Where the 1.43 trillion goes, and why the state cannot fully account for its distribution by nationality

Germany’s social budget worked out to 45,375 euros per second in 2025. 70 percent of social protection benefits went to old age and sickness. 1.4 million civil service pensioners received on average considerably more than recipients of a statutory old age pension. The total social contribution rate formed for a parent outside Saxony from the applicable contribution rates and the membership weighted supplementary rate stood at 42.29 percent at the start of 2025, and on the upper projected stress path it passes the 50 percent mark by 2040

I was on the autobahn toward Munich when the file came in. I pulled onto the next rest stop, killed the engine, and read the numbers on the phone with trucks idling 20 meters away. It took me maybe 4 minutes to understand that I had been asking the wrong question for months, and about 40 seconds after that to understand that almost everyone I know has been asking the same wrong question.

The wrong question is where the money goes. Everybody asks it, I asked it, and it produces a fight about the smallest part of the bill.

A few days earlier I had published a piece on the German state taking in 2,140.2 billion euros in 2025 and still ending the year 119.1 billion short. I sent it to the chancellor, to the federal minister of labor, and to the minister president of my state, with a short note saying that these figures worry me. No reply came, and I did not expect one, because those offices receive a thousand mails a day and mine was one of them.

The replies came from somewhere else. The first call was from the friend who appears in that piece, the one who left this country and sounds happier every year, and she was not amused this time, because what she had read had genuinely shaken her. She had grown up with the same number everyone here grows up with, the one about Germany collecting roughly a trillion in taxes, and nobody had ever put the other side of the ledger in front of her.

Then came the calls from acquaintances. One asked, in a tone I would describe as careful, whether I was going into politics now as an agitator. Another asked outright whether I vote for a particular party, which I do not, because I have not voted for any party in years. Then came the messages, dozens of them, and I am not going to quote from those. What matters is that roughly every second one contained the same open question, sometimes at the end, sometimes buried in the middle where people put the thing they actually want to know.

Where does all that money go.

So I sat down and took the whole thing apart, line by line, system by system, and what I found on that rest stop is worse than the version people are shouting at each other about. Not because the number is bigger than advertised. Because of what nobody can tell you about it.

Here is the thesis of this piece, stated plainly so nobody has to assemble it: Germany runs a welfare state of 1.43 trillion euros a year. For its largest insurance systems it cannot fully account for how that spending divides by nationality. At the same time the demographic arithmetic underneath it stopped working several years ago. That is not an argument against having a welfare state. It is an argument against running one blind.

The four numbers that never appear on the same page

Before any of this makes sense, four figures have to be separated, because German public debate throws them into one pot with impressive consistency.

The first is what the entire state took in and spent. Under the framework of the national accounts, which covers the federal government, the states, the municipalities, and the social insurance funds, Germany collected 2,140.2 billion euros in 2025 and spent 2,259.3 billion. The financing deficit came to 119.1 billion euros, or 2.7 percent of economic output. [2] The second is the social budget, which is the figure this piece is about. The Federal Ministry of Labor and Social Affairs puts it at 1,430.95 billion euros for 2025, with a data status of May 2026, and the 2025 values are explicitly estimates rather than final results. [1] Of that, 1,373.075 billion are social protection benefits, 51.189 billion are administration, and 6.686 billion are other spending. The social benefit ratio, meaning the social budget measured against economic output, rose by 0.7 points to 32.0 percent, and it rose mostly because the economy grew nominally by only 3.3 percent.

The third is the federal budget, which is a much smaller thing than most people assume and which is not the same as the social budget at all.

The fourth is the national accounts figure for monetary social benefits paid by the state, which came to 751.2 billion euros in 2025. [3] That number counts only cash transfers and uses a different boundary again.

Anyone who adds the social budget to the state accounts double counts. Anyone who describes the social budget as a pot the finance minister can reallocate has misunderstood the document. The social budget includes continued wage payment by employers during illness, occupational pensions, and the basic benefits of private health insurance, none of which are state spending. It is a consolidated national accounting of social security across all carriers, and its purpose is to show the size of the machine, not to hand anybody a savings list.

What can be said cleanly is this. In pure size terms the social budget corresponds to roughly 63 percent of state spending as measured in the national accounts. That quotient is not a spending share, because the two boundaries overlap and because the social budget additionally contains non state employer and insurance benefits. A large share of that total consists of insurance benefits financed by contributions and earned entitlements. [1] [2] A pension is not welfare. A hospital stay paid by a fund you have paid into for 40 years is not welfare either. Confusing the two is the single most common error in this entire debate, and both political camps commit it daily.

Now the arithmetic that made me stop the car.

Social budget 2025	Amount
Total	1,430.95 bn EUR
Per inhabitant	17,144 EUR
Per employed person	31,108 EUR
Per day	3.920 bn EUR
Per second	45,375 EUR

Own calculation, based on 83,467,117 inhabitants at December 31, 2025, and roughly 46.0 million employed persons on annual average. Day equals annual amount divided by 365, second equals annual amount divided by 31,536,000. [1] [4] [5] Those per capita values are not payments to anybody. They are a sense of scale, and the scale is the point. In the time it takes to read this paragraph, the German social state has committed roughly a quarter of a million euros.

Where the money actually goes, and it is not where the fight is

Here is the functional breakdown, which is the single table that should be on the front page of every newspaper in this country and appears on none of them.

Function	2025	Share of social protection benefits	Per inhabitant
Old age	492.608 bn EUR	35.9 %	5,902 EUR
Sickness	469.062 bn EUR	34.2 %	5,620 EUR
Children	155.312 bn EUR	11.3 %	1,861 EUR
Invalidity	95.578 bn EUR	7.0 %	1,145 EUR
Survivors	70.680 bn EUR	5.1 %	847 EUR
Unemployment	47.562 bn EUR	3.5 %	570 EUR
Housing	27.854 bn EUR	2.0 %	334 EUR
Remaining functions	14.419 bn EUR	1.0 %	173 EUR

Old age and sickness together come to 961.67 billion euros. That is 70.0 percent of all social protection benefits in this country, and it is the first finding any honest debate has to survive.

The mass of the money does not go to unemployment benefits. It does not go to refugees. It goes to pensions, to hospitals, to medication, to care homes, and to the consequences of a society whose median age keeps climbing. It goes, in very large part, to the same people who call me asking where it goes, and to the generation directly above them.

Unemployment, the function the entire country argues about, accounts for 3.5 percent. Housing accounts for 2.0 percent. Together the two functions come to 75.416 billion euros. The sickness function alone grew within a single year from 432.665 to 469.062 billion euros, an increase of 36.397 billion, which is close to half of those two permanently contested functions combined. [1] Now the financing side, which is equally absent from public discussion.

Source of financing 2025	Amount	Share
Employer contributions	512.806 bn EUR	34.5 %
Contributions of the insured	461.096 bn EUR	31.0 %
State subsidies	488.086 bn EUR	32.8 %
Other revenue	26.252 bn EUR	1.8 %
Total financing	1,488.239 bn EUR	100.0 %

The financing total exceeds the spending total because the system shows a financing balance of 57.289 billion euros for 2025. That is not an error, it is the other side of the same consolidated accounting. [1] Two things follow from this table, and both cut against a slogan somebody likes.

Contribution financed does not mean workers alone carry it, because state subsidies pay for almost a third of the whole apparatus, and a large part of those subsidies flows into the insurance funds. Tax financed does not mean means tested either, because child benefit is partly a tax compensation mechanism rather than a transfer to the needy.

And the employer share being larger than the employee share is not generosity. It contains imputed contributions for civil service pensions, for the health cost reimbursement scheme covering civil servants, and for continued wage payment. Somebody produces all of it, and that somebody works.

The number that is missing, and the fact that it is missing on purpose

Now to the part that made me put the phone down on the rest stop and stare at a fence.

I wanted to know how those 1.43 trillion euros divide between German nationals and everybody else. Not because the answer would settle any moral question, because it would not. I wanted to know because the entire political argument in this country is conducted as though the answer were known.

It is not known, it cannot be produced from the data that exist, and in the three largest systems it is not collected or published as a complete euro breakdown of spending by nationality.

Area	2025 total	Complete euro spending by nationality available?	What can actually be said
SGB II payment entitlements	46.653 bn EUR	Euro and persons	Non-Germans 21.752 bn, EU 3.294 bn, Ukraine 5.990 bn
Child benefit	55.324 bn EUR	Euro by nationality of the claimant	Non-Germans 12.684 bn, EU 4.225 bn, Ukraine 1.197 bn
Asylum seeker benefits	6.2 bn EUR gross	No euro amounts by nationality	Only totals and recipient counts
Social assistance SGB XII [22] [23]	21.5 bn EUR net	No euro amounts by nationality	Recipient counts only
Housing benefit	5.0 bn EUR	None	No usable payment sum by nationality
Statutory health insurance	352.4 bn EUR	None	Spending recorded by service type only
Long term care insurance	73.82 bn EUR	None	Spending recorded by benefit type only
Statutory pension insurance	426.50 bn EUR	None	Stock figures and a cautious calculation only

Read that third column twice, because it carries the whole argument. For roughly 850 billion euros a year, meaning the three large insurance systems, no such breakdown exists at all. The federal government has confirmed this in writing in response to a parliamentary question: health insurance financial statistics record expenditure by type of service and by provider, not by nationality or membership group, physicians’ billing data contain no nationality field, and no such breakdown is planned. [24] Anybody naming an officially exact share of all German social spending by nationality is either using the small measurable slice as though it were the whole, or claiming a precision the official data do not support. Model estimates are possible, a complete spending statistic they are not.

I want to be exact about what that does and does not mean, because this is the sentence people will rip out of context. Not available does not mean zero. It means the claim cannot be derived from official data, in either direction. It is just as unserious to assert a huge share as to assert a negligible one.

And now the part that I did not expect and that changed the piece. The missing number is not the scandal. The missing number is a symptom. The scandal is that Germany runs a system of social security with a social budget of 1.43 trillion euros whose largest carriers cannot answer basic questions about the distribution of their spending, and that the federal government has stated it is not planning such a survey.

That is what I mean by loss of control. It is not theft and it is not a conspiracy, it is something considerably duller and much worse, an apparatus that has grown past the point where anybody inside it can see the whole of it.

1.4 million people in a pension system without individual contributions

Let me start with the part of the bill that almost never appears in this argument, because it belongs to the group that writes the rules.

On January 1, 2025 Germany had 1,418,800 pensioners of the public service under civil servant and soldier pension law. A further 382,100 survivors drew benefits from the same system. The average pension came to 3,416 euros gross per month, which was 5.4 percent more than a year earlier. [6] Ten years before that the same figure stood at 2,730 euros.

Spending on those pensions came to 56.9 billion euros in 2024. Survivor benefits added 9.0 billion. Together that is 65.9 billion euros, roughly 1.5 percent of economic output, and it is financed out of current public budgets without individual pension contributions for that entitlement.

Now hold that against the system the large majority of working people pay into. The statutory pension insurance paid out 379.437 billion euros net in pension spending in 2025, against total spending of 426.498 billion. At the end of 2024 there were 18.919 million statutory old age pensions in payment, and their average gross pension amount came to 1,296 euros per month. [7] [8]

System	People	Average monthly amount	Financing
Civil service pensions, January 2025	1,418,800 pensioners	3,416 EUR gross pension	public budgets, no individual pension contributions
Survivor benefits, January 2025	382,100	not comparable with the pension figure	public budgets, no individual pension contributions
Statutory old age pensions, end of 2024	18.919 million	1,296 EUR gross pension amount	Contributions plus federal funds

The plain quotient of those two gross figures is 2.64. That is not a comparison of like working lives, and it does show the order of magnitude of the gap. [16] The pension comes from a system into which no individual pension contribution was paid for that entitlement, because the German civil service pension follows from the employment relationship itself. And the health cost reimbursement scheme for retired civil servants comes on top of that figure, financed the same way.

I want to be careful here, because this is where a text like this usually goes stupid. The systems are not comparable in construction. Civil servants receive lower gross salaries in some brackets precisely because the pension promise is part of the deal. Many of them worked in jobs that are genuinely hard, and a police officer who spent 35 years on shift work has earned every cent. None of that is in dispute and I am not disputing it.

What I am recording is something narrower and, to me, decisive. In the entire pension debate of the past 3 years, in the commission report of June 2026, in the coverage of the contribution rate rising toward 20 percent and beyond, one system has been discussed relentlessly and the other has been almost invisible. The one that is discussed is the one the large majority of working people pay into. The one that is invisible is the one that pays 2.64 times as much on a pure gross comparison, out of current public budgets and without individual pension contributions for that entitlement, to a group that includes the people drafting the legislation.

I am not accusing anybody of anything, and I am explicitly not alleging any wrongdoing. I am pointing out that the group least exposed to the arithmetic is the group holding the pen. That is a structural observation, and structural observations do not require intent.

And the number is going the wrong way. The federal government alone expects its own pension spending, including the postal successor bodies, to rise from 19.8 billion euros in 2025 to roughly 30.6 billion by 2060, with the number of federal pension recipients climbing from 193,000 in 2025 to around 230,000. The states carry the far larger share, and their spending is driven by hiring decisions taken decades ago that cannot now be unwound.

There is a second layer to this that almost never surfaces, and it is larger than the pension itself. Retired civil servants remain inside the state’s own health cost reimbursement scheme, which reimburses a set rate of their eligible medical expenses directly from current public budgets, with private insurance covering the remainder. Those payments are not part of the 65.9 billion euros above. They sit in a separate line of the public accounts, they rise with medical inflation exactly like everybody else’s health costs, and they attach to a group whose average age keeps climbing.

Then there is the liability that has already been incurred and simply has not been paid yet. Pension entitlements earned by today’s serving officials are a promise the state has already made, and unlike a fully funded scheme there is no capital stock covering those obligations in full, because the federal government and the states do maintain pension reserves and pension funds, and those cover only part of the obligations. Estimates of the accumulated obligation run into several hundred billion euros, and I am deliberately not printing a single figure here, because the published estimates differ by a factor that makes any one of them look like advocacy. What is not in dispute is the direction. For the immediate federal sphere the government expects a rise from 193,000 pension recipients in 2025 to 230,000 in 2060. Including the expiring stocks of the former rail and postal sectors, the federal total falls over the same period from 595,000 to 298,000. The states, which employ teachers and police among others, form pension systems separate from that. [15] The reason this matters for the rest of the piece is mechanical rather than moral. Every euro of civil service pension comes out of the same public budgets that also have to fund the pension subsidy, the health fund subsidy, the basic income support reimbursement, and the debt service. It competes directly with all of them. And unlike a contribution rate, it cannot be adjusted by turning a dial, because it rests on individual legal entitlements protected by constitutional principle.

Both systems create legally protected entitlements. The civil service pension is additionally shaped directly by the constitutional principle of maintenance, while the contribution rate, the pension formula, and the federal subsidies of the statutory pension insurance are adjusted politically far more often. I am not going to pretend that is an accident, and I am also not going to claim I can prove it is not.

