Inequality between households in the national accounts Wider gaps in income than in consumption

Maryse Fesseau, National Accounts Department, INSEE Vanessa Bellamy and Émilie Raynaud, Consumer Prices, Household Income, and Living Conditions Department, INSEE

In 2003, the most affluent 20% of households received more than 40% of total disposable income, compared with an 8% share for the most modest 20% of households. This diversity also concerns the structure of consumption expenditures: the most modest households post a high share of food expenditures and the smallest share of recreational and cultural expenditures. Household saving is overwhelmingly concentrated in the high-income bracket. On average, the least well-off households−like the youngest ones−have negative savings. However, private transfers partly offset the weak income of the poorest households.

Publications for general public
Insee Première – No 1265
Paru le : 11/17/2009

Over 80% of disposable income spent on consumption

The average annual disposable income of households in metropolitan France (excluding overseas départements : DOMs) measured per consumption unit , was EUR24,910 in 2003. Of the total, 83% went on consumption expenditures, of which two-thirds were concentrated on four items: housing, food, transport, and recreational activities. The remaining 17% was saved. However, these figures, drawn from the national accounts, differ from households’ perceptions based on personal experience, for an "average" household is merely a statistical construct.

The analysis of the household account by living standards, household composition, age, and socio-occupational category provides a link between macroeconomic description and individual real-life experience (box).

The best-off households have five times as much disposable income as the most modest

We classify households by increasing living standards (definitions) into five quintiles, each representing 20% of the total, i.e., 5 million out of 25 million households. The top 20% have 40% of total disposable income ; the bottom 20% have 8% (table 1). Average disposable income per consumption unit of the most affluent households is thus 5.0 times as high as that of the most modest ones, at EUR50,030 versus EUR10,080. For households in the lowest quintile, social benefits (retirement pensions, unemployment benefits, family allowances, and basic income support) make up 52% of disposable income, compared with 25-35% for the other quintiles. By contrast, the distinctive feature of the best-off households is property income (returns on financial investments and on real estate), which constitutes 32% of their disposable income.

The share of property income rises with living standards: from the first quintile (8%) to the fourth (18%), it consists mainly of imputed rents for owner-occupiers. For the top quintile, property income is divided almost evenly between real-estate income and financial income. In fact, 81% of financial income is concentrated at the top of the distribution.

Table 1 – Household disposable income by living standards in 2003

Household disposable income by living standards in 2003
Q1 Q2 Q3 Q4 Q5 Total
Total (billion euros) 78.7 130.3 169.2 218.0 397.3 993.4
Average annual income per CU (euros) 10,080 16,410 21,040 26,750 50,030 24,910
Structure (%)
Income from economic activity, of which: 67 94 102 104 89 93
employers' social contributions 14 23 26 26 20 22
social contributions by employees and self-employed 8 10 11 11 10 10
Property income, of which: 8 11 14 18 32 21
financial income 1 2 2 4 17 9
Taxes -5 -7 -10 -13 -20 -14
Social contributions -22 -33 -36 -37 -30 -32
Social benefits, of which: 52 35 30 27 25 30
old age 21 19 19 18 20 19
unemployment 6 4 3 3 2 3
other (family benefits, basic income support, etc.) 25 12 8 6 3 8
Other transfers 1 0 0 1 4 2
  • How to read this table: For 100 euros of disposable income, the most modest households (first quintile, Q1) received, on average, 67 in earned income, 53 in benefits and transfers, and 8 in property income (real-estate and financial income), less 27 in taxes and social contributions. The total disposable income of households in the first quintile was EUR78.7 billion in 2003, i.e., an average EUR10,080 per year and per consumption unit (CU).
  • Scope of coverage: individual households residing in metropolitan France; excludes financial intermediation services indirectly measured (FISIM).
  • Sources: INSEE, 2003 national accounts, 2004 SILC Survey, 2003 Tax Income Survey, and 2006 Family Budget Survey.

