Please use this identifier to cite or link to this item: http://hdl.handle.net/10419/100401
Authors: 
Graetz, Georg
Feng, Andy
Year of Publication: 
2014
Series/Report no.: 
Beiträge zur Jahrestagung des Vereins für Socialpolitik 2014: Evidenzbasierte Wirtschaftspolitik - Session: Job Polarization D14-V2
Abstract: 
Job polarization the rise in employment shares of high and low skill jobs at the expense of middle skill jobs occurred in the US not just recently, but also in the late nineteenth and early twentieth centuries. We argue that in each case polarization resulted from increased automation, and provide a theoretical explanation. In our model, firms deciding whether to employ machines or workers in a given task weigh the cost of using machines, which is increasing in the complexity (in an engineering sense) of the task, against the cost of employing workers, which is increasing in training time required by the task. Insights from artificial intelligence and robotics suggest that some tasks do not require training regardless of complexity, while in other tasks training is required and increases in complexity. In equilibrium, firms are more likely to automate a task that requires training, holding complexity constant. We assume that more-skilled workers learn faster, and thus it is middle skill workers who have a comparative advantage in tasks that are most likely to be automated when machine design costs fall. In addition to explaining job polarization, our model makes sense of observed patterns of automation and accounts for a set of novel stylized facts about occupational training requirements.
JEL: 
J23
J31
O33
Document Type: 
Conference Paper

Files in This Item:
File
Size





Items in EconStor are protected by copyright, with all rights reserved, unless otherwise indicated.