Hospital competition, spillovers and provision of uncompensated care

ABSTRACT

Using data for California from 2005 until 2010, we investigate to what extent market competition and the presence of non-profits in the area may play a role in equilibrium uncompensated care (UC) levels, allowing those effects to differ according to the hospital’s ownership type. Previous studies have not explored the potential spillover effects from non-profit hospitals into the hospital decision of UC provision. We find evidence that regions with more non-profits experienced larger increases in UC levels, and even more in less concentrated markets. Our results also indicate that UC provision by for-profit hospitals decreases the larger the presence of non-profits in the region, and this effect is magnified when competition is more intense. We, therefore, find no positive spillover effects of non-profits into the hospital decision of UC provision, which may help us to understand the recent trends in UC levels.

KEYWORDS:

• Uncompensated care
• non-profit hospitals
• hospital competition
• spillover

JEL CLASSIFICATION:

• I11
• I18
• L39

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

¹ Available at https://www.cbo.gov/system/files/115th-congress-2017-2018/reports/53300-individualmandate.pdf.

² Source: The Kaiser Commission on Medicaid and the Uninsured. Available at https://kaiserfamilyfoundation.files.wordpress.com/2014/05/8596-uncompensated-care-for-the-uninsured-in-2013.pdf.

³ Clemens-Cope, Garrett, and Buettgens (2010) estimated that the uninsured population would decrease from 49.1 million to 15.1 million (a decrease of 69.3%), once ACA was fully implemented, whereas uncompensated care costs were projected to decline 58.7% (from 61.1 to 25.2 billion dollars).

⁴ In March 2015, the Assistant Secretary for Planning and Evaluation (ASPE) revised their prediction for the decline in uncompensated care in 2014, estimating uncompensated care costs would drop 7.4 billion dollars (from 34.7 billion dollars in 2013 to 24.3 billion dollars), rather than the initial projection of a 5.7 billion dollar decrease (available at https://aspe.hhs.gov/sites/default/files/pdf/139226/ib$\_$UncompensatedCare. pdf). In contrast, a Congressional Budget Office report from 2013, projected an 11 million (from 55 million to 44 million), or 20%, decrease in the uninsured population in 2014 (available at https://www.cbo.gov/sites/ default/files/51298–2013-02-ACA.pdf), while recent data from the Current Population Survey indicated a 8.8 million (from 41.8 million in 2013 to 33 million in 2014), or 21%, decrease in uninsured in 2014 (available at https://www.census.gov/content/dam/Census/library/publications/2015/demo/p60-253.pdf).

⁵ Source: Internal Revenue Service (2009) IRS Exempt Organizations Hospital Study. Available at https:

⁶ This geographic definition was developed by Makuc et al. (1991) together with the Department of Human and Health Services for health planning purposes based on patient flows.

⁷ As stated by the OSHPD, MDCs are mutually exclusive categories containing all possible principal diagnosis areas. There are 25 different MDCs.

⁸ In results not shown, these variables indeed become insignificant when HSA fixed effects are included.

⁹ Nonprofit is the omitted category.

¹⁰ See the next section for an explanation of why we estimate hospital price.

¹¹ This instrumenting procedure is common in studies that examine market structure and related outcomes. See, for example, Berry, Levinsohn, and Pakes (1995) and Copeland (2014).

¹² We should note that results presented in Tables 3 and 5 are robust to alternative measures of UC. If we adjust UC to the cost-to-charge ratio, as suggested by the OSHPD, we get similar results than the ones in Tables 3 and 5. We also consider the ratio between UC adjusted by contractual deductions and net patient revenue as an additional dependent variable. Once we consider the UC level relative to the hospital size, overall results in Tables 3 and 5 still hold, but the interactions between the for-profit dummy and our three variables of interest become insignificant. This result indicates that the effect of market competition and non-profit presence would be no different for non-profits and for-profits. This result is somewhat expected as the ratio between UC and net patient revenue is very similar for both non-profit and for-profit hospitals.