The 46.65 billion everybody is fighting about

Now to the part where the shouting happens, and where, unusually, we do have exact figures.

The federal government answered a parliamentary question in July 2026 with the complete 2025 breakdown of payment entitlements under SGB II, the German system of basic income support for jobseekers. The data status is June 2026. [17]

Group 2025	Regular benefit recipients, annual average	Share	Payment entitlements	Share
All	5,316,378	100.0 %	46,652,758,234 EUR	100.0 %
Germans	2,802,938	52.7 %	24,900,792,076 EUR	53.4 %
Non-Germans	2,513,386	47.3 %	21,751,578,677 EUR	46.6 %
EU excluding Germany	391,912	7.4 %	3,294,426,960 EUR	7.1 %
Third countries	2,120,365	39.9 %	18,446,897,161 EUR	39.5 %
Ukraine	679,291	12.8 %	5,989,845,053 EUR	12.8 %
Top 8 asylum origin countries	884,177	16.6 %	7,233,221,710 EUR	15.5 %

The groups overlap and must never be added. Ukraine sits inside third countries. Romania and Bulgaria sit inside both the EU line and the non-German line. Anyone summing those rows produces a number that does not exist.

Administrative costs of 6.788 billion euros come on top and cannot be split by nationality at all.

Now put that against the social budget, which is the calculation nobody performs on television.

The SGB II payment entitlements, including housing costs and the health and care contributions paid on behalf of recipients, amount to 3.3 percent of the German social budget. Including the separately reported administrative costs the figure is roughly 3.7 percent. [1] [17] The share attributable to Ukrainian nationals amounts to 0.42 percent of the social budget. The share attributable to all non-Germans in this one system amounts to roughly 1.5 percent.

The country has spent 4 years arguing about 3.3 percent of the bill, measured on payment entitlements.

I am not saying those figures are trivial in absolute terms, because 21.75 billion euros is not trivial by any standard, and neither is 5.99 billion. I am saying that anybody who presents this as the cause of a 1.43 trillion euro problem is either innumerate or dishonest, and I no longer care which.

There is a second reading of the same table, and it is the uncomfortable one for the other camp. Non-Germans make up roughly 15 percent of the population and 47.3 percent of regular benefit recipients. That is not a rounding artifact and it will not be explained away by pointing at the small euro share. It is a benefit, integration, and labor market problem of the first order, and I will come back to it, because it is the actual finding of this piece. For a pure labor market statement the relevant subgroup is employable benefit recipients rather than the total including children.

Note also what the payment entitlement figure contains. Divide 5.99 billion euros by 679,291 Ukrainian recipients and you get 8,818 euros per person per year. That is not a cash handout. It contains the standard rate, housing and heating, and the health and long term care insurance contributions paid on the recipient’s behalf, and the recipients live in households where children are counted as individuals. I am putting the calculation here because somebody will do it anyway, and I would rather it be done correctly.

Child benefit, and the export that is not the scandal

Child benefit in Germany is not primarily a welfare payment. It is partly a constitutionally required offset that keeps the subsistence minimum of a child free of income tax, which is why it appears in the social budget and in the tax code at the same time.

In 2025 the family benefits office paid out 55,323,952,897 euros in child benefit for roughly 17.57 million children, plus 3,279,064,536 euros in the supplementary child allowance for low income households. [18] [19]

Nationality of the claimant	Child benefit	Share	Child allowance	Combined
All	55.324 bn EUR	100.0 %	3.279 bn EUR	58.603 bn EUR
Germans	42.640 bn EUR	77.1 %	1.607 bn EUR	44.247 bn EUR
Non-Germans	12.684 bn EUR	22.9 %	1.672 bn EUR	14.356 bn EUR
EU excluding Germany	4.225 bn EUR	7.6 %	0.378 bn EUR	4.603 bn EUR
Romania	0.854 bn EUR	1.5 %	0.100 bn EUR	0.954 bn EUR
Bulgaria	0.410 bn EUR	0.7 %	0.060 bn EUR	0.470 bn EUR
Ukraine	1.197 bn EUR	2.2 %	0.076 bn EUR	1.273 bn EUR

The EU figures are the sum of the 26 foreign EU nationality rows in the official table, and the non-German figures are the difference between the total and the German row. That is reproducible arithmetic from an official table, not a model.

One detail matters more than any of these numbers. The statistic classifies the nationality of the claimant, not the residence or nationality of the child. So this table says nothing whatsoever about money leaving the country, and anybody using it that way is misreading the column header.

For money actually leaving the country there is a separate figure, and it is small. Child benefit transferred to accounts abroad came to 528 million euros in 2025, roughly 1 percent of the total, with 99 percent of it going to EU member states, led by Poland, Romania, and Czechia. The payments follow the European coordination rules. They concern children resident abroad whose claimants meet the German eligibility conditions. Abuse has been documented in individual cases. [18] [19] I include it because it is the single most reliably misquoted figure in German social policy. It amounts to roughly 0.8 percent of what the civil service pension system costs in a year.

Asylum seeker benefits belong in the same category of things that are loudly discussed and quantitatively modest. Germany spent 6.2 billion euros gross under the asylum seeker benefits act in 2025, which is 8.2 percent less than the year before. In 2024 the figure was 6.718 billion gross and 6.423 billion net, with roughly 461,000 people receiving regular benefits at year end, of whom about 25,200 were Ukrainian nationals. [20] [21] Recognized refugees leave this system and move into the general one, which is why asylum seeker, asylum origin country, and refugee are not interchangeable words.

What comes back, and I am putting it in because leaving it out would be cowardice

If I only print the payout side of this ledger, somebody will take this piece apart in 2 minutes, and they will be right to.

In June 2025 Germany had roughly 34.9 million people in employment subject to social insurance contributions. Roughly 29.1 million were German nationals and roughly 5.8 million were foreign nationals, meaning 16.6 percent of the contributing workforce. In May 2012 that share stood at 7.5 percent. [55] Social contributions are deducted from their contributory earnings up to the applicable assessment ceilings. Where income tax liability and taxable consumption arise, they pay the same taxes under the same rules as German nationals.

The employment gap has narrowed substantially, and the direction is the opposite of what the debate assumes.

Employment rate, June	2015	2024	2025
German nationals	58.5 %	65.4 %	65.3 %
Foreign nationals	44.4 %	57.0 %	58.3 %
Gap	14.1 points	8.4 points	7.0 points

The foreign employment rate rose by 13.9 points in a decade. [53] It still sits 7 points below the German rate, and that gap has a price, which I will quantify in a moment. But the sentence that foreigners do not work is factually false, and I will not print a false sentence to please anybody.

On the fiscal balance there is one serious study I am willing to cite. Researchers at the Center for European Economic Research computed net fiscal contributions per capita for Germany using household panel data. In the model calculation for the reference year 2018, German nationals contributed a net 629 euros per head, EU foreigners contributed a net 3,175 euros, and non-EU foreigners came out at minus 2,633 euros. Combined, all foreign nationals produced a net position of roughly minus 2.5 billion euros, which is about 0.07 percent of economic output at the time. In that model calculation the net position of EU foreigners was clearly positive and that of non-EU foreigners negative. [74] That result is inconvenient for both sides, which is exactly why it is here. The combined effect is small compared to almost everything else in this piece. And the Ukrainian arrivals are not included in it, so the study does not allow any conclusion about what the same calculation would produce today. My personal expectation is that it would come out less favorably now, and that is expressly an expectation rather than a finding. I am saying that openly rather than hiding behind a study that flatters my argument.

Here is the honest summary of this chapter. The fiscal problem of German migration policy is not that people come. It is the 7 point employment gap, the clear overrepresentation in the benefit system, and the years that pass between arrival and the first tax payment. Those are policy failures with names and addresses, and they are fixable, which is precisely why leaving them unfixed is unforgivable.

Ukraine and Poland. Where does the gap come from?

This is the chapter people asked me about most, and it is also the chapter where the most quoted number in the entire debate fails to survive a check.

At the end of 2025 Germany had 1,327,268 people registered as having come from Ukraine in connection with the war since February 24, 2022. Of those, 1,290,526 were Ukrainian nationals, 36,549 were third country nationals, and 193 were stateless. [55]

Age group	People	Share
Under 18	357,158	26.9 %
18 to 21	115,690	8.7 %
22 to 39	342,163	25.8 %
40 to 59	344,547	26.0 %
60 to 64	49,866	3.8 %
65 and over	117,844	8.9 %
Total	1,327,268	100.0 %

Of that total, 757,398 were female and the remainder male or unrecorded. Children and people aged 65 and over together come to 475,002 individuals. Anyone dividing the number of employed Ukrainians by 1.33 million is treating children and pensioners as failed workers, and that calculation is worthless.

Germany, the hard facts

The comparison circulating in German debate puts Germany at 31 percent against Poland at 78 percent in the OECD comparison. The 78 percent for Poland is correct and refers to 2024, computed by the OECD from Polish central bank survey data, with Lithuania at 72 percent and the United Kingdom at 69 percent behind it. [66] The 31 percent is not Germany. In that same OECD table, 31 percent is Norway.

The German figure that gets quoted at 36 or 37 percent comes from the employment statistics of the federal employment agency, which counts only dependent employment, excludes the self employed and civil servants, includes Ukrainians who lived here long before the war, and is calculated against the register of foreign nationals. The administrative employment rate of Ukrainian nationals stood at 37.4 percent in September 2025. [54] [56] It is a real number and it measures something real, but it is not the same construct as a survey of displaced persons.

The comparable German figure comes from survey research. The Institute for Employment Research found an employment rate of roughly 50 percent for the Ukrainian refugee cohort that arrived between February and August 2022, measured roughly three and a half years after arrival. [58] A separate survey by the Federal Institute for Population Research arrived at 51 percent for the end of 2025. [59] [60] So that this debate stops being conducted with a single number, here are the available German figures side by side, each with its reference date and definition.

Measure	Value	Reference date and definition
Ukrainian nationals in dependent employment	340,550	June 2025, employment statistics
of whom subject to social insurance	285,504	June 2025
of whom in marginal employment only	55,046	June 2025
of whom working full time	182,145	June 2025, insured employment only
of whom working part time	103,359	June 2025, insured employment only
Self employed	data gap	no figure verifiable for this edition
Administrative employment rate	37.4 %	September 2025, denominator is the Ukrainian population aged 15 to under 65 in the register of foreign nationals
Unemployment rate	36.3 %	December 2025, share of the unemployed in the civilian labor force of the same nationality, not the share of all Ukrainian nationals without work
Exits from unemployment	518,000	annual total of exit cases in 2025, of which 92,000 into employment, multiple counting of the same person during the year possible
Employable benefit recipients under SGB II	478,320	December 2025
Non employable benefit recipients	179,357	December 2025
Regular benefit recipients including children	657,677	December 2025
Regular benefit recipients, annual average	679,291	annual average 2025
SGB II payment entitlements	5.990 bn EUR	full year 2025
Child benefit to claimants of Ukrainian nationality	1.197 bn EUR	full year 2025
Child allowance to the same group	0.076 bn EUR	full year 2025

The frequently quoted figure of roughly 320,000 Ukrainians in work refers to an earlier reference date and is out of date. For June 2025 the same statistics show roughly 341,000 people in dependent employment. Both figures cover Ukrainian nationals in general, including people who lived and worked in Germany long before 2022. [17] [18] [54] [55] [57] So the honest comparison reads as follows.

Ukrainian war refugees	Poland	Germany
Survey based employment rate	roughly 78 % (2024)	roughly 50 % for the early refugee cohort after about three and a half years
Administrative dependent employment rate	not comparable	37.4 % (September 2025)
Gap between two survey based but not fully harmonized measurements		roughly 28 points

28 points instead of 47. I would rather publish the smaller number that holds than the larger one that collapses under the first competent objection. And 28 points is still a catastrophe. Applied as a pure order of magnitude calculation to roughly 850,000 Ukrainians of working age in Germany, that gap corresponds to just under 240,000 additional people in work. That is a mechanical counter calculation and not a causal demonstration that Germany could reach the Polish rate exactly under its own conditions.

So where does that gap actually come from. There are 5 reasons for it and only one of them is comfortable.

What Poland does differently

The first is composition, and it explains a great deal. Poland had a large Ukrainian labor migration long before 2022. Employers, recruitment channels, language skills, and social networks already existed. The Polish figures must not be mixed, because they count three different things.

Measure	Value	Source and reference date
People under temporary protection	993,665	Eurostat, July 31, 2025
Ukrainian nationals performing work	739,400	Statistics Poland, August 31, 2025
Ukrainians insured with the social insurance institution	roughly 857,000	Polish labor ministry on the basis of ZUS data, December 31, 2025
of whom carrying the temporary protection marker	roughly 251,000	Polish labor ministry on the basis of ZUS data, December 31, 2025

So only about 251,000 of the insured Ukrainians carried the temporary protection marker. The remainder consisted in particular of earlier labor migration and people holding other residence statuses, so the insurance figure is not a pure refugee cohort. [61] [62] [63] Germany took in relatively more people with no existing labor market anchor at all, and a demographic skewed heavily toward women with children.

The second is language and recognition. Polish and Ukrainian are both Slavic languages, which does not remove the barrier but lowers it considerably. Germany chose a different route and put language and qualification recognition in front of employment rather than alongside it. In February 2026, roughly 64,900 Ukrainian nationals who were not registered as unemployed were attending integration courses. The hours spent there were not available for paid work. That approach produces better jobs eventually. It also produces years in which nothing is produced at all.

The third is benefit design, and the Polish central bank names it explicitly as a driver of high labor force participation. Poland’s child benefit program pays 800 zloty per child per month, roughly 2,264 euros per year at the 2025 average exchange rate, and it is not means tested. But since June 2025 it has been tied to school or preschool attendance for children with protection status, and since February 2026 renewed entitlement generally requires proof of economic activity, legal residence, actual residence in Poland, and the child’s school attendance, with exceptions for disability and certain care situations. Roughly 150,000 people were affected by the reapplication requirement. [67] [68] Germany, since June 2022, has given Ukrainian protection holders in need full access to the general benefit system with no such linkage.

Two further findings belong to the size of the Polish system. In December 2024 roughly 292,000 Ukrainian children received the 800 plus family benefit. As an expressly mechanical twelve month calculation, 292,000 times 800 zloty times 12 comes to 2.803 billion zloty, which at the 2025 average exchange rate is roughly 661 million euros. That is not an official annual figure, because both the number of children and the period of entitlement fluctuate during the year.

