Introduction

The Affordable Care Act (commonly termed Obamacare) has and continues to experience substantial political and economic challenges to the “Exchange” market for individual (non-job-related) insurance it is trying to build to cover the uninsured. One of the law’s most popular features by far is the promise to make insurance available to those who have already become above-average risks , given age. The challenge to this feature is where to find the money to pay the difference between the average premium charged and the much higher expected value of benefits for high risks , for any nominal insurance policy. The mechanism built into the law, so-called “modified community rating” in which the same premiums are charged to below average risks has become the most serious flaw in the ACA framework, as defections by lower risks from insurance markets have resulted in high and growing premiums for Exchange coverage.

Another feature of Obamacare , which is less popular because it is perceived to be weak, is its efforts to control the rate of growth of medical spending. While it did make some apparently successful efforts to slow the growth of government payments for Medicare , it has not managed to bring private sector spending growth down to the level of GDP growth, and specifically been ineffective in curtailing the feature most commonly agreed to cause excessive private spending growth , the tax subsidy to employment-based premiums.

Finally, the cost of subsidies to the formerly uninsured has provoked taxpayer backlash to such an extent that repeal of the entire program almost succeeded. Is there a way to redesign subsidies and the distribution of methods to pay for them that might have had a different effect? James Buchanan ’s work offers insights on each of these three issues. We focus here on his contribution to “positive political economy ,” first outlined in 1949 (Buchanan 1949), especially the idea of “fiscal illusion ” treated masterfully in Chapter 7 in Public Finance in Democratic Process (Buchanan 1967).

For the first problem, Buchanan ’s thoughts can explain why there is concern for insurance coverage and premiums for people who are not poor but who happen to be above average risks , why Obamacare selected the politically expedient but flawed solution of modified community rating, and what alternatives might repair that flaw but still have political challenges. It can also help to explain why the most effective cost containment device—closing the tax loophole that provides high benefits to the rich while encouraging the use of expensive low value medical care has proven so politically durable. Finally, it can suggest alterations in the design of subsidies and taxes to increase the number of citizens who obtain net benefit from the program and thus improve its future political prospects.

Individual insurance markets before Obamacare generally operated in ways consistent with competitive insurance markets: premiums for a given policy with given coverage varied across buyers depending on insurer estimates of their health risk and other determinants of use of high priced care. Adverse selection did occur but was limited by insurance provisions that asked applicants about all prior use of medical care or medical conditions for up to 5 years, and rescinded coverage in the case of fraud. Among highest observable risks (say, the top 2%) there was more danger of adverse selection since the very high premiums they would have to be charged would only be paid by those who were likely to use care–so such persons were typically rejected for coverage. The great bulk of applicants including those at much-above-average risk could obtain coverage at some premium if they searched (Pauly 2010). This was less true in a small minority of states that required insurers to charge the same premiums to all risks (termed “community rating”), where either many more applicants were rejected for coverage (if that was permitted) or individual insurance markets virtually disappeared.

One of the major goals of the ACA, with support across the political spectrum, was to prevent this discrimination in favor of low risks and against high risks by requiring insurers to cover everyone buying an individual policy at community rated premiums that could vary only by age, location, and smoking status. Why was there support for this provision (which has gone on to cause so much mischief in the operation of Obamacare ), especially since risk rated markets had extended coverage to many high risks, to such an extent that, controlling for income, there was very little relationship between risk level and being insured? After all, because of the shared nature of the distribution of risks, there were many more low risks who stood to lose a little and relatively few high risks who would gain a lot.

Buchanan ’s work offers several insights which, when combined, help to explain the peculiar political economy of this policy. There was an intense minority (of chronic illness advocacy groups) lobbying for high risks, but no one to speak for non-poor lower risks (who were not all young) who were being grossly overcharged relative to the benefits they could expect to collect. However, the stronger argument focuses on the rest of the population. It seems to have been the case that many people who would not be affected by these changes in the individual insurance market (those on Medicare or obtaining insurance through an employment-based group) favored doing something about this “unfairness” (of possibly actuarially fair premiums). They were moved by stories of families with a high-risk member who suffered either from higher than average premiums or higher than average out of pocket bills, depending on whether they bought insurance or remained uninsured. At one level, the concern of these members of the community for others in the community who were either under consuming medical care or under consuming other key items of consumption was a kind of “externality per se in consumption” that Buchanan had identified, even though both health insurance and medical care (except for care for contagious conditions) appear to be pure private goods (see Buchanan 1968). More deeply, they may have felt that the “rules of the game” that they would have chosen behind a veil of ignorance might have embodied at least some transfers from the lucky who stay healthy to the unlucky who are hit with a chronic condition. Probably their willingness to pay more taxes to finance these transfers (even taxes on their insurance premiums that were part of the ACA since repealed) was modest, but they would prefer that high risks be covered by insurance compared to the situation with unsubsidized and largely unregulated individual insurance markets.

