Article Regulating Employment-Based Anything † Brendan S. Maher INTRODUCTION Benefit regulation has been called “the most consequential subject to which no one pays enough attention.”1 It exhausts judges, intimidates legislators, and scares off theorists.2 That need not be so. The reality is less complicated than advertised. Governments often consider intervention if markets fail to make some socially desirable Good X—such as education, health care, home mortgages, or pensions, for example— sufficiently available. One obvious fix is for the government to provide the good itself. A less obvious intervention is for the government to regulate employment-based (EB) arrangements that provide Good X as a benefit to employees and their fami- lies. In the United States, such employment-based interven- tions are massive: they affect trillions of dollars, billions in tax † Associate Professor and Robert D. Paul Scholar, University of Con- necticut School of Law; J.D. Harvard Law School; A.B. Stanford University. Earlier drafts of this work were presented and received valuable criticism at three events. I would like to thank commenters from the 2014 Har- vard/Stanford/Yale Junior Faculty Forum (particularly Joseph Bankman); from the University of Connecticut Faculty Workshop series (particularly Pe- ter Siegelman); and from the Boston College Faculty Workshop series (particu- larly Brian Galle). I would also like to thank—for everything—my father, Rob- ert James Maher (19452014), who recently died after a long struggle with ALS. All errors are mine alone, Dad. Copyright © 2016 by Brendan S. Maher. 1. On March 22, 2013, the Michigan law, business, and graduate schools held a joint conference entitled “Regulation of Benefit Plans: The Most Conse- quential Subject to Which No One Pays Enough Attention.” See Conference Schedule, U. Mich., Regulation of Benefit Plans: The Most Consequential Sub- ject to Which No One Pays Enough Attention (Mar. 22, 2013), http://www.bus .umich.edu/Conferences/US-Benefits-Law-A-Meta-Assessment/GetFile.aspx? paper_ord=635205. 2. Justice Souter apparently retired from the Supreme Court rather than adjudicate more ERISA cases. Jess Bravin & Evan Perez, Justice Souter To Retire from Court, WALL ST. J., May 1, 2009, at A1 (noting that one of Justice Souter’s reasons for retirement was a desire to be free of “numbingly technical cases involving applications of pension or benefits law”). 1257 1258 MINNESOTA LAW REVIEW [100:1257 breaks, and millions of people.3 They have been written into federal law for decades and generate constant litigation before the United States Supreme Court.4 Yet, while other regulatory interventions are well- theorized, employment-based interventions are not. There is no coherent account of employment-based interventions as a con- cept independent from the peculiarities of Good X or the rele- vant implementing statutes. This is a significant failure, and 3. The numbers are staggering. At the end of 2013, private pension as- sets totaled approximately 8 trillion dollars. See BD. OF GOVERNORS OF THE FED. RESERVE SYS., FEDERAL RESERVE STATISTICAL RELEASE Z.1, FINANCIAL ACCOUNTS OF THE UNITED STATES, THIRD QUARTER 2014, 84 tbl.L.117 (2014) (reporting the value of both “defined benefit” and “defined contribution” plan assets). And annual funding and paying for all employee benefits (not just re- tirement) totaled almost 3 trillion dollars in 2010, the last year in which relia- ble estimates are available. See EMP. BENEFIT RESEARCH INST., EBRI DATABOOK ON EMPLOYEE BENEFITS ch.2 (2012) (on file with author) (provid- ing an annual contribution and expenditure estimate of $2.8 trillion for 2010). Approximately 150 million and 60 million people receive EB health and retirement benefits, respectively. See CONG. BUDGET OFFICE, CBO AND JCT’S ESTIMATES OF THE EFFECTS OF THE AFFORDABLE CARE ACT ON THE NUMBER OF PEOPLE OBTAINING EMPLOYMENT-BASED HEALTH INSURANCE tbl.2 (2012), http://www.cbo.gov/publication/43082 (estimating that over 150 million people would receive EB insurance in 2012); Craig Copeland, Employment-Based Re- tirement Plan Participation: Geographic Differences and Trends, 2012, EBRI ISSUE BRIEF, Nov. 2013, at 1, http://www.ebri.org/ pdf/briefspdf/EBRI_IB_011 -13.No392.Particip.pdf (indicating that there were 61.6 million beneficiaries of EB retirement benefits in 2012). Because many EB arrangements are tax- favored, the tax expenditure associated with them is very high: projected to be almost 3% of the nation’s GDP over the 2013–2022 period. See Tax Expendi- tures Have a Major Impact on the Federal Budget, CBO BLOG (Feb. 3, 2012), http://www.cbo.gov/publication/42919 (estimating tax revenue loss for EB re- tirement and insurance plans). 