Jonathan Bachman / Reuters

With the national rate of uninsured people at historic lows, why isn’t Obamacare more popular? Over the course of the last two years, which saw around 20 million Americans gain health insurance coverage, public opinion of the Affordable Care Act nevertheless remains negative. Its unfavorability rating in a Kaiser Family Foundation poll hovered at 47 percent in both September 2014 and 2016, and over the same time span favorability has increased from 35 percent to just 44 percent, though it was higher in 2015 than today.

That public ambivalence has made Obamacare a thorny issue on the campaign trail. Congressional races often find Democratic candidates positioning themselves as reformers and fixers of the health law, a stance necessitated by its enduring unpopularity. Last week, Bill Clinton called the law a “crazy system” in a campaign speech for Hillary Clinton—a characterization that has been overblown, but still contains a legitimate criticism embedded in the former president’s praise of the necessity of health reform and the importance of expanding coverage. If Donald Trump’s candidacy weren’t so unusual—and if he wasn’t currently fighting off a monumental and unparalleled campaign of self-sabotage—anxiety about Obamacare probably would have been one of the most important issues in the months leading to the election.

So what gives? At Vox, Sarah Kliff points out that most Americans simply don’t know that the uninsured rate is the lowest that it’s ever been. In fact, according to a Kaiser Family Foundation poll, about 20 percent of all Americans think there are actually more uninsured people on average than last year. While a majority of Democrats and supporters of the law don’t know that uninsured rates are at historic lows, nearly all independents, Republicans, and people who oppose the law are unaware. Kliff’s conclusion of a “negativity bias” among media toward Obamacare's failures—and toward disproportionately covering populations that have been harmed by the law—is supported by those numbers. Judging anecdotally by the range of health-policy stories that I’ve covered for The Atlantic, there’s a rather loud ring of truth here.

It is almost certainly the case that some of the dissatisfaction with the health-care system’s costs has been driven by some negativity bias in media and politics. For example, a 2015 Harvard University poll found that support for expansive Medicare drug-pricing reforms increased independently of actual exposure to price increases among those polled, suggesting respondents feel the anxiety over drug prices that’s highlighted in media coverage and political campaigns. As Medicare prescription-drug reforms have become a key point of health-policy contention between Republicans and Democrats, it’s clear that dissatisfaction, while probably rooted in real personal, financial anxieties and health issues, has been affected by the political climate and coverage of that climate.

In a column for The Wall Street Journal, Kaiser Family Foundation president and CEO Drew Altman largely agreed with Kliff’s diagnosis of negativity bias and political spin, although he added one possible alternative hypothesis: that maybe the labels of “uninsured” and “insured” used in health-policy circles and journalism don’t actually mean much to the daily lives of Americans. In the column, Altman states:

It could also be that categories experts refer to—uninsured, insured, under-insured—aren’t as meaningful for non-experts. People who are uninsured and people who have insurance both struggle to pay medical bills. Categorizing insurance status for statistical purposes may be more meaningful to experts who can recite study after study about how much insurance matters than it is to consumers struggling with medical bills even if they have coverage.

That’s an important point. Much of the existing historical research on health-insurance “churn,” or the number of times people change insurance coverage types over a given year, indicates that people’s level of coverage falls along a continuum of statuses—more volatile for some than others—rather than in a solid state of existence. Even over the course of a single year, people’s insurance status often changes as they gain or lose jobs; slide in and out of eligibility for public insurance; experience major life events; face employer changes in coverage offered; or make choices to change insurance based on price or benefits. Many of these transitions are accompanied by gaps in insurance coverage or major changes in costs, doctors, and health-care regimens. Those things—the actual cost and experience of seeking and receiving necessary health-care services—are the most salient effects of any health policy on the daily lives of people.

A new study from Ben Sommers and researchers at the Harvard T. H. Chan School of Public Health suggests that at least some of the daily lived experience of people has not changed, even as fewer and fewer of them have become uninsured. That study finds that churn in three states in 2015 after the major coverage expansions of Obamacare was not significantly different from churn in 2013 before those expansions came into effect. For both years, roughly about one-fifth of the population of low-income adults in Texas, Arkansas, and Kentucky changed their insurance. Among those, Texas did not opt to expand Medicaid to all low-income adults, which suggests that reforms have so far done little to change the latent instability of coverage for the most vulnerable people.

From a policy perspective, these results are actually promising, because such a sweeping law with so many policy changes, new requirements, and insurers dropping in and out of markets should be expected to increase churn by a great deal. That it hasn’t means President Obama’s unrealistic promise about keeping health-insurance plans has actually been less damaging than it could be. But for millions of Americans—even those who switch between coverage types and don’t enter the ranks of the long-term uninsured—churn is inconvenient, if not dangerous. The Harvard study finds that 56 percent of all low-income people who changed coverage or status in 2015 faced some temporary or long-term gap in coverage where they were essentially uninsured. Almost 20 percent of all these churners had to change physicians, and 40 percent reported that the quality of their health care declined.

The results from this study may not be generalizable across the country and only speak to the experience of low-income people who are generally at the highest risk for instability, but those are also the people most likely to be uninsured and affected by coverage expansions. For most other people, the coverage gains of Obamacare are even less apparent as immediate benefits. For employer-based private coverage, under which most Americans receive their insurance, costs have increased at about the same rate over inflation and wages as they always have. For people who enroll for coverage in the private insurance exchanges, premiums jumped by between 13 percent and 21 percent over the last year. While most people who receive coverage this way are protected from insurance cost increases by government subsidies, those who have to pay full price have to shoulder the entire burden of that increase. Among people with exchange coverage, dissatisfaction with costs has risen for each of the years since the exchanges were created.

The expansion of health-insurance coverage has had real, positive effects on the ability of families to pay for medical bills, but those effects might not move the enthusiasm needle much. Results for the end of 2015 and first quarter of 2016 are forthcoming, but the National Health Interview Survey indicates that the proportion of non-elderly people in households that had trouble paying for medical bills dropped from 18 percent in 2014 to 16.5 percent in early 2015. That’s a significant decrease in the number of people hampered by medical bills, but the decrease is part of a gradual, neatly sloped decrease predating Obamacare’s passage, which might indicate that the rising tide of economic recovery simply makes families more economically resilient. The Affordable Care Act played a role in that recovery and in macroeconomic effects leading to those families’ improving fortunes, but that point is likely obscure where public opinion is concerned.

Many of the economic protections and personal benefits of health insurance, like any insurance, are long-term in nature, and will probably elude detection in public opinion over the short term. Stability for families should increase, if incrementally. Unlikely catastrophic health events should become less financially catastrophic over time, the average ability of families to handle crises should improve, and the cumulative effects of better access to care should lead to much healthier people. But those effects occur over the horizon of years, and the lives of those people that coverage can affect the most are still turbulent now.

Overall, the fact that 46 percent—a plurality—of Americans think the uninsured rate is about the same as it was before is telling. There have certainly been thousands of lives that have been saved by sudden access to insurance, and stories like the tearful enthusiasm of newly eligible poor black citizens in Louisiana signing up en masse for Medicaid are only snapshots of the wide-scale positive effects of increasing coverage. But for many people, even those now in line to receive better care than ever before, the inherent instability and cost issues of insurance largely remain as major obstacles to navigate in their daily lives. Obamacare’s coverage expansions should whittle away at all of these issues, but the point of success probably won’t be whenever public opinion of Obamacare or acknowledgement of the uninsured rate spikes—if it ever does—but when people don’t even think about it at all. That might take years.

We want to hear what you think about this article. Submit a letter to the editor or write to letters@theatlantic.com.