How to Know Who Won While Pennsylvania Dithers
Florida Points the Way Towards Victory
Pennsylvania sucks at counting votes. In 2000, Philadelphia didn’t finish counting mail-in ballots until the Saturday after the election. Even though Biden eventually won Pennsylvania by over 69,000 votes, it wasn’t obvious he would win Pennsylvania until Thursday morning. Pennsylvania has not gotten its act together since 2020. Its law still forbids election workers from opening mail-in votes until the polls open on election day. There will probably be fewer mail-in votes in 2024 than during the pandemic, and Philadelphia will probably hire more workers this time than last time. Still, I doubt the winner in Pennsylvania will be apparent on election night. In just over 15% of my simulations, Pennsylvania was essential to a Harris victory. Pennsylvania is essential to a Trump victory about as often. It is the tipping point state about a third of the time.
Fortunately, Florida is good at counting votes. It learned from bitter experience in 2000 and has enacted the most efficient vote-counting procedures in the country. Despite having only three-fifths of California’s population, Florida will count more votes on election day and will count the lion’s share of its votes within a few hours of the polls closing.
Florida is useful for its 30 electoral votes. Indeed, Harris has a 99.9% chance of winning the election if she wins Florida. However, even if Haris loses, Florida is a good bellwether because its voting patterns are strongly correlated with most other swing states. Strong correlation does not mean the margins will be identical or even similar. It means Florida will tend to move in the same direction as the states with which it is strongly correlated. A correlation coefficient of 1 represents perfect correlation. If Florida and Michigan had a correlation coefficient of 1, variations in Florida’s margin would explain all of the variation in Michigan’s margin. A correlation coefficient of 0 means two variables are statistically independent, and a correlation coefficient of -1 indicates perfect inverse correlation. Here are the correlation coefficients between Florida’s results and those of other swing states in Presidential elections from in 2000 to 2020.
AZ 0.8
GA 0.41
MI 0.81
MN 0.57
NV 0.86
NC 0.71
PA 0.8
WI 0.68
These correlations are strong enough — and Biden lost Florida badly enough — that a narrow loss by Harris in Florida would increase her victory odds from their current, 58% level. Here are Harris’s win probabilities based upon different margins of loss in Florida.
FL Loss
Margin Harris Victory Odds
>5-6 pts 2%
3-4 pts 22%
2-3 pts 49%
1-2 pts 75%
0-1 pts 91%
A close loss in Florida would be a bullish sign for Harris.