Oh. No? We can agree that most Stein or Johnson voters above their normal baseline would otherwise vote Democratic, yea?
No, we can't agree on that at all. Stein voters should be blue, 100%, yes. Johnson voters would normally be majority ashamed red voters (probably 80-20) who refuse to vote for the strange social policy of the right. In this election, I could see how maybe it's 50-50, at best. I've met some people who are left leaning libertarian-esque. In fact, the ideology more accurately aligns with the left. But the voter base absolutely does not.
I'm just thinking of some examples here. My roommate for the last two years was a Republican. He was a Florida Gary Johnson vote this year. My Michiganese friend has Republican parents. I've been to his house in Michigan. Met his parents. His dad is a radio show host for a conservative radio channel. He voted Trump. His wife voted Johnson, because she couldn't stomach Trump. My brother-in-law and his whole family are Republicans through and through. Their house was divided this election cycle, Trump-Johnson.
These are just some examples. Not enough to paint a picture of the whole nation. But, I can't think of anyone who was a Democrat voting Libertarian. I can't think of them because the primary purpose for voting Libertarian, should you not actually align with those views, is because you can't find a candidate on the right who is socially acceptable.
Hmm...
Michigan 2012 - 3rd party candidates accounted for 1% of vote
Michigan 2016 - 3rd party candidates accounted for 6% of vote.
Trump wins state by .3%
Wisconsin 2012 - 3rd party candidates account for 1% of vote
Wisconsin 2016 - 3rd party candidates account for 5.3% of vote
Trump wins state by 1%
Pennsylvania 2012 - 3rd party candidates account for 1.5% of vote
Pennsylvania 2016 - 3rd party candidates account for 3.6% of vote
Trump wins state by 1.2%
Ignoring the fact that Gary Johnson supporters are 50-50, at absolute best (which probably isn't even true), the bottom line is voters, Democrats in specific, failed to show up.
Wisconsin
Wisconsin Total, 2012: 3,068,000
Wisconsin dem, 2012: 1,621,000
Wisconsin 3rd party, 2012: 40,000
Wisconsin Total, 2016: 2,976,000
Wisconsin dem, 2016: 1,382,000
Wisconsin 3rd party, 2016: 100,000
Change in voter turnout: -3%
Dem voters lost: 239,000
3rd party voters gained: 60,000
3% of the Wisconsin voting age population is ~130,000 voters that didn't show up versus the last election. You're shocked at the 60,000 voters that shifted to 3rd parties. Just as a scenario, if all 100,000 voters who went third party voted Democrat, while Clinton would have won, a whopping 139,000 democrat voters are still missing from 2012. Either that, or they voted for Trump. 139,000 is over two times greater than the amount of voters who shifted for a third party. Conclusion? Far more Democrat votes simply stayed home or went Trump than the amount who opted to vote 3rd party. And again, that's the ideal case assuming that every single third party voter was drawn from the Clinton camp. Your theory that 3rd party voters shifted this state's election more non-voters is bad, and the 3rd party stockpile of votes conveniently fills the gap and is a scapegoat argument.
Michigan
Michigan Total, 2012: 4,731,000
Michigan dem, 2012: 2,565,000
Michigan 3rd party, 2012: 51,000
Michigan Total, 2016: 4,799,000
Michigan dem, 2016: 2,269,000
Michigan 3rd party, 2016: 251,000
Change in voter turnout: N/A, estimated to be roughly +/- 0% and around 63-64% overall turnout
Dem voters lost: 296,000
3rd party voters gained: 200,000
Gary Johnson wasn't even on the ballot in 2012. The Republicans gained some 160,000 votes between these two elections, whereas the Democrats lost nearly twice that. Sure, if 'some' third party voters went to the Clinton camp instead of voting for a 3rd party candidate, Clinton would have won. We again can agree on that oversimplification. But when you actually take into consideration that at most you can only split the Gary Johnson camp, to roughly 50-50 red and blue, you're left with some ~196,000 voters that stayed home or voted for Trump, while some 51,000 Jill Stein voters should be ashamed at themselves for not even being the size of the gap in this particular election. Again, conclusion here? Yes, on paper, third party voters represent the gap in volume of votes. In reality, the amount of voters who went blue in 2012 drastically decreased.
In fact, had Michigan's third party voters been distributed evenly across their respective political ideologies, and Jill Stein's voters thrown the election to Clinton... the gap would be roughly 10,000-20,000 votes. That leaves another 96,000 missing votes, or between 5 to 10 times more blue votes that stayed home or voted red. So, these voters had far more of a share in the result than third party voters.
Pennsylvania is a totally different story. I don't know why I even got so into this, but I hardly wish to re-dissect Pennsylvania. There was a surge in angry red voters, in pair with the same story as Wisconsin and Michigan. If you look into the numbers there, it's easy to say that either Dem voters stayed home, or they abandoned ship for Trump. Probably the latter.