Edition 2026.1 — income, spending committed in advance and housing of households in France, 2025 national accounts and prices as of 6 October 2026

What French households really have left

Purchasing power, money available, the end-of-month feeling: three stories that do not always match — and for whom the gap is widest

State of the data and of the study as of 6 October 2026. The 2026 values are provisional or forecast, and each carries its date.

A household can earn more, read that purchasing power is rising, and still find that it has less left at the end of the month.

The statistic and the household do not contradict each other: they do not measure the same thing. The first describes what the whole country's income can buy; the second looks at a bank account, once the rent, the loan, the bills and the fuel have been paid.

This study puts the three measures side by side, explains what each counts and leaves out, then looks for whom the gap is widest.

The answer

The gap between measurement and the end of the month comes mainly from the housing one occupies, the make-up of the household, where one lives, the way one counts — per household or per person — and a memory of prices that retains the rises. Not from an average constraint that keeps climbing: on average, the share of income committed in advance has stopped rising since the mid-2000s, as defined by the national accounts, which count neither food nor fuels in it.

For whom the gap is widest, and why, is demonstrated in the full analysis.

Where do things stand today?

Purchasing power per consumption unit, last closed year: −0.7%
Price rise over a year, latest month published: +3.0%
Household confidence (long-run average: 100): 86

These three benchmarks do not add up: neither the same unit nor the same date.

The criteria that would bring this answer down were written before drafting; they are set out with the method.

Back to the answer

Chapter 1 · open to all

Three measures people confuse

Three quantities answer the question “how much do I have left?”. Two are measured; the third, the one everyone has in mind, is measured by nobody.

Discretionary income, the second measure, removes from income the pre-committed spending, whose biggest line is an imputed rent. The third, the one everyone has in mind, exists in no official statistic: that is why this study gives no average “money left to live on” in euros.

Purchasing power

measured

What the income of all households can buy, once prices are taken into account.

Contains: wages, the income of the self-employed, property income, cash social benefits, minus taxes and contributions — and a rent that owners are deemed to pay themselves

Does not contain: what loan interest has already taken before it, and everything the State provides free of charge — school, hospital — which is not money

INSEE, annual national accounts

Discretionary income

measured

What remains of income once the spending that cannot be renegotiated from one month to the next has been paid: housing, home energy, insurance, subscriptions, canteen.

Contains: the rent that owners pay themselves, on both sides but not for the same amount: as spending at its full value, as income for what is left of it once maintenance, property tax and loan interest have been paid

Does not contain: the monthly payment on a mortgage — neither its interest, already removed from income, nor its principal, which is not consumer spending

INSEE, annual national accounts (table 2.102)

The money left in the account

measured by no source

What a household actually sees at the end of the month, after paying its rent or mortgage, its bills and its subscriptions.

Contains: the whole mortgage payment, principal included — which the household does pay every month

Does not contain: the rent that owners pay themselves: it never went through their account

No public producer at national level

A few words recur throughout the study. Disposable income is the starting point; purchasing power corrects it for prices; the consumption unit makes it possible to compare households of different sizes, and living standard ranks households by tenth. An buyer repays a loan for their home; financial services also count what banks earn on rates; a gap is read in percentage points; and a scenario is never a forecast.

What to remember

When the statistic and the household do not say the same thing, they are often not measuring the same thing.

The series

What do household accounts show over the long run?

One chart open to everyone, 5 in the full analysis. Each carries its question, what to see in it, what cannot be concluded from it and what it does not measure.

1 free chart5 premium chartsThe premium panels work — they are deliberately closed. Click one to see what it contains.

Does average purchasing power still rise every year?

FreeObserved 1960-2025% annual change

Reading

Purchasing power per consumption unit

Source — INSEE, 2025 annual national accounts, table 2.101 (income and purchasing power). Axis: The axis is centred on zero: above it, income buys more than the year before; below it, less.

What this chart cannot show, and how to use it

Limits

The 2025 accounts are provisional and have already been revised: INSEE first announced a rise.

Interactions

Hover over the plot: a tooltip gives the point's values, unit and status. Click to hold the selection — it does not disappear when the pointer moves away, so that the reading stays displayed while you read it. On the keyboard, the tab key brings focus to the plot, the arrow keys move the selection, Home and End go to the ends, and Escape closes. On a touch screen, one press selects a point, a second releases it.

Back to the answer

Chapter 2 · reserved

What becomes of a hundred euros of income?

Before discussing rises or falls, one needs to know where the money goes. Scaled to a hundred euros, the income of all households is divided up as follows.

Reserved · full analysis

A hundred euros of income, item by item: 9 items, what is committed in advance, what is committed without being paid — the rent nobody pays —, other spending and what is not consumed

Back to the answer

Chapter 3 · reserved

Is the share of income committed in advance still rising?

It is the most widespread explanation for difficult ends of the month. The national accounts make it possible to check it year by year, since 1959.

Reserved · full analysis

The share of income committed in advance from 1959 to 2025, year by year: when it rose, when it stopped, what the measurement of banking services has added since 2023, and the same rise measured from several starting years

Back to the answer

Chapter 4 · reserved

At equal income, does housing separate households?

At equal income, two families do not have the same margin. The first reason lies in how they are housed.

Reserved · full analysis

Housing at equal income: the housing cost ratio of 4 tenures in 4 living-standard bands, 4 calculated cases for the same family, and how much principal a mortgage payment repays

Back to the answer

Chapter 5 · reserved

2021-2026: where did the shock go?

Prices have risen sharply since 2021. So have incomes. The result depends on the year one looks at, and on what one calls income.

Reserved · full analysis

2021-2026, where the shock went: prices item by item, purchasing power year by year, discretionary income, the minimum wage and the basic wage against prices, and why 2025 and 2026 do not fall for the same reasons

Back to the answer

Chapter 6 · reserved

For whom is the gap widest?

An average adds up winners and losers. INSEE's studies make it possible to tell them apart — by living standard, family type, age and place of residence.

Reserved · full analysis

Not the same story for all: the price rise experienced by income, age and housing tenure, the weight of the 2023 shock in 10 living-standard bands, in euros and as a share of income, what income rises offset in 2022 then in 2023 — single-parent families, young households and rural municipalities included —, poverty and saving by living standard

Back to the answer

Chapter 7 · reserved

Why does the feeling remain bad?

Even when the average rises, many households say the opposite. Several reasons are documented; none suffices on its own.

Reserved · full analysis

Why the feeling stays bad: the same income measured per person, per consumption unit and per household, smaller households, household confidence and the inflation they perceive

Back to the answer

Chapter 8 · reserved

Are firms making households pay for the shock?

In 2026, energy also raises firms' costs. The accounts show who absorbs the shock.

Reserved · full analysis

When firms' margins become households' margins: what the accounts show, what INSEE says about it, and what no source quantifies

Back to the answer

Chapter 9 · reserved

What could still change the end of the month?

The study does not predict. It translates the scenarios published by institutions, and the shocks already observed, into effects on the end of the month — and it says, for each, what is established and what is not.

Reserved · full analysis

The shocks that could change the end of the month, as conditional calculations: 3 Banque de France scenarios turned into real wages, 3 oil prices turned into inflation by INSEE's rule, 2 rate rises turned into monthly payments — each element classified: observed, documented mechanism, projection, scenario or not established

Back to the answer

Chapter 10 · reserved

What remains of the starting ideas?

9 hypotheses written before the investigation, and the criteria that would overturn the answer — each assessed on the data, no verdict written by hand.

