Was Perot A Spoiler Candidate?

Since I have little better to do, I've been trying to figure out if Perot actually cost Bush Sr. the 1992 election. So first things first, if Perot had dropped out of the election, how many of his voters would have to vote for Bush to give him the win? Doing some quick math, I found that 66% of Perot voters would have to vote for Bush to allow him to win. So now I have to find if more or less than 66% of Perot voters preferred Bush to Clinton.

To determine this next step, I compared the 1992 and 1996 election. Assuming that there were no changes in preferences, what affect did the decline in votes for Perot have? To get the 1996 result, 67% of Perot voters would have to have voted for Clinton. Closer inspection of this number showed some flaws though. In several states, Dole received fewer votes than Bush despite the 33% of former Perot votes he should have gained, meaning that Clinton gained former Bush voters. Also, such a split would have given Clinton an implausible lead in the 1992 election if Perot had dropped out.

My next idea was to do some complicated and questionable mathematics to determine the split in Perot's lost votes that best fit the 1996 election result. This showed that only 47% of former Perot voters chose Clinton in 1996. Still, that's enough in favor for Clinton to have kept Bush from winning in 1992. However, this split only explains the 1996 election. Perot voters may have split differently in 1992.

Trying to find other studies of the 1992 election, I found two things: First, polls asking for Perot voters second preference generally showed them evenly split between Bush and Clinton. Second, polls before the election generally showed large changes in support for Clinton and Perot, but Bush support remained fairly constant.

So my conclusion, it is unlikely that Perot acted as a spoiler candidate. He probably cost Bush a few states, but not enough to change the result.