ILR School

Models of Linked Employer-Employee Data Conference 2019

Permanent URI for this collection

In 1999, John Abowd, Francis Kramarz, and David Margolis published "High Wage Workers and High Wage Firms." Their model and econometric techniques were pioneering in the analysis of labor markets, and many other markets with high-dimensional interactions between two sides. The conference will celebrate the continuing influence of the AKM (1999) model and its foundational role in the analysis of labor markets using linked employer-employee data. The goal of the conference is to highlight the range of new questions for which the AKM decomposition provides an effective way to understand data. The conference will also focus attention on key econometric and modeling issues that are common to applications of AKM. The program is posted at


Recent Submissions

Now showing 1 - 4 of 4
  • Item
    Do firm effects drift?
    Lachowska, Marta; Mas, Alexandre; Saggio, Raffaele; Woodbury, Stephen A. (Presented at the Models of Linked Employer-Employee Data Conference 2019, 2019-10-13)
    Firm effects in the AKM model are typically assumed to be constant over time. But what if they aren't constant? We look at a Time-Varying AKM model to show that firm effects are highly persistent and that variance components vary with the business cycle.
  • Item
    It Ain’t Where You’re From, It’s Where You’re At..
    Kline, Patrick; Saggio, Raffaele; Sølvsten, Mikkel (Presented at the Models of Linked Employer-Employee Data Conference 2019, 2019-10-12)
    The AKM model of wage determination was loosely motivated by wage posting models (e.g., Burdett and Mortensen, 1998) that feature a stable wage ladder. Using the AKM model, we looked at gender differences in sorting. Where workers "are at" - their current jobs - has been found to be more important than where they "are from" (previous jobs) in the gender gap in hiring wages.
  • Item
    Social Connections and the Sorting of Workers to Firms
    Eliason, Marcus; Hensvik, Lena; Kramarz, Francis; Norström Skans, Oskar (Presented at the Models of Linked Employer-Employee Data Conference 2019, 2019-10-13)
    We assess the impact of social connections on the sorting of workers to firms (and the presumption that connections increase sorting inequality) by first examining the distribution of displaced workers’ social connections to employed workers and their firms; using Swedish data, we measure multiple types of networks, of both the strong and weak sort (family, former co-worker, former classmate, current neighbor); we estimate an AKM decomposition (to assess sorting inequality). Then, we examine the causal impact of connections on hiring for these displaced workers and how connections and their strength affect sorting inequality. Finally, we look at how connections affect sorting for all job-to-job movers. In this way, we find that our measured social connections display homophily: positive sorting in terms of earnings capacity; high-wage workers are connected to high-wage workers who tend to be employed by high-wage firms. Thus we can conclude that social connections matter when looking for jobs.
  • Item
    Firm Pay Dynamics
    Engbom, Niklas; Moser, Christian (2019-10-13)
    We investigate firm pay in cross-section and over time by combining linked employer-employee data from Sweden, register data on individual characteristics for all workers, and income statement and balance sheet data for all firms in order to estimate AKM equation augmented with firm-year fixed effects, relate (dynamics of) firm pay to (dynamics of) firm financials, and measure static and dynamic sorting between workers and firms.