Reserved · full analysis

The assessment: the full answer, the 9 starting hypotheses and their calculated verdict, the 5 refutation criteria evaluated on the data, and the 5 other common beliefs

Open to all

What is said, and what the figures make of it

Each verdict is calculated from the position of the study's frozen conclusions. None is written by hand.

2 of the 7 claims in circulation, set against the 41 frozen conclusions — settled before the writing — of this study; the others are in the full study. 3 others were set aside, and the reason is written further down.

The verdict is not written: it is calculated. Each conclusion of the study is declared as going in the direction of the claim, against it, or relevant without settling — and a single rule turns these positions into a verdict. Conclusions on both sides give “misleading”; no conclusion that settles gives “cannot be settled”.

Misleading · 1
False · 1

Misleading
“Constrained spending keeps on climbing.”
They rose a great deal until the early 1990s, mainly because France became a nation of owners. Since the mid-2000s, their share of income has stopped rising on average, and part of the recent jump comes from the way banking services are measured.
What this verdict does not cover: A national average, as defined by the national accounts: food and fuels are not in it, and it says nothing about any particular household.
See why— 3 conclusions of the study
Why this idea circulates: Rent, energy and subscriptions leave every month and get noticed; and for some households — private tenants, recent buyers — their weight has indeed increased.
- supports it · value withheld
  The share of income committed in advance rose sharply between the late 1950s and the early 1990s.
- goes against it · value withheld
  Since the mid-2000s, the share committed in advance has stopped rising on average.
- goes against it · value withheld
  Depending on the starting year, the share committed in advance rose, stalled or fell: the rise belongs to the 2000s, and part of the recent jump comes from how banking services are measured.
What would overturn the verdict: A rise of at least two points in the share excluding financial services over the last published decade, or since the start of the 2021 shock.
Read the chapter that demonstrates it
False
“Purchasing power is what I have left at the end of the month.”
Purchasing power measures what the income of all households can buy; it counts a rent owners do not pay, and it ignores the principal a buyer repays every month. What remains in the account is measured by no official statistic.
What this verdict does not cover: The two published measures — purchasing power and discretionary income — frame the money that remains; they do not give it.
See why— 2 conclusions of the study
Why this idea circulates: The word is the same in households' mouths and in press releases, and there is no other national measure to set against it.
- goes against it
  No official statistic measures, for all households, the money left in the account once constrained spending and the whole loan have been paid.
- goes against it · value withheld
  A large part of the spending committed in advance is a rent nobody pays: the one owners are deemed to pay themselves.
What would overturn the verdict: An official statistic of the money left once constrained spending and the whole loan have been paid.
Read the chapter that demonstrates it

Three claims that this study cannot test

“The end of the month is harder than thirty years ago.”
No official series measures the money left in households' accounts, neither today nor thirty years ago.
“Single-parent families have lost the most since 2021.”
The offset of the shock by household type is published only for 2023; no series follows these families over the whole period.
“The Hormuz blockade is depriving France of fuel.”
No public source says what share of the energy France imports passes through the strait; the documented channel is the world price.

Open to all

Where do these figures come from?

Each family of sources, with its producer, its role, its cut-off date and what it does not measure.

Source	What it contributes	What it does not measure
INSEE — annual national accounts, 2020 base (tables 2.101 to 2.104 and detailed pre-committed spending)release of 29 May 2026	income, purchasing power per person, per household and per consumption unit, pre-committed spending item by item, discretionary income and saving, from 1959 to 2025	an average of all households: it says nothing about any particular household, nor about the mortgage payment, nor about what is left in an account
INSEE — tax and social income survey (Insee Première No. 2117)2024 incomes, released on 9 July 2026	median living standard, the bottom and top deciles and monetary poverty, from 1996 to 2024	neither spending nor cash: a living standard is an income per consumption unit
INSEE — France, portrait social 2025, Statistics on income and living conditions survey2023 incomes and spending	the share of income spent on housing by tenure status and living standard — the best public approximation of the money that remains	a single income year (2023), as a median: it traces no change and is not comparable with the previous edition, which published averages
INSEE — France, portrait social 2024 (microsimulation) and price indices by household category2023 for the microsimulation, 2025 for the indices	the weight of the 2023 price shock by living standard, in euros and as a share of living standard, and the price rise experienced by age, tenure and income from 2020 to 2025	prices actually paid by each household: the category indices apply the same prices to different baskets
INSEE — distributed national accounts (Insee Analyses No. 119, Insee Focus No. 338)2023 and 2022	2023 purchasing power and 2022 saving by fifth of living standard	a direct measurement: the consumption of low-income households is imputed, and the negative saving of the bottom fifth is not debt
INSEE — consumer price indices, 2025 base, and monthly household confidence surveySeptember 2026, provisional · to be re-queried before publication	prices month by month up to September 2026, and what households say they feel	the last month is provisional; an opinion is not a price measurement
Banque de France — September 2026 projections; INSEE — September 2026 economic outlook15 September 2026 · to be re-queried before publication	the inflation and wage scenarios for 2026 to 2028, and the rule linking the oil price to inflation	no projection is a certain forecast, and no scenario is announced
Directorate-general for energy and climate, Energy Regulatory Commission, Banque de Franceweek of 2 October 2026 for fuels · to be re-queried before publication	fuel prices at the pump, regulated tariffs and interest rates on new loans	readings at a date: they do not say what a given household paid over the year

Open to all · by necessity

What do official statistics not allow us to conclude?

This block says what this study cannot settle. Reserving it would amount to selling the only information that prevents mistakes about what one is buying.

The share committed in advance, by income, since 2017
the last measurement by living standard uses the 2017 Household Budget Survey (France Stratégie, 2021): it sets spending committed in advance, loan repayments included, against the household's total spending — not its income, and not in the national accounts definition. No later survey has been exploited in this way.
What its absence forbids concluding: that a given share of low-income households' income is committed in advance today, or that the gap with well-off households has widened further since 2017.
What would close it: the exploitation of the next Household Budget Survey, with the same definition, by INSEE, DREES or France Stratégie.
What imputed rent brings in as income
the tables used give imputed rent as spending, but not its counterpart as income: the surplus owner-occupiers draw from it, once maintenance and property tax have been paid, is mixed with that of households who let a dwelling to others.
What its absence forbids concluding: that imputed rent is neutral for owners' discretionary income, or on the contrary that it reduces it by a given amount.
What would close it: the publication of the production account of owner-occupied housing, separate from letting to third parties.
The money left in the account, at national level
no public series measures, for all households, what remains once constrained spending AND the whole mortgage payment have been paid, with imputed rents removed.
What its absence forbids concluding: an average “money left to live on” for the French in euros, or its change.
What would close it: a publication combining household accounts and principal repayments on housing loans, or a recent budget survey made public.
The share committed in advance in 2026
pre-committed spending and discretionary income are published only once a year. The last vintage is 2025; the 2026 energy rise is not in it.
What its absence forbids concluding: a share committed in advance for 2026.
What would close it: the 2026 annual national accounts, expected at the end of May 2027.
The principal repaid by home buyers
none of the tables used gives the annual amount of mortgage principal repaid by households, nor its share of their income.
What its absence forbids concluding: the share of home buyers' income absorbed by principal alone, or its change.
What would close it: a public series of principal repayments on housing loans, set against borrowers' incomes.
The change in the housing cost ratio by tenure
the 2025 edition of France, portrait social publishes MEDIAN housing cost ratios by quartile of living standard; the 2024 edition published AVERAGES by quartile of disposable income. The two are not comparable, and the 2024 Housing survey has not yet been exploited.
What its absence forbids concluding: that the housing cost ratio of tenants or home buyers has risen or fallen.
What would close it: a homogeneous series of housing cost ratios by tenure, or the exploitation of the 2024 Housing survey.
An official series of perceived inflation in France
INSEE asks the quantified question but publishes only opinion balances; the Banque de France publishes four annual survey waves, from 2022 to 2025.
What its absence forbids concluding: a month-by-month change in perceived inflation, or its value in 2026.
What would close it: the publication of the quantified answers of the household confidence survey.
What, in the energy France imports, passes through the Gulf
no public source says what share of the oil, refined fuels and gas imported by France passes through the Strait of Hormuz or the Red Sea. The published origins are countries of extraction, not routes.
What its absence forbids concluding: that a closure of the strait would deprive France of a given quantity of energy.
What would close it: a publication by customs or the energy statistics service by maritime route.
How fast government bond rates pass through to mortgage rates
the Banque de France and INSEE describe the mechanism — higher long-term rates pass through gradually to the cost of credit — without publishing its size or its lag.
What its absence forbids concluding: that a given rise in government bond rates makes a household's loan dearer by a given amount.
What would close it: an estimate published by the Banque de France of the pass-through of sovereign rates to rates on new housing loans.

