Did Obamacare Cost Hillary the Election?

Back in 2011 I noted over at National Review the work of three Stanford political scientists who dove deep into the data of the 2010 election to discern the causes of the wave that saw Republicans win 63 House seats from Democrats, when none of the pre-election prediction models forecast that Republicans would come anywhere close to that large a gain. Their article, “The 2010 Elections: Why Did Political Science Forecasts Go Awry?“, concluded that 40 House Democrats lost their seats because of their Yes votes in favor of Obamacare and the Waxman-Markey cap and trade bill to fight global warming climate change.

One of the amusing parts of this study was how the political scientists who predicted only modest losses for Democrats before the 2010 election seemed to have let their own ideological biases affect their findings to some extent. The Stanford authors (who actually lean to the right) noted, “[I]t is clear from the commentary accompanying the forecasts that most of the authors were more concerned about overpredicting the Democratic losses than underpredicting them.” Heh.

But that kind of bias needs to be kept in mind when reading a new paper just out at the Social Science Research Network from two Ohio State political scientists that concludes that the unpopularity of Obamacare may have been the crucial factor in Trump’s victory in the key swing states. And you can tell by the abstract that the authors are not happy about this:

Did Obamacare Implementation Cost Clinton the 2016 Election?

Vladimir Kogan, Thomas Wood, Ohio State University


We combine administrative records from the federal health care exchange with aggregate- and individual-level data on vote choice in the 2016 election. We show that personal experiences with the Affordable Care Act informed voting behavior and that these effects could have altered the election outcome in pivotal states, suggesting that Republican efforts to undermine the law’s implementation paid tangible political dividends. We also offer evidence that consumers purchasing coverage through the exchange were sensitive to premium price hikes publicized shortly before the election — despite most receiving a federal tax credit that shielded them from the increases. We attribute this to the design of the HealthCare.gov website, which reduced the salience of federal subsidies and likely made consumers needlessly sensitive to media coverage focusing on rising premiums. Placebo tests using survey responses collected before the premium information became public suggest that these relationships are indeed causal.

The conclusion in the body of the study is more straightforward:

Our estimates provide strong evidence that the implementation of Obamacare may have indeed cost Hillary Clinton the presidency. Of the states in the table, four did not expand Medicaid. In two of those states (Florida and Wisconsin), our estimates predicts that the expansion would have delivered enough Clinton votes to change the winner of the popular vote there, and thus give Clinton an extra 39 electoral votes — one more than would be necessary to win. In addition, some combination of higher exchange enrollment and modestly lower premium growth would have been sufficient to flip the outcome in Michigan, giving Clinton an additional 16 electoral votes.

Some parts of the full study are really fun to ponder, such as this passage:

[T]he behavioral responses we document are far from rational. For example, more than 85 percent of individuals who purchased insurance through the federal exchange in a typical county received federal tax credits to offset some of the costs, which meant they were largely insulated from the widely reported premium increases, yet their voting behavior appears to have nevertheless responded to rising prices.

Maybe people who receive government subsidies are still capable of reaching a complex and “rational” judgment that they don’t like a coercive policy anyway? This possibility seems somehow strange or foreign to many political “scientists” these days. (Though the authors do hint at this possibility in their concluding discussion: “[T]hese groups of voters may have seen large premium increases as symptomatic of broader problems with the Affordable Care Act or as a signal of likely deterioration in their local individual insurance market. . .” Ya think? Nah, because the next sentence is: “While possible, this explanation strikes us unlikely because we find no evidence that these same subgroups similarly punished Democrats when all but one insurer had pulled out of their local exchange — arguably a much more worrying sign about the future of the local marketplace.”)

The other explicit message of this study is that the residual unpopularity of Obamacare with swing voters is all the result of those mean Republicans who just won’t get on the side of history. Here, for example, from the “discussion” of the findings at the end of the study:

[O]ur results reinforce the political logic of the Republican Party’s efforts to undermine the implementation of Obamacare, suggesting that these likely hurt Democrats at the polls to the extent that they reduced Medicaid and exchange enrollments and increased premiums. . .

One troubling implication is that electoral considerations can encourage policy sabotage if responsibility for implementation is shared across different offices or levels of government. This is particularly relevant in the context of American federalism, where state and federal governments share both legislative and administrative responsibility across many policy domains. Even when one party controls the federal government, it is likely that at least some states remain under the control of the opposition, a dynamic Brown (2010) describes as “divided federalism.” Because state elections are largely referenda on the incumbent president (see Rogers 2016), the opposition party in these states may be tempted to use its influence in perverse ways, with the hope that the president’s party ultimately pays the electoral price.

Of course, perhaps the “troubling implication” should be that parties should not attempt to implement sweeping social policy changes affecting 15 percent of the economy on a partisan party line vote, and ignore the history of major social policy changes like Social Security and Medicare and welfare reform that were passed on a bipartisan basis. In other words, heaven forbid that political science entertain the thought that it was Obama and the Democrats who behaved irresponsibly in the creation of Obamacare. The second lesson is that, given the role of Obamacare in the 2010 election referenced above, Obamacare is the gift that keeps on giving for Republicans. Maybe they were smart after all not to repeal it.

Finally, the last conclusion is especially delicious:

[O]ur finding that exchange users who were largely shielded from the rising costs of health insurance premiums nevertheless punished Democrats for these increases showcase how the implementation process can shape the political “feedback” effects that policies ultimately produce. . .  The results also highlights the important role played by media coverage in translating objective conditions into voters’ retrospective evaluations and, ultimately, votes.

Translation: Damn Fox News!

Of course, this entire study is completely wrong, because it ignores the real reason everyone knows Trump won: Racism. I hope these two authors have tenure already, since they’ve obviously ignored the memo on how to report the causal factors of 2016.