Classic municipal social assistance is considerably narrower. Statistics Poland counted roughly 24,200 foreign recipients of local social assistance in 2024, of whom 89.3 percent were Ukrainian nationals. Including special Ukraine measures, the value of recorded benefits to foreigners came to 101.0 million zloty, which is 1.5 percent of all recorded social assistance. At the 2025 exchange rate that would be roughly 23.8 million euros, and because the benefit figure is from 2024, that euro conversion is only an illustrative approximation. [64] [69] [70]

Polish benefit to Ukrainian claimants	Value	Boundary
800 plus family benefit, children	roughly 292,000	December 2024
800 plus, twelve month calculation	2.803 bn PLN, roughly 661 m EUR	own mechanical calculation, not an official annual figure
Municipal social assistance to foreigners	101.0 m PLN, roughly 23.8 m EUR	2024, converted at the 2025 rate, approximation only
Accommodation	no consolidated spending figure available
Education	no consolidated spending figure available
Health care access	no consolidated spending figure available

Those 101 million zloty exclude the large family program and therefore do not prove that Poland pays almost nothing. They do show that ongoing subsistence social assistance for foreigners is small in the Polish system.

The fourth is job structure, and this is where Poland’s number gets less flattering. Poland moved people quickly into production, logistics, construction, services, and simple work. In the Polish central bank survey, 36 percent of employed war refugees described their work as below their qualification level, rising to 58 percent among those in simple occupations. [65] Poland bought its employment rate partly with underemployment, and it is fair to say so.

The fifth is gender and care. The displaced population is mostly women and children. Single parenthood, missing childcare, and care responsibilities cap working hours and mobility, in Poland as much as here. In my assessment that explains part of the gap, but not the full 28 points.

My reading of all of this, marked as my reading: Poland accepts fast and sometimes unsuitable employment and ties benefits tightly to activity. Germany invests in security, language, and formal qualification, and has tolerated for far too long that a very large number of employable people remain in the transfer system while it does so. Humanity explains the goal, and it does not excuse the absence of any serious measurement of whether that goal is being reached.

The arithmetic that does not care what anybody votes

Everything up to this point was about a single year. Now comes the part that made me stop treating this as a policy argument and start treating it as a countdown.

Between 2015 and 2025, Germany recorded 2,568,888 more deaths than births. The total fertility rate fell further to 1.32 children per woman in 2025. [51] In 2025 alone the natural deficit came to 352,347 people, from 654,241 births against 1,006,588 deaths. [48]

Year	Births	Deaths	Natural balance
2015	737,575	925,200	−187,625
2016	792,141	910,902	−118,761
2017	784,901	932,272	−147,371
2018	787,523	954,874	−167,351
2019	778,090	939,520	−161,430
2020	773,144	985,572	−212,428
2021	795,492	1,023,687	−228,195
2022	738,819	1,066,341	−327,522
2023	692,989	1,028,206	−335,217
2024	677,117	1,007,758	−330,641
2025	654,241	1,006,588	−352,347
Sum 2015 to 2025	8,212,032	10,780,920	−2,568,888

In 2025 Germany still recorded net immigration of 235,000 people. It was no longer enough. Births minus deaths plus net migration comes to minus 117,347, and the officially recorded population decline came to roughly 110,000, with the small difference arising from rounding and statistical correction. The population stood at 83.5 million at the end of 2025, the first fall since 2020. [4] [47] Now the official projection, which is harder than any headline I have read about it. The 16th coordinated population projection of the Federal Statistical Office is not a forecast but a system of scenarios, and even its friendly variants are grim.

Measure	2024	2040	Change
Population 20 to 66	51.2 m	45.4 to 48.0 m	−3.2 to −5.8 m
Population 67 and over	16.7 m	at least 20.5 m	at least +3.8 m
People 20 to 66 per person 67 and over	3.07	2.21 to 2.34	−24 to −28 %

The spread between 45.4 and 48.0 million rests mainly on the migration assumption, either 150,000 or 350,000 net arrivals per year in the long run. Note carefully what that spread actually means, because it is the load bearing sentence of this whole section. Even the high migration variant does not prevent the decline. None of the officially calculated migration variants makes this table go away by 2040, because the people who will be 67 in 2040 have already been born and are already here. [49] The relevant formula is not total inhabitants. It is the number of benefit cases multiplied by the benefit per case, divided by the number and income of the people paying contributions. More inhabitants help only if they are of working age, actually work, earn enough, and pay in. Children are a future investment and present cost. On average, health and care spending rises steeply with age, and the contributions paid by pensioners do not cover that spending in full.

Now put the cost side next to it. Combined spending by statutory health insurance and long term care insurance rose from 242.68 billion euros in 2015 to 426.22 billion in 2025. That is 75.6 percent in a decade, an average of 5.79 percent per year. [25] [26] [27] [46] And here I have to correct a claim I have made myself in conversation and that circulates constantly. Revenue did not fail to keep up. Combined revenue of the same two systems rose by 76.5 percent over the same decade, marginally faster than spending. The sentence that revenue is not growing is nominally false.

What is true is how that revenue was produced. It came from higher wages, more people in work, higher contribution rates, higher assessment ceilings, and public money. The care insurance contribution rate rose from 2.35 percent in 2015 to 3.6 percent in 2025, and the surcharge for the childless rose from 0.25 to 0.6 points. The health insurance supplementary rate averaged 2.94 percent at the end of 2025, although an average supplementary rate of 2.5 percent had been announced for 2025. Revenue kept pace because the burden was increased, repeatedly, on the same shrinking group.

So the honest question is not whether the money arrives. It is how much further the burden per contributor has to rise as the number of people of working age falls. For converting nominal amounts into 2025 prices, the scenarios below assume a constant 2.0 percent inflation per year.

Cost scenario to 2040	Nominal growth per year	Spending 2040 nominal	In 2025 prices
Dampened	3.0 %	664 bn EUR	493 bn EUR
Middle burden path	4.5 %	825 bn EUR	613 bn EUR
Historical rate 2015 to 2025	5.79 %	992 bn EUR	737 bn EUR

Converted into a burden indicator, meaning combined health and care spending divided by the population aged 20 to 66, the 2025 figure is 8,325 euros. That starting value combines 2025 spending with the population aged 20 to 66 from 2024 and is therefore an approximation. For 2040, in 2025 prices, it looks like this.

Cost scenario	At 48.0 m people 20 to 66	At 46.7 m	At 45.4 m	Real increase against 2025
3.0 % nominal	10,279 EUR	10,565 EUR	10,868 EUR	+24 to +31 %
4.5 % nominal	12,768 EUR	13,124 EUR	13,500 EUR	+53 to +62 %
Historical rate	15,356 EUR	15,784 EUR	16,236 EUR	+85 to +95 %

That is not a contribution notice and nobody will receive a bill for it. It is a stress measure, total spending divided by the potential working age population, and it exists to show a direction rather than a payment. The direction is that health and care alone, in real terms, could demand between a quarter and nearly double again per person of working age within 15 years.

The government’s own health finance commission has already put a figure on the near term. Without countermeasures it expects a funding gap of more than 15 billion euros in 2027, rising to more than 40 billion by 2030. [31] And the pension system is not waiting either. The statutory pension insurance closed 2025 with a deficit of 3.9 billion euros. Its sustainability reserve stood at 41.3 billion euros at year end, equal to 1.38 months of spending. The pension insurance report projects the contribution rate holding at 18.6 percent through 2027, then rising to 19.8 percent in 2028, 20.0 percent in 2029, 20.1 percent by 2030, and 21.2 percent by 2039. [8] Read those two paragraphs together. Health, care, and pensions all come under financing pressure at the same time, on the same shrinking base, and reserves are already being drawn on in several of these systems.

What the birth rates support and what they do not

One demographic finding is misused in this debate almost every time, which is why it belongs here. Yes, immigrant women have more children on average than German women. In 2024 the total fertility rate stood at 1.84 for foreign women and 1.23 for German women. That figure has also fallen almost continuously since 2017, and it sits well below the replacement level of roughly 2.1. [50] Immigration raises the number of births and can soften the birth deficit. The fertility rate of foreign women alone does not close the gap either. Anybody selling it as a complete solution is calculating as badly as the person presenting immigration as the sole cause of the costs.

Whether demographic considerations played a part in the decisions of 2015 I do not know. I will not attribute a motive to a real person that I cannot document. What I can say is that the demographic figures were already available at the time and that every government has known them since.

What could have gone better

The aging of this society has many causes, and most of them are not scandals. They are the consequences of free decisions.

Women have become more independent, they study, they work. That is what puts the choice between a career and children on the table in the first place, and the fact that it is on the table is progress rather than a problem. It becomes a problem only where a state makes that choice expensive.

And that is exactly what I have been saying for decades. A state has to make it possible to have children without carrying a financial penalty for it. When I look at what young parents around me go through with childcare places, schools, and costs, I understand every decision against having a child. That went wrong, and it went wrong across decades.

The second point is less comfortable. A single parent with 4 children can often take on paid work only to a limited extent, especially when reliable childcare is missing. That produces additional public spending. I say this without any reproach, because the reproach would be aimed at the wrong address. A family policy that fails to secure childcare pays twice in the end, once through forgone employment and once through transfer payments.

And here I deliberately turn my own calculation around. Seen through this piece, those 4 children are not a cost center. They are the only item on the entire bill that can later move to the revenue side by itself. I wrote above that children are an investment in the future and an expense in the present, and that is exactly the case here. Somebody raising 4 children inside a system whose core problem reads 2.21 to 2.34 working age people per person aged 67 and over is delivering precisely what that system lacks. [49] Whether 4 contributors actually come out of it depends on childcare, school, training, and the labor market, meaning on exactly the places where this country has been economizing for decades. That is my assessment and not a finding, and it follows from the same formula I used for everything else above.

I am not trying to be clever here, and I have no prescription I would call complete. I only know that a society which financially punishes children has no business being surprised by a shortage of them.

Who actually paid for the revenue that kept pace

I wrote a paragraph above saying that revenue in health and care insurance grew slightly faster than spending, and that the claim revenue is not growing is nominally false. That paragraph is correct and it was incomplete, and the missing half is the part that matters most to the person reading this.

Revenue did not grow by itself. It was mobilized, deliberately, by widening and loading the base that pays for it. Higher supplementary rates in health insurance, higher contribution rates in care insurance, rising assessment ceilings, rising contributory wages, continued and additional payments from the federal budget, and in individual years loans and special subsidies. Every one of those instruments ends at the same address.

The federal budget is not a third party standing outside society. Federal subsidies come from taxes or from borrowing, and borrowing is a financing burden moved into the future. Employer contributions do not appear as a deduction on anybody’s payslip, and they are part of the cost of a job all the same. So the precise sentence is this. Contribution revenue rose because the contributory and taxable base was widened and burdened more heavily, which has stabilized the financing so far, and which was not free.

The following calculation uses the average supplementary rate announced for each year together with the rates in force in the other branches of social insurance.

Year	Pension	Unemployment	Health, general	Average health supplement	Care, parents	Total
2015	18.70 %	3.00 %	14.60 %	0.90 %	2.35 %	39.55 %
2017	18.70 %	3.00 %	14.60 %	1.10 %	2.55 %	39.95 %
2019	18.60 %	2.50 %	14.60 %	0.90 %	3.05 %	39.65 %
2021	18.60 %	2.40 %	14.60 %	1.30 %	3.05 %	39.95 %
2023	18.60 %	2.60 %	14.60 %	1.60 %	3.40 %	40.80 %
2024	18.60 %	2.60 %	14.60 %	1.70 %	3.40 %	40.90 %
2025	18.60 %	2.60 %	14.60 %	2.50 %	3.60 %	41.90 %

On a membership weighted basis the rate actually charged stood at roughly 0.83 percent at the start of 2015, at roughly 2.89 percent at the start of 2025, and at 2.94 percent by the end of December 2025. From those figures the total social insurance contribution actually charged comes to 39.48 percent at the beginning of 2015 and 42.29 percent at the beginning of 2025. [32] [33] [34] [35] [46] That is 2.81 points more, a relative increase of 7.1 percent.

Two components fell over that decade. The pension rate came down by 0.1 points and the unemployment rate by 0.4 points. That combined relief of 0.5 points was overwhelmed several times over by health and care.

Now the part that lands on the payslip. Since 2019 the supplementary health rate is shared equally between employer and employee again, whereas in 2015 the employee carried it alone. That is why the employee share rose by less than the headline rate.

An employed parent outside Saxony paid roughly 20.155 percent of contributory gross pay in 2015 and roughly 21.145 percent in 2025, an increase of 0.99 points. A childless employee aged 23 or over went from roughly 20.405 percent to 21.745 percent, an increase of 1.34 points. Isolating the rate effect alone and applying every component to the same contributory base of 60,000 euros, the result is 594 euros more per year for the parent and 804 euros more for the childless employee, before any effect of changed assessment ceilings and before a single cent of income tax. [46] Somebody will now object that inflation explains all of it, so here is the calculation with purchasing power held constant. The consumer price index stood at 94.5 in 2015 and 121.9 in 2025, an increase of 29.0 percent. An income of 50,000 euros in 2015 corresponds to 64,497 euros in 2025 at identical purchasing power. Counting only the direct employee share of social insurance and nothing else, the result looks like this.

Model person	Contribution 2015	2015 in 2025 prices	Contribution 2025	Real additional burden
Parent	10,078 EUR	12,999 EUR	13,638 EUR	639 EUR, or 4.9 %
Childless, 23 or over	10,203 EUR	13,161 EUR	14,025 EUR	864 EUR, or 6.6 %

At the ceiling, meaning the maximum monthly employee contribution using each year’s assessment limits, the real increase comes to 6.9 percent for a parent and 8.3 percent for a childless employee. That is the cleanest possible confirmation of the complaint people make at the kitchen table. After adjusting for inflation, the direct contribution burden did rise.

The assessment ceilings themselves moved as follows. The health and care ceiling went from 49,500 euros a year to 66,150 euros, up 33.6 percent nominally and 3.6 percent in real terms. The pension and unemployment ceiling in the west went from 72,600 to 96,600 euros, up 33.1 percent nominally and 3.2 percent in real terms. Those ceilings are legally indexed to wage growth, so their nominal rise is not by itself a hidden rate increase. For anybody earning between the old ceiling and the new one, the amount subject to contributions rose all the same. The 2025 health ceiling alone jumped 6.5 percent against 2024, and the government’s health finance commission names that unusually large step as the reason contributory revenue grew faster in 2025 than the total wage and salary bill. [31] [34] [35] [36] [37] Now step back from the payslip to the whole economy, because this is where the argument is usually lost.

Year	Tax ratio	Social contribution ratio	Total revenue ratio
2015	23.4 %	16.2 %	39.8 %
2019	24.1 %	16.9 %	41.2 %
2021	24.4 %	17.2 %	41.9 %
2023	23.0 %	17.0 %	40.1 %
2024	23.1 %	17.6 %	40.9 %

Between 2015 and 2024 the total ratio of taxes and social contributions to economic output rose by 1.1 points. The tax ratio fell by 0.3 points. The social contribution ratio rose by 1.4 points. In this framework the entire increase in the aggregate burden came from social contributions and none of it from taxes. [38] That finding has to be stated in both directions, because half of it is inconvenient for me.