Open to all

What exactly does this study assert?

Each conclusion, with its nature, period, scope, source, formula and limits. The values of the reserved chapters appear there without their figure.

Official data
a value published as such by its producer
France ÉcoScan calculation
a ratio or cumulative figure calculated from published values; the formula is given
Calculation under stated assumptions
a calculation that depends on assumptions written alongside it, and changes if they change
Institutional projection
what an institution forecasts in its central scenario
Conditional scenario
what would happen if an assumption came true — never an announcement
Reported perception
what respondents say they feel
Established absence
a quantity sought that no public source gives — it is a finding, dated

The 41 claims, one by onevalue, unit, period, scope, formula, revision, sources, limits

Official datahigh confidence
In 2025, the purchasing power of disposable income per consumption unit fell by 0.7%.
Value:
−0.7%
Unit:
% annual change
Period:
2025
Scope:
all households, including sole proprietors
Formula:
published by INSEE
Where to read it:
table 2.101, row “purchasing power of gross disposable income per consumption unit”
Revision:
provisional, revisable in May 2027
Not comparable with:
an average of all households; neither per household nor per person
Sources:
INSEE, 2025 annual national accounts, table 2.101 (income and purchasing power)
- INSEE attributes it mainly to the after-effect of the exceptional dividends paid in 2024 — a property income, which does not weigh in the same way on all households.
Official datamedium confidence
In September 2026, consumer prices rose by 3.0% over a year, according to INSEE's provisional estimate.
Value:
+3.0%
Unit:
% year on year
Period:
September 2026
Scope:
consumer price index, all households
Formula:
published by INSEE
Where to read it:
series 011814632, latest reading; Informations rapides No. 242, page 1
Revision:
provisional; final result on 15 October 2026
Not comparable with:
a twelve-month change, not an annual average
Sources:
INSEE, monthly consumer price indices, 2025 base · Informations rapides No. 242 — consumer price index, provisional results for September 2026
- An end-of-month estimate, based on a restricted set of prices.
Reported perceptionhigh confidence
In September 2026, the household confidence indicator stands at 86, against a long-run average of 100.
Value:
86
Unit:
index, 1987-2025 average = 100
Period:
September 2026
Scope:
about 2,000 households surveyed each month
Formula:
published by INSEE; average recomputed over 1987-2025: 100.0
Where to read it:
series 001587668, seasonally adjusted
Revision:
slightly revisable each month
Not comparable with:
a summarised opinion, not a measurement of income or prices
Sources:
INSEE, monthly household confidence survey
- An opinion proves what people feel, never what prices did.
Official datahigh confidencedemonstrated in a reserved chapter
The share of income committed in advance rose sharply between the late 1950s and the early 1990s.
Value:
reserved
Unit:
% of gross disposable income
Period:
1959 and 1993
Scope:
all households; “pre-committed” spending as defined by INSEE, imputed rents included
Formula:
published by INSEE
Where to read it:
table 2.103, row “share of pre-committed spending in gross disposable income”
Revision:
final accounts up to 2023, semi-final in 2024, provisional in 2025
Not comparable with:
INSEE's notion is not that of budget surveys, which also count loan repayments
Sources:
INSEE, 2025 annual national accounts, table 2.103 (pre-committed spending)
- A national average: it says nothing of any particular household's share.
France ÉcoScan calculationhigh confidencedemonstrated in a reserved chapter
A large part of this doubling comes from the spread of home ownership and the measurement of banking services, not from new bills.
Value:
reserved
Unit:
% of the rise in the share, 1959-1993
Period:
1959 and 1993
Scope:
all households
Formula:
rise in (item / gross disposable income) between 1959 and 1993, set against the rise in the published share
Where to read it:
detailed pre-committed spending, rows “imputed housing rents” and “financial services”
Revision:
final accounts up to 2023, semi-final in 2024, provisional in 2025
Not comparable with:
an accounting decomposition, not a causal explanation
Sources:
INSEE, 2025 annual national accounts, pre-committed spending by item · INSEE, 2025 annual national accounts, table 2.103 (pre-committed spending)
- Imputed rent rises with the number of owners and the value of dwellings: it says France became a nation of owners, not that it pays more bills.
France ÉcoScan calculationhigh confidencedemonstrated in a reserved chapter
Since the mid-2000s, the share committed in advance has stopped rising on average.
Value:
reserved
Unit:
% of gross disposable income
Period:
2005-2025
Scope:
all households
Formula:
smallest and largest published value since the first year from which the share no longer falls below 29%
Where to read it:
table 2.103, same row, vintages 2005 to 2025
Revision:
final accounts up to 2023, semi-final in 2024, provisional in 2025
Not comparable with:
a stable share is not stable spending: both quantities rise at the same pace
Sources:
INSEE, 2025 annual national accounts, table 2.103 (pre-committed spending)
- The gap between these two bounds comes from the years, not from uncertainty: it is neither a margin of error nor a confidence interval.
- The stability of the average may hide widening gaps between households: from 2001 to 2017, the only measurement by living standard shows them widening; none allows this to be checked since.
France ÉcoScan calculationhigh confidencedemonstrated in a reserved chapter
Depending on the starting year, the share committed in advance rose, stalled or fell: the rise belongs to the 2000s, and part of the recent jump comes from how banking services are measured.
Value:
reserved
Unit:
percentage points of income since 2013
Period:
2001-2025
Scope:
pre-committed spending minus financial services (including financial intermediation services indirectly measured)
Formula:
(pre-committed spending − financial services) / gross disposable income
Where to read it:
detailed pre-committed spending, value and price-index sheets, rows “pre-committed spending” and “financial services”
Revision:
final accounts up to 2023, semi-final in 2024, provisional in 2025
Not comparable with:
the financial services counted by INSEE include an imputed interest margin, which does not leave an account like a bill
Sources:
INSEE, 2025 annual national accounts, pre-committed spending by item
- Removing all financial services also removes bank fees actually paid: it is a bound, not an exact measure of the “paid” share.
France ÉcoScan calculationhigh confidencedemonstrated in a reserved chapter
A large part of the spending committed in advance is a rent nobody pays: the one owners are deemed to pay themselves.
Value:
reserved
Unit:
% of gross disposable income
Period:
2025
Scope:
imputed rents of all dwellings occupied by their owners
Formula:
imputed rents / gross disposable income; imputed rents / pre-committed spending
Where to read it:
detailed pre-committed spending, row “imputed housing rents”
Revision:
final accounts up to 2023, semi-final in 2024, provisional in 2025
Not comparable with:
income contains this rent net of charges; spending contains it gross: the two do not subtract exactly
Sources:
INSEE, 2025 annual national accounts, pre-committed spending by item
- This rent is not false income: it measures the service the dwelling renders to its owner, and makes owners and tenants comparable.
France ÉcoScan calculationhigh confidencedemonstrated in a reserved chapter
The study splits a hundred euros of income between what is committed in advance, what is committed without being paid, other spending and what is not consumed.
Value:
reserved
Unit:
euros committed in advance excluding imputed rents, per €100 of income
Period:
2025
Scope:
all households
Formula:
each item / gross disposable income × 100; the sum equals 100
Where to read it:
detailed pre-committed spending; table 2.104, individual consumption expenditure
Revision:
final accounts up to 2023, semi-final in 2024, provisional in 2025
Not comparable with:
specified in the reserved chapter
Sources:
INSEE, 2025 annual national accounts, pre-committed spending by item · INSEE, 2025 annual national accounts, table 2.104 (household account and saving)
- An average of all households: no household lives this split.
- Insurance is counted net of the claims paid out: what households pay in premiums is higher.
France ÉcoScan calculationhigh confidencedemonstrated in a reserved chapter
Depending on whether it is measured per person, per consumption unit or per household, the same national income does not tell the same growth.
Value:
reserved
Unit:
% cumulative change
Period:
2008-2025
Scope:
all households; three ways of dividing the same income
Formula:
product of (1 + published annual change), from the year after the start to the end year
Where to read it:
table 2.101, rows per person, per household and per consumption unit
Revision:
final accounts up to 2023, semi-final in 2024, provisional in 2025
Not comparable with:
