Automatic translation made using Deepl
“Nicolas” is a french meme, i.e., a viral phenomenon on the internet, often used by the right or far right, which sums up the supposed discontent of a generation of young workers who feel they pay a lot and receive little in return. Nicolas targets “Bernard and Chantal” or “Isabelle and Sylvie” retired couple on a cruise, and “Karim”, who benefits from redistribution, all supposedly thanks to the taxes paid by Nicolas.
The distributed accounts produced by the World Inequality Database are a means (imperfect, see the following section) to give a more precise meaning to the sentiment evoked by Nicolas’s imagination. graphic 1 and table 1 illustrate the comparison that can be made between countries and over time, and the identification of a particular point in redistribution: those who receive as much as they pay.
We take into account all direct and indirect taxes, excluding pension contributions (gross income), on the one hand, and redistribution through benefits and public spending on the other (adjusted income).
Six important points can be identified on graphic 1:
The income scale before and after redistribution is greater in the United States than in other countries. The average income of the top decile is just under €20,000 per month per person1 after redistribution, whereas it is half that amount (€9,000/month/person for the same decile) in France and slightly lower in Germany (€12,500/month/person for the same decile)2. The graphic 2 (based on the same data) confirms this point: after redistribution, the interdecile ratio in the United States is much higher (above 40) than in all other countries in the sample (around 10).
Between 1994 and 2024, the curves tend to shift upwards and to the right, due to gains in average per capita GDP growth over the last thirty years. However, this shift is not enough to upset the hierarchies: an individual in the top deciles of income in 2024 would have been in the top deciles of income in 1994. In most countries, individuals in the top deciles (or top 10 percentiles) will have incomes before redistribution in 2024 that are comparable to or lower than those in 1994. Only greater redistribution has led to a significant increase in income after redistribution in most countries, with the exception of the United States.
The graphic 2 and graphic 3 indicate that income inequality (measured by the change in the interdecile ratio) increased in most countries before redistribution, except perhaps in the United Kingdom, where the primary incomes of the poorest moved closer to the median income. Redistribution has stabilized the interdecile ratio in the United Kingdom, Germany, and Italy, and reduced it in France and Japan. In the United States, the sharp increase in inequality before redistribution has been offset by redistribution, but since 2020, income inequality has been declining significantly. In all countries except Japan, redistribution plays a major role in limiting the poverty rate, which would otherwise reach nearly 30% in many countries.
The income of the poorest after redistribution is much lower in the United States than in other countries where the concept of minimum or universal income leads to a significant income (including public expenditure allocated to each individual) (more than €1,500 per month per person in France, the most generous country in this respect, and just under €500 per month per person in Japan). Japan and the United States therefore have very high poverty rates after redistribution, which also corresponds to a high intensity of poverty (graphic 4).
The equilibrium point between income before taxation and redistribution (after pension contributions) and income after redistribution (including cash benefits and imputed public expenditure) is at the 70th percentile in France, i.e., a gross income excluding pension contributions of more than €4,200/month/person or, for a single employee without children, a net income after income tax of €2,550/month3. Below this income, individuals are net beneficiaries of redistribution extended to public spending. In Germany, the United Kingdom, and Italy, this equilibrium income is lower (around €3,000/month/person for gross income excluding social security contributions) and is located lower in the income distribution (around the 60th percentile). In Japan and the United States, the equilibrium point is significantly higher in the income distribution. Correspondingly, there are more winners from redistribution in the United States than in other countries.
In most countries, the break-even point before and after redistribution corresponds to an income (in constant 2024 euros) that is higher in 2024 than in 1994, as well as a higher position in the income distribution. Redistribution thus benefits the middle class more and places a greater burden on the wealthiest, who in most countries are becoming increasingly wealthy. Trickle-down is neither spontaneous nor produced by market forces, but is forced by redistribution. In Japan and Italy, the equilibrium point has remained the same over a 30-year period, which may be a consequence of low per capita income growth, preventing this forced trickle-down.
1 All amounts are in constant 2024 euros, in purchasing power parity. These amounts are supposed to be directly comparable across time and space. However, price and purchasing power parity measures raise some reservations as usual.
2 Data by decile is available in the downloadable data and is displayed on the hovering thumbnails of graphic 1.
3 In France, for this level of income, an employee pays approximately 12% of their gross salary (super gross) in mandatory pension contributions (general and supplementary schemes). A gross income of €4,200/month excluding retirement therefore corresponds to a gross salary of €4,740 per month, or for a single person (or a couple with the same income) without children, a net income after income tax withholding of €2,550 per month.
Distributed national accounts (André, Germain and Sicsic, 2023a ; Bellamy et al., 2009 ; Piketty, Saez and Zucman, 2018) are a systematic way of extending national accounting to inform this balance sheet. Distributed accounts have many limitations: they distinguish households or individuals according to their income and rarely according to other dimensions. Based on income distributions before and after redistribution, an approximate assessment of the effect of redistribution is made4. This simplifies construction and interpretation, allowing income inequality issues to be addressed directly, but it masks the complexity of non-progressive transfers or transfers between categories such as retirees and workers, single people and families, or men and women.
4 Strictly speaking, two distributions are compared, one before redistribution and the other after full redistribution (compulsory levies plus public expenditure). Assuming that full redistribution does not alter the order of the income hierarchy, comparing the two distributions provides an approximation of the net beneficiaries (winners) and net contributors (losers) of redistribution. For a more accurate approach to this issue, a complete micro-simulation model is needed for each year and each country. See Madec et al. (2018) for a full discussion and Pierre Madec’s analyses of various reforms (e.g., “Stop à la dette”, “Blank year,” “Tax relief,” “Electricity, gas”).
