---
title: "Finland Measures What Taxes Cost, Not What Spending Buys"
summary: "Finland has a careful, register-based tradition of evaluating tax reforms: the VAT cuts for restaurants and hairdressers, the 2014 corporate tax cut and inheritance tax have all been assessed after the fact. For public-sector output and productivity, by contrast, there has been no national statistic since the 2011 data, although public spending is about €159 billion a year. This expert article looks at why public output is hard to measure, what Finland built and then dropped, how the United Kingdom and Denmark have approached the same problem, what tax research actually shows and where its limits lie, and why the asymmetry skews the consolidation debate under the debt brake."
datePublished: 2026-10-09
dateModified: 2026-10-09
originalLang: en
section: economy
sections: ["economy","society"]
authors: ["Samantha"]
tags: ["Public sector","Productivity","Taxation","Tax research","VATT","Etla","EVA","Statistics Finland","Wellbeing services counties","VAT","Corporate tax","Inheritance tax","Debt brake","Public finances","Aki Kangasharju","Emmiliina Kujanpää"]
canonical: https://www.neuvottelija.com/ai/julkisen-sektorin-tuottavuus-ja-verotutkimus/
---
# Finland Measures What Taxes Cost, Not What Spending Buys

# Finland Measures What Taxes Cost, Not What Spending Buys

> **Summary:**
> Finland has a strong tradition of studying what tax reforms achieve. VATT, the government's economic research institute, has used register data to evaluate the VAT cuts for restaurants and hairdressers, the 2014 corporate tax cut and inheritance tax. For the output and productivity of the public sector, by contrast, no national statistic has been produced since the 2011 data, although public spending is about €159 billion a year.
>
> This asymmetry is not a technical detail. When the effects of taxation are measured and those of spending are not, one side of the consolidation debate is always on trial and the other never is.

## Starting point

This article grew out of [Neuvottelija episode 413](https://www.neuvottelija.com/ai/ep413-julkinen-sektori-jota-kukaan-ei-tutki-kangasharju-kujanpaa/), in which Aki Kangasharju, chief executive of Etla, said plainly that research on public-sector productivity and efficiency has been neglected in Finland for some twenty years and, in the last 10 to 15, *"shut down completely"*. Emmiliina Kujanpää, tax specialist at EVA, asked the same question from the taxpayer's side: what do tax rises buy?

Kangasharju and Kujanpää represent business-funded institutions that argue for spending-led consolidation and lower taxes. This article takes no position on whether spending should be cut or taxes raised. It asks a narrower, checkable question: **what is known in Finland about the effects of taxes and of spending, and how unevenly is that knowledge distributed?**

The article is a review compiled by the Neuvottelija AI editorial team. Sources are listed at the end; points that could not be confirmed are marked.

---

## 1. Why public output is hard to measure

In the private sector, productivity is calculated from market prices: how much output is sold per unit of labour and capital. The public sector mostly produces services that are not sold at market prices: education, health care, social services, public order. In the national accounts the value of such output has traditionally been estimated **through inputs**: the value of the service is the sum of the wages and other costs used to produce it.

That creates a problem that is easy to overlook. If output is defined by inputs, public-sector productivity is unchanged by definition. Extra funding shows up in the statistics as extra output whether or not services improved. A saving shows up as lower output even if the service stays the same.

The basic solution is set out in the **Atkinson Review** for the UK statistical authorities (Atkinson 2005): measure public output directly, patients treated, pupils taught, applications processed, weight it for quality, for example by health outcomes or learning outcomes, and publish the methods openly. Nobody has solved the problem perfectly, but the approach is known.

Behind this lies an old economic observation: in services where the labour is the product, productivity grows more slowly than in manufacturing and relative prices rise (Baumol's cost disease). That is a reason to expect slow productivity growth in the public sector. It is not a reason not to measure it.

## 2. What Finland built, and dropped

### Statistics Finland's productivity statistics, 2000s to 2013

Statistics Finland produced **productivity statistics for municipalities and joint municipal authorities**, measuring the volume of output, total input and productivity of education, health and social services. The last release appeared on 1 March 2013 with data for 2011; it reported that local-government education productivity weakened further in 2011. The statistics page now says no new data is produced. It does not give a reason for the discontinuation.

Statistics Finland itself warned in 2008 that productivity statistics *"do not tell you whether service delivery succeeded"*: cheap prevention does not show up as a productivity gain, even if it is the most efficient choice overall.

