Notes from 5/21 Meeting

We had a great discussion on the paper “Striking Oil: Another Puzzle” by Driesprong, Jacobsen, and Maat.

To recap, the central point of this paper is that stock markets take time to react to changes in oil prices in ways that cannot be explained by reasons consistent with efficient markets, such as time-varying risk premia, or transitory effects due to changes in the oil market itself. The main reasons the authors give is that it take time for information about oil markets to reliably diffuse to broader market investors; hence there is an under-reaction; this effect is supposedly also the reasons for momentum factors elsewhere in the stock market (See Hong-Harrison).

However, someone at the meeting pointed out that Ben Bernanke recently (2016) wrote how this effect might be complicated than claimed above. Specifically, there are times when oil prices and stocks are *positively* correlated rather than negatively. Bernanke writes:

“Much of this positive correlation can be explained by the tendency of stocks and oil prices to react in the same direction to common factors, including changes in aggregate demand and in overall uncertainty and risk aversion.”

In other words, if there is a reduction in predicted world economic output , there will be a lower demand for oil, which in turn will lower the price of oil and stock prices together.

This is by no means the end of the story. This effect continues to be an active area of research. For example see Chiang, who  “explores stock return predictability by exploiting the cross-section of oil futures prices. Motivated by principal component analysis, we find the curvature factor of the oil futures curve predicts monthly stock returns: a 1% per month increase in the curvature factor predicts 0.4% per month decrease in stock market index return.”

As a test to see if I could build a crude trading model, I took relatively small datasets of USO and SPY (period of 1 year) and auto-regressed them with a lag of 1-10 days. There was a negative relationship when a lag of 3-5 days was introduced similar to what the paper described, but it’s not clear how persistent this is across different periods, as Bernanke predicts.

We should revisit these topics in future blog posts.

Leave a Reply

Please log in using one of these methods to post your comment: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s