We believe that growth prospects for the commercial HMOs (e.g. CI, AET, WLP, UNH) are substantially weaker than growth prospects for the Medicaid HMOs (e.g. MGLN, WCG, CNC, MOH), but that valuations fail to fully reflect these disparate fundamentals – making the Medicaid HMOs undervalued by comparison

Relative Fundamentals: Commercial v. Medicaid HMOs

Fundamentally, demand for the commercial HMOs’ services is the combination of enrollment gains from expansion of coverage under the Affordable Care Act (ACA) and from rising employment, net of enrollment losses as employer sponsors move beneficiaries to private health exchanges[1], and also net of falling average contract values as beneficiaries choose cheaper plans. In the most generous possible case (ACA enrollment meets CBO expectations, and beneficiaries do not choose cheaper forms of insurance), we would expect commercial premium growth (5 year CAGR) of 7.5%. In the least generous case (ACA enrollment is 1/3^rd the CBO estimate, and average contract values fall by 20 percent[2]), we would expect 5 year commercial premium growth of 1.8% (Exhibit 1). N.B.: these estimates reflect anticipated commercial premium growth across the entire commercial market, regardless of where the commercial beneficiaries go for coverage (employers v. ACA exchanges, or employers v. private exchanges). To the extent that the large ‘national account oriented’ commercial HMOs lose share (beneficiaries shift from employers to either ACA or (more likely) private exchanges), their realized rates of premium growth will be lower

Conversely, growth fundamentals for the Medicaid HMOs are almost the exact opposite case – overall growth in potential beneficiaries is rapid, the Medicaid HMOs are gaining share of the growing number of Medicaid beneficiaries; and, average contract values for Medicaid beneficiaries stand to rise as dually eligible (Medicare and Medicaid) beneficiaries increasingly find their way into the Medicaid HMOs

Central to our Medicaid HMO thesis is the belief that all states ultimately will expand their Medicaid programs – either to an upper income eligibility limit of 100 percent of the federal poverty level (100 FPL, most likely scenario), or to 138 FPL. At present, some states have expanded to 138 FPL, and other states have not expanded at all. Former HHS Secretary Sebelius established a policy restricting the availability of enhanced federal matching for expansion enrollees only to states that had expanded ‘fully’ to 138 FPL. The purpose was to create an incentive for states to expand fully; and, to block states’ otherwise optimal strategy of expanding Medicaid only to 100 FPL[3]. We presume her strategy was intended to maximize states’ participation in the Medicaid expansion; however because hold-out states with the majority of potential expansion enrollees remained on the sidelines, it’s now apparent that refusing the enhanced federal match to states that might be willing to expand only partially (to 100 FPL) is leaving more than 9 million potential Medicaid beneficiaries without coverage. Former Secretary Sebelius’ policy was set to expire at the end of 2016. Because the policy arguably has not worked – states with most of the potential expansion enrollees refused to expand coverage – we see at least some chance that the Secretary nominee will revisit the policy. Whether she changes the policy or not, we still hold a mid- to long-term expectation of an ‘equilibrium’ wherein most states expand to 100 FPL; however with the change in HHS leadership we see at least some odds of the hold-out states expanding to 100 FPL even sooner

Assuming an eventual expansion across all states to 100 FPL, we would expect 5 year Medicaid spending growth of 11 percent; if all states expanded to 138 FPL (less likely) we would expect 5 year Medicaid spending growth of 17 percent (Exhibit 2)

Valuations in the Context of Growth Expectations

Relative longer-term growth expectations arguably are best reflected in forward price-earnings ratios; ignoring risks, higher fPE’s reflect higher growth expectations and vice versa. In Exhibit 3, the y-axis represents relative fPE’s on FY+3 consensus for the ten HMOs. The x-axis represents relative growth prospects as calculated by us … all we’ve done here is to multiply the percent of each HMO’s current revenues coming from any given segment (commercial, Medicare, or Medicaid) by our 5 year expectation for premium growth in each of these segments. In Exhibit 3, we assume none of the HMOs gain or lose share in any segment – so the commercial HMOs don’t lose share of employer sponsored beneficiaries as these beneficiaries shift to private exchanges, and the Medicaid HMOs don’t gain any further share of Medicaid beneficiaries. Exhibit 4 is the same as Exhibit 3, except that in this case we assume Medicaid HMOs continue to gain share of Medicaid beneficiaries – effectively what this represents is an assumption that states with lagging shares of Medicaid beneficiaries in HMOs reach the point of having an average share of Medicaid beneficiaries in HMOs. In both exhibits the 45 degree line is ‘par’, i.e. the point at which relative fPE (x-axis) accurately reflects relative growth prospects (y-axis). Stocks below the par line appear undervalued, and vice versa. Whether we assume the Medicaid HMOs continue gaining share of Medicaid (Exhibit 4) or not (Exhibit 3), it’s clear the Medicaid HMOs generally appear undervalued, both relative to ‘par’ and in particular relative to the commercial HMOs