Letter to Honorable Bill Monning on Estimated Costs Related to Certain ACA Provisions

April 19, 2013 Hon. Bill Monning, Chair Senate Budget Subcommittee No. 3 on Health and Human Services Room 5019, State Capitol Sacramento, California 95814 Dear Senator Monning: Your budget subcommittee staff asked our office to analyze the administration’s February fiscal estimates and assumptions related to changes made by the federal Patient Protection and Affordable Care Act (ACA) that the administration is characterizing as the “mandatory” Medi- Cal expansion. We agreed to focus our analysis on the estimated costs related to ACA provisions that will likely result in additional Medi-Cal enrollment from persons who are currently eligible for the program, but not enrolled. (The administration’s mandatory expansion fiscal estimates include other relatively minor fiscal effects, such as the cost of expanding Medi-Cal coverage to former foster youth up to age 26.) We provide our analysis below. LAO Bottom Line. We agree with the administration that several ACA provisions will likely increase Medi-Cal enrollment from persons who are currently eligible, but unenrolled in the program. There is substantial uncertainty about the state costs associated with these changes. Based on our analysis of a range of potential fiscal estimates using various sets of enrollment assumptions, we find the administration’s fiscal estimates to be plausible, but likely too high. We have no major issues with the administration’s assumptions about per-enrollee costs for the mandatory expansion population. Rather, the primary difference between our fiscal estimates and the administration’s fiscal estimates relates to differing underlying assumptions about the number of additional enrollees under the mandatory expansion. We are unclear on the basis for some of the administration’s assumptions in developing its estimates of additional enrollment, and find that the administration’s enrollment estimates, while plausible, are likely high— particularly in the short term. ACA WILL LIKELY INCREASE ENROLLMENT FROM CURRENTLY ELIGIBLE PERSONS The ACA contains several provisions that are expected to increase the number of persons who are eligible for Medi-Cal under current eligibility standards that enroll in the program. We describe the major provisions below. It is important to note that, unlike health care services provided to individuals who may become eligible under the optional Medi-Cal expansion authorized under the ACA (that is, the expansion that is due to changes made by the ACA intended to expand the number of people eligible for Medi-Cal), the state will generally be Preprinted logo will go here Hon. Bill Monning 2 April 19, 2013 responsible for 50 percent of the costs for services provided to individuals who would have been eligible under current Medi-Cal eligibility standards. Greater Use of Electronic Data to Verify Medi-Cal Eligibility. Under the ACA, where feasible, states must verify an applicant’s information electronically. The amount of paperwork that an applicant will need to submit will likely be reduced, thereby potentially encouraging more people to apply for the program. In addition, the state must use electronic information to make annual eligibility redeterminations. If the available electronic information suggests that an individual is still eligible for the program, he or she will automatically remain enrolled in Medi- Cal without the need for additional action by the individual. Simplification of Methodology Used to Determine Financial Eligibility. Beginning January 1, 2014, the ACA makes changes to the methodology used to calculate income when determining Medicaid Program eligibility for most beneficiaries—excluding certain populations, such as seniors and persons with disabilities. Currently, the methodology used to determine financial eligibility for Medi-Cal is complex—often involving verification of an applicant’s assets and accounting for a variety of income deductions and exemptions. The ACA generally simplifies the methodology used to determine financial eligibility. The two major changes to the methodology are:  Requiring the use of a new methodology to calculate income, known as Modified Adjusted Gross Income (MAGI). As part of this change, various adjustments to applicant income that are currently made would no longer be made.  