Quick follow up. I got the following mail from the Economist Intelligence Unit, who are you know, business analysts who are paid to be right, not ideologically correct (emphasis added in third paragraph): [quote]The Economist Intelligence Unit expects Hillary Clinton to win the 2016 presidential election, but this may prove to be a poisoned chalice. A one-term administration is likely as the new president struggles with high unfavourability ratings, a divided Congress and a predicted mild recession as the business cycle comes to an end. These are the key findings from our latest free report available to download below, Election 2016: The unpopularity contest. In the report we forecast that in November the Democrats will win with a majority in the Senate, but the Republicans will retain control of the House of Representatives. Nonetheless, the unpopularity of both candidates means that the next president will start from a position of weakness and a divided Congress will make passing key pillars of their agenda even more difficult. With possible concessions on immigration policy and the toning down of nativist rhetoric, a Republican victory is therefore likely in 2020. [b]We believe that Mr Trump's plans for a huge wall on the southern border with Mexico and mass deportations of millions of undocumented workers are completely unrealistic.[/b] Ms Clinton will be successful in promoting clean energy and healthcare reforms, but will struggle to tighten gun control laws. Whoever wins, filling the vacancy in the Supreme Court will be the most consequential act by the new president.[/quote]

As the Atlantic says, it seems to be a vicious cycle: people are frustrated so they vote for ideologues (Tea Party, Sanders) whose ideas are unworkable and fail, making people even more frustrated, so they vote for more extreme ideologues...