France’s retirement age is supposed to rise from 62 to 64 under a pension reform that President Emmanuel Macron argues is necessary to protect the financial sustainability of the country’s pension system. The policy itself has been debated for months. What has transformed the dispute into a much larger political crisis is how the government chose to enact it.
On 16 March, Prime Minister Élisabeth Borne invoked Article 49.3 of the French Constitution, allowing the pension legislation to proceed without a final vote in the National Assembly unless parliament removed the government through a no-confidence motion. On Monday, the principal motion came extraordinarily close to succeeding. It received 278 votes, only nine short of the 287 required. The government survived, the reform remained alive, and the political confrontation moved increasingly onto the streets.
By Thursday, enormous demonstrations were taking place across France. Trade unions organised strikes affecting transport, schools, refineries and other services, while protests in Paris and several other cities produced confrontations with police. The government insists that reform is necessary and constitutionally enacted. Opponents argue that using Article 49.3 to force through such an unpopular and consequential policy reveals a government unable to persuade either parliament or much of the public.
Both claims contain something important.
France has not ceased being a democracy because its government used Article 49.3. The provision is written into the Constitution of the Fifth Republic. Parliament retained a mechanism for stopping the government: remove it through a vote of no confidence. MPs attempted exactly that and failed.
The constitutional sequence therefore worked according to the rules.
Yet constitutional legality and democratic legitimacy are not identical concepts.
That distinction has appeared repeatedly throughout these posts. Ranil Wickremesinghe became Sri Lankan president through a constitutional parliamentary procedure while possessing weak direct popular legitimacy. Dina Boluarte became Peru’s president constitutionally while substantial parts of the country rejected the political process producing her government. Rishi Sunak became British prime minister entirely legally without a new general election.
France now provides another version.
Macron won an election.
His government possesses constitutional authority.
The reform can become law through procedures established by the constitution.
None of those facts automatically answers whether citizens consider the process sufficiently representative.
This is not an argument for replacing parliament with street demonstrations. If every government policy required approval from whichever side could mobilise the largest protest, democratic equality would disappear quickly. People with more time, organisational resources or geographical proximity to capitals would possess greater political influence than citizens who stayed home.
Elections solve that problem imperfectly by assigning every citizen one vote.
Protests measure something different.
They measure intensity.
A voter may oppose a pension reform mildly but still vote for Macron because they prefer his position on Europe, Ukraine, taxation or immigration. Another voter may care about pensions so intensely that they are willing to lose wages by striking for several days.
The election records both people through one ballot.
The strike reveals the second voter’s intensity.
A functioning democracy needs mechanisms capable of processing both kinds of information without confusing them.
France’s problem is that its parliamentary system currently struggles to convert either signal into compromise.
Macron was re-elected president in April 2022 but lost his absolute parliamentary majority two months later. His centrist coalition remains the largest bloc in the National Assembly but cannot reliably pass legislation alone.
That result contained political information.
French voters gave Macron another presidential term.
They did not give him unrestricted legislative control.
The constitutional system therefore produced divided authority.
Macron’s government has responded by negotiating some legislation with opposition parties and, on several occasions, using Article 49.3 where parliamentary support could not be guaranteed.
That is constitutionally available.
But repeated reliance on a mechanism designed to bypass a final parliamentary vote raises a democratic question precisely because the government’s lack of a majority is itself an electoral outcome.
If voters deliberately or accidentally produced a parliament in which compromise is necessary, using constitutional tools repeatedly to reduce the need for compromise can appear to circumvent the message contained in the parliamentary election.
Again, legality remains.
The political meaning changes.
The government’s strongest argument is that governing sometimes requires doing unpopular things.
That is true.
If democratic leadership meant implementing only policies receiving majority support in every opinion poll, long-term reform would become extremely difficult.
Pension systems illustrate why.
