Kiri

1

NET SOCIAL EXPENDITURE INDICATORS: QUESTIONNAIRE

1. Introduction

1. The OECD Social Expenditure database (SOCX) contains information on public and private

social expenditure in OECD countries (Public social expenditure reflects budgetary resource allocation by

general government (i.e. different levels of government and social security funds), as for example old-age

cash benefits and housing allowances. Private social expenditure involves mandatory private social

protection arrangements that although legislated are not financed through funds that are part of general

government. Voluntary private social expenditure includes benefits accruing from privately operated

programmes that involve the redistribution of resources across households as well as benefits accruing

from tax advantaged individual and collective plans (often towards pension or health coverage).

2. These gross expenditure data fail, however, to account for the impact of the tax system on

spending. In this regard, there are three items that should be accounted for:

1. Tax advantages with a social purpose;

2. Direct taxation of benefit income;

3. Indirect taxation of consumption by benefit-recipients.

As the impact of each of these three items differs markedly across countries, a comparison of gross social

expenditure to GDP ratios can lead to misleading analysis of public social effort across countries. Further

details are provided in, for example, DAFFE/CFA/WP2(2002)33, Adema, Fron and Ladaique, 2011 or see

section 5. Net (after tax) social expenditure in the OECD SOCX Manual 2019 Edition: A guide to the

OECD Social Expenditure Database.

3. At their meetings in November 2002 and June 2003, delegates to the Committee on Fiscal

Affair’s Working Party No. 2 on Tax Policy Analysis and Tax Statistics, agreed to collect relevant

information for all OECD countries. Such information was last gathered in 2022 for years 2018/19 for all

38 OECD countries. Following the usual bi-annual data collection practice, this new version of the

questionnaire intends to gather information on direct taxation and tax breaks with a social purpose for 2020

to 2022. Countries are asked to return the completed questionnaire by Tuesday 30th April 2024.

4. This document first provides a concise summary of what social expenditure involves. Then, in

section 3, it provides an overview of the framework for calculating net (after tax) social expenditure

indicators. Section 4 sets out the type of information we would like to obtain from you for 2020-22.

2

2. Defining Social expenditure

5. The OECD defines social expenditures as:

The provision by public and private institutions of benefits to, and financial contributions

targeted at, households and individuals in order to provide support during circumstances which

adversely affect their welfare. Provided that the provision of the benefits and financial

contributions constitutes neither a direct payment for a particular good or service nor an

individual contract or transfer. Such benefits can be cash transfers, or can be the direct (in-kind)

provision of goods and services. Since only benefits provided by institutions are included,

transfers between households - albeit of a social nature, are not.

Social expenditure1 includes cash benefits (e.g. unemployment benefit), social services (e.g. daycare

subsidies) and tax breaks with a social purpose (e.g. tax expenditures towards families with children).

However, these tax breaks with a social purpose are generally not included in public social expenditure

totals in the National Accounts or the OECD Social Expenditure Database (SOCX).

2.1 Social vis-à-vis non social

6. Two main criteria have to be simultaneously satisfied for some expenditure item to be classified

as “social”. First, the benefits have to be intended to address one or more social purposes. Second, In order

to be considered social, programmes regulating the provision of benefits have to involve a) inter-personal

redistribution, or b) compulsory participation.

7. The purpose of the benefits is one factor in the delineation of what is social and what not. The

OECD Social Expenditure Database groups benefits with a social purpose in 9 policy areas: old-age,

survivors, incapacity related benefits, health, family, active labour market policies, unemployment

compensation, housing and other contingencies, e.g., cash transfers to those with low incomes (OECD

SOCX Manual 2019 Edition: A guide to the OECD Social Expenditure Database ). In most OECD

countries, public benefits cover most, if not all, of the aforementioned policy areas. On the other hand,

public support to general savings programmes or life insurance policies are not considered social, and

while fiscal support towards children is considered social, favourable tax-treatment because of marital

status is not.

8. Benefits are “social” if entitlement to receive benefits with a social purpose involves the inter-

personal redistribution of resources among programme participants or compulsory participation. If benefits

are the result of direct market transactions by individuals given their individual risk profiles, then the

payment is not social, even though its purpose may be. Thus, individual pension plans taken out at

prevailing market prices that are not subject to compulsion or redistribution (e.g. without fiscal support) are

not considered social. By contrast, public benefits with a social purpose practically always involve

redistribution across households, as they are either financed through general taxation or compulsory social

security contributions, leading to the redistribution of resources across the population or within population

groups (e.g. all adherents to an unemployment insurance fund).

1. Social expenditure or social spending concerns an aggregate of all (or a group of) social benefits. It does

not include contributions and other payments by households that finance social programmes. Such

payments are considered to be “social contributions”.

3

2.2 Public vis-à-vis private social expenditure

9. The distinction between public and private is on the basis of who controls the relevant financial

flows; public institutions or private bodies. Social benefits are regarded as public when general government

(that is central, state, and local governments, including social security funds) controls relevant financial

flows. Sickness benefits financed by mandatory employer and employee contributions to social insurance

funds (receipts) are by convention considered public. All social benefits not provided by general

government are within the private domain.

10. Within the group of private social benefits, a further distinction is made concerning the nature of

provision. Sometimes, governments mandate - force by legislation, employers to provide benefits to their

employees, or mandate individuals and/or employers to make contributions to private funds from which

benefits accrue. All such benefits are regarded as mandatory private social benefits. Private benefits with a

social purpose made by employers on a voluntary basis or that do not derive from mandatory contributions

to insurance plans that redistribute resources among the insured population, are considered voluntary.

Table 1: Some examples of benefits with a social purpose

Social policy area Spending item

Public (Different levels of government,

including social insurance funds)

Old age Old-age cash benefits

Services to the elderly

Age-related tax allowances

Mandatory private (employers, private

funds)

Incapacity-related Legislated continued (partial) wage-payments

in case of absence due to illness

Legislated occupational injury benefits paid by

private funds

Voluntary private (employers, private

funds)

Health Employer-financed health insurance or

occupational pensions

Other contingencies Welfare services provided by NGOs

3. A concise overview of Net Social Expenditure adjustments2

11. Gross social expenditure data can give a misleading picture of social effort across countries as

they do not account for the impact of tax systems. In this context, there are broadly speaking three features

that play an important role.

