Edited by Gerrit De Geest
Chapter 6: Health insurance
[In: Volume 2, Kenneth G. Dau-Schmidt, Seth D. Harris and Orly Lobel (eds) Labor and Employment Law and Economics]
In 2006, approximately 60 percent of the US population received health insurance coverage through their place of employment, or that of a family member (US Census 2006). Apart from Hawaii, which has an employer mandate, employers provide such coverage voluntarily.
Employers offer health benefits to help them attract and retain qualified workers, lower absenteeism, sick pay, and disability costs, and increase productivity. The substantive content of employer’s health insurance offerings, how many employees qualify for such coverage, and the allocation of its direct costs changed over time, in response to rising health care costs and larger macroeconomic trends.
From the employee’s perspective, health insurance helps fund necessary health care, promotes health, and protects employees from the financial costs associated with serious illness or injury (Gabel and Marquis 2002). Employees view health insurance as one of the most important fringe benefits they receive.
Whether employees have access to employment-based insurance varies tremendously. Roughly 60 percent of all employers offered such coverage to employees in 2007 (Kaiser Family Foundation and Health Research and Education Trust 2007). Employment-based coverage is much less available to those who work in certain industries (for example, agriculture, retail, and food service), and temporary and part-time employees (Farber and Levy 2000; Hyman and Hall 2001). Unionized employees are more likely to have access to coverage as well – sometimes through multi-employer plans operated by their union (Hyman and Hall 2001)
Size does matter; only 45 percent of employers with between three and nine workers offered such coverage, compared to 99 percent of employers with more than 200 employees (Kaiser Family Foundation and Health Research and Education Trust 2007). Most employers will also cover dependants of an employee for a premium that varies based on the number of covered dependants, although some employers do not cover dependants at all (Hyman and Hall 2001).
Health care coverage is expensive; average annual total premium cost in 2007 was roughly $4,500 for individuals and $12,000 for families, with direct contributions by the employer of $3,800 and $8,800, respectively (Kaiser p. 173Family Foundation and Health Research and Education Trust 2007). In 2006, health insurance accounted for almost 8 percent of total compensation paid by employers – roughly equal to the combined total paid by employers for Medicare, Social Security, unemployment insurance and workers’ compensation, and substantially exceeding the amount contributed by employers for employee retirement benefits (Bureau of Labor Statistics 2008).
Over the past several decades, employment-based insurance has evolved away from a fairly uniform model (broad coverage of all providers and most conditions; fee-for-service payment of providers; minimal or no deductibles and limited cost-sharing) to a much broader array of coverage options. The mid-1990s saw employers experimenting with various forms of managed care in response to rising health care costs. Modifications to the standard model included the use of capitation, coverage limited to restricted panels of providers, tiered benefits (depending on which provider one went to), prospective approval (or disapproval) of coverage, and dramatically increased direct cost-sharing. After these changes led to labor unrest and a legislative backlash, employers and insurers abandoned many of the more aggressive tactics. However, employees now directly pay more for – and receive substantially different coverage than was previously the case.
Many large employers offer ‘cafeteria’ plans to help employees pay for health-related expenditures. These plans allow employees to set aside a fixed dollar amount per year on a tax-free basis. The amount can be used to pay for co-payments and deductibles not covered by insurance, and health-related services that are excluded from coverage.
Roughly 20 percent of those who can obtain employment-based health insurance choose not to do so (Kaiser Family Foundation and Health Research and Education Trust 2007). Some are covered by a spouse’s plan, but many others simply opt not to secure coverage. This phenomenon has increased over the past two decades, and is related to the cost of coverage, and the options offered by the employer (Blumberg and Banthin 2001; Bundorf 2002; Chernew et al. 2005; Cooper and Vistnes 2003). Many others cannot obtain insurance through their place of employment at all, and are unable or unwilling to secure it through other means (for example, private (that is, individual or small group) and public (for example, Medicaid and SCHIP) insurance). The result is that approximately 45.7 million people are uninsured at any given point in time – and that figure (both absolute and relative) has trended up over the past decade. Roughly 85 percent of the uninsured live in families headed by an individual who works at some time during the year. Approximately 40 percent of the uninsured are full-time, full-year workers, or their family members.
2 Factors encouraging employment-based health insurance
p. 174Unlike most other developed countries, a majority of Americans obtain health insurance through their place of employment, or that of a family member. Prior to World War II, less than 3 percent of the US population received health insurance through the place of employment (Helms 1999). Since 1960, employment-based health insurance has covered between 60 percent and 68 percent of the US population, with a modest downward trend over the last two decades. What explains the rise, dominance, and erosion of employment-based insurance in the US?
