After the 2008 primary season, I analyzed Ron Paul’s performance in each state to see how institutional factors such as caucus and primary form affected his electoral success. This exercise turned out to be useful for estimating the size of the pro-liberty electorate in each state. In this post, I do the same with the 2012 results.
The dependent variable in this analysis is the percentage of the vote obtained in each state’s statewide primary or caucus. If a state held both a primary and a caucus or convention, I used the primary results. The independent variables are as follows. (All variables were taken from Dave Leip’s Atlas of U.S. Presidential Elections.)
First, I use a dummy variable for whether the state had a caucus or convention versus a primary. As an outsider candidate with a committed band of activist supporters, Paul tended to do much better in caucuses than primaries.
Next, I also included dummy variables for who is eligible to vote: a dummy for an open caucus/primary, in which all voters are allowed to vote, and a dummy for a closed caucus/primary, in which only registered Republicans are allowed to vote. The excluded category consists of “modified-open” elections, in which independents and Republicans are allowed to vote. Since Paul was more popular among self-identified independents than among self-identified Republicans, it stands to reason that he would do best in open primaries and worst in closed primaries.
I also tried variables for the number of other candidates running active races and appearing on the ballot. In this regard, I counted only “significant” candidates that might have had a chance of winning, that is, the following six (other than Paul): Mitt Romney, Rick Santorum, Newt Gingrich, Rick Perry, Jon Huntsman, and Michele Bachmann. Bachmann dropped out of the race after Iowa and Perry and Huntsman after New Hampshire. Thereafter, Santorum withdrew only after the April 10 primaries and Gingrich held on until May 2. However, in many states – almost entirely primary states – candidates that had withdrawn remained on the ballot and often received significant support. We should expect both of these variables to have a negative influence on Paul’s vote share (for instance, in Virginia, where only Romney and Paul were on the ballot, Paul benefited from tactical voting by supporters of other candidates and received 40%), but in fact the number of candidates in the race had a statistically significant, positive association with Paul’s support (candidates on the ballot has the expected negative association, as shown below). This result is theoretically implausible and apparently simply captures Paul’s otherwise unexplained success in Iowa and New Hampshire. Therefore, in the estimations reported here, I have included solely the variable for number of non-Paul candidates on the ballot.
Next, I included a three-point variable for the extent to which Mitt Romney had clinched the Republican nomination when the primary in question was held. After February 28, it became very unlikely that Romney would lose, as Nate Silver noted here. After April 3, it became a near mathematical certainty that Romney would win. Therefore, the “clinching” variable takes on values of 0.5 for contests held after February 28 through April 3 and values of 1.0 for contests held after April 3.
Finally, I include a variable intended to capture home-state effects for active candidates. It is coded zero for all states except Massachusetts, Georgia, and Texas. For the former two, home states of Romney and Gingrich, respectively, it is coded -1, and for the last, Paul’s home state, it is coded 1. (Santorum dropped out of the race before the Pennsylvania vote, for which reason it is coded zero.)