France, like much of Europe, is ageing. People live longer, and the ratio between workers contributing to pension systems and retirees receiving benefits changes over time. Pay-as-you-go pension arrangements therefore face increasing pressure unless governments adjust some combination of contribution rates, retirement ages, benefit levels, taxation or employment.
There is no political mechanism capable of repealing demographics.
Someone has to absorb the adjustment.
Raise the retirement age and workers bear more of it through additional years of employment.
Raise payroll taxes and workers and employers bear more through contributions.
Reduce pensions and retirees bear it through lower benefits.
Finance the system increasingly through general taxation and taxpayers absorb more.
Increase borrowing and future taxpayers inherit more debt.
Economic policy often becomes politically difficult because avoiding one cost simply moves the cost somewhere else.
This is why slogans such as “protect pensions” are incomplete.
Protect them from what?
Higher retirement ages?
Higher contributions?
Lower benefits?
Fiscal pressure?
Every design protects one dimension by adjusting another.
Macron’s government has therefore identified a real policy problem.
That matters because democratic legitimacy cannot simply mean every unpopular reform should be abandoned.
Governments are elected partly to make decisions citizens would prefer not to confront individually.
The harder question is whether this particular reform distributes the adjustment fairly and whether alternatives were genuinely considered.
France’s retirement debate contains strong distributional differences because the burden of working two additional years is not equal across occupations.
A lawyer working in an office at 63 experiences ageing differently from a construction worker performing physically demanding labour.
A worker who entered employment at 18 has already contributed for many more years by 62 than someone who completed postgraduate education and began professional work at 26.
Life expectancy itself varies by income and occupation.
A uniform retirement age can therefore be formally equal while producing unequal substantive burdens.
This is why pension policy cannot be evaluated only through the average retirement age.
Distribution matters.
The government’s reform includes provisions intended to address long careers and certain difficult working conditions, but opponents argue these protections remain inadequate.
That is an empirical policy dispute.
The larger principle is that equal rules can have unequal effects when citizens begin from different positions.
Economics often calls this incidence.
Who formally pays a tax is not necessarily who economically bears it.
Likewise, who formally receives the same retirement age does not necessarily experience the same cost.
This is why labour unions have become central to the opposition. They represent workers whose bargaining power individually would be relatively weak but collectively becomes much larger.
A single railway employee refusing to work has little influence.
Thousands striking simultaneously can disrupt national transport.
Collective action converts dispersed workers into a political actor.
This is exactly why strikes are controversial.
Supporters view them as necessary counterweights to concentrated economic and governmental power.
Critics point out that strikers impose costs on citizens who did not participate in the dispute.
Both are true.
A transport strike pressures government partly by inconveniencing passengers.
A refinery strike pressures government partly by reducing fuel availability.
The political leverage comes from the externality.
If a strike affected only the employer and employees involved, many public-sector strikes would possess much less bargaining power.
That creates another democratic legitimacy question. Why should workers in strategically important sectors possess more political leverage than workers whose absence does not disrupt national infrastructure?
There is no perfect answer.
The right to organise protects workers from employers and governments that otherwise possess much greater concentrated power.
But democratic societies also need rules preventing essential services from becoming permanently hostage to small organised groups.
Again, pluralism means balancing forms of power rather than pretending one is inherently legitimate.
France has a particularly strong protest tradition, and street mobilisation has repeatedly shaped policy under governments of both left and right. This can function as an informal check on executive power.
It can also encourage governments to believe that every reform will produce demonstrations regardless of its merits, reducing the informational value of protest.
If every government concludes protesters will always say no, government becomes less likely to listen when no actually contains important information.
That is another repeated-game problem.
Trust affects interpretation.
Macron’s personal governing style intensifies the issue. Critics have long accused him of exercising power in an excessively top-down or technocratic manner, while supporters argue that France’s political system often requires presidents willing to make difficult reforms despite entrenched opposition.
Technocracy itself contains a democratic tension.
Experts can identify policies that improve long-term fiscal sustainability.
