Vik Amar and Evan Caminker have posted this draft on SSRN (forthcoming, University of Illinois L. Rev.). Here is the abstract:

Pursuant to the so-called Purcell doctrine, lower federal courts (and perhaps the U.S. Supreme Court itself) are supposed to refrain from issuing remedies that would alter the rules for election administration in the runup to Election Day.  Whatever may be invoked to explain or support the instincts behind the Purcell rule (e.g., concerns about voter confusion, candidate and campaign expectations, smooth operation of election logistics, etc.), one tremendously problematic entailment of Purcell is that elections are held (and candidates are elected and policies are determined) under (often substantial) legal clouds.  Even if those clouds can be removed for future election cycles (an outcome itself not always easy under current justiciability rules), the fact remains that, because of Purcell, binding elections (with momentous consequences) take place even when serious doubt exists about the legality of the contests under federal statutes and the Constitution.  Because of these deleterious consequences (and even assuming the Purcell doctrine makes some sense or in any event is entrenched at the Court), we seek, making use of “compensation” theory, to identify ways that justiciability doctrines (e.g., mootness, first- and third-party standing, ripeness, vagueness and overbreadth) and related constraints on access to federal fora can and should be overtly modified to offset Purcell’s undesirable effects, and facilitate earlier and easier federal adjudication and remediation in election-related challenges.  It turns out the foundations of such an election exceptionalism as regards access to federal courts has been partially (if inconsistently and haphazardly) laid by the Supreme Court, but the Court has never meaningfully tried to tie the various foundation beams together in a structurally sound and coherent way, much less describe and explain what the doctrinal edifice should look like and why.  That is what we seek to do in this Article.