The centrality of employment-based health insurance is an accident of history, largely explained by wartime labor policy, compounded by tax incentives. Dramatic increases in employment-based coverage came as a result of wage and price controls imposed by the Office of Price Administration during World War II. Employer contributions to insurance and pension funds were not counted as wages, and were accordingly excluded from the wage controls. The freezing of cash wages forced employers to compete for scarce labor by enhancing their fringe benefit packages. Health insurance offered a straightforward way for employers to sweeten their compensation package in a manner that would be appealing to potential employees, but not run afoul of the controls on wages.
As employers ramped up their use of employment-based coverage, the Internal Revenue Service issued a ruling in 1943, stating that the amounts paid by employers for insurance for employees did not constitute income to employees. However, employers could deduct these amounts as ordinary and necessary business expenses, and were not required to count them as wages for purposes of employment tax.
The IRS withdrew this ruling in 1953, but Congress quickly amended the Internal Revenue Code and reinstated the exclusion. The relevant provision of the statute flatly declares that ‘gross income does not include contributions by the employer to accident or health plans for compensation (through insurance or otherwise) to his employees for personal injuries or sickness’. Subsequent amendment made it possible for employees to pay for their direct contributions to coverage with pre-tax dollars as well, and allowed self-employed individuals to deduct health insurance premiums to the extent they did not exceed earned income from self-employment.
The result is that employees who obtain employment-based insurance (and self-employed individuals who qualify to deduct their premiums) can purchase coverage with pre-tax dollars, while those who obtain insurance through other channels must purchase it with post-tax dollars. The precise amount of the subsidy is a function of the marginal tax rate for any given taxpayer, but because the federal tax system has progressive rates, the benefits of this tax treatment are greater for higher-income taxpayers. Overall, p. 175this subsidy has been estimated to range as high as $200 billion in foregone tax revenue per year (Selden and Gray 2007; Sheils and Haught 2004).
For individuals who do not receive employment-based health insurance, health care spending (including insurance premiums) is generally deductible only to the extent it exceeds 7.5 percent of adjusted gross income (AGI). This deduction is available to those who itemize their deductions. For most taxpayers, the standard deduction is larger than their itemized deduction; in 2004, about one-third of all individual income tax returns had itemized deductions, but only 17 percent of these claimed a medical expense deduction, accounting for about 6 percent of all tax returns (Lyke 2005).
In an attempt to broaden the tax subsidy beyond employment-based health insurance, Congress created Health Savings Accounts (HSAs) in 1996, and broadened their availability in 2003. Similar to an Individual Retirement Account, individuals can deduct from taxable income contributions made to a HSA if they have health insurance with a deductible of at least $1,050 for individual coverage or $2,100 for family coverage. The insurance must also impose a $5,250 maximum out-of-pocket limit for individual coverage, and a $10,500 limit for family coverage. Unspent balances in an HSA grow tax free, and distributions from an HSA are tax free when used for qualified medical expenses and certain premiums.
Although there have been modest steps to broaden the tax subsidy beyond employment-based health insurance, the tax code creates a substantial financial incentive for employees to obtain coverage through their employer if at all possible. Many employees will also prefer to receive income in the form of tax-subsidized benefits, instead of as taxable salary – and will prefer richer and broader benefit packages, since an additional dollar of benefits costs them much less than a dollar, with the exact savings dependent on which tax bracket they are in (Gruber 2002). The savings are skewed in a variety of ways, by income level, firm size, and industry sector (Selden and Gray 2007; Selden and Bernard 2004). There is broad consensus that this tax subsidy results in employees receiving an inefficiently high level of health care coverage (Feldstein 1973; Feldstein and Friedman 1977; Pauly 1986).
Labor unions were a third factor driving the rise of employment-based insurance. During the post-war period, unions aggressively bargained for richer benefit packages, with health insurance at or near the top of their list. In industries in which unions were strong (for example, manufacturing and public-sector employment), the result was that employment-based health insurance became the rule. Employers with non-unionized workforces offered rich benefits to discourage their employees from unionizing.
As health care costs have increased dramatically over the past several decades, employer attitudes regarding employment-based health insurance p. 176have become substantially more negative. Many prominent employers see the relative and absolute cost of health insurance as a global competitiveness issue. Some are striving to use their purchasing power to improve the performance of the health care delivery marketplace, while others are wondering why they are in the business of purchasing health insurance at all. Annual increases in health insurance costs running at twice the rate of inflation have forced employers to decide between ‘hollowing out’ the coverage they offer, using payroll dollars to pay for premium increases instead of raises for employees, or dropping insurance entirely. Changes in FASB accounting rules have compounded this dynamic, by forcing employers to confront the expected costs of retiree health benefits.
