So, I got to thinking about the polls around the presidential election. The first thought that came to my mind is the polls are bunk. If you look at Real Clear Politics you can see different polls that each have 1,000 plus sample sizes, that are taken right around the same time frame, where the averages, plus or minus the error margin, do not overlap. This is ridiculous, a sample of 1,000 plus people of should give the average of a population of 122 million (the voting public) 95% of the time, inside the margin of error.
Then I had another thought. I wonder if the market is correlated to the expectations of the election? It should. The two candidates have very different ideas on regulation, free markets, corporate taxes, and various other factors that influence the market. So, the market should care about the outcome of the election, and the prices should reflect expectations of the election.
So, here’s what I did, I took the daily average of the Real Clear Politics average and the closing price of the Dow since August 1st. I then took the change in the poll average and the change in the Dow. I then checked, on a one day lag, if the market reacted to the movements of the poll average. I used a one day lag because I assume that the traders, analysts, and other market participants generally digest their political information after the NYSE and the NASDAQ close. One other consideration that I made was this: the change in poll numbers over the weekend is a Monday minus Friday delta, just as the change in the Dow close is a Monday minus Friday delta. I’ve also looked at a two day lag, a three day lag, a average of the next three days’ delta, and a four day delta average.
What I found was this: on the days John McCain closed the gap in the polls, the following day the stock market would climb. On the days when Barack Obama gained in the poll average, the market would go down. The numbers are this: average gain for McCain in the polls 0.74% for an average gain of 1.86 points in the Dow. For Obama, his gains in the polls were an average of 0.67% for an average drop in the Dow of 66 points. A two day lag reverses the effect, and a three day lag re-reverses the effect. The average delta for three and four days shows gains by both candidates have a negative effect on the Dow; however, an Obama gain has at least twice the negative effect that a McCain gain has.
This study is not scientific, and there was little if any effort to try to control for outside factors. I did this little exercise because I was curious and I am a huge dork. Still, the data is what the data is. (The data are what the data are, I suppose) And, if one only takes th
e lead Obama has in the RCP Average and runs a correlation of it on the Dow Price (not the movement in price, just the price) for that time frame, one will find it is negative correlated. I also ran a covariance and found the Obama’s lead has a negative covariance to the Dow Price.
I made a pretty graph of the One Day Lag scenario. Yay for me.
Later,
B
Then I had another thought. I wonder if the market is correlated to the expectations of the election? It should. The two candidates have very different ideas on regulation, free markets, corporate taxes, and various other factors that influence the market. So, the market should care about the outcome of the election, and the prices should reflect expectations of the election.
So, here’s what I did, I took the daily average of the Real Clear Politics average and the closing price of the Dow since August 1st. I then took the change in the poll average and the change in the Dow. I then checked, on a one day lag, if the market reacted to the movements of the poll average. I used a one day lag because I assume that the traders, analysts, and other market participants generally digest their political information after the NYSE and the NASDAQ close. One other consideration that I made was this: the change in poll numbers over the weekend is a Monday minus Friday delta, just as the change in the Dow close is a Monday minus Friday delta. I’ve also looked at a two day lag, a three day lag, a average of the next three days’ delta, and a four day delta average.
What I found was this: on the days John McCain closed the gap in the polls, the following day the stock market would climb. On the days when Barack Obama gained in the poll average, the market would go down. The numbers are this: average gain for McCain in the polls 0.74% for an average gain of 1.86 points in the Dow. For Obama, his gains in the polls were an average of 0.67% for an average drop in the Dow of 66 points. A two day lag reverses the effect, and a three day lag re-reverses the effect. The average delta for three and four days shows gains by both candidates have a negative effect on the Dow; however, an Obama gain has at least twice the negative effect that a McCain gain has.
This study is not scientific, and there was little if any effort to try to control for outside factors. I did this little exercise because I was curious and I am a huge dork. Still, the data is what the data is. (The data are what the data are, I suppose) And, if one only takes th
e lead Obama has in the RCP Average and runs a correlation of it on the Dow Price (not the movement in price, just the price) for that time frame, one will find it is negative correlated. I also ran a covariance and found the Obama’s lead has a negative covariance to the Dow Price.I made a pretty graph of the One Day Lag scenario. Yay for me.
Later,
B
No comments:
Post a Comment