Food, recreational activities, and culture as "social features"

Consumption expenditures rise with living standards: average per consumption unit ranges from EUR9,930 for households in the first quintile to EUR33,140 at the top of the distribution (table 2). Consumption gaps between modest and well-off households are therefore narrower than income gaps, at 1:3.3 versus 1:5.0. Not only does an affluent household spend three times as much as a low-income household, but the structure of its consumption differs as well. The least well-off households devote 20% of their consumption expenditures to food, i.e., 8 points more than the most affluent. In contrast, they devote 7% of their spending to recreational activities and culture, 4 points less than the most affluent.

The share of housing expenditures varies relatively little with living standards, but the differences are far greater when we distinguish between imputed rents and actual rents. Imputed rents account for 17% of expenditures by the best-off households, 10 points more than for the least well-off. The most affluent households often own their homes: 84% of the wealthiest households are owner-occupiers or  home-buyers, whereas 69% of the most modest households are tenants.

Table 2 – Household consumption expenditure by living standards in 2003

Household consumption expenditure by living standards in 2003
Q1 Q2 Q3 Q4 Q5 Total
Average annual expenditure (euros) 9,930 15,450 19,760 24,420 33,140 20,590
Structure (%)
Food and non-alcoholic beverages 20 17 16 14 12 15
Alcoholic beverages and tobacco 5 4 4 3 2 3
Apparel and footwear 5 5 5 5 5 5
Housing, water, gas, electricity, and other fuels, of which: 24 23 25 25 25 25
actual rents 7 6 5 3 2 4
imputed rents 7 10 14 16 17 14
water, gas, electricity, etc. 10 7 6 6 6 7
Furnishings, routine home maintenance 5 6 5 6 8 6
Health 5 4 4 3 2 3
Transport 12 14 16 15 14 14
Communication 4 3 3 3 2 3
Recreational activities and culture 7 9 9 10 11 10
Education 1 0 1 1 1 1
Hotels, cafés, and restaurants 4 5 5 6 7 6
Other goods and services 9 9 9 9 10 9
  • How to read this table: In 2003, the most modest households (first quintile, Q1) spent an average EUR9,930 per consumption unit (CU); 20% of this amount went on food products, compared with 15% for the average household.
  • Scope of coverage: individual households residing in metropolitan France; excludes FISIM (see table 1).
  • Sources: INSEE, 2003 national accounts, 2004 SILC Survey, 2003 Tax Income Survey, and 2006 Family Budget Survey.

One third of disposable income goes to precommitted expenditures, except among the best-off households

Some consumption expenditures−such as housing, telecommunication services, insurance, and financial services−are incurred under the terms of a contract that is hard to renegotiate or replace by another spending item at the household’s discretion in the short term. These expenditures, known as "precommitted" (or non-discretionary), take up one-third of income among the least well-off households, versus one-fifth among the best-off (table 3). The reason is that, while precommitted expenditures increase with living standards, they rise more mildly than income for the most affluent households.

Once these precommitted items have been expensed, the remaining income−called "discretionary" (definitions) −serves to pay for the other consumption expenditures (notably food, clothing, transport, and recreational activities) or is saved.

Average discretionary income per consumption unit ranges from EUR6,780 among the least well-off households to EUR39,730 among the best-off, a ratio of 1:5.9 versus 1:5.0 for disposable income. In other words, the disparities are greater for the portion of income that can be more freely allocated between consumption and saving.

Table 3 – Precommitted expenditures by living standards in 2003

Precommitted expenditures by living standards in 2003
Average annual expenditure (euros) Q1 Q2 Q3 Q4 Q5 Total
Disposable income (DI) 10,080 16,410 21,040 26,750 50,030 24,910
Precommitted expenditures 3,300 4,840 6,280 7,780 10,300 6,510
Discretionary income 6,780 11,570 14,760 18,970 39,730 18,400
Precommitted expenditures (% of DI) 33 29 30 29 21 26
  • Note: Our study offers an approximation of the concept of precommitted expenditures without, however, reproducing its exact boundaries. For example, life-insurance expenditures cannot be separated from total insurance expenditures, although they lie outside the scope of precommitted expenditures (see definitions).
  • How to read this table: In 2003, the most modest households (first quintile, Q1) spent an average EUR3,300 euros per consumption unit (CU) on precommitted expenditures, i.e., 33% of their average disposable income (EUR10,080).
  • Scope of coverage: individual households residing in metropolitan France; excludes FISIM (see table 1).
  • Sources: INSEE, 2003 national accounts, 2004 SILC Survey, 2003 Tax Income Survey, and 2006 Family Budget Survey.