Citizens possess values and distributional preferences that technical models cannot determine.
An actuary can estimate how changing retirement age affects pension finances.
The actuary cannot decide how society should distribute additional years of work between a cleaner, a teacher, a banker and a retiree.
That is politics.
Evidence can constrain the feasible options.
It cannot select among them without normative assumptions.
This is why “the reform is necessary” should always invite another question:
which part is necessary?
Some adjustment may be necessary.
This exact adjustment may not be.
Governments sometimes move too quickly from identifying a genuine problem to declaring their preferred solution inevitable.
That rhetorical move depoliticises a choice that remains political.
“There is no alternative” is attractive because it transforms opposition into ignorance.
There are usually alternatives.
They have costs.
The serious debate is over which cost society prefers.
The government should therefore make the counterfactual explicit. If retirement remains at 62, what taxes rise, what benefits change or what fiscal burden results? How large is the gap? Over what time horizon? Which assumptions drive the forecast?
Transparency improves democratic debate because it forces both sides to confront trade-offs.
Opposition parties should face the same standard.
Rejecting Macron’s reform is not a complete pension policy.
If the retirement age remains lower, how should the system be financed?
Democratic opposition needs alternatives, not merely vetoes.
This is where fragmented parliaments can become dysfunctional. Every opposition party receives electoral benefits from defeating the government’s proposal, while no opposition party individually bears responsibility for constructing a replacement.
The government therefore faces concentrated responsibility.
The opposition enjoys distributed veto power.
That can produce policy paralysis.
Article 49.3 exists partly to solve that problem.
The Fifth Republic was deliberately designed in 1958 to create stronger executive government after the instability of the Fourth Republic, when fragmented parliamentary coalitions repeatedly collapsed.
The constitutional architecture therefore reflects a historical judgement: France needed more governability.
Article 49.3 is one instrument through which the executive can overcome parliamentary obstruction while remaining theoretically accountable because the Assembly can respond by removing the government.
That design makes more sense when viewed through its institutional purpose.
The problem is that mechanisms created to increase governability can weaken deliberation if used too aggressively.
Every constitution balances decisiveness and constraint.
Too many veto points and government cannot act.
Too few and government stops needing consent.
France’s current crisis sits directly on that frontier.
The no-confidence vote demonstrates the architecture functioning exactly as designed. The government used Article 49.3. Opposition parties attempted to remove it. They came nine votes short.
From a constitutional perspective, the result is clear.
From a political perspective, a government surviving by nine votes after forcing through an enormously unpopular reform has received a much more ambiguous signal.
It has enough authority to continue.
It may not have enough legitimacy to govern comfortably.
That distinction affects policy implementation.
Laws do not operate only because parliament approves them.
Citizens, workers, employers, bureaucracies and institutions have to comply.
A policy implemented against intense opposition may generate strikes, administrative friction and future electoral reversal.
Political capital therefore affects the real cost of implementation.
A government can win the legal battle and make the policy more expensive politically than expected.
This is why persuasion matters even when persuasion is not constitutionally necessary.
Consent reduces enforcement costs.
That principle applies far beyond pensions.
Tax systems work partly because most citizens comply voluntarily.
Public-health rules work better when citizens trust authorities.
Environmental regulations work better when businesses believe rules are legitimate.
Governments possessing legal authority still benefit enormously from social cooperation.
Legitimacy is therefore not merely philosophical.
It is an input into state capacity.
This may explain why Macron gave a television interview this week defending the reform and saying he was prepared to accept political unpopularity in order to pursue what he considers necessary for the national interest.
There is something admirable in politicians accepting electoral cost for policies they genuinely believe are needed.
There is also something dangerous in leaders becoming convinced that public opposition demonstrates their courage rather than providing information.
Unpopularity is not proof of seriousness.
Sometimes governments are unpopular because they are making necessary decisions.
Sometimes they are unpopular because they made poor decisions.
The politician does not get to determine which interpretation is true simply by invoking leadership.