3.1 Direct taxation and social security contributions on transfers

12. In some OECD countries almost all benefits are paid net of tax; in others most benefit income is

taxed in the same way as income from work. Similarly, there are marked differences in the taxation of

benefit types. Benefits for families with children are generally not taxed, even when other transfer income

is taxed in line with earnings. Thus, there is considerable variation in the extent of taxation across countries

and across different types of benefit income.

13. In countries where governments levy direct income tax and social security contributions on cash

transfers to beneficiaries, social spending and redistribution of resources is lower than suggested by gross

spending indicators. For example, in Austria the recipient of an unemployment benefit who previously had

2. Over the years, a methodology has been developed for collating net social expenditure indicators, see

Adema, et al (1996), Adema and Einerhand (1998), Adema (1999, and 2001), and Adema and Ladaique

(2009 and 2005).

4

earnings equivalent to average earnings and who lives in a couple-family with two young children received

approximately the equivalent of EUR 19 100 in 2020, on which he or she did get a tax credit of EUR 700.

By contrast, a similar person in Sweden received approximately annual income support of EUR 23 700 but

paid EUR 7 000 in income taxes and social-security contributions so that net benefit income was EUR

16 700 (On-line Tax-Benefit calculator via www.oecd.org/els/social/workincentives). Thus, net income for

such a family in Sweden is lower than in Austria, although gross income was much higher. In aggregate

spending terms, this means that countries that tax transfer incomes rather heavily divert a significant part of

transferred income to flow back into the coffers of the Treasury. As a result, net (after tax) public spending

on unemployment benefits is about 75% of the level suggested by gross indicators in Sweden.

3.2 Indirect taxes on consumption out of benefit income.

14. Recipients of social benefits generally use their benefit income to finance consumption of goods

and services, much of which is subject to consumption taxes. For example, in Germany in 2010, duties on

coffee amounted to just under EUR 1.0 billion (OECD, 2012b), part of which was paid by benefit

recipients. Consumption taxes reduce the real value of consumption which can be financed out of a given

level of benefits; establish another flow of back in tax receipts to the Exchequer; and, affect the gross

social spending levels. For example, in order to provide benefit recipients with a net income of 100 units, a

country like the US with an average consumption tax rate close to 4% needs to pay a gross benefit of about

104 units. In EU-countries, where the average consumption tax rate is around 20% a gross payment would

have to be around 125 units to have an equivalent net value. Thus, gross spending totals in the US can be

somewhat lower than in EU-countries while generating a similar net value for benefit-recipients.

3.3 Tax breaks for social purposes.

15. Many governments of OECD countries pursue social policy objectives through the tax system.

Two main types of such measures can be distinguished. One is reduced taxation on particular sources of

income or types of household (e.g. an age-related tax allowance). This sort of tax relief is equivalent to a

variation in direct taxation of benefit income, and it is thus not necessary to account for these again. For

example, income tax exemptions for those receiving “Long-term incapacity benefit” in the

United Kingdom are accounted for while establishing the amount of direct taxes paid on benefit income. In

general, exemptions of benefits from taxation are not recorded again in as TBSP.

16. The second group of tax measures with social effects concern Tax Breaks for Social Purposes

(TBSPs) and are defined as:

“those reductions, exemptions, deductions or postponements of taxes, which: a) perform the

same policy function as transfer payments which, if they existed, would be classified as

social expenditures; or b) are aimed at stimulating private provision of benefits".

17. Tax breaks which mirror the effect of cash benefits often concern fiscal support to families with

children. Such support can be provided in different forms. First, tax systems may include tax credits for

children or associated costs (e.g. childcare-related tax credits). Sometimes, cash transfers and fiscal support

are part of one and the same programme. For example, tax relief in Germany for families with children

amounted to EUR 46.2 billion in 2019, of which EUR 27.2 billion was off-set against tax liabilities (and

thus recorded as a TBSP) and EUR 19.0 billion paid out in transfer income – to be recorded as a cash

transfer in the OECD Social Expenditure Database. The United States’ Earned Income Tax Credit (EITC)

is similar in providing benefits partly as a TBSP and partly as cash transfers.

18. Second, in many OECD countries children are included in the tax unit, benefiting the household

in question. Such measures are not a deviation from the national standard tax system, and thus may not be

5

considered as “tax expenditures” (OECD, 1996). Nevertheless, such support is clearly important, and an

estimate on the value of revenue foregone provides a useful indicator of financial support to children in

society, and so should be included in reported TBSPs.

19. Tax breaks for social purposes can also be granted with the aim to stimulate private social

protection. In addition to tax reliefs for individuals and employers who contribute to private social

insurance funds, such fiscal measures can include exemptions of donations to non-government

organisations that provide social services or support to employers for different reasons, e.g. when

supporting employee childcare needs or health coverage, or when hiring long-term unemployed workers,

etc.

4. Request for Data

20. This section discusses the type of information that we would like to obtain from countries. The

items to be considered are:

− Direct taxation of benefits;

− Tax breaks for social purposes;

4.1 Direct taxes and social security contributions on transfers

21. Across countries, information on direct tax (including Soc. Sec. Cont.) on transfer income can be

obtained from a variety of sources. These include: official administrative information; microsimulation

models based on annual income or monthly income; microsimulation based on survey data; adjustments

made on the basis of average tax rates based on microdatasets.

4.1.1 Administrative data

22. Ideally, administrative data is the preferred source of information of tax paid on benefit income

as such information is likely to be the most reliable. It may be feasible to obtain such information when

national social protection systems involve only a few taxable benefits (e.g. Germany), and/or when

completed tax forms allow for identification of tax paid on (some specified items of) transfer income. Also,

administrative information may give a picture of the aggregate of taxes paid on all public social cash

transfers, but not facilitate an allocation of tax paid on different types of transfers (e.g. unemployment vis-

à-vis pension benefits).

6

Question 1: Do you have administrative information on direct tax (including social security

contributions) paid on transfer income by benefit recipients. If yes, please complete Table Q1?

Table Q1.