The issue of retirement-related transfer income raises many problems. In a recent publication (André, Germain and Sicsic, 2023b), INSEE treats retirement benefits as a transfer and concludes that redistribution has a particularly strong impact, reducing the interdecile ratio from 13 before redistribution (primary income including social security contributions for retirement) to 3 after redistribution. The approach we have chosen here is that of Bozio et al. (2024) and the World Inequality Database (WID) (Blanchet et al., 2024), where pension benefits and annuities paid by pension funds are considered as primary income (from capital for pension benefits from a pension fund) and, conversely, income is net of pension contributions. This convention has the advantage of treating deferred income more accurately by not ignoring the contributory status of pension contributions and makes it easier to compare different countries that may have very different public and private pension systems.
In doing so, Nicolas’s “Bernard and Chantal” problem is not taken into account, i.e., the possible transfer between generations when the imbalance in the pension system is resolved either by future retirees (lower pensions, later retirement) or by future contributors (higher contributions). Intergenerational accounting à la Auerbach-Kolitkoff (Auerbach, Gokhale and Kotlikoff, 1992) is supposed to answer this question: the work of Hippolyte d’Albis and his co-authors does not support the idea that young people are being sacrificed (d’Albis, 2024) — the pension system deficit remains moderate, younger generations enjoy higher average incomes than previous generations and have received more transfers due to their later entry into the labor market and longer periods of study (and therefore a shorter period of activity and a larger transfer due to their studies than the baby boomer generations). Inequalities exist between individuals of the same generation and tend to increase, but methodological issues are critical and difficult (Bonnet, 2002). We leave this important aspect of redistribution aside.
Using distributed account data (called DINA) from the World Inequality Database, we can take into account both transfers made by the fiscal and social security system—on the one hand, all direct and indirect taxes paid, and on the other, monetary benefits received—as well as public spending received, either individually (such as healthcare) or collectively (such as security spending). In the WID’s current methodology, public expenditure is distributed proportionally to income, except for public health expenditure, which is distributed on a flat-rate basis to each individual.
These choices are rather crude, and other approaches are possible. The distributed national accounts produced by INSEE for France for certain years use more sophisticated conventions (André, Germain and Sicsic, 2023b). For example, education expenditure is allocated to households according to the age and number of children in the household. Health expenditure is allocated on the basis of reimbursement data and therefore incorporates the effects of age or social category on the consumption of health goods and services. It should be noted that, moreover, the consideration of public expenditure does not take into account any element of the quality of this expenditure.
The life cycle approach, as in Allègre, Melonio and Timbeau (2012), or in generational accounting, allows for a more relevant allocation: education expenses are allocated to the individuals receiving the education (rather than to their families), and the redistributive balance is calculated over their life cycle rather than considering generations as having a fixed age. This profoundly changes the conclusions in terms of transfer, avoiding certain obvious pitfalls. Unfortunately, it requires both much more complex data processing (having micro-simulation models for each year or generation) and uncertain assumptions (what will be the taxes and transfers of generations in the coming years) that largely determine the result (Bonnet, 2002).
Identifying the Nicolas is complex, to say the least, and requires a comprehensive assessment of redistribution. The approach presented in section 1, based on distributed national accounts, masks intergenerational aspects, which, according to the latest research, remain weak (or even very weak, but this needs to be investigated in order to take into account the redistributive and intergenerational effect of the pension system) compared to inequalities within a generation, which redistribution aims precisely to reduce. Other aspects of redistribution over the life cycle—conditions of entry into the labor market, the effects of real estate prices or interest rates on mortgages—can have significant effects.
Nevertheless, the distributed income approach allows for comparisons between countries and over time. In France, as in many countries, Nicolas is at the top of the income distribution (among the richest 30%). France therefore does not have a particularly redistributive system, or at least redistributes widely through extensive socialization [of education and health] to a large proportion of individuals. Over the last thirty years, the weight of redistribution in France, as in many countries, has shifted towards the top of the income distribution (with the exception of Italy and the Netherlands, see table 1), in response to an increase in primary inequalities and still somewhat in line with the initial principles of redistribution (à la Beveridge): maintaining social cohesion, guaranteeing a decent standard of living for all, and offering opportunities to everyone.
<theme> List of 1
$ plot.background: <ggplot2::element_rect>
..@ fill :List of 9
.. .. $ f :function ()
.. .. $ x : 'simpleUnit' num 0.5npc
.. .. ..- attr(*, "unit")= int 0
.. .. $ y : 'simpleUnit' num 0.5npc
.. .. ..- attr(*, "unit")= int 0
.. .. $ width : 'simpleUnit' num 1npc
.. .. ..- attr(*, "unit")= int 0
.. .. $ height: 'simpleUnit' num 1npc
.. .. ..- attr(*, "unit")= int 0
.. .. $ hjust : num 0.5
.. .. $ vjust : num 0.5
.. .. $ extend: chr "none"
.. .. $ group : logi TRUE
.. .. - attr(*, "class")= chr [1:2] "GridTilingPattern" "GridPattern"
..@ colour : NULL
..@ linewidth : NULL
..@ linetype : NULL
..@ linejoin : NULL
..@ inherit.blank: logi FALSE
@ complete: logi FALSE
@ validate: logi TRUE