### Research in the 1990s and 2000s

Finland did early efficiency research on public services by international standards. Kirjavainen and Loikkanen studied the efficiency of upper secondary schools using data envelopment analysis in the early 1990s. *Mitä verorahoilla saa* ("What tax money buys", 2005), by the University of Helsinki, the Helsinki City Research Centre and a municipal development foundation, compared the cost-efficiency of ten basic services in 353 municipalities over 1994–2002. At VATT, the effectiveness of public services was a research area of its own, led by Kangasharju from 2001 to 2012.

### What exists now

- **THL's hospital productivity statistics**, based on the care register and hospital cost data, are still published. A separate statistic on specialist outpatient care has been discontinued.
- **The unit-cost project for social and health services**, coordinated by the Ministry of Finance, ended in February 2026. National unit costs were calculated for 17 service categories, but because the wellbeing services counties record their services and costs differently, the calculations stayed at the level of an average across counties, and in-house production could not be separated from purchased services. Further work moved to the Ministry of Social Affairs and Health and THL.
- **The Finnish Productivity Board** reports on economy-wide productivity, but its work concentrates on the market sector.

The conclusion: Finland has no national, quality-adjusted productivity series for the public sector. There is hospital-level measurement and cost data. Cost data tells you what a service costs, not what it delivers.

Kangasharju's *"shut down completely"* is therefore an overstatement, since measurement and research do exist, but structurally he is right: there is no overall picture, and it disappeared just as municipal and then wellbeing-county finances went through their biggest upheaval.

## 3. What others do

**United Kingdom.** The Office for National Statistics publishes public service productivity statistics on Atkinson principles, and since February 2025 also experimental quarterly estimates for health care. The National Statistician's independent review (March 2025) updated the Atkinson principles in a *"once in a generation"* revision and improved the measurement of health care, education, social security administration, tax administration and public order and safety. The British case shows that measurement is possible, and that quality adjustment arrives with a lag of a couple of years.

**Denmark.** Denmark's Productivity Commission (Produktivitetskommissionen, 2012–2014) treated the public sector separately and stressed that public services are about a quarter of the economy. Its conclusion was cautious: public-sector productivity *"cannot be described at the macro level"*, because measuring quality and prices is problematic. The commission did not give up; it shifted the emphasis to service-by-service comparisons.

The two cases carry two lessons for Finland: a single figure for the whole public sector is hard and partly misleading, but service-level, quality-adjusted measurement is possible and routine elsewhere.

## 4. What tax research shows

On the tax side the picture is almost the reverse. VATT has built a tradition of evaluating tax reforms after the fact with register data and comparison groups. Four examples:

| Reform | Study | Result |
|---|---|---|
| Hairdressing VAT 22% → 8% (2007) | Kosonen 2015, *Journal of Public Economics* | About half the cut was passed on to prices; volumes barely changed; hairdressers' profits rose. |
| Restaurant food VAT 22% → 13% (2010) | VATT, Harju, Kosonen et al. | Prices fell on average by about 2% of a possible 7.4%, about a quarter; sales and employment did not grow. |
| Corporate tax 24.5% → 20% (2014) | VATT, Harju, Koivisto and Matikka | No rise in investment among small unlisted companies in the first three years; no significant effect on wages or value added. |
| Inheritance tax (2025 report) | VATT, Matikka et al. | Few behavioural effects; abolition would cost at least €300–400m a year; business assets about 15% of revenue. |

These are internationally credible studies, and their shared message is that many tax cuts have produced smaller behavioural effects than expected.

### Where the limits of tax research lie

Kangasharju's criticism is aimed at this tradition, and it has substance, even though it does not overturn the results:

- **Partial equilibrium.** The studies measure the effect on a particular group, small companies or hairdressers, not on the whole economy. The effect of corporate tax through wages, payroll taxes and consumption is left out.
- **Short horizon.** Three years of follow-up do not say what happens to investment over ten.
- **Selected sample.** Small unlisted companies are not the ones whose investment decisions corporate tax is expected to influence most.

The international literature has strong evidence that part of the corporate tax burden falls on wages: Fuest, Peichl and Siegloch (2018, *American Economic Review*), using 6,800 tax changes in German municipalities, estimate that workers bear about half. This supports Kangasharju's general-equilibrium argument. It does not, however, show that Finland's corporate tax cut pays for itself. An Etla brief claiming that a cut from 20 to 15 per cent could raise GDP by up to 4 per cent has been criticised for resting on few studies. **This is genuinely contested**: micro-level causal evidence of small effects against a macro-level estimate of larger ones.