Tax paid by benefit recipients on transfer income,

2020 to 2022

Amount

(in national currency)

Total tax paid (including soc. sec. cont.) on public1 transfer

income

of which:

- Income tax

- Social security contributions

Total tax paid (including soc. sec. cont.) on private1 transfer

income

of which:

- Income tax

- Social security contributions

1) Public transfer income concerns all cash benefits paid by general government (different levels of

government and social security institutions). Other social benefits, e.g. occupational pension

payments, are considered private.

Question 2: Do you have administrative information on direct tax (including social security

contributions) paid on specified transfer income items? If yes, please complete Table Q2 by

providing information at the greatest possible level of detail (e.g. by items as listed in Table Q3).

Table Q2.

Tax paid by benefit recipients on specific transfer income items,

2020 to 2022

Amount of tax

(including soc. sec. cont.) paid

(in national currency)

Total tax paid (including soc. sec. cont.) on transfer income

- item 1…

- item 2…

- item 3…

- etc.

For example, on basis of the US income-tax form, one could think of the

following list of items:

Total tax paid (including soc. sec. cont.) on transfer income

- item 1… - all social security benefits (OASDI)

- item 2… - unemployment compensation

- item 3… - pensions and annuities

- item 4 - Tax-advantaged individual retirement accounts

7

4.1.2 Estimates based on microdata sets

23. In the absence of administrative data, “microsimulation-models” and micro data sets (e.g. the

survey of personal incomes for the UK) which contain detailed information on both the incomes received

by households and their taxation can be used to generate average itemised tax rates (AITRs): e.g. the

average tax rate on public pension benefit or unemployment benefits. Indeed, the concept of AITRs has

been developed to facilitate identification of different tax levies on different social benefits. The AITR can

be defined as the total taxes paid by those receiving a given benefit, divided by the total income (from all

sources) of those receiving that benefit. Formally, the relevant calculations are:

AITR i = tu =1, n TI i/ tu=1, n I i

where: I is the amount of income-type “i”, and TI is the amount of tax paid on that particular amount of income, “i” is

the type or category of income, “tu” is a tax unit with income-type “i”, and “n” is the number of tax units in the

sample with income of type “i”. The broad income categories “i” include old-age cash benefits, unemployment

compensation, wage income, etc (see table Q3).

24. As is clear from the formula, calculation of the AITR requires information on the tax paid on

each particular benefit. If a given benefit is subject to a constant tax rate (possibly zero), this information is

straightforward to calculate. However, if it is subject to a progressive tax schedule (possibly applied to the

total of several income sources), the calculation needs more care. The methodology that we ask countries

to follow is to calculate the average tax rate a particular household faces on that income (or group of

incomes), as illustrated in Annex 1. This tax rate is applied to all household benefit income subject to the

progressive schedule. This methodology does not imply an ordering of different parts of income nor does it

require a more or less arbitrary decision on what part of household income should be “taxed” at a higher or

lower rate.

25. Depending on the nature of national social protection systems (the number of different social

programmes and taxable benefit payments) and the nature of information in microdatasets, the number of

identifiable AITRs will vary from country to country. Please complete Table Q3 in line with what is

possible given the source of information to which you have access.

Question 3: If your country has many taxable benefit items and administrative data do not allow for

identification of tax paid by benefit recipients of broad income groups, we ask you to generate

Average Itemised Tax rates per social expenditure programme (example line 1a in Table Q3) or

if this level of detail is not available the suggested broad social expenditure categories (for

example line 1 in Table Q3).

8

Table Q3.

Average Itemised Tax Rates,

2020 to 2022

Item

n° Broad Income Category 'i' 1

AITR

(%)

Amount of tax

(including soc. sec. cont.) paid

(in national currency)

1 Old-age cash benefits

1a - public pensions 2

1b - early retirement benefits

1c - private pensions 3

2 Survivors' benefits

2a - public pensions 2

2b - private pensions 3

3 Incapacity-related benefits

3a - Disability pensions

3b - Occupational Injury benefits

3c - Sickness payments 4

4 Family cash benefits

4a - Family benefits

4b - Maternity and parental leave payments

4c - Sole parent benefits

5 Active labour market policies

5a - benefits while on training

6 Unemployment

6a - unemployment insurance benefit

6b - unemployment assistance benefit

7 Housing

7a - rent subsidies

8 Other contingencies

8a - Low Income benefits

9 Wage income 5

1) Certain broad income categories may not apply to your country, please indicate. Not dissimilarly,

please indicate if pension income includes old-age cash benefits, survivor benefits and/or disability

pensions.

2) Public transfer income concerns all cash benefits paid by general government (different levels of

government and social security institutions). Other social benefits, e.g. occupational pension payments,

are considered private.

3) All pension income (old-age cash benefits) paid by the state or a social security fund.

4) If sickness benefits paid through social insurance funds (please indicate whether or not his covers

maternity and parental benefits in your data-set).

9

5) This category is included for reference purposes, but can be used to estimate tax paid by recipients of

continued wage payments in case of absence because of illness.

26. In some countries, ignoring local taxation may lead to underestimating the value of tax paid by

benefit recipients. In such cases, the preferred option is to “model” the situation analogous to Taxing

Wages (OECD, 2023). For example, for Switzerland, the situation in Canton Zürich could be used to

generate AITRs.

27. These AITRs are subsequently applied to the relevant spending items in the OECD Social

Expenditure Database leading to estimates of tax paid on particular spending items. Information derived

from microsimulation models/micro data sets can also be used to verify available administrative

information, or enrich it by adding more detail to its aggregate nature (see the previous section). In general,

the Secretariat will only publish aggregate indicators on net social expenditure, and differences taxes paid

across broad social expenditure items to illustrate the different tax burdens on different benefits.

4.2 Tax breaks with a social purpose

28. Please provide information on the revenue foregone of TBSPs other than pensions (see below) in

as much detail as possible in Table Q4, which includes different examples from different countries.

Question 4: Please provide information on the revenue foregone (accruals basis) of Tax Breaks for Social

Purposes in your country in Table Q4. Please note if you have already accounted for a Tax

breaks that is similar to cash benefits in the direct tax calculations, there is no need to record this

item again. If appropriate, please do include information on provincial and/or municipal TBSPs

and the revenue foregone related to including children in the tax unit.

10

Table Q4.

