A piece earlier this week in The Hill by Merrill Matthews made for interesting reading. In a nutshell, Dems are polling really well in key swing states when it comes to the Senate races. And yet when it comes to Trump vs Harris it is basically neck and neck in these states.

Here are the respective RCP averages:

For Senate:

Arizona – Gallego (Dem): +7.3

Nevada – Rosen (Dem): +10.7

Michigan – Slotkin (Dem): +4.8

Penn – Casey (Dem): +3.0

Wisconsin – Baldwin (Dem): +6.4

New Mexico – Heinrich (Dem): +9.5

For POTUS (Trump vs Harris):

Arizona – Trump: +1.0

Nevada – Harris: +0.6

Michigan – Harris: +1.1

Penn – Harris: Tie

Wisconsin – Harris: +1            .4

New Mexico – Harris: +10*

Matthews argues that if these Senate race numbers are any kind of reflection of how voters will actually vote in the presidential election then Trump could well be really losing in all these crucial swing states. And he thinks that post-Labor Day, and this close to the election, it is hard to see what could dramatically change any of the Senate polling numbers.

His key point is that although voters sometimes split their presidential and local, down ticket votes between the parties this has been trending down significantly in recent elections. So, we could all be underplaying the glaring klaxon that is the excellent Dem team Senate polling.

Back in June, Axios also noted this disparity between Biden’s polling in swing states and the Senate candidate’s polling. At the time Biden was between 2 and 5 points down in these states and doing far worse than the Senate candidates. Axios suggested this all reflected voters’ frustration at Biden being the candidate again. Democratic analysts they reached out to suggested, or at least were hoping, that once voters seriously focused on Trump’s likely return to the White House, they would move more toward their Senate Democratic preference and swing back behind Biden.

This all seemed very interesting as a theory, and — as I am personally desperate for any information that shows Trump 2.0 is actually not going to come to pass — I thought I would take a look at what happened in 2020.

Senate polling in these swing states just before election day in 2020**:

Arizona – Dem: +5%

Nevada – No election

Michigan – Dem: +7.6%

Penn – No election

Wisconsin – No election

New Mexico  – Dem: +8%

And the rolling POTUS averages on 3^rd September, 2020:

Arizona – Biden:  +5%

Nevada – Biden: +7.5%

Michigan – Biden: +2.8%

Penn – Biden: +4.2%

Wisconsin – Biden: +4.0%

New Mexico – Biden: +13%***

Biden’s final winning margin results in 2020 in those states:

Arizona – +0.3%

Nevada – +2.4%

Michigan – +2.8%

Penn – +1.1%

Wisconsin – +0.6%

New Mexico – +10.6%

Does this tell us anything? Arizona’s Senate final few polls averaged at a 5% Dem lead and 3^rd September POTUS polls gave him the same and yet Biden only won the state by a squeaky 0.3%. Michigan gave the Senate race Dem a 7% lead and Biden was coming in over half as much lower at around 3% on 3^rd September which was pretty much the final margin. In New Mexico, the pollsters reckoned the Dems would rock into the Senate with an 8% lead and yet gave Biden a 3^rd September margin of a bigly 13%. In the end he got a 10% win in the Land of Enchantment, which is kind of in the middle.

Unfortunately, with three of the states not having Senate elections in 2020 this comparison is somewhat lacking in data. But at best the picture, it seems to me, is rather mixed. There’s no clear sign that a fairly good Dem polling figure in 2020 in a swing state Senate race near the end of the race pointed the way to a better performance for Biden than his own polling was showing just before the election.

Sadly, I don’t think I can hang too much hope on the good polling for Dems for Senate races lighting the way towards bigger wins in these states for Harris on the day.

But I would be interested in other PBer’s opinion on all this.

Rotten Borough

* This is the Emerson/Hill poll of 22^nd Aug. RCP don’t seem to be doing an average for NM.

**RCP annoyingly doesn’t seem to have archived senate polling, so I used 270towin.

*** RCP did not bother with averages for New Mexico, so I used a poll dated in early September

https://thehill.com/opinion/4859013-swing-state-senate-races-trump