German tax revenue rose from 673.3 billion euros in 2015 to 989.8 billion in 2025, an increase of 47.0 percent in nominal terms. Adjusted for the 29.0 percent rise in consumer prices, roughly 14.0 percent of real growth remains. That is genuinely more money for the state, driven by more employment, higher real incomes, profits, and consumption. It does not prove that the average individual pays a higher share of income in tax, and anybody using it that way is misusing it. [39] [41] Because on the income tax side there were real reliefs, and leaving them out would make this piece dishonest. The basic tax free allowance rose from 8,472 euros in 2015 to 12,096 euros in 2025, an increase of 42.8 percent, well ahead of the 29.0 percent price rise. The threshold where the 42 percent bracket begins rose from 52,882 to 68,481 euros, an increase of 29.5 percent, roughly matching inflation. Since 2016 the bracket thresholds have been shifted regularly to offset bracket creep. The solidarity surcharge has been abolished entirely since 2021 for roughly 90 percent of those who previously paid it, with a further 6.5 percent partly relieved, and its revenue fell from 15.9 billion euros in 2015 to 12.9 billion in 2025 despite much higher nominal incomes. Both value added tax rates are unchanged at 19 and 7 percent. The renewable energy levy disappeared from electricity bills in July 2022 and was replaced from the federal budget. [42] [43] Against that stands one genuinely new burden. The national emissions trading scheme started in 2021 at 25 euros per ton of carbon dioxide and reached 55 euros by 2025, with a corridor of 55 to 65 euros for 2026. Assuming full pass through and including value added tax, 55 euros per ton corresponds to roughly 15.3 cents per liter of petrol, roughly 17.3 cents per liter of diesel or heating oil, and roughly 1.3 cents per kilowatt hour of natural gas. [44] It is a steering levy whose proceeds fund climate and relief measures. For a household that burns fossil energy and cannot switch quickly, it remains a real additional payment.

The individual burden on labor, measured properly, tells the same story with a nasty twist at the end. The OECD calculates the share of income tax plus employee and employer contributions in total labor costs for standardized households.

Single, no children	2015	2024	2025
67 percent of average wage	45.3 %	44.0 %	45.5 %
100 percent of average wage	49.4 %	47.9 %	49.3 %
167 percent of average wage	51.3 %	49.1 %	50.0 %

Over the full decade the wedge for a single average earner moved by 0.1 points, downward. So the blanket claim that every model worker has been squeezed harder every year is false, and I am printing that even though it weakens a line I would otherwise like to use.

What is true is the level and the direction of travel. At 49.3 percent Germany ranked second among the 38 OECD members in 2025 for a single average earner. And the single year jump from 2024 to 2025 came to 1.34 points, which the OECD decomposes into 0.53 points of employee contributions, 0.52 points of employer contributions, and 0.28 points of income tax. In the most recent year, work was taxed substantially harder, and the larger part of that came from social contributions. [45]

Where the contribution rate goes from here

None of the above is the problem. The problem is the direction, and this part is no longer historical.

How little historical it is became clear while this piece was being written. The national association of statutory health and long term care insurers published the half year figures of the social long term care insurance on August 30, 2026, 3 days before this piece appeared. In the first half of 2026 revenue rose by 5.4 percent to 38.7 billion euros while spending rose by 11 percent to 39.5 billion euros, which produces a half year deficit of 770 million euros. From October onward, current revenue will no longer cover current benefits according to that projection. For the full year 2026 the association expects a deficit of 1.2 billion euros, and that figure already contains a federal loan of 3.2 billion euros which has to be repaid between 2029 and 2033. Without that loan the shortfall would be 4.4 billion euros. For 2027 the association puts the additional funding requirement at 10 billion euros, of which 7.5 billion is the projected deficit. Its chairman sums the situation up in one sentence that is better than anything I could write about it: “Care in Germany is living on credit.” [28] Cause and effect are undisputed here for once, and both already sit in the tables above. The cause is the rising number of people needing care, meaning the same age structure that shows at least 20.5 million people aged 67 and over for 2040. The effect is a spending dynamic running many times faster than the revenue dynamic in the current year. And the response is exactly the one this piece describes: first a loan, then a call for more federal money, and if neither is enough, a higher contribution rate.

How far spending is running ahead depends on which revenue measure you hold against it. Against contribution revenue, which rose 3.9 percent to 36.71 billion euros in the first half year, spending grew almost 3 times as fast. Against total revenue of 38.71 billion euros, which already contains half of the federal subsidy of 1.6 billion euros, it was a good 2 times as fast. Both comparisons are legitimate, and the second looks friendlier only because tax money is already inside it. [29] No decision by the federal government on additional federal funds or a higher contribution rate existed as this piece went to press. The ministerial draft of a care reorganization act dates from June 2026, the hearing of the associations took place on June 10, 2026, and a cabinet decision has been postponed repeatedly since. The ministry changed hands in July 2026. At the end of August 2026 the new federal health minister announced that he would not support the reduction of pension contributions for family carers contained in the draft, which is a political announcement rather than a change in the law. The Bundestag was in summer recess until early September, and a first reading is expected in autumn 2026 at the earliest. [30] That makes this case a textbook illustration of what I describe further down as a decision problem. The figures have been on the table for months, the draft has existed since June, and the fund expects a funding gap from October. What is missing here is not the analysis.

The mechanism of this piece is playing out in fast motion, and faster than I assumed while writing. My scenarios for 2040 assume an unchanged care contribution of 3.6 percent on the reform path. A system already dependent on loans and special subsidies in the current year does not make that assumption more likely. On the contrary, it makes it look optimistic.

The health finance commission published a middle scenario for the years ahead, and it is the most concrete official statement anybody in this debate currently has.

Year	Health spending growth	Supplementary rate needed to cover spending	Gap if the rate stays at 2.9 percent
2027	5.3 %	3.65 %	15.3 bn EUR
2028	5.5 %	3.91 %	21.5 bn EUR
2029	4.9 %	4.36 %	31.9 bn EUR
2030	4.7 %	4.69 %	40.4 bn EUR

The uncertainty band for the 2030 supplementary rate runs from 4.1 to 5.3 percent. The commission expects only about 3 percent annual revenue growth from 2027, largely stagnant employment numbers, and higher spending growth. Without lower spending or additional tax financing, the difference lands on the contribution rate. [31] Put that together with the official pension path, holding care and unemployment unchanged, and 2030 looks like this. Pension 20.1, unemployment 2.6, general health 14.6, supplementary health 4.69, care for parents 3.6, giving a total of 45.59 percent. Using the full health band the range is 45.0 to 46.2 percent. No increase in the care contribution is included in that figure, and nobody seriously expects the care contribution to hold.

For 2040 there is no legally fixed rate and there will not be one, so what follows is my own stress calculation with every assumption on the table. Pension at 21.2 percent, matching the official projection for 2039. Unemployment unchanged at 2.6 percent. General health unchanged at 14.6 percent. Care for parents at 3.6 percent unless stated otherwise.

Scenario 2040	Health supplement	Care	Total social contribution
Reform and stabilization, health flat after 2030	4.69 %	3.60 %	46.69 %
Lower observed dynamic continues, plus 0.3 points a year	7.69 %	3.60 %	49.69 %
Upper observed dynamic continues, plus 0.4 points a year	8.69 %	3.60 %	50.69 %
Hard stress, upper dynamic plus 1 point on care	8.69 %	4.60 %	51.69 %

The 0.3 to 0.4 points a year correspond to the annual dynamic the commission itself describes through 2030. Extending it to 2040 is not an official forecast and I am not presenting it as one. It is a warning calculation, and reforms, benefit cuts, higher tax subsidies, a broader assessment base, or productivity gains can all change the path.

Translated into a payslip, the numbers look like this. All euro amounts are in 2025 prices. The calculation assumes a real gross income of 60,000 euros for an employed parent and an assessment ceiling indexed accordingly, so that this income remains fully contributory in 2030 and 2040 as well.

Year and scenario	Employee share	Direct employee contribution	More than 2025
2025, weighted actual	21.145 %	12,687 EUR
2030, middle health scenario	22.795 %	13,677 EUR	+990 EUR
2040, reform and stabilization	23.345 %	14,007 EUR	+1,320 EUR
2040, lower dynamic continues	24.845 %	14,907 EUR	+2,220 EUR
2040, upper dynamic continues	25.345 %	15,207 EUR	+2,520 EUR
2040, hard stress	25.845 %	15,507 EUR	+2,820 EUR

A childless employee would add 0.6 percent of contributory gross under current law, meaning a further 360 euros a year at that salary. [46] And the employer carries the same additional amount on the shared components, which is the detail that makes this worse rather than better. Labor gets more expensive twice over. Take home pay falls and the cost of employing the person rises, at the same moment, from the same cause.

I am deliberately not adding a projected income tax figure on top of these. Tax rates, allowances, the solidarity surcharge, value added tax rates, assessment ceilings, the split between contributions and federal subsidies, the scope of insured benefits, copayments, the carbon price after it moves into the European system, and the participation rate are all politically variable and none of them is fixed to 2040. A combined number would look devastating on the page and would be methodologically worthless.

One more thing belongs here, because it is the piece the public debate keeps getting backwards. Federal money is not a way out of this. The regular federal subsidy to health insurance rose from 11.5 billion euros in 2015 to 14.5 billion a year from 2017, with extra tranches of 3.5 billion in 2020, 5 billion in 2021, 14 billion in 2022, and 4 billion split across 2025 and 2026 for hospital transformation costs. Its share of health insurance revenue nonetheless fell, from roughly 6 percent to roughly 5 percent, so contributions had to carry the growing remainder. The commission states plainly that without the recent short term federal measures of nearly 10 billion euros, a further jump of about 0.5 points in the supplementary rate would have been necessary for the start of 2026, and that the federal contributions paid on behalf of basic income recipients fall short of the actual cost by roughly 12 billion euros a year, with the difference financed by everybody else’s premiums.

On the pension side, federal money has accounted for roughly 30 percent of revenue for years. In 2025 federal subsidies to the general pension insurance came to roughly 93.2 billion euros, with a further 19.2 billion in contributions paid by the federal government for child raising periods. [8] None of that money falls from anywhere. Shifting a cost from a contribution to a tax does not reduce the burden on society by a single euro. It changes who pays and by which formula, and it is often chosen precisely because the new formula is harder to see.

That is the whole mechanism of this piece in one sentence. The state can only ever distribute what citizens and companies produce as taxes, contributions, prices, or future debt service.

What happens when the base that pays for it shrinks

I can put numbers on 2 movements that hollow out this base at the same time. A third one I cannot calculate responsibly, and that is precisely why it belongs in this picture.

The first is the industrial one. German passenger car production fell from 5,708,138 units in 2015 to 4,148,836 in 2025, a decline of 27.3 percent. New registrations in Germany fell from 3,206,042 to 2,857,591 over the same period, a decline of 10.9 percent. Employment in the automotive industry fell in 2025 alone from 772,949 to 731,928 people, minus 5.3 percent, with parts and accessories suppliers hit hardest at minus 11.1 percent. [75] [76] [77] The convenient explanation is that too many brands now share the same domestic cake. That explanation does not survive contact with the data. The five largest brands held roughly 53.4 percent of the German market in 2015 and roughly 52.7 percent in 2025, a shift of 0.7 points. German group brands together still held close to 70 percent of the domestic new car market in 2025. BYD, the brand that generated the loudest coverage, jumped from 2,891 registrations in 2024 to 23,306 in 2025 and thereby reached 0.82 percent of the market.

The real finding sits outside Germany. World passenger car production in 2025 came to 82.55 million units. China produced 29.92 million of them, roughly 36 percent. Germany as a production location produced 4.15 million, roughly 5 percent. German group brands produced 13.30 million cars worldwide and thereby held roughly 16.1 percent of world production, but their foreign production fell 3.9 percent in 2025, and their share of the Chinese market fell to 16.2 percent, down 2.7 points in a single year. [78] A country that keeps losing employment at one of its largest industrial employers is not losing prestige. It is losing contributors, and contributors are the only input the formula above accepts.

The second movement is the one people write angry columns about, and it is smaller and more ambiguous than those columns claim.

In 2025, 288,579 German nationals moved away and 191,890 moved in, producing a net loss of 96,689. Over 2015 to 2025 the cumulative net loss of German nationals comes to 780,543 people, on 2,760,785 departures against 1,980,242 arrivals. [47] Two qualifications belong here and both weaken the alarmist reading. In the German Emigration and Remigration Panel Study of 2018 and 2019, 62.2 percent of the emigrants surveyed were considering returning to Germany for a longer period or for good. The population research institute therefore deliberately speaks of circular mobility rather than a demonstrated permanent loss of skilled workers. In the same survey 75.5 percent of the emigrants held an academic qualification, against 24.7 percent in the non mobile comparison group. [52] That is a serious warning signal and it is not a proof that the departing group consists of top rate taxpayers.

Because it cannot be a proof. The migration statistics record no income. The tax statistics record no subsequent departure. There is no linked data set, so the sentence that the top 5 percent of taxpayers are leaving cannot be quantified from official German data in either direction.

What can be quantified is how concentrated the tax base is, and it is concentrated to a degree most people do not believe when they first read it. According to the finance ministry’s tax policy data collection, projected for the 2025 assessment year on the basis of the wage and income tax statistics, the top 1 percent of taxpayers carry 23.6 percent of wage and income tax revenue, the top 5 percent carry 43.6 percent, the top 10 percent carry 56.3 percent, and the bottom 50 percent carry 6.5 percent. [40] The top 5 percent begin at a total income of 144,932 euros, and jointly assessed couples count as one taxpayer.

A stress calculation makes the exposure visible without pretending to precision. If the 2025 net German emigration repeated unchanged every year to 2040, and if those leaving had the same income distribution as everybody else, the cumulative loss would correspond to roughly 2.6 percent of a static annual income tax base. If an implausibly high 20 percent of them came from the top 5 percent, the figure would be roughly 5.7 percent. Those are not forecasts, because they exclude income growth, replacement by other taxpayers, and behavioral change. They are the outer walls of the room.

The honest finding is therefore threefold. The tax base is extremely concentrated. German emigrants are above average in qualification. And the specific revenue loss has never been measured, so anybody converting it into a billion euro figure is inventing precision.

I have left artificial intelligence and robotics out of every calculation in this piece. I cannot responsibly quantify how many contributory jobs they will replace by 2040, how many new ones they will create, or how much productivity will rise as a result. If they displace more well paid work than they generate, my burden paths are too friendly. If the productivity gains arrive broadly, they can absorb part of the load. I do not know, and that is why no invented value went into the tables. My calculation does not contain this risk. That makes the outlook less reassuring rather than more.

Fraud, and the difference between a suspicion and a finding

I have been asked repeatedly in the past weeks whether this state is being defrauded. The answer requires three categories that get mixed constantly.

Confirmed fraud, for the purposes of this analysis, requires solid administrative or judicial proof of deliberate deception and unlawful receipt. An investigation is grounds to check and not proof. An overpayment or clawback can arise from a late notification, an administrative error, or changed circumstances, with no fraud involved at all.

For 2025, the federal employment agency reported 133,640 initiated proceedings or suspected cases of benefit abuse in the jointly run job centers. The federal government warned explicitly that this does not permit the conclusion that fraud or unlawful receipt was established in all of those cases. Job centers run by municipalities alone are not included in that figure, and no breakdown by nationality is possible.