three different quantities: the consumption unit measures living standards; the household measures what a home receives, not what each of its members lives on
Sources:
INSEE, 2025 annual national accounts, table 2.101 (income and purchasing power)
- The consumption unit takes account of a household's economies of scale; it is the measure INSEE considers the most relevant.
- That this gap changes what households feel is measured by no source: it is a possible explanation, not a finding.
Official datahigh confidence
A household had on average 3.08 people in 1968 and 2.14 in 2023; 38.9% of homes are occupied by a single person.
Value:
2.14
Unit:
ratio: people per main residence
Period:
1968-2023
Scope:
household population, censuses
Formula:
published by INSEE
Where to read it:
figures 1 and 2
Revision:
final
Not comparable with:
the census of year N is an average of five annual surveys
Sources:
INSEE, household size, censuses from 1968 to 2023
- Fewer people per household means more households to equip and heat for the same national income.
France ÉcoScan calculationhigh confidencedemonstrated in a reserved chapter
The study measures the price rise since the start of the shock, with and without the rent owners pay themselves.
Value:
reserved
Unit:
% cumulative rise
Period:
January 2021 to August 2026
Scope:
all households
Formula:
ratio of final monthly indices
Where to read it:
series 011814630 and 011814633
Revision:
final; the following month is provisional
Not comparable with:
the price index counts neither loan repayments nor, in its usual version, imputed rents
Sources:
INSEE, monthly consumer price indices, 2025 base
- A cumulative price rise is not a loss: incomes rose too.
France ÉcoScan calculationhigh confidencedemonstrated in a reserved chapter
The study measures average purchasing power since the start of the inflation shock, per consumption unit and per household, year by year.
Value:
reserved
Unit:
% cumulative change
Period:
2021-2025
Scope:
all households
Formula:
product of (1 + published annual change)
Where to read it:
table 2.101
Revision:
final accounts up to 2023, semi-final in 2024, provisional in 2025
Not comparable with:
an average, which adds up winners and losers
Sources:
INSEE, 2025 annual national accounts, table 2.101 (income and purchasing power)
- The year 2021, a post-pandemic rebound, is left out of the cumulative figure: it would inflate growth without saying anything about the shock.
Official datahigh confidencedemonstrated in a reserved chapter
In 2025, what remains after spending committed in advance fell faster than income itself.
Value:
reserved
Unit:
% annual change
Period:
2025
Scope:
all households, per consumption unit
Formula:
published by INSEE; discretionary income is deflated by the price of non-committed spending
Where to read it:
table 2.102, row “purchasing power of discretionary income per consumption unit”
Revision:
provisional, revisable in May 2027
Not comparable with:
two different deflators: the comparison is between two measures, not two parts of one income
Sources:
INSEE, 2025 annual national accounts, table 2.102 (discretionary income)
- Over the long run, the two measures move almost in step: the 2025 gap is not a trend.
Official datahigh confidencedemonstrated in a reserved chapter
The 2025 fall and the one expected in 2026 do not have the same origin, and do not hit the same households.
Value:
reserved
Unit:
% annual change
Period:
2025
Scope:
all households, excluding property income
Formula:
published by INSEE
Where to read it:
September 2026 economic outlook, section on household purchasing power, and the focus on dividends on pages 49-50
Revision:
September 2026 estimate
Not comparable with:
three measures for all households, not per consumption unit; the measure excluding property income does not replace the purchasing power of disposable income
Sources:
INSEE, 2025 annual national accounts, table 2.101 (income and purchasing power) · Economic outlook (Note de conjoncture) — September 2026
- Dividends go mostly to households that hold shares: their after-effect says nothing about the end of the month of the others.
- The gap between the overall measure and the measure excluding property income comes from all property income — interest, rents received, dividends —; dividends are only part of it.
Institutional projectionmedium confidencedemonstrated in a reserved chapter
The 2026 quarterly accounts and INSEE's forecast point to a second fall, of a different kind from that of 2025.
Value:
reserved
Unit:
% quarter on quarter
Period:
first and second quarters of 2026
Scope:
all households
Formula:
published by INSEE
Where to read it:
Informations rapides No. 212, page 3, table on gross disposable income and ratios; September 2026 economic outlook, page 69, figure 8
Revision:
provisional quarters; detailed third-quarter accounts on 27 November 2026
Not comparable with:
specified in the reserved chapter
Sources:
Informations rapides No. 212 — quarterly national accounts, detailed results for the second quarter of 2026 · Economic outlook (Note de conjoncture) — September 2026
- A forecast is not a result: the 2026 annual accounts will be released at the end of May 2027.
France ÉcoScan calculationhigh confidencedemonstrated in a reserved chapter
The study compares the minimum wage and the basic wage with prices since the start of the shock: they did not follow the same path.
Value:
reserved
Unit:
% cumulative purchasing power
Period:
January 2021 to August 2026
Scope:
gross hourly minimum wage; basic monthly wage in firms with 10 or more employees, excluding bonuses
Formula:
(1 + wage rise) / (1 + price rise) − 1
Where to read it:
minimum wage: series 000822484 and prices excluding tobacco 011814056 (monthly path of the “minimum wage purchasing power” study); basic wage: series 010562695 and price index 011814630 as a quarterly average
Revision:
final
Not comparable with:
a statutory hourly rate and an average basic wage; neither is a household's income
Sources:
DARES and INSEE, basic monthly wage (ACEMO survey) · INSEE, monthly consumer price indices, 2025 base
- The basic wage excludes bonuses and overtime, which may have offset part of the gap.
Official datamedium confidencedemonstrated in a reserved chapter
Since spring 2026, rising prices go first through energy and fuels again.
Value:
reserved
Unit:
% year on year
Period:
August to October 2026
Scope:
energy price index; average diesel price reported by operators
Formula:
published
Where to read it:
Informations rapides No. 242, page 1, Energy row; Informations rapides No. 218, diesel row; weekly DGEC survey, page 6
Revision:
September provisional; weekly readings
Not comparable with:
a pump price in one week is neither an annual average nor a household's spending; the gap between two readings by the energy directorate is not the year-on-year change in the price index, which measures something else
Sources:
Informations rapides No. 242 — consumer price index, provisional results for September 2026 · Informations rapides No. 218 — consumer price index, final results for August 2026 · Prices and margins of petroleum products in France and the European Union — note of 2 October 2026
- Fuels are not part of spending committed in advance as defined by INSEE: the 2026 shock goes first through the “free” spending of households that drive.
Official datahigh confidencedemonstrated in a reserved chapter
The share of income spent on housing varies threefold depending on whether one is an outright owner, a home buyer or a tenant.
Value:
reserved
Unit:
% of income, median, net of housing benefits
Period:
2023 incomes and spending
Scope:
households in ordinary dwellings, excluding those housed free of charge
Formula:
published by INSEE
Where to read it:
figure 3, “All” column
Revision:
final
Not comparable with:
2023 medians; the previous edition published averages, and the two are not comparable
Sources:
INSEE, France, portrait social 2025, Housing chapter
- The spending counted includes home buyers' loan repayments, property tax, charges, water, energy and home insurance.
France ÉcoScan calculationhigh confidencedemonstrated in a reserved chapter
The gap between tenures does not come from income alone: it persists at equal living standard — and income, at equal tenure, separates just as much.
Value:
reserved
Unit:
percentage points
Period:
2023
Scope:
twelve comparisons: three tenures against outright owners, in each of the four quartiles
Formula:
smallest of the gaps (median rate of the tenure − median rate of outright owners), quartile by quartile
Where to read it:
figure 3
Revision:
final
Not comparable with:
gaps between medians, not between matched households
Sources:
INSEE, France, portrait social 2025, Housing chapter
- Age hides behind tenure: outright owners are mostly older, private tenants mostly young.
Official datahigh confidencedemonstrated in a reserved chapter
Tenure follows age: comparing tenures partly means comparing generations.
Value:
reserved
Unit:
% of households in the age group
Period:
2020
Scope:
households in ordinary dwellings, by age of the reference person
Formula:
published by INSEE
Where to read it:
figure 1 (source: 2020 Housing survey)
Revision:
final; the 2024 Housing survey has not yet been exploited
Not comparable with:
a 2020 distribution, while the housing cost ratios are from 2023
Sources:
INSEE, France, portrait social 2025, Housing chapter