4. The seminal statute is the Employee Retirement Income Security Act of 1974, Pub. L. No. 93-406, 88 Stat. 829 (codified as amended in scattered sec- tions of 26 & 29 U.S.C.). Legal recognition of EB approaches occurred before that, of course, but ERISA is the landmark statute. See, e.g., PRIVATE PEN- SIONS AND PUBLIC POLICIES vii (William G. Gale et al. eds., 2004) (“Tax incen- tives for employer-based pensions originated in 1921.”). Since 2004 alone, the Supreme Court has decided fourteen ERISA cases. See Fifth Third Bancorp v. Dudenhoeffer, 134 S. Ct. 2459 (2014); Heimeshoff v. Hartford Life & Accident Ins. Co., 134 S. Ct. 604 (2013); US Airways, Inc. v. McCutchen, 133 S. Ct. 1537 (2013); CIGNA Corp. v. Amara, 563 U.S. 421 (2011); Hardt v. Reliance Standard Life Ins. Co., 560 U.S. 242 (2010); Conkright v. Frommert, 559 U.S. 506 (2010); Kennedy v. Plan Adm’r for DuPont Sav. & Inv. Plan, 555 U.S. 285 (2009); Metro. Life Ins. v. Glenn, 554 U.S. 105 (2008); LaRue v. DeWolff, Boberg & Assocs., Inc., 552 U.S. 248 (2008); Beck v. Pace Int’l Union, 551 U.S. 96 (2007); Sereboff v. Mid Atlantic Med. Servs., Inc., 547 U.S. 356 (2006); Aetna Health Inc. v. Davila, 542 U.S. 200 (2004); Cent. Laborers’ Pension Fund v. Heinz, 541 U.S. 739 (2004); Ray- m ond B. Yates, M.D., P.C. Profit Sharing Plan v. Hendon, 541 U.S. 1 (2004). 2016] EMPLOYMENT-BASED ANYTHING 1259 one that has obscured clear thinking on the subject for decades. This Article offers a simple theory of employment-based inter- ventions that (1) explains the common conceit of all such inter- ventions and (2) provides a non-technical framework for evalu- ating any particular intervention, regardless of Good X. The gist of the theory (which for convenience I call “EB theory”) is easily stated. Regulatory interventions occur be- cause the government concludes there is a problem with the quantity, quality, or distribution of Good X; employment-based interventions occur because the government believes regulating the labor deal is an attractive way to fix it. EB theory offers a systematic way for observers to determine whether the gov- ernment is right. In so doing, EB theory makes the relative ap- peal (or insufficiency) of employment-based interventions vast- ly easier to understand. It also brings order and lucidity to a famously untidy subject. By gathering under the same tent ob- servations made across eras and disciplines, by laying bare as- sumptions about the “right” way to provide certain goods, and by suggesting promising paths for reform, EB theory promotes clarity in the national conversation. *** Part I provides some brief but necessary background. It supplies a short history of EB interventions in the United States, and considers the relevant features of the two most im- portant federal EB statutes: the Employee Retirement Income Security Act of 1974 (ERISA) and the Patient Protection and Affordable Care Act of 2010 (ACA).5 Even though employment-based interventions can include a variety of socially desirable goods, the original Good X, as Part I explains, was pensions. Abstract thinking was largely trained on questions peculiar to pensions, rather than on the inherent consequences of conscripting the employment bargain to convey some socially desirable Good X beyond wages. No or- ganized attempt to theorize the latter appears in the literature. That pattern repeated for the second great Good X: health care. Intense criticism has been leveled at EB health care approach- es, but with little consideration that such problems could be predictable manifestations of EB interventions generally. The 5. The Affordable Care Act is actually two pieces of legislation: the Pa- tient Protection and Affordable Care Act, Pub. L. No. 111-148, 124 Stat. 119 (2010), and the Health Care and Education Reconciliation Act of 2010, Pub. L. No. 111-152, 124 Stat. 1029. I follow custom by referring to them both as the “ ACA.” 1260 MINNESOTA LAW REVIEW [100:1257 ACA, while signaling continuing legislative fascination with EB interventions, has prompted little general theorizing on EB in- terventions beyond their role in health care. Part II pivots from the historical to the conceptual to offer a theory of EB interventions. The theory identifies the essential characteristics of EB interventions and builds a vocabulary to describe their advantages and disadvantages. The theory next