That is why independent evidence and eventual elections matter.
France’s next national electoral judgement remains years away.
In the meantime, parliament, protests, courts, unions, media and opinion polls provide intermediate feedback.
None alone possesses the authority of an election.
Together they tell the government something about the political environment in which its mandate operates.
The correct response is neither surrendering automatically nor ignoring them.
It is updating.
Could implementation improve?
Could protections for physically demanding occupations become stronger?
Could the government explain alternatives more transparently?
Could opposition parties offer a fiscally credible alternative?
Could parliament negotiate rather than simply survive confidence votes?
Those questions are less dramatic than “Macron versus the streets”.
They are how representative government should work.
France also demonstrates why democracy cannot be reduced to a binary classification. Nobody serious should describe France as authoritarian because its government used a constitutional provision.
The relevant question is about democratic quality within a democracy.
How much deliberation occurs?
How responsive are institutions?
How concentrated is executive power?
How meaningful is parliamentary scrutiny?
Can opposition change policy without overthrowing government?
These variables exist on continuums.
Democracies can become more or less representative without ceasing to be democracies altogether.
That analytical vocabulary is more useful than constantly asking whether every controversial action means democracy is “dying”.
Institutions can deteriorate incrementally.
They can improve incrementally too.
The protests themselves should be judged by the same standard. Most demonstrators are exercising legitimate rights to assembly and political opposition. Violence, vandalism or attacks on police do not become legitimate because the underlying cause is popular.
The state has a responsibility to distinguish violent individuals from peaceful demonstrators and use proportionate force.
Again, behaviour rather than political identity should determine enforcement.
This principle has become almost repetitive across Peru, Brazil, Georgia and France because it is foundational: democratic governments must protect opposition while policing unlawful conduct within it.
The harder the political conflict, the more important the distinction becomes.
Perhaps the deepest issue underneath France’s pension confrontation is temporal.
Democracies are very good at representing current preferences.
Pension systems involve obligations extending decades into the future.
Future taxpayers do not vote today.
People not yet born certainly do not.
Governments therefore sometimes need to protect long-term interests against current majorities.
This is one of the strongest arguments for fiscal institutions and long-term reform.
But politicians can abuse the language of future generations to justify almost any present sacrifice.
Again, evidence matters.
How large is the future pension problem?
Who bears it under different scenarios?
What assumptions about employment, productivity, immigration and life expectancy determine the forecast?
Intergenerational justice requires numbers, not slogans.
Young French workers have an interesting position in this debate because they may simultaneously oppose working longer and face the greatest long-term burden if the pension system becomes increasingly expensive.
Political preferences do not always align neatly with actuarial interests.
People value fairness, expectations and social rights alongside lifetime financial calculations.
That is another reminder that economic rationality contains more than maximising income.
Institutions themselves create expectations.
If citizens spend decades believing retirement at 62 forms part of the social contract, changing the rule imposes a loss relative to that expectation even if the previous arrangement eventually becomes fiscally difficult.
Governments therefore need credibility not only when making promises but when revising them.
The social contract cannot mean every policy remains permanent.
It should mean changes occur through procedures citizens consider sufficiently fair.
That brings us back to Article 49.3.
The government’s strongest defence is that the procedure is constitutional and parliament had the opportunity to remove it.
Its critics’ strongest argument is that a constitutional mechanism can still be used in a way that weakens the deliberative legitimacy of a major reform.
Both can be true.
The question is not whether France followed the constitution.
It did.
The question is whether constitutional minimums should always be treated as democratic maximums.
I do not think they should.
A constitution establishes what government is legally permitted to do.
Political legitimacy asks whether citizens continue believing the exercise of that authority deserves compliance.
Good government needs both.
Macron has probably won the constitutional battle over whether his government can continue pursuing pension reform.
The demonstrations suggest he has not won the political argument.
Those are different victories.
A durable reform may eventually require both.
