Imagine you've been complaining for years that your apartment lease forbids you from rearranging the furniture. You finally break the lease, move out — and rearrange almost nothing. That's what German employers do when they exit collective bargaining agreements (Tarifverträge): they gain the freedom to restructure working hours and barely use it. The committed claim here is sharp: employer exit from collective bargaining in Germany is NOT primarily driven by working-time constraints, despite what employer associations have loudly insisted. Using IAB Establishment Panel data spanning 2008–2024 — one of the richest employer surveys in any advanced economy — Pyka and Rieder find that leaving collective coverage produces a working-time increase of just 0.3 to 0.6 percent. In absolute terms, that's eight to thirteen additional minutes per week per worker. Not per day. Per week. The identification strategy is solid applied micro: firm fixed-effects plus event-study designs that let you see the time profile of adjustment before and after exit. The event-study plots are the key exhibit. If firms were straining against binding working-time ceilings, you'd expect a sharp, sustained jump after exit — a coiled spring releasing. Instead, there's a small bump that appears shortly after exit and then flattens. No delayed ramp-up, no gradual expansion of weekend or shift work. Saturday and Sunday work margins show no significant change post-exit. This matters because German employer associations — particularly the BDA — have for decades cited working-time rigidity as a primary reason firms abandon sectoral bargaining. This narrative has been policy-relevant: it feeds into arguments for decentralizing Germany's bargaining system, which has seen coverage fall from roughly 70% of western German workers in the mid-1990s to around 43% today. If hours aren't the real reason firms leave, the political argument for further erosion of collective bargaining rests on shakier ground. The paper's architecture is straightforward panel econometrics — nothing exotic — but the data advantage is substantial. The IAB Establishment Panel is a stratified sample of roughly 16,000 establishments annually, giving enough statistical power to detect small effects and enough time depth (17 waves) to build credible pre-trends. The authors are explicitly honest that they cannot rule out all endogeneity: firms that choose to exit may differ in unobservable ways. But the pre-treatment parallel trends in the event-study plots provide reasonable reassurance. The integrity picture is decent but not airtight. The IAB panel is a well-known, institutionally maintained dataset — not a convenience sample — and the econometric specification is standard. The weakness is that this is observational, not experimental. You can't randomly assign firms to exit collective bargaining. And while the paper debunks the hours narrative, it doesn't conclusively answer what IS driving exit. Wage costs, administrative burden, opening clauses, firm-level competitive pressure — these candidates are flagged but not systematically tested. The successor question is the obvious one: if not hours, then what? The authors gesture toward wages and general flexibility but don't run a parallel decomposition. My read: this is likely paper two in a sequence. The clean negative result on hours earns its own publication; the messy positive story about what actually drives exit is harder and probably in progress.