evaluates in depth those advantages and disadvantages. Part II.A offers a simple conceptualization: an EB inter- vention is a legislative decision to improve Good X by use of the labor deal. This conceptualization incorporates two key premis- es: first, that the market has failed to optimize the quality, quantity, or distribution of Good X, and second, that it is desir- able to improve Good X through regulation of the employment relationship. Part II.A also describes common variations in the implementation of EB interventions. Part II.B considers why EB approaches might be justified. It builds the preliminary case for EB interventions by asking what Good X would look like in “the baseline world,” i.e., an unregulated market, and imagining how an EB intervention might improve things. Compared to the baseline, EB interven- tions admit of three potential advantages: (1) the use of sophis- ticated parties to aid employees in obtaining Good X; (2) the power of groups in purchasing Good X; and (3) use of the com- pensation deal as a natural decision point to promote Good X acquisition and planning. Comparison of EB interventions with other regulatory so- lutions is more difficult, because a conclusive comparison de- pends on the details of those competing options. One can, how- ever, identify intuitive advantages EB interventions might have over other families of regulatory solutions. Part II.B iden- tifies two. First, the labor deal seems a robust regulatory tar- get. People need to make these deals and will have trouble abandoning them merely because of some additional regulatory burden. Employers, moreover, are familiar with serving as compliance bureaucrats. Second, EB interventions are regula- tory solutions that preserve a meaningful role for market forc- es, and thus arouse less skepticism in many stakeholders than more invasive approaches. Part II.C builds the general case against EB interventions. Rarely will any seriously considered EB intervention make matters worse than the baseline world. Instead, the relevant brief against EB interventions consists of two strains of argu- 2016] EMPLOYMENT-BASED ANYTHING 1261 ment. The first is that the purported advantages of EB inter- ventions are weaker than they appear. The second is that there are good reasons to be skeptical of EB interventions compared to other regulatory alternatives designed to solve the problems of Good X. Stated briefly, little is to be gained from using employers and the labor deal as a regulatory nexus. Employers are not, on matters of Good X, meaningfully sophisticated; their expertise lies elsewhere. Extensive regulation will likely be necessary to prevent employers from making unwise decisions regarding Good X, particularly when non-expert employers engage truly sophisticated third-party providers of Good X, as they often will. To the extent employers are more sophisticated on Good X than workers, extensive regulation will be needed to prevent employers from taking advantages of employees, with whom they have, on the matter of compensation, an adversarial and unbalanced relationship. The likely quantum of regulation needed to mitigate potential employer incompetence or exploi- tation suggests that a more direct solution—cutting out em- ployers and directly regulating the providers of Good X—may be preferable to an EB intervention. EB interventions also suffer from two more subtle nega- tives. The first, regulatory fragility, comes from the strategic advantage employers have by being able to deliver a Good X marginally better than the baseline world. Regulatory efforts to improve Good X will be bound by employer threats to no longer offer Good X, a threat that increases in magnitude the worse the baseline world is. Given that an EB intervention is justified in the first instance by some problem with Good X in the base- line world, this threat will virtually always have currency. It gains further power still when employers remind regulators that they are not, generally speaking, in the business of provid- ing Good X. Such incidental providers of Good X are those we would most expect to abandon doing so if the going gets rough. The second subtle negative, opacity, operates to confuse stakeholders about who, precisely, is paying for Good X (and at what cost), and thus impedes accurate consideration of alterna- tive solutions that are more transparent about the cost of Good X. EB interventions also promote the mistaken belief that Good X is an employment issue, rather than a social issue. EB inter- ventions, however, are