Asset tests will no longer be used to determine eligibility. These changes will likely make the application process simpler and may result in more individuals enrolling in the program. Creates Penalties for Certain Individuals Without Health Insurance Coverage. Beginning January 1, 2014, the ACA requires most U.S. citizens and legal residents to have health insurance coverage or pay a penalty—a requirement that is commonly known as the individual mandate. Certain individuals who are eligible for Medi-Cal are exempt from the individual mandate, including those exempt from filing federal taxes due to their low-income status. However, the penalties—or the perceived threat of the penalties—will likely encourage some currently eligible but not enrolled persons to enroll in the program. Enhanced Outreach Activities to Encourage Individuals to Obtain Health Coverage. In addition to the eligibility changes identified above, the ACA also includes funding for enhanced outreach activities to encourage individuals to obtain health coverage. For example, the California Health Benefit Exchange (the Exchange) will be conducting outreach activities aimed at enrolling uninsured individuals in health insurance coverage. Its outreach efforts will be targeted primarily to persons who would potentially purchase coverage on the Exchange. However, the outreach efforts may also encourage individuals who are eligible for Medi-Cal to enroll in the program. Hon. Bill Monning 3 April 19, 2013 ADMINISTRATION’S FISCAL ESTIMATES The Governor’s January budget included a $350 million General Fund placeholder to account for costs related to the so-called mandatory expansion in 2013-14, until a more refined estimate could be developed. In February, the administration provided an updated fiscal estimate for the mandatory expansion that replaced the $350 million placeholder. Figure 1 provides a summary of the administration’s February fiscal estimates for the additional General Fund costs for the currently eligible but uninsured population under the ACA. The administration’s mandatory expansion fiscal estimates included some other fiscal effects associated with the ACA, such as the cost of expanding Medi-Cal coverage to former foster youth up to age 26. However, these other fiscal effects are relatively minor compared to the estimated costs of providing services to currently eligible but uninsured persons. We do not discuss these other fiscal effects in this analysis. LAO ASSESSMENT OF ADMINISTRATION’S FISCAL ESTIMATES Below, we provide our assessment of the administration’s estimated General Fund costs for the currently eligible but unenrolled population under the ACA. (We note that the administration estimated costs for only the currently eligible and uninsured population, rather that the entire unenrolled population. As discussed later, estimating costs for only the uninsured population would likely have a relatively minor effect on overall cost estimates.) ACA-Related Estimates Are Subject to Substantial Uncertainty We agree with the administration that the ACA will increase enrollment from persons who are currently eligible for the program. However, the type and scope of changes made by the ACA are largely unprecedented and, thus, there is very little evidence about how these changes will affect state costs for persons who are eligible for Medi-Cal under current eligibility standards. Therefore, any estimate of these additional costs is subject to considerable uncertainty. Several national and state-level studies have estimated these additional state costs associated with the ACA—with significantly varying results. In addition, our office has provided a range of cost estimates (discussed in more detail below). Each of these estimates depend on several key assumptions—each of which is subject to uncertainty. We discuss these key assumptions later. Hon. Bill Monning 4 April 19, 2013 Side-by-Side Comparison of LAO and Administration’s Estimates Analyst’s Range of Estimated Annual Costs. Given the significant uncertainty surrounding costs associated with the mandatory expansion, in our February report, The 2013-14 Budget: Analysis of the Health and Human Services Budget, we provided the Legislature a range of estimated costs for these additional enrollees under three different scenarios. Each scenario involved a different set of key assumptions that drives the estimated costs under the scenario. Figure 2 provides a summary of our estimated costs under the three different scenarios. The potential for the particular assumptions used in each of these three scenarios is based on our review of a wide variety of studies and reports. Of the three LAO scenarios presented, we believe the moderate-cost scenario is most likely. We believe the low-cost and high-cost scenarios are plausible, but not likely. Under the moderate-cost scenario, we estimate that the health care costs associated with this population would be $104 million in 2013-14—about $84 million less than the $188 million estimated by the administration. Using different, but still plausible assumptions, we estimate state costs could potentially be as low as $30 million or as high as $254 million in 2013-14. We also estimate annual costs will likely be over $350 million within a few years—potentially ranging from about $100 million to nearly $1 billion annually. Hon. Bill Monning 5 April 19, 2013 Administration’s