Efficient Inefficiency in the Hospital Sector?

Recent research demonstrates that productivity dispersion in health care is comparable to other sectors (Chandra et al., 2016.) The hospital industry is comprised of for-profit, not-for-profit, and public providers, and there is some evidence that markets with higher for-profit penetration deliver care with greater efficiency (Kessler and McClellan, 2002.) Kahn (2017) finds that liberal cities hire more public-sector workers than conservative cities, and pay these workers more. Against this backdrop, we investigate the sources of inefficiency in the hospital sector, as well as the potential benefits of inefficiency. Exploiting a previously developed measure of hospital productivity (Romley et al., 2015), we first document that public hospitals are less efficient in delivering care than nonprofit and for-profit hospitals. Thus, as inpatient utilization continues to decline, closure of public hospitals would efficiently rationalize excess capacity within the industry. Indeed, from 1970 through 2001, exit rates were highest for public hospitals (Chakravarty et al., 2006). However, when we examine the period 2002-2011 using American Hospital Association annual surveys, public hospitals were least likely to close. Yet, in a discrete time hazard analysis with hospital fixed effects, we find that nonprofits were more likely to close in areas in which land values appreciated (as measured by the Zillow Home Value Index), and for-profit hospitals were even more likely to exit in the face of rising land values; this pattern corroborates and complements the earlier results of Chakravarty. Going forward, we will build on our current findings by assessing the degree to which the inefficiency of public hospitals stems from “excess” utilization of labor. We will also assess the political economy of inefficiency, in terms of potential benefits to particular communities. To begin with, we will explore the degree to which the spatial distribution of public hospitals serves to mitigate spatial mismatch in employment opportunities, using PUMA-level Census data. In addition, we will explore the degree to which the location of public hospitals provides timely and convenient access to care to disadvantaged communities.