Work & Contribution
Strong social partnership must actively shape the transitions brought by AI in office work.
Where does Germany stand, and where does it break?
In 2024, 49 per cent of employees still worked in companies covered by a sectoral or firm-level collective agreement (Destatis), a share that has been falling for years. The social partnership between trade unions and employers shapes wages and working conditions. Short-time work has proven its worth as a bridge through demand downturns during the financial crisis and the pandemic. At the same time, skills shortages persist across many sectors, while the automotive industry is in the midst of profound transformation. By the end of the third quarter of 2025, the sector employed around 48,700 fewer workers than a year earlier (Destatis). The German model is built on a scarce and irreplaceable factor of labour. Collective bargaining power, co-determination, and short-time work all presuppose that employment is preserved and that people are needed.
Short-time working helps bridge a temporary demand downturn, not the permanent elimination of activities. If AI automates administrative processing, accounting, customer service, and parts of engineering work, the result is not a cyclical employment problem but a structural one that conventional instruments are not designed to address. Skills shortages and automation pull in opposite directions. In some fields there are too few people; in others they are becoming redundant. The decisive question is no longer the number of jobs, but whether upskilling and social adjustment can keep pace with the speed of change. At the same time, declining collective bargaining coverage is eroding precisely the negotiating power that would need to broker a fair transition.
Paths for Germany
The state and the social partners are substantially expanding continuing education, retraining, and qualification short-time work, channelling employees out of automated activities and into fields where demand is strong. For Germany, this is the most natural path, since skills shortages and automation coexist and retraining addresses both sides. The Qualifizierungsgeld and the Weiterbildungsgesetz already provide the requisite instruments. Success depends on whether people can actually transition into more productive activities and on whether the pace of AI adoption does not overtake them.
Trade-off: Employment security against realistic retrainability
Feasibility: Only effective if sufficient compatible roles emerge
The remaining paid work is distributed across a larger share of the population through shorter working weeks or shorter working lives, in order to maintain broad-based employment. For Germany, this draws on a long tradition of collectively bargained working-time reductions and could help preserve social cohesion. At the same time, it sits in tension with the acute shortage of skilled workers and with the financing of a social security system reliant on contribution volumes. Without productivity gains sufficient to sustain wages, a loss of prosperity becomes a real risk.
Trade-off: Social equity against economic output
Feasibility: Only viable if productivity gains are sufficient to sustain wages
Collective bargaining and co-determination are strengthened and extended through transformation agreements that shape automation at company level in a socially equitable manner, with employment bridges, upskilling commitments, and a share in productivity gains. This is central to Germany because the declining collective bargaining coverage of 49 per cent is eroding precisely the negotiating power that a fair transition requires. The transformation of the automotive industry is the real test case. Without a reversal of the trend in collective bargaining coverage, however, the reach will be insufficient.
Trade-off: Orderly transition against waning organisational power
Feasibility: Requires a reversal of the decline in collective bargaining coverage
Social security is becoming more decoupled from employment, through a reformed basic income guarantee, a negative income tax, or elements of a universal basic income, financed from returns on capital and value creation. This is relevant for Germany if structurally less paid work is available and participation must be redefined. It does, however, break with the contribution-based and performance-linked ethos of the German welfare state, and it is costly. The political and cultural hurdle is high.
Trade-off: subsistence security versus work incentives and fiscal sustainability
Feasibility: Only becomes relevant once paid employment declines permanently and broadly
Trigger-based phases
Trigger: AI first takes on well-standardised office and service tasks; certain occupational profiles shrink as a result.
The skills shortage initially masks the displacement, as those released move into vacant positions. Short-time working schemes and retraining still function. The automotive industry already illustrates how structural change and job losses can coincide.
Trigger: Job losses driven by automation exceed the number of viable new positions across several industries.
Short-time working reaches its limits because the downturn is not cyclical. Pressure on requalification, working hours, and social security mounts. Weak collective bargaining coverage makes a negotiated adjustment harder to achieve.
Trigger: Germany redefines contribution, participation, and social protection beyond the traditional model of full-time employment.
Social partnership, requalification, and social security are aligned to an economy with less paid work, distributed differently. This fork in the road will determine social cohesion and the legitimacy of the model through to 2040.
Real-world pilot programmes
Status: In force since 2024. Supports retraining in the context of structural change.
The Qualifizierungsgeld supports employees whose roles are being eliminated by structural transformation while they undergo retraining. It represents Germany's real-world attempt to evolve the short-time work principle from a cyclical bridge into a tool for structural reskilling.
Status: Ongoing. In the third quarter of 2025, employment was down by 48,700 compared with the prior year.
Regional transformation networks and collectively bargained future agreements — notably in the automotive and supplier industries — represent the real test case for socially managed structural change. They reveal whether social partnership can navigate deep disruption without rupture.
Status: Proven and available; designed for cyclical downturns.
Short-time work saved millions of jobs through the financial crisis and the pandemic. As a reference point, it also illustrates its limits: it works for temporary shortfalls in demand, not for the permanent elimination of entire activities through automation.
Status: Established in Sweden; to be assessed as a model for Germany.
In Sweden, transition councils run by the social partners swiftly provide individualised support to redundant workers, helping them move into new employment or further training. For Germany, this is a reference model showing how social partnership can organise active labour market transition rather than merely defending the status quo.
What we are watching
Collective bargaining coverage measures how broadly working conditions are negotiated collectively. The threshold matters because the path to the economy ahead requires strong negotiating parties to shape automation in a socially responsible way. If coverage continues to fall, the organisational capacity for an orderly transition is absent. For social stability, this is an early indicator.
The automotive industry is Germany's leading industrial sector and the most visible case of transformation. The threshold matters because job cuts without replacement employment demonstrate that structural change is destroying jobs on a net basis. A tipping point would signal that social partnership is failing to cushion the transition. For regions and suppliers, this is an existential issue.
This indicator measures whether qualifications and demand are aligned. It matters because a growing mismatch signals that workers are being displaced without the skills required for available positions. A tipping point would mean that retraining is lagging behind the pace of automation. For policymakers, it is the signal to accelerate requalification efforts.
Take-up reveals whether these instruments serve as a temporary bridge or become a permanent condition. The threshold matters because workers remaining on short-time work without returning to employment marks the dividing line between cyclical and structural crisis. A tipping point would indicate that conventional instruments can no longer absorb technological change — forcing a rethink of the underlying model of income protection.