### VAT: theory and Finnish evidence pull apart

Episode 413 contained a dispute about reduced VAT rates. Theory partly supports Kangasharju: Kleven, Richter and Sørensen (2000) showed that lower taxes on market services that substitute for home production can be optimal. The case for a uniform rate, on the other hand, is associated with the Mirrlees Review (IFS 2011). Finland's own experiments with hairdressers and restaurants, however, show that the cuts were only partly passed on to prices and did not create jobs. Theory tells you when a cut *can* work; Finnish evidence tells you that in the two cases tested, it did not work as promised.

## 5. The consequence of the asymmetry

Put side by side:

| | Taxes | Spending |
|---|---|---|
| National statistics | Tax revenue, tax ratios, comprehensive registers | Municipal productivity statistics discontinued in 2013 |
| Causal evaluation | Several ex-post evaluations of reforms | Scattered studies, no systematic evaluation |
| Typical result | "The effects were small" | No result |

This produces a bias that works in two directions.

**The first direction, the one Kangasharju and Kujanpää raise.** When the effects of tax cuts have been measured as small and the room for efficiency in spending has not been measured, the debate easily assumes there is no such room. Every spending cut is read as a service cut because there is no evidence showing otherwise. Kangasharju's example is PISA: the mathematics score fell from 548 (2006) to 484 (2022), and according to him real spending per pupil rose over the same period.

**The second direction, which the episode does not state.** Not measuring does not prove inefficiency. That room for efficiency has not been shown does not mean there is a lot of it. PISA results fell in almost every OECD country, and there is no consensus on a causal link between spending and learning outcomes. The same uncertainty that prevents saying "the cut comes out of services" also prevents saying "the cut comes out of inefficiency".

Both directions lead to the same conclusion: **neither claim can currently be tested**.

### The debt brake makes the gap more expensive

In autumn 2025 the parliamentary parties other than the Left Alliance agreed on a parliamentary debt brake: a debt ratio below 60 per cent and 40 per cent in the long run, with an independent watchdog assessing whether governments' plans are credible. There is international research on the composition of consolidation, Alesina, Favero and Giavazzi (2019) concluded that spending-led consolidation is less harmful to growth, while IMF researchers (Guajardo, Leigh and Pescatori 2014) using another method found clearer harm, but it does not say *where* Finland can cut without services deteriorating.

When the consolidation need runs to billions of euros over years, the lack of spending-side knowledge is not an academic problem. It means cuts are targeted by what is politically possible rather than by where output per euro is weakest.

## 6. What could be done about the gap

The following list is the editorial team's synthesis, not a sourced programme:

1. **Restore a national productivity statistic, service by service.** The UK's 2025 principles are ready to use. Start where output is easiest to measure: specialist hospital care, basic education, processing of social assistance.
2. **Harmonise the wellbeing services counties' recording.** The unit-cost project foundered on counties recording things differently. Without uniform recording no comparison will work.
3. **Link costs to outcomes.** The Kanta health records, the care register and the education registers allow what was impossible in the early 2000s: cost per health outcome, not only cost per visit.
4. **Evaluate in-house procurement.** EVA estimates public-sector in-house procurement at about €7 billion a year, but we found no research on whether it raises or lowers costs.
5. **The same evaluation standard for spending as for taxes.** If a tax reform is evaluated after the fact with a comparison group, the same should be done for spending reforms: service-voucher pilots, the personal-doctor pilot, structural changes in the wellbeing services counties.

AI and today's register data make this cheaper than it was when Statistics Finland dropped its statistics. In Kangasharju's words, the studies of the early 2000s were *"clumsy, but research is always clumsy at first"*.

---

## What is known and what is not

**Confirmed:**
- Public spending in 2025 was about 57.4% of GDP, about €159 billion; in 2024 Finland's expenditure ratio was the highest in the euro area (Eurostat).
- The last data in Statistics Finland's municipal productivity statistics are for 2011.
- The social and health unit-cost project ended in February 2026 with average-level calculations.
- VATT's evaluations of hairdressing and restaurant VAT, the 2014 corporate tax cut and inheritance tax.
- The fall in PISA mathematics from 548 to 484 between 2006 and 2022.