simply ways to solve problems with Good X. They signify no deeper relationship between employ- ment and Good X. 1262 MINNESOTA LAW REVIEW [100:1257 Part III demonstrates the utility of EB theory. The theory establishes and explains EB interventions as a distinct species of regulatory intervention, with identifiable traits. It therefore immediately makes coherent the inquiry into the merits of any proposed EB intervention. What is the problem with Good X that demands action? How will regulating the labor deal ame- liorate the problem? Is the ameliorative rationale based on em- ployer sophistication, group advantage, the behavioral effects of tweaking the compensation bargain, or some combination thereof? Will an EB approach lead to a deteriorating equilibri- um, i.e., where the alternative world of Good X is so undesira- ble that regulators have little leverage? How tolerable is the opacity that plagues many EB interventions? Not only does the theory promote fruitful comparison of EB solutions to other regulatory approaches, it distinguishes among possible EB interventions. For example, ERISA might be undesirable, but perhaps some other version of an EB re- tirement intervention is superior to both ERISA and its non-EB regulatory competitors. EB theory provides a roadmap for both types of inquiry. Importantly, the answers to the questions EB theory poses are not pre-ordained. They depend on the empirical judgment and priorities of those doing the asking. That important caveat aside, in Parts III.A, III.B, and III.C, I briefly consider some specific implications of EB theory. Part III.A uses EB theory to identify a potentially promis- ing path for reform. A perhaps under-appreciated strength of EB interventions is their ability to promote “segregative push- es.” Segregative pushes segregate a portion of wages and com- mit employees to spending those funds on Good X. A segregative push combined with a particular style of regula- tion—a “conduit” approach—might be the breed of EB interven- tion most likely to capture the attention of sincere reformers. A conduit system is one in which the primary role of the employer is to transparently withhold and transfer some amount of the employee’s pay to an account that the employee can only spend in a regulated, non-EB market for Good X. Conduit systems combine the attractive front-end aspects of EB approaches with a regulated but private non-EB market on the back end. Fur- ther study by scholars is warranted. Part III.B uses EB theory to shed light on ERISA and the ACA, suggesting comprehensible narratives for two infamously complex statutes. ERISA is a statute that underestimated the 2016] EMPLOYMENT-BASED ANYTHING 1263 dangers of conscripting the labor deal as a regulatory nexus, and fell victim to regulatory fragility. The ACA is a statute that attempted to create a non-EB market for health insurance while simultaneously choosing, perhaps for political reasons, to perpetuate an EB approach. The latter decision is certain to be reexamined in the not-distant future. Part III.C uses EB theory to offer a very short thought ex- periment about why education is a good provided outside of employment, and what it may say about the degree to which our views of EB interventions are colored by hidden assump- tions, which may or may not survive more considered analysis. Part III.D shifts from the specific to the general and ex- plains the broad value—to scholars, decision-makers, and stakeholders—in having an accessible EB theory. One pleasing side effect of EB theory is that it organizes a great mass of seemingly unrelated scholarship (on pensions, health care, in- surance, disability, and so on) that stretches back over a centu- ry. It is also a crucial first step in dispelling the fog that im- pairs mainstream understanding and discussion of EB approaches. Rarely has a mechanism so central to the processes of everyday life been shrouded in mystery and obscurity for so long. Understanding is valuable in and of itself, but particular- ly so when the subject is something that touches so many dol- lars and so many lives. I. HISTORY AND BACKGROUND The story of EB interventions in the United States can be told in three parts. The first part traces the organic, unregulat- ed rise of employment-based retirement and health care ap- proaches. The second part is the enactment of ERISA, which governs most private EB arrangements. The third part is the enactment of the ACA, which regulates EB health arrange- ments. Missing from the story is any overarching theory about