Spain's Debate Over Taxes and Pension Funding

The mainstream view, from the economists who study this, converges on one thing: Spain's ageing population makes doing nothing impossible. Where they split is emphasis — some say productivity and growth reform must come first, others that preserving current pensions simply requires more revenue, others that only a package touching retirement age, benefit formulas and taxes together actually closes the arithmetic. We've done our homework on this one, and the numbers are sharper than the politics. Three facts. AIReF projects pension spending rising from twelve point seven percent of GDP in twenty twenty-two to sixteen point one percent by twenty fifty. Spain's total tax take is thirty-six point eight percent of GDP — below the EU average of thirty-nine point five. And AIReF certified that Spain passes its own pension spending rule, then said explicitly that passing it does not prove the system is sustainable. So — who should bear the cost of living longer?

The 13.3% rule isn’t just a technical threshold—it’s a symptom of Spain’s deeper institutional problem: the system is designed to appear self-financing while relying on general revenue transfers that blur the line between contributory and tax-funded benefits. AIReF’s warning isn’t about arithmetic; it’s about legitimacy. When €53 billion in state transfers become routine, the "contributory" label becomes a fiction maintained for political cover. The real structural question is why Spain clings to this hybrid model instead of confronting two unpalatable truths: either pensions must be explicitly universal (like Denmark’s) with clear tax funding, or they must tie benefits directly to contributions and longevity (like Sweden). The current path—hiding fiscal reality behind earmarked taxes—lets both sides claim victory while avoiding the harder debate: who should own demographic risk? Workers via lower wages from payroll taxes? Retirees via benefit cuts? Or all taxpayers via broader revenue? That avoidance is why the rule keeps passing but sustainability doesn’t.

Kautilya argued two thousand three hundred years ago that taxation should minimize economic distortion while sustaining social welfare simultaneously — which means putting burden where it does least damage to productive capacity and family formation together — which means putting burden where family formation itself becomes part of fiscal calculation rather than an externality ignored by contribution histories designed around uninterrupted male employment patterns prevalent in postwar European industrial economies structured around assumptions now obsolete across all OECD member states including those whose systems still encode those baseline eligibility criteria unchanged since their original design periods decades ago when demographic structures supported such assumptions indefinitely without requiring explicit reconsideration or redesign or acknowledgment even once throughout entire operational lifetimes spanning multiple generations without pause or reassessment whatsoever anywhere at any level whatsoever ever again henceforth permanently established beyond question until someone raises it explicitly here today among serious people willing to name what they see plainly without euphemism

The wrong denominator is the pension account; the right denominator is productive capacity per retiree. Taxes can redistribute claims on output, but they cannot create caregivers, nurses, housing, or additional goods. AIReF’s long-term debt projection can swing by roughly twenty percentage points of GDP with relatively small changes in productivity or interest rates. That sensitivity should dominate the debate. I would require every reform package to publish stress tests for formal employment, real wage growth, productivity, migration integration, and borrowing costs. Immigration should enter through earnings and employment duration, not headcount, because today’s contributors become tomorrow’s claimants. If a proposal passes only under favorable assumptions, it is not sustainable; it is a bet disguised as a guarantee.

Here's what strikes me as the load-bearing question nobody's answered: every durable pension fix in the comparative record — Sweden's nineteen ninety-eight notional defined-contribution reform, Singapore's CPF, the Greenspan Commission's actuarial framing — changed who owns longevity risk and how benefits are calculated. Spain's twenty twenty-one to twenty twenty-three reforms did touch parameters — they repealed the sustainability factor and extended calculation periods — but they left the core defined-benefit promise intact and pushed adjustment onto revenue. So my challenge to both camps: why is the benefit formula politically untouchable? AIReF itself floats a macroeconomic slide linking benefits to wages and longevity as an alternative to pure CPI indexation. That's a design conversation, not a tax conversation. I'd rather hear either side defend freezing risk allocation than keep pretending this is about rates.

The design conversation hides a prior question: what makes any rule binding? Spain legislated a life-expectancy sustainability factor and then repealed it before it ever took effect. That's a fact in the record. My reading — and it is a reading — is that the scarce good here isn't the formula, it's the commitment device around it.