Cost Estimates Are Within the Range of LAO Estimates, but Likely Too High. As shown in Figure 3, the administration’s estimates are higher than our moderate-cost scenario, but lower than our high-cost scenario. Therefore, the administration’s cost estimates are generally plausible as they are within the range of LAO estimates. However, as discussed in detail below, we believe state costs will likely be lower than what was estimated by the administration. LAO Assessment of Administration’s Underlying Assumptions Estimated costs for a given year depend on two key factors: (1) the number of additional enrollees (which is itself made up of an estimate of the size of the eligible, but unenrolled population and the percent of the eligible population that will enroll—commonly referred to as the “take-up” rate), and (2) the cost of providing health coverage to each additional enrollee. Below, we provide our assessment of the methodology and assumptions used by the administration to estimate these two key factors. Hon. Bill Monning 6 April 19, 2013 Administration’s Enrollment Estimates Likely Too High; Basis for Some Assumptions Unclear Enrollment Estimates Are Generally Plausible, but Likely Too High. As shown in Figure 4, the administration’s enrollment estimates are slightly lower than our high-cost scenario enrollment estimates. Therefore, we believe the administration’s enrollment estimates are generally plausible, yet likely too high. In addition, we are unclear on the basis for certain assumptions used by the administration to estimate the additional enrollment, which we discuss in more detail below. Administration’s Methodology Differs From Most Available Studies and Reports. We generally relied on prior studies and reports when estimating additional enrollment. For example, there are several research teams that have developed economic models (microsimulation models) that attempt to estimate how the ACA will affect the distribution of health insurance coverage in California, including the University of California’s California Simulation of Insurance Markets, the Urban Institute’s Health Insurance Policy Simulation Model, and the Lewin Group’s Health Benefits Simulation Model. Each model used somewhat different data and assumptions. However, given the significant amount of research and expertise that was used to develop these Hon. Bill Monning 7 April 19, 2013 models, in our view, they provide a reasonable starting point for estimating the number of additional Medi-Cal enrollees from the currently eligible but unenrolled population under the ACA. The administration developed its own, very different model to predict how certain ACA changes would affect Medi-Cal enrollment. Since the administration’s methodology is fundamentally different from other studies and reports, it is difficult to directly compare certain assumptions made by the administration with those made in the other studies and reports that informed our estimates. For example, the administration’s enrollment estimates do not make any assumptions about the overall size of the currently eligible but unenrolled population or the overall take-up rate under the ACA—key assumptions that are frequently cited in other reports. Mandatory Expansion Population Consists of Two Groups. The eligible but unenrolled population consists of two general groups: (1) individuals who were formerly enrolled in Medi- Cal, but then disenrolled, and (2) individuals who were never enrolled in Medi-Cal. The administration’s estimates make assumptions with respect to both of these groups. We discuss the administration’s assumptions related to both of these groups below. Unclear Basis for Administration’s Assumptions About Program Retention Under ACA. Individuals generally disenroll from Medi-Cal coverage for one of three basic reasons: (1) they lose eligibility, (2) they acquire other health insurance, or (3) they “drop out”—when a beneficiary remains eligible, but becomes uninsured. Individuals who drop out are eligible to participate in Medi-Cal, but for various reasons have not completed the redetermination process—perhaps because the process was too burdensome. A simulation model developed by the Department of Health Care Services (DHCS) was used to predict how the new redetermination process would reduce the number of people who drop out and, thus, increase the number of people who remain in the program. Based on results from DHCS’ simulation, the administration estimates that over 600,000 additional Medi-Cal enrollees will remain in the program by 2014-15 as a result of the changes in the redetermination process. The simulation model developed by DHCS incorporates the following key assumptions about program retention under the ACA.  100 percent of individuals who would otherwise have a gap in Medi-Cal coverage of six months or less would now remain in the program.  