The other feature that doubtless contributed to the popularity of this provision was uncertainty—ranging up to total ignorance about who would pay to make up the difference between the average premiums and the much-above-average medical claims of the high risks. Some people, it is clear, thought it both likely and desirable that rich insurance companies should pay for the high risks they were forced to cover. Other more reflective people may have believed that premiums would need to be increased (compared to risk rated premiums) for low risks buying individual insurance, a small and diffuse minority of all citizens. This phenomenon too was anticipated by Buchanan in Cost and Choice (1969), where he notes that “cost” is only meaningful if agents have a choice that they understand.

Thus even in a simple majority rule model one has all the ingredients to combine both fiscal illusion with an influential intense minority. Those bearing the bulk of the cost of this increased generosity and fairness were a tiny misinformed and disorganized group while the great majority both mildly favored it and did not expect to pay.

As is often the case, the aftermath of putting this provision into action did not exactly match the rosy scenario envisioned by its supporters. For one thing, the increase in premium required an increase in taxpayer support for subsidies to lower income people so they could afford and be willing to buy insurance—so the rest of the population did pay, along with the small number of high income lower risks that kept buying. For another thing, though this widespread pattern of subsidies did to some extent diminish adverse selection , there apparently were enough non-poor low risks who realized they were facing a bad deal and so did not take individual coverage. Insurers initially had difficulty in covering their claims—and so many of them pulled out.

The so-called “individual mandate” in the ACA (really a modest penalty for failure to purchase coverage by the middle class and above) was not very effective at keeping low risks in the ACA pool (despite having had some earlier success in Massachusetts)—many unsubsidized low risk buyers apparently dropped or avoided overpriced coverage even at the cost of the penalty, and others sought coverage through employment groups of mostly low risk well off workers for whom premiums were allowed to be low.

Here again, there was a public choice puzzle. Apparently, there was little support for a large penalty that would have stabilized the community rated market covering high risks that so many claimed to favor. “Policy” illusion, a failure to connect the dots and therefore to blame insurance companies, rather than the policies they themselves had chosen, probably contributed to this behavior by an “irrational majority.” The problems associated with defection of low risks continue to bedevil individual insurance markets; the premiums charged by the remaining firms have now gotten high enough to cover claims, but the abolition of the individual mandate may destabilize those markets for purchases by the remaining unsubsidized buyers whose current behavior is not well understood and whose future behavior is unpredictable.

For the record, there are alternative ways to generate stable subsides to low and middle income high risks if that is what externality-affected voters want to do. One could make the penalty attached to a mandate high enough to guarantee coverage, require individuals who are low risk to purchase guaranteed renewable coverage at moderately higher premiums that will keep premiums down should they become high risk as long as they have maintained continuous coverage, or simply fund a high-risk pool with general revenue financing. Yet none of these has as yet garnered political support.

We still have a puzzle then. The low risks in individual markets who are not poor have dodged much of the burden for paying for the high risks. Instead, once one unpeels the complex structure in Obamacare , it turns out that much of the payment to help high risk was actually being made by the general taxpayers who are funding the subsidies to lower income average and low risks in exchanges. Wouldn’t they prefer to pay the subsidy directly to high risks rather than go through the channels of first subsidizing exchanges in order to subsidize middle and upper class high risks? Perhaps the explanation is precisely that the subsidy is so filtered through so many channels that taxpayers cannot see the tax price of their support and distinguish it from the subsidy they might want to pay to lower income average and lower risks on externality grounds; they cannot change what they cannot unravel—and a promise to lower exchange premiums for other risks if high risks could be skimmed off into a well-supported high risk pool is just not credible. The confusion caused by complexity benefits politicians who can point with pride to benefits from Obamacare —reasonably priced coverage for high risks, insurance at no cost for children up to age 26 whether dependents or not—and claim that no identified group has to pay for this.

