Private Health Exchanges: Why They’re Coming; What They Mean

Print Friendly, PDF & Email

Richard Evans / Scott Hinds / Ryan Baum


203.901.1631 /.1632 / .1627 richard@ / hinds@ /


April 2, 2013

Private Health Exchanges: Why They’re Coming; What They Mean

  • We believe commercial health insurance markets will shift toward an exchange-based format for employees of all income ranges and employers of all sizes, and that this transition will occur for most employees within three to five years
  • Rather than choosing from one or two relatively generous plans from a single (often national) issuer, the typical employee will have the choice of both less generous and more generous levels of coverage from multiple issuers, many of whom are likely to be local
  • This implies smaller average contract values (we believe employees will choose plans with average AVs of 65, a 21 pct decline from the current average of 82), share gains by local insurers at the expense of large national underwriters, and rising MLRs as each beneficiary is now both ‘in range’ of more underwriters, and more willing to switch plans each year based on price
  • The larger commercial HMOs (e.g. CI, AET, UNH, WLP) are most negatively affected; falling average contract values and rising average MLRs being compounded by a loss of share to local insurers. Because of these risks we now favor the smaller Medicaid-predominant HMOs (e.g. CNC, MGLN, MOH, WCG)
  • Where they exist, integrated health networks (IHNs) may have better reputations and generally do have cost advantages relative to other local issuers, and are likely to see large share gains. Being newly empowered to choose among multiple issuers, beneficiaries will demand revolutionary improvements in customer service from all health insurers
  • Pharmacy benefit managers (PBMs) are likely to see a shrinking market for pharmacy benefit carve-outs sold to large national accounts. Issuers of comprehensive exchange-based health plans are likely to outsource administration of pharmacy benefits to PBMs, but at ‘third party administrator’ margins below those in PBMs’ direct-to-employer relationships. Being less reliant on large employer-based national accounts, we believe CTRX is far less negatively affected than ESRX
  • Hospitals are negatively affected by the decline in average actuarial values, particularly as this affects collections. A beneficiary receiving care under an 82AV plan (the current ESI average) pays (out-of-pocket) $0.18 per dollar of care received, and hospitals’ collection rate on the beneficiary’s portion of the obligation is +/- 40pct over 90 days. The same beneficiary receiving care under a 65AV plan is responsible out-of-pocket for $0.35 per dollar of care received, and collection rates on this larger beneficiary obligation are likely to fall from the current level
  • Non-Rx Consumables companies (e.g. CFN and OMI in particular because of US focus, and also BDX, BCR, COV) offer exposure to rising per-capita utilization as coverage expands, without direct or even significant exposure to Hospitals’ pricing and collection pressures; accordingly we favor these names over Hospitals
  • Operators of private HIEs gain considerable market power relative to insurers, and are in a position to extract a higher share of each beneficiary’s contract value than is possible in their current roles as intermediaries / consultants between insurers and large employers. There are already far more private HIE operators than necessary, making it even less clear who will survive the inevitable shakeout; and, with the potential exception of TW, the publicly traded private HIE operators (MMC, AON) are far too large for their private HIE initiatives to be a primary feature of these companies’ earnings prospects

Everybody In?

We’re convinced that a majority of employees will soon purchase their health insurance on an exchange – of one form (public) or another (private)

Elsewhere we’ve shown that grossing up wages and having subsidy-eligible employees[1] purchase their health coverage on the state-based (i.e. ‘public’) health insurance exchanges is more efficient for both employers and employees[2], with the relatively narrow[3] exception of employees buying single coverage and having household incomes above 250 FPL

Thus given a binary choice between the current approach to employer-sponsored insurance (ESI) and the upcoming alternative of sending subsidy-eligible employees to the public health insurance exchanges (‘public HIEs’), employers should choose the latter. In reality the choice isn’t binary; private health insurance exchanges (‘private HIEs’) are rapidly emerging as a third option

Beyond the negative tax and penalty effects[4], the scenario of employers sending employees to the public exchanges suffers three substantial frictions:

1) social / political norms[5];

2) the question of where to find coverage for employees who are not subsidy eligible; and

3) some risk that tax-deductibility of ESI premiums is lost or limited for households who are not subsidy-eligible

And, from the employer’s perspective, less critical shortcomings of the public HIEs include an inability to exert influence over either the exchange or the insurance issuers, and the likelihood of excessively rapid household-level premium growth for beneficiaries on the public HIEs[6]

Why Private Exchanges, and Why Now?