The value of Tax breaks for social purposes,

2020 to 2022

11

4.2.1 Tax treatment of pensions

29. The notion of tax breaks with a social purpose includes those measures aimed at stimulating

private pension take-up. However, tax breaks on occupational and individual pension programmes are

difficult to deal with as intervention (taxation and/or tax reliefs) occurs at, various stages of this form of

contractual savings. Tax treatment needs to be considered in three different areas:

30. Due to the complexity of calculations arising from these issues, there is no comparable data-set

available on the value of tax advantages on pensions. Nevertheless, available information suggests,

however, that the value of favourable tax-treatment of private pension arrangements can be considerable.

For example, in Germany and the United States such favourable tax treatment pensions was worth about

1% of GDP in 2019 and accounted for about 2% of GDP in Australia.( see detailed information on the

impact of the tax system on social expenditure by country) We would like to present the cost to public

budgets of the current tax system in the current financial year on tax breaks for pensions, regardless of

what effects the current tax system may have on revenues in future years. Because of the limitations in

comparability of information, such information would be presented as a memorandum item only.

Question 5: Please provide information on the revenue foregone (accruals basis) of Tax Breaks for

Pensions in your country in Table Q5.

Table Q5.

The value of Tax breaks for pensions,

2020 to 2022

− Contributions to programmes could be by employers or employees, out of taxed or untaxed income;

− The funds which invest the pension contributions on behalf of those contributing could be taxed or

untaxed;

− The payment of pension or annuity or lump-sum benefits at the end of the contributions period could

be taxed or untaxed.

12

ANNEX 1: AN EXAMPLE OF CALCULATING AVERAGE ITEMISED TAX RATES

Microsimulation models underlie the estimates on direct taxation of benefits in many countries,

especially when most transfer income is tax (e.g. Denmark and Sweden). It is relatively straightforward to

calculate average itemised tax rates (AITRs) on benefit income if households have only one source of

income. For example, if a retiree receives pension income worth 100 units at the “standard” tax rate of

20%, net transfer income is 80 units, and if all households were the same (and had no income from other

sources) the AITR on public pension income would be 20%. If the retiree were to receive non-taxable child

benefit, that income would be disregarded for the calculation of the tax rate on his/her household income,

while the AITR on child benefit would be nil for that country. It may well be that for older retirees (e.g.

over 75) a more favourable tax rate applies. If possible, it would be preferable to account for this age-

related measure while calculating the household tax rate, rather than it being separately treated as a tax

break with a social purpose.

Frequently pensioners receive income from different sources. Consider the case of a retiree who also

receives a supplementary (private) pension worth 50 Units. In the absence of progressivity in the tax

system, the household tax rate remains as before. However, the substantial increase of income may well

lead to the retiree being in a higher tax bracket, so that the “average” tax rate increases. This “higher” tax

rate is applied to all household income: different tax rates are not applied to different parts of household

income. (If the retiree were to receive child benefit that would be taxed at the same rate, unless the tax

benefit system explicitly regards this income as non-taxable). Thus, the methodology does not imply an

ordering of different parts of income, whereby different income sources are taxed differently according to

an arbitrary decision on which part of income is “basic” or “supplementary” (see Table A1). Differences in

AITRs are thus associated with income groupings wherein benefit recipients typically find themselves.

Table A1. Example of calculating AITRs on two types of income

household Basic

pension

Supplementary

pension

Total

household

income

Income

tax rate

Tax paid Allocation of tax over

pension income

components

Basic Supplementary

1 50 25 75 10% 7.5 5.0 2.5

2 75 50 125 15% 13.8 8.3 5.5

3 100 0 100 10% 10.0 10.0 0.0

4 50 100 150 15% 17.5 5.8 11.7

5 50 250 300 15% 40.0 6.7 33.3

Total 325 425 750 88.8 35.8 53

AITR basic pension income = tax paid over basic pension/total basic pension

income

11.1%

AITR supplementary pension income = tax paid over supplementary

pension/total supplementary pension income

12.5%

Assumed: standard tax rate is 10% when Income is less than 100 Units, and 15% of income over and above the 100 unit threshold.

13

BIBLIOGRAPHY

Adema, W., B. Eklind, J. Lotz, M. Einerhand and M. Pearson (1996), “Net Public Social Expenditure”,

Labour Market and Social Policy Occasional Papers, No. 19, OECD, Paris.

Adema, W. and M. Einerhand (1998), “The Growing Role of Private Social Benefits”, Labour Market and

Social Policy Occasional Paper, No. 32, OECD, Paris.

Adema, W. (1999), “Net Social Expenditure”, Labour Market and Social Policy Occasional Paper, No.

39, OECD, Paris

Adema, W. (2001), “Net Social Expenditure – 2nd edition”, Labour Market and Social Policy Occasional

Papers, No. 52, OECD, Paris.

Adema, W. and M. Ladaique (2005), “Net Social Expenditure, 2005 Edition”, Social, Employment and

Migration Working Papers, No. 29, Paris (www.oecd.org/els/workingpapers).

Adema, W. and M. Ladaique (2009), “How Expensive is the Welfare State? Gross and Net Indicators in

the OECD Social Expenditure Database (SOCX)”, OECD Social, Employment and Migration

Working Papers No. 92.

Adema, W, P.Fron and M. Ladaique (2011), Is the European Welfare State Really More Expensive?

Indicators on Social Spending, 1980-2012; and a Manual to the OECD Social Expenditure Database

(SOCX) Social, Employment and Migration Working Papers, No. 124, Paris

(www.oecd.org/social/expenditure.htm).

Adema, W, P.Fron (2019), The OECD SOCX Manual – 2019 edition

A guide to the OECD Social Expenditure Database (www.oecd.org/social/expenditure.htm)

OECD (1996), Tax Expenditures: Recent Experiences, Paris.

OECD (2023), Benefits and Wages, OECD, Paris and On-line Tax-Benefit calculator via

www.oecd.org/els/social/workincentives .

OECD (2023), The OECD Social Expenditure database (SOCX), www.oecd.org/social/expenditure.htm.

OECD (2023), Taxing Wages, Paris. (https://www.oecd.org/tax/taxing-wages-20725124.htm )

OECD (2023), Revenue Statistics 2023: Tax Revenue Buoyancy in OECD Countries, OECD Publishing,

Paris, https://doi.org/10.1787/2522770x.

