Commodity sensitivity has recently incited increased concern in the MLP space, notably the subset that comprises midstream oil and gas companies. The potential for this factor to become seriously weighty in driving MLP unit/share price changes has always existed. However, a long-standing statistical disconnect between price swings in a major crude oil indicator such as USO (NYSEARCA:USO) and those in an index of the midstream MLPs (NYSEARCA:AMLP) has led a many investors, including me, to give the issue a low profile. What a shock, then, to see the recent emergence of a relatively high correlation between the two indices!
Chart 1 shows that this high correlation is a development of the last few months, at least if we go back no earlier than the days when the major decline in crude oil prices began during 2014. This chart compares the time patterns of weekly percentage change in closing prices between USO and AMLP, from early January 2015 to late January 2016.
A key take-away from this chart is that since late summer 2015, the swings in AMLP began to follow quite closely those in the crude oil index . This is a distinct contrast to the earlier period covered by Chart 1. This take-away from the chart is not news. Experts writing in Speaking Alpha have already discussed it. (See here and here.) What this article contributes is a concrete statistical documentation of the new correlation cited above.
Less well covered in the existing literature, and perhaps as important for the strategy shifts that we all might contemplate, is the matter of change in the pattern of volatility within the MLP space. (Click here for a related discussion.) This too is recent and quite striking, as Chart 2 shows.
This chart illustrates a new statistical measure that I call the "high probability trading band". Generally, it is a band within which almost all of the trading took place during a given time period, such as a week or a month. At its center is the modal price, the price at which you are most likely to have made a trade (in random sampling) during that period. (Of course, 100% of trading took place between the period's High and Low prices; but often tiny fractions of trading happen near these prices.)
In Chart 2, the band is the vertical distance between two corresponding curves at any given point on the horizontal axis of the chart. For example, it is the vertical distance between the two red curves (at a given point on the horizontal axis), or the same distance between the two blue curves. The red curves represent AMLP, and the blue curves pertain to USO. Each curve thus contains many bands. The chart gives you the high probability trading bands from early January 2015 to late January 2016.
Each point on a curve in Chart 2 summarizes the situation for an entire month of trading. I drew upon Yahoo Finance's tables of daily Open, Close, High and Low to compute the modal price for each month. The computation takes daily volumes into account. A trading band is set by computing the distance from (1) the mode plus two times the volume-weighted mean deviation to (2) the mode minus two times the volume-weighted mean deviation. In most cases at least 90% trading would have taken place within that band, which is more representative of a day's action than the distance from High to Low.
Chart 2 shows that the time path of the trading bands for USO has become more narrow in the past few weeks, as we approach late January 2016. Generally speaking, the bands for USO have become less erratic in their month-to-month swings compared to earlier in 2015.
The situation is quite different for AMLP; because there is a period of relative tranquility up to late summer of 2015. Then, in a sense, all hell has broken loose; because the bands have widened substantially, and we are now in the presence of enhanced volatility among the AMLP bands.
Let us turn to some key strategy-related questions suggested that by Charts 1 and 2.
How is your particular MLP positioned on this issue of commodity sensitivity, if you already have units/shares in an MLP? If you are planning to buy some units, where should you go if you wish to avoid great exposure to commodity sensitivity?
Various authors have already written about the importance of this question. Most of the existing literature tells the reader in qualitative terms about the high or low sensitivity of particular MLPs. It would also be helpful to provide statistical data that measure this relative sensitivity, so as to show distances between different pairs of MLPs regarding commodity sensitivity.
Here is the second strategy-related question. If a high correlation between the MLP Index and crude oil prices is relatively new, what explains this development? Some authors have put forward the hypothesis that investors have suddenly awakened to the prospect that oil prices are so low that distributable cash flows among MLP companies are now under unusual threat. While this looks like a possibly key explanatory variable, it would be a good idea to consider other relevant variables.
One example is the matter of who are the dominant traders in the two camps. It is now widely reported that the dominant traders in USO are very large institutions, including particularly big banks. If this is true, then we know already that the dominant traders are in fact the "robots" owned by those institutions (by "robots" I mean various forms of computerized algorithmic trading). (How the robots tend to behave has been the subject of many articles. See, e.g., relevant chapters in Dark Pools by Scott Patterson, Random House 2013).
A key thing to remember about these robots is that their tolerance of high losses in their trading battles may be much greater than that of human beings. Thus, they are in a position to push the price of USO, for example, down to what might have been previously unimaginable levels. (For a discussion on the implications for traditional market valuation, especially the possibly reduced utility of robot-driven market prices to become a factor in capital allocation in the real economy, see this article at marketstatsanalytics.com. )
In contrast, in the MLP space it has been widely thought that trading was dominated by retail traders. However, there exist reports that institutional buying and selling in the MLP space is happening increasingly. It would be good to have more information on this development, since to the extent that institutional money comes into the MLP space, the robots will not be far behind.
My point here is that if we are seeking to explain the new high correlation between price swings in USO and those in AMLP, we should encourage brainstorming discussions about potentially relevant variables. Such discussions could be the source of useful insights in connection with our third strategy-related question.
What should we assume concerning the longevity of the new high correlation cited above? The answer to this question will influence how we position (or re-position) ourselves as buyers or holders of MLP units. No one knows the future, course; but we will need to do some kind of forecasting on the question of how long the situation will last.
To the extent that there is a growing presence, and perhaps eventual dominance, of institutional traders in the MLP space, this will tend to give the high correlation a long life. The holders of data on trends in institutional incursion into buying/selling stock in the MLP space could provide us with a useful service here. In the meantime, we might find it helpful to keep a close watch on price patterns of the sort provided in Charts 1 and 2, so as to get rough and indirect indications as to whether the said correlation is fading away or is intensifying.
This article has been provided for information purposes only. I have done my best to avoid adding to any errors that may be in the raw data, whose accuracy I cannot guarantee. Apologies in advance for my own errors. Also, I am not an investment advisor. I do hope that this article will be useful as a part of a process of arming yourself to deal with the enhanced levels of volatility in the New Stock Market.
Disclosure: I am/we are long USO.
I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.