EV calculation is:
+ 40% (day 1 town lynch) × 25% (town election given day 1 town lynch)
+ 30% (day 2 town lynch) × 33% (town election given day 2 town lynch)
+ 20% (day 3 town lynch) × 50% (town election given day 3 town lynch)
+ 10% (day 4 town lynch) × 100% (town election given day 4 town lynch)
= 10% + 10% + 10% + 10% = 40%.
There's actually a very simple pattern here: the setup's mathematically equivalent to making it a normal Nightless (except with dead players voting) and town always having to lynch all but one Elect (because electing one player is equivalent to lynching everyone else); not only is there a 1:1 correspondence between how the games end, town also have the same amount of information as of each lynch. As such, all that actually matters is the alignment of the player who's lynched last, and thus the EV is purely based on the town:scum ratio among the Elects (in this case, the ratio is 40% town, so the EV is 40%).
It'd be interesting to see how this compares psychologically to the "2 Innocent Child, 2 VT, 3 Goon, Nightless, dead players can vote" setup. In that setup, the "default" would be to try to lynch scum repeatedly every lylo, as opposed to picking a townie and lynching everyone else. In this setup, the elect is likely to be determined via townhunting rather than by eliminating all the scum. There's no difference, though, because choosing wrong means the game is over, and thus you don't get any information when you choose right (i.e. you could line up your future lynches pre-flip without any issues, because the flip doesn't help and scum don't have any night choices that you could be helping them with).