The income of the least well-off households barely covers their consumption

The saving ratio, which broadly increases with living standards, is particularly high−at 35% or so−for the most affluent 20% of households (chart 1). Indeed, the top quintile generated 80% of total household savings in 2003. Depending on the estimation method, the saving ratio of the lowest quintile ranged between -11% and 1%. This means that consumption by the lowest-income households very likely exceeded their total income in 2003.

At individual level, various circumstances may explain such a gap between income and current consumption in a given year. Examples include a non-recurring item such as a car purchase, tuition costs, or a drop in income due to an unemployment or inactivity spell, at a time when the household needs to maintain a given consumption level. Households may finance this excess consumption by means of a loan, prior savings, or family support.

The saving ratio generally rises with the age of the head of household (chart 2). However, it decreases in the 60-69 age bracket, then rises again. The peak among the oldest households may reflect a desire for precautionary saving or to build an estate, but also the fact that older generations tend to consume less. Like the least well-off households, the youngest households post a negative total saving ratio (-10%).

Chart 1 – Saving ratios by living standards in 2003

  • Note: Given the imprecision of saving ratios by living standards, particularly for the most modest households, we give two estimates. Their difference is due to adjustments performed in the Family Budget Survey.
  • Scope of coverage: individual households residing in metropolitan France; excludes FISIM (see table 1).
  • Sources: INSEE, 2003 national accounts, 2004 SILC Survey, 2003 Tax Income Survey, and 2006 Family Budget Survey.

Chart 2 – Saving ratios before and after private transfers, by age of head of household in 2003

  • Note: Results may differ significantly according to the method used to estimate private transfers. We give two estimates of the saving ratio after private transfers for households aged under 30, as the results diverge by two percentage points or more.
  • Scope of coverage: individual households residing in metropolitan France; excludes FISIM (see table 1).
  • Sources: INSEE, 2003 national accounts, 2004 SILC Survey, 2003 Tax Income Survey, and 2006 Family Budget Survey.

Private transfers partly offset low income among the youngest households

The introduction of financial aid between households (such as child support and parents’ allowances to children who have left home to live on their own) alters the saving-ratio profile (chart 2). These private transfers, largely from households aged 50 or more, significantly raise the saving ratio of the youngest households, moving it from negative to positive. Coupled with their income, the financial aid received by the youngest households therefore allows them to cover their consumption expenditures. Single-parent families also benefit greatly from these transfers, which include child support and alimony (chart 3).

Chart 3 – Saving ratios before and after private transfers by household composition in 2003

  • Note: Results may differ significantly according to the method used to estimate private transfers. We give two estimates of the saving ratio after private transfers for single-parent families, as the results diverge by two percentage points or more.
  • Scope of coverage: individual households residing in metropolitan France; excludes FISIM (see table 1).
  • Sources: INSEE, 2003 national accounts, 2004 SILC Survey, 2003 Tax Income Survey, and 2006 Family Budget Survey.

Encadré

Going beyond the “average” household

The household account published in the national accounts shows household income, consumption, and saving. We can thus measure the change in household purchasing power, within a harmonised conceptual framework that allows comparisons between countries. Built at macroeconomic level, i.e., for all households in the aggregate, the account depicts conditions for an “average” household but does not describe differences between households. This partly explains the gap between the change in average purchasing power measured by INSEE and households’ perceptions of their living standards− perceptions influenced by personal circumstances that are, by definition, highly diverse. That is why several reports (Moati-Rochefort, Quinet, and Stiglitz) have recommended enhancing macroeconomic statistics with indicators that are more individualized but consistent with the overall framework.

