One of my favorite social campaign slogans of all time is "Neighbors for more neighbors" -- the mantra of supporters of upzoning in Minneapolis-St. Paul. And to co-opt it, Democrats should support policies that create more Democrats.
At one level, that's obvious; at another, it's obscure. What does it mean for a policy to "create" more Democrats? It'd be nice if "good policy that makes people's lives better" had a direct connection to getting more votes, but I'm dubious. Typically, the process through which people become members of a political party is a little less straightforward -- working through cultural affinity and other group dynamics as much if not more so than policy preferences. And on the other side, we should not support a policy that's objectively unethical just because it might redound to the transient political advantage of the Democratic Party. All politics is, in a sense, a trade-off between what's right and what's expedient, but the very best political moves -- the true no-brainers -- are those which are both right and expedient. What we'd want, then, are policies that are both (a) objectively good and (b) are likely to inject more Democratic voters into the polity.
Statehood for DC (and the other colonies) is an obvious one -- it rectifies a clear injustice of areas under permanent American jurisdiction which lack political representation, and most of the relevant places are strongly blue-leaning (at "worst", places like Puerto Rico are swingy) and so would add more Democrats into American politics.
Immigration reform is, potentially, another. Again, it is correct on the ethics, but it also is likely that many (not all) of the immigrant populations will be inclined to vote blue -- particularly if Republicans insist on declaring loudly and consistently that the immigrants aren't welcome here. Accelerating paths to citizenship -- basically, creating a fatter spigot of naturalized U.S. citizens -- will likely yield more Democratic voters.
A less obvious play is policies which enhance college accessibility ("free college" or related programs), resulting in more Americans getting college-educated. The big story in American voter behavior over the past decade is that partisanship is now sorted almost entirely along the dimension of education -- higher-education cohorts voting blue, lower-education cohorts voting red (this holds even accounting for differences in wealth -- high-ed/low-income voters are still blue, high-income/low-ed voters are still red).
Does this mean that, if more Americans go to college, they'll come out Democrats? Not necessarily -- it could be that "people who are Democrats are more inclined to go to college" rather than "going to college makes people more inclined to become a Democrat" -- if that's the case, then adding new college attendees won't change the underlying partisan composition of the electorate. But I'm inclined to think that the causal arrow does flow in the direction of "college attendance --> Democrat" rather than vice versa. One hint that this is right is that we're seeing a big shift in voting patterns from college-educated voters who are long-since removed from college, which seems more compatible with college attendance --> Democrat than Democrat --> college attendance.
But what makes the pattern work? It's not because lefty professors are successfully indoctrinating students (as we often remark, we can't even get them to read the syllabus!). In part, it may be that college exposes students to people from a wider range of backgrounds and experiences than might otherwise be the case; that horizon-broadening experience fits better with political progressivism. But right now, I think the larger answer is simply a form of cultural affinity (or, to be a little cruder, tribalism): college-educated persons now are far more likely to be liberals than not, and that very consensus makes it more likely that each marginal member of the college-educated cohort will also be liberal (the same is true for non-college educated voters, but in reverse). People tend to adopt the politics of their surrounding community; if their community is fellow college-educated persons, they'll trend towards the predominant views of that set.
What this means is that if Democrats make a big push to increase the number of Americans who get college degrees, it is likely that the result will be more Democratic voters. It's not going to be everyone, of course. But I suspect if one randomly assigned a sample of Americans who were not planning to attend college into two groups -- one sent to college, one not -- the former would in four years have more Democratic voters than the latter.
It's good to give representation to places under American sovereignty. It's good to welcome immigrants who want to make their home here into the fabric of America. And it's good to increase college accessibility and affordability for Americans of all backgrounds. But each of these policies, in addition to their moral goods, may have the additional happy consequence of creating more Democratic voters. Democrats for more Democrats, please.