The dependent variable is rather skewed, as the following histogram shows:
As a result, statistical analysis with this variable revealed the presence of heteroskedasticity, which violates the assumptions necessary for Ordinary Least Squares regression. Therefore, I transformed the variable using the natural log, which gave it a roughly normal distribution and eliminated heteroskedasticity in the regression analysis:
The results of the regression analysis are reported below the fold:
. regress lnrp caucus open closed ocb clinch home Number of obs = 51 F( 6, 44) = 7.68 Prob > F = 0.0000 R-squared = 0.5114 Adj R-squared = 0.4448 Root MSE = .36564 ------------------------------------------------------------------------------ lnrp | Coef. Std. Err. t P>|t| [95% Conf. Interval] -------------+---------------------------------------------------------------- caucus12 | .4508634 .1559367 2.89 0.006 .1365937 .7651332 open12 | .0994477 .1677612 0.59 0.556 -.2386528 .4375483 closed12 | -.0407699 .154206 -0.26 0.793 -.3515516 .2700119 ocb | -.1784982 .045436 -3.93 0.000 -.2700685 -.0869279 clinch | -.3356797 .1908377 -1.76 0.086 -.7202878 .0489283 home | .126041 .237436 0.53 0.598 -.3524799 .6045618 _cons | -1.229249 .3189755 -3.85 0.000 -1.872101 -.5863958 ------------------------------------------------------------------------------
The results basically fit our prior expectations: Paul did better in caucuses* and open contests, worse in closed contests, when there were more candidates on the ballot*, and when Mitt Romney had clinched*, and better when he benefited more or suffered less from a home-state advantage for a candidate. (The asterisks (*) indicates statements derived from statistically significant coefficient estimates.) Somewhat surprisingly, these variables alone are sufficient to explain just over half of the variance in Paul’s support across states, as indicated by R-squared, even though we have included no measures of ideology or campaign effort. (However, one of the reasons Paul did better before Romney clinched and in caucus states was surely that the Paul campaign dedicated more resources to these contests.) The residuals of the model should largely capture ideology and aspects of relative campaign effort not picked up by other variables. Here I rank the states by how “pro-Ron Paul” they were in 2012, when factoring out the apparent effects of institutions:
| State | residual |
| New Hampshire | 0.825768 |
| Vermont | 0.815371 |
| Rhode Island | 0.667037 |
| Pennsylvania | 0.642602 |
| Virginia | 0.571283 |
| Iowa | 0.348832 |
| Maine | 0.337433 |
| Massachusetts | 0.243393 |
| New York | 0.217729 |
| Michigan | 0.189426 |
| Oklahoma | 0.168561 |
| South Carolina | 0.15903 |
| Connecticut | 0.13723 |
| Indiana | 0.136645 |
| Washington | 0.128566 |
| Texas | 0.113333 |
| North Dakota | 0.111803 |
| South Dakota | 0.102509 |
| Alaska | 0.100775 |
| Oregon | 0.083904 |
| Kentucky | 0.064148 |
| Montana | 0.063033 |
| Wisconsin | -0.0036 |
| Arkansas | -0.00969 |
| Kansas | -0.0119 |
| Maryland | -0.02353 |
| Tennessee | -0.03488 |
| Wyoming | -0.04629 |
| Ohio | -0.08236 |
| Missouri | -0.08555 |
| Minnesota | -0.08936 |
| North Carolina | -0.096 |
| West Virginia | -0.10322 |
| Delaware | -0.10785 |
| Hawaii | -0.12206 |
| New Mexico | -0.12313 |
| Georgia | -0.13006 |
| California | -0.1328 |
| District of Columbia | -0.13284 |
| New Jersey | -0.16776 |
| Idaho | -0.18677 |
| Nebraska | -0.21019 |
| Colorado | -0.25117 |
| Louisiana | -0.27987 |
| Florida | -0.3154 |
| Nevada | -0.31986 |
| Illinois | -0.36137 |
| Arizona | -0.46823 |
| Alabama | -0.63111 |
| Mississippi | -0.75494 |
| Utah | -0.9466 |
Ron Paul had surprising strength in northern New England and Rhode Island. All four of those states were in the top seven. Indeed, Paul did rather well in Massachusetts and Connecticut as well. Perhaps there is an untapped libertarian constituency in New England? Paul’s strength in New Hampshire was perhaps not surprising, as it was his top state in 2008. One can say much the same about his performance in Pennsylvania (despite the fact that this was one the few states where Paul actually did worse in 2012 than in 2008, dropping from 15.9% to 13.1%). As a rule, Paul did badly in strongly Protestant and Mormon states and in left-leaning, highly urban states like Illinois, New Jersey, D.C., California, Hawaii, and Delaware. The former were likely to go for Santorum or Gingrich (or Romney if Mormon), while the latter went heavily for Romney.
