Originally posted at PredictWise.com

… The markets think Trump will win anyway.

The Cruz and Kasich campaigns announced yesterday that they will collaborate to try to lower Trump’s odds at reaching a majority of pledged delegates and keep their respective campaigns alive.

The deal makes sense. While Trump is in an enviable position, he cannot afford to lose very many of the remaining delegates if he wants to become the nominee. The deal signals to Cruz and Kasich voters to start voting strategically: vote Cruz in Indiana, and Kasich in Oregon and New Mexico–all to deny Trump delegates and the nomination.

What could go wrong?

Despite making sense and being worth a try, the Cruz-Kasich deal is unlikely to work for two main reasons: it’s unlikely to meaningfully shift public opinion, and it helps Trump’s pitch to unbound delegates in the event he needs them.

CRUZ IS UNLIKELY TO GAIN MANY VOTES FROM DEAL

To have any meaningful effect, the Cruz-Kasich deal needs to convince enough Kasich supporters in Indiana to switch over and vote Cruz. Oregon and New Mexico are strictly proportional, so strategic voting in those states will do nothing to lower Trump’s delegate haul. So it all comes down to Indiana (and maybe California too, since new reports suggest Cruz and Kasich may do a similar deal there).

But there are three obstacles to getting voters to vote strategically:

1) Coordination Problem: For this deal to work, it will require careful coordination between the Cruz and Kasich campaigns. To really move the needle, Kasich should probably tell his supporters in Indiana to vote Cruz. Instead, Kasich has already told his supporters in Indiana to vote for him–and not Cruz. This undermines the entire point of the deal.

2) Backfire: A new addition to Trump’s campaign message over the past couple weeks is that the system is rigged against him. Trump has been spinning Cruz’s successful delegate maneuvering at county and state conventions to argue the party is trying to steal the nomination away from him. Nate Silver was already arguing that this messaging was causing Trump’s polling numbers to rise, and Cruz and Kasich colluding to stop Trump can only reinforce this narrative.

Indeed, Dave Weigel shared today that Kasich canvassers in Indiana talked to some Kasich supporters who heard about the deal and switched their vote–to Trump.

While this is merely an anecdote, Trump and Kasich do typically compete over self-identified moderate voters. If those voters get the message that Kasich isn’t competing in Indiana, these moderate voters may just as well cross over to the guy who wins the moderate vote: Trump.

This is the worst case scenario for the #NeverTrump movement, but a neutral effect may just be Kasich voters deciding to stay home–increasing both Trump and Cruz’s vote share equally, and therefore making a Trump win in Indiana look even more impressive.

3) Not many voters will hear the cue: While the Cruz-Kasich deal was exciting on political Twitter and in the media, most voters are unlikely to hear and internalize the news.

You can think of these three obstacles this way:

Kasich currently has about 20% of Indiana GOP primary voters in his corner. Of those 20%, only a fraction will hear and internalize the message of the deal. Of those that do, some will stay home, some will vote Kasich anyway, some will vote Trump, and some will vote Cruz. Because of those obstacles, it’s unclear that Cruz will net a meaningful number of voters in Indiana.

After initially dipping the markets, Trump’s probability of winning Indiana is almost back to where he was prior to the deal.

But let’s assume it works out. Let’s assume Cruz does gain a significant number of former Kasich voters, who end up lifting Cruz above Trump in Indiana. And let’s even assume this loss in Indiana ends up being critical in preventing Trump from reaching 1,237 delegates.

TRUMP’S CASE TO UNBOUND DELEGATES MAY BECOME AN EASIER SELL

In an earlier post, I argued that 1,237 pledged delegates is less of a magic number than it seems. The reasoning was simple: there will be 150+ unbound delegates that Trump can and will try to convince vote for him on the first ballot. Many of them from Pennsylvania, a state Trump is expected to win convincingly. And, Trump will have six weeks after California and before Cleveland to convince these unbound delegates. Could this deal help his case?

If this deal shows that a meaningful number of voters do vote strategically if given the cue (a big if), and if Cruz honors the deal and doesn’t try to win Oregon and New Mexico, then Trump’s odds at winning go up in each state.

Cruz is Trump’s biggest challenge in both states. According to Nate Cohn’s demographics model, Cruz should be expected to beat Trump by about 2.5 points in Oregon. According to a model that uses Google Correlate data, Cruz is +6 against Trump. For New Mexico, Trump is up +4 and +1 according to each model.

Kasich, meanwhile, trails Trump between 12-14 points in Oregon. It’s even worse in New Mexico, where Kasich trails between 22 and 28 points. With such big deficits, it’s possible that strategic voting by Cruz voters could end up elevating Kasich to 2nd place…and Trump to 1st.

Why does this matter to unbound delegates? If Trump is a few delegates short of 1,237, his case will be that he won the most votes, the most delegates, the most states, and leads in the national polls, and that the only reason he isn’t the presumptive nominee is because the party is trying to steal it from him. Trump winning Oregon and New Mexico (plus California) in a two-against-one contest will only strengthen his argument that he’s the true winner.