INTRODUCTION Jonathan Gruber's article deals primarily with political and economic issues posed and resolved by health insurance reform recently adopted in Massachusetts, which gives only a tiny role to employer mandates. For good reasons, some of them special to Massachusetts, Gruber follows form of state plan in giving only small consideration to employer mandates to pay for workers' insurance; he pays much more serious attention to state design and control of insurance options and state determination of financing and subsidy distributions. In my remarks I will to some extent reverse relative weights by discussing and commenting briefly on design and financing but then discuss at some more length the dog that almost did not bark in Massachusetts: pay or play employer mandates and their economic and political implications for other states. Some Preliminaries Before I turn to question of employer mandates, I want to express some concern about two of points made in Gruber's article, one an obiter dictum that nevertheless raises an important issue about net benefits of employment-based coverage as it is usually offered, and other that links that issue to a key feature of Massachusetts plan, an organization called Connector that collects and assembles a variety of state-approved insurance options for employers and individuals seeking coverage. Gruber mentions (p. 453) low 1.3 percent administrative cost reported for universal coverage in Canada and suggests that there might be some merit in single-payer road not taken in Massachusetts. (A comparable figure for U.S. Medicare, of about 2 percent, is also often cited.) And, indeed, one of key factors determining desirabfiity of a set of insurance options is insurer administrative costs and profits that might occur. As others have noted, there are some statistical and definitional objections to imagining that very low administrative loading percentage reported for Canadian plan could be exhibited by arrangements like those in Massachusetts that rely on private insurance for administration. One might note, for example, that costs of research and design efforts, which are included in private costs, are largely overlooked in government's budgets. Research done directly by staff or debates in federal and provincial legislatures about health insurance, have a cost that should be included. The more serious costs that a tax-financed single-payer system avoids are high selling and billing costs that accompany voluntary plans that display great variety in form and type of coverage. The great bulk of cost of single-payer coverage comes from involuntary taxes. Collecting taxes is generally much less costly than persuading workers or employers to part with premiums, but this apparent advantage conceals a much larger cost. Because practical taxation links agents' tax liabilities to a tax base (wage earnings plus other income) that can be altered by individual choices about how much to work and how to receive income, they generate what economists call an burden: a loss in value of total economic activity traceable to distorted incentives. To be sure, it is overall pattern of taxes and redistribution imposed by government at all levels, which determines marginal excess burden, but if a publicly provided health benefit were to be financed by increases in taxes following usual distributional pattern (rather than uniform lump sum head tax add-ons), moving nearly a trillion dollars of national health care spending from private to public sector would generate a substantial distortion cost. This cost is estimated empirically to be at least 25 percent of total collection at prevailing tax rates in United States, and probably would be similar in Canada. Obviously, this hidden but real cost of transferring financing from private payment to government taxation swamps any savings from lower explicit insurer administrative costs. …