For the narrowly defined category of organized gang benefit abuse, the jointly run job centers recorded 229 new proceedings in 2023, 421 in 2024, and 195 from January to May 2025. That definition captures mainly fictitious employment or fictitious self employment by non German EU citizens used to circumvent the benefit exclusion. Other organized constellations are not counted separately, and job centers do not always receive feedback on how criminal proceedings ended. [71] [72] Sanctions tell their own story and it is not the story either camp wants. In 2025 the job centers imposed roughly 461,400 benefit reductions. 85.5 percent of them rested on missed appointments. 31,000 concerned refusal to take up or continue work, training, or an integration measure. Across the full year roughly 224,100 employable recipients received at least one newly imposed reduction. In the average monthly stock, 0.9 percent of employable recipients were subject to at least one ongoing reduction. Those two measures must not be treated as the same thing. [73] A missed appointment is not fraud. But an average monthly affected rate of 0.9 percent is not evidence of firm enforcement either.

And here is the uncomfortable balance. Abuse exists, organized models are officially documented, and Germany does not possess a consolidated damage figure for 2025 across the employment agency, the municipal job centers, the social welfare offices, the family benefits offices, and the remaining carriers. Anyone turning 133,640 suspicions into 133,640 fraudsters is manipulating. Anyone concluding from the absent total that the problem does not exist is manipulating in the other direction.

That absent total is the same phenomenon as the absent nationality data. A web of social security systems spending 1.43 trillion euros a year cannot say what total damage benefit abuse causes, and is not organized in a way that would let it establish that figure.

What it costs to run the machine, and why nobody can tell you that either

The social budget contains a line most readers skip, and it is the second thing on that rest stop that made me sit still for a while.

Administration cost 51.189 billion euros in 2025. [1] That figure is not a social protection benefit at all. It is the cost of deciding, processing, checking, appealing, archiving, and paying out the benefits, spread across pension carriers, health funds, care funds, the employment agency, the job centers, the municipal welfare offices, and everything attached to them.

For scale, that single line is larger than the entire unemployment function of the social budget, which came to 47.562 billion euros. Germany spends more on running the welfare state than it spends on unemployment inside it.

One component of that is visible in isolation. Administrative spending in the basic income support system alone came to 6.788 billion euros in 2025, against payment entitlements of 46.653 billion. That is roughly 14.6 euros of administration for every 100 euros of benefit, and I am computing that ratio myself rather than quoting it, because nobody publishes it in that form.

I want to be fair about this figure before somebody weaponizes it. A welfare state needs people to run it, and supervising more than 5 million beneficiaries with legally reviewable individual decisions is not something anybody does casually. Every one of those decisions can be appealed, and the appeal itself has to be processed by someone. A cheap administration would be an arbitrary one, and arbitrary administration is the thing constitutional courts exist to prevent.

What I object to is not the existence of the cost. It is that between the euro collected and the euro delivered there is a distance nobody in public life wants to discuss, and that the distance has never been measured across the whole apparatus. I went looking for it. I found no published account comparable to the social budget that consolidates all administrative costs of the federal government, the states, the municipalities, and the social insurance funds, because those costs sit scattered across dozens of individual budgets, insurance carriers, and municipal accounts. I looked for that figure and could not find one I was willing to print.

I could write down an estimate here that sounds forceful and that nobody could check. I am leaving it out, because an invented number in a piece about bad arithmetic would be a special kind of self sabotage.

There is one more line worth naming. Interest spending by the German state rose 8.1 percent in 2025 against the previous year. That money buys nothing at all. It is the price of having spent money in earlier years, and every increase in it permanently reduces the room available for everything above.

Put the three together and the picture sharpens. The German system of social security reports spending of 1.43 trillion euros, of which 51.189 billion euros covers the administration recorded in the social budget. Alongside that sit further administrative costs of the state that are not consolidated in any comparable account, and rising interest spending. Then it declares that there is no scope for savings anywhere. In an overall system of that size, the claim that nothing can be cut is not a finding. It is a decision, and decisions can be justified. Somebody would only have to do it.

Where the belief in this system actually dies

I want to put the mechanism of collapse in one place, because it is not the dramatic one people expect.

Nothing about this process detonates. A pension fund does not run out of money on a Tuesday. What happens instead is that the burden on the people who still carry it rises, year after year, in increments each of which is individually defensible.

The contribution rate rises because the reserve is empty. The supplementary health rate rises because hospital spending rose 9.6 percent in one year. The care contribution rises because the number of people needing care rises. The assessment ceiling rises because that is easier to legislate than a structural reform. Each of those steps is justifiable on its own terms, and every one of them lands on the same person.

That person then does the arithmetic at the kitchen table, and at some point the arithmetic tells him that additional effort is not worth it. He works less, or he declines the extra contract, or he stops expanding, or he leaves. None of that is illegal, none of it requires organization, and no instrument exists against it, because there is nothing to seize except a willingness that departed years ago.

That is the collapse, and it does not look like a collapse. It looks like a very slow shrug.

What a serious reform would have to be able to measure

Criticism without a proposal is cheap, so here is what I would want before anybody drafts another law, and none of it requires agreeing with a single opinion of mine.

It would need consistent outcome metrics rather than activity metrics. At present a success in labor market policy is frequently reported as a course entry, a measure taken up, or a case leaving the register. None of those tell you whether a person is supporting himself. The metrics that matter are employment after 6, 12, and 24 months, hours worked, contributory income, and the end of dependency, and they need to be published by cohort rather than aggregated into a monthly headline that hides everything.

It would need clean register matching, done lawfully and with an audit trail. Whether somebody is actually resident, actually employed, actually earning what he declares, whether a child actually attends school, and whether parallel entitlements exist in more than one system are all questions whose underlying data largely already exist. They sit with different authorities, and they are used only for their designated purpose rather than brought together into any continuous check of outcomes. I am aware that this is exactly where data protection concerns begin, and I take them seriously, because a state that can match everything against everything is a different kind of state. The answer to that is purpose limitation and logging, rather than a permanent inability to know what is happening.

It would need employment to start early rather than after the training pipeline finishes. Language alongside work instead of before it, recognition of qualifications running in parallel instead of sequentially, and no multi year waiting loop before the first job. The Polish comparison is not a recommendation to copy Poland, because Poland bought its employment rate partly with underemployment. It is evidence that the sequence can be reversed and that the reversal produces a very large effect.

It would need clear activity obligations with equally clear protection. Whoever can work should have to demonstrate reasonable work, qualification, or verifiable care responsibilities. Children, the sick, people with disabilities, and those caring for relatives need protection rather than general suspicion, and the difference between those two groups has to be established by a person looking at a case rather than by a slogan.

It would need enforcement aimed at the profiteers rather than at the weakest participant. In the organized constellations that are officially documented, the economic winners are rarely the benefit recipients. They are the intermediaries, the fictitious employers, and the landlords renting substandard property at rates the state reimburses. Those are the parties with assets, and those are the parties an enforcement apparatus can actually reach.

And it would need a published loss account. Suspicion, confirmed fraud, clawback ordered, clawback enforced, and money actually recovered, separated by program and by carrier. Germany publishes no consolidated series across carriers covering suspicion, confirmed fraud, clawback ordered, clawback enforced, and money actually recovered, which is why every public argument about abuse consists of two sides shouting numbers that measure different things.

None of this is radical. All of it is measurement. A system this size that cannot measure itself will keep producing laws that look reasonable on the day of the press conference and land somewhere nobody predicted 3 years later.

Why the next election changes nothing about this arithmetic

Ahead of the coming state elections in Saxony-Anhalt on September 6, 2026 and in Mecklenburg-Vorpommern on September 20, 2026, I will say this as plainly as both sides will dislike hearing it. [79] [80] Voting for the AfD cannot solve this problem, and it will not solve it. That is my opinion, but it rests on the figures in this piece. Anybody who believes the arithmetic works out as soon as you cut spending on immigrants has not read the tables. The entire SGB II payment entitlements amount to 3.3 percent of the social budget. The Ukrainian share of 0.42 percent is already contained in that figure. Even under a purely arithmetical extreme assumption that set all SGB II payment entitlements to zero, roughly 96.7 percent of the social budget would remain unchanged. The 70 percent for old age and sickness would be entirely untouched, and so would the civil service pension system. The actual financing problem would not be solved.

The federal government of CDU, CSU, and SPD cannot solve it either, and that is the second uncomfortable part. Inside a coalition every party has its own convictions, its own advisers, and its own findings. Those collide, and what gets implemented in the end is whatever everybody can still agree on, meaning almost nothing. A country that has known the same figures for years and still touches no structure does not have a knowledge problem. It has a decision problem.

What I consider unavoidable instead, and this is opinion as well, fits into 3 sentences. The administrative apparatus must not keep growing, neither here nor in Brussels, although for the European level I have expressly verified no reliable total and therefore quote none. Pressure belongs on the manufacturers when it comes to medicine prices, and which instruments actually work has to be decided against measurable prices, spending, and treatment outcomes. And the red pen belongs everywhere that spending is not attached to a measurable result.

I am not the guru with the solution, and I do not claim to be. But anybody who reads the figures in this piece and can interpret them will not get past the problem sitting in these tables.

I am convinced that no coming government will solve this structural arithmetic. Any effective solution would have to limit benefits immediately, redistribute contributions, cut administration, and touch protected entitlements. Every coalition will fail at exactly that point, long before it fails at the mathematics. It can shift contribution rates, raise subsidies, and take on new debt. It cannot legislate away the age structure, the entitlements already earned, or the contributors missing by 2040. To be blunt about it, I believe this arithmetic is going to go wrong.

What I actually think

Up to this point I have marked official data, my own calculations, and the assessments derived from them as what they are. What follows is my personal conclusion, plainly marked, and you should treat it as exactly that.

I think this system is finished in its present form. Not doomed in some vague future, finished now, in the sense that the promises currently on the books cannot be kept on the base that is currently paying for them, and the arithmetic is visible to anybody with access to the figures.

I think a welfare state is one of the genuine achievements of this country and I want it to exist. Somebody who becomes ill, loses work, raises a child alone, or grows old must not fall through into nothing. That is decent and it is also economically sound, because a society that catches people in crisis preserves their capacity to work.

And precisely because I want it to exist, I refuse to be sentimental about how it is run. My position is one sentence long. Everyone who can work must work, and I mean everyone without exception. Not as a punishment and not as a slogan, but because a system financed by a shrinking group of contributors has no capacity left to carry anybody who could carry himself. Children, the ill, the disabled, and those caring for relatives need protection rather than general suspicion. Everybody else needs an obligation, and that obligation needs to be checked rather than assumed.

I also think this system is being misused, and I am stating that as my assessment rather than as a finding, because as shown above Germany does not measure the total. I have watched too many constellations professionally to believe that 133,640 suspected cases in 2025 and 421 newly recorded organized abuse proceedings in 2024 represent the full picture of a system whose social budget covers 1.43 trillion euros and whose carriers cannot consolidate their own loss figures. I could be wrong about the scale. Nobody can prove me wrong, and that is the actual problem.

Two demands follow from that, and both are my opinion rather than a finding. The first is that the number of civil servants must not keep rising, because every appointment creates a pension commitment across decades that no later budget can vote away. The second is that pension payments in the public service belong frozen, at least above a threshold that would have to be discussed seriously. If this country loads its working population more heavily year after year, it cannot at the same time exempt a group whose payments rose 5.4 percent in a single year and whose pension entitlements are financed out of current public budgets without any individual pension contributions being paid for those entitlements.

Here I have to correct myself, because I had this wrong in my head for a long time. It is not true that civil servants never paid anything in. During the civil service relationship they are in principle exempt from statutory pension insurance, so no pension contribution flows for that period of service. Many of them did pay contributions before their appointment, during insured training or employment. Statutory pension entitlements can arise from that, and under section 55 of the civil service pension act they can be offset against pension payments where the stated conditions apply. Anybody leaving service without a pension entitlement is in principle insured retroactively, unless a retirement allowance rule or another exception applies, and the details differ between the federal government and the states. The version of my criticism that holds is therefore narrower and still lands: no individual contribution is paid for the pension entitlement itself. It follows from the employment relationship and is financed out of public budgets. [9] [10] [11] [12] [13] [14] One more word on the debate this country is conducting in parallel, because it shows the problem in its purest form. The argument is about retirement at 63.

Even the name is wrong, and that is symptomatic. The old age pension for the particularly long term insured was introduced on July 1, 2014, and the reduced age threshold of 63 applied exclusively to insured people born before the turn of the year 1952 to 1953. Since 2016 it has risen by 2 months with every birth cohort. Anybody born in 1964 or later retires without deductions at 65 after 45 contribution years, and the standard retirement age is 67. With 35 contribution years you can still get out at 63, but then with a permanent deduction of up to 14.4 percent. [81] [82] [83] A country that has spent more than 10 years arguing about a rule that in the form under discussion no longer exists for the vast majority does not have a pension problem at this point. It has a debate problem.

And now my opinion on it, expressly as opinion. That reduction should never have been decided. Life expectancy has been rising for decades, many people stay healthy considerably longer, and among the people I know there are self employed men and women working at 85 because they can and because they want to. Anybody who lowers the retirement threshold in such a society, while over the same period the number of people aged 67 and over climbs from 16.7 million toward at least 20.5 million, simply has not done the arithmetic. I place great value on logic and on systems that carry themselves, and at this point this system does not. Somebody could not do the math from the very beginning, and it would be time to admit that mistake.

I deliberately print no figure for how much longer a 65 year old lives on average today, because I did not check it against the life tables for this piece. The direction alone carries the statement.

And the sharpest thing I have to say is not about anybody receiving money. It is about the group deciding. A country that discusses the statutory pension every day and the civil service pension almost never, while the second pays 2.64 times as much per head on a pure gross comparison from pure tax revenue, is not conducting an honest debate. Whether that is deliberate or merely convenient I cannot say, and I am not going to speculate about the motives of people I do not know.

What I will say is that I no longer believe it is intent. Intent requires a plan, and a plan requires that somebody looked at the whole picture. That is exactly what is missing. Every single measure is defensible in isolation. Compulsory coverage is defensible against old age poverty. Higher contributions are defensible against deficits. Full benefit access for war refugees is defensible against destitution. Language courses before work are defensible against dead end employment. And the sum of these individually defensible measures is a country dismantling the group that finances all of them.

Nailing it down

One question I cannot shake. Why does no chancellor stand up and tell this country that it has a problem of this magnitude? The central figures are officially published, and even the decisive data gaps are officially confirmed. There is no shortage of warning signals. What is missing is the plain sentence. That is what I am denouncing, not the sheer size of the spending but the absent disclosure: honesty about the figures, honesty about the data gaps, and honesty about where this country will very probably end up as it keeps aging.