As a response to rising health costs, employers are in the midst of a steady transition in health benefits, from defined benefit to defined contribution – just as occurred in years past in retirement benefits with the shift from pensions to 401k’s. The ACA both enables and catalyzes this shift to a defined contribution model for health insurance. The Act enables more efficient individual and small group insurance markets through a number of measures (e.g. guaranteed issue) that help shift the basis of competition among underwriters in health insurance markets – from an overwhelming emphasis on risk-selection pre-ACA, to a much greater emphasis on price and quality post-ACA. Also, the Act catalyzes a shift to defined contribution in a number of ways. Inadvertently, ACA encourages employers to shift subsidy-eligible employees to the public HIEs, in which case any related gross-up of wages is in effect a defined contribution. The existence of the public HIEs and the availability of premium subsidies for what is in effect two-thirds of working-age households brings into life an alternative market whose more desirable features – such as the option to buy cheaper (or more expensive) coverage from a potentially broad choice of issuers – arguably must spill over into the current ESI market. And, the Cadillac tax motivates plan sponsors (in this case employers) to both reduce effective premium growth and reduce their direct odds of having to pay the tax – both of which can be accomplished by shifting to health benefits to a defined contribution[7]model

A defined contribution is of course useless without a reasonably efficient marketplace for buying coverage as an individual. Employers cannot efficiently shift employees with household incomes above 400 FPL to the public HIEs; and, for many employers whose salary mix is heavily tilted to >400 FPL households, shifting their <400 FPL employees to the public HIEs may not be practical. Also, because the public HIEs generally will be state-specific, multistate employers presumably will face difficulties producing consistent benefit packages in a public HIE-based approach (44.5 pct of US employees work for a multi-state employer)[8]. Enter the private HIEs

Employees with household incomes <400 FPL are not eligible for premium subsidies on the public HIEs, so employers taking this option lose that potential arbitrage[9]. Other advantages of the private HIE approach may make foregoing this arbitrage worthwhile, particularly for employers with relatively small percentages of subsidy-eligible employees. On a private HIE, employers can:

1) shift to a defined contribution funding approach;

2) retain the tax deductibility of premiums (in the form of defined contributions);

3) have some level of influence over the management of the private HIE and/or the offerings of the issuers;

4) offer issuers on the private HIE far better risk pools than are likely to exist on the public HIEs[10];

5) continue to integrate health coverage seamlessly with the broader benefits package; and

6) realize higher participation rates[11] than if employees are purchasing coverage on the public HIEs

On top of these benefits, shifting employees to an exchange – private or public – is likely to reduce the amount of health insurance coverage the average employee purchases, thus immediately lowering average premiums

Most if not all of the major benefit consultancies[12] have announced and/or are operating private HIEs, and are joined in this emerging market by a large number of independent efforts[13]. If employers want to move to private HIEs, there are plenty of options, and plenty of apparent capacity

What This All Means: Individual, Local, Cheap and Comprehensive

Beyond the obvious implications for federal spending, we’re not terribly concerned with whether employees go to public or private HIEs – what matters most is that we expect a majority of employees ultimately will be in an HIE of one type or the other. Thus the relevance of the trend to private HIEs is:

1) this raises the odds that exchanges become the standard architecture of health insurance markets for a majority of beneficiaries;

2) this raises the number of beneficiaries to whom an exchange-based option is applicable; and

3) this accelerates the rate at which exchanges (are likely to) become the prevailing standard

Being convinced exchanges are the emerging ESI standard, the question becomes what this means for the broader health care landscape, i.e., what are the knock-on effects of employees buying their own coverage on an exchange, rather than having someone else choose it for them?

