Does rising oil make the market fall?

OK, now that I've gotten my oil rant out of my system, it's time to take a more analytical look at the situation. I'm going to do the same thing I did with the VIX last month - look for correlations between the market and the price of oil. Does rising oil really make the market go down? Let's see. First of all, here's a chart of crude oil futures (the CL M1) in red, for the last 300 days, 12/18/09 through 3/4/11 alongside the Dow (in blue). The Dow values have been divided by 1000 to make them scale nicely. At first glance, it sure looks like there's actually a high positive correlation between the market and the price of oil. Oil up, market up, and vice versa.

Now let's do a cross covariance of oil and the Dow. Yup, pretty much what you'd expect from looking at the raw data. There is actually a strong positive correlation in the sample period between the price of oil and the price of the market. Here's the chart. Recall that the halfway point of the x-axis represents the point in time where both charts are completely lined up.

Now let's do the same thing, but on just this past month of trading. Here are the last 20 days of the Dow vs. CL. Whoa! Almost the exact opposite. Ignoring the slope of the curve for the moment, we see that what we have here is a high negative correlation: oil up, market down, and vice-versa. So sometimes oil and the market correlate but other times they don't. It might be useful to know when this flip-flop happens.

So I ran the 20 day correlation in a loop, testing every 20 day (one trading month) period from 12/28/09 to today, and then plotted the result of each correlation, ie. how well oil and the market correlated for every 20 day period.The results are interesting. We can see that over the course of the last year, oil and the markets have pretty much moved together, though by varying degrees.

The greatest correlation came in the 20 day period beginning 4/30/10. After that, oil started to become generally less correlated with the market but at no time did it exhibit an inverse correlation (except for two dips just below zero). Until just under two months ago: January 21st of this year. After that, oil and the markets started moving in opposite directions, and the trend has been getting constantly larger ever since. Ie., the more oil goes up, the more the market goes down.

Unfortunately, this all seems to raise more questions than it answers. Why should oil suddenly reverse its correlation with the markets around the end of January? I tried poking around the news archives around that time but didn't find anything particuliarly significant. The unrest in Tunisia was already going on at that time, but it was a month later than when it all started. And why the big spike in positive correlation at the end of April 2010? Quite frankly, I have no clue.

It might be worth revisiting this to see what happens when oil and the markets go back into positive correlation. I wish I had a better conclusion than this, but it's still pretty interesting. Perhaps the main lesson here is that rising oil is never necessarily always either good or bad for the market. And whatever relationship it has is subject to change in a non-random fashion.