Does capitated managed care affect budget predictability? Evidence from Medicaid programs

Abstract

As the second largest item in the budget of every US state, Medicaid budget stability and financial transparency have significance for every state. This study is the first to test whether managed care enrollment reduces the variance of Medicaid spending, in contrast to the focus of the existing literature on spending levels. This variance bears directly on whether budget constrained states whether budget constrained states benefit from managed care in the form of stabilized spending, leading to improved budget predictability. Capitated payments stabilize spending at the margin, but the effects may be unobservable in aggregate due to variation in enrollment, which is directly measured in the analysis, or selection bias, which is unobserved. Although the majority of Medicaid enrollees are in managed care, the study shows that managed care use has been concentrated among the enrollees with the most stable spending, resulting in only small gains to budget predictability. This finding is robust to the exclusion of the claims expenditures that exhibit the most variance.

Appendices

Appendix A: Measures of managed care enrollment by type of enrollee

The purpose of this section is to control for changes in the Medicaid managed care population. This analysis uses additional data extracted from MSIS to include the share of managed care enrollees by basis of eligibility. While this data provides more nuanced insights into composition of managed care populations in theory, the broad categorization of “managed care” from this data source limits the interpretability of these results. Enrollees in any form of managed care, including PCCMs that are reimbursed on a volume basis and largely function as traditional fee-for-service Medicaid coverage (Tables 8, 9, 10, 11).

Table 8 Effect of Medicaid managed care on the accuracy of projected medicaid spending

Table 9 Effect of Medicaid managed care on the magnitude of mid-year adjustments

Table 10 Effect of Medicaid managed care on medicaid spending variance

Table 11 Effect of Medicaid managed care on measures of fiscal transparency

Appendix B: Marginal effects of managed care enrollment by type of enrollee

The purpose of this section is to test differences in the marginal effect of managed care enrollment by enrollee type, overall Medicaid enrollment, and revenue volatility. This analysis uses the same data as in the main specification; it differs in that the specification includes interaction terms (Tables 12, 13, 14, 15).

$$\begin{aligned} Overage_{jst}= & {} \alpha +\beta _1 \%CapitatedManagedCare_{st} \nonumber \\&+ \beta _2 \%NoncapitatedManagedCare\nonumber \\&+\beta _3 Enrollment_{st}+\beta _4 BOE+\beta _5RevenueChange_{st}\nonumber \\&+\beta _5 Enrollment_{st}* \%CapitatedManagedCare_{st} \nonumber \\&+\beta _6 Enrollment_{st}* \%NonCapitatedManagedCare_{st} \nonumber \\&+\beta _7 BOE* \%CapitatedManagedCare_{st} \nonumber \\&+\beta _8 BOE* \%NonCapitatedManagedCare_{st} \nonumber \\&+\beta _9 RevenueChange_{st}* \%CapitatedManagedCare_{st} \nonumber \\&+\beta _{10} RevenueChange_{st}* \%NonCapitatedManagedCare_{st} \nonumber \\&+s+t+\epsilon _{jst} \end{aligned}$$

(5)

Table 12 Effect of Medicaid managed care on the accuracy of projected medicaid spending

The interactions between the model of managed care and the distribution of enrollees within the Medicaid program is intended to test the hypothesis that managed care may affect the dependent variable differentially based on the overall risk pool of the Medicaid population. For example, higher managed care enrollment in a population with more volatile spending, such as enrollees dually eligible for SSI due to disability, should yield higher gains in budget forecasting accuracy than increased managed care among patients with more stable spending, such as children. The interactions are largely statistically insignificant.

Table 13 Effect of Medicaid managed care on the magnitude of mid-year adjustments

Table 14 Effect of Medicaid managed care on medicaid spending variance

Table 15 Effect of Medicaid managed care on measures of fiscal transparency

Appendix C: Sensitivity checks for outcome measures

See Tables 16 and 17.

Table 16 Effect of Medicaid managed care on measures of projection accuracy using different thresholds of accuracy

Table 17 Effect of Medicaid managed care on measures of Medicaid spending variance based on NASBO data