The Family Paradox — Why Germany Subsidises Marriage Instead of Families
Germany spends €25 billion a year on a tax instrument that requires no family as a condition. The Ehegattensplitting is not a design flaw. It is the design.
Germany spends €25 billion a year on a tax instrument that requires no family as a condition. The Ehegattensplitting is not a design flaw. It is the design.
The Tax Benefit That Asks No Questions
Germany spends roughly €25 billion a year on a tax instrument that requires no family as a condition.
The Ehegattensplitting — Germany's marital income-splitting regime — benefits married couples with a large income gap, regardless of whether children are in the household, whether any care work is being done, or whether either condition will ever apply. The benefit grows with the income differential: a single-earner couple where one partner earns around €60,000 gross and the other earns nothing saves roughly €4,600 a year through the splitting mechanism. At €100,000 gross with a non-working partner, the benefit exceeds €10,000. The theoretical maximum — €18,938 per year — is only accessible above a taxable income of €556,000, which affects a vanishingly small fraction of the population. But the principle holds at every income level: the benefit scales upward, uncapped, with no requirement for children and no requirement for care work of any kind.
What makes no difference: children. Elder care. Any form of dependent support. Two married adults without children and with a large income gap receive the full benefit. Two unmarried parents raising three children together receive nothing.
This is not a design flaw. It is the design.
On 25 March 2026, Lars Klingbeil — chairman of the SPD, Germany's Social Democrats — announced at a Bertelsmann Foundation event in Berlin that he wanted to abolish the splitting regime for new marriages. His words were unambiguous: "I want to abolish the Ehegattensplitting in its current form for future marriages." The direction is right. The argument is skewed. What should come next is entirely absent.
How the Splitting Mechanism Actually Works
The mechanics are straightforward. Both spouses' taxable incomes (zu versteuerndes Einkommen, or zvE — not gross pay, but income after social security contributions, standard deductions, and allowances) are added together, halved, taxed at the standard rate, and the result doubled. Because Germany's income tax schedule is progressive — higher incomes face higher marginal rates — a couple with unequal incomes pays less under joint assessment than under individual taxation. The difference is the splitting benefit (Splittingvorteil).
One distinction matters for understanding the numbers: zvE is substantially lower than gross pay. At a gross salary of around €60,000 — roughly the average for a male full-time employee in Germany by 2022 tax statistics — standard deductions produce a taxable income of around €48,000, about 20 per cent less than what the employer pays out. It is this lower figure on which the splitting calculation operates. A single-earner couple where the primary earner reaches a zvE of €48,000 saves roughly €4,600 a year. If the second partner earns €10,000, the benefit falls to around €1,800. If both partners earn equally, it is zero.
Two variables drive the benefit: total household income and the size of the gap between the two earners. The larger the gap and the higher the income, the stronger the progression effect. What drives it: children, care work, dependency — none of these.
The system is internally consistent. It does exactly what it was built to do: it treats the married couple with an income differential as a single tax unit. The question is why that effect is tied to marital status rather than to the family reality behind it.
Three Things the Splitting Fails to Do
The Non-Marriage Penalty
The splitting ties the tax benefit to marital status. Care work as such — raising children, caring for elderly relatives, running a household — does not appear in the calculation. What matters is: marriage. And income differential.
This produces a structural contradiction. Two unmarried parents raising children together receive no splitting benefit, regardless of how large their income gap is or how much care work they provide. Two married adults without children, with a substantial income gap, receive the full benefit. The instrument does not distinguish between these situations because it is not designed to. It checks marital status and income gap — nothing more.
A frequent objection is that around 91 per cent of the total splitting volume flows to couples with children. That is true. But it is a demographic correlation, not a design feature. Married couples statistically have children more often than unmarried couples — the splitting does not check for this. It requires no children. Where the benefit flows depends on marital status and income; who actually performs the care work is irrelevant to the calculation.
The same logic applies when a partner reduces paid work to care for an ageing parent or in-law. The splitting benefit exists where the marriage exists — not where the care is delivered. An unmarried adult in the identical care situation receives nothing. The instrument is indifferent to the care.
The Employment Effect
The splitting creates a structural incentive for the lower-earning partner to remain out of paid work — and an equally structural incentive not to change that.
The mechanism is direct: under income splitting, the second earner's income is taxed at the marginal rate that applies to the combined household income, not to their own. A partner whose spouse earns €60,000 gross faces approximately 30 per cent income tax on their first euro of own employment income. Add social security contributions, and OECD calculations place this combined burden at 55 per cent of gross earnings upon entering the workforce — for a household where the primary earner is around two-thirds of average wages. Only Belgium ranks higher in the OECD. The OECD names the Ehegattensplitting as a contributing cause.