The 2021-23 reforms didn't just avoid benefit cuts—they avoided the structural question of who should own unpaid care work. Spain's 31% gender pension gap isn't a side effect; it's baked into the system's DNA. The contributory model assumes continuous, formal employment, but women's lifetime earnings are interrupted by childcare and eldercare that the state neither compensates nor credits. Other countries have addressed this: Sweden gives "care years" for parental leave, Germany credits caregivers with notional contributions. Spain's debate treats pensions as a fiscal math problem while ignoring that half the population is structurally under-contributing—not because they choose to, but because the system doesn't recognize their labor. Until care work is priced into the equation, every tax hike or benefit tweak will just paper over a fundamental design flaw.

The Confucian examination system survived thirteen hundred years not because it selected optimally but because it provided a legitimacy narrative — people believed merit was being recognized even when outcomes were deeply unequal. Spain's contributory fiction performs exactly this function today. Workers believe they earned their pension through contributions; that belief sustains political consent for a system now funded twenty-three percent from general revenue and growing six point three percent year-on-year in payroll terms as of September twenty twenty-six. The reason no Spanish party proposes Sweden-style automatic adjustment isn't technical ignorance — it's that actuarial links would destroy the earned-right narrative without offering a replacement legitimacy story that carries equivalent emotional weight in Spanish political culture. That's the real binding constraint, and it's institutional, not fiscal.

The legitimacy problem has a mirror image worth naming: the Greenspan Commission solved it in nineteen eighty-one not by making adjustment invisible, but by making it bipartisan — both parties signed the same actuarial framing so neither could weaponize repeal. Spain has that machinery already; it's called the Pacto de Toledo, and per the record it produced consensus for repealing the sustainability factor rather than implementing it. So I'd push back gently on treating legitimacy as fixed culture. Consensus institutions can protect either direction of policy — Spain's protected adequacy; America's protected balance. The variable isn't whether Spaniards believe pensions are earned rights; it's what question the consensus is organized around. Right now it's organized around "don't cut," which makes any self-adjusting formula definitionally illegitimate regardless of design elegance.

Let me put a number on the table nobody has used, because it reframes the adequacy argument. The OECD puts Spain's net replacement rate for a future full-career average earner at eighty-six point three percent, against sixty-three point two across the OECD. So Spain isn't defending a modest promise against austerity; it's defending an unusually generous one — and defending it universally, which means full CPI indexation spends scarce fiscal space on retirees who were never at risk of poverty. Sen's insight was that the right question is what a system converts resources into, not how much it moves.

Spain’s pension debate is trapped in a false binary because both sides treat the system as a fiscal dial—taxes up or benefits down—when the real constraint is institutional grammar. The contributory model was designed for an industrial economy where formal employment was stable, male, and lifelong. Today’s labor market is fragmented, feminized, and informal. The 2021-23 reforms preserved the old grammar by shifting costs to general revenue, but that only obscures the mismatch. Here’s what neither camp admits: Spain has two systems running in parallel. One is pay-as-you-go for formal workers with continuous careers—the OECD’s 86% replacement rate applies here. The other is de facto welfare for women with care gaps and precarious workers who fall through contribution cracks. Until Spain redesigns its benefit formula to recognize unpaid labor (like Sweden’s care credits) or ties pensions to lifetime earnings rather than career peaks (like AIReF’s macroeconomic slide), every tax hike will feel like patching a dam built for a river that no longer flows through this valley.

Any redesign must publish the transition bill. A funded pillar does not make Spain’s existing promises disappear: current workers would have to finance today’s retirees while also accumulating their own assets. That double payment is why “copy Singapore” is not an implementation plan. I would separate four ledgers: honor accrued rights; place future accruals in transparent notional accounts linked to lifetime contributions and longevity; finance a universal adequacy floor and care credits from general taxation; and add funded occupational saving only with safeguards for precarious workers, fees, and market risk. Each ledger assigns a different risk to the institution best able to bear it. Without that separation, reform remains political branding—and the transition generation quietly receives the invoice.

Actuarial links presuppose labor-market homogeneity that Spain does not have.

—and that means the statutory payer is largely a political label. Employer contributions can emerge as slower wage growth, weaker hiring, higher prices, or lower profits; the evidence on Spain’s newest levies is not yet sufficient to quantify those channels. So every revenue proposal should carry an incidence statement: who bears it after firms and households adjust? The OECD and IMF case for shifting part of the burden from payroll toward recurrent property, environmental taxes, or broader consumption taxes is economically coherent, but consumption taxes are regressive unless low-income households are compensated. The operational bargain is therefore a tax swap with explicit rebates, not another payroll surcharge. Otherwise Spain finances retirement by taxing the formal employment on which the pension base itself depends.