1

NET SOCIAL EXPENDITURE INDICATORS: QUESTIONNAIRE

1. Introduction

1. The OECD Social Expenditure database (SOCX) contains information on public and private

social expenditure in OECD countries (Public social expenditure reflects budgetary resource allocation by

general government (i.e. different levels of government and social security funds), as for example old-age

cash benefits and housing allowances. Private social expenditure involves mandatory private social

protection arrangements that although legislated are not financed through funds that are part of general

government. Voluntary private social expenditure includes benefits accruing from privately operated

programmes that involve the redistribution of resources across households as well as benefits accruing

from tax advantaged individual and collective plans (often towards pension or health coverage).

2. These gross expenditure data fail, however, to account for the impact of the tax system on

spending. In this regard, there are three items that should be accounted for:

1. Tax advantages with a social purpose;

2. Direct taxation of benefit income;

3. Indirect taxation of consumption by benefit-recipients.

As the impact of each of these three items differs markedly across countries, a comparison of gross social

expenditure to GDP ratios can lead to misleading analysis of public social effort across countries. Further

details are provided in, for example, DAFFE/CFA/WP2(2002)33, Adema, Fron and Ladaique, 2011 or see

section 5. Net (after tax) social expenditure in the OECD SOCX Manual 2019 Edition: A guide to the

OECD Social Expenditure Database.

3. At their meetings in November 2002 and June 2003, delegates to the Committee on Fiscal

Affair’s Working Party No. 2 on Tax Policy Analysis and Tax Statistics, agreed to collect relevant

information for all OECD countries. Such information was last gathered in 2022 for years 2018/19 for all

38 OECD countries. Following the usual bi-annual data collection practice, this new version of the

questionnaire intends to gather information on direct taxation and tax breaks with a social purpose for 2020

to 2022. Countries are asked to return the completed questionnaire by Tuesday 30th April 2024.

4. This document first provides a concise summary of what social expenditure involves. Then, in

section 3, it provides an overview of the framework for calculating net (after tax) social expenditure

indicators. Section 4 sets out the type of information we would like to obtain from you for 2020-22.

2

2. Defining Social expenditure

5. The OECD defines social expenditures as:

The provision by public and private institutions of benefits to, and financial contributions

targeted at, households and individuals in order to provide support during circumstances which

adversely affect their welfare. Provided that the provision of the benefits and financial

contributions constitutes neither a direct payment for a particular good or service nor an

individual contract or transfer. Such benefits can be cash transfers, or can be the direct (in-kind)

provision of goods and services. Since only benefits provided by institutions are included,

transfers between households - albeit of a social nature, are not.

Social expenditure1 includes cash benefits (e.g. unemployment benefit), social services (e.g. daycare

subsidies) and tax breaks with a social purpose (e.g. tax expenditures towards families with children).

However, these tax breaks with a social purpose are generally not included in public social expenditure

totals in the National Accounts or the OECD Social Expenditure Database (SOCX).

2.1 Social vis-à-vis non social

6. Two main criteria have to be simultaneously satisfied for some expenditure item to be classified

as “social”. First, the benefits have to be intended to address one or more social purposes. Second, In order

to be considered social, programmes regulating the provision of benefits have to involve a) inter-personal

redistribution, or b) compulsory participation.

7. The purpose of the benefits is one factor in the delineation of what is social and what not. The

OECD Social Expenditure Database groups benefits with a social purpose in 9 policy areas: old-age,

survivors, incapacity related benefits, health, family, active labour market policies, unemployment

compensation, housing and other contingencies, e.g., cash transfers to those with low incomes (OECD

SOCX Manual 2019 Edition: A guide to the OECD Social Expenditure Database ). In most OECD

countries, public benefits cover most, if not all, of the aforementioned policy areas. On the other hand,

public support to general savings programmes or life insurance policies are not considered social, and

while fiscal support towards children is considered social, favourable tax-treatment because of marital

status is not.

8. Benefits are “social” if entitlement to receive benefits with a social purpose involves the inter-

personal redistribution of resources among programme participants or compulsory participation. If benefits

are the result of direct market transactions by individuals given their individual risk profiles, then the

payment is not social, even though its purpose may be. Thus, individual pension plans taken out at

prevailing market prices that are not subject to compulsion or redistribution (e.g. without fiscal support) are

not considered social. By contrast, public benefits with a social purpose practically always involve

redistribution across households, as they are either financed through general taxation or compulsory social

security contributions, leading to the redistribution of resources across the population or within population

groups (e.g. all adherents to an unemployment insurance fund).

1. Social expenditure or social spending concerns an aggregate of all (or a group of) social benefits. It does

not include contributions and other payments by households that finance social programmes. Such

payments are considered to be “social contributions”.

3

2.2 Public vis-à-vis private social expenditure

9. The distinction between public and private is on the basis of who controls the relevant financial

flows; public institutions or private bodies. Social benefits are regarded as public when general government

(that is central, state, and local governments, including social security funds) controls relevant financial

flows. Sickness benefits financed by mandatory employer and employee contributions to social insurance

funds (receipts) are by convention considered public. All social benefits not provided by general

government are within the private domain.

10. Within the group of private social benefits, a further distinction is made concerning the nature of

provision. Sometimes, governments mandate - force by legislation, employers to provide benefits to their

employees, or mandate individuals and/or employers to make contributions to private funds from which

benefits accrue. All such benefits are regarded as mandatory private social benefits. Private benefits with a

social purpose made by employers on a voluntary basis or that do not derive from mandatory contributions

to insurance plans that redistribute resources among the insured population, are considered voluntary.

Table 1: Some examples of benefits with a social purpose

Social policy area Spending item

Public (Different levels of government,

including social insurance funds)

Old age Old-age cash benefits

Services to the elderly

Age-related tax allowances

Mandatory private (employers, private

funds)

Incapacity-related Legislated continued (partial) wage-payments

in case of absence due to illness

Legislated occupational injury benefits paid by

private funds

Voluntary private (employers, private

funds)

Health Employer-financed health insurance or

occupational pensions

Other contingencies Welfare services provided by NGOs

3. A concise overview of Net Social Expenditure adjustments2

11. Gross social expenditure data can give a misleading picture of social effort across countries as

they do not account for the impact of tax systems. In this context, there are broadly speaking three features

that play an important role.