The second earner is predominantly the woman. This is not coincidental: the splitting stabilises, in tax law, exactly the model it was historically designed to reflect — one full-time earner, one non-working or part-time partner. Every increase in the second partner's working hours reduces the splitting benefit incrementally.
This is not a political value judgement. It is the result of a simulation: Germany's Institute for Economic Research (DIW) has calculated that even a compromise reform — not a complete system change, merely a so-called Realsplitting with a transferable basic allowance — would increase married women's working hours by 1.7 per cent and their labour market participation by 0.6 percentage points. The effects of full abolition would be proportionally larger. The Federal Finance Ministry's own scientific advisory board reached the same conclusion in a 2018 report: the splitting "favours specialisation in marriage in the sense of one partner working and the other providing domestic services" and "thereby works against the politically desired goal of combining work and family life." That is not the opposition speaking. That is the ministry's in-house advisory body.
The Distribution Problem
Because the splitting benefit grows with income and with the income gap, its distribution is predictable: upper-income households receive the largest absolute tax advantage. According to DIW calculations — based on a simulation of individual taxation with a transferable basic allowance, not full abolition — the top three income deciles would each face an additional burden of €3–4 billion under reform, corresponding to the benefit they currently receive. The bulk of the €25 billion total flows to households that need it least as a transfer.
The necessary precision: most splitting recipients are not high earners. The splitting is not an instrument exclusively for the wealthy. It is an instrument whose value rises with income — and which reaches its highest value precisely where it is hardest to justify as family support. A child-based alternative instrument that achieves the same policy goal at lower cost would free up fiscal space for childcare infrastructure. A concrete proposal follows below.
Two Lines of Defence, and Why They Fall Short
Horizontal Tax Equity
The most serious objection is normative: two households with the same total income should bear the same tax burden. This claim of horizontal equity cannot simply be dismissed.
Germany's Federal Constitutional Court (Bundesverfassungsgericht, BVerfG) accepted it in 1982 (BVerfGE 61, 319, 3 November 1982). The splitting, the Court held, corresponded to the economic reality of marriage as an Erwerbs- und Verbrauchsgemeinschaft — an earning-and-spending community — in which both partners share income and burdens and the non-working partner makes an equivalent contribution. The splitting was therefore a constitutionally appropriate form of taxation — but, and this is the operative word in the ruling, "without prejudice to the legislature's more detailed discretion to design alternatives." The Court validated splitting as permissible, not as uniquely correct. The legislative space for alternatives was explicitly preserved.
What has changed since 1982: the dual-earner marriage is now the majority. The assumption that married couples typically pool their incomes and that the non-working partner makes an equivalent contribution describes the average marriage of 1982 — not of 2026. The horizontal equity argument rests on a premise that has shifted. The Constitutional Court has not updated that premise. It has, however, expressly authorised the legislature to do so.
The Solidarity Community Argument
The second line of defence is more cautious: marriage is a solidarity community with mutual legal obligations — maintenance, pension equalisation (Versorgungsausgleich), community of accrued gains (Zugewinngemeinschaft). Those who enter this commitment take on real burdens that justify special tax treatment.
This objection describes what the splitting does — not why it is the right instrument for doing it. A solidarity community can be reflected in tax law in multiple ways. France has done this since 1945 with the quotient familial: married couples file jointly, but the tax advantage beyond a basic two-person assessment comes from children, not from the income gap. Each child adds a further share to the denominator of the tax calculation; the additional benefit grows with the number of children. A childless married couple receives no splitting bonus beyond a simple two-person assessment. The instrument matches the declared policy: it benefits children — not marital status as such. France's birth rate of around 1.6 is substantially higher than Germany's (1.36); the quotient familial is not the only factor, but the instrument logic is a different one.
That the splitting reflects solidarity community is a description. Whether it benefits the performance it claims to reward is a separate question. An instrument that checks marital status rather than the care itself is poorly calibrated — regardless of whether the normative goal is justified.
The Test Case the Argument Cannot Handle
A few days after Klingbeil's announcement, a comment appeared on a social network from a young doctor that framed the core problem more precisely than many an editorial: "Marriage is a legally binding contract with substantial financial obligations: maintenance, pension equalisation, accrued gains. Those who deliberately do not enter this commitment can hardly argue for identical tax treatment. Equality requires comparable obligations."
The argument is internally consistent. It is nonetheless wrong — on three counts.
First, it is circular. Pension equalisation and accrued gains are consequences of marriage, not independent reasons for it. The argument runs: marriage justifies the splitting because marriage generates obligations that the splitting reflects. That is a construction that supports itself.
Second, the history is wrong. The splitting was not introduced to reflect maintenance obligations in tax law. It emerged in 1957 as one of several constitutionally permissible responses to a Federal Constitutional Court ruling — and was justified in the political debate of the postwar period by the model of the housewife marriage, not by the theory of mutual maintenance. The claim that it reflects a solidarity community is a rationalisation that came later.
