Summary and Conclusions

Despite the ACA’s razor-thin margin of passage and ideologically narrow base of support, more Americans want to see the Act implemented than to see it dismantled, or even dramatically modified. Poll data indicate that this remains true despite difficulties with HealthCare.gov, and the widespread cancellation of pre-ACA policies in the individual market

Even though website difficulties, and especially individual market cancellations, may further reduce support for the Act in the coming weeks[1], these are both relatively short-lived sources of controversy

Far longer-lasting, and far more fundamental to the effective implementation of the Act, is the matter of whether the uninsured will enroll in individual market coverage in sufficient numbers to both meaningfully reduce the total number of uninsured, and to create functioning risk pools

It’s much too early to make any firm judgments, but the evidence isn’t good. In the 15 states operating their own exchanges, and thus avoiding some but not all of the problems that underlie the problems with HealthCare.gov, the number of persons selecting plans (as distinct from paid enrollees) in the first month is just 0.78pct of persons who are currently covered by individual policies in those states – and presumably this number has to reach 100pct for these states to accomplish nothing more than maintaining their current number of insured. Plan selection as a percentage of the uninsured in these states is just 0.73pct. In the remaining states that rely on HealthCare.gov, these percentages are even smaller – 0.13pct (of current individual market beneficiaries) and 0.09pct (of uninsured)

We believe the website difficulties will pass, and that as they do, enrollment will continue to be weak. Subjectively, the uninsured appear to have relatively modest interest in purchasing coverage; and objectively, becoming newly insured under the Affordable Care Act – despite the subsidies –requires the uninsured to increase their (average) health spending. We believe the uninsureds’ view of the ACA comes down to this: it’s a good deal if you’re sick (or expect to be), but not if you’re well. This implies weak enrollment into 2014, continued high costs of uncompensated care for hospitals (the de facto catastrophic coverage provider for the uninsured), and mounting adverse selection pressures in subsequent enrollment periods

The Affordable Care Act isn’t likely to be repealed, or even dramatically modified – voters want the Act implemented

Unlike the nation’s longstanding cornerstone social programs, the Affordable Care Act (ACA) was born into partisan controversy, and only born at all by the thinnest of margins – 3 more votes than necessary in the House, and exactly the number of votes necessary in the Senate (Exhibit 1). Opinions of the ACA still fall sharply along party lines (Exhibit 2)

However despite the highly publicized shortcomings of HealthCare.gov (which opened October 1, 2013), Americans’ views of the Act after the failed website launch are much as they were in the months or even years preceding (Exhibit 2, again; also Exhibits 3 and 4)

Polls attempting to determine whether a majority supports or opposes the Act tend to show a modest majority either in favor (Exhibit 3) or in opposition (Exhibit 4), all of which suggests that roughly similar numbers of persons support or oppose the Act. However when the question is framed in terms of whether Congress should keep / expand v. repeal / replace the Act (Exhibit 5, latest observations taken the week of October 17 to 23), or whether Congress should use its budgetary authority to defund the Act (Exhibit 6, latest observations also taken the week of October 17 to 23), it is apparent that, despite recent problems, a majority of the electorate want the Affordable Care Act implemented. Despite the Act’s narrow margin of passage, it’s ideologically narrow support base, and its host of recent problems, there is no voter support for dismantling the ACA

Yes, this is important for the Administration – but let’s all keep our bearings

Clearly the Affordable Care Act is important to the President’s legacy; this plus the fact that voters want to see the Act implemented argue that the Administration will do everything within its power to see the Act succeed – such as extending cancellation dates on ‘pre-ACA’ policies, as occurred last Thursday. It’s nevertheless worthwhile to put the importance of the ACA to the President in some context, as we sense a tendency to view ACA-related policy matters as having more relative (to other policy matters) gravity than data would support. The President’s approval rating and voters’ approval of his handling of healthcare are correlated (Exhibit 7a); however the President’s approval is also similarly correlated with the other issues Americans define as most important (Exhibits 7b thru 7e). As further context, ‘Obamacare’ and ‘Affordable Care Act’ have – since the President’s first inauguration – been sharply less common search terms than any of ‘Immigration’, ‘Syria’, ‘Afghanistan’, or ‘Shutdown’. And, at its recent peak, ‘Obamacare’ / ‘Affordable Care Act’ reached about 60 pct of the ‘Syria’ peak value, and only 35 pct of the ‘Shutdown’ peak value (Exhibit 8). The Affordable Care Act is a cymbal-crash that punctuates the Obama presidency, but – as with most administrations – the economy, budget deficit, and foreign affairs provide the steady political drumbeat

Website dysfunction and policy cancellations will fade in relevance; leaving the enduring question of whether the uninsured actually want to buy coverage…