I am dating these statements so they can be held against reality. The following 5 paragraphs contain different kinds of prediction. The pension path and the supplementary health rate follow from official projections. The threshold years for health and care spending follow from the scenario calculations disclosed above. The statements on assessment ceilings, working age population, automotive employment, and statistics that will still be missing are my personal predictions without a model path of their own. All of them are worded so that they can be checked later.

The contribution rate to the statutory pension insurance will be above 19.8 percent by 2029, and the assessment ceilings will have risen faster than average wages in at least 3 of the intervening years.

Combined health and care spending will exceed 550 billion euros in nominal terms in 2031 on the middle burden path, in 2030 if the historical rate continues, and only in 2034 on the dampened 3.0 percent path. The average supplementary health insurance rate will pass 3.5 percent before 2030.

The population aged 20 to 66 will be below 49 million by 2032, and no migration policy adopted in the meantime will prevent it.

German automotive employment will be below 650,000 by 2032, and the discussion will still be about brands rather than about production location.

Germany will still not publish a consolidated damage figure for benefit abuse in 2030, and still not publish social spending by nationality for the insurance systems, and the reason given will still be that the data are not collected in that form.

I would rather have been wrong about this. I have checked it three times, in two different frameworks, against the primary publications rather than against somebody’s summary of them, and it holds. Read the tables for yourself. Every source is listed below, every calculation of mine is marked as mine, and every gap is named as a gap. Then decide for yourself what you want to do with it.

And one more prediction, for which I need no official source. I am going to leave this country. Not tomorrow, and not in anger, but within the next few years. It hurts to leave a country in the last decade of your life that shaped that life across decades. The people are here, the memories are here, and everything that cannot be written into a spreadsheet is here. I am not leaving because I hate this country. I am leaving because I no longer trust the arithmetic by which it distributes its future.

Why I am going can be said soberly. I was born in 1970, I lived through the 80s and the 90s, and the period around 2000 was still all right. What has grown since then is a bubble of bad mood, absent prospects, and the settled impression that work no longer pays.

And then comes the part where I almost turn cheerful. Out of whatever is left to a working person after taxes and social contributions, that person cheerfully keeps paying. Value added tax on more or less everything. Energy tax on the fuel, electricity tax on the electricity, and the carbon price on the heating and the tank. Motor vehicle tax on the car bought with money that was already taxed, insurance tax on the insurance for that car, and property tax for the roof over the head. Tobacco tax, beer tax, alcohol tax, and coffee tax, the last of these levied since 1949 and in Europe still collected in only a handful of countries. Air travel tax, in case that person would like to leave the whole thing behind for a few days. And then, my personal favorite, the sparkling wine tax.

That one has to be savored. The Reichstag passed it in 1902 so that Kaiser Wilhelm II could pay for his war fleet. The fleet has been gone for more than 100 years. The tax is still there, it amounts to 1.02 euros per bottle, and it brings the federal government roughly 400 million euros a year. Anybody raising a glass in this country is still financing ships that have long been sitting on the seabed. [84] [85] If somebody finds a more compact description of this country, I will gladly take it.

There is no sugar tax here yet, and that is precisely what is about to change. The health finance commission recommended a levy on sweetened drinks staggered by sugar content, roughly 26 cents per liter above 5 grams per 100 milliliters and roughly 32 cents above 8 grams. The federal cabinet adopted the benchmark figures for the 2027 federal budget on April 29, 2026, and those list the sugar levy from 2028 onward as a consolidation measure. The act on stabilizing contribution rates in the statutory health insurance, adopted the same day, additionally records that the federal government will introduce it through a legislative procedure. The exact thresholds are still being negotiated. [86] [87] And then you can watch the thing grow. At the end of August 2026 a benchmark paper from the finance ministry became known that would also have covered sugar free drinks, meaning light, zero, oat drinks, and beer mixes. After the protest it was described as an internal working draft. Meanwhile the expected revenue circulates simultaneously as 650 million euros from the coalition agreement, roughly 2 billion from the agriculture ministry, and nearly 4 billion from the calculations of the industry associations. A levy whose estimated revenue varies by a factor of 6 depending on who you ask is obviously not a prevention instrument. [88] Do you notice what is happening here? A levy justified by health protection appears in the budget benchmarks under consolidation measures. It starts as steering, and it ends as a revenue source. The sparkling wine tax took 124 years to travel that road, and nobody has abolished it since.

The status of this levy is in motion, and that belongs in the record. The benchmark paper was one week old as this piece went to press, the debate about including sugar free drinks was still running, and the thresholds are not fixed. The legal and procedural position of August 31, 2026 underlies everything in this section. If something has been decided by the time of publication, that changes the direction not at all, only the decimal place.

I feel like I am living inside a completely surreal matrix in this country. It is a farm with millions of hamsters, each in its own wheel, each running until it drops, and above the hall hangs a sign reading “solidarity.” For me that is no longer a tolerable condition. God did not put me on this earth to pay sparkling wine tax for a war fleet that sank more than 100 years ago and, from 2028, sugar tax on a lemonade on top of it.

The decision was made by a spreadsheet rather than by a mood. When everything lay in front of me on the table, what I felt was no longer concern. It was fear, which is a word I use precisely and almost never. Fear because I can read the direction of this arithmetic, and because the years in which it could have been corrected quietly are over.

I will never retire, at least not in the sense of stopping work. I could not imagine it at all. In the United States people work at supermarket checkouts at 80. I am expressly not glorifying that. Life is no picnic, and work is part of life as far as I am concerned.

But work has to pay, and that is exactly the point. For me, additional work in this country no longer pays. So I am scaling back my working hours, jumping out of the hamster wheel, and thinking hard about every assignment before I take it. I am not going to ruin my health for mistakes that were made in the past. That is not a fit of defiance, it is the consequence of my own arithmetic.

This is not a figure from a table but my personal conclusion: anybody not born with a silver spoon fights a battle in this country for their entire life that they will never win. A welfare state in which work no longer pays destroys, in the end, exactly the foundation that carries it.

In 10 years I will pull this text out again. I am not looking forward to it.

Afterword

I thought about it for a long time before publishing this in the form you have just read.

Politics repels me and is otherwise a matter of indifference to me. I have never voted, and I am not going to vote. This site stands for forensics and for science, not for party politics. But numbers are my life, I live, think, and feel in numbers, and for years the same question has occupied me. What comes in, what goes out, and why do so many people believe of all things that a single party could solve it?

None of these parties solves it. Just look at the figures for a moment. Who is supposed to pull a cart out of the water that weighs 1,400 billion? The matter is settled for good.

What would have to be done fits into 4 sentences, and all 4 are expressly my opinion. The standard retirement age rises to 70, and that is not up for negotiation. Pension payments in the public service are not raised until 2040, and that is not up for negotiation either. Whoever can work and still does not work receives nothing. And manufacturers of medicines are reimbursed only 70 percent of the cost, or the product loses its reimbursability.

On the third of those sentences I have to correct myself, in a way that only confirms my own point. On November 5, 2019 the Federal Constitutional Court ruled that human dignity under article 1 in conjunction with article 20 of the constitution guarantees everybody a subsistence minimum compatible with dignity, and that this dignity is not lost through supposedly undignified behavior. Complete withdrawal is therefore excluded. What is permitted is the principle of subsidiarity, reasonable obligations to cooperate, and benefit reductions, under that ruling provisionally capped at 30 percent. [89] [90] But that is exactly my point. Under today’s form of government none of it can be pushed through, because every measure has to be talked palatable for every group before it even reaches a cabinet meeting.

And that is precisely why I am publishing this text anyway. For everybody who wants to know how things really look here. In my own hard words: this country is in an absolutely shit state. No miracle will change that, no single party, and no coalition in whatever combination of colors.

It only works through a drastic reduction in spending, and there is no other way.

Anybody who wants to vote for the AfD can do that. It will change nothing about this arithmetic, not the slightest thing.

One thing I would like to add to that, and it matters to me. In the hospitals, in the clinics, and in the care homes of this country it is long since not only the Maria from next door who works there. In 2025 roughly 20 percent of all employees in nursing occupations held a foreign nationality, meaning 352,600 out of 1.76 million people, and 10 years earlier it was roughly 7 percent. In hospitals and clinics the share is 17 percent, in outpatient and residential care 24 percent. Since 2022 the entire growth in employment in nursing has come exclusively from foreign staff, and the number of employees with German nationality in elderly care has recently even fallen. In the Munich metropolitan region and in Frankfurt am Main more than half of the workforce in elderly care now holds a foreign nationality. [91] [92] This country needs these people by now. Deporting a person who works here and is integrated is therefore not merely stupid. It ensures that the collapse this country is facing arrives earlier rather than later.

A postscript to the afterword, written on the day of publication

This piece was finished, the files were ready to upload, and then a message arrived from an acquaintance who emigrated to Malta 2 years ago. My buttered pretzel fell out of my hand, and for once that is not a figure of speech.

The Initiative for a New Social Market Economy has analyzed the federal government’s staffing plans. The draft budget for 2027 contains 310,487 posts. Against 2024, the last full budget year of the previous government, that is 12,347 more posts, an increase of 4.1 percent. For 2026 the figure is 305,753 posts, meaning 7,613 more than in 2024. The entire increase between 2024 and 2026 falls on established civil service posts, 9,008 additional ones and therefore a rise of 4.6 percent, and a further 4,316 established posts are planned for 2027. The number of ordinary employee posts fell slightly over the same period. [93] [94]

Federal personnel spending rose from 44.97 billion euros in 2024 to 48 billion in the current year and is set to reach 51.62 billion in 2027. Posts grow by 1.6 percent in 2027 while spending grows by 7.5 percent. The longer view makes the direction plainer still. In 2017 the federal government had 257,069 posts, in 2026 it has 305,753, an increase of 18.9 percent. Established civil service posts grew by 35.6 percent over that period, employee posts fell by 5.3 percent, and planned personnel spending rose by 50.1 percent. [93] [94]

And what does the coalition agreement say? It says the workforce will be cut by at least 8 percent by 2029. That would mean roughly 274,300 posts. Starting from the level planned for 2027, roughly 36,200 posts would have to disappear, around 18,000 in each of the years 2028 and 2029. [93]

A word on the source, because I demand this of others and therefore say it about myself. The analysis comes from a business aligned initiative that holds a position on this question. The underlying figures come from the federal government’s own draft budget, so they are not that initiative’s opinion but this country’s plan.

Read that against the chapter above in which I wrote that the number of civil servants must not keep rising, because every appointment creates a pension commitment across decades that no later budget can vote away. While I was writing that sentence, the 4,316 additional posts for 2027 were already in the plan. Every single one of them is a pension falling due from roughly 2060 onward, and nobody deciding this today will still be in office then.

Anybody inflating the state apparatus in this situation has either learned nothing or has not read their own figures. A third possibility does not occur to me, however hard I try.

And that is why I am revising, in the last paragraph of my own text, a statement from the chapter on the election. I wrote there that voting for the AfD cannot solve this problem. I stand by that, because arithmetic on this scale is not changed by any election result. What I would no longer let stand is the conclusion that it is therefore all the same. A party does not have to be able to solve a problem of 1.43 trillion euros in order to stop somebody from planning 12,347 additional posts in precisely this situation. There is a difference between healing and stopping the cutting, and the second is by far the less demanding of the two skills.

Whether they can manage that, I do not know, and personally I expect nothing from it. At this point I am hoping only very modestly that somewhere in this country there is a person who can operate a pocket calculator without first appointing a commission to establish whether they are the competent authority for it, and without evaluating the exercise afterward in a benchmark paper. That is expressly my personal assessment and not a voting recommendation, and as written above, I do not vote at all.

And now comes the part I have described nowhere in this entire piece, although it sits above everything else in it.

It is not only the acquaintance in Malta, a lawyer, who asks. The friend in the United States asks as well, and so do 3 acquaintances now living in Cyprus and audibly breathing out there. The question is the same every time, and it consists of 3 words: “Just come over.”

That is not a private detail, it is the mechanism itself. Whoever leaves lowers the cost for the next one. He knows the tax adviser, the bank, the school, the authority, and every pitfall along the way, and he passes all of it on. The first leaves out of conviction, the second leaves because it worked for the first, and the fifth leaves because half of his circle is gone anyway. Emigration therefore does not run in a straight line but in waves, and every wave shortens the path for the next. That is my reading and not a finding, but it follows the same logic I used for everything else above.

Now hold that against the financing table from the start of this piece. Employer contributions of 512.806 billion euros, insured persons’ contributions of 461.096 billion euros. Together that is 973.902 billion euros and therefore 65.4 percent of the entire financing, and both halves hang on one and the same thing, namely an employment relationship inside Germany. [1]

The contribution rate itself is split almost evenly in the large branches, roughly half each for employer and employee, with the familiar exceptions for the childless surcharge in long term care insurance, the special rule in Saxony, and accident insurance, which the employer carries alone. [32] [33]

And this is exactly where the question sits that I have asked nowhere in this piece. What actually happens when it is not the employee who leaves but the employer? Then it is not 50 percent that disappears but both halves at once. Along with the wage tax on that job, the value added tax on whatever was produced and sold there, the municipality’s trade tax, and the consumption of everybody who worked there. An employee who emigrates takes one contributor along. A company that relocates or closes takes along everybody who worked there.

Every scenario in this text assumes that the base shrinks slowly, because people are getting older and fewer children were born. None of them assumes that employers withdraw that base in blocks. My figures are therefore friendlier still than they should be, and I would rather say that myself than have it said to me later.

That leaves one single quantity at the end which appears in none of my tables, because it cannot be entered into a balance sheet. That quantity is simply time itself.

Contribution rates can be raised, subsidies can be voted through, and debt can be taken on. Commissions can be appointed, benchmarks can be agreed, and legislation can be postponed. The people who will be 67 in 2040 have already been born. The children who would have to be paying in by 2040 have not. And the years in which this country could still have turned either of those around are not running alongside, they are running out.

That is exactly why I dated this text.

Calculation appendix

So that every figure marked as my own calculation can be checked, here are the formulas and input values used.

Inflation adjustment, consumer price index 94.5 in 2015 and 121.9 in 2025: a 2015 value in 2025 prices equals the 2015 value times 121.9 divided by 94.5, meaning times 1.289947.

Average annual growth of combined health and care insurance spending 2015 to 2025: 426.22 divided by 242.68, raised to the power of 1 divided by 10, minus 1, giving 5.7937 percent.

Spending in 2040: 426.22 times the quantity 1 plus the growth rate, raised to the power of 15.

Conversion into 2025 prices: the nominal 2040 value divided by 1.02 raised to the power of 15. The assumed 2.0 percent inflation is a calculation anchor rather than an inflation forecast.

Burden indicator: real combined health and care spending divided by the population aged 20 to 66. The 2025 starting value uses 426.22 billion euros and 51.2 million people.

Employee contribution in the model: contributory income times the employee contribution rate. For 2015 that is 50,000 euros times 20.155 percent for a parent and times 20.405 percent for a childless employee. For 2025 it is 64,497.35 euros times 21.145 and 21.745 percent respectively.

Administrative share in SGB II: 6.788 divided by 46.653 times 100, giving 14.55 euros of administration per 100 euros of payment entitlement.

