Werner Eichhorst: "If we thin out the management layer, we save costs and even increase productivity"

Middle management used to look like the safest rung in corporate life: well paid, close to decision-making, rarely the first target of restructurings. That assumption is now being challenged. We discuss the shift with Prof. Dr. Werner Eichhorst, an Honorary Professor at the University of Bremen and a long-time senior figure at the IZA – Institute of Labor Economics, where he has held leading roles in Labour Policy Europe and in coordinating labour-market and social-policy research with a strong EU policy focus.

In the conversation, Werner explains why companies are thinning out “the middle”: a mix of cost pressure, overgrown process bureaucracy, and AI-driven rationalisation that can automate routine coordination and administrative work. Using the Bayer example, he also explores what greater employee self-organisation can unlock – and where the risks lie, from status loss for displaced managers to overload and stress when teams are expected to carry more responsibility.

Across many industries – also in Germany – you get the sense that middle management is being thinned out. Broadly speaking, what’s behind that?

Werner Eichhorst: First, it’s striking that this layer is being cut. Traditionally, it’s a part of organisations that tends to grow. Managers sit high up the food chain – and usually it’s others who are downsized before they are. That this is changing is new, and it introduces insecurity into a group that long considered itself privileged.

I see two main reasons. One is the expansion of management roles in good times. What was built up over many years is now judged too expensive in a weaker economy – these are well-paid roles, often focused on protracted internal processes. The second is that top management increasingly questions whether the pyramid is productive. The logic is: if we thin out this layer, we save costs and perhaps even increase productivity or innovation, because hierarchy and process complexity are seen as a drag.

Another explanation is that companies are reversing pandemic overhiring: they hired too many managers and are now correcting. Does that matter?

It can. But we don’t yet have a representative empirical basis – most reports focus on large companies. And large companies tend to develop large “overheads” because they’ve grown structurally. Coordination tasks became more complex over time. Add to that external contract management, managing subcontractors, steering outsourced projects – work that grew more demanding as companies outsourced more.

That model seems to be hitting limits. It’s costly, cumbersome, perceived as slow – and therefore comes under attack. Companies want to be cheaper, but also more decisive, flexible, and “agile.” And that puts a layer in the crosshairs that long felt safe.

Is this mainly about the economy – or also about technology? AI can automate administrative work.

You can read it as another rationalisation wave. We’ve seen these waves whenever competitiveness dominated, and new technical solutions entered the market – management techniques or production techniques. With AI, we now have tools that can take on cognitive routine work: parts of management or secretarial work, standardised coordination, documentation, and reporting.

The key is to distinguish between routine management and overgrown process complexity on one side – and strategic decision-making under uncertainty on the other. The latter is the core management competence: assessing complex situations and making decisions. There, I expect support through better data, not straightforward replacement. But simpler, repetitive tasks are clearly at risk.

“Management” is a broad term. Are we only talking about administrative office work?

Administrative work is particularly vulnerable, yes, but it has been for a while. But I’m thinking of routine types of management rather as coordinating and organising work – structuring what others do, steering teams and interfaces, including external contributors.

For many leaders, that’s the heart of their identity: running a team, setting tasks and resources, ideally growing the team. And you’re saying this is now under pressure?

The growth of that layer is under pressure – because of costs and because companies want to reduce complexity. It’s often unclear what the productivity contribution of very complex structures and cumbersome processes actually is; sometimes they’re even counterproductive. And parts of coordination, project alignment, and even elements of project control can now be handled technically. When repetitive coordination and monitoring fall away, management shifts: it becomes more about initiating projects, setting priorities, deciding why you do something, and for what purpose. AI doesn’t generate that on its own.

To be clear, management tasks remain: entrepreneurial decisions – or delegated entrepreneurial decisions – organisation, and communicating priorities to the workforce. And human work remains necessary where no technical solution exists, or where tools aren’t good enough.

In the AI bubble, people share diagrams of an “hourglass organisation”: leadership at the top, execution at the bottom, less middle in between.

That’s exactly what we’re discussing. I hope that the hourglass doesn’t simply become more top-heavy – that top executives aren’t spared. At the top, too, you should ask: how do these roles contribute to operational success? If leaders keep each other busy, it may be better to organise work differently and put more “power on the road” lower down.

Bayer, an example I wrote about before, suggests self-organisation can work, supported by an AI tool: employees find projects, plan development, “build” their careers – without the need for managers who oversee this. As a consequence, thousands of managers in the company were made redundant. Is that a radical exception – or a model for the future?

Bayer may be among the most radical and fastest right now – triggered by a “revolution from above.” And under pressure from serious business problems rooted in earlier management mistakes, including costly legal disputes in the US. The pressure was real.

It does mean more self-organisation and more responsibility for employees. That matches the promise of the New Work debate – but it isn’t total self-fulfilment. It happens within projects, ultimately triggered from above or by external demands. A group can’t simply decide to work on product Y if strategy demands product X. Constraints remain.

But you can cut the middle layer that translates goals and micromanages.

That can be positive. Micromanagement – especially without technical or social legitimacy – creates dissatisfaction and demotivation. Autonomy within a clear mission supports motivation and engagement. Still, it doesn’t work without human communication: explanations, conflicts and negotiations, performance assessments – those remain. And “building your own career” also means making a case for yourself, demonstrating performance, and navigating competition. Decisions about who takes coordinating roles will still be made – and they still require competent people.

Many hope technology can make such decisions more objective, less based on networks and biases. But this is still an experiment. Where are the risks?

Two big challenges: what happens to managers, and what happens to those who are expected to self-organise.

For managers, downward mobility becomes a real risk. Many went through study and career, expecting permanent roles of that kind. If the trend spreads, there are fewer opportunities to “do management.” Entry becomes harder, and trainee programmes may shrink. Even the signalling value of traditional business degrees and MBAs is being renegotiated – especially given studies that question their contribution and point to self-occupation and managerial self-inflation.

Companies should not treat exits as purely financial. Coaching and outplacement can increase acceptance and protect reputation. It can also open new fields of work – for example, former managers helping others transition.

On the other side, employees may need to learn self-organisation – often from scratch. It also means self-assertion in a dynamic environment and unlearning bureaucratic routines internalised over decades. More freedom can be unsettling when old certainties disappear. This is a qualification and enablement issue – and it’s in the companies’ own interest.

And self-organisation can turn into self-exploitation: overload and multitasking generate stress, ultimately causing health issues. These risks must be identified early and addressed preventively. Otherwise, people withdraw, shift to part-time, or drop out through illness.

You’ve emphasised corporate responsibility. Is there also a role for politics?

The primary responsibility lies with companies. Regulation matters indirectly – rules on part-time work, access to training. In a welfare state, I also see a need to strengthen prevention, especially around psychological strain and related illnesses. Germany already has many protective rules; we shouldn’t dismantle them lightly. Protection can also create incentives for a more productive way of managing pressure.

And finally, on a current note: The right to part-time work is under pressure in Germany. Should it remain?

High part-time rates aren’t mainly caused by the legal claim; they result from life circumstances and individual agreements – parenting, caring responsibilities, and personal preferences. I wouldn’t impose a norm that everyone must work full-time. What matters is that part-time workers aren’t discriminated against because an “ideal worker” model based on full-time (and overtime) dominates. AI could even help assess performance more fairly – without a full-time bias.

And one last point: more work does not automatically increase productivity. Studies suggest, for example, that people – such as mothers working part-time – can be extremely efficient in their paid hours. We need to become more productive and innovative, but not necessarily by working longer.