The Loophole That Refuses to Close, and Spending Growth That Refuses to Stop

The recent tax reform legislation has been controversial, but there is considerable support for its successful effort to curtail at least some of the tax loopholes that provide poorly designed tax subsidies to some activities, such as high state and local taxes and home mortgages for second homes. What is perhaps surprising is that reform failed to tackle (or even mention) what is in the opinion of many tax experts the most troubling loophole of them all: the exclusion from taxation of insurance premiums arranged as part of employment-based group insurance. Indeed, the new law further postponed a poorly designed partial limit on this subsidy—the Cadillac tax.

The original Obamacare design (and subsequent Congressional follow-up) did have another feature designed to limit growth in government spending on the Medicare program, perhaps because (as will be discussed in more detail below) the bulk of the direct funding for subsidies to insurance for uninsured under 65 was to come from “savings” on Medicare . Specifically, the ACA put ad hoc limits on the growth in reimbursements for hospital care under Medicare , some motivated by an expectation that fewer uninsured would reduce the charity care burden on hospitals and some motivated by a hope that hospitals would find new “productivity improvements” that would allow them to survive lower prices. This strategy was extended to physician services in the Medicare physician payment reforms enacted in 2016: a lower growth of payments (especially after 2025 or so), linked to incentives to maintain quality and yet reduce the volume of services.

Things have not worked out perfectly—the failure of many states to expand Medicaid coverage left their hospitals with the same charity care burden but lower Medicare payments, and hospital financial status on average has worsened somewhat—but the slower growth of Medicare outlays has materialized without obvious reductions in quality or access, at least for the moment. The government’s ability to cut payments without doing much harm implies that pre-reform payments were higher than they needed to be, but the main message for voter choice is that lower payments to providers generate little objection from the majority of voters—only those tied to or employed in the health care industry potentially suffer, and lower wage growth and job growth has yet to materialize.

However, much of the burden for cost containment in the private sector was to come from changing the tax treatment of employment based health insurance. Previous research had shown that the generosity of health insurance chosen responded to the price of insurance, and that more generous insurance led even non-poor consumers to use substantially more medical care and to make less effort to find low priced sellers. Whether spending growth would be higher with more generous coverage once any effect of expanding coverage wore off was less definitively established, but there was some evidence in support of an effect on long run growth of spending. Insurance obtained as a work-related group benefit was excluded from most federal, state and local taxes, so the consequence was to lead to more generous coverage than would have been the case had compensation received in the form of partial payment of premiums been taxable and had any explicit worker premium not also been tax shielded under a cafeteria plan.

Phelps and Parente (2018) have recently produced updated estimates of the effect of the tax subsidy on amount of insurance demanded and the subsequent effect of more coverage on medical care spending; they find that the short run effect of removing the tax exclusion at the margin would be to reduce spending on insured services by about 20%, with the lower spending (or sacrificed wages) going for insurance offset by higher money income spendable on other goods and services of value.

However, as with any tax advantage, those who are benefitted by that advantage—compared to its removal and other taxes remaining the same—will resist attempts to constrain or eliminate it. Obamacare did surprisingly include an awkward attempt to offset the exclusion, through a provision that would impose a 40% sales tax (to be paid by insurers, or employers in the case of self-insurance) on premiums in excess of some target amount at some future date. This Cadillac tax probably survived into law because the revenue it would generate (either from the tax itself or from higher income taxes if money wages rose as employer-paid premiums were cut back) helped to finance the subsidies to lower income individual insurance, and because the primary beneficiaries of the subsidy were higher income workers (with high marginal tax rates). However, fierce opposition from unions (who generally negotiated for benefits-rich compensation packages) and from people living in areas where medical costs were high has led to postponement of implementation of the tax until 2024. Over the longer term, the tax provision would have eventually hit a large number of households (Herring and Lentz 2011).

So the median person (whatever that would mean in this context) might be worse off if the exclusion disappeared and nothing else was changed. No voter was fooled by the delusion that the additional tax collections would come out of insurer or employer profits; in this case, workers and their unions pierced the cloud of fiscal illusion to register a definite “no.” However, the fact that the exclusion leads to a non-(Pareto) optimal outcome means that it should still be possible to find an alternative arrangement which could be beneficial to many, even to those who use the exclusion. Phelps and Parente have suggested one: the simple idea of eliminating the exclusion entirely but then using the proceeds to lower marginal income tax rates, with the reductions roughly proportional to the average value of the exclusion by income bracket and possibly other demographic characteristics. Given that the self-interested taxpayer-voter who is at the core of Buchanan ’s theory of public choice would have preferred the exclusion to its removal without side payments, would this proposed exchange of one way of taxing for another that raises about the same amount from almost everyone but allows a dividend of higher spending on other types of consumption (rather than low value spending on health insurance) be accepted by a majority?