Some employers have responded by dropping or drastically limiting retiree health benefits – and made those changes effective immediately. Large industrial firms have struggled with the legacy costs of collective bargaining agreements that promised generous insurance benefits. In 2007, General Motors divested itself of the costs of health care coverage for retirees (totaling $51 billion) by creating a Voluntary Employee Benefit Association and funding it with $30 billion. Other automobile companies appear likely to follow. Although most employers plan to continue offering health insurance to their employees for the foreseeable future, the ever-increasing costs place considerable pressure on that decision.
3 Consequences of linking employment and health insurance
The linkage between employment and health insurance has both benefits and costs. On the benefit side, employers serve a useful pooling function, by creating reasonably stable risk pools that are not created in order to obtain insurance (thus diminishing adverse selection problems). Employees are also lower risk than the general population, since the individual obtaining coverage must be well enough to work. Because employment-based coverage is underwritten on a group basis, it is less expensive than individual coverage. Finally, employers serve an administrative and record-keeping function, again lowering the costs associated with obtaining coverage.
Linking employment and health insurance also has substantial costs. There can be sequencing difficulties when an employee changes jobs, or loses their job (Gruber and Madrian 1977). Many health insurance policies contain waiting periods or exclusions on pre-existing conditions, which chill job mobility (‘job-lock’), although federal legislation has lowered the frequency and severity of this problem. A worker might also choose to stay in his current job if the substantive terms of insurance coverage are particularly valuable to the worker or his family, even though another job might offer greater opportunities or a higher salary.
There are also agency problems with having employers make decisions p. 177about coverage for employees. Changes in coverage design can induce disruption and dislocation costs for employees, with the burdens of those costs disproportionately falling on those with chronic conditions requiring specialized care. An employer may care greatly about conditions that affect its most highly valued employees, but show less consideration for conditions that disproportionately affect employees who are fungible, or work in a division slated for sale or closure. Incentive mismatches also affect issues of quality. Because employers internalize only a portion of the benefits of better quality care, they have less incentive to favor any particular quality enhancement than do employees as a group. Stated more concretely, because plans are a ‘bundled’ product aimed at a diverse workforce, the alternatives that any given employer offers frequently do not include desired and desirable features from the perspective of any given employee, while also including features an individual employee may regard as a waste of money.
The linkage between employment and health insurance creates one final difficulty. Economists agree that employer contributions are just another form of compensation to employees – and increased costs of coverage in the long run result in smaller wages (and wage increases) for employees. However most employees (and some employers) believe that employers are footing the bill for the coverage that employees receive. The result is that employees are relatively indifferent to the cost of their health care coverage (at least to the extent that their employer is the one writing the check), while employers are extremely concerned about the cost of providing coverage for their employees. This lack of transparency creates a set-up for conflict between employers and employees about the nature and cost of coverage.
4 Regulation of employment-based health insurance
Employment-based health insurance is regulated by both state and federal law, with the exact regulatory framework determined by coverage structure. Pursuant to the McCarran-Ferguson Act, states have primary regulatory authority over insurance sold to state residents. The Employee Retirement Income Security Act (ERISA) has three relevant provisions: the preemption clause, the savings clause, and the deemer clause. The pre-emption clause broadly pre-empts state regulation of employment-based health insurance. The savings clause saves from pre-emption bona-fide state regulation of insurance – determined by whether the statute is directed at insurer’s insurance practices, and by whether it affects insurer-insured risksharing arrangements. The deemer clause precludes employee benefit plans from being treated as insurers for purposes of the savings clause.
In combination, these three provisions mean that employment-based p. 178health insurance is subject to state-level regulation only to the extent the employer provides coverage by purchasing a state-regulated health insurance contract (‘insured plan’). Conversely, employment-based health insurance is not subject to state regulation if the employer self-funds the coverage it provides to its employees (‘self-funded plan’). Of the 177 million Americans who obtain health insurance through their employer, 45 percent are in an insured plan, while 55 percent are in a self-funded plan (Kaiser Family Foundation and Health Research and Education Trust 2007).
In practical terms, this framework means that the 80 million Americans who obtain coverage individually or through an employer’s insured plan are subject to both state and federal regulation, while the 97 million Americans who obtain coverage through an employer’s self-funded plan are subject only to federal regulation.