Third, and most sharply: the argument cannot handle the test case it itself raises.
Apply the principle to two unmarried parents raising children together. They carry comparable social obligations: maintenance duties towards the children, care work, economic commitment. They have waived the maintenance obligations towards each other — not the responsibility for the children. Should they receive equal tax treatment to the married couple?
Those who say yes have adopted the position of this article. Those who say no are no longer arguing for solidarity community — they are arguing that marriage as such deserves a reward. That is a different, harder position.
The category error underneath: the legitimate state interest lies in children and care work — not in whether the people performing that work have signed a particular contract.
The Constitutional Boundary
The most common response to reform proposals is constitutional: the splitting is required by the Basic Law; abolition is therefore impossible. This is not accurate.
Germany's Federal Constitutional Court declared the then-existing joint assessment without a splitting tariff unconstitutional in 1957 — because it placed married couples at a disadvantage relative to individuals with equivalent income. Article 6(1) of the Basic Law (Grundgesetz) prohibits a marriage penalty: the state may not punish marriage through the tax system. What the Court did not do was prescribe income splitting as the only permissible solution. It named it as one of several constitutionally sound options, alongside deductions for maintenance payments and joint assessment without a splitting tariff. The Court confirmed this in 2013 (BVerfGE 133, 377): to reflect an earning-and-spending community in tax law, "other legislative designs would also have been available."
The constitutional boundary can be stated precisely. Article 6(1) prohibits a marriage penalty — a tax disadvantage for married couples relative to individual assessment. It does not mandate a marriage premium. From BVerfGE 105, 313 (2002): "To satisfy the constitutional duty of protection and promotion, it is sufficient that marriage is not placed at a disadvantage." Marriage may be favoured. It need not be.
The Federal Finance Ministry's own scientific advisory board put this in writing in 2018. In the report "On the Reform of the Taxation of Spouses" (September 2018), the Board concluded that constitutional law is "less restrictive regarding the constitutionality of alternative models than is commonly assumed." That is not a reform pamphlet — it is the internal advisory body's formal opinion.
Article 6(5) of the Basic Law adds a further dimension: the constitution explicitly requires the legislature to create equal conditions for children born outside marriage and those born within it. An instrument that channels family support through marital status creates tension with this requirement — not as a clear constitutional violation, but as an argument that a child-based reform better reflects the spirit of Article 6 as a whole.
The constitutional vulnerability in Klingbeil's proposal is not abolition itself — it is abolition without a replacement. The constitutional minimum is narrow: the unused basic personal allowance of the lower-earning partner must remain transferable to the primary earner, so that no new tax disadvantage is created for married couples. If abolition of the splitting produces a situation where married couples pay more than comparable individuals, that would be precisely the marriage penalty the Court prohibited in 1957. Klingbeil has not addressed this.
What Klingbeil Should Have Said — and What Is Actually Needed
I would put what was delivered here considerably more directly: no discernible concept, no clearly named replacement instrument, poor framing, weak communication, visibly reactive to a poor election result. But: § 188 of the German Criminal Code — which provides enhanced penalties for defamation of public officeholders, including for value judgements that are factually grounded — exists.
The criticism is therefore directed at the proposal.
Three structural failures can be attributed to the announcement of 25 March 2026.
The framing is wrong. Klingbeil framed the splitting as a labour market measure — a lever for releasing women from the part-time trap. That is not false, but it is the wrong question. A reform of the Ehegattensplitting is not a labour market measure. It is a fundamental decision about the principle of family taxation: should the state support care work and children — or marital status? Framing this as an employment problem invites a counter-question that cannot be answered: what if women do not want to work full-time?
The replacement is absent. The abolition is announced without naming what should follow. No family splitting, no child-based tax credits, no childcare expansion as a structural accompaniment. The Federal Finance Ministry's scientific advisory board described the problem clearly in 2018: a system change without expanding childcare infrastructure shifts the burden without resolving it — because the employment that the splitting suppresses can only increase once the care is available. A reform that removes the splitting without putting a child-based instrument in its place does not improve the situation of the women who currently depend on it.
The reach is marginal — as the proposal stands. It applies only to future marriages. Germany registers around 400,000 marriages a year. The effect accumulates over decades — maximum political resistance for minimum near-term impact. A choice model, giving existing marriages the option to switch, would fundamentally change this. That was the proposal that should have been made.
The Alternative Model
What would the right answer look like? Not abolition without replacement — but reallocation: away from the status bonus, towards care and dependency. Consistently.
The existing splitting ties the benefit to the marriage certificate. Anyone arguing that the problem is the wrong point of attachment — marital status instead of care work — cannot propose as a solution a model that still requires the marriage certificate. The contradiction would be obvious.
The consistent solution is simple, simpler than what currently applies.