The final bit of analysis I’ve done is to add variables for Libertarian (Barr and Phillies – and half of Paul’s Constitution Party vote in Montana) general election support from 2008 and Democratic and Green general election support from 2008. These indicators of state ideology should help improve our estimates of the associations of primary institutions with Ron Paul’s electoral support. Here are the results:
. regress lnrp caucus open closed ocb clinch home lp08 demgr08 Number of obs = 51 F( 8, 42) = 6.74 Prob > F = 0.0000 R-squared = 0.5621 Adj R-squared = 0.4787 Root MSE = .3543 ------------------------------------------------------------------------------ lnrp | Coef. Std. Err. t P>|t| [95% Conf. Interval] -------------+---------------------------------------------------------------- caucus12 | .5364822 .1576557 3.40 0.001 .2183201 .8546443 open12 | .0962812 .1636184 0.59 0.559 -.233914 .4264765 closed12 | -.0960194 .1518778 -0.63 0.531 -.4025213 .2104824 ocb | -.1500122 .0459341 -3.27 0.002 -.242711 -.0573134 clinch | -.2519858 .1931687 -1.30 0.199 -.641816 .1378445 home | .1946671 .2333168 0.83 0.409 -.2761853 .6655195 lp08 | .0554998 .2246781 0.25 0.806 -.3979189 .5089185 demgr08 | .0109411 .005007 2.19 0.035 .0008366 .0210457 _cons | -1.977219 .4672169 -4.23 0.000 -2.920101 -1.034337 ------------------------------------------------------------------------------
The results, surprisingly, show no statistically significant association of Libertarian support in 2008 with Paul’s primary support in 2012. On the other hand, Democratic-Green support in 2008 is positively and statistically significantly associated with Paul support in 2012. One might infer from this result that Paul’s 2012 candidacy to a significant extent relied on attracting progressives into the Republican primaries, as improbable as that sounds. However, another possibility is that Republicans and independents in left-leaning states have become more libertarian over the last four years. To distinguish between the two hypotheses, I interact the variable for “open contest” with Democrat-Green support in 2008. If it’s true that Paul actually got lots of votes from people who normally vote Democratic, then the nature of the primary should have made a difference for his ability to do this. Those results are as follows:
. regress lnrp caucus open closed ocb clinch home lp08 demgr08 demgropen Number of obs = 51 F( 9, 41) = 6.32 Prob > F = 0.0000 R-squared = 0.5812 Adj R-squared = 0.4892 Root MSE = .3507 ------------------------------------------------------------------------------ lnrp | Coef. Std. Err. t P>|t| [95% Conf. Interval] -------------+---------------------------------------------------------------- caucus12 | .5419239 .1561067 3.47 0.001 .22666 .8571877 open12 | -.7515828 .6415346 -1.17 0.248 -2.047188 .5440226 closed12 | -.0875847 .1504633 -0.58 0.564 -.3914516 .2162822 ocb | -.1413276 .0459104 -3.08 0.004 -.2340456 -.0486097 clinch | -.2146438 .1931531 -1.11 0.273 -.6047245 .1754368 home | .2160615 .2314797 0.93 0.356 -.2514213 .6835443 lp08 | .0383642 .2227516 0.17 0.864 -.4114919 .4882202 demgr08 | .0083372 .0053102 1.57 0.124 -.002387 .0190614 demgropen | .0170368 .0124733 1.37 0.179 -.0081536 .0422271 _cons | -1.896288 .4662559 -4.07 0.000 -2.837911 -.954665 ------------------------------------------------------------------------------ . lincom demgr08+demgropen ( 1) demgr08 + demgropen = 0 ------------------------------------------------------------------------------ lnrp | Coef. Std. Err. t P>|t| [95% Conf. Interval] -------------+---------------------------------------------------------------- (1) | .025374 .0116714 2.17 0.036 .0018031 .0489449 ------------------------------------------------------------------------------
The hypothesis test following the regression results tests whether the sum of the coefficients on Democratic-Green support and the “interaction term” with open contests is statistically significantly different from zero. It is. In other words, when there is an open primary, we can be highly confident that Paul’s support was increasing in a state’s left-wing tilt. When there is not an open primary, we cannot be so confident of this association (it is not quite statistically significant). The way I judge this is that, indeed, disaffected Democrats did tend to vote for Ron Paul in Republican primaries in numbers large enough to make an observable difference, when they were allowed to. But it’s also probably the case that more Republicans in more left-wing states are pro-Paul.
