Not sure how many more times I can say "the impact is debatable" in recognition that individuals voting as other individuals would be an inefficient way to sway an election. However, it is fairly well documented that this happens to some degree:
https://ballotpedia.org/Dead_people_voting
But, since you asked, here is a scenario that I think is entirely plausible: A group of over-zealous activists travel to a very tight state with a handful of close precincts. They look up recent deaths in the areas and get a list of 1000 names and addresses that they then use to vote, each voting multiple times. They simply go in, give the name and address, and vote, then go to another polling location and repeat. Other groups decide on other tactics. A poll worker decides to stuff a box. A hacker changes a few 100 votes just to see if she can. No one technique alone would have swayed the election, but in concert they push it to the opposite of what a 100% fair election would have been.
In this scenario, note that there is no big evil conspiracy - just a handful of misguided people taking actions with a collective big consequence.
Obviously the risk of fraud has to be weighed against the risk that someone due a vote will be unable to do so because they couldn't obtain an ID or lost it. It's also worth pointing out that before the general election, the various shenanigans of the primaries and two parties mean that the general is not really "fair" before it even starts. Again, my point is that it's not "ridiculous" to posit that US elections have systemic flaws that could sway the outcome.
It still requires some cooperation, which means it's subject to game theory and defector risk. And there are numerous after the fact approaches to enforcement.
Just make a $1,000,000 reward for anyone exposing a conspiracy to rig an election, and audit the voter rolls on close elections, and impose stiff penalties for voter fraud.
These are all cheaper and probably more effective than voter ID laws. Strangely, these kinds of approaches come up far less than imposing an ID requirement. Also, mail in voter fraud is a far greater risk, yet is also ignored.
and that it is completely feasible to pretend to be another person when voting: https://youtu.be/pZMU2oMJNHU
But, since you asked, here is a scenario that I think is entirely plausible: A group of over-zealous activists travel to a very tight state with a handful of close precincts. They look up recent deaths in the areas and get a list of 1000 names and addresses that they then use to vote, each voting multiple times. They simply go in, give the name and address, and vote, then go to another polling location and repeat. Other groups decide on other tactics. A poll worker decides to stuff a box. A hacker changes a few 100 votes just to see if she can. No one technique alone would have swayed the election, but in concert they push it to the opposite of what a 100% fair election would have been.
In this scenario, note that there is no big evil conspiracy - just a handful of misguided people taking actions with a collective big consequence.
Obviously the risk of fraud has to be weighed against the risk that someone due a vote will be unable to do so because they couldn't obtain an ID or lost it. It's also worth pointing out that before the general election, the various shenanigans of the primaries and two parties mean that the general is not really "fair" before it even starts. Again, my point is that it's not "ridiculous" to posit that US elections have systemic flaws that could sway the outcome.