Working Papers:
with Simon Jäger and Patrick Nüß, NBER Working Paper 31988
Abstract
We study how labor market conditions affect unionization decisions. Tight labor markets might spur unionization, e.g., by reducing the threat of unemployment after management opposition or employer retaliation in response to a unionization attempt. Tightness might also weaken unionization by providing attractive outside alternatives to engaging in costly unionization. Drawing on a large-scale, representative survey experiment among U.S. workers, we show that an increase in worker beliefs about labor market tightness moderately raises support for union activity. Effect sizes are small as they imply that moving from trough to peak of the business cycle increases workers' probability of voting for a union by only one percentage point. To study equilibrium effects, we draw on three quasi-experimental research designs using data from across U.S. states and counties over several decades. We find no systematic effect of changes in aggregate labor market tightness on union membership, union elections, and strikes. Overall, our results challenge the notion that labor market tightness significantly drives U.S. unionization.
Labor Movements and Electoral Representation
with Felix Kersting
Abstract
This paper studies the relationship between workers' electoral representation and organized labor movements. We show that they are substitutable: a sharp decline in the political representation of the working class leads to an increase in union membership and a rise in the labor share. These results are based on a historical episode in 19th-century Germany, where universal male suffrage for state elections in Saxony was replaced by a restrictive, class-based system that allocated voting power based on tax payments. To establish that this reform was salient to the electorate, we first examine its effects on turnout. In a triple-difference framework, we compare over time the difference between turnout in state and national elections in Saxon constituencies with the same object in Prussian constituencies (electoral rules for all national elections and Prussian state elections remained unchanged). We document that voter turnout fell by 13.5 pp (about 25%) in Saxon state elections affected by the reform, with the decline driven by lower-income workers abstaining from the vote. Next, we show how the reform affected the composition of state parliament by drawing on a novel biographical dataset that captures the socioeconomic background of all members of parliament (MPs) between 1869 and 1918. Beyond decreasing the number of seats held by the Social Democratic Party (SPD) to zero, the antidemocratic suffrage reform yielded a sharp increase in the number of newly elected MPs that can be classified as “industrialists.” We then study whether this negative shock to workers' electoral representation led to a change in union membership – an alternative channel through which workers could collectively express their demands. To that end, we draw on newly digitized panel data on county-industry level union membership for the German Empire matched with employment information from the 1895 census. Using a difference-in-differences design, we show that the decrease in electoral representation of the working class led to a 4 pp increase in union membership and a 3 pp rise in the labor share, suggesting that the labor movement more than compensated for the loss in political representation.
Work in Progress:
Efficient Layoff Prioritization
with Nathan Lazarus
Abstract
We examine the welfare consequences of requiring firms to prioritize workers with low displacement costs for layoffs. Using linked employer–employee administrative data, we exploit a reform that raised the firm-size threshold for the applicability of German layoff-prioritization rules in a difference-in-differences design. Firms newly exempt from the rules shift layoffs toward older, longer-tenured workers and parents, indicating that the law previously constrained their layoff decisions along these dimensions. These displaced and newly unprotected workers experience larger earnings losses, spend more time unemployed, and draw more on unemployment insurance benefits. At the same time, newly exempt firms recover faster from shocks, with quicker rebounds in profits and employment. To evaluate welfare consequences implied by these patterns, we develop a search-and-matching model with worker heterogeneity in productivity and job-finding rates and introduce dismissal rules that require firms to dismiss high job-finding-rate workers before any low job-finding-rate workers. The model shows that prioritization can raise welfare by reallocating displacement toward workers with shorter expected nonemployment spells, but can also directly reduce output and create dynamic distortions in hiring. Moving forward, we plan to combine the model with our empirical estimates to characterize when layoff-prioritization rules increase social welfare.
Unions and Technology
with Daron Acemoglu, Kjell Salvanes and Alexander Willén
Taxes And Directed Technical Change
with Nathan Lazarus
Abstract
We study how changes in relative labor-to-capital taxation affect the direction of technological change, particularly the adoption and innovation of labor-saving versus labor-augmenting technologies. To study these effects, we draw on a stacked difference-in-differences design and variation in the effective relative tax rate across U.S. states. We link these changes to (i) patent classifications into labor-augmenting and labor-saving technologies to study innovation outcomes and (ii) classified microdata from the Longitudinal Foreign Trade Transactions Database to study technology adoption. Our preliminary results so far show limited reduced form effects of increases in relative taxation on the share of labor augmeting patents. We are moving the analysis to the firm level using census data, estimating the adoption of automating technology and using time-varying production function estimation to estimate the effect of taxation on the exponent on labor in the production function.