Every person has a basic personal allowance. This is already law: roughly €12,000 of income is tax-free because it represents the tax-law subsistence minimum (Existenzminimum). Those earning below this threshold pay no income tax.
Whoever finances another person's basic allowance should be able to deduct it. The relevant criterion is not the living arrangement but the economic reality: B earns below the subsistence minimum; A contributes to B's subsistence. Whether B lives in the same household, studies in another city, or lives in a care home is irrelevant to the tax logic. The unused basic allowance of B is the available tax volume — it does not belong economically to A's income, because A is effectively spending it on B. It is therefore transferable to A. This is not a reward for marital status. It is the tax reflection of an actual expenditure. The maximum benefit is the marginal tax rate times the unused allowance — for a single-earner couple at a median income, roughly €2,000–3,000 per year, well below today's splitting maximum. The amount should not be set politically but adjusted annually to track the subsistence minimum — following the existing mechanism of the Federal Government's Existenzminimumbericht, which already updates the basic allowances on a regular cycle.
Several supporters can claim shares of B's unused allowance. Three siblings jointly financing their mother each receive tax recognition for exactly the share they actually contribute — not just the sibling who happens to claim the deduction. B tells the supporters their respective shares; each records the amount in their own tax return, and B confirms the total allocation in their own return. The tax office adds the claimed shares and verifies against B's actually unused allowance. Whoever uses this mechanism must file electronically — via ELSTER, Germany's digital tax filing system, or equivalent platforms. The ID cross-referencing runs automatically; paper-based returns cannot support this linkage. This is not a barrier — it is the technical precondition for a model that operates without bureaucracy.
Every child brings an additional allowance — for all parents, regardless of marital status. Per child, a further allowance is transferred to the parent, offset against child benefit (Kindergeld) through the same Günstigerprüfung (favourable-option test) that applies today. Two unmarried parents raising three children together receive the same benefit as a married couple with three children. The bonus follows the care work, not the register office. The level should likewise be indexed and empirically grounded — based on actual child cost research, not as a political estimate. Conceptually, a figure in the order of two-thirds of the adult subsistence minimum is plausible; the precise calibration is for statisticians and tax researchers, not coalition negotiators.
Blended families and shared custody are handled through proportional allocation of the child allowance — following the existing half-split mechanism for the Kinderfreibetrag under § 32(6) of the Income Tax Act. The mechanism exists; it simply needs to be extended to the new instrument.
The model is constitutionally sound. Article 6(1) prohibits a marriage penalty — through the transferability of the unused basic allowance, married couples are not placed at a disadvantage relative to individual taxpayers. A marriage premium beyond this is constitutionally permissible but not mandated; the Federal Constitutional Court has confirmed this consistently, most recently in 2013 (BVerfGE 133, 377). Article 6(5), which guarantees equal development conditions for children born outside marriage, is not merely satisfied by a status-neutral child allowance — it is actively implemented.
The transition for existing marriages: annual choice. Those married today who have arranged their lives in reliance on the existing splitting deserve protection — and receive more than that. No compulsion, no deadline, no one-time decision binding for decades. Marriages concluded before the new system comes into force choose each year, with their tax return, which system applies to them. This is not an invention — § 26 of the Income Tax Act already allows married couples to choose annually between joint and individual assessment. The mechanism exists; it simply needs to be extended to the system choice. In a year when children are present and the new child allowance is more advantageous, a couple takes the new system. The following year, if circumstances change, they choose again. The tax law reflects the actual life situation — every year. This also resolves the reach problem that Klingbeil's proposal leaves open: the new standard becomes the majority once it is the better deal for the majority. Not after decades.
The model costs substantially less than the existing system, because the bonus only arises where children are present — not for every income gap between childless married couples. The fiscal space this creates is available for childcare infrastructure — the precondition without which any reform remains theoretical.
Those who object that this is an attack on marriage: it is the opposite. It is the demand that the state supports the performance — not the contract.
Those on the other side who want to abolish the splitting without offering this replacement: that is not progress. It is a spending cut with progressive branding.
Conclusion
This article is not an argument against marriage. It is an argument against using marriage as a proxy for something else.
The state has a legitimate interest in children and care work. That interest can be pursued directly through the tax system — without routing it through marital status, without excluding parents who cannot or do not want to present a marriage certificate. The constitutional obstacles are smaller than claimed. The Federal Constitutional Court preserved the legislative space in 1982 and confirmed it explicitly in 2013. The obstacles are political, not legal: coalition pressure, church lobbying, protection of acquired positions.
The Ehegattensplitting has been unchanged since 1958. Several serious reform attempts have failed against precisely these obstacles — not in Karlsruhe.
The next question — what marriage means as an institution once the special tax status is removed — is a different analysis. The answer will be more uncomfortable than the question.
Published: 2026-04-07