Practical political economy suggests some potential impediments. Since tax rate reductions cannot be perfectly tailored to prior tax benefits, some people who had very large benefits from the subsidy may object—either because their former tax subsidy is much in excess of their prospective tax reduction, or because it seemed unfair to them that they should get less net benefit than others. Moreover, one of the other efficiency dividends from the exchange—less deterrence of work effort because of lower marginal tax rates—may not be realized if previously workers had counted on high excluded spending when their incomes rose. Finally, there is an ideological constituency in the medical care sector that views with alarm any attempt to reduce insurance coverage of medical care, for fear that some may go without “needed” care (Gladwell 2005). Nevertheless, my conclusion is that public choice theory argues for eventual removal of this long-standing unfair and inefficient stimulus to medical spending as a likely outcome. One possibility might be to allow workers (individually or as the workforce of a firm) to choose voluntarily whether they want to trade in the exclusion for lower tax rates. There would be some bias in this choice (workers who got a lower than average tax exclusion would be more likely to volunteer), but it might break the political logjam.

The final feature of Obamacare that might usefully be viewed through the lens of Buchanan ’s insights was the signature feature of reform: making subsidized insurance available on an individual basis to lower and middle-income people on Exchanges. This attempt to nationalize the insurance broker business for people who do not get job-based insurance was not very successful, and led to a rocky start in signing up customers, which has remained somewhat unstable. And the number of people who actually bought insurance through these exchanges remains well below initial projections even from the Congressional Budget Office. Although that has meant a lower level of government spending on subsidies, it is fair to say that, after the law squeaked through Congress, it has not achieved strong support from a majority of voters or a majority of taxpayers ever since and is still in political jeopardy.

What does Buchanan ’s political economy theory tell us about the reason for lukewarm support and likely future developments of subsidized insurance for the non-poor? The policy goal of many economist commentators on the uninsured and medical care use and spending envisions “trading in” the pro-rich and cost-increasing tax exclusion for subsidies to those mostly lower income individuals who are likely to be uninsured and suspected of failing to use beneficial care. I have already commented on the political impediments to getting rid of the tax loophole, but what about the subsidy program present in Obamacare ? That program rejected a uniform universal public program for health insurance for under-65 people at all income and wealth levels (in the form of original Medicare and Social Security) in favor of subsidies based on “need,” and hence much more generous for relatively lower income people given risk and at higher risk given income. The qualitative outlines of this subsidy program I have argued can be based on Buchanan ’s notion of “consumption externalities per se,” and fit the pattern I suggested in my doctoral work done under his direction (Pauly 1971). That approach began with the idea that persons other than the direct consumer may get utility benefits from assuring that beneficial care that might be worth less than its cost to the direct consumer but worth something positive at the margin to other concerned persons should be encouraged. Per unit subsidies to medical care consumption are de facto insurance benefits (given the stochastic nature of illness), and so provide both health- and wealth-protection features. But totally covering the cost of care may lead to use of care worth less than its cost to everyone because of moral hazard (Pauly 1970). So the ideal would be a pattern of subsidies to insurance intended to move consumption from the individually optimal level to the socially optimal level—and because both insurance and health appear to be normal goods, would require generous subsidies to generous coverage for the poor, and then phasing down both subsidies and required coverage generosity as income rose.

That was the pattern incorporated into law in Obamacare for the modest share of the population not already covered by employment-based health insurance. However, the theoretical mutual welfare gains to recipients of subsidies (from more care and less risk ) and to taxpayers who are concerned about the medical and financial health of that population have thus far failed to assure majority support. The simple public choice model involving voter comparison of marginal utility benefits from a public program with marginal tax cost can, I believe, explain the current opposition and suggest a way to think about how to determine whether this program should be stabilized and how to do so.

