The Effect of Physician Fees and Density Differences on Regional Variation in Hospital Treatments

We use a panel data set of about 1.7 million hospital records in 4,000 Dutch zip code regions for the years 2006-2009. We estimate the effect of physician fees and physician density on regional variation in hospital care for nine different treatments. Our results show that a 1 percent increase in the total number of physicians, if these extra physicians are all paid according to an output-based reimbursement scheme, would increase the number of treatments on average by 0.40 percent. For salaried physicians we find a significantly lower average effect of 0.15 percent. We find no or weak effects for hip fractures, which is included in the analysis as a control treatment. Our data allows us to deal with reverse causality, excess demand, border crossing, and availability effects. Our findings lend support to the existence of supplier induced demand for the majority of the analyzed treatments.