Benefit Managers Aren’t Real People

The US has roughly 6 million firms with paid employees, and the average firm has about 20 employees. If we exclude the 3.6 million firms with fewer than five employees, the average US firm has about 50 employees

Generally speaking the larger the firm the better the benefit package, and the greater the odds an employee chooses to enroll in employer-sponsored health insurance (ESI). As a result about 69 pct of ESI premiums are paid by firms with more than 100 employees, and nearly 40 pct of ESI premiums are paid by firms with more than 1,000 employees (Exhibit 1)

Within these (particularly larger) firms, health plans are chosen by a benefit manager or managers – thus by the time a typical employee chooses their health coverage, they have few (or no) options. In 2011, 84 pct of firms offered only one plan choice, 15 pct offered 2 choices, and 1 pct offered three or more health plan choices[14]. Larger firms are more likely to offer more than one option; as a result about half of employees have a choice of 2 or more plans, and about half can choose only one plan[15]. For most employees it’s Plan A, maybe Plan B, or nothing

On an HIE, decision-making shifts from the benefit manager to the employee. Assuming the HIE carries multiple plan options from multiple issuers, the employee’s choice set is considerably expanded – he or she can buy any of several plan types from any of several issuers. Thus on an HIE, plan selection begins to reflect employee preferences, rather than the preferences of benefit managers. As a result, at the very least we expect average AVs purchased to be smaller, ‘beneficiary-to-plan interfaces’ to become far more beneficiary-friendly (and more electronic), and average MLRs to rise[16]

Buy Local

Healthcare is locally produced and consumed, however ESI typically is bought and sold on a multi-state or even national level

It isn’t remotely practical for an employer with many employees in multiple states to negotiate health care arrangements in all of the communities in which its employees live, thus the natural tendency of such employers to use larger multi-state or national health insurance underwriters. For obvious reasons this optimal (buy national) approach for the employer may nor may not be optimal for the employee

In many cases the employee may prefer a more local underwriter for their health coverage; this is especially likely to be true when the preferred local underwriter and preferred local provider is the same organization – as occurs in the case of integrated health networks (IHNs) that both underwrite and provide healthcare. We often use the example of a national employer’s employee living in Salt Lake City, and receiving care from Intermountain Healthcare. Being a local provider (and underwriter), Intermountain isn’t efficiently able to contract with the multiple out-of-state home offices of the many national employers that employ persons in greater Salt Lake City, so many of the patients using Intermountain will not be Intermountain ‘members’; instead, these members pay for their care at Intermountain with coverage obtained from their employer’s national underwriter. However if these employees of national employers are shifted from traditional ESI to a private HIE, and assuming the private HIE offers employees the opportunity to buy local, it’s highly likely that they will[17]

Nationally, among the subset of IHNs currently underwriting coverage[18], roughly 40 pct of patient capacity is used for non-members. It’s very reasonable to believe many of these non-members would purchase coverage directly from the IHN if they had the option – which they probably will have if they are shifted to an HIE

Thus any shift from traditional ESI to an HIE carries with it a ‘localization’ or at least a ‘de-nationalization’ effect – beneficiaries will be more able to buy locally, and we expect many will prefer to

As an index of the relative risk faced by the larger health insurers, we calculated what would happen if as many of that insurers’ members as possible left to become members of local IHNs – the constraint being the amount of ‘non-member’ capacity represented by IHNs in that insurers’ markets. We express the result as a percent decline in enrollment, and the results are shown in Exhibit 2. To be clear this index overstates the absolute risks – mathematically, it assumes the major currently underwriting IHNs in each insurer’s markets fill themselves to membership capacity by stealing members only from the insurer in question. The value of the index is relative – some insurers are more at risk of an HIE-catalyzed shift to local purchasing than others

I’ll Have the Value Meal (Right-Size Me)

On average, the coverage offered under ESI has an actuarial value (AV) of 82, meaning the plan pays 82 pct of allowed medical expenses. For most employees this is quite a lot, for two reasons. For a large majority of employees, the dollar value of premiums paid into the plan will be quite a bit greater than the dollar value of care received as a result of being in the plan[19]. Second, because health costs are growing faster than wages, health cost inflation is reducing the remaining ‘non-health’ purchasing power of the household’s wages. Because of this, each year there will be a margin of households who need (or at least want) the marginal wage more than they want a corresponding margin of health coverage. The point is that an 82 AV plan is more coverage than many (we think most) households would buy if they had the option of something less expensive

