Efficiency of American states after implementation of the patient protection and affordable care act (PPACA) from 2014 to 2017

ABSTRACT

We assess the impact of healthcare efficiency among U.S. states after implementation of the PPACA. A Malmquist Index decomposes productivity changes of state healthcare systems since 2014. Results indicate productivity regression of 0.14-percentage points from 2014 to 2017. In 2017 the average state is 4.9-percentage points inefficient. Using the conditional order-m estimator and nonparametrically regressing the ratio of conditional to unconditional order-m efficiency scores on secondary environmental variables, we find that behavioural, socioeconomic, and healthcare utilization factors play a role in explaining state inefficiency. The average state becomes nearly fully efficient after controlling for behaviour (efficiency improves by 3.7-percentage points), socioeconomic (4.1-percentage points), and healthcare utilization (3.2-percentage points) variables. Using a second-stage regression framework, we find that higher levels of obesity, adult smoking, and diabetes lead to lower healthcare efficiency. Likewise, we see that low and high levels of unemployment are associated with improved healthcare efficiency, which is in line with contradictory studies. We also find further support for the link between income and health outcomes, through the vehicle of improved health efficiency. Lastly, we find that ensuring that low-cost populations engaging in health treatments have improved health outcomes, as they prevent high-cost morbidities in the future.

KEYWORDS:

• Conditional order-m
• nonparametric frontier estimation
• affordable care act
• healthcare efficiency
• nonparametric econometrics

JEL CLASSIFICATION:

• C14
• I11
• I12
• I18

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

¹ Unfortunately, having to choose between an input-orientation and an output-orientation leads to an issue surrounding the order-α and order-m estimators. Often, as noted in Wheelock and Wilson (2009), the choice between input- or output-orientation is often arbitrary. Wheelock and Wilson (2008) developed an unconditional hyperbolic order-α estimator, which allows for input contraction and output expansion simultaneously, which ignores the issues posed by having to pick an input- or output-orientation.

² In our analysis, the demographic and economic variables that would lead us to reject the separability assumption, based on the analysis shown in Figure 1, are: the labour force participation rate, natural log of GDP per household, the fraction of individuals age 65 and over, and the immigrant share in the population. The social protection variables that would lead us to reject the separability assumption are: current healthcare expenditures on healthcare as a percent of GDP and per capita out-of-pocket healthcare expenditures. These figures are available upon request from the authors.

³ https://www.cdc.gov/diabetes/statistics/meduse/fig2.htm.