There were several serious flaws in the support for and design of the program as it was (barely) passed into law. The first is that, to my knowledge, neither the pattern of subsidies to insurance (proportion of premium covered) or subsidies to care (proportion of unit prices of care covered) were based on defensible empirical evidence. Three key and knowable but currently unknown pieces of empirical information never came up in the discussion of the design of the program: how subsidies at different levels for people at different incomes would affect their take up of insurance, how insurance coverage of different degrees of generosity would influence the use of additional care, and how much of an improvement in population health might be expected based on enhanced coverage. Moreover, the more fundamental question of how much improvements in health were worth to those covered and to those subsidizing the coverage was not asked. It is true that measures of demand responsiveness for insurance and care did figure prominently in estimating the cost implications of coverage chosen, and some recent research has tried to estimate the value of coverage to those subsidized (and found, unsurprisingly, that is worth less to them than the full premium cost and often less than the subsidized cost (Finkelstein et al. 2017).

But the key design parameters of the plan—subsidies up to 400% of poverty line incomes, minimum coverage at least 60% of average expected health spending, seem to have been selected on an ad hoc basis involving consideration of what was done in Massachusetts’ prior plan, concern for the fairness of distribution of subsidies across income levels, and the need to hit some spending impact targets. In particular, there was no attempt either to demonstrate what health improvements might accrue along the income distribution or that the premiums and cost sharing of the Affordable Care Act actually were affordable (meaning able to motivate purchase, with or without complaints).

The other sketchy aspect of the law was its financing. There was not an earmarked addition to taxes that each person could estimate—except for a modest share of financing to be generated by higher Medicare taxes imposed on higher income taxpayers of all ages. As already noted, the largest share of financing was to come from lower growth of payments for hospital and related care for Medicare beneficiaries, the Cadillac tax (eventually) and a grab bag of excise taxes on insurance, medical devices, and (believe it or not) tanning salons. There were also mandates on employers and the uninsured that would return revenues from the tax penalties imposed on those who did not comply.

For the recipients of subsidies, deciding in favor of the program was generally an easy call. But what about a kindly and concerned insured middle-class taxpayer, asked to support a program with no rigorous evidence of the health improvement it might generate with financing whose eventual impact (if any) on that taxpayer was impossible to guess? Buchanan ’s public choice model would not predict that such persons (not to mention those higher up on the income scale) would support the program on benefit cost grounds. If “uncertain personal cost” could be translated into “no personal cost,” small whiffs of benefit from the program (covering 25-year-old slacker “children,” guaranteeing coverage to high risks under 65) would generate positive sentiments, but uncertainty about effects on taxes and private premiums would lead to second thoughts. And polls suggest that a large slice of the population is in this uncertain swing group, opposed to almost everything they hear whether it is continuation or canceling of the program.

What needs to be established to help to make progress? The key thing is whether there is some way of distributing the actual cost of the program over taxpayers in such a way that their perceived benefits from the program (compared to some alternative) are higher than their incremental tax cost. That leads to a big but important “to-do” list for researchers and politicians.

One is to produce conclusive evidence to establish that insurance coverage does cause improved health. A randomized controlled trial of Medicaid expansion to poor able-bodied adults in Oregon found no major effects on health but some benefit in terms of reducing high medical bills; a less-robust study found positive effects on health in Massachusetts. This was coverage expansion for the least well off, and presumably, effects of coverage on health would be smaller for those who have higher incomes. Suffice it to say that there is no unequivocal evidence of consistently large magnitude effects of insurance on health outcomes that might persuade a kindly but skeptical taxpayer to be eager to pay higher taxes for such a program.

The other unknowns involve taxes. One is how the final tax payments to cover the uninsured under Obamacare actually are distributed. The other is how values of improvements in health for subsidy recipients (assuming such increments can be demonstrated) are distributed across the population based on characteristics that might plausibly be used in a tax system. (Buchanan attributes this idea to Knut Wicksell .) There is for example evidence in cross state analysis of Medicaid programs that higher taxpayer income promotes more generosity to a given poor population. If the “income elasticity of demand for subsidies for health insurance” by non-recipients of the subsidy could be known, taxes could be made to vary with income in the same fashion, potentially increasing support for the program if it is of enough value in the first instance. Some other characteristics will be harder to incorporate: we know that Southern taxpayers are less supportive of public programs but more generous with private charity, but there is no obvious way to build such regional differences in preferences into the tax system.

About a third of the financing for the subsidies newly extended to purchasers on exchanges and states expanding Medicaid eligibility were to be financed by “cuts in Medicare .” That did not literally mean year over year reductions in spending, but rather slowing the rate of growth in spending below what had previously been forecast. What insights might Buchanan ’s work offer for this method of financing?

