or add yourself.
  • Some assumptions are estimates that don't have a source yet. They are marked with low confidence and a note.
  • The tax figures are direct (static) effects. Behavioural effects – for example changes in consumption or tax planning – are included as separate cause-and-effect effects, of which the disputed ones are off by default.
  • On the current path the deficit narrows slightly after 2030, mainly because there are fewer children and education spending falls. In this calculation education spending falls about twice as fast as in the EU Ageing Report, and in reality the fall may be slower (schools' fixed costs), so this improvement is probably somewhat too optimistic.
  • For comparison: in the Economic Policy Council's assessment public debt rises to about 118% of GDP by 2040 if no new decisions are made after 2028. This calculation's current path is of the same order.
Read more about the model and its limitations