(CNN) Republicans gained a seat in the Connecticut state Senate this week. They won a special election in a district that President Joe Biden won by more than 20 points in 2020.
Any individual special election such as this one comes with its caveats, but the trend in special state elections this year is becoming noticeable. Republicans are doing better than they were in the beginning of the year, and it could foretell their party's success in 2022.
Across more than 30 special state legislative and federal elections during the Biden presidency, Republicans are doing 4 points better on average than former President Donald Trump did in these same districts last year.
Some of these races included multiple Democratic and/or Republican candidates running at the same time in what is known as a jungle election. If we focus only on non-jungle races in data compiled by Ethan Chen, Republicans are outperforming the 2020 presidential baseline by 3 points.
Either way you look at it, there has been a slight Republican overperformance compared to where things stood last year.
Such a change is not too large. Remember, too, that Trump lost nationally by 4 points, so a 4-point swing toward Republicans suggests a neutral national environment.
But this would likely be enough for Republicans to take back the US House of Representatives, especially considering that they are in a good position for redistricting.
What really jumps out, though, is the trend.
When you look at the first 17 special elections this year (through early April), the Republican overperformance over Trump was just a point. Examining the last 17 special elections, the overperformance has been 7 points. When you splice the data even further, Republicans have been outperforming the 2020 baseline by double-digits since the beginning of July.
Whether such a shift sustains itself can't be known at this point. Things may shift back to Democrats.
It shouldn't be too surprising if the Republican overperformance does hold. Back in 2009, there was a big movement away from Democrats in special elections toward the middle of the year. This foretold Republicans doing very well in the 2010 midterms.
The one thing that is clear is that the special elections this year look nothing like they did four years ago at this point. Back then, Democrats were outperforming Hillary Clinton's performance by 14 points in the average special state legislative and federal elections.
Democrats went on to take back House control in 2018.
What we saw in 2010 and 2018 fits with what we know about special elections. Individually, they can be all over the place. Together, special elections can be a leading indicator of House results in midterm elections.
They were also an indicator that the good polls for Democrats heading into 2020 might have been a bit too rosy.
Importantly, the shift in favor of Republicans in special election results comes as other indicators suggest that the environment is getting better for the GOP.
The Democratic advantage on the generic congressional ballot is down to an average 2 points over the last month, after being 4 points for much of the year.
Biden's job approval rating has been tilting downward. His net approval (approval - disapproval) rating is only about +3 points at this point. It had almost always been +10 points or above before July.
Combined, however, these numbers tell a clear story of Republicans picking up ground.
The story may become even clearer a month from now in the California recall gubernatorial election. You might think of it as the biggest non-regularly scheduled election of the year. Biden won California by nearly 30 points.
A Democratic governor in California shouldn't be recalled or be close to being recalled, if Democrats are doing well nationally. The polling suggests that the race could, in fact, be close.
It's important to note that the shift we're seeing across these metrics now is what we normally expect in the lead up to a midterm election. The president's party almost always loses ground in midterm elections.
Perhaps the ultimate takeaway from this data is that we're more likely to experience a normal outcome in 2022 than something unique.