- A gap between tenures therefore mixes the effect of housing with that of age; none of the sources used separates them.
Calculation under stated assumptionsqualified confidencedemonstrated in a reserved chapter
A calculated case shows what the gap between tenures means, every month, for the same family at equal income.
Value:
reserved
Unit:
euros per month
Period:
2024 living standard, 2023 housing cost ratios
Scope:
illustrative household: 2.1 consumption units, median living standard, median housing cost ratio of the tenure between the 2nd and 3rd quartiles
Formula:
median living standard / 12 × 2.1; × median housing cost ratio of the tenure
Where to read it:
median living standard: supplementary table 2; housing cost ratio: figure 3
Revision:
final
Not comparable with:
a calculated order of magnitude, not an observed household: the median of a rate is not the rate of the median household
Sources:
INSEE, Insee Première No. 2117, living standards and poverty in 2024 · INSEE, France, portrait social 2025, Housing chapter
- For the home buyer, part of the amount repays principal: it is a real cash outflow AND saving.
- The amounts also include energy, water, charges, property tax and home insurance, not just the rent or the loan payment.
Calculation under stated assumptionsmedium confidencedemonstrated in a reserved chapter
On a typical recent loan, a large part of the monthly payment repays principal, that is, builds wealth.
Value:
reserved
Unit:
% of the monthly payment in the first year
Period:
2026 conditions
Scope:
amortising loan with constant payments, excluding insurance
Formula:
constant payment of an amortising loan (engine of the “becoming a homeowner” study)
Where to read it:
Overview of housing finance, page 14, chart G11; Stat Info on loans to individuals, page 1
Revision:
July 2026 rate provisional
Not comparable with:
a typical NEW loan, not the situation of established home buyers, whose loans are older and often cheaper
Sources:
Overview of housing finance — April 2026 · Stat Info — loans to individuals, July 2026
- No public source quantifies the principal repaid by all home buyers.
- A loan built on three published averages — amount, rate, term — is nobody's loan.
France ÉcoScan calculationhigh confidencedemonstrated in a reserved chapter
The price rise experienced since 2020 depends more on age and housing tenure than on income level.
Value:
reserved
Unit:
percentage points of difference in cumulative rise
Period:
2020-2025
Scope:
price indices by household category: same prices, different baskets
Formula:
ratio of published annual indices, 2025 over 2020
Where to read it:
series 011818266 to 011818295
Revision:
final
Not comparable with:
an index gap is neither a gap in euro spending nor a weight in the budget
Sources:
INSEE, consumer price indices by household category, 2025 base
- The indices apply the same prices to everyone: they ignore that a low-income household sometimes pays more, or less, for the same product.
- The gap between tenths depends on the window: it widens during the surge in food and energy prices, then closes.
Official datamedium confidencedemonstrated in a reserved chapter
In euros, the price shock cost well-off households more; as a share of living standard, it weighed far more heavily on low-income households.
Value:
reserved
Unit:
% of living standard
Period:
2022 and 2023
Scope:
households in ordinary dwellings; INSEE microsimulation
Formula:
published by INSEE
Where to read it:
2024 edition, figures 1a and 1b, column “All additional spending”; 2023 edition, figure 5b, column “Additional spending”
Revision:
model estimate, published November 2024
Not comparable with:
with unchanged consumption: the calculation does not account for the purchases households cut back
Sources:
INSEE, France, portrait social 2024, study of 2023 inflation and living standards · INSEE, France, portrait social 2023, study of 2022 inflation and living standards
- The weight of the shock follows first the share of income that is consumed: a household that saves little spends almost all it earns on goods whose price is rising.
- A model calculation, which imputes each household's consumption from the 2017 Household Budget Survey; there is no published equivalent for 2026.
Official datamedium confidencedemonstrated in a reserved chapter
The offset of the shock was the same neither at the bottom and top of the scale, nor from one year to the next.
Value:
reserved
Unit:
% of the extra cost offset
Period:
2022 and 2023
Scope:
same microsimulation
Formula:
published by INSEE
Where to read it:
2023 edition, figure 5b, last column; 2024 edition, figure 3, “Balance” column
Revision:
model estimate
Not comparable with:
specified in the reserved chapter
Sources:
INSEE, France, portrait social 2023, study of 2022 inflation and living standards · INSEE, France, portrait social 2024, study of 2023 inflation and living standards
Official datamedium confidencedemonstrated in a reserved chapter
Depending on family make-up and age, rising incomes offset the 2023 shock very unevenly.
Value:
reserved
Unit:
% of the extra cost offset
Period:
2023
Scope:
households in ordinary dwellings, by household type and age of the reference person
Formula:
published by INSEE
Where to read it:
figure 4, “Balance” column
Revision:
model estimate, published November 2024
Not comparable with:
a share offset, not a living standard: it says who caught up with the shock, not who lives best
Sources:
INSEE, France, portrait social 2024, study of 2023 inflation and living standards
- A single year of microsimulation, and one characteristic at a time: the source does not separate the effect of household type from that of income or housing.
Official datamedium confidencedemonstrated in a reserved chapter
Outside cities, the car made the price shock weigh more heavily.
Value:
reserved
Unit:
% of living standard
Period:
2022
Scope:
people in ordinary dwellings, by size of urban unit
Formula:
published by INSEE
Where to read it:
figure 2b, columns “Fuels” and “All additional spending”
Revision:
model estimate, published November 2023
Not comparable with:
the extra cost before any offset; it does not say what remained once incomes had been raised
Sources:
INSEE, France, portrait social 2023, study of 2022 inflation and living standards
- The 2024 edition does not publish this breakdown for 2023; in 2026, the shock goes through fuels again, but no equivalent measurement is published.
Official datamedium confidencedemonstrated in a reserved chapter
A slightly rising average can conceal a clear fall for the poorest fifth.
Value:
reserved
Unit:
% annual change
Period:
2023
Scope:
national accounts distributed by fifth of living standard
Formula:
published by INSEE
Where to read it:
figure 4b, rows “Purchasing power of adjusted disposable income per CU” and “Purchasing power of gross disposable income per consumption unit”
Revision:
semi-final account
Not comparable with:
the 2024 edition of the distributed accounts is not published; 2023 is the latest year
Sources:
INSEE, Insee Analyses No. 119, household account by category in 2023
- Splitting an aggregate between fifths relies on surveys and imputations; its small gaps are not significant.
- The 2023 fall follows the end of the exceptional aid of 2022, which had mainly supported the bottom of the scale (see the offset of the shock).
Official datahigh confidencedemonstrated in a reserved chapter
Monetary poverty remains at its highest level since the mid-1990s — a relative measure, which does not say the poorest are getting poorer.
Value:
reserved
Unit:
% of people, threshold at 60% of median living standard
Period:
2024
Scope:
people in ordinary dwellings
Formula:
published by INSEE
Where to read it:
supplementary table 3
Revision:
final
Not comparable with:
RELATIVE poverty: it rises when bottom incomes diverge from the median, even if their purchasing power does not fall
Sources:
INSEE, Insee Première No. 2117, living standards and poverty in 2024
- The scope excludes homeless people and those living in institutions.
- The series is linked across several changes in survey method: a gap of a few tenths between distant years is not significant.
Official datamedium confidencedemonstrated in a reserved chapter
A high national saving rate covers opposite situations depending on living standard.
Value:
reserved
Unit:
% of disposable income
Period:
2025; fifths: 2022
Scope:
gross saving for the aggregate; net saving for the fifths
Formula:
published; 2009-2019 average recomputed
Where to read it:
table 2.104, row “household saving rate”; Insee Focus No. 338, figure 2
Revision:
final accounts up to 2023, semi-final in 2024, provisional in 2025
Not comparable with:
gross and net saving are not comparable; negative saving of the bottom fifth is not debt
Sources:
INSEE, 2025 annual national accounts, table 2.104 (household account and saving) · INSEE, Insee Focus No. 338, consumption and saving by household category in 2022
- The consumption of low-income households is imputed in these accounts: the bottom-fifth figure is fragile, its sign less so.
- The overall rate is a 2025 gross rate; those of the fifths are 2022 net rates: two measures side by side, not one figure split up.
Reported perceptionmedium confidencedemonstrated in a reserved chapter
