Please cite this article in press as: King, J., et al. Computerization and wage inequality between and within German work establishments. Research in Social Stratification and Mobility (2016), http://dx.doi.org/10.1016/j.rssm.2016.05.002 ARTICLE IN PRESS G Model RSSM-299; No. of Pages 11 Research in Social Stratification and Mobility xxx (2016) xxx–xxx www.sciencedirect.com Research in Social Stratification and Mobility journal homepage: http://www.elsevier.com/locate/rssm Computerization and wage inequality between and within German work establishments Joe King a,∗ , Malte Reichelt b , Matt L. Huffman c a Department of Sociology University of California, Irvine Irvine, CA 92697-5100, United States b Institute for Employment Research (IAB), Regensburger Strasse 104, 90478 Nuremberg, Germany c Department of Sociology & The Paul Merage School of Business University of California, Irvine Irvine, CA 92697-5100, United States a r t i c l e i n f o Article history: Received 15 July 2015 Received in revised form 10 May 2016 Accepted 11 May 2016 Available online xxx Keywords: Wage inequality Workplace heterogeneity Computerization Matched employer-employee data a b s t r a c t Recent evidence has revealed that a significant share of the rise in wage inequality has occurred at the establishment level, underscoring the importance of workplace-level analyses for understanding grow- ing inequality. Using longitudinal matched employment data from Germany, we provide new insights into how investments in information and communication technologies (ICT) affect earnings inequality between and within establishments over time. Focusing on the mechanisms of inequality, cross-sectional estimates provide evidence of both skill- and class-biased technological change; however, establishment fixed effects models reveal that this relationship is driven by unobserved establishment heterogene- ity. Despite a strong relationship between computerization and the rise in workplace heterogeneity, we find little evidence of a causal effect of computers on changes in establishment-level inequality. Rather, establishments that invest more greatly in ICT pay on average better wages and exhibit higher within-establishment inequality. These results challenge dominant explanations about the role of com- puterization in rising inequality, while also reinforcing the necessity of using organizational data to study inequality processes. Published by Elsevier Ltd. The rise in wage inequality across many advanced societies has been dramatic, spurring considerable scholarly and public inter- est in its causes. Most existing studies have investigated rising inequality using individual-level data (e.g. Bound and Johnson, 1992; Juhn, Murphy, & Pierce, 1993; Katz and Murphy, 1992; Krueger, 1993; Western & Rosenfeld, 2011), industry-level data (e.g. Autor, Katz, & Krueger, 1998; Berman, Bound, & Griliches, 1994; Kristal, 2010, 2013; Lin & Tomaskovic-Devey, 2013), or occupational-level data (e.g. Firpo, Fortin, & Lemieux, 2011; Goos & Manning, 2007). Recently, Card, Heining, and Kline (2013) showed that establishments have also played an important role in ris- ing wage inequality. Using West German administrative data, they demonstrated that a large proportion of growing earnings differentials across occupations, educational groups, and indus- tries as well as the general increase in inequality is explained by rising establishment-specific wage premiums and increased worker sorting across workplaces. Given that considerable growth in inequality has transpired between workplaces, this suggests ∗ Corresponding author. E-mail addresses: kingj2@uci.edu (J. King), malte.reichelt@iab.de (M. Reichelt), mhuffman@uci.edu (M.L. Huffman). that establishment-level analyses may be particularly helpful in explaining rising inequality. A leading explanation for growing inequality has been the spread of information and communication technologies (ICT). Although myriad studies report an association between ICT and inequality, the effect of computers on the wage structure remains elusive. According to the well-known skill-biased technological change hypothesis, spread of these technologies is believed to have increased demand for highly skilled workers and reduced demand for low-skilled workers, leading to skill-based inequal- ity (Card & DiNardo, 2002; Morris & Western, 1999). Much of the rise in inequality along the skill gradient has been attributed to ICT enhancing demand for certain job tasks, while reducing demand and ultimately displacing other tasks through automation (Autor, Levy, & Murnane, 2003; Firpo et al., 2011; Goos & Manning, 2007; Spitz-Oener, 2006). Alternatively, in addition to this direct, skill- based channel, class-biased technological change (i.e. Kristal, 2013; Kristal & Cohen, 2015) contends that ICT investments contribute to inequality indirectly through furthering the decline in collective bargaining. Still others (e.g. Bresnahan, 1999; DiNardo & Pischke, 1997; Doms, Dunne, & Troske, 1997; Handel, 2007) argue that ICT has no influence on wages and that the observed relationship is driven by unobserved heterogeneity. http://dx.doi.org/10.1016/j.rssm.2016.05.002 0276-5624/Published by Elsevier Ltd.