For Medicare beneficiaries, this change literally meant that some of the money that would have been spent on care for them would now be spent on other groups in the population. While they were reassured that hospitals and to some extent physicians would be able to continue to provide the same access to care and quality of care by “productivity improvements,” reducing prior technical inefficiency in the face of reimbursement constraints, beneficiaries then and now are concerned about impacts of these cuts, with provider organizations predicting potential adverse consequences in terms of access to care and hospital closure.

Beneficiaries (or their self-appointed lobbyists, such as AARP) could have argued that they would have preferred any improvements in technical efficiency to go toward higher quality care for themselves or lower beneficiary cost sharing, but they did not. They might also have argued, and a few did, that any slowdown in Medicare spending growth should go toward paying forward some of the future costs of Medicare which were forecasted by the government actuary to run ahead of Medicare payroll taxes and the growth of GDP; they could have argued that the savings should have been used for the benefit of future Medicare beneficiaries who would otherwise get less or have to pay more.

Transparency of this process was impeded by the fiscal accounting by the Medicare trust fund that pays for hospital (Part A) benefits for all seniors regardless of income; the slowdown in spending growth (relative to what had been forecasted) actually improved the “health” of the trust fund in that it delayed the date at which it would achieve zero balance, even though it meant that some of the proceeds from the payroll tax earmarked for hospital care were being diverted to paying subsidies to others (in place of the higher taxes or greater budget deficits that would otherwise have been required). The best explanation of this apparent paradox is to note two points: (1) slowing growth of benefits payments always delays exhaustion of the trust fund but (2) taking all of the excess tax collections over benefits cost to buy bonds for the trust fund postpones this date further than diverting some of those funds to other purposes. So beneficiaries were not harmed compared to no reduction in spending growth , but diversion of funds to others harmed them relative to how well off they would have been with lower spending growth and no diversion. All of this assumes, of course, that providers were able to absorb lower payments over the long term without changing access or quality.

An aspect of this Medicare change in the ACA, and Medicare in general, may deserve comment. There is a view that passage of the original Medicare legislation in 1965 along with its trust fund financing of the then-dominant hospital spending represented a promise to current and future elderly that the benefits then described—no limits on access, no interference in provider-patient relationships—were to be perpetual. Economists have taken great glee in pointing out that, despite appearances of piling up assets to buttress such promises, there actually is no trust fund and the great bulk of Medicare (and Social Security) taxes go out to pay for current federal spending of all kinds. People who believe in the “promise” model have been duped.

However, Buchanan ’s notion of constitutional rules limiting current short-term government behavior may have an application here. The application is different from those he typically made, where tax rules were put in place in part to simplify political bargaining and in part to constrain politicians and bureaucrats who want to spend more. The stability over time of the rate structure in the personal income tax was the usual example of a “quasi-constitutional” rule. Admitting that determining empirically what is “constitutional” or not is a much more complex and deeper problem than can be covered here, one might imagine that Medicare (and Social Security) embody a longstanding constitutional rule that promises not to spend less on the elderly—and certainly not to spend less on this program in order to divert funds for programs that benefit others. In effect, the ACA violated a long-standing policy on which people may have relied in planning retirement and living in retirement; it violated what many perceived as rules of the game—and it did so based on a single Congressional bill passed with a razor thin margin. We do not want to go too far here: we have no definitive basis for deciding which types of government actions have constitutional protection, and what sanctions are implied if such rules are broken. It is probably fair to say that nothing obviously terrible has yet happened to Medicare beneficiaries (who have yet to experience the access problems associated with low physician payments in Medicaid) but many of the consequences will only take place more than a decade after the legislation. Still, both the future trajectories of Social Security and Medicare spending relative to federal tax collections (both earmarked and total) are so dire that circumstances at some point may force a constitutional convention in which decades-long promises to seniors are reconsidered—probably with delayed application behind a veil of ignorance. The unpredictability and plasticity of Medicare spending will press the issue.

Conclusion

At this point using a public choice model largely raises unanswered questions about the problems it identifies, and does not provide an immediate recipe for what “we” should do. But that phenomenon was very much part of Buchanan ’s intellectual style, seeking first to identify and then begin to address the fundamental questions of what individual preferences are for collective action and how they might be translated (or mistranslated) into government efforts. “Looking through a new window” and “analyzing before action” were two bedrock characteristics of his approach. It is as much needed in this field as ever.