The inflation the French feel remains far above measured inflation, even when the latter falls back.
Value:
reserved
Unit:
% over twelve months, median of answers
Period:
November-December 2025
Scope:
more than 4,000 people aged 18 and over
Formula:
published by the Banque de France
Where to read it:
Bulletin No. 265/4, page 3, table of survey waves, median and mean rows
Revision:
closed survey wave
Not comparable with:
the value retained is the median: half the answers are above it; the mean, pulled up by a few very high answers, is given alongside. The reference measure is the European harmonised price index
Sources:
Banque de France Bulletin No. 265/4 — the French and inflation: the after-effects of the 2022-2023 inflation episode
- A perception is not a mistake to correct: it says what people remember — prices that stayed high, rises more striking than falls.
Reported perceptionmedium confidencedemonstrated in a reserved chapter
A large majority of French people underestimate the rise in their own income since 2021: it is the best-documented cause of the gap.
Value:
reserved
Unit:
% of respondents
Period:
2021-2025
Scope:
income reported by respondents; gross disposable income of all households, divided by the population
Formula:
product of (1 + change in disposable income) / (1 + change in population), from 2022 to 2025
Where to read it:
Bulletin No. 265/4, page 8; table 2.101, rows “gross disposable income” and “population”
Revision:
closed survey wave; provisional accounts for 2025
Not comparable with:
specified in the reserved chapter
Sources:
Banque de France Bulletin No. 265/4 — the French and inflation: the after-effects of the 2022-2023 inflation episode · INSEE, 2025 annual national accounts, table 2.101 (income and purchasing power)
- The Banque de France notes the same gap in surveys in other countries: it is not a French peculiarity.
- The lower the income, the less the rise is perceived, according to the same survey: the underestimation is not randomly spread.
Reported perceptionmedium confidencedemonstrated in a reserved chapter
Set against income actually received, perceived income is more often underestimated than overestimated.
Value:
reserved
Unit:
% of people
Period:
2023 to 2024 incomes, early-2025 survey
Scope:
people surveyed by the Statistics on income and living conditions survey, whose income is known from other sources
Formula:
published by INSEE
Where to read it:
September 2026 economic outlook, page 52, focus on the perception of income
Revision:
study published in September 2026
Not comparable with:
a single year of change, from 2023 to 2024; it is not the period of the 2022 shock
Sources:
Economic outlook (Note de conjoncture) — September 2026
- According to INSEE, underestimation is more frequent among well-off households and retirees; according to the Banque de France, it is low incomes that perceive the rise least. The two surveys do not ask the same question.
- The question can be understood as an annual average or from one start of year to the next; INSEE judges the two readings close.
France ÉcoScan calculationmedium confidencedemonstrated in a reserved chapter
In 2026, firms and households share the energy shock — but not in equal parts.
Value:
reserved
Unit:
% of value added
Period:
the fourth quarter of 2025 to the second quarter of 2026
Scope:
non-financial corporations
Formula:
gross operating surplus / value added, recomputed; it matches the published value; average of the twenty quarters from 2015 to 2019
Where to read it:
series 011794733 (gross operating surplus) and 011794722 (value added); September 2026 economic outlook, page 17 (“largely borne by them”) and page 20 (margin rate)
Revision:
provisional, revisable at each quarterly release
Not comparable with:
an aggregate ratio; it says nothing about any particular firm or sector
Sources:
INSEE, quarterly national accounts, 2020 base · Informations rapides No. 212 — quarterly national accounts, detailed results for the second quarter of 2026 · Economic outlook (Note de conjoncture) — September 2026
- The firms → households channel is described, not quantified: no source says what share of the weakness of real wages comes from margins.
- The 2015 to 2019 average is a benchmark chosen by the study — the pre-pandemic period —, not a norm.
Conditional scenarioqualified confidencedemonstrated in a reserved chapter
The study calculates what the real wage would become under the energy scenarios published by the Banque de France.
Value:
reserved
Unit:
% real wage
Period:
2026-2027
Scope:
average wage per head against the harmonised price index
Formula:
(1 + wage in the central projection) / (1 + prices in the scenario) − 1
Where to read it:
September 2026 projections, page 6, table 2, and page 5, table of scenarios
Revision:
projections of 15 September 2026
Not comparable with:
the Banque de France does not publish wages by scenario: assuming they do not react gives a bound for the employee, not a forecast
Sources:
Macroeconomic projections for France — September 2026 (interim projections of 15 September 2026)
- A scenario is never an announcement: it says what would happen IF.
- It is a bound, not a path: the minimum wage is indexed to prices, and wages follow inflation with a lag.
Conditional scenarioqualified confidencedemonstrated in a reserved chapter
The study translates oil prices into inflation using the rule INSEE publishes, without turning it into a forecast.
Value:
reserved
Unit:
points of inflation
Period:
price in the week of 2 October 2026
Scope:
direct and quick effect on the overall index; neither gas nor second-round effects
Formula:
0.1 × (price − 90) / 5
Where to read it:
March 2026 economic outlook, footnote on Brent; weekly DGEC survey, page 2
Revision:
March 2026 rule; weekly price
Not comparable with:
a one-week price applied to an order-of-magnitude rule: it is not 2026 inflation
Sources:
Economic outlook (Note de conjoncture) — March 2026 · Economic outlook (Note de conjoncture) — September 2026 · Prices and margins of petroleum products in France and the European Union — note of 2 October 2026
- A Gulf crisis reaches France through the world oil price; the share of its imports passing through it is published nowhere.
- A minimum: the rule contains neither refining margins, which drove the price of diesel in 2026, nor gas, which goes through other markets.
- The 90-dollar convention and the September projections are already overtaken by the observed price: their figures are dated, not void.
Conditional scenarioqualified confidencedemonstrated in a reserved chapter
The study calculates what a rise in loan rates would change to the monthly payment of a household borrowing today.
Value:
reserved
Unit:
euros per month
Period:
2026 conditions
Scope:
average 2026 loan, excluding insurance
Formula:
constant payment at the raised rate, minus the payment at the observed rate
Where to read it:
same documents as the average loan
Revision:
conditional calculation
Not comparable with:
specified in the reserved chapter
Sources:
Overview of housing finance — April 2026 · Stat Info — loans to individuals, July 2026
- How fast government bond rates pass through to loans is not published: this calculation says neither whether nor when the rise would come.
Official datamedium confidencedemonstrated in a reserved chapter
The rates at which the State borrows rose sharply in 2026; the channel to household credit is described, not quantified.
Value:
reserved
Unit:
%
Period:
early September 2026
Scope:
10-year sovereign bonds
Formula:
cited by INSEE
Where to read it:
September 2026 economic outlook, page 7, section on bond yields
Revision:
market rates, moving
Not comparable with:
a market rate at one date; neither the average cost of State debt nor the rate of a mortgage
Sources:
Economic outlook (Note de conjoncture) — September 2026
- The gap with Germany converts into no amount per household: no published channel allows it.
Established absencehigh confidence
No official statistic measures, for all households, the money left in the account once constrained spending and the whole loan have been paid.
Value:
not published
Unit:
—
Period:
as of 6 October 2026
Scope:
national accounts, tax and social income surveys, income and living conditions statistics, Household Budget survey
Formula:
—
Where to read it:
documentary search of 6 October 2026
Revision:
to be reassessed at each edition
Not comparable with:
INSEE's discretionary income is not this quantity: it counts imputed rents and ignores principal repaid
Sources:
INSEE, 2025 annual national accounts, table 2.102 (discretionary income) · INSEE, France, portrait social 2025, Housing chapter
- That is why this study gives no average “money left to live on” in euros.
Established absencehigh confidence
No public source says what share of the energy imported by France passes through the Strait of Hormuz or the Red Sea.
Value:
not published
Unit:
—
Period:
as of 6 October 2026
Scope:
crude oil, refined products, liquefied natural gas
Formula:
—
Where to read it:
September 2026 economic outlook, page 6, section on world trade
Revision:
to be reassessed at each edition
Not comparable with:
the WORLD share passing through it (about 20% of oil and liquefied gas) is not the French share
Sources:
Economic outlook (Note de conjoncture) — September 2026
- The documented channel is the world price, which France pays like all importers.