How have the federal and state governments exercised this regulatory authority? At the federal level, there has been relatively limited direct regulation of health insurance. When ERISA was enacted in 1974, it focused on pension plans, and imposed no substantive regulations on employment-based health insurance. Over the intervening 34 years, there have been a few new substantive regulations, including requirements prohibiting ‘drive-through’ deliveries,1 requiring parity in coverage of mental health treatment,2 imposing limits on the use of pre-existing condition exclusions,3 and prohibiting genetic discrimination4 Because ERISA pre-empts state law, but does not impose much in the way of substantive regulation, this framework means that self-funded employers have operated in a virtual regulatory vacuum. Self-funded employers are more likely to operate in multiple states, so this regulatory vacuum has meant such employers can implement uniform coverage arrangements without worrying about stateby-state regulatory variation. Stated differently, the current framework provides self-funded employers with virtually complete freedom to design and implement whatever health care coverage they desire – including spending as little or as much as they want.
At the state level, there has been a massive amount of regulation. As Figure 6.1 demonstrates, there are three distinct relationships that can be regulated: the relationship between the insurer and the physician/provider (Type I regulation); the relationship between the physician/provider and the patient (Type II regulation), and the relationship between the patient and the insurer (Type III regulation).
p. 179Examples of Type I regulation include ‘any willing provider’ legislation, restrictions on compensation mechanisms and prohibitions on ‘gag clauses’. Type II regulation includes mandated disclosure of qualifications, results and incentives to limit care. Type III regulation includes mandated coverage of certain benefits, such as alcohol treatment, and post-partum stays, and provisions affecting the circumstances and price at which insurance may be offered (including guaranteed issue and community rating).
States have adopted numerous Type I and Type III regulations, but relatively few Type II regulations. The number of Type I and Type III regulations also appears to have grown dramatically over time. The most common Type I regulations are any willing provider/freedom of choice legislation covering chiropractors (46 states), psychologists (44 states) and optometrists (43 states) (CAHI 2006). The most common Type III regulations are mandated coverage of newborns (50 states), alcoholism treatment (45 states), diabetic supplies (47 states), breast reconstruction after mastectomy (48 states) and mammograms (50 states) (CAHI 2006).
To summarize, the regulatory framework that applies to employment-based health insurance varies tremendously, depending on whether the p. 180coverage is insured or self-funded, and if the coverage is insured, by which state the policy is purchased in.
5 Employer mandates and ERISA
As noted previously, Hawaii is the only state with an employer mandate. Employees who work 20 hours a week or more for more than four consecutive weeks must be given coverage that does not cost them more than 1.5 percent of their salary. This employer mandate is not pre-empted by ERISA because of an express statutory exclusion passed by Congress in 1983.
Legislators in other states and in Congress have expressed considerable interest in using a similar approach to solve the problem of the uninsured. Such ‘pay or play’ strategies require employers to either provide health insurance to their employers or pay a tax/penalty. Maryland and Massachusetts have already enacted such legislation, as has the City of San Francisco and Suffolk County, New York.
These pay or play statutes required employers to spend at least a specified percentage of their payroll (Maryland) or a specific amount per worker per hour (San Francisco and Suffolk County) or their ‘fair share’ of the cost of coverage (Massachusetts). The first three were challenged as pre-empted by ERISA. The Fourth Circuit Court of Appeals held that Maryland’s statute was pre-empted.5 A similar San Francisco ordinance was held pre-empted by the Federal District Court, but the Ninth Circuit reversed, and upheld the ordinance.6 The Suffolk County ordinance was also enjoined by the District Court.7 The Massachusetts statute has not yet been challenged, and the modest imposition imposed on employers makes it less likely it will be pre-empted under ERISA (Monahan 2007).
It is important to note that the Maryland statute and Suffolk County ordinance, although facially neutral, were nonetheless carefully targeted: the Maryland statute was widely known as the ‘Wal-Mart bill’, since it was drafted so as to exclude every other company in the state, and the Suffolk County ordinance targeted non-unionized grocery stores. Both were strongly backed by local and national unions.
The future of pay or play initiatives will vary greatly, depending on what the Supreme Court does (if anything) with the Ninth Circuit decision in the p. 181San Francisco ordinance case. Congress could also amend ERISA to allow such initiatives, as it did with Hawaii.
Newborns’ and Mothers’ Health Protection Act of 1996, 42 USC §300gg-4 (2003).
Mental Health Parity Act of 1996, 29 USC §1185a and 42 USC §300gg-5 (2003).
26 USC §9801 (2003); 29 USC §1181 (2003); 42 USC §300gg.
Genetic Information Non-Discrimination Act of 2008, Public Law 110-233.
Retail Industry Leaders Association v. Fiedler, 475 F.3d 180 (4th Cir. 2007).
Golden Gate Restaurant Association v. City and County of San Francisco, 546 F.3d 639 (9th Cir. 2008).
Retail Industry Leaders Association v. Suffolk County, 497 F. Supp. 2d 403 (EDNY 2007).