At least on the public HIEs, there will be a range of AVs offered – 60, 70, 80, and 90. Logically, people will buy the plan most likely to give them back more value in coverage (and risk mitigation) than they pay in premiums. For the well, this is the 60 AV option; for the sick this is almost always the 90 AV option. We project an enrollment-weighted average AV of about 65 on the public HIEs – way below the prevailing 82 AV under ESI. If a similar breadth of AVs is offered on private exchanges (the range is mandatory on the public exchanges, but not on the private), we would also expect population-weighted AVs under ESI to fall to around 65

In 2012, Aon Hewitt developed a multi-carrier (UNH, Kaiser, Health Net, multiple BCBS providers) private HIE that enrolled more than 100,000 employees. Thirty-nine percent of enrollees chose consumer-directed (CDHP) plans on the private HIE for 2013 coverage; in 2012 only 12 pct of these employees had chosen CDHPs[20]. Forty-two percent of enrollees reduced their costs by selecting cheaper coverage than they had in 2012; 32 pct chose coverage of similar cost, and 26 pct chose more expensive coverage[21]

On top of buying less generous coverage on average, we also expect beneficiaries to purchase comprehensive ‘all-in’ packages of health coverage on the HIEs, rather than buying individual components (e.g. medical, prescription). Under ESI, employers often ‘carve-out’ elements of the broader health benefit package – particularly the prescription benefit. On the HIEs, we expect this to be much less common, with obvious implications for PBMs[22]