75 percent of individuals who would otherwise have a 7-11 month gap in Medi-Cal coverage would now remain in the program.  40 percent of individuals who would otherwise lose coverage for a period of more than a year would now remain in the program. The administration cites several studies and reports as the basis for these assumptions. Based on our review of these studies, we are unclear how they support the administration’s assumptions. For example, one study cited by the administration found that, of the children who disenroll from Medicaid and remain uninsured 12 months later, 45 percent were still eligible the next year. It is not clear how this result supports an assumption that 75 percent of individuals who would otherwise have a 7-11 month gap in coverage would remain eligible and enrolled in Hon. Bill Monning 8 April 19, 2013 the program. In addition, the administration assumes all of the individuals who lose coverage for less than six months were drop outs—an assumption that we believe is unreasonable. While we believe it is reasonable to assume many of the individuals who would otherwise have a gap in coverage for six months may retain coverage under the new redetermination process, it is highly unlikely that 100 percent of those individuals would retain coverage. (This particular assumption operates to make the administration’s enrollment estimates higher than they would be under a more reasonable assumption.) Unclear Basis for Administration’s Assumptions About Additional First-Time Enrollees. The administration’s estimates assume that by 2014-15 roughly 200,000 individuals who would otherwise never enroll in the program will enroll as a result of certain ACA provisions, such as enhanced outreach and the individual mandate. The administration assumes that the size of this population will be 33 percent of the roughly 600,000 individuals who are estimated to remain on the program by 2014-15 as a result of the simplified redetermination process (discussed above). The basis for this assumption is not clear. Administration’s Assumed “Phase-In” Period Is Likely Too Quick, Leading to Higher Short-Term Enrollment Estimates. The administration assumes that the additional enrollment associated with the ACA will be complete—or “phased-in”—by September 2014. In contrast, our estimates assumed enrollment would be phased-in through June 2016. We believe it will take at least a couple of years for the effects of the ACA on Medi-Cal enrollment to be fully realized. For example, the size of the financial penalties associated with the individual mandate increase from 2014 to 2016. If financial penalties associated with the individual mandate encourage more people to enroll in Medi-Cal, it is reasonable to expect the mandate to encourage additional people to enroll in 2015 and 2016. In addition, the Exchange currently has a plan to conduct marketing and outreach activities through 2017. We believe it is reasonable to expect to see these outreach activities gradually increase Medi-Cal enrollment over a period of at least a couple of years. The phase-in assumption that is incorporated into the administration’s estimates has the effect of likely overestimating enrollment in the next couple of fiscal years. Administration’s Assumptions About Underlying Caseload Growth Are Likely Too High. Once the phase-in period for additional enrollment associated with the ACA is complete, both of our estimates assume that the number of enrollees will continue to grow annually—generally reflecting underlying caseload growth in the Medi-Cal Program. The administration assumes 3 percent annual caseload growth, which is higher than the 1 percent caseload growth assumed in our moderate-cost scenario and the 2 percent caseload growth assumed in our high-cost scenario. Based on our review of historic caseload trends for Medi-Cal families with children, we believe the administration’s caseload growth assumption of 3 percent is likely too high. However, this difference in assumptions has a relatively minor impact on the overall cost estimates, particularly in the first few years. Certain Administration Assumptions Tend to Underestimate Additional Enrollment. While we believe the administration’s overall enrollment estimates are likely, on balance, too high, there are certain assumptions in its analysis that tend to underestimate additional enrollment. For example, the administration includes costs for additional enrollment among individuals who are currently eligible but uninsured, instead of unenrolled. Most other analyses assume that there Hon. Bill Monning 9 April 19, 2013 will be some additional enrollment among individuals who would otherwise have some other type of private coverage, such as job-based coverage. This effect is often known as “crowd-out” of private insurance. The administration does not explicitly assume any additional enrollment due to such