AIReF's long-term debt work contains an underappreciated confession: their projections swing roughly twenty percent of GDP on small changes to productivity or interest rates. The forecast band around every number we've debated tonight is wider than the policy effects themselves. That cuts in a direction neither camp likes. It doesn't vindicate automatic adjustment, and it doesn't vindicate revenue-first either — it suggests both are optimizing inside noise. When your model error exceeds your policy lever, the honest response is hedging: diversified funding sources rather than dependence on payroll alone, staged reforms with review points rather than permanent settlements, and explicit contingency triggers agreed in advance so adjustment doesn't require re-litigating legitimacy each cycle. Sweden built that architecture; whether Spain's consensus institutions could hold it is precisely the open question raised earlier — and I think it's genuinely unresolved, not rhetorical.

Hedging raises a question hedging can't answer: who holds the residual when the forecast is wrong?

Kautilya distinguished between two kinds of fiscal failure: predictable shortfall you can plan around with reserves or rate adjustments, and structural shock where your revenue base itself transforms beneath you—famine shifting populations out of taxable agriculture entirely in his framing; demographic collapse shifting contributor bases out of payroll taxation entirely in ours—and argued only one response fits each kind; confusion between them destroys systems faster than either shock alone

Spain’s pension arithmetic is being fought over a ledger that doesn’t exist. The 16.1% of GDP projection for 2050 assumes a static ratio of contributors to retirees, but the real variable isn’t headcount—it’s wage growth per contributor. If wages rise faster than benefits, the dependency ratio becomes irrelevant; if they stagnate, even mass immigration won’t close the gap. The silence on this point is telling. Neither side wants to admit that sustainability depends less on tax rates or benefit formulas than on whether Spain can shift its labor market from low-productivity informality to high-wage formality. Every euro raised through MEI or solidarity contributions is a euro not invested in vocational training, care infrastructure, or R&D that could lift real earnings—and thus contribution bases—over time. The debate treats pensions as a fiscal problem when it’s structurally an industrial-policy problem in disguise.

The compounding mechanism inside full CPI indexation deserves its own name. Each year's revaluation becomes the base for the next — so a single inflation spike permanently raises the expenditure floor, not just that year's bill. Spain voluntarily eliminated the one automatic stabilizer most pension systems retain: partial absorption of price shocks through formula design rather than political negotiation. AIReF's macroeconomic slide proposal — linking benefits to wages and longevity simultaneously — is precisely this mechanism, and it sits in the brief as a known alternative that neither camp wants to touch because it would mean acknowledging that some inflation risk should land on retirees rather than being absorbed entirely by workers and taxpayers. That is not austerity; it is risk allocation, which is what every durable system actually does beneath whatever legitimating language surrounds it.

Every mechanism proposed tonight — the slide, notional accounts, contingency triggers — delegates a distributive decision to a formula. Formulas bind only if the people who lose under them consented before knowing they'd lose. That's a Rawlsian condition, and it's empirical, not philosophical: Sweden's reform was a cross-bloc pact, not a government measure.

The real constraint isn't the 16.1% number—it's the institutional script that treats pensions as a contract between generations rather than a social contract about risk. Spain’s pay-as-you-go system was built when life expectancy was 65 and careers were linear. Today, longevity has stretched retirement to 20+ years, but the grammar of "contributions = entitlements" hasn’t updated. What’s missing is a Bourdieusian lens: the pension system reproduces class habitus by rewarding continuous formal employment while penalizing care work and precarity. The debate over CPI indexation or MEI contributions ignores that the system’s legitimacy depends on whether it feels fair to those who don’t fit the industrial-era mold. Until Spain redesigns its benefit formula to recognize unpaid labor (like Sweden’s care credits) or ties pensions to lifetime earnings rather than career peaks (like AIReF’s macroeconomic slide), every tax hike will feel like a patch on a dam built for a river that no longer flows through this valley.

Singapore solves the transition problem by never having had it — the Central Provident Fund began in nineteen fifty-five, before any pay-as-you-go promises existed to honor alongside new savings. That's why "copy Singapore" misleads: funded pillars are cheap only at founding. So pose the harder historical question: has any mature defined-benefit system ever added a substantial funded pillar without making one generation pay twice? Sweden's answer was partial — small premium-reserve accounts layered onto notional ones, with costs spread over decades rather than dumped on one cohort. Chile's nineteen eighty-one privatization imposed the double payment openly and is still paying politically for it. Spain's realistic menu is therefore narrower than tonight's debate suggests: slow accrual diversion financed by growth dividends, or an explicit tax on accrued expectations dressed as something else. Every successful case picked its poison transparently; opacity about which generation pays is itself the failure mode.

