I know that chicken little can be a useful role to play, and I want to see the health care debate pushed left as much as the next blogger, but I also feel bound to abide by certain numbers. The collective blow-up over Baucus’s bill (see TPM, Pandagon, NMMNB, & LGM), while completely justified on the merits — the bill sucks — is completely off base regarding the electoral impact.
Assuming Democrats do pass a bill where “the poor, the unemployed, the working class are forced to pay large sums they don’t have to insurance companies for “junk insurance” with high deductibles” (aimai) or “the problem of the uninsured [is solved] by passing a law forcing them to buy health insurance which, by definition, most a) cannot afford or b) are gambling they won’t need because they’re young and healthy” (TPM), how potentially damaging is the offended constituency? This is a crude question and obviously the morality of a policy has nothing to do with political power. But I’m tired of the ambiguous boogie man of electoral backlash. Let’s quantify:
My chart of voting behavior for insured vs. uninsured people per US HHS Overview of the Uninsured in the United States, KaiserEDU Public Opinion: Health Care and the 2008 Election, and Gallup 2009 Detailed Political ideology.
Uninsured, politically moderate, likely voters (UPMLV). That’s my definition of the demographic who will be directly adversely affected if Democrats pair individual mandates with low levels of government subsidy. At most, that means 4% of voters. 4% isn’t nothing. Plenty of elections have been decided by less. Given current patterns, it’s 2/3 of the 6% swing Republicans potentially need to retake the House, and if Republicans did win 26 seats, that would be a big deal.
However, that number assumes a group so outraged as to produce a 100-0 split. It assumes no subsidies reach the UPMLV to dull the anger. It assumes that the uninsured are moderate at the same rate as the general population, when they likely skew liberal. I don’t feel comfortable trying to quantify these factors, since the combined margins of error become an order of magnitude greater than the size of the population we’re talking about, which of course is the larger point: we’re debating the electoral importance of Microtrends-sized group. Outside of politicians still being scammed by Mark Penn, I think we can agree that this is not going to be a winning argument for better health care reform.
What’s most noteworthy about this whole reform process is that universal/expanded access to insurance is the core of all the health care reform bills, even Baucus’s lame one. Democratic politicians have largely ignored the fact that there’s no real political margin among swing voters in reducing the uninsured. This actually says something pretty positive about the bulk of the Democratic political class.
A few notes about chart methodology:
- “% who voted in 2002” is from KaiserEDU and refers to the percent of uninsured and insured people, respectively, who self-reported having voted in the 2002 midterm congressional elections. I could not locate 2006 data, and in general it is a tragedy that every 2006 & 2008 exit poll didn’t record respondents health insurance status. For shame, pollsters, for shame.
- “% of 2002 voters” numbers are determined using the following equation: [ “% of total insured population” x “% insured who voted in 2002” ] / [(“% of total insured population” x “% insured who voted in 2002” ) + (“% of total uninsured population” x “% uninsured who voted in 2002” )]
- Insured vs. uninsured as a percent of total population comes from the US HHS 2007 population survey, and given recent job losses the uninsured share is undoubtedly higher, likely meaning that uninsured voters will make up a higher percentage of the 2010 electorate.
- Final set of bars overlays Gallup’s partisan identification data over the Insured/Uninsured bars for “% 2002 voters”. This is a bad assumption given the income, age, and race statistical disparity between insured and uninsured populations, but is methodologically conservative for estimating “uninsured likely swing voters.”