My take on the PA-SEN race

*P.S. I wrote this a day before the 2022 midterm election. My analysis and intuition didn't sit well with Sabato Crystal Ball's PA-SEN projection, so I decided to explain my reasoning through what has turned into a mighty long article. Now, I am definitely glad I did it! Japanese readers, just plug this into DeepL and you're going to be able to get the gist of it!

アメリカの大手メディア・最強政治アナリスト集団 Crystal Ball に異論を唱えて、ペンシルベニア州上院選にFetterman (民主）勝利の予想を下したところ、大当たりしたので選挙前日（日本時間11月8日）に書き上げておいた文章をここに載せておきます。

I’m writing this on the morning of 11/8(JST, 11/7 US time), a day before the 2022 midterm elections.

This morning, one post on my Twitter feed immediately caught my attention:

Crystal Ball rates PA-SEN as Leans Republican

For those who are new to American politics, Larry Sabato’s Crystal Ball is a nonpartisan political analysis and newsletter produced by Larry J. Sabato, Kyle Kondik, and J. Miles Coleman. Simply put, some of the most prominent household names in election analysis today. So prominent, in fact, that Kyle’s analysis of the 2020 presidential race was quoted on the front page of the biggest newspaper in Japan, The Yomiuri Shimbun. Considering that American political coverage in Japanese media usually whittles down to cheap copies of “In this rural diner” stories run by the likes of WSJ and extremely surface-level race projections, this is a huge deal.

When Crystal Ball released their final electoral college projections in 2020, I want to be on the record that I was, unfortunately, one of the people who scoffed at their GA-Lean D and FL-Lean R prediction, and oh boy, was I wrong. They went on to call 49/50 states correctly, only missing North Carolina while I was sitting in my chair deluding myself into thinking that Ohio was actually competitive. In 2021’s VA-GOV election, I was ever so confident that the people of my Biden +10 home state would keep the Governor’s mansion blue. Turns out, not everybody in Virginia is from Fairfax county (albeit it did see a lot of Biden voters ticket splitting for Youngkin) and Crystal Ball’s Lean R rating was vindicated yet again.    

Albert Einstein once said; “Insanity is doing the same thing over and over and expecting the same results.”

I hate to do it, I really do, but my brain and gut think that PA-SEN is Leans D.

Before I explain my reasoning for Pennsylvania, let’s take a look at the state of play in Georgia. The reason that I bring Georgia up is because it’s a swing state with a competitive senate & governor election WITH a significant amount of ticket-splitting expected. Incumbent governor Brian Kemp (R) is running against Stacy Abrams (D) and incumbent senator Raphael Warnock is running against his controversial opponent Herschel Walker (R). Both incumbents are great candidates who have over-performed baseline partisanships in races before who are up against (for reasons I will not delve into here) less-than-ideal candidates. The candidate quality difference in these two races is stark, and yet, polls show the delta between these two races suggests their delta to be only about 5 points.

Let’s now shift our focus back to Pennsylvania. In the governor’s race, State Attorney General Josh Shapiro (D), who is pound for pound one of the strongest Dem recruits this cycle, someone with a proven track record of stellar electoral performances in the past, is up against far-right state senator Doug Mastriano (R) who attended the infamous January 6th riots. Shapiro is taking advantage of his massive war chest and his name ID, establishing a dominating lead in the polls in the very early stages of the race. On the other hand, saying Mastriano’s campaign is a dumpster fire would be the understatement of the century. Not only does he lack funding or anything resembling a coherent campaign message, but it also took him until VERY recently for him to put up his first statewide ad, and instead of engaging in actual GOTV operations, his campaign announced it will engage in a “40-day fasting and prayer”… 40 days removed from Election Day. Lakshya Jain, one of Election Twitter’s most respected forecasters, joked that he is doing it out of necessity because his campaign is so broke that he cannot even afford food for his team. Even worse for him, he is running against Shapiro, who consistently outperforms his fellow Dems, especially in the northeastern part of the state (In 2020, he significantly ran ahead of Biden even though the current President is from the Scranton area and likely had a home turf-boost.) As expected, the FiveThirtyEight polling aggregate has him up by 10.4 at the time of this writing, and suffice to say, I don’t expect this race to be all that competitive in the end.

Now, here is the confusing part. The FiveThirtyEight aggregate for the senate race has Mehmet Oz (R) leading  John Fetterman (D) by 0.5 points. That is a polling delta of approximately 11(!!!) points, and I will explain why I am very skeptical that it manifests tomorrow. Businessman/TV celebrity Dr. Oz is running against Lieutenant Governor John Fetterman for an open senate seat vacated by Pat Toomey (R). Fetterman led the race for a time and hammered Oz with effective attacks, but since suffering a stroke and being sidelined from campaigning for a period of time, Oz has picked up massive momentum and took the lead in the polls just a couple of days ago. However, I myself am not on the Oz bandwagon, here’s why.

1: The Polls in Question
The polls that have shown Oz leads recently and have skyrocketed him in the aggregates are from the following polling organizations; Remington Research Group, InsiderAdvantage, Trafalgar, and Patriot Polling. In short, polling firms that are known to have been very bullish on Republicans in the past, show extremely rosy polling for Republicans in other states that are contrary to what independent polling suggests. To make matters worse, all of these polls have Shapiro up by at least 6 if not more, and have him at around 52% of the vote. Even if we assume that undecideds break heavily Republican, that will translate to a Shapiro victory of over 4 points.

2: Candidate Quality
Simply put, Oz is not good enough of a candidate and Fetterman is not a bad enough one for the delta between the races to be this significant. Oz has had negative favorables since the beginning of the campaign, with even recent polls showing his disapproval crack 50%. Meanwhile, even while getting hit with negative ads and his health condition deteriorating, Fetterman’s net approval is above water, even outpacing current governor Tom Wolf. As I stated above, even in Georgia, the delta between the two races is going to be 5 points give or take, and there’s no evidence that R’s have a candidate quality edge that will enable Oz to pull off something similar to Brian Kemp in Georgia.

1.+ 2. = 3. Overestimated Delta:
In conclusion, I have a hard time seeing Shapiro’s victory being under ~3 points and the delta between the two races being less than 4 points, which has led me to the conclusion that Fetterman is favored conditional on Shapiro winning by more than 4 (which I think is likelier than not).

In summary, I expect a Shapiro victory of more than 4, which will be coupled with a narrower Fetterman win against Oz. The following is how I expect the race to go tomorrow night.

Shapiro + >4 = Lean Fetterman
Shapiro + 3.5~4 = pure tossup
Shapiro + <3.5 = Lean Oz
Shapiro + <2 = Likely Oz

Again, bucking Crystal Ball is never a good idea, but, the third time’s a charm?…

Yu Yamada

追記　＝　最終的にShapiroは15ポイント差、Fettermanは5ポイント差で勝利しました。