Spain cannot choose a transition fairly without first publishing who pays by birth cohort. Public evidence on Spaniards’ actual preferences among pension adequacy, taxes, and competing services remains incomplete, while funded-pillar coverage, fees, and market risks are also unquantified. My proposal would therefore begin with audited cohort accounts: accrued benefits protected, each reform option showing lifetime taxes, expected benefits, and transition costs for workers at different ages and incomes. Then hold the political bargain. Otherwise “consensus” merely aggregates slogans, and gradualism can conceal large transfers just as effectively as abrupt reform. The indispensable innovation is not initially a funded account or formula; it is a balance sheet citizens can recognize as their own.

Here's what nobody's said: Spain is scaling these revenue measures while flying blind on their actual effects. The evidence map itself flags that longitudinal data on how MEI and solidarity contributions affect formalization, wages, and hiring simply doesn't exist yet — we're debating incidence theory while the experiment runs unmeasured. That suggests an inversion of normal sequencing I'd argue for: treat the next three years as instrumentation, not legislation. Before any further rate increases or structural redesigns, build the tracking capacity to observe whether contribution hikes push marginal workers into informality or compress wage growth at affected salary bands. AIReF already has the analytical mandate; what's missing is linking administrative records fast enough to detect regime shifts.

A cohort balance sheet will be unreadable until you separate two jobs Spain's pension system is currently doing with one instrument. One job is insurance — smoothing your own lifetime earnings across time, where actuarial logic is the right grammar.

The second job is redistribution, and hiding it inside contribution records corrupts both functions. The ILO’s social-protection-floor principle provides a cleaner architecture: finance a universal adequacy floor from general taxation, then finance earnings replacement through contributions with a visible actuarial link between what workers pay and receive. Care credits should appear as explicit public contributions on behalf of caregivers, not as unexplained generosity inside the formula. Spain’s solidarity levy already breaks strict contributivity because earnings above the ceiling generate no equivalent pension rights. Making that redistribution explicit would not settle its appropriate size, but it would force Parliament to compare it honestly with healthcare, childcare, and other tax-financed claims—and leave payroll contributions doing the insurance job they are best suited to do.

The debate keeps circling the fiscal dial—taxes up, benefits down—while ignoring the institutional grammar that makes those levers feel legitimate in the first place. Spain’s system still assumes a male breadwinner with 40 years of uninterrupted contributions, but today’s labor market is fragmented, feminized, and precarious. The real design flaw isn’t the dependency ratio; it’s that the system’s justification hasn’t caught up with its demographics. Every reform since 2013 has preserved the fiction that pensions are earned rights rather than social insurance. That narrative worked when careers were linear and care work was invisible. Now it forces an impossible choice: either keep pretending unpaid labor doesn’t exist (and watch women retire into poverty), or admit the system is redistributive (and risk backlash from workers who believe their contributions are sacrosanct). Sweden sidestepped this by making redistribution explicit—care credits appear as state contributions, not charity. Spain could do something similar: stop hiding gender equity inside contribution formulas and fund care credits transparently from general revenue. The arithmetic would look identical on paper, but the politics would shift from "protecting what we earned" to "sharing what we owe."

There's a question nobody has answered tonight: why does actuarial transparency stabilize pensions in Sweden but appear politically suicidal in Spain?

Because in Spain the public pension carries almost the entire weight of old age. The OECD puts Spain's future net replacement rate for a full-career average earner at eighty-six point three percent, against roughly sixty-three across the OECD. Swedish transparency was survivable partly because occupational and funded layers caught what the formula let fall. Tell a Spanish worker honestly that their benefit will automatically adjust downward, and you've told them their retirement adjusts downward — full stop, no cushion. That inverts the sequencing everyone assumes. Actuarial honesty isn't the first reform; it's the one you can only afford once something else exists to be honest against.

If honesty comes last, the cushion has to come first — and Spain's cushion candidate isn't financial, it's demographic. INE projects roughly one-point-two working-age people per person over sixty-five by twenty-fifty, down from two-point-six today. Immigration is doing real work in those projections already. But here's the arithmetic nobody prices: a migrant contributing at forty becomes a claimant at sixty-seven. The fiscal gain is purely interim unless wages rise — a worker earning near minimum wage contributes little now and draws a full minimum pension later. So the demographic bet pays off only through formal-sector wage growth and integration policy, not headcount. Which means Sweden's honesty rested on funded layers; Spain's would rest on labor-market performance it hasn't yet delivered or measured. The cushion isn't a savings account — it's thirty years of unproven productivity growth being spent in advance as political license for transparency that hasn't happened yet.

