It's happening again.
You know what I mean- the bold predictions that Hillary Clinton will in the end come out on top of the polular vote, along with analyses that states she will not. As much as I respect Chris Bowers and his analysis at the latter link- this falls into what we know is a false narrative.
There is no such thing as the "popular vote" as a measure of support for the Democratic nomination.
For those who cannot bring themselves to agree with that observation above the fold, this diary is for you.
The 14 states which have held caucuses have done so in good faith- under the assumption that there is no inherent weight given to the type of contest- open or closed, primary or caucus and accordingly the pledged delegate count is an effective means of normalizing between the type of contest. One could argue that one type of contest is superior to another... I don't believe any of these arguments are particularly accurate. If the question is whether the candidate can organize strength among party activists and voters willing to get deeply involved with the process, caucuses are superior. If the Democratic nomination is predicated upon palatability within the party, a closed primary may be best. If we want general election viability only, perhaps we would prefer a nationwide open primary? If we want to assess "popular support" we cannot take all these different varieties of contest and tally up their raw counts between the states. The pledged delegate count would, in fact, be a more accurate portrayal of that sentiment.
If, on the other hand, we require numbers that reflect the sheer number of voters who influence this election, let's consider another measure. What would the caucuses have looked like if they were primaries- and how would they have influenced the "popular vote?"
Over the course of the campaign, the sum of the votes cast for Obama and Clinton in primaries average 70% of the 2004 vote for John Kerry in the general election. This ratio varies substantially based on how significantly the race in question (a standard deviation of 20%). Turning out a higher number of voters compared to the Kerry numbers does not appear to preferentially help either candidate. If there is a better predictor of primary turnout out there at this point I would be interested in hearing about it. For the meantime- if the 14 states that held caucuses instead turned out voters at the mean percentage of Kerry in 2004, where would the race be?
State | Obama | Clinton | margin | 2004 Kerry | frac04 | proj turnout | Obama | Clinton | |
ND | 11625 | 6948 | 0.6259 | 111052 | 0.16725 | 76975 | 48180 | 28796 | |
ID | 16880 | 3655 | 0.8220 | 181098 | 0.11339 | 125528 | 103185 | 22343 | |
AK | 6674 | 2194 | 0.7526 | 111025 | 0.07987 | 76957 | 57917 | 19040 | |
NM | 71396 | 73105 | 0.4941 | 370942 | 0.38955 | 257118 | 127038 | 130079 | |
KS | 27172 | 9462 | 0.7417 | 434993 | 0.08422 | 301514 | 223638 | 77877 | |
CO | 80113 | 38839 | 0.6735 | 1001732 | 0.11875 | 694348 | 467637 | 226712 | |
MN | 142109 | 68994 | 0.6732 | 1445014 | 0.14609 | 1001608 | 674256 | 327352 | |
NE | 26126 | 12445 | 0.6773 | 254328 | 0.15166 | 176287 | 119408 | 56879 | |
HI | 28347 | 8835 | 0.7624 | 231708 | 0.16047 | 160608 | 122445 | 38163 | |
WY | 5378 | 3311 | 0.6189 | 70776 | 0.12277 | 49058 | 30364 | 18694 | |
IA | 940 | 737 | 0.5605 | 741898 | | 514245 | 288247 | 225998 | |
WA | 21629 | 9992 | 0.6840 | 1510201 | | 1046792 | 716014 | 330779 | |
NV | 4773 | 5355 | 0.4713 | 397190 | | 275311 | 129745 | 145566 | |
ME | 2079 | 1396 | 0.5983 | 396842 | | 275070 | 164567 | 110503 | |
| | | | | | | | | |
Total | | | | | | 5031421 | 3272642 | 1758779 | 1513863 |
| | | | | | | 0.65044 | 0.34956 | |
This is the calculation with the original margin of victory. The number at the bottom right indicates Obama's advantage with those assumptions, yielding roughly a 1.5 million vote advantage over Clinton in these states. Clinton supporters would (and should) protest that the margin of victory is unlikely to be the same with the states holding primaries. For that purpose, let's take the difference between the Texas primary and caucus (8%) and subtract that, across the board, from Obama's support. This is a fairly crude way of adjusting the results, but is quite a friendly reduction for Clinton. In the case where we have a regional analogue, bringing Obama down to 60% in Minnesota is fairly comparable to his advantage in Wisconsin. We will know, perhaps, on the 20th whether Oregon matches up to Washington.
| | | | | | 0.69315 | adjusted | | |
State | Obama | Clinton | margin | 2004 Kerry | frac04 | proj turnout | Obama | Clinton | |
ND | 11625 | 6948 | 0.5459 | 111052 | 0.1672 | 76975 | 42022 | 34954 | |
ID | 16880 | 3655 | 0.7420 | 181098 | 0.1134 | 125528 | 93143 | 32385 | |
AK | 6674 | 2194 | 0.6726 | 111025 | 0.0799 | 76957 | 51761 | 25196 | |
NM | 71396 | 73105 | 0.4141 | 370942 | 0.3896 | 257118 | 106469 | 150649 | |
KS | 27172 | 9462 | 0.6617 | 434993 | 0.0842 | 301514 | 199517 | 101998 | |
CO | 80113 | 38839 | 0.5935 | 1001732 | 0.1187 | 694348 | 412089 | 282259 | |
MN | 142109 | 68994 | 0.5932 | 1445014 | 0.1461 | 1001608 | 594128 | 407481 | |
NE | 26126 | 12445 | 0.5973 | 254328 | 0.1517 | 176287 | 105305 | 70982 | |
HI | 28347 | 8835 | 0.6824 | 231708 | 0.1605 | 160608 | 109596 | 51011 | |
WY | 5378 | 3311 | 0.5389 | 70776 | 0.1228 | 49058 | 26440 | 22619 | |
IA | 940 | 737 | 0.4805 | 741898 | | 514245 | 247107 | 267138 | |
WA | 21629 | 9992 | 0.6040 | 1510201 | | 1046792 | 632270 | 414522 | |
NV | 4773 | 5355 | 0.3913 | 397190 | | 275311 | 107720 | 167591 | |
ME | 2079 | 1396 | 0.5183 | 396842 | | 275070 | 142562 | 132509 | |
| | | | | | | | | |
Total | | | | | | 5031421 | 2870128 | 2161292 | 708836 |
| | | | | | | 0.5704 | 0.4296 | |
Even under this scenario, Obama leads by a 700000 vote margin, meaning that the RealClearPolitics popular vote tally (at about 280000) considerably underestimates Obama's support. It would take a rather drastic reduction in Obama's support in these states to change that story. If one uses the same model on the remaining primaries, giving Clinton 2/3 of the votes in West Virginia and Kentucky and splitting Oregon, Montana and South Dakota 50/50 (which, on the whole, is vairly absurd) then Clinton makes up 240123 votes- a bit more than half of what is being unreasonably excluded from Obama's totals, leaving Puerto Rico to make up the difference.
The only reason to cite popular vote totals as they are tallied by RCP, outside of giving people a vague notion of the turnout and support in select large states, is to exclude the caucus states from consideration of the nominee. When one wishes to assess which candidate has popular sentiment in favor of the nomination, Barack Obama has an insurmountable lead. You can include the contest in Florida and it will not change that conclusion. You can include Michigan without crediting Obama with uncommitteds, and it will not change the result.
You cannot exclude 14 states and consider your math to be "popular" or "democratic."