  1. Household incomes <400 pct of the Federal Poverty Level, or FPL
  2. Assume an employee only wants a given amount of after-tax and after-health spending income, plus a given level of health coverage – and is agnostic to whether her employer meets her expectations by: 1) providing cash wages plus employer-sponsored coverage; or 2) providing a higher cash wage but no health coverage, qualifying her for subsidized coverage on that state’s public health insurance. For employees with household incomes < 400FPL and buying either family coverage or single plus one coverage, the employer’s cost of funding a benefit package that meets her after-tax / after-health premium and health coverage expectations is lowest under option 2 – i.e. grossing up wages and sending her to the exchange is the most efficient solution. This is true even after accounting for penalties and tax effects. In reality, given two compensation packages with identical after-tax and after-health premium income and also equal health coverage, the employee will not be indifferent. On average, she will prefer option 2, as it provides more optionality (she can buy more coverage, less coverage, and/or coverage from a broad selection of underwriters). For full details please see: “Why Employers are Likely to Drop Health Insurance – A Simplified View”, SSR llc, July 11, 2011
  3. Single coverage accounts for about 26 pct of ESI premiums. We don’t know precisely what pct of single coverage premiums are paid by households above 250FPL, but see ½ as a reasonable guess. Thus in very rough terms, we estimate that moving subsidy-eligible employees onto the exchanges is the most efficient path for policies that represent +/- 80 pct of ESI premiums in the subsidy-eligible income ranges
  4. Which for most incomes and plan types are more than offset by federal premium subsidies
  5. We think these are more a matter of inertia than friction. The socio-economic reality is that marginally higher cash wages and federally subsidized exchange-based coverage should be preferred by employees shifted to the private exchanges; the political reality is that the first employers to take these steps almost certainly will be pilloried
  6. For two reasons: 1) indexing provisions within the Affordable Care Act (ACA); and 2) adverse selection
  7. Growth in the defined contribution amount is at least partially de-linked from the rate of growth in health costs: exchange-based employees arguably will purchase cheaper plans on average; and, for those employees who hit the Cadillac Tax threshold in a defined contribution model, the tax obligation falls on the employee rather than the employer
  8. US Census, 2009 County Business Patterns
  9. Which in 2010 dollars equals +/- $2500 per subsidy-eligible employee
  10. How much of an advantage this creates is not entirely clear. Because of ACA premium regulations issuers offering coverage both on and off of the private exchanges may be unable – or only partially able — to reflect the (presumably better) risks of private exchange enrollees with correspondingly lower premiums
  11. On a private exchange, employers place a fixed subsidy for purchasing coverage at the exchange, for the benefit of the employee. If the employee enrolls in coverage on the private exchange that fixed subsidy is applied to the employee’s total premium cost, and the employee pays any remaining balance. The employer can designate that (all or part of) the unused subsidies are returned to the employer rather than given to the employee. In the case of employers grossing up wages and leaving employees to purchase federally subsidized coverage on the public exchanges, employees have – and many would take – the option of going without coverage, and keeping the marginally higher wages
  12. E.g. Mercer (as part of Mercer Market Place), Aon Hewitt (The Corporate Health Care Exchange), Buck Consultants (RightOpt), Towers Watson (OneExchange, which incorporates Extend Health), and others
  13. E.g. Array Health (, Bloom Health (, acquired by WLP and a consortium of Blues), Choice Administrators (, Connected Health (, hCentive (, and many others
  14. Paul Fronstin: “Health Plan Choice: Findings from the 2011 EBRI/MGA Consumer Engagement in Health Care Survey”, Employee Benefit Research Institute (EBRI) Notes, July 2012, Vol 33, No. 7
  15. Ibid. Fronstin
  16. The current ESI market is dominated by relatively few larger underwriters who are able, by virtue of their geographic breadth, to offer coverage to multi-state employers. Pricing patterns in ESI are more reflective of cooperative than competitive pricing, i.e. with few exceptions pricing reflects a consistent mark-up on underlying medical costs. In competitive models, capacity and costs of capital tend to serve as constraints on the pricing cycle – in ESI they plainly do not. The ESI to HIE shift implies a parallel shift from relatively few larger issuers pricing multi-state offerings to larger employers, to substantially more (smaller and more local) issuers pricing local offerings to individuals. We would expect classic service industry pricing dynamics in this latter setting – i.e. we expect a shift to more competitive pricing (and thus higher average MLRs)
  17. If Intermountain is the family’s preferred provider, it’s almost certainly cheaper to be a member of Intermountain (i.e. buy health insurance from Intermountain) than it is to buy insurance from a third party for use at Intermountain
  18. A number of the largest national IHNs comprise this group (e.g. Kaiser Permanente, University of Pittsburgh Medical Center); collectively they account for about 8% of national commercial enrollment
  19. In the context of this specific argument we’re agnostic as to the proportion of premiums paid by the employer or the employee – in reasonably efficient compensation markets both are a reduction to cash wages. We appreciate that having more net payers than net recipients is very much the point with risk pools, and that our simple cash accounting ignores the economic utility of beneficiaries’ having off-loaded a non-trivial block of health risk. Still the point remains that under ESI’s current take-it-or-leave-it architecture, a lot of employees are having to buy more insurance than they really want
  20. Stephen Miller, “On Private Health Exchange, Choice Drives Satisfaction”, Society for Human Resource Management (SHRM) HR Topics and Strategy, March 21, 2013. (
  21. Ibid. Miller
  22. There are several moving parts here. As written, the ACA seems to point to only comprehensive packages being sold on the public HIEs. The private HIEs are less restricted; however the ACA definition of a qualifying medical benefit (one that allows the employer and beneficiary to avoid penalties) incorporates prescription drug coverage – thus we expect beneficiaries to be buying comprehensive packages on the private HIEs as well, though we’re less sure here than we are in the case of the public HIEs. Finally, the ACA’s MLR floors create an incentive for all health plans to carve out parts of the benefit – and the prescription benefit is an easy one to carve out. What this likely means is that plans still sell a comprehensive benefit to the beneficiary on the HIE, then turn around and carve out parts of the benefit to sub-contractors. Thus while there’s still a role for PBMs, the percentage of PBM business that is direct to employer presumably will fall dramatically – once HIEs are fully operational, the majority of PBMs contracts are likely to be with health insurance issuers
Print Friendly, PDF & Email