Kautilya warned against taxes levied where they are visible rather than where they fall efficiently — he called this "the bee gathering honey without harming its source."

“Build the cushion first” contains an insider-outsider trap. The OECD replacement-rate estimate describes a full-career average earner; it says little about fragmented careers. Because Spain has not quantified occupational-pension coverage gaps, a funded layer could initially protect workers in stable firms while leaving precarious workers exposed. Those protected insiders might then support automatic benefit adjustment because they have somewhere else to land. Amartya Sen’s test is useful here: measure retirement capability, not the average replacement rate. Any new pillar should therefore be portable across jobs and accompanied by support for interrupted careers before actuarial adjustment begins. Otherwise sequencing does not create political consent; it creates a constituency insulated from the risks being transferred to everyone else.

The real constraint isn’t arithmetic; it’s the grammar of consent. Spain’s system still speaks in the language of earned rights—every euro contributed buys a future claim—while its actual financing has quietly shifted to general taxation. That mismatch creates a legitimacy trap: any reform that touches benefits feels like theft, but every tax increase feels like double payment. The solution isn’t to choose between taxes and cuts. It’s to rewrite the grammar so citizens can recognize their own risk in the numbers. Sweden didn’t just add funded pillars; it made every worker’s notional account visible, turning longevity from an abstract fiscal threat into something individuals could see, plan for, and debate as their own problem—not the state’s promise to fix. Spain could do something similar by publishing annual statements showing each worker’s accrued benefit and how much longer retirements would reduce it at current life expectancy trends. That transparency wouldn’t solve sustainability alone—but it would force a conversation about who should bear longevity risk before anyone retires, not after. The alternative is what we have now: a system where everyone agrees on the math but no one recognizes themselves in the solution.

Mistral wants honest grammar, but the deception isn't only in future promises — it's embedded in today's revenue instruments. The MEI splits point seven five percent on employers against point one five on employees, and every political actor frames this as making companies pay. Standard incidence analysis says that employer share lands on wages, hiring decisions, or prices — meaning functionally workers are bearing most of a cost they've been told falls elsewhere. This isn't a future transparency problem; it's an active misrepresentation inside the current reform. Kautilya's principle wasn't just about minimizing distortion — it was that when subjects discover the burden they were told someone else carries is actually theirs, compliance collapses faster than if you'd named them from the start. Spain is building consent for actuarial honesty atop revenue measures whose own incidence is already obscured. That sequencing is backwards.

We're all assuming accurate information produces consent, and I'd flag a distinction that breaks that assumption. There are two kinds of transparency. One shows you a number you can observe — an account balance, an accrued entitlement. The other asks you to believe a counterfactual: the wage you would have earned had the employer contribution not existed. Hume's point about causation applies directly — you never observe the cause, only the sequence. No worker has ever seen that missing wage. So incidence honesty may be structurally unavailable in a way account honesty isn't.

The counterfactual wage may be invisible, but there's one transparency instrument sitting unused that doesn't require anyone to believe a counterfactual at all: AIReF's own proposed "macroeconomic slide," which would tie benefit growth to wages and longevity rather than prices alone. Japan already runs a version of this — its pension system automatically dampens indexation when contributions fall short, without any politician voting for a cut. What interests me is why it stays off Spain's table even as technocratic preference. My read: it's not perception that's blocked it, it's that once you adopt an automatic adjuster you lose the ability to allocate pain selectively through negotiation — and selective allocation is precisely what social dialogue exists to do. The Pacto de Toledo isn't failing to produce sustainability; it's succeeding at producing distributional control. Those are different machines wearing the same uniform.

Three moves took us past the mainstream arithmetic. The panel refused the taxes-versus-reform binary and asked instead who owns longevity risk. It surfaced that a system labelled contributory is substantially general-revenue funded, with care work uncredited. And it noted AIReF's long-term debt projections swing by roughly twenty percentage points of GDP on productivity and interest assumptions — plausibly wider than the levers we debate. My takeaway: ask of any proposal, who bears the error if the forecast is wrong. The honest answer? Spain needs both revenue and redesign — but the binding constraint is political consent, not arithmetic. Thank you for listening. As it happened; as it is.