Conversion of the Polish family benefit: 800 zloty times 12 divided by 4.2397 zloty per euro, giving 2,264.31 euros per year at the 2025 average rate published by the European Central Bank.

Counter calculation on the employment gap: 28 percentage points times roughly 850,000 people of working age gives roughly 238,000 people. The 850,000 follow from 1,327,268 displaced persons minus 357,158 minors and minus 117,844 people aged 65 and over.

Emigration stress calculation: the inputs are a net German migration loss of 96,689 people in 2025, an employment share of 73.6 percent among emigrants taken from the panel study sample, roughly 41.2 million taxpayer units as a rough order of magnitude from the income tax statistics, and 15 projection years from 2026 to 2040. The tax distribution comes from the finance ministry projection for the 2025 assessment year. The annual loss share equals 96,689 times 73.6 percent divided by 41.2 million, multiplied by the quantity 8.72 times p plus 0.594 times the quantity 1 minus p, where p is the unknown share of emigrants belonging to the top 5 percent. The factor 8.72 follows from a 43.6 percent revenue share divided by a 5 percent share of persons, and the factor 0.594 from 56.4 divided by 95. Cumulated over 15 years this gives roughly 2.6 percent of a static annual tax base at an identical income distribution, and roughly 5.7 percent if 20 percent of those leaving come from the top 5 percent.

World production share of the German group brands: 13.30 divided by 82.55 times 100, giving 16.11 percent.

This piece is based on publicly available data as of August 2026. Every figure is given with its source and reporting period. Where a figure has been estimated or approximated, that is indicated in the text. Passages identified as personal opinion are personal opinion and not established fact.

References

Every source carries a quality grade. P means an official primary source, S a secondary source, and E a calculation of my own based on the input values named. All internet sources were retrieved between August 28 and August 31, 2026. Titles of German language sources are given in the original.

Social budget, public finances, and the breakdown by function

• [1] P, Bundesministerium für Arbeit und Soziales (Federal Ministry of Labor and Social Affairs), Sozialbudget 2025, data status May 2026, 2025 values marked as estimates, S. 8, Tabelle I-2, Leistungen und Finanzierung, sowie S. 22, Tabelle III-2, Leistungen nach Funktionen und Institutionen. https://www.bmas.de/SharedDocs/Downloads/DE/Publikationen/a230-26-sozialbudget-2025.pdf?__blob=publicationFile&v=2
• [2] P, Statistisches Bundesamt (Federal Statistical Office), Staatsdefizit, Einnahmen und Ausgaben des Staates 2025, press release 060 of February 25, 2026. https://www.destatis.de/DE/Presse/Pressemitteilungen/2026/02/PD26_060_813.html
• [3] P, Statistisches Bundesamt (Federal Statistical Office), Monetäre Sozialleistungen des Staates 2025 um 5,9 Prozent gestiegen, press release 179 of May 19, 2026. https://www.destatis.de/DE/Presse/Pressemitteilungen/2026/05/PD26_179_813.html
• [4] P, Statistisches Bundesamt (Federal Statistical Office), Bevölkerung Deutschlands am Jahresende 2025, press release 203. https://www.destatis.de/DE/Presse/Pressemitteilungen/2026/06/PD26_203_124.html
• [5] E, own calculation of the per capita, per day, and per second values from [1] und [4], method in the calculation appendix.

Civil service pensions and the statutory pension insurance

• [6] P, Statistisches Bundesamt (Federal Statistical Office), 0,9 Prozent mehr Pensionärinnen und Pensionäre im öffentlichen Dienst im Jahr 2025, press release 455 of December 17, 2025. https://www.destatis.de/DE/Presse/Pressemitteilungen/2025/12/PD25_455_742.html
• [7] P, Deutsche Rentenversicherung (German Pension Insurance), Rentenversicherung in Zahlen, S. 34 bis 35, pension stock at December 31, 2024, and S. 70 bis 71, gross pension amounts by pension type. https://www.deutsche-rentenversicherung.de/SharedDocs/Downloads/DE/Statistiken-und-Berichte/statistikpublikationen/rv_in_zahlen.pdf?__blob=publicationFile
• [8] P, Deutsche Rentenversicherung (German Pension Insurance), Kennzahlen zur Finanzentwicklung der gesetzlichen Rentenversicherung, annual result 2025. https://www.deutsche-rentenversicherung.de/DRV/DE/Experten/Zahlen-und-Fakten/Kennzahlen-zur-Finanzentwicklung/kennzahlen-zur-finanzentwicklung_node.html
• [9] P, section 5 of the Sixth Book of the German Social Code, exemption from compulsory insurance. https://www.gesetze-im-internet.de/sgb_6/__5.html
• [10] P, section 8 of the Sixth Book of the German Social Code, retroactive insurance. https://www.gesetze-im-internet.de/sgb_6/__8.html
• [11] P, section 181 of the Sixth Book of the German Social Code, Berechnung und Tragung der Beiträge bei retroactive insurance. https://www.gesetze-im-internet.de/sgb_6/__181.html
• [12] P, section 55 of the German civil service pension act, concurrence of civil service pensions with statutory pensions. https://www.gesetze-im-internet.de/beamtvg/__55.html
• [13] P, general administrative regulation on the civil service pension act, retroactive insurance and pension surcharge. https://www.verwaltungsvorschriften-im-internet.de/bsvwvbund_11022021_D4303011423.htm
• [14] S, Erläuterung zum Zusammentreffen von Versorgungsbezügen und Renten, supplementing [12]. https://beamtenversorgung-online.de/bv_online_zusammentreffen_von_versorgung_und_renten The retirement allowance rules of the federal government and the states differ from one another.
• [15] P, Deutscher Bundestag (German Bundestag), printed paper 21/1040, Achter Versorgungsbericht der Bundesregierung, table IV-1. https://dserver.bundestag.de/btd/21/010/2101040.pdf
• [16] E, own calculation of the quotient 3.416 geteilt durch 1.296 equals 2.64, from [6] und [7].

Basic income support, child benefit, and asylum seeker benefits

• [17] P, Deutscher Bundestag (German Bundestag), printed paper 21/6855 of July 2, 2026, Entwicklung der Ausgaben im SGB II in den Jahren 2025 und 2026, data status June 2026. https://dserver.bundestag.de/btd/21/068/2106855.pdf
• [18] P, Bundesagentur für Arbeit (Federal Employment Agency), Kindergeld und Kinderzuschlag, Jahreszahlen 2025. https://statistik.arbeitsagentur.de/Statistikdaten/Detail/202512/famka/famka-jz/famka-jz-d-0-202512-pdf.pdf?__blob=publicationFile&v=1
• [19] P, Bundesagentur für Arbeit (Federal Employment Agency), Familien im Blick, die Bilanz der Familienkasse 2025, press release 4 of January 19, 2026. https://www.arbeitsagentur.de/presse/2026-04-familien-im-blick-die-bilanz-der-familienkasse-2025
• [20] P, Statistisches Bundesamt (Federal Statistical Office), Asylbewerberleistungen, gross spending, revenue, and net spending, status July 29, 2026. https://www.destatis.de/DE/Themen/Gesellschaft-Umwelt/Soziales/Asylbewerberleistungen/Tabellen/4-2-zv-il-bruttoausg-einn-netto.html
• [21] P, Statistisches Bundesamt (Federal Statistical Office), Asylbewerberleistungen 2024, number of recipients down 10 percent, press release 328. https://www.destatis.de/DE/Presse/Pressemitteilungen/2025/09/PD25_328_222.html
• [22] P, Statistisches Bundesamt (Federal Statistical Office), Sozialhilfeausgaben 2025 um 6,1 Prozent gestiegen, press release 300 of August 21, 2026. https://www.destatis.de/DE/Presse/Pressemitteilungen/2026/08/PD26_300_221.html
• [23] P, Statistisches Bundesamt (Federal Statistical Office), 1,3 Millionen Empfängerinnen und Empfänger von Grundsicherung im Alter und bei Erwerbsminderung im Dezember 2025, press release 104 of March 30, 2026. https://www.destatis.de/DE/Presse/Pressemitteilungen/2026/03/PD26_104_228.html
• [24] P, Deutscher Bundestag (German Bundestag), printed paper 21/4736 of March 13, 2026, no statutory health insurance or asylum seeker benefit spending by nationality. https://dserver.bundestag.de/btd/21/047/2104736.pdf

Health and long term care insurance

• [25] P, Bundesministerium für Gesundheit (Federal Ministry of Health), Finanzentwicklung der gesetzlichen Krankenversicherung im Jahr 2025, preliminary accounting result. https://www.bundesgesundheitsministerium.de/presse/pressemitteilungen/finanzentwicklung-gkv-2025
• [26] P, GKV-Spitzenverband (National Association of Statutory Health Insurance Funds), GKV-Kennzahlen. https://www.gkv-spitzenverband.de/presse/zahlen_und_grafiken/gkv_kennzahlen/gkv_kennzahlen.jsp
• [27] P, Bundesministerium für Gesundheit (Federal Ministry of Health), Finanzentwicklung der sozialen Pflegeversicherung 2025. https://www.bundesgesundheitsministerium.de/fileadmin/Dateien/3_Downloads/Statistiken/Pflegeversicherung/Finanzentwicklung/03-Finanzentwicklung-der-sozialen-Pflegeversicherung__2025.pdf
• [28] P, GKV-Spitzenverband (National Association of Statutory Health Insurance Funds), Die Pflege lebt auf Pump, press release of August 30, 2026, half year figures 2026 of the social long term care insurance and funding requirement 2027, retrieved August 31, 2026. https://www.gkv-spitzenverband.de/gkv_spitzenverband/presse/pressemitteilungen_und_statements/pressemitteilung_2270740.jsp
• [29] S, Deutsche Handwerks Zeitung, Pflegeversicherung droht 2027 ein Milliarden-Defizit, 31. August 2026, breakdown of the half year figures into contribution revenue and total revenue. https://www.deutsche-handwerks-zeitung.de/pflegeversicherung-droht-2027-ein-milliarden-defizit-382919/
• [30] S, compilation on the procedural status of the care reorganization act, ministerial draft of June 2026, hearing of the associations, pending cabinet decision, and the federal health minister’s announcement of August 2026. https://www.deutscherpflegebeistand.de/pflegereform-2027/
• [31] P, Bundesministerium für Gesundheit (Federal Ministry of Health), Erster Bericht der FinanzKommission Gesundheit of March 30, 2026, S. 61, Tabelle 4.2.3.1, overview of the central results of the medium term projection. https://www.bundesgesundheitsministerium.de/fileadmin/Dateien/3_Downloads/F/FinanzKommission_Gesundheit/FinanzKommissionGesundheit_Erster_Bericht_20260330.pdf
• [32] P, Bundesministerium für Gesundheit (Federal Ministry of Health), contributions of the statutory health insurance. https://www.bundesgesundheitsministerium.de/beitraege/seite
• [33] P, Bundesministerium für Gesundheit (Federal Ministry of Health), financing of the long term care insurance. https://www.bundesgesundheitsministerium.de/themen/pflege/online-ratgeber-pflege/die-pflegeversicherung/finanzierung

Contribution rates, assessment ceilings, taxes, and prices

• [34] P, Bundesministerium für Arbeit und Soziales (Federal Ministry of Labor and Social Affairs), Sozialversicherungsrechengrößen-Verordnung 2025. https://www.bmas.de/DE/Service/Gesetze-und-Gesetzesvorhaben/sozialversicherungs-rechengroessenverordnung-2025.html
• [35] P, Bundesministerium für Arbeit und Soziales (Federal Ministry of Labor and Social Affairs), Rechengrößen der Sozialversicherung 2015. https://www.bmas.de/DE/Service/Presse/Pressemitteilungen/2014/rechengroessen-sozialversicherung-2015.html
• [36] P, Deutsche Rentenversicherung (German Pension Insurance), assessment ceilings of the pension insurance. https://rvrecht.deutsche-rentenversicherung.de/SharedDocs/rvRecht/07_AktuelleWerte/A-C/awert_bybmgz.html
• [37] P, Deutsche Rentenversicherung (German Pension Insurance), assessment ceiling for health and long term care insurance. https://rvrecht.deutsche-rentenversicherung.de/SharedDocs/rvRecht/07_AktuelleWerte/A-C/awert_bybmgzkvpv.html
• [38] P, Bundesministerium der Finanzen (Federal Ministry of Finance), development of the tax and total revenue ratios, Monatsbericht Juni 2025. https://www.bundesfinanzministerium.de/Monatsberichte/Ausgabe/2025/06/Inhalte/Kapitel-6-Statistiken/6-1-s13-entwicklung-der-steuer-und-abgabequoten.html
• [39] P, Bundesministerium der Finanzen (Federal Ministry of Finance), development of tax revenue 2000 bis 2025. https://www.bundesfinanzministerium.de/Content/DE/Standardartikel/Themen/Steuern/Steuerschaetzungen_und_Steuereinnahmen/entwicklung-der-steuereinnahmen.html
• [40] P, Bundesministerium der Finanzen (Federal Ministry of Finance), Datensammlung zur Steuerpolitik 2025, S. 22, Abschnitt 2.2, Beitrag der Steuerpflichtigen zum Lohn- und Einkommensteueraufkommen 2025. https://www.bundesfinanzministerium.de/Content/DE/Downloads/Oeffentliche-Finanzen/Datensammlungen-zur-Steuerpolitik/datensammlung-zur-steuerpolitik-2025.pdf?__blob=publicationFile
• [41] P, Statistisches Bundesamt (Federal Statistical Office), Verbraucherpreisindex 1991 bis 2025, GENESIS table 61111-0001. https://genesis.destatis.de/datenbank/online/table/61111-0001/
• [42] P, Bundesministerium der Finanzen (Federal Ministry of Finance), offsetting of bracket creep. https://www.bundesfinanzministerium.de/Content/DE/FAQ/kalte-progression.html
• [43] P, Bundesministerium der Finanzen (Federal Ministry of Finance), solidarity surcharge from 2021. https://www.bundesfinanzministerium.de/Monatsberichte/2021/01/Inhalte/Kapitel-3-Analysen/3-1-wichtigste-steueraenderungen-2021.html
• [44] P, Umweltbundesamt (Federal Environment Agency), national emissions trading and the carbon price for heating and transport. https://www.umweltbundesamt.de/der-nationale-emissionshandel
• [45] P, Organisation für wirtschaftliche Zusammenarbeit und Entwicklung, Taxing Wages 2026, effective tax wedges 2000 to 2025. https://www.oecd.org/en/publications/taxing-wages-2026_3a5169ef-en/full-report/evolution-of-effective-tax-rates-on-labour-income-2000-25_5aed550b.html
• [46] E, own calculations of the total contribution rate, the inflation adjusted model, the scenarios, and the burden indicator, from [25] bis [45], method fully disclosed in the calculation appendix.