EB interventions. I should offer an important clarification. Other goods be- yond health care and retirement income have been, could be, or are provided through EB arrangements. For example, one very important good provided in significant part through an EB model is disability insurance, i.e., wage replacement for those who cannot work because of a disability. And one good that could conceivably be provided through an EB approach, but is not, is unemployment insurance, i.e., temporary wage replace- 1264 MINNESOTA LAW REVIEW [100:1257 ment for involuntary loss of employment.6 Retirement and health care, however, are the two most important EB goods in America, as well as the ones that best convey the necessary background in advance of the theoretical analysis offered in Parts II and III. A. HISTORY (PRE-1974) The history of EB interventions orbits around two familiar things: retirement and health care. While today those subjects are on the minds of every aging voter and ambitious politician, in historical terms they are relatively recent problems. The EB story begins with retirement. Retirement. Retirement income, broadly, includes any in- come one relies upon after aging out of the workforce. Original- ly, however, the conception of retirement income was narrower: the pension. A pension is a fixed stipend paid by the govern- ment or one’s former employer. Pensions assumed social signif- icance during the late 19th century.7 Prior to that, most work- ers were farmers or artisans who participated in family businesses; if they lived long enough, they relied on younger relatives to continue the business and provide for them in se- nescence. As people began to work for enterprises they did not own (and began to live long enough to survive their careers), the need for post-employment income became apparent.8 Pensions—a temporal transfer of wages—were a market reaction to a workforce with a post-employment need. As work- ers realized the need for retirement income, the promise of a pension advantaged employers in the labor market. Pensions also appealed to employers for their own benefit, because pen- sions encouraged employees to make firm-specific investments of human capital that benefited employers.9 Pension arrange- 6. See infra note 184. 7. Civil War veterans received government pensions for service, and in 1875, American Express offered the first private pension. See Peter Blanck & Chen Song, “Never Forget What They Did Here”: Civil War Pensions for Get- tysburg Union Army Veterans and Disability in Nineteenth-Century America, 44 WM. & MARY L. REV. 1109, 1116 (2003) (describing the Civil War pension intervention); STEVEN A. SASS, THE PROMISE OF PRIVATE PENSIONS: THE FIRST HUNDRED YEARS 23 (1997) (describing the American Express pension). 8. See SASS, supra note 7, at 10–25; see also DAN M. MCGILL ET AL., FUNDAMENTALS OF PRIVATE PENSIONS 46 (9th ed. 2010) (discussing the transformation of the American labor force in the late nineteenth century). 9. MCGILL ET AL., supra note 8, at 23; see also JONATHAN BARRY FORMAN, MAKING AMERICA WORK 225 (2006) (explaining that a pension plan w ill typically “provide large financial incentives for workers to stay with a firm 2016] EMPLOYMENT-BASED ANYTHING 1265 ments, moreover, helped increase voluntary departure when the employee’s productivity declined due to age.10 Voluntary departure is for employers preferable to termination because the former furthers amicable relations with the workforce. It soon became clear that unregulated private pensions were insufficient to provide adequate retirement security for workers or, obviously, for the citizenry as a whole.11 The pre- World War II solution to the problem was to sidestep EB inter- ventions entirely; the passage of the Social Security Act in 1935 provided broad-based government support for retired and disa- bled Americans.12 The passage of the Social Security Act, however, did not obviate the need or desire for workplace pensions. Other fac- tors, such as the steady increase of income tax rates, as well as the wage and price controls of World War II, increased the ap- peal of EB pensions. Indeed, employee (and union) realization of the tax benefits and inherent value of occupational pensions contributed to their sharp rise in the 1950s and 1960s.13 By 1974, almost thirty million workers (approximately forty-four at least until they are eligible for early retirement”). 