crowd-out. In addition, the simulation model developed by DHCS was used to estimate retention among families and children in certain Medi-Cal aid codes. The aid codes analyzed by the administration likely make up the large majority of persons who may remain in the program as a result of the new redetermination process. However, some individuals in other aid codes may remain in the program as a result of these changes. Excluding these other aid codes from the analysis may result in an underestimate of enrollment. For both of the examples provided, the effect is likely relatively minor compared to the other enrollment-related assumptions we discussed above. Administration’s Cost-Per-Enrollee Assumptions Generally Appear Reasonable Per-Enrollee Cost Estimates Generally Appear Reasonable. The administration’s overall weighted average per member per month (PMPM) cost assumptions appear reasonable. (Separate PMPM cost assumptions are made for individuals who are eligible for federal Medicaid matching funds and individuals who are eligible for federal Children’s Health Insurance Program [CHIP] matching funds.) As shown in Figure 5 (see next page), the administration’s PMPM cost assumptions are between the LAO moderate-cost and high-cost scenarios. Compared to the LAO moderate-cost scenario, the administration assumes slightly higher PMPM costs for individuals eligible for Medicaid matching funds and slightly lower costs for individuals eligible for CHIP matching funds. Our only significant concern with the administration’s PMPM assumptions is that they assume that the costs for new enrollees will be nearly identical to costs for families and children who are currently enrolled. We agree with the administration that it is reasonable to use families and children already enrolled in Medi-Cal as the starting point for estimating PMPM costs for the new enrollees. However, we also believe it is reasonable to assume that the PMPM costs for new enrollees would be slightly less than the PMPM costs for the already enrolled families and children because new enrollees will likely have somewhat lower health care needs and thus costs. This is because most individuals who are eligible but not enrolled likely do not have an immediate need for health care services. It may take a couple of years for these slightly lower PMPM costs to be reflected in the state budget because it may take a couple of years for the costs for the new enrollees to be fully incorporated into the capitation rates paid to managed care plans. Hon. Bill Monning 10 April 19, 2013 Administration’s Annual Medical Inflation Assumption Appears Reasonable. The administration assumes that PMPM costs will increase by 5 percent annually as a result of medical inflation. Our office’s estimates assume annual medical inflation of roughly 4 percent to 5 percent. While there are fundamental differences in the manner in which our office and the administration arrived at each of our annual medical inflation assumptions, we believe the administration’s assumption of 5 percent annual medical inflation is generally a reasonable one. ADDITIONAL ISSUES TO CONSIDER Below we identify a couple of other important issues to consider related to expected state costs associated with the so-called mandatory expansion under the ACA. ACA May Reduce Long-Term Average Administrative Costs for Eligibility Determinations. In the long run, we expect streamlined eligibility and enrollment processes to reduce the average per-applicant cost of Medi-Cal eligibility determinations. The conversion to MAGI eligibility and other changes to the eligibility processes described above will likely result in some efficiencies and lower per-applicant eligibility costs. Several important implementation details are still being determined at the state and federal levels. At this point, neither the LAO analysis nor the administration’s estimates make any explicit assumptions about reduced administrative costs. Hon. Bill Monning 11 April 19, 2013 Remaining Implementation Details Will Likely Affect State Costs. Several remaining ACA implementation decisions could have a significant impact on state costs. Some of these remaining implementation details include:  Eligibility Verification Plan. The standards and electronic data sources that will be used to conduct eligibility determinations and redeterminations are still being determined.  MAGI Conversion Methodology. The methodology the state will use to convert current income standards into MAGI-equivalent income standards is uncertain.  Methodology for Claiming Federal Funds. The methodology used to identify who is currently eligible versus newly eligible (assuming adoption of the optional expansion) is uncertain. If you have any further questions on these issues, please contact Ross Brown of my staff at (916) 319-8345 or Ross.Brown@lao.ca.gov. Sincerely, Mac Taylor Legislative Analyst