Herewith a plan for Democratic Party victories in elections.
Let us begin with several observations, and a few simple assumptions.
1. Only about 50% of the eligible voters actually vote in a presidential year, and rather fewer in the off years.
2. The geographically based winner-take-all structure of the American electoral system, combined with the high degree of residential segregation by income, by race, and to a lesser extent by ethnicity and religion, produces a patchwork of communities -- wards, precincts, parishes -- that are lop-sidedly either Democratic or Republican, even in Congressional districts or states that are fairly evenly divided.
Since this is central to my plan, let me take a moment to make sure this is clear. At the ward and precinct level, one sees electoral districts that are heavily Democratic or Republican, because voting is closely correlated with income, education, race, and to a lesser extent religion and ethnicity, and communities small enough to comprise a single ward or precinct tend to be quite homogeneous in this respect. Rich neighborhoods and poor neighborhoods, Catholic neighborhoods and Jewish neighborhoods, Black neighborhoods and White neighborhoods. Sometimes these homogeneous areas are located in Congressional districts that are largely homogeneous all the way through. These districts are so lop-sidedly Democratic or Republican that frequently they are not even contested and the incumbent runs unopposed. There is no point in the Republican Party putting up a candidate against Charley Rangel, for example.
Many relatively evenly balanced districts, however, have intensely one-sided pockets of Republican and Democratic voters, and this is especially true at the level of Senatorial campaigns. Hence the familiar phenomenon, on election night, of waiting for the vote from upstate or downstate or the inner city or the suburbs to come in. Seasoned election watchers know where the pockets of votes are that favor their candidates, and will not concede a seat until those precincts have been heard from.
3. Now an assumption, based on the combination of residential segregation and the correlation between voting patterns and the characteristics on which that segregation is organized: I think it is extremely likely that the non-voters in a ward or precinct, were they to vote, would break down pretty much in the same proportions as those who do actually vote. If a ward goes 75% for the Republican candidate, and half the eligible voters go to the polls, then it is very likely that the other half would vote Republican in roughly that proportion, were they to vote.
Do I have evidence for this assumption? No, but I think it is sufficiently plausible to serve as the basis for a serious experiment.
4. It is a fact that in many Republican Congressional districts, or Republican Senatorial states, there are wards and precincts that are heavily Democratic -- and vice versa, of course.
5. We come finally to the crucial question: Are there Republican Congressional districts or Republican Senatorial states that in recent years have been closely enough contested by the Democrats so that if a dramatic increase took place in the proportion of voters coming to the polls in the heavily Democratic wards and precincts, the balance of new Democratic voters over new Republican voters from those wards and precincts would be enough to swing the seat into the Democratic column?
If the answer is yes, then a precisely targeted get out the vote campaign might produce dramatic results. What is more, such a campaign could circumvent campaign finance laws by being strictly non-partisan. No attempt would need to be made to persuade new voters to vote Democratic. By targeting districts already known to be heavily Democratic, we could pretty well count on most of the new votes going for our candidates. Now, to be sure, this effort would result in an increase in the Republican vote total. If a district is 75% Democratic, and we get a thousand new voters to the polls, then 750 or so will vote with us, and 250 against us, so the net gain will be only 500. But that is fine, because all we care about is whether we are adding more Democratic than Republican votes.
Such a campaign would have no use for television advertising, which is so expensive, because there is no way to confine television advertising precisely to a single ward or precinct, and that is crucial to the success of the plan. Instead, the campaign would have to rely on hordes of foot soldiers, drawn if possible from the district itself, who would go door to door and try to bring people to the polls.
To prepare for such a campaign, we would first need a precise computerized database of the election results for the past two cycles or so from every single ward, precinct, and parish in America. This is not readily available, as I discovered when I tried to get this data just for Massachusetts. The Office of the Secretary of State in Boston stores the results in useful form only at the level of towns and cities, which is not fine-grained enough. But precinct breakdowns tend to appear in local newspapers a day or two after an election. Through the magic of the internet, it would not be hard to mobilize a nation-wide effort to secure the results for every single voting unit. Then some fairly simple data manipulation would suffice to tell us whether a 10% increase in turnout in selected wards and precincts would suffice to tilt an election, whether a 20% increase would suffice, and so forth.
Because the campaign would be strictly non-partisan in operation, I think it would circumvent financing restrictions, and since it would not use television, it would be inexpensive enough to be financed by several left-leaning multi-millionaires [or billionaires].
Couldn't the Republicans do the same thing? of course. But I think we would probably have an edge, because not only party loyalty but also voting behavior is strongly correlated with income. To put it another way, among that 50% who don't vote, there are almost certainly more Democrats than Republicans. That by itself does not settle the issue, of course, because the real question is this: Are there more Republican districts vulnerable to this sort of targeted get out the vote campaign than Democratic districts? I simply do not know, but it might be worth doing the data collection necessary to find out.