Jump to:Page Content
The past decade has seen an explosion in the use of big data and algorithm-based learning. But can this information be used by regulators to target their interventions more effectively?
Professor David I. Levine (PhD 1987), Eugene E. and Catherine M. Trefethen Professor of Business Administration at the Haas School of Business at the University of California, Berkeley presented his perspective during a Regulatory Policy Program seminar on Feb. 4.
In one of the largest randomized group studies ever conducted, Professor Levine, along with Professors Michael Toffel of Harvard Business School and Matthew Johnson of Boston University, found clear evidence that work site inspections by the Occupational Safety and Health Administration (OSHA) had a positive and lasting impact in reducing workplace injuries. But how could our understanding that inspections reduce workplace injuries be applied to help OSHA do its job better? By combining randomized trials like this one with machine learning algorithms, Professor Levine suggested that OSHA could isolate the types of workplaces that would benefit most from an inspection.
Machine learning uses large amounts of data to recognize patterns and make predictions, and it is already being applied with great success by companies like Google, Uber, and Amazon to anticipate user behavior. Randomized experiments allow organizations to pinpoint the effectiveness of an intervention. Credit card companies like CapitalOne have used this technique to test the success of new products and marketing efforts.
The combination of these two techniques, while common in the private sector, has yet to be applied in the public sector. By randomizing workplace inspections that already take place, OSHA could determine what types of characteristics are most predictive of workplace injury. This information would then be plugged into an algorithm to determine how to allocate scarce inspector time and energy. Especially in the current political environment where evidence-based policy is the emerging gold standard for regulation, this could be a very appealing proposition.
Joseph Aldy, associate professor of public policy and faculty chair of the Regulatory Policy Program, moderated the presentation. The Spring 2016 New Directions in Regulation Seminar series is sponsored by the Regulatory Policy Program at the Mossavar-Rahmani Center for Business and Government.
For more information about the seminar series, visit https://www.hks.harvard.edu/centers/mrcbg/programs/rpp/seminars.
David I. Levine, Eugene E. and Catherine M. Trefethen Professor of Business Administration at the Haas School of Business at the University of California, Berkeley
By randomizing workplace inspections that already take place, OSHA could determine what types of characteristics are most predictive of workplace injury.