How would the major parties and their candidates behave if they had to win the national popular vote plurality to become president? Making Every Vote Count will hold a conference on October 7, 2019 in Washington, DC to discuss answers to this question.

Here are my somewhat speculative beliefs.

We have to start with a point of view about what happens in the general election of the president. Mine is that this is right: “boosting turnout [in the general election], primary election persuasion, and perhaps persuasion in special elections [such as for ballot measures] are possible…But evidence of persuasion in general elections remained negligible.”

If turnout is the chief goal of the major parties in the general election, then it follows that each of the two will define their bases in a new and different way. Namely, in the first instance they will pay little attention to the geographic location of their likely voters. No longer will they focus on how to obtain a plurality in a handful of states. No more will they identify voters primarily according to zip code and census tract.

Instead, the campaigns, with their billions of dollars, will determine how to achieve a majority of the votes (they need a plurality but will aim at majorities, not knowing how many votes a third or fourth party candidacy may siphon away from the two major parties). To do this calculation, they will analyze the tendencies of all those likely to vote, and also those who are registered but less likely to vote, in every precinct in the country.

Based on the analysis, the Republicans and Democrats will determine the left-right political leaning of the likely and less likely voters. Presumably they will discover that about 45% of the possible national electorate is inclined one way or the other, and about 10% is undecided. That is what most polls suggest.

For these two different bases of their leaners, the Republicans and Democrats will then divide the potential voters into demographic segments. Each segment will be defined by age, gender, religion, race, favorite spoken language. They will poll each segment nationally in order to determine policy preferences.

For example, do Democratic-leaning old Hispanic females attach great priority to preserving Social Security and Medicare? If so, then the Democratic nominee will highlight his or her commitment to that policy, perhaps in Spanish language advertisements, and bewail the risk to these benefits that the Republican nominee will present.

As another example, do Republican-leaning middle aged white males fear an assault weapon ban by a Democratic president? If so, the Republican nominee will want to ask the Democratic nominee in a debate whether he or she will seek such legislation.

Policy will follow pragmatism: the parties are not going to push policies that deny them the possibility of winning the national vote.

The question then presented is how the parties’ policies will change if they must win the national vote to get the prize. My view is that it depends on how many segments need to be bonded into a coalition in order to compose a majority. The Democrats and Republicans have needed only to achieve this feat as against each other in less than ten states. In 2020 they are going to try to build winning coalitions in not many more than Florida, Pennsylvania, Michigan, Wisconsin, and New Hampshire, while paying relatively little attention to their respective bases in the rest of the country. But to win a national election, the job of building winning coalitions necessarily will require a different calculation.

It may be that opposition to abortion will still be critical to the Republican Party whereas hostility to immigrants may be politically disastrous. It could be that supporting aggressive government action to move to a clean power platform might be a central plank of a Democratic platform, while support for an open border is anathema. Polling and practice will have to answer these questions.

After figuring out the composition of hypothetical national bases, the parties then will decide how to reach their likely voters. Cost is king. I imagine that neither party has unlimited funds, and in any case during the general election timing matters a great deal. My guess is that for the first time in modern political history the parties will each spend money on national television in order to present – “brand” – their candidates during the summer. They will buy advertising time during the Olympics, the World Series, and nationally broadcast football games. Historically, almost no money has gone to national television, because of course only a handful of states have constituted the battleground and it would be a waste to spend advertising dollars on national shows.

On the other hand, the cost for reaching demographic segments through local radio, local television and local newspapers is fairly low in most regions of the country. This is not true in major urban markets, but it would be very cheap for the two parties to buy local media advertising space in the Dakotas, for instance. For the first time in history, the outlets in thinly populated non battleground states would be a valuable way for either party to reach their bases. While their messages would be different, both parties would be able to generate several hundred thousand turned out voters in the Dakotas very cheaply. In most elections the Republicans have left many voters unsolicited and not encouraged to vote in the thinly populated states. They would not pass up the opportunity to run up the score in these places, while Democratic leaning segments would be less numerous but perhaps even cheaper to reach.

In most states it is probable that the general national election would cause the gap between the two parties to close. This might be uncomfortable for state politicians in whichever party now dominates in a particular state. For instance, there’s little doubt the Republicans would seek turnout in California, as opposed to dedicating no advertising or organizing effort to the state now. The result might be revival of the two party system in the Golden State. Perhaps Democrats would not like that. On other hand, in pursuit of this goal the Republican nominee might favor a compromise policy on immigration.

Social media permits advertisers to target consumers by individual identity, advertising with versioned content to each pair of eyes. Such a capability would be used by both parties to encourage their bases to turn out. Mass media does not enable nearly as much individuation in advertising so it is reasonable to think the current trend of spending proportionally more in campaign dollars on social media would continue.

But the targeted voters would be identified without regard to geography. Social media advertising therefore would enable either party to reach all farmers, all Mormons, all Native Americans – all of any segment. As a result, demographic segments that are ignored, and sometimes derogated, by the current system would have much more importance in the election. They would still be subject to the inexorable law of numerosity: if your group is small, your electoral weight is small. But today many populous groups – African Americans are the leading example – have an inequitably minor impact in the general election, because many of their members are located in states where they are outvoted by people in a different political party and so they choose no electors from those states.

It might be that with national voting suddenly relevant, the Republican Party would need to adopt policies more sympathetic to the concerns of African Americans. Just like in Lincoln’s day.

Surely there are many other consequences to adopting a national system. But for sure many myths would be dispelled. All could see that a national election would empower people in states with low population and rural inhabitants, whereas they are taken for granted now. All would recognize that minorities of many kinds – based on religion, race, etc.—would have to be assembled into coalitions to support either of the major parties, while now they are ignored or taken for granted.

Some will argue too that a national election will threaten the viability of national parties. But there’s no way a regional party can win a national election unless and until the country’s electorate decomposes into at least four or five different regional parties. The existing media are not likely to facilitate this development – they present candidates the way they sell burgers, trucks, and software. Everyone gets the same products, with scarcely any regional variation.

However, in the unlikely event that national elections for president cause the deterioration of the existing Republican and Democratic parties, one can ask two questions. How much will they be missed? And doesn’t that mean the antagonists could agree at last on amending the Constitution to create a run-off so that under all circumstances the winner of the presidency has won a majority of the votes cast in the country?