Open to all

How is this study made?

The calculation, the starting hypotheses, the sentences the study forbids itself and those an automatic check prevents from appearing.

national accounts read as INSEE publishes them, without recalculating what it publishes; ratios and cumulative figures with a written formula; results found again by an independent second calculation from the archived documents; scenarios taken from the institutions that publish them.

Back to the answer

What would refute this answerwhat would bring the answer down, written before the drafting

If the share of income committed in advance — excluding financial services, whose measurement changed with interest rates — had gained at least 2 percentage points since 2015 or since 2021.
Consequence: the average constraint would again become an explanation of the current end of the month, and the study's answer would fall.
If, at comparable living standard, the median housing cost ratio of tenants and home buyers exceeded that of outright owners by less than 5 percentage points.
Consequence: housing would cease to be the main dividing line, and the chapter saying so would be wrong.
If, since 2008, the changes in purchasing power per household and per consumption unit differed by less than 1 percentage point.
Consequence: shrinking household size would not separate what a household gains from what the average per consumption unit gains, and the chapter on units would fall.
If, in 2022 as in 2023, rising incomes had offset the extra cost of inflation in at least as large a proportion for the poorest 10% as for the richest 10%.
Consequence: the shock would have been experienced in the same way everywhere, and the answer “not the same story” would fall.
If the 2026 national accounts, expected at the end of May 2027, showed purchasing power per consumption unit rising over the year.
Consequence: the reading of a second shock, after 2022, would fall for the year 2026; this criterion cannot be evaluated before that date.

The 9 starting hypothesesset before the calculation; the review says what remains of them

The share of income committed in advance has risen and keeps rising.
The constraint has become concentrated on some households rather than worsening on average.
Housing explains a decisive share of the gaps.
Imputed rents move the measurement away from the money actually available.
Principal repayment sharply reduces some households' cash while building wealth.
The 2021-2026 shock hit differently depending on spending patterns.
A high saving rate masks very different situations.
Shrinking household size changes how income gains are felt.
Firms' difficulties pass through to household incomes.

The 12 sentences this study forbids itself to writethe shortcuts a hurried reader would take

Never present a “money left to live on” of the French as an official statistic, nor give it a value in euros or as a percentage: the title of the study names a question, not an aggregate.
Never write or imply that the measurement of purchasing power is false or misleading: it is accurate for what it measures, and the study explains what it does not measure.
Never call imputed rents fictitious or false income: they measure the service a dwelling renders to its owner, and make it possible to compare owners and tenants.
Never present the repayment of loan principal as a loss of purchasing power: it is a real cash outflow AND a build-up of wealth.
Never infer from a doubling of the share of pre-committed spending a halving of room for manoeuvre: the share doubled from 1959 to 1993, and discretionary income nevertheless rose in volume.
Never present saving as having become impossible for the French: the aggregate saving rate is high, and it is its distribution — not its level — that is at issue.
Never express the borrowing gap between France and Germany, or the State's debt burden, as a quantified cost for each home: no published channel allows it.
Never present a price rise identical for all households, nor, conversely, that low-income households suffered a markedly higher price rise: the indices by living standard are close; it is the weight of the shock in the budget that differs.
Never write that strain on State debt mechanically causes inflation.
Never present a future closure of the Strait of Hormuz or the Red Sea as a fact, nor a scenario as a forecast.
Never use a reported opinion as evidence of an economic change: it proves what people feel, not what prices did.
Never compare a change per household, per person and per consumption unit as if they were the same quantity, nor a quarterly change with an annual change.

