Subnational government spending

Subnational governments account for 40% of public expenditure, corresponding to 16% of GDP, a proportion that has increased in recent decades for most countries.

In 2016, subnational government (SNG) expenditure stood at USD 6 820 per capita on average in the OECD area, accounting for 16.2% of GDP and 40.4% of total public expenditure (Figure 5.1). Among OECD countries, the SNG share of total public expenditure varied from less than 8% in Greece and Ireland, to 76.2% in Canada. SNG spending may vary according to whether the country is federal or unitary, its size and territorial organisation, the level of decentralisation and the nature of responsibilities for certain sectors (Figure 5.1).

5.1. Subnational government expenditure as a percentage of GDP and total public expenditure, 2016

Note: Federal countries: dark brown markers.


In federal countries, SNG expenditure reached USD 8 940 per capita, which account for 19.2% of GDP and 50.0% of public expenditure. In Canada, the value was USD 14 140 per capita, i.e. 31.6% of national GDP. In federal countries, the share of expenditure carried out by the local government compared to that of state government varied in 2016: while 47% of SNG expenditure is carried out by municipalities in Austria and 38% in Germany, municipalities represented only 14% of SNG expenditure in Australia and 16% in Mexico (Figure 5.2).

5.2. Public expenditure per capita by level of government (USD PPP, 2016)


In unitary countries, local government expenditure is lower than in federal countries, representing on average USD 4 250 per capita, or 9.2% of GDP and 28.7% of public expenditure in 2016. While in Chile, Greece, Ireland, New Zealand and Turkey local governments have limited competencies and spending capacity, in Japan and European Nordic countries local expenditure amounts for a significant share of public expenditure. In Denmark, for example, SNG expenditure amounts to USD 17 070 per capita (higher than in Canada), corresponding to 34.8% of GDP and 65.0% of public expenditure, which is primarily due to the fact that municipalities administer a number of social security transfers.


General government includes four sub-sectors: central/federal government and related public entities; federated government (“states”) and related public entities; local government i.e. regional and local governments and related public entities; and social security funds. Data are consolidated within these four sub-sectors. Subnational government is defined as the sum of state governments and local/regional governments.

Expenditure comprises: “current expenditure” and “capital expenditure” (see Annex D for a detailed definition).

The OECD averages are presented as the weighted average of the OECD countries for which data are available, unless otherwise specified (i.e. unweighted average, arithmetic mean, OECD UWA). Data in USD use Purchasing Power Parities.

SNG spending responsibilities have changed over the past 20 years, notably as a result of decentralisation processes that have transferred responsibilities to the subnational level in sectors such as education, health, social protection, economic development, urban and spatial planning, etc. This was the case in Spain, Sweden, Denmark, Finland, Belgium, Germany and Poland. In Belgium, the significant increase of the weight in GDP (6 percentage points) and in public expenditure (11 points) is explained by the implementation of the 6th State reform of 2011 (in effect since 2014), which has devolved new responsibilities from the federal government to the regions and communities. Some OECD countries, however, have recentralised and thus the share of SNG expenditure has decreased over the last 20 years and especially since the crisis (e.g. Ireland, Hungary) (Figure 5.3).

5.3. Changes in subnational expenditure, as a % of total public expenditure and of GDP (1995-2006)


The share of SNG expenditure as an indication of spending autonomy should be interpreted with caution. While it often provides a valuable macroeconomic overview of the level of decentralisation, it can also lead to an overestimation of subnational expenditure autonomy. In fact, it does not always assess the real degree of decision-making power and action that SNGs have in terms of spending. In some countries, the subnational “spending autonomy” can be restricted because of mandatory spending (acting as “paying agent” for example for teachers’ salaries or social security benefits), regulatory constraints or budget norms.


OECD (2018), National Accounts Statistics (database), and estimates from IMF Government Statistics for Australia and Chile.

OECD (2018), “Subnational Government Finance”, OECD Regional Statistics (database),

See Annex B for data sources and country-related metadata.

Reference years and territorial level

2016: National Accounts; Levels of government; 2015: Mexico, New Zealand and Turkey.

Further information

OECD (2018), Subnational Governments in OECD Countries: Key data (brochure)

Figure notes

5.1: Federal countries are in dark brown markers.

5.3: Iceland 95-2016, EU28 2000-2016, Mexico 2003-2016, Japan: 2005-2016, New Zealand 1995 -2015. No data for Australia and Turkey due to lack of time-series.

OECD9 and OECD26 refer to average for OECD federal countries for OECD unitary countries.

End of the section – Back to iLibrary publication page