3.1 Direct taxation and social security contributions on transfers

12. In some OECD countries almost all benefits are paid net of tax; in others most benefit income is

taxed in the same way as income from work. Similarly, there are marked differences in the taxation of

benefit types. Benefits for families with children are generally not taxed, even when other transfer income

is taxed in line with earnings. Thus, there is considerable variation in the extent of taxation across countries

and across different types of benefit income.

13. In countries where governments levy direct income tax and social security contributions on cash

transfers to beneficiaries, social spending and redistribution of resources is lower than suggested by gross

spending indicators. For example, in Austria the recipient of an unemployment benefit who previously had

2. Over the years, a methodology has been developed for collating net social expenditure indicators, see

Adema, et al (1996), Adema and Einerhand (1998), Adema (1999, and 2001), and Adema and Ladaique

(2009 and 2005).

4

earnings equivalent to average earnings and who lives in a couple-family with two young children received

approximately the equivalent of EUR 19 100 in 2020, on which he or she did get a tax credit of EUR 700.

By contrast, a similar person in Sweden received approximately annual income support of EUR 23 700 but

paid EUR 7 000 in income taxes and social-security contributions so that net benefit income was EUR

16 700 (On-line Tax-Benefit calculator via www.oecd.org/els/social/workincentives). Thus, net income for

such a family in Sweden is lower than in Austria, although gross income was much higher. In aggregate

spending terms, this means that countries that tax transfer incomes rather heavily divert a significant part of

transferred income to flow back into the coffers of the Treasury. As a result, net (after tax) public spending

on unemployment benefits is about 75% of the level suggested by gross indicators in Sweden.

3.2 Indirect taxes on consumption out of benefit income.

14. Recipients of social benefits generally use their benefit income to finance consumption of goods

and services, much of which is subject to consumption taxes. For example, in Germany in 2010, duties on

coffee amounted to just under EUR 1.0 billion (OECD, 2012b), part of which was paid by benefit

recipients. Consumption taxes reduce the real value of consumption which can be financed out of a given

level of benefits; establish another flow of back in tax receipts to the Exchequer; and, affect the gross

social spending levels. For example, in order to provide benefit recipients with a net income of 100 units, a

country like the US with an average consumption tax rate close to 4% needs to pay a gross benefit of about

104 units. In EU-countries, where the average consumption tax rate is around 20% a gross payment would

have to be around 125 units to have an equivalent net value. Thus, gross spending totals in the US can be

somewhat lower than in EU-countries while generating a similar net value for benefit-recipients.

3.3 Tax breaks for social purposes.

15. Many governments of OECD countries pursue social policy objectives through the tax system.

Two main types of such measures can be distinguished. One is reduced taxation on particular sources of

income or types of household (e.g. an age-related tax allowance). This sort of tax relief is equivalent to a

variation in direct taxation of benefit income, and it is thus not necessary to account for these again. For

example, income tax exemptions for those receiving “Long-term incapacity benefit” in the

United Kingdom are accounted for while establishing the amount of direct taxes paid on benefit income. In

general, exemptions of benefits from taxation are not recorded again in as TBSP.

16. The second group of tax measures with social effects concern Tax Breaks for Social Purposes

(TBSPs) and are defined as:

“those reductions, exemptions, deductions or postponements of taxes, which: a) perform the

same policy function as transfer payments which, if they existed, would be classified as

social expenditures; or b) are aimed at stimulating private provision of benefits".

17. Tax breaks which mirror the effect of cash benefits often concern fiscal support to families with

children. Such support can be provided in different forms. First, tax systems may include tax credits for

children or associated costs (e.g. childcare-related tax credits). Sometimes, cash transfers and fiscal support

are part of one and the same programme. For example, tax relief in Germany for families with children

amounted to EUR 46.2 billion in 2019, of which EUR 27.2 billion was off-set against tax liabilities (and

thus recorded as a TBSP) and EUR 19.0 billion paid out in transfer income – to be recorded as a cash

transfer in the OECD Social Expenditure Database. The United States’ Earned Income Tax Credit (EITC)

is similar in providing benefits partly as a TBSP and partly as cash transfers.

18. Second, in many OECD countries children are included in the tax unit, benefiting the household

in question. Such measures are not a deviation from the national standard tax system, and thus may not be

5

considered as “tax expenditures” (OECD, 1996). Nevertheless, such support is clearly important, and an

estimate on the value of revenue foregone provides a useful indicator of financial support to children in

society, and so should be included in reported TBSPs.

19. Tax breaks for social purposes can also be granted with the aim to stimulate private social

protection. In addition to tax reliefs for individuals and employers who contribute to private social

insurance funds, such fiscal measures can include exemptions of donations to non-government

organisations that provide social services or support to employers for different reasons, e.g. when

supporting employee childcare needs or health coverage, or when hiring long-term unemployed workers,

etc.

4. Request for Data

20. This section discusses the type of information that we would like to obtain from countries. The

items to be considered are:

− Direct taxation of benefits;

− Tax breaks for social purposes;

4.1 Direct taxes and social security contributions on transfers

21. Across countries, information on direct tax (including Soc. Sec. Cont.) on transfer income can be

obtained from a variety of sources. These include: official administrative information; microsimulation

models based on annual income or monthly income; microsimulation based on survey data; adjustments

made on the basis of average tax rates based on microdatasets.

4.1.1 Administrative data

22. Ideally, administrative data is the preferred source of information of tax paid on benefit income

as such information is likely to be the most reliable. It may be feasible to obtain such information when

national social protection systems involve only a few taxable benefits (e.g. Germany), and/or when

completed tax forms allow for identification of tax paid on (some specified items of) transfer income. Also,

administrative information may give a picture of the aggregate of taxes paid on all public social cash

transfers, but not facilitate an allocation of tax paid on different types of transfers (e.g. unemployment vis-

à-vis pension benefits).

6

Question 1: Do you have administrative information on direct tax (including social security

contributions) paid on transfer income by benefit recipients. If yes, please complete Table Q1?

Table Q1.