10. MCGILL ET AL., supra note 8, at 6; see LOUIS D. BRANDEIS, Our New Peonage: Discretionary Pensions, in BUSINESS—A PROFESSION 65, 67 (Small, Maynard & Co. 1914) (explaining the desire of employers to use pensions to amicably hasten departure of aged employees from the workforce). 11. See, e.g., Economic Security Act of 1935: Hearing on H.R. 4120 Before the H. Comm. on Ways & Means, 74th Cong. 219–21 (1935) (statement of Mur- ray Latimer, Chairman, Railroad Retirement Board), http://www.ssa.gov/ history/pdf/hr35latimer.pdf (testifying as to the inadequacy of the private pen- sion intervention and the pressing need for enactment of what would become Social Security). See generally MURRAY W. LATIMER, INDUSTRIAL PENSION SYSTEMS IN THE UNITED STATES AND CANADA (1932) (examining and criticiz- ing the private pension system). 12. Social Security Act of 1935, Pub. L. No. 74-271, 49 Stat. 620 (1935) (codified as amended at 42 U.S.C. ch. 7). Social Security was amended in 1939 to resemble more closely what it is today. See Kathryn L. Moore, An Overview of the U.S. Retirement Income Security System and the Principles and Values It Reflects, 33 COMP. LAB. L. & POL’Y J. 5, 8 (2011) (explaining the Social Secu- rity Act Amendments of 1939). Interestingly, a rarely discussed historical fact is that an EB “opt-out” was rejected by the 74th Congress. See Tamela D. Jerrell, A History of Legally Required Employee Benefits: 19001950, 3 J. MGMT. HIST. 193, 199–200 (1997) (discussing the failed Clark Amendment, a proposed amendment to the Social Security Act allowing companies who pro- vided comparable retirement benefits to opt out of the federal system). 13. See Alfred M. Skolnik, Private Pension Plans, 1950–74, 39 SOC. SECU- RITY BULL. 3, 4 (1976), http://www.socialsecurity.gov/policy/docs/ssb/v39n6/ v39n6p3.pdf (reporting that the percentage of private sector workers covered by pension plans grew from 22% in 1950 to 44% in 1974). 1266 MINNESOTA LAW REVIEW [100:1257 percent of the private sector workforce) were covered by pen- sion plans.14 By the 1960s, however, the uniformly good economic news of post-World War II America began to change for the worse, and pensions became a source of concern. Knowledgeable ob- servers warned of an emerging pension crisis, in which difficult economic circumstances would undermine promises made in headier times. Congressional study of the problem began in earnest and would reach fruition the following decade with the passage of ERISA in 1974.15 Health care. The other key subject in the history of EB in- terventions is health care. The relative sophistication of mod- ern medicine makes it difficult to forget how recently medicine was a primitive enterprise. The germ theory of sickness was not widely accepted until the later stages of nineteenth centu- ry; more Civil War soldiers, for example, perished from disease and illness than enemy weaponry.16 Subsequent advancement of medical knowledge, however, led to confidence that purchas- ing medical services was worthwhile.17 In the 1860s, several companies experimented with provid- ing “sickness funds” for injured workers.18 In the 1870s, other companies that required physically demanding and dangerous work—such as railroads, mines, and manufacturers—began providing company physicians to workers (for deducted wag- es).19 Successful direct maintenance of these programs proved challenging.20 14. Id. at 4 tbl.1 (reporting data on the number of private pension partici- pants). 15. See infra Part I.B. 16. See, e.g., Amir Attaran & Kumanan Wilson, A Legal and Epidemiolog- ical Justification for Federal Authority in Public Health Emergencies, 52 MCGILL L.J. 381, 400 n.65 (2007) (“[I]t was not until the nineteenth century that scholars such [as] John Snow, Louis Pasteur, and Robert Koch estab- lished the precepts of epidemiology, vaccinology, and germ theory—and that is where the scientific understanding begins.”). Numerous casualty estimates of the Civil War rank disease as a killer of more people than battle. See, e.g., Wil- liam H. Neinast, United States Use of Biological Warfare, 24 MIL. L. REV. 1, 11 (1964) (estimating a disease to battle death ratio of 1.45 to 1). 17. See PAUL STARR, THE SOCIAL TRANSFORMATION OF AMERICAN MEDI- CINE 259–60 (1982) (discussing the improving effectiveness of, and confidence in, medical services in the early twentieth century). 18. See JOHN E. MURRAY, ORIGINS OF AMERICAN HEALTH INSURANCE: A HISTORY OF INDUSTRIAL SICKNESS FUNDS 74–76 (2007). Several of the em- ployers, interestingly, hoped to make a profit on the funds. Id. at 75. 19. See STARR, supra note 17, at 200. 20. One reason may have been the primitive state of risk classification at