Update: Deepest apologies. I’ve been staring at too many spreadsheets lately and made a major screw-up. The previous version of this included the election day tallies in the early vote calculation which is why there was such a large spike on the final day. The numbers below have been corrected.
In 2008 Barack Obama won Nevada by 121k votes. Nevada is unique to most states in that it is really no more than 3 areas, Clark County (Las Vegas), Washoe County (Reno) and everyone else in between. In 2008 Clark accounted for 67% of the vote, Washoe accounted for 19% of the vote and the rest of the state accounted for 14%. President Obama won Clark by 123.7k and Washoe by 22.8k but lost the rest of the state by 25k (effectively cancelling out Washoe). For these reasons Clark County gets the lion’s share of attention although Washoe gets its fair share. The increased focus on early voting also shares this focus with Republican’s chances of carrying Nevada depend greatly on their ability to stay close in Democrat-heavy Clark (“minding the gap’), while trying to flip Washoe back into the GOP fold and running up leads everywhere else.
Greatly assisting the Democrats’ win were huge margins in early voting turnout. By election day in 2008 the Democrats’ aggregate advantage over Republicans was 83,633 ballots cast, 69% of their final margin. We know a cast party ballot doesn’t necessarily mean it was cast for that party but for simplicity sake we assume each side has the same base and crossover support. The key for Republicans obviously will be to “mind the gap” in early voting where Obama achieved most of his success. Early voting in Nevada only started a couple days ago and after the first day of returns I created a rather crude model for how the rest of early voting could go based on the 2008 historical information. Although the model was basic, assumptions were honest based on precedent and conservative based on the available data.
The assumptions were as follows:
- Democrat turnout after Day 1 would average 97% of the day 1 result consistent with 2008
- Republican turnout as a share of the Democrat turnout would increase it at a rate of 2.4 percentage points (pp) every day consistent with 2008
After Day 1 our model expected the Democrat advantage in Clark County heading into election day would be 81,786, below the 86,607 margin in 2008. This would eat into Obama’s overall margin by approximately 4%.
We now have the Day 2 results for Clark County and my conservative assumptions have proven to be just that. Rather than a 2.4pp increase in the day-over-day change in Republican vote proportional to the Democrat vote, it jumped 13pp. Additionally Democrats saw a steep drop-off in the day-over-day change in turn-out falling -29pp in Day 1 while Republicans only trailed off by -11pp.
By simply plugging the new actual #s into the model and changing no other assumptions, we now crudely forecast a final early vote margin in favor of Democrats of 76,963. This is 4,823 less than yesterday and 6670 less than 2008 — an 8% decline.
Not shown in the charts is that in raw #s the aggregate lead of Democrats on Day 2 in 2008 was 16,385 versus today where it is 13,295 — a 3090 actual ballot decrease in only the first two days of early voting.
We see after two days of early voting in Clark County Republicans are slowly eating into what proved to be Democrats’ insurmountable lead in 2008. This is all before considering the two other segments of Nevada where all indications show Republicans greatly outstripping their 2008 efforts.
If anyone has Washoe’s 2008 daily early voting #s by party I’d be happy to run them as well. [Thanks to one hell of a guy, I have the Washoe #s] Two days of early voting doesn’t mean Nevada will flip from blue to red, but Republicans are clearly on the right path towards “minding the gap” in Clark County.