Looking at it in this way suggests Perot had a regional impact rather than a national impact.  For example, Perot probably did flip Montana and Nevada to Clinton due to his strength among whites, but in the Midwest, it is harder to gauge.

But, this leads to another thought and analysis.  If Perot cost Bush the election, he probably did it before election day.  The reasoning for this is fairly simple.  Horse race polling before Perot declared showed Bush leading handily.  After Perot declared, his support ebbed and shrunk.  Just looking at this period of time in the polls shows Clinton was stuck at 27 percent before the Democratic Convention when Perot dropped out.  Even after Perot came back into the race Clinton had established a double-digit lead over Bush.

This also has to be coupled with the fact candidates and their campaigns will react and adapt to the dynamics of campaigns.  What Perot likely did is make it easier for Clinton to slice and dice the electorate in ways favorable to the candidate.  Clinton, with a base in the South and not needing to worry about the swing Northeast was able to pivot from the 80s liberal Democratic mindset and moderate his views to target younger, Reagan supporting, culturally moderate voters.  Who can forget Clinton’s Sister Soulja moment?  The rebuke of the rapper by Clinton was not to win over Southern voters.  It was to ensure a slice of Republicans, Clinton was a new, moderate kind of Democrat.

Further, Perot forced Bush into a two-front war.  Instead of being able to hit the Clinton campaign, Bush had to pivot to rebuke Perot attacks.  Critically, Perot hit Bush on the deficit and allowed Clinton to argue he was the actual deficit hawk in the race (a switch from the 80’s electoral dynamic).

This, in turn, gave Republican and conservative leaning voters an excuse to vote for Clinton.  Regionally, in the Northeast, it permanently shifted the region to Democrats in combination with generational and demographic changes.

The best we can say even with the best analysis is that Perot MIGHT have cost Bush reelection.  There is empirical evidence he did not cost Bush the election, but if Perot did fundamentally reset the electoral map then the evidence does not prove he didn’t cost Bush the election.

What we can know is Perot voters were demographically similar to Clinton voters, but that is only after Perot probably had broken them away in the first place.  But, they still didn’t vote for Clinton.

It also is possible none of the above would have mattered.  Clinton could still have appeared on MTV, played the saxophone, still had a good convention and logged solid debate performances.  Most importantly from the fundamentals perspective, the economic recovery still would have been slow.

But, it is also possible, absent Perot, Bush would have been able to double down on attacking Clinton as beholden to the Dukakis’s of his party and forced Clinton into looking desperate.  Bush would have been able to early on set the narrative of the election as a steady hand vs. a liberal and been able to deflect attacks on the deficit and economy.

Of course, elections are also often decided by variables beyond a candidate’s control.  More than twenty years ago, that was an economic recession and a sluggish recovery.  Either way, Clinton won in 1992 and 1996, ushering in a new generation of leaders in electoral politics (Baby-Boomers).  In our system, a President who wins a majority in the Electoral College becomes President, and they have no less authority to lead than if third-party candidates do play spoiler.  Just ask Clinton and Trump about that.