The Likelihood of "Repealing Health Care"

The subject of “repealing” the health care reform package is of considerable importance for political scientists, pundits and lawmakers, both as a political issue and for the next congressional session.  It just so happens that I have a forthcoming paper (see here for links and a summary) that addresses the latter issue.  For this project, I used David Mayhew’s list of landmark legislation enacted from 1951-2006, tracked each law after enactment (using Westlaw) and recorded formal repeals to provisions the original law.  My initial work with this dataset was largely ecumenical.  The central question was: What are the determinants of policy repeal?  I was primarily interested in whether divided party control results in the passage of more durable policies (I found that it is, but only when we observe legislation for a sufficient duration).  In my opinion, this is an interesting finding because a number of studies find that divided government results in the passage of fewer policies (e.g. Sarah Binder’s excellent work).  Thus, with this paper, I argue that divided party control has a complementary effect on the overall supply of legislation. Consistent with this blog’s original aims, I figured it fitting to take a more rigorous look at the prospects of “repealing health care.”  The phrase “repealing health care” is in quotes for two reasons: (1) the bill enacted into law is unlikely to be repealed in its entirety, and thus talk of repealing “health care” is a misnomer; (2) the term “repeal” has been used as a blanket phrase for the range of options available for reversing or modifying the slate of health care policies.  Some of this is discussed in the paper.  For example, alternative routes to reversing or modifying public policy include altering or eliminating funding, invalidation by the courts, amending activity and allowing sunset provisions to expire.  A future post will explore these alternatives in greater detail.

Here are some highlights from the paper that are relevant to the current political rhetoric.  The chart to left is the hazard of repeal for all landmark legislation over time (measured as the number of subsequent 2-year congresses).  Peaks indicate a higher likelihood of repeal in time t while lower values indicate a lower likelihood of repeal.  As you can see from the hazard function, repeals increase in likelihood up to about five subsequent congresses (ten years) and then decreases monotonically for the reminder of a policy’s lifecycle.  Thus, the health care package recently enacted is most “at risk” over a decade from now (less at risk in the next legislative session).  Second, based on the event history results, there are some factors that will constrain the ability of Republicans to formally repeal provisions of the original law.  Government will be divided next term and Democrats will likely maintain control of the Senate.  These factors—divided party control and the subsequent majority status (in at least one chamber) of the enacting party—are estimated to decrease the hazard of repeal by 25% and 20% respectively.  Also, the generally high levels of polarization in recent Congresses—among individual members and across chambers—will further constrain efforts to repeal provisions of the new law.  Basically, if (when?) Republicans regain unified control of Congress, it will be difficult for them to garner enough Democratic support (namely to overcome a filibuster in the Senate) for repeal.  Finally, though some polls find a slight majority of registered voters who disapprove of the health care law (I recall a Quinnipiac poll a few months ago), it still remains a low priority according to Gallup’s Most Important Problem question (the data used in my analysis).  Though, admittedly, the Gallup data is an imperfect measure of public opinion on this issue, the point is that health care is not as big a concern for Americans as other issues such as the economy and terrorism.  Finally, I also find in the paper that health care policy is generally resistant to repeal compared to other public policies (though this effect is not statistically significant).  This is may be somewhat surprising given the short history of the Medicare Catastrophic Coverage Act.  Still, I think this makes sense as entitlement policies are notoriously difficult to reverse.  In summary, there are some real obstacles to repealing health care reform.  I think Republicans’ confidence in their ability to “repeal” the newly enacted law is either (1) designed to stir up the base for the midterm or (2) an imprecise term for the range of options available to reverse or modify the original enactments.

Can we quantify the overall likelihood of repealing health care?  Yes; but let me offer an important disclaimer.  In the context of my study, repeal is measured as the formal repeal of at least one provision of the original law.  This approach is designed to model the decision of a subsequent governing coalition to repeal policies enacted by their predecessors rather than the magnitude of repeal.  Modeling the magnitude of repeal in a “number-of-provisions” fashion is conceptually problematic (see the paper).  This point is important because with a massive bill like the recently passed health care law, there is an increased risk of some repeal.  In fact, there is already discussion of Democrats favoring the repeal of one original provision (section 9006, which requires businesses to file tax forms for vendors from whom they purchases more than $600 in goods).

Here are some important factors I’m using to estimate the likelihood of repeal:

  • 7% of Americans over the next two years will say that health care is the “most important problem” facing the nation.  This is the current estimate as of August, according to Gallup.
  • The “public mood” will trend in the conservative direction at 2% every 2 years.
  • I’m estimating that the law has 1300 separate legal provisions.  When collecting the data on repeal I recorded the total number of provisions for each of Mayhew's laws.  Here I am guesstimating based on the number of pages.
  • Bicameral and ideological polarization remains at current levels.  Recall that greater polarization on both measures decreases the likelihood of repeal
  • Finally, and most importantly, I estimate three models based on the potential results of the 2010 midterm and 2012 presidential election.  In the model, these scenarios represent alternative specifications of the Divided Government Sub and Same Party variables.  In the first scenario, the Republicans regain only the House in the upcoming election (leaving divided government with the enacting party in control of at least one chamber).  In the second scenario, Republicans regain both chambers in the upcoming midterm (still divided government).  Finally, in the worse-case-scenario for proponents of the new law, Republicans regain unified control of government in 2012.

Here are the estimates:

In the first scenario, the estimated likelihood of at some repeal during the 112th Congress is 52%.  In the second scenario, the estimated likelihood of at some repeal is 59%.  And in the third scenario, the estimated likelihood of some repeal in the 113th Congress is 69%.

Again, let me point out that these numbers are estimates of the likelihood of at least one repealed provision.  The size of these estimates is driven largely by the law’s massive size.  Rather than focusing on the estimated likelihoods, it’s more important to take into consideration the effect of the various factors and their combinations on the likelihood of repeal.  In this way, we can have a meaningful, empirically driven discussion of when repeal is most likely.

So what are the conclusions?  First, discussion of “repealing” health care is a clever bit of issue framing (in my opinion).  There are better options, from the Republican perspective, to altering the original law.  Second, the newly enacted law will be most “at risk” not in the next Congress, but a decade from now.  So sit tight.  Third, there are some serious constraints the Republican Party will have to overcome to make major changes to the original law.  True, some repealing activity is likely to succeed.  But this is more a function of the law’s size than anything else.  Still, if Democrats can maintain the Senate in 2010, and hold the White House in 2012, their prospects of defending the law from “repeal” is significantly greater.