Note as well that the institutional results are largely unchanged. The model still sees Paul as doing much better in caucus states and when fewer of the significant candidates were on the ballot. The only change in statistical significance is that the “clinching” variable is no longer statistically significant, although it remains negative.
In summary, the patterns in Ron Paul’s electoral success across the states once again support the case that he was an outsider candidate with an unusually motivated activist base. They also show just how different Paul supporters are from average Republicans. Nevertheless, Republicans in left-leaning states might learn something from Paul’s campaign about how to win in unfavorable circumstances.
PILEUS 
As always, your analysis is greatly appreciated. Here are a few thoughts I had as I read through your post.
– In looking back at 2008, you created a hypothetical situation in which each state had the “average” conditions of, in that case, “primary not caucus, pre-clinching, with 5 candidates in the race, and a turnout of 6.27%.” Any chance you can look at the result for 2012 if you used the median value of the variables? I know this isn’t a true apples to apple comparison of 2008 and 2012, but it might be one way of looking at ways in which Ron Paul’s performance shifted from 2008.
– Why not include Ron Paul’s 2008 primary performance as a variable?
– Finally, your conclusion, “Nevertheless, Republicans in left-leaning states might learn something from Paul’s campaign about how to win in unfavorable circumstances.”
I have seen a fair amount of analysis, pushed by the Cato Institute and Reason Magazine, that tries to identify the “libertarian” wing of the Tea Party movement, and what this movement means for Republicans as a whole. The problem is that it ends up focusing on roughly half, or less, of the Tea Party movement as the last, best hope for libertarianism within the GOP, and ignores that according to their own analysis of American voters libertarian-leaning voters split almost equally between the Tea Party/GOP and Democrats/Independents/Others.
I’ve never been able to get a response on how exactly libertarians in the Democrats/Independents/Others category breaks down, but a Republican in a left-leaning state would benefit from knowing both that base Republicans in the state lean more libertarian AND someone like Ron Paul has demonstrated a history of pulling in some Democrats/Independents into the primary process. Maybe you have some thoughts on how to go about looking at this subset of libertarians?
The hypothetical situation would be interesting. I’ll see about getting that info soon. As far as including Paul’s 2008 support goes, I didn’t feel comfortable using the raw number because the calendar changed so much – however, I could have used the “adjusted” vote share variable I came up with last time. But the models were different, so I was nervous about that too. And lazy. But if I have a moment that week, I’ll give that a try too.
P.S. on 2012 vs. 2008, I looked at rank ordering the states by performance, removed Wyoming and Hawaii as they weren’t in the 2008 list, and looking at the shifts in where the states ranked.
Some of the big improvements were a little confusing, like Indiana, Oklahoma, South Carolina, and Kentucky. As were some of the declines, like Idaho, Nevada, and Montana.
Oddly, the best state was New Hampshire. And the bottom didn’t move around much. Alabama was 47th both times, Mississippi shifted from 48th to 49th, and Utah dropped from 44th to 49th.
I don’t know if you’d have any more comments on this.
Wow…what an analysis. You must have taken a special high school stats class!
There is a difference between open primaries (in which any voter can choose any party’s primary ballot) and semi-closed primaries (in which independents are free to choose any party’s primary ballot, but members of other qualified parties don’t have that freedom). When this article uses the term “open primary” I am not sure which type of primary is meant.
The variables I used were for fully open and fully closed primaries. The excluded category consisted of the semi-open/semi-closed contests in which independents can vote. Sorry if that wasn’t clear.
Amazing work. Thankyou very much.