INSEE household surveys collect microeconomic information on income and consumption that allow a study of inequalities between individuals. These surveys can yield results that diverge from macroeconomic aggregates. The reason is that their scope of coverage is confined to individual households−a narrower field than the one used in national accounting. Moreover, their figures are inevitably surrounded by some uncertainty: the numbers are obtained from sample surveys, and they are adjusted for under-reporting−or even non-reporting−by households.

Combining macroeconomic and microeconomic approaches

This study combines the two approaches− macroeconomic and microeconomic−for the first time, using 2003 data. Our goal is to break down disposable income and consumption expenditure, as measured in the national accounts, by various socioeconomic criteria: living standards, household composition, and age or socio-occupational category of the head of the household. We have used five surveys (sources) conducted at different frequencies (annual or every five years). The year 2003 is the most recent "median" year for all the surveys. The study concerns individual households in metropolitan France (i.e., excluding overseas départements [DOMs]), as they constitute the common scope of coverage for the five surveys.

For more details on the methodology, and to obtain the detailed results, see "National accounts - Public finances" - "Income, purchasing power and household consumption" - "Income, consumption and saving by household category in 2003" sub-section on the www.insee.fr website.

Sources

The breakdown of the household account is based on 2003 national-accounts data and on five INSEE surveys on income and consumption: the 2004 statistics on income and living conditions (SILC), the 2003 Tax Income Survey, the 2006 Family Budget Survey, the 2002 Housing Survey, and the 2003 Health Survey. Sociodemographic data are drawn from the Labour Force Survey and the 2003 satellite account for housing.

Définitions

In national accounting, gross disposable income (referred to here as disposable income) includes income from economic activity (gross wages and salaries of households ; profits of self-employed business owners), property income excluding unrealized and realized capital gains (dividends, interests, and rents), transfers (notably insurance benefits net of premiums paid) and social benefits (retirement pensions, unemployment benefits, family allowances, basic income support, etc.). From those incomes, national accountants subtract taxes (mainly income tax, occupancy tax, and two levies on taxable income earmarked for social-insurance funds: the Contribution Sociale Généralisée [CSG] and the Contribution au Remboursement de la Dette Sociale [CRDS]) and social contributions.

In the national accounts, all transfers between resident households are neutral and are therefore not measured. They include transfers in cash (child support, financial aid), and purchases and sales of goods. To break down the household account by category, we have had to introduce income flows designated as "private transfers", for they are not uniformly distributed among households. They consist solely of transfers in cash. Transfers in kind are excluded, as are estates and gifting.

Household consumption expenditure comprises expenditures funded directly by households. It does not include the share of health, education, and housing expenditures paid by government. By contrast, it includes imputed rents. The reason is that the national accounts treat owner-occupier households as producers of a housing service provided to themselves. Their income and consumption are accordingly increased by imputed rents, which are defined as the rents that would be charged in the private rental sector for dwellings with similar characteristics.

Some expenditures known as precommitted (or non-discretionary) are incurred under the terms of contracts that are hard to renegotiate in the short term. They include housing-related expenditures (real and imputed rents and other related expenses such as water, gas, and electricity), canteen expenses, telecommunication services, insurance (excluding life insurance) and financial services. Discretionary income is disposable income minus precommitted consumption expenditures.

To compare living standards of households of different sizes and composition, statisticians use a measure of income adjusted by consumption unit (CU) with the aid of an equivalence scale. The scale sets the following weights: 1 CU for the first adult household member ; 0.5 CU for the other members aged 14+ ; 0.3 CU for children under 14.

The share of disposable income not used for consumption expenditure constitutes savings. The saving ratio is household saving divided by gross disposable income.

Individual households are persons living in independent dwellings. The term therefore excludes persons living in institutions, such as boarding schools, young workers’ hostels, retirement homes, and prisons.