The 13 assertions that cannot appear on this pagean automatic check rejects them

an official “money left to live on” measured by INSEE
Why: no official statistic carries this name or this quantity
an average value of the “money left to live on” of the French
Why: household cash is measured by no source in aggregate
an accusation of lying or manipulation against official statistics
Why: the measurement is accurate for what it measures
owners who would receive a false income
Why: imputed rent measures a service rendered by the dwelling, not a fiction
a loan's principal treated as lost purchasing power
Why: it is saving in the national accounts sense
room for manoeuvre reduced by half
Why: a share that doubles does not halve what remains
French people who could no longer save
Why: the aggregate saving rate remains above its 2009-2019 average; it is its distribution that is at issue
a cost for each home derived from the borrowing gap with Germany
Why: no published channel links this gap to a household's spending
an identical inflation for all households
Why: the price rise experienced differs clearly depending on age and housing tenure
a markedly higher price rise for low-income households
Why: price indices by living standard remain close, even at the height of the shock; it is the weight of the shock in the budget, and what offset it, that differ
a mechanical link between debt strain and inflation
Why: no published mechanism establishes it
a future closure of the Strait of Hormuz announced as a fact
Why: a scenario is not a forecast
average purchasing power down since 2021
Why: per consumption unit, it is higher in 2025 than in 2021 according to the national accounts; a fall can only be written by naming its unit and its years

Open to all

Reading aids

Each term is also explained where it is used. This block is a reminder, never the only place where the word is defined.

Disposable income

What households have left once they have received their wages, self-employment income, property income and social benefits, and once they have paid their direct taxes and contributions.

What it is not: It is not net pay, nor what is left at the end of the month: rent and bills have not yet been taken out of it.

Purchasing power

The change in disposable income once the price rise is taken out: how much more or less the income can buy than before.

What it is not: It is not a living standard, nor an amount in euros: it is a change, as a percentage.

Consumption unit

A way of counting household members that takes account of what they share: 1 for the first adult, 0.5 for each other person aged 14 or over, 0.3 for each younger child. A couple with two young children counts 2.1 units.

What it is not: It is neither a person nor a household: it lies in between, and it is the measure INSEE considers most relevant.

Spending committed in advance (“pre-committed” as defined by INSEE)

Spending settled by a contract that is hard to renegotiate from one month to the next: housing, its water and energy, insurance (excluding life insurance), banking services, phone and internet subscriptions, television, school or work canteens.

What it is not: Neither groceries, nor fuel, nor loan repayments are part of it. And it includes a rent nobody pays: imputed rent.

Discretionary income

Disposable income minus spending committed in advance: what a household really keeps a choice over — food, clothing, transport, leisure, saving.

What it is not: It is not the money left in the account: a mortgage payment is not fully taken out of it.

Imputed rent

The rent an owner would pay if they rented the home they live in. Official statistics count it both in their income and in their spending, so that owners and tenants are measured the same way.

What it is not: It is not false income: it measures the service the home renders to its owner. But it goes through no bank account.

Housing cost ratio

The share of income spent on housing: rent or loan repayments, charges, property tax, water, energy and home insurance, with housing benefits deducted.

What it is not: It is not a borrower's debt-service ratio, which banks calculate on income before tax and without charges.

Home buyer

An owner who is still repaying the loan taken out to buy their home.

What it is not: It is not an outright owner: these two kinds of owner live very different ends of the month.

Living standard

A household's disposable income divided by its number of consumption units: it makes it possible to compare a single person and a family.

What it is not: It is not the wage, nor the income of the whole household.

Tenth, fifth, quarter

A way of ranking the population from the lowest living standard to the highest and cutting it into ten, five or four equal parts. “The poorest 10%” are the first tenth.

What it is not: They are not social categories: only income per consumption unit separates them.

Indirectly measured financial services

Official statistics consider that a bank renders a service when it pays less on your deposits than a reference rate: it counts this gap as household spending. When rates rise, the gap grows, and the “spending” with it.

What it is not: It is not a bill: no debit appears on a bank statement.

Percentage point

The gap between two percentages. Going from 29% to 31% means gaining 2 percentage points — not 2%.

What it is not: It is not a percentage rise: 2 points on a 29% share make a rise of nearly 7%.

Scenario

A calculation that describes what would happen if an assumption came true. Institutions publish several, from the most favourable to the most adverse, around their central forecast.

What it is not: It is not a forecast, still less an announcement.

Microsimulation

A calculation made household by household on a large representative sample: INSEE applies the year's price rise, benefits and taxes to each household's consumption and income, then adds them up.

What it is not: It is not an observation: each household's consumption is estimated from an older survey.

Harmonised price indexIPCH

The price measure calculated the same way in every euro area country; it is the one followed by the Banque de France and the European Central Bank.

What it is not: It is not INSEE's national price index, which weights some spending differently, such as health.

Firms' margin rate

The share of the wealth produced by firms that remains to them once wages and taxes on production have been paid — what serves to invest, repay and remunerate capital.

What it is not: It is not a profit: interest, corporate tax and the wear of equipment have not yet been taken out of it.

What this study contains

What does this study contain?

This list is generated by the study's contract: a surface removed disappears from it, and a surface added enters it.

Open to all

The question, the answer in one sentence and the criteria written before the text that would refute it
3 dated benchmarks of the current situation — purchasing power in the last closed year, the latest month's price rise, household confidence — each with its source and status
1 interactive example chart — purchasing power per consumption unit, year by year from 1960 to 2025
The three measures people confuse — purchasing power, discretionary income, the money left in the account — what each counts and what it leaves out
The 9 quantities official statistics do not provide — including the money left in the account at national level — with, for each, what its absence forbids concluding
2 common beliefs of the 7 the study sets against its frozen conclusions, each with a calculated verdict, not one written by hand
The 8 families of sources, each with its producer, its cut-off date and what it does not measure
The study's 41 quantified claims, each with its nature, period, scope, source and limits; the value of 6 of them
The method: the 12 shortcuts the study forbids itself, the 13 sentences an automatic check prevents from appearing, and the 9 starting hypotheses
The 16 reading aids, each defined where it is used, with what it is not
The reference to copy, the edition, the version and the data cut-off date

Reserved

A hundred euros of income, item by item: 9 items, what is committed in advance, what is committed without being paid — the rent nobody pays —, other spending and what is not consumed
The share of income committed in advance from 1959 to 2025, year by year: when it rose, when it stopped, what the measurement of banking services has added since 2023, and the same rise measured from several starting years
Housing at equal income: the housing cost ratio of 4 tenures in 4 living-standard bands, 4 calculated cases for the same family, and how much principal a mortgage payment repays
2021-2026, where the shock went: prices item by item, purchasing power year by year, discretionary income, the minimum wage and the basic wage against prices, and why 2025 and 2026 do not fall for the same reasons
Not the same story for all: the price rise experienced by income, age and housing tenure, the weight of the 2023 shock in 10 living-standard bands, in euros and as a share of income, what income rises offset in 2022 then in 2023 — single-parent families, young households and rural municipalities included —, poverty and saving by living standard
Why the feeling stays bad: the same income measured per person, per consumption unit and per household, smaller households, household confidence and the inflation they perceive
When firms' margins become households' margins: what the accounts show, what INSEE says about it, and what no source quantifies
The shocks that could change the end of the month, as conditional calculations: 3 Banque de France scenarios turned into real wages, 3 oil prices turned into inflation by INSEE's rule, 2 rate rises turned into monthly payments — each element classified: observed, documented mechanism, projection, scenario or not established
The assessment: the full answer, the 9 starting hypotheses and their calculated verdict, the 5 refutation criteria evaluated on the data, and the 5 other common beliefs

Open to all

Cite this study

The reference to copy, with the edition, the version and the data cut-off date.

France ÉcoScan v9.0.14, “What French households really have left”, edition 2026.1 — income, spending committed in advance and housing of households in France.

Edition:: 2026.1
Product version:: France ÉcoScan v9.0.14
Window covered:: 1959–2025 for the annual accounts, prices and confidence up to September 2026, data as of 6 October 2026
Data as of:: 6 October 2026