Opportunities and risks of data-driven labour market policies
Data-driven decision models are seen as a commitment to an efficient digital society. This project looks into the opportunities and risks presented by such models for employment services.
Thanks to increasing computing power and growing data sets, machines can make accurate predictions and even solve complex decision-making problems such as «What product might you like in an online shop». Such data-driven decision models are already being used on a large scale by technology companies such as Google or Facebook. However, it is not yet clear whether and to what extent they can also be useful in more policy-relevant contexts. We aim to identify the opportunities that targeted application of these methods may offer employment services: Can the models provide personnel consultants with decision-making support? How do the choice of the database or the formulation of the algorithm's goals influence the quality of the data-driven decisions?
In recent years, research concerning data-driven decisions in the field of economics has progressed well. However, this research has so far been mainly theoretical. The new methods must therefore be subjected to practical evaluation in a realistic employment service setting before they are implemented.
The goal of the project is to systematically determine whether the latest machine learning methods can improve decision-making. We aim to demonstrate how to use a data-driven decision model to best achieve decision-makers’ objectives.
The project aims to answer practical questions that arise when data-driven decision models are used for job placement. These questions also being of interest to the State Secretariat for Economic Affairs (SECO), a potential implementer of the models, we will be working closely with this office. However, our results will not be limited to the labour market, but will also be applicable in other areas characterised by similar decision-making processes.
Chances and risks of data-driven decision making for labour market policy