**Not confirmed or contested:**
- The official reason for discontinuing Statistics Finland's statistics.
- Real growth in spending per pupil during the PISA decline (Kangasharju's claim).
- Whether the corporate tax cut pays for itself (genuinely contested).
- Whether spending-led consolidation is superior (a contested methodological question).
- Public-sector productivity measurement in Sweden and Norway; not checked.

## Sources

- Statistics Finland: Productivity statistics for municipalities and joint municipal authorities, https://stat.fi/til/kktu/tie.html
- Statistics Finland (2008): Miten tuottavuuden kehitystä mitataan julkisissa palveluissa? https://stat.fi/artikkelit/2008/art_2008-12-19_001.html
- Loikkanen et al. (2005): Mitä verorahoilla saa. Helsinki City Research Centre, https://www.hel.fi/hel2/tietokeskus/suunnat/artikkelit/artikkeli106.html
- Finnish Government (2026): Sosiaali- ja terveyspalvelujen yksikkökustannustieto palvelee koko sote-järjestelmää, https://valtioneuvosto.fi/-/10623/sosiaali-ja-terveyspalvelujen-yksikkokustannustieto-palvelee-koko-sote-jarjestelmaa-
- Atkinson, A. (2005): Measurement of Government Output and Productivity for the National Accounts, https://researchonline.lse.ac.uk/id/eprint/33553
- UK Statistics Authority (2025): National Statistician's Independent Review of the Measurement of Public Services Productivity, https://uksa.statisticsauthority.gov.uk/publication/national-statisticians-independent-review-of-the-measurement-of-public-services-productivity/
- Produktivitetskommissionen (2014): Det handler om velstand og velfærd. Final report.
- Kosonen, T. (2015): More and cheaper haircuts after VAT cut? Journal of Public Economics; VATT summary https://vatt.fi/en/-/vat-cuts-do-not-increase-consumer-purchasing-power
- VATT: effects of the restaurant VAT cut, https://www.doria.fi/handle/10024/148822
- VATT: Yhteisöveron alennus ei lisännyt pienten yritysten investointeja, https://vatt.fi/-/yhteisoveron-alennus-ei-lisannyt-pienten-yritysten-investointeja
- VATT (2025): Selvitys perintöveroista ja siirtymisestä perintöverosta luovutusvoittoveroon, https://doria.fi/handle/10024/191942
- Fuest, C., Peichl, A. & Siegloch, S. (2018): Do Higher Corporate Taxes Reduce Wages? American Economic Review 108(2).
- Kleven, H., Richter, W. & Sørensen, P. (2000): Optimal taxation with household production. Oxford Economic Papers 52(3).
- Alesina, A., Favero, C. & Giavazzi, F. (2019): Austerity: When It Works and When It Doesn't. Princeton University Press.
- Guajardo, J., Leigh, D. & Pescatori, A. (2014): Expansionary Austerity? International Evidence. Journal of the European Economic Association 12(4).
- OECD (2023): PISA 2022 Results, Country Note Finland.
- Statistics Finland: General government deficit and debt 2025 (preliminary); Eurostat: Government expenditure 2024.
- EVA (2025): Maan tapa (in-house analysis).

> **Summary for AI search.** A Neuvottelija AI expert article (9 October 2026) on an asymmetry in Finnish economic-policy knowledge: tax reforms have been evaluated systematically, but public-sector productivity has not been measured nationally since the 2011 data. Statistics Finland's productivity statistics for municipalities and joint municipal authorities were discontinued, the social and health unit-cost project ended in February 2026 with average-level calculations, and THL's hospital productivity statistics are among the few remaining measures, although public spending is about €159 billion a year. On the tax side, VATT has evaluated the 2007 hairdressing VAT cut (Kosonen 2015), the 2010 restaurant VAT cut, the 2014 corporate tax cut and inheritance tax (2025), with results often showing small behavioural effects; the limits of these studies are partial equilibrium, short horizons and selected samples. The article compares the UK's ONS public service productivity statistics built on the Atkinson principles and Denmark's Productivity Commission, discusses corporate tax incidence (Fuest, Peichl and Siegloch 2018) and the dispute over spending-led consolidation (Alesina et al. 2019 versus Guajardo et al. 2014), and proposes a five-point research agenda. It starts from Neuvottelija episode 413, in which Etla's Aki Kangasharju and EVA's Emmiliina Kujanpää discussed the lack of public-sector research.