David Wittekopf

A large literature documents that negative labor-demand shocks scar affected communities through persistent job loss, earnings declines, and political radicalization. Yet a theoretical tradition going back to Schumpeter's notion of creative destruction predicts that the closure of unproductive firms can also free resources for more productive uses. Whether the cleansing or scarring force dominates in any given episode plausibly depends on what set off the destruction. Existing studies have largely confirmed the scarring hypothesis for market-driven shocks – where demand or technology forces firms to exit – but the policy-relevant case of a government choosing to let firms fail remains almost entirely unaddressed. We study one such setting. After reunification, the German Treuhandanstalt liquidated thousands of state-owned East German firms as part of the transition from a planned to a market economy. Combining firm-level data on 6,608 Treuhand firms with administrative records for 251,812 workers from September 1989 onward, we find that a one-standard-deviation increase in a county's exposure to Treuhand liquidations raises labor earnings by 2.7 log points. These effects are not driven by selection through unemployment or migration. We attribute this premium to faster reallocation of workers from declining to expanding sectors, and higher rates of new firm establishments which provide higher average pay. Further, high-exposure counties register persistently lower support for far-right parties across three decades of federal elections. Our results provide a clean existence proof that creative destruction can dominate scarring, and suggest that the adverse consequences associated with labor-demand shocks reflect not destruction per se but the failure to reallocate.

Sorting into careers—occupations and the educational pathways leading to them—is known to account for much of intergenerational earnings persistence. Yet in Swedish cohorts born 1960–1978, we document that about a fifth of the aggregate parent–child earnings association survives conditioning on narrowly defined career entry cells (sex, education level and institution, and occupation at career start), and the residual is concentrated above the 80th percentile of parental earnings. This is driven by what we call the “class pay gap within careers”: within the same career, children with higher combined parental earnings increasingly out-earn their peers as careers progress, and the gradient steepens toward the top—bottom-quintile children earn about 7% less than top-quintile peers at age 40 (0.25 within-cell standard deviations). The gap appears in graduate and non-graduate careers alike, but is muted where the public sector dominates. Dimensions of class beyond parental earnings—education, wealth, and occupation-derived social class—further stratify earnings within careers, with children from working-class households in the bottom earnings quintile facing the largest gaps. Firm and geographical sorting and non-cognitive skills together account for about 40% of the gap. We show that the remainder largely reflects differential earnings growth within firms rather than higher switching rates or larger gains at switches (net of firm effects), pointing to differences in progression and promotion inside firms.

How far do sectoral wage settlements reach beyond the workers they nominally cover, and where along the wage distribution do they bite? I link agreement-level data on negotiated wages to German linked employer-employee data and the IAB Establishment Panel to measure establishment-level exposure to sectoral settlements. Using a shift-share design with exposure-robust inference, I estimate the pass-through of negotiated wage increases to actual wages, the response of the wage cushion, and spillovers to establishments not bound by an agreement. The estimates inform tools that forecast wage growth from collective agreements, which observe only the covered segment of the labor market.