The unintended effects of the Medicare Part D low income subsidy
Introduction
Medicare is a public health insurance program for the elderly and disabled in the United States that covers over 50 million beneficiaries. Medicare consists of several parts. Parts A and B cover hospital and outpatient services, respectively, under a fee-for-service model. Part C allows consumers to switch from fee-for-service to government-subsidized managed care administered by private insurers. Part D, introduced in 2006, is a voluntary program that provides insurance for prescription drugs.
In 2014, Part D had an enrollment of 37 million individuals and its cost for the government was estimated to be $75 billion. The distinguishing feature of this program is the delivery of insurance exclusively through a choice-based private insurance market. The public intervention is limited to paying subsidies and setting the rules under which the insurers operate. Hence, Part D is an important testing ground for how the government can regulate a publicly financed privately delivered health insurance program.
This paper focuses on the intended and, especially, the unintended effects of the Part D subsidies on insurer pricing strategies. In Part D, subsidies are in various forms and account, overall, for 90% of insurer revenues, while premiums paid by enrollees only constitute the remaining 10% of revenues [20]. This paper shows how the regulations involving the low income subsidy (LIS), which Medicare pays to plans enrolling beneficiaries of limited financial resources, might induce distortions in plan pricing choices. By linking together two parts of the regulation, that is the algorithm through which the LIS amount is calculated and the rule according to which low income enrollees are assigned to plans, I argue that insurers are capable of and do distort premiums. In particular, since Medicare pays the premium of low income enrollees in full as long as their premium is not higher than a threshold amount called the Low Income Premium Subsidy Amount (LIPSA), premiums can be increased up to the LIPSA without losing low income enrollees. Hence, for plans with a high share of LIS enrollees, insurers will try to forecast the LIPSA and set premiums equal to it. Furthermore, since the LIPSA is endogenously determined as a weighted average of plan premiums, a second type of distortion can result when insurers offering more than one plan use some of their plans to bolster the LIPSA. Using plan level data for the years between 2006 and 2013, I show evidence consistent with the presence of both types of distortions. I then conclude discussing the negative effects of premium distortions and some remedies.
This study contributes to a small but growing literature on the determinants of premiums in Part D. In particular, the idea presented in this paper is further developed in [4], [5] to quantify the effect of the LIS-induced distortion on premium growth and consumer's welfare. Meanwhile, the current paper is concerned exclusively with establishing the presence of premium distortions and their evolution as regulations change between 2006 and 2013. The focus on supply side issues distinguishes this work from the majority of studies which focus on demand-side questions [1], [11], [14], [10], [15], [19]. Finally, since Part D low income enrollees are mainly Medicare–Medicaid dual eligibles, this study contributes to the analysis of the mechanism used to provide drugs to this important population group [17], [7], [9], [18].
Section snippets
Relevant regulations
In Part D, enrollees are divided into two groups, LIS receivers (35% of all enrollees) and “regular enrollees.” LIS beneficiaries are the dual Medicare–Medicaid eligibles as well as certain institutionalized enrollees and enrollees with combination of assets and income below certain thresholds. Every year, regular enrollees choose a plan and pay its premium. In contrast, the Center for Medicare and Medicaid Services (CMS) randomly assigns LIS enrollees to plans where they are charged a zero
Expected effects of the LIS regulations
Insurers can submit a single bid per plan that cannot be made conditional on the LIS status of the enrollee. Thus, the simultaneous presence of both regular and LIS enrollees requires that, to discuss the impact of LIS regulations, a more general framework premium determinants be presented first. To simplify, I will consider three sets of premium determinants – cost, demand and competition – and briefly describe their interaction with the LIS regulations.
First, premiums are linked to insurer
Data and method
This study uses publicly available data released by CMS describing enrollment and plan features for all plans offered between 2006 and 2013. These plan-level data allow us to observe enrollment and several plan characteristics, like the basic and enhanced components of the premium, the type of PDP and MA plan, the deductible, the type of coverage in the gap, the identity of the insurer, the drug formulary and the pharmacy network. Insurers are required to offer different plans across the 34
Results
(1) Passive distortion: Fig. 2 reports four histograms that illustrate premiums concentration at the LIPSA. The top-left plot reports the difference between premium and the LIPSA for all plans. The absolute frequency of plans that are within $1 below the LIPSA is twice that of plans that are between $1 and $2 below the LIPSA and three times that of plans that are within $1 above the LIPSA. No other bin in the histogram has a similar abrupt increase in its frequency. The next two plots show a
Conclusions
The evidence presented in this study reveals that the LIS regulations are associated with premium distortions. The natural question is then: What is the extent of harm from these distortions? Although a detailed answer is beyond the scope of this paper, there are at least two motives why premium distortion should be a major concern. First, one of the main pillars of a choice-based insurance market is that prices must guide consumers to make the best choice. This, in turn, is a pre-requisite for
References (20)
- et al.
Plan selection in medicare part D: evidence from administrative data
Journal of Health Economics
(2013) - et al.
Choice inconsistencies among the elderly: evidence from plan choice in the medicare part D program
American Economic Review
(2011) Government payments and insurer benefit design in medicare part D
(2014)Competition and the cost of medicare's prescription drug program. Report
(2014)Medicare part D: are insurers gaming the low income subsidy design?
American Economic Review
(2015)- et al.
The welfare effects of supply side regulations in medicare part D
(2014) - et al.
The effect of medicare part D on pharmaceutical prices and utilization
American Economic Review
(2010) - et al.
The distortionary effects of government procurement: evidence from medicaid prescription drug purchasing
The Quarterly Journal of Economics
(2006) Consumer inertia and firm pricing in the medicare part D prescription drug insurance exchange
American Economic Journal: Economic Policy
(2014)- et al.
Should drug prices be negotiated under part D of medicare? and if so, how?
Health Affairs
(2008)