Thus if the ACA is not going to die politically – it’s not – the question still remains whether it can be brought to life administratively. We believe that website shortcomings and voter anger over policy cancellations both will be short-lived, and that these issues ultimately will not meaningfully affect the trajectory of the ACA. The fundamental IT challenges of the website (multiple highly sensitive data inputs from multiple domains with multiple source formats – all in real time) should not be underestimated, but we live in an information age where these challenges are routinely overcome. And, as consumers of technology and participants in e-commerce we’ve all braved crashing websites (eBay), privacy offenses (Facebook), dropped calls (AT&T iPhones), and orders that never arrived (e-commerce before Amazon) to place ourselves into e-interactions that ultimately are far more convenient than doing things the old-fashioned way. For anyone that wants individually purchased coverage, waiting for a balky website to get on its feet is still far better than calling insurers for bids by phone and enrollment materials by mail. All of which means we reject the idea that if HealthCare.gov isn’t fixed soon, then people who want individually purchased coverage won’t come back

This having been said, it’s nevertheless interesting to see just how much HealthCare.gov’s problems have done to stall first-month enrollment. The 15 states with their own exchanges are at least partially insulated from the problems of HealthCare.gov, so by comparing these states’ enrollment rates to those of the remaining states that rely on the federal exchange website (HealthCare.gov) we can get a sense of just how much of an impact HealthCare.gov’s problems are having on enrollment

Exhibit 9 shows the absolute number of persons selecting an exchange-based plan by state (grey dotted line); and, more importantly, the number of persons selecting an exchange-based plan as a percent of that state’s uninsured (solid green line). States on the left are operating their own exchanges; states on the right participate in HealthCare.gov. In states with their own exchanges, the number of persons selecting an exchange-based plan was on average 0.73% of the number of uninsured persons (excluding those Medicaid eligible); in states using HealthCare.gov, this number was just 0.09%. These obviously are tiny numbers, and it’s only the first month of enrollment – but still. Exhibit 10 takes a slightly different look, comparing the number of persons selecting an exchange-based plan to the number of persons currently insured in the individual market, by state. For states with state-based exchanges, the percentage is 0.78%; in states relying on HealthCare.gov the percentage is 0.13%

These data support two obvious generalizations – enrollment in all states is very slow; and the HealthCare.gov states are enrolling at about one-sixth the rate of the states that operate their own exchanges

Far more importantly: Do (enough of) the uninsured want coverage?

Poll data indicate that the uninsured are generally aware of the requirement to obtain insurance (Exhibit 11), but that many still do not intend to obtain coverage (Exhibit 12); and that despite the overwhelming news coverage only (roughly) one in four uninsured persons are familiar with the health insurance exchanges (Exhibit 13). The uninsured are more aware of the requirement to have insurance than of the means to obtain it, and subjective measures (i.e. poll responses) of demand for coverage show relatively modest interest

Subjectivity aside, generally speaking (i.e. assuming limited adverse selection) the average uninsured person who purchases health insurance on an exchange will spend more on health care than if they had remained uninsured – even after accounting for the effect of subsidies.   A recent RAND research report compared total (net premiums + out-of-pocket costs) health spending for persons who were either uninsured, or who purchased health insurance on an exchange. The study looked at change in total health spending for persons who were currently insured (under policies purchased in the ‘pre-ACA’ individual market), and for persons who are currently uninsured

For the average person who is currently uninsured (Exhibit 14, top 3 rows), total health spending (before tax and penalty effects) increases if they become newly insured on the exchanges. In net (of tax and penalty effect) terms, total health spending rises for persons with incomes below 400pct of the federal poverty threshold (aka 400FPL). What’s happening here is that the uninsured currently spend very little (on average) for healthcare, and – despite subsidies – their share of premiums and out of pocket costs under exchange-based policies will be (again, on average) more than their current spending. The marginal cost of being insured on the exchanges is significant for persons in these income ranges – in the neighborhood of 3 to 4 pct of income. Or think of it this way – persons below 138FPL already have very negative savings, with expenditures that exceed incomes by $9,015 – and buying exchange based coverage will require finding an additional $559. For persons with incomes in the 138FPL to 400FPL range, the $1,567 marginal cost of purchasing exchange-based insurance adds to their current household spending deficit of $2,928

Conversely, if you’re already covered – i.e., if you were going to purchase individual coverage anyway even if the exchanges didn’t exist – then the exchanges will save you money (Exhibit 14, bottom 3 rows). For persons who would have purchased coverage anyway, those with incomes below 138FPL will save $2,565 (fully 20pct of income) because of the exchanges; for those with incomes between 138FPL and 400FPL the savings are $894, or 4.4pct of incomes. All that’s happening here is that subsidies are lowering the cost of coverage for those that would have purchased coverage anyway

The bottom-line: exchanges make health insurance cheaper than it otherwise would have been for persons in the subsidy-eligible (138-400FPL) income range; however being uninsured is still (on average) less expensive than being insured for persons in this income range. The healthier and/or younger you are, the more this is true