4. Consider again the basic Hoteling-Downs model with two non-ideological vote share maximizing

candidates. Again the ideological space is [0,1] and there is a continuum of voters uniformly distributed along it, with the total number normalized to 1. We typically assume that both candidates choose their policy platform at exactly the same time. However, this may not be that realistic.

Sometimes, one candidate or party may release their party manifesto before the other one. So let's suppose the game is sequential: First candidate 1 chooses a platform, then candidate 2.

(i) Let's work backwards. In the second stage, candidate 2 observes candidate 1's choice before

making a decision. Given candidate 1 chooses platform x, what is the best choice of platform for candidate 2?

(ii) Candidate 1 moves first and therefore does not know candidate 2's choice when deciding.

However, candidate 1 expects candidate 2 to behave rationally and try to win the election.

What platform is the best choice for candidate 1?

(iii) Is it better to be candidate 1 or candidate 2? What does this imply for the `principle of

minimum didifferentiation'?