Tax paid by benefit recipients on transfer income,

2020 to 2022

Amount

(in national currency)

Total tax paid (including soc. sec. cont.) on public1 transfer

income

of which:

- Income tax

- Social security contributions

Total tax paid (including soc. sec. cont.) on private1 transfer

income

of which:

- Income tax

- Social security contributions

1) Public transfer income concerns all cash benefits paid by general government (different levels of

government and social security institutions). Other social benefits, e.g. occupational pension

payments, are considered private.

Question 2: Do you have administrative information on direct tax (including social security

contributions) paid on specified transfer income items? If yes, please complete Table Q2 by

providing information at the greatest possible level of detail (e.g. by items as listed in Table Q3).

Table Q2.

Tax paid by benefit recipients on specific transfer income items,

2020 to 2022

Amount of tax

(including soc. sec. cont.) paid

(in national currency)

Total tax paid (including soc. sec. cont.) on transfer income

- item 1…

- item 2…

- item 3…

- etc.

For example, on basis of the US income-tax form, one could think of the

following list of items:

Total tax paid (including soc. sec. cont.) on transfer income

- item 1… - all social security benefits (OASDI)

- item 2… - unemployment compensation

- item 3… - pensions and annuities

- item 4 - Tax-advantaged individual retirement accounts

7

4.1.2 Estimates based on microdata sets

23. In the absence of administrative data, “microsimulation-models” and micro data sets (e.g. the

survey of personal incomes for the UK) which contain detailed information on both the incomes received

by households and their taxation can be used to generate average itemised tax rates (AITRs): e.g. the

average tax rate on public pension benefit or unemployment benefits. Indeed, the concept of AITRs has

been developed to facilitate identification of different tax levies on different social benefits. The AITR can

be defined as the total taxes paid by those receiving a given benefit, divided by the total income (from all

sources) of those receiving that benefit. Formally, the relevant calculations are:

AITR i = tu =1, n TI i/ tu=1, n I i

where: I is the amount of income-type “i”, and TI is the amount of tax paid on that particular amount of income, “i” is

the type or category of income, “tu” is a tax unit with income-type “i”, and “n” is the number of tax units in the

sample with income of type “i”. The broad income categories “i” include old-age cash benefits, unemployment

compensation, wage income, etc (see table Q3).

24. As is clear from the formula, calculation of the AITR requires information on the tax paid on

each particular benefit. If a given benefit is subject to a constant tax rate (possibly zero), this information is

straightforward to calculate. However, if it is subject to a progressive tax schedule (possibly applied to the

total of several income sources), the calculation needs more care. The methodology that we ask countries

to follow is to calculate the average tax rate a particular household faces on that income (or group of

incomes), as illustrated in Annex 1. This tax rate is applied to all household benefit income subject to the

progressive schedule. This methodology does not imply an ordering of different parts of income nor does it

require a more or less arbitrary decision on what part of household income should be “taxed” at a higher or

lower rate.

25. Depending on the nature of national social protection systems (the number of different social

programmes and taxable benefit payments) and the nature of information in microdatasets, the number of

identifiable AITRs will vary from country to country. Please complete Table Q3 in line with what is

possible given the source of information to which you have access.

Question 3: If your country has many taxable benefit items and administrative data do not allow for

identification of tax paid by benefit recipients of broad income groups, we ask you to generate

Average Itemised Tax rates per social expenditure programme (example line 1a in Table Q3) or

if this level of detail is not available the suggested broad social expenditure categories (for

example line 1 in Table Q3).

8

Table Q3.

Average Itemised Tax Rates,

2020 to 2022

Item

n° Broad Income Category 'i' 1

AITR

(%)

Amount of tax

(including soc. sec. cont.) paid

(in national currency)

1 Old-age cash benefits

1a - public pensions 2

1b - early retirement benefits

1c - private pensions 3

2 Survivors' benefits

2a - public pensions 2

2b - private pensions 3

3 Incapacity-related benefits

3a - Disability pensions

3b - Occupational Injury benefits

3c - Sickness payments 4

4 Family cash benefits

4a - Family benefits

4b - Maternity and parental leave payments

4c - Sole parent benefits

5 Active labour market policies

5a - benefits while on training

6 Unemployment

6a - unemployment insurance benefit

6b - unemployment assistance benefit

7 Housing

7a - rent subsidies

8 Other contingencies

8a - Low Income benefits

9 Wage income 5

1) Certain broad income categories may not apply to your country, please indicate. Not dissimilarly,

please indicate if pension income includes old-age cash benefits, survivor benefits and/or disability

pensions.

2) Public transfer income concerns all cash benefits paid by general government (different levels of

government and social security institutions). Other social benefits, e.g. occupational pension payments,

are considered private.

3) All pension income (old-age cash benefits) paid by the state or a social security fund.

4) If sickness benefits paid through social insurance funds (please indicate whether or not his covers

maternity and parental benefits in your data-set).

9

5) This category is included for reference purposes, but can be used to estimate tax paid by recipients of

continued wage payments in case of absence because of illness.

26. In some countries, ignoring local taxation may lead to underestimating the value of tax paid by

benefit recipients. In such cases, the preferred option is to “model” the situation analogous to Taxing

Wages (OECD, 2023). For example, for Switzerland, the situation in Canton Zürich could be used to

generate AITRs.

27. These AITRs are subsequently applied to the relevant spending items in the OECD Social

Expenditure Database leading to estimates of tax paid on particular spending items. Information derived

from microsimulation models/micro data sets can also be used to verify available administrative

information, or enrich it by adding more detail to its aggregate nature (see the previous section). In general,

the Secretariat will only publish aggregate indicators on net social expenditure, and differences taxes paid

across broad social expenditure items to illustrate the different tax burdens on different benefits.

4.2 Tax breaks with a social purpose

28. Please provide information on the revenue foregone of TBSPs other than pensions (see below) in

as much detail as possible in Table Q4, which includes different examples from different countries.

Question 4: Please provide information on the revenue foregone (accruals basis) of Tax Breaks for Social

Purposes in your country in Table Q4. Please note if you have already accounted for a Tax

breaks that is similar to cash benefits in the direct tax calculations, there is no need to record this

item again. If appropriate, please do include information on provincial and/or municipal TBSPs

and the revenue foregone related to including children in the tax unit.

10

Table Q4.

