Headlines broke last week about Midwestern Democrats doubling down on their push for the 2020 Democratic National Convention (DNC) to be held in Milwaukee, Wisconsin. Milwaukee is one of three finalists under consideration by the Democrats along with Houston and Miami, and is the only city among the three to have never held a party convention before. Meanwhile, the Republicans landed on Charlotte, North Carolina as the host city of the 2020 Republican National Convention (RNC) back in 2018 after briefly considering Las Vegas (which also bid for the DNC but withdrew).

Of course, it would be naive to think a decision made at this high of a level is ever nonpartisan. Having a convention in Florida boosts the party in a large key swing state that seems to be rapidly running away from Democrats in the Trump era. Holding the convention in Houston would capitalize on the momentum built by Beto O’Rourke’s senatorial campaign towards turning Texas purple and put Republicans on the defensive on their own turf. For the Republicans, both Charlotte and Las Vegas are located in states that are moving away from them; North Carolina will hold what is likely to be a fairly competitive and critical Senate race in 2020, and Republicans have suffered major losses in Nevada in the last four years. And then there’s Milwaukee, the largest city in Wisconsin, the “tipping-point” state in the 2016 election, which dealt the death blow to Hillary Clinton in the Electoral College. Democrats are desperate to shore up their support in the Midwest and repair their reputation after experiencing major electoral collapses in states like Wisconsin, Michigan, Pennsylvania, Iowa, and Ohio (all states Barack Obama carried in both 2012 and 2008) in 2016. Hosting their convention in a major Midwestern city is a signal that Democrats are willing to put in the work to win those states back.

This begs the question – does the selected host city actually make a marked difference when it comes to the party’s performance in that state in that presidential election cycle? It feels like it would; choosing a region, a state, and a city are not arbitrary decisions – they are made by the party with an eye towards messaging, scheduling, and logistics, and electoral advantage – the latter item being the ostensible goal of political parties: to win elections so they can execute policy and so forth.1There is a lot of interesting – and I’d argue more accurate in the American political system – data supporting the inverse, that politicians actually support policy in order to win elections.

To test this, I collected data on how each party performed in each state where a convention was held (as far back as 1948) as well as their performance in the election on a national level compared to that state, and then comparable data from the prior election cycle in that state. Subtracting the year in which the convention was in the state from the previous year (adjusted for national performance by that party in each year) would give us a rough calculation of how much of an effect the location of the convention might have had in that state. For example, in 2016 the Republicans held their convention in Ohio. Donald Trump then received 51.69% of the vote in Ohio and 46.09% of the vote nationally (a 5.6% over-performance in that state). In the previous cycle, Mitt Romney received 47.69% of the vote in Ohio and 47.2% of the vote nationally (a 0.49% over-performance in that state in a year where there was not a convention there), bringing us to a 5.11% bump adjusted for national performance in the state where that party’s convention was held in 2016.

Running a regression to determine statistical significance, no matter how I twisted it, proved fruitless – a conclusion I eventually accepted as perhaps a more interesting one. The best I could do was find an uninspiring mean of 0.22% increase for the party in the convention host state, adjusted for national performance (this is not a statically significant number, the spread is all over the place and does not seem to have a relationship to whether or not the party held its convention there). But, while this may mean that convention location doesn’t lead to a significant increase in performance in the host state, that doesn’t mean that it is a useless exercise. In 2004, the Republican National Convention was held in New York City, a city in a deep blue state that Bush didn’t have much of a chance of winning even if they did see an increased performance of a percentage or two – but that misses the forest for the trees. Holding the convention in New York was symbolic; the convention contained many references to the 9/11 attacks and featured a tribute to those killed, speeches from the relatives of victims, and was organized around a theme titled “Fulfilling America’s Promise by Building a Safer World and a More Hopeful America.” I would argue this is a good use of convention location, and far more relevant to determining electoral performance overall than the choice of having it in a particular state in order to woo that state itself, which is where convention location may matter.

What does this mean for the Democrats and Republicans and their 2020 conventions? It means conventions are probably more useful for sending a subjective national signal, rather than achieving an objective numerical benefit locally. If the central theme of 2020 Democratic Convention is remembering their roots – the workers, underprivileged, and “left behind” that likely cost them the Midwest and a presidential victory in 2016 – then holding a convention in Wisconsin or Michigan (which are looking more and more like the major battlefront of the 2020 election) is a better optical choice than holding it in New York or California. In other words, Milwaukee is probably the Democrats’ best bet, unless they want to lean far in the other (probably in the long-term more successful) direction and host a convention in Arizona, Texas, or North Carolina to take advantage of those states’ shifting demographics and build a platform around globalization, moderation, and progress. The bolder and more offense oriented option for Democrats would be to send an inclusionary message (which has dominated their platform and messaging for the last few years) and hold their convention in Arizona, New Mexico, or Texas, emphasizing their stance on immigration, inclusivity, diversity, and on courting suburban voters (a major block in their 2018 midterm victory).

For Republicans, North Carolina is about the most boring choice possible. It’s a mildly Republican state (it’s shifted left in this century, and even went for Obama in 2008, but is still fairly red for the time being), and would be akin to Democrats holding a convention in Oregon: bland, in line with party orthodoxy, and not striking a serious point. Whereas in 2016, Republicans held their convention in Cleveland, Ohio, a huge symbolic gesture towards reclaiming lost ground that fit in very nicely with the Trump campaign’s themes of fighting for the left behind and unheard blue collar workers (Trump also knocked Ohio out of the park, the state was barely even in play for Clinton to begin with, which should have been a sign of things to come across the region). If I were the RNC, I’d have gone with something a bit bolder – like holding the convention in Alabama as a rebuke to the Doug Jones’s shocking 2017 Senate special election victory, setting themselves up to retake the seat and entrench their Southern roots in 2020 after experiencing a series of electoral setbacks since Trump’s election. Another great option would be to send a policy signal since the Trump administration has been largely focused on very specific policies: building the wall, protectionism, immigration curbing, etc. Hosting the convention in southern California, Arizona, or Texas would be a good way to capitalize on the Republicans’ immigration messaging and set them up to be a bit more assertive in places that almost slipped away from the party in the 2018 midterms (this parallels the philosophy of Democrats hosting their convention in Wisconsin in response to the 2016 election). Were this a different version of Republican Party, more in line with the Bush-Cheney era, Virginia or Colorado would be easy picks (whereas now they’re boring picks leaning more Democratic) that would emphasize the defense industry, educated suburban white voters, and a broader “compassionate conservatism” instead of  “building the wall” and North Carolina’s solidly Republican base.

But in general, I think both parties can rest easy that their choice of location will probably not help or hurt them anyway. This is not an era centered on sending a national message, like it was post-9/11 or during the Great Recession. The suggestions I’ve made reflect the demographic and policy futures of each of those parties. Democrats need to come to terms with whether they want to appeal to working class voters, or try to take the cake on the Republicans’ former and rapidly-eroding hold amongst white suburban well-educated voters and build a coalition of urban, diverse, millennial voters. And Republicans will need to come to terms with whether they are firmly going to move in Trump’s direction, with immigration and protectionism in the spotlight, or come back around to the Wall Street-types and traditional conservatives of the 1980s to try and connect with a more diverse and centrist nation.