Demography and migration

• [47] P, Statistisches Bundesamt (Federal Statistical Office), Wanderungen zwischen Deutschland und dem Ausland 1950 bis 2025. https://www.destatis.de/DE/Themen/Gesellschaft-Umwelt/Bevoelkerung/Wanderungen/Tabellen/wanderungen-zwischen-deutschland-und-dem-ausland-jahr-02.html
• [48] P, Statistisches Bundesamt (Federal Statistical Office), Lebendgeborene und Gestorbene, time series. https://www.destatis.de/DE/Themen/Gesellschaft-Umwelt/Bevoelkerung/Sterbefaelle-Lebenserwartung/Tabellen/lrbev04.html
• [49] P, Statistisches Bundesamt (Federal Statistical Office), 16. koordinierte Bevölkerungsvorausberechnung, assumptions and results. https://www.destatis.de/DE/Themen/Gesellschaft-Umwelt/Bevoelkerung/Bevoelkerungsvorausberechnung/annahmen_ergebnisse_16te_kBv.html
• [50] P, Statistisches Bundesamt (Federal Statistical Office), Rückgang der Geburtenziffer schwächte sich 2024 deutlich ab, press release 259 of July 17, 2025, total fertility rates of 1.84 and 1.23 and the trend since 2017. https://www.destatis.de/DE/Presse/Pressemitteilungen/2025/07/PD25_259_12.html
• [51] P, Statistisches Bundesamt (Federal Statistical Office), Zusammengefasste Geburtenziffer sinkt im Jahr 2025 auf 1,32 Kinder je Frau, press release 230. https://www.destatis.de/DE/Presse/Pressemitteilungen/2026/07/PD26_230_12.html
• [52] P, Bundesinstitut für Bevölkerungsforschung (Federal Institute for Population Research), analysis of the German Emigration and Remigration Panel Study 2018 and 2019, return intention and qualification structure. https://www.bib.bund.de/Publikation/2019/pdf/Bevoelkerungsforschung-Aktuell-6-2019.pdf?__blob=publicationFile&v=2

Labor market, Ukraine, and Poland

• [53] P, Bundesagentur für Arbeit (Federal Employment Agency), Monatsbericht Dezember 2025, employment rates of German and foreign nationals. https://statistik.arbeitsagentur.de/Statistikdaten/Detail/202512/arbeitsmarktberichte/monatsbericht-monatsbericht/monatsbericht-d-0-202512-pdf.pdf
• [54] P, Bundesagentur für Arbeit (Federal Employment Agency), Arbeitsmarktsituation von Staatsangehörigen der Asylherkunftsländer und der Ukraine, fact sheet with the definition of the employment and unemployment rates. https://statistik.arbeitsagentur.de/Statistikdaten/Detail/Aktuell/arbeitsmarktberichte/faktenblatt-asyl8-ukr/faktenblatt-asyl8-ukr-d-0-pdf.pdf?__blob=publicationFile
• [55] P, Deutscher Bundestag (German Bundestag), printed paper 21/5316, Ukrainer auf dem Arbeitsmarkt und im Sozialsystem, employment, age structure, and gender distribution. https://dserver.bundestag.de/btd/21/053/2105316.pdf
• [56] S, Institut für Arbeitsmarkt- und Berufsforschung (Institute for Employment Research), Zuwanderungsmonitor November 2025. https://doku.iab.de/arbeitsmarktdaten/Zuwanderungsmonitor_2511.pdf
• [57] S, Institut für Arbeitsmarkt- und Berufsforschung (Institute for Employment Research), Zuwanderungsmonitor Februar 2026, unemployment rate of Ukrainian nationals in December 2025. https://doku.iab.de/arbeitsmarktdaten/Zuwanderungsmonitor_2602.pdf
• [58] S, Institut für Arbeitsmarkt- und Berufsforschung (Institute for Employment Research), Arbeitsmarktintegration ukrainischer Geflüchteter, cohort arriving between February and August 2022. https://iab.de/presseinfo/arbeitsmarktintegration-ukrainischer-gefluechteter-verlaeuft-deutlich-schneller-als-bei-frueheren-fluchtkohorten/
• [59] P, Bundesinstitut für Bevölkerungsforschung (Federal Institute for Population Research), Drei Jahre nach Ankunft in Deutschland, etwas mehr als die Hälfte der Ukrainerinnen und Ukrainer ist erwerbstätig, November 4, 2025. https://www.bib.bund.de/DE/Presse/Mitteilungen/2025/2025-11-04-Drei-Jahre-nach-Ankunft-Etwas-mehr-als-die-Haelfte-der-Ukrainer-ist-erwerbstaetig.html
• [60] P, Bundesinstitut für Bevölkerungsforschung (Federal Institute for Population Research), underlying study. https://doi.org/10.12765/bro-2025-04
• [61] P, Eurostat, persons under temporary protection, status July 31, 2025. https://ec.europa.eu/eurostat/web/products-eurostat-news/w/ddn-20250910-1
• [62] P, Statistics Poland, foreigners performing work in Poland in August 2025. https://stat.gov.pl/files/gfx/portalinformacyjny/pl/defaultaktualnosci/6329/15/34/1/cudzoziemcy_wykonujacy_prace_w_polsce_w_sierpniu_2025_r..pdf
• [63] P, Ministry of Family, Labor and Social Policy of the Republic of Poland, information on the employment of foreigners in Poland to December 31, 2025, figures based on ZUS data, S. 22. https://psz.praca.gov.pl/documents/10828/183932/Informacje%2Bo%2Bzatrudnieniu%2Bcudzoziemc%C3%B3w%2Bw%2BPolsce%2Bdo%2B31%2Bgrudnia%2B2025%2Br.pdf/342f72c1-d683-78ca-166b-3d50d5134eb8?t=1776935297384
• [64] P, Statistics Poland, Social Assistance Beneficiaries 2024, annex on foreign recipients. https://stat.gov.pl/files/gfx/portalinformacyjny/en/defaultaktualnosci/3399/2/9/1/social_assistance_beneficiaries_2024.pdf
• [65] S, Narodowy Bank Polski (National Bank of Poland), living and economic situation of migrants from Ukraine, 2025 survey. https://nbp.pl/wp-content/uploads/2025/11/Sytuacja_imigrantow_z_Ukrainy-2025.pdf
• [66] S, Organisation for Economic Co-operation and Development, International Migration Outlook 2025, employment rates of displaced persons. https://www.oecd.org/en/publications/international-migration-outlook-2025_ae26c893-en.html
• [67] P, Zakład Ubezpieczeń Społecznych (Polish Social Insurance Institution), the 800 plus family benefit. https://www.zus.pl/pl/swiadczenia/swiadczenia-dla-rodzin/swiadczenie-wychowawcze-800-plus
• [68] P, Zakład Ubezpieczeń Społecznych (Polish Social Insurance Institution), changes to the conditions from February 2026. https://www.zus.pl/pl/swiadczenia/swiadczenia-dla-rodzin/swiadczenie-wychowawcze-800-plus/zmiany-w-swiadczeniu-wychowawczym-program-rodzina-800-
• [69] P, European Central Bank, annual average exchange rate of zloty per euro 2025. https://data.ecb.europa.eu/data/datasets/EXR/EXR.A.PLN.EUR.SP00.A
• [70] S, Sejm of the Republic of Poland, labor ministry figures on the number of Ukrainian children in the 800 plus program. https://www.sejm.gov.pl/sejm10.nsf/wypowiedz.xsp?dzien=2&posiedzenie=27&view=1&wyp=92

Abuse, sanctions, and the fiscal balance

• [71] P, Deutscher Bundestag (German Bundestag), plenary record 21/79, proceedings on suspicion of benefit abuse in 2025. https://dserver.bundestag.de/btp/21/21079.pdf
• [72] P, Deutscher Bundestag (German Bundestag), printed paper 21/966, organized gang benefit abuse. https://dserver.bundestag.de/btd/21/009/2100966.pdf
• [73] P, Bundesagentur für Arbeit (Federal Employment Agency), Zahl der Leistungsminderungen ist 2025 gegenüber dem Vorjahr gestiegen, press release 13. https://www.arbeitsagentur.de/presse/2026-13-zahl-der-leistungsminderungen-ist-2025-gegenueber-dem-vorjahr-gestiegen
• [74] S, Gnaedinger, K., Le Floch, M., und Stichnoth, H., The Net Fiscal Contribution of Foreigners in Germany, WeLaR Deliverable D6.3, Zentrum für Europäische Wirtschaftsforschung, reference year 2018. https://projectwelar.eu/wp-content/uploads/2024/10/D6.3-The-net-fiscal-contributions-of-foreigners-in-Germany.pdf

Automotive industry

• [75] P, Verband der Automobilindustrie e. V. (German Association of the Automotive Industry), passenger car production in Germany. https://www.vda.de/de/aktuelles/zahlen-und-daten/jahreszahlen/automobilproduktion
• [76] P, Verband der Automobilindustrie e. V. (German Association of the Automotive Industry), new registrations and ownership transfers. https://www.vda.de/de/aktuelles/zahlen-und-daten/jahreszahlen/neuzulassungen-und-besitzumschreibungen
• [77] P, Verband der Automobilindustrie e. V. (German Association of the Automotive Industry), turnover and employment. https://www.vda.de/de/aktuelles/zahlen-und-daten/jahreszahlen/allgemeines
• [78] P, Verband der Automobilindustrie e. V. (German Association of the Automotive Industry), development of domestic and foreign production and exports 2025. https://www.vda.de/de/themen/Automobil-Insight-2025/2025-Entwicklung-Inlands-und-Auslandsproduktion-und-Exporte

Election dates

• [79] P, Landtag of Saxony-Anhalt, state election on September 6, 2026. https://www.landtag.sachsen-anhalt.de/alle-dossiers/landtagswahl-am-6-september-2026
• [80] P, State Office for Internal Administration of Mecklenburg-Vorpommern, state election on September 20, 2026, election day set by decision of the state government of July 29, 2025. https://www.laiv-mv.de/Wahlen/Landtagswahlen/2026/

Retirement at 63 and consumption taxes

• [81] P, Bundesministerium für Arbeit und Soziales (Federal Ministry of Labor and Social Affairs), Abschlagsfreie Altersrente für besonders langjährig Versicherte, raising of the age threshold from 63 to 65 years. https://www.bmas.de/DE/Soziales/Rente-und-Altersvorsorge/Leistungen-Gesetzliche-Rentenversicherung/Altersrenten/abschlagsfreie-altersrente.html
• [82] P, Federal Government Press and Information Office, Rente mit 63, die Fakten, age thresholds by birth cohort. https://www.bundesregierung.de/breg-de/aktuelles/rente-mit-63-2195784
• [83] P, Deutsche Rentenversicherung (German Pension Insurance), Renten für langjährig und besonders langjährig Versicherte, deduction of up to 14.4 percent. https://www.deutsche-rentenversicherung.de/DRV/DE/Rente/Allgemeine-Informationen/Rentenarten-und-Leistungen/Altersrente-fuer-langjaehrig-Versicherte/altersrente-fuer-langjaehrig-versicherte_node
• [84] S, account of the sparkling wine tax, the Reichstag decision of 1902 to finance the imperial war fleet, the current rate, and annual revenue. https://de.wikipedia.org/wiki/Schaumweinsteuer
• [85] S, overview of rarely discussed consumption taxes, including the coffee tax levied since 1949 and its spread in Europe. https://www.insm.de/aktuelles/oekonomenblog/die-ueberfluessigsten-steuern-deutschlands
• [86] P, Deutscher Bundestag (German Bundestag), answer of the federal government to a parliamentary question, printed paper 21/6290, sugar levy from 2028 in the benchmark figures for the 2027 federal budget and in the act stabilizing health insurance contribution rates. https://www.bundestag.de/presse/hib/kurzmeldungen-1185020
• [87] S, Deutsches Ärzteblatt, Politik plant Ausweitung der Zuckersteuer auf andere Getränke, staggering by sugar content. https://www.aerzteblatt.de/news/politik-plant-ausweitung-der-zuckersteuer-auf-andere-getranke-8c855bc6-cd1b-48df-9b95-78d3c32dbd26
• [88] S, Deutsches Ärzteblatt, Bundesregierung erteilt Zuckersteuer auf zuckerfreie Getränke eine Absage, revenue estimates and the benchmark paper of August 2026. https://www.aerzteblatt.de/news/bundesregierung-erteilt-zuckersteuer-auf-zuckerfreie-getranke-eine-absage-023d3483-722b-4eb2-a968-45794aafdf4b
• [89] P, Bundesverfassungsgericht (Federal Constitutional Court), Urteil of November 5, 2019, 1 BvL 7/16, sanctions in social law, guarantee of a subsistence minimum compatible with human dignity. https://www.bundesverfassungsgericht.de/SharedDocs/Entscheidungen/DE/2019/11/ls20191105_1bvl000716.html
• [90] P, Bundesverfassungsgericht (Federal Constitutional Court), press release 74/2019, sanctions enforcing obligations to cooperate in receipt of unemployment benefit II partly unconstitutional. https://www.bundesverfassungsgericht.de/SharedDocs/Pressemitteilungen/DE/2019/bvg19-074.html
• [91] P, Bundesagentur für Arbeit (Federal Employment Agency), Tag der Pflege, jede fünfte Pflegekraft hat ausländische Staatsangehörigkeit, press release 16 of May 2026. https://www.arbeitsagentur.de/presse/2026-16-tag-der-pflege-jede-fuenfte-pflegekraft-hat-auslaendische-staatsangehoerigkeit
• [92] S, Mediendienst Integration, Factsheet ausländische Pflegekräfte und Ärzte 2026, Daten der Bundesagentur für Arbeit zum reference date June 30, 2025. https://mediendienst-integration.de/fileadmin/user_upload/Factsheets/Factsheet_Pflege_2026_Mediendienst_auslaendische_Pflegekraefte_Aerzte.pdf
• [93] S, Initiative Neue Soziale Marktwirtschaft, study on the increase in posts in the federal administration 2026, press release of August 31, 2026, federal posts and personnel spending 2017 to 2027 based on the budget plans and the 2027 draft budget. https://www.presseportal.de/pm/39474/6342355
• [94] S, Initiative Neue Soziale Marktwirtschaft, Neue Studie zur Stellenmehrung in der Bundesverwaltung 2026, breakdown by established civil service and employee posts and by personnel spending. https://www.insm.de/aktuelles/pressemitteilungen/neue-studie-zur-stellenmehrung-in-der-bundesverwaltung-2026

PDF Read and download this article as PDF

Preview Download PDF

Disclaimer: This article reflects the author’s own research and opinion as of the publication date shown above; later findings or legal changes may have overtaken it, so always check that date. Sources are cited for independent verification, and no liability is accepted for third-party studies. This is general information, not medical, legal, or professional advice: for medical questions see a doctor, for legal questions a lawyer, and in an acute crisis contact emergency services or a crisis helpline.

Read the full disclaimer

About the header image: it is AI-generated. Cheaper than a photo shoot, and I have made my peace with the age of AI. Everything inside the article is real, the diagrams, the skulls, the findings, and every word. The machine gets the opening shot and not one inch past it.