The value of Tax breaks for social purposes,

2020 to 2022

11

4.2.1 Tax treatment of pensions

29. The notion of tax breaks with a social purpose includes those measures aimed at stimulating

private pension take-up. However, tax breaks on occupational and individual pension programmes are

difficult to deal with as intervention (taxation and/or tax reliefs) occurs at, various stages of this form of

contractual savings. Tax treatment needs to be considered in three different areas:

30. Due to the complexity of calculations arising from these issues, there is no comparable data-set

available on the value of tax advantages on pensions. Nevertheless, available information suggests,

however, that the value of favourable tax-treatment of private pension arrangements can be considerable.

For example, in Germany and the United States such favourable tax treatment pensions was worth about

1% of GDP in 2019 and accounted for about 2% of GDP in Australia.( see detailed information on the

impact of the tax system on social expenditure by country) We would like to present the cost to public

budgets of the current tax system in the current financial year on tax breaks for pensions, regardless of

what effects the current tax system may have on revenues in future years. Because of the limitations in

comparability of information, such information would be presented as a memorandum item only.

Question 5: Please provide information on the revenue foregone (accruals basis) of Tax Breaks for

Pensions in your country in Table Q5.

Table Q5.

The value of Tax breaks for pensions,

2020 to 2022

− Contributions to programmes could be by employers or employees, out of taxed or untaxed income;

− The funds which invest the pension contributions on behalf of those contributing could be taxed or

untaxed;

− The payment of pension or annuity or lump-sum benefits at the end of the contributions period could

be taxed or untaxed.

12

ANNEX 1: AN EXAMPLE OF CALCULATING AVERAGE ITEMISED TAX RATES

Microsimulation models underlie the estimates on direct taxation of benefits in many countries,

especially when most transfer income is tax (e.g. Denmark and Sweden). It is relatively straightforward to

calculate average itemised tax rates (AITRs) on benefit income if households have only one source of

income. For example, if a retiree receives pension income worth 100 units at the “standard” tax rate of

20%, net transfer income is 80 units, and if all households were the same (and had no income from other

sources) the AITR on public pension income would be 20%. If the retiree were to receive non-taxable child

benefit, that income would be disregarded for the calculation of the tax rate on his/her household income,

while the AITR on child benefit would be nil for that country. It may well be that for older retirees (e.g.

over 75) a more favourable tax rate applies. If possible, it would be preferable to account for this age-

related measure while calculating the household tax rate, rather than it being separately treated as a tax

break with a social purpose.

Frequently pensioners receive income from different sources. Consider the case of a retiree who also

receives a supplementary (private) pension worth 50 Units. In the absence of progressivity in the tax

system, the household tax rate remains as before. However, the substantial increase of income may well

lead to the retiree being in a higher tax bracket, so that the “average” tax rate increases. This “higher” tax

rate is applied to all household income: different tax rates are not applied to different parts of household

income. (If the retiree were to receive child benefit that would be taxed at the same rate, unless the tax

benefit system explicitly regards this income as non-taxable). Thus, the methodology does not imply an

ordering of different parts of income, whereby different income sources are taxed differently according to

an arbitrary decision on which part of income is “basic” or “supplementary” (see Table A1). Differences in

AITRs are thus associated with income groupings wherein benefit recipients typically find themselves.

Table A1. Example of calculating AITRs on two types of income

household Basic

pension

Supplementary

pension

Total

household

income

Income

tax rate

Tax paid Allocation of tax over

pension income

components

Basic Supplementary

1 50 25 75 10% 7.5 5.0 2.5

2 75 50 125 15% 13.8 8.3 5.5

3 100 0 100 10% 10.0 10.0 0.0

4 50 100 150 15% 17.5 5.8 11.7

5 50 250 300 15% 40.0 6.7 33.3

Total 325 425 750 88.8 35.8 53

AITR basic pension income = tax paid over basic pension/total basic pension

income

11.1%

AITR supplementary pension income = tax paid over supplementary

pension/total supplementary pension income

12.5%

Assumed: standard tax rate is 10% when Income is less than 100 Units, and 15% of income over and above the 100 unit threshold.

13

BIBLIOGRAPHY

Adema, W., B. Eklind, J. Lotz, M. Einerhand and M. Pearson (1996), “Net Public Social Expenditure”,

Labour Market and Social Policy Occasional Papers, No. 19, OECD, Paris.

Adema, W. and M. Einerhand (1998), “The Growing Role of Private Social Benefits”, Labour Market and

Social Policy Occasional Paper, No. 32, OECD, Paris.

Adema, W. (1999), “Net Social Expenditure”, Labour Market and Social Policy Occasional Paper, No.

39, OECD, Paris

Adema, W. (2001), “Net Social Expenditure – 2nd edition”, Labour Market and Social Policy Occasional

Papers, No. 52, OECD, Paris.

Adema, W. and M. Ladaique (2005), “Net Social Expenditure, 2005 Edition”, Social, Employment and

Migration Working Papers, No. 29, Paris (www.oecd.org/els/workingpapers).

Adema, W. and M. Ladaique (2009), “How Expensive is the Welfare State? Gross and Net Indicators in

the OECD Social Expenditure Database (SOCX)”, OECD Social, Employment and Migration

Working Papers No. 92.

Adema, W, P.Fron and M. Ladaique (2011), Is the European Welfare State Really More Expensive?

Indicators on Social Spending, 1980-2012; and a Manual to the OECD Social Expenditure Database

(SOCX) Social, Employment and Migration Working Papers, No. 124, Paris

(www.oecd.org/social/expenditure.htm).

Adema, W, P.Fron (2019), The OECD SOCX Manual – 2019 edition

A guide to the OECD Social Expenditure Database (www.oecd.org/social/expenditure.htm)

OECD (1996), Tax Expenditures: Recent Experiences, Paris.

OECD (2023), Benefits and Wages, OECD, Paris and On-line Tax-Benefit calculator via

www.oecd.org/els/social/workincentives .

OECD (2023), The OECD Social Expenditure database (SOCX), www.oecd.org/social/expenditure.htm.

OECD (2023), Taxing Wages, Paris. (https://www.oecd.org/tax/taxing-wages-20725124.htm )

OECD (2023), Revenue Statistics 2023: Tax Revenue Buoyancy in OECD Countries, OECD Publishing,

Paris, https://doi.org/10.1787/2522770x.