Are the Polls Valid?

The latest polls show Clinton and Trump pretty much tied. The polls seem reliable (meaning they seem to ratify each other), but how valid are they? Validity has to so with the extent to which the poll accurately measures what it purports to measure. 

The bookie odds also seem to predict a Clinton victory. In fact, I have heard of a shop in the UK that has already started paying out the bets on the basis of a Clinton victory. But how valid is that indicator?

Polls are done using samples, and samples are selected, as I understand, on the basis of the electorates for the previous two elections. For example, in 2012, several million evangelicals abstained from voting because they didn’t want to vote for a Mormon (Romney), so now evangelicals (usually voting Republican) are not the same proportion of the samples as they are of the electorate. Conversely, African Americans (mostly voting Democrat) turned out in high numbers for the last two elections, so they are disproportionately represented in the samples. This means that samples are stacked in favor of Clinton. Polls are using samples that have been selected on the basis that Democrats will be more supportive of their candidate than Republicans will be of theirs on November 8.  In other words, poll samples include a higher percentage of Democrats than the electorate includes.

Gambling odds can be contrived also. They reflect the bets placed. I am told that the big majority of bets are for Trump to win, but the big dollars are on Clinton.  Dollars determine the odds, not number of bets, whereas in an election the number of votes are the determinant. 

Clinton and Trump tied in the polls probably means a Trump win.  Let’s hope that he surrounds himself with capable advisors to whom he is willing to listen. Perhaps in the next day or two I’ll comment on why I think some of his policies are problems rather than solutions.

I think that the markets understand that a Trump victory is problematic and that markets are priced now for a Clinton victory. If she loses, I expect the stock market to take a big hit…..not just in weeks, but in hours.  An initial drop like that could cascade and lead eventually to a real estate collapse. They often go hand in hand.  For example, the 1837 stock market crash was followed by the Chicago real estate of 1841.  The Florida real estate crash of 1926 was followed by the stock market crash of 1929. In Canada, in living memory, we saw a real estate crash of the early-mid 1980s followed by the 1987 stock market wipe-out. History does have patterns. That Florida beachfront property you’ve been wanting may be a lot cheaper in another year or two.