Oil market observers will appreciate that the oil price has become more volatile of late with daily movements of several percent common place. A friend suggested I could look into this to see if past patterns of price volatility had any predictive powers.
Figure 1 Daily d$ for WTI since January 1988. Periods of enhanced volatility are clearly visible (click all charts to get a large readable version).
I am using the EIA oil price series. Their reporting of Brent begins mid 1987, hence my charts begin in January 1988. The starting point was to create a time series of d$ by subtracting the price one day from the day before (Figure 1). Periods of enhanced variability are clearly visible and there is a trend towards greater variability with the passage of time. This simply reflects the increase in oil price.
Figure 2 Same as Figure 1 but for Brent.
Figure 2 is included to illustrate that Brent shows pretty much the same thing as WTI. This is the only Brent chart to be presented.
Figure 3 Figure 1 data transformed to % of day before price. This has the effect of removing trends in oil price from the variance that is being plotted.
In order to normalise to a common datum the next step was to convert d$ to d$%. That is the change each day is expressed as % of the day before (Figure 3). Four periods of enhanced volatility are now clearly visible as described below.
Figure 4 The data from Figure 3 have been converted to a 21 day trailing standard deviation (SD). That means that each column represents the SD of that day + the 20 preceding days.
The final stage was to smooth the data. There are a number of ways of doing this. I elected to express the results as standard deviation (which is a standard measure used to characterise market volatility) and chose 21 days as an appropriate time period. Each point has now become 1 SD of the day and the 20 preceding days. This measure that is the SD of percentage changes in oil price is rather abstract. I have named it OPV (oil price volatility).
OPV >4 may be viewed as periods of high volatility while <2 as periods of low volatility. Four periods of high volatility stand out with a couple of lesser events. The first is linked to Iraq’s invasion of Kuwait on 2 August 1990 followed by the initiation of Operation Desert Storm on Jan 17 1991. This was the climax in volatility and the oil price settled down quickly after that.
The second period of high volatility is linked to the oil price crisis of 1998/99 when the oil price fell below $10/bbl. Notably, this volatility is not visible on the oil price trace.
A couple of lesser events in 2001/2/3 are linked to price corrections.
The third major event is the finance crash of 2008/09. And the final event is the one that is in progress today beginning in the autumn of 2014. Can these data for tell what the oil price will do next? Is the price bottom in and are we about to witness a sharp rise? Or are further falls in store? Let’s take a closer look.
Figure 5 Detail from Figure 4 plotting January 2008 to present day. See text for further explanation.
Figure 5 shows detail from Jan 2008 to present. The two orange lines mark where volatility began to rise in the lead up to the 2008 and 2014 oil price crashes. An important observation is that in both cases the oil price had already fallen substantially before volatility rose. In 2008, the quietly falling oil price, that began months before The Crash, was the prophet of doom.
In 2014, the oil price was falling quietly towards its support level at around $80. It was only when that support was broken that the market became more disordered.
In each of these cases, rising volatility foretold of further falls to come but in neither case did volatility foresee the market cycle before it began.
The blue lines mark the volatility peaks. In 2009, this marks a “double bottom” in the price. I’m not sure I’d have been comfortable making that call back then.
The volatility peaks in 2015/16 are rather more interesting and instructive. During 2015 the oil price traced out a head and shoulders pattern that I followed in multiple Vital Statistics updates. We can note that on each shoulder volatility rose while during the formation of the head it fell significantly.
My conclusion from all of the foregoing is that oil price volatility signifies market uncertainty. It is normally associated with a fall in the oil price and occurs when traders get involved in a tug-of-war, trying to guess and place bets on when the bottom is reached.
We are currently in a third near-term volatility spike (Figure 5) which means that the market is as yet undecided. So all we know is that we do not yet know which way the market is heading. The oil price chart technicals point towards recovery, while in my opinion the fundamentals point towards on-going weakness.
Figure 6 Detail from Figure 4 of the 1990 to 2004 time interval.
Taking a closer look at an earlier 1990 to 2004 period we find that volatility spikes are normally associated with lows (sometimes price corrections) in the oil price. The three volatility spikes marked on Figure 6 with blue lines were all followed by very decent Bull legs, although in 1998 there were two false dawns.
Looking at technicals is one thing, looking at what is going on in the real world is another. US rig utilisation has resumed its free fall and US production must follow the way down. But the joker in the pack is Iran. Recent OPEC + Russian diplomacy aimed at trying to get Iran to peg its production at sanctions depleted levels, is crude in the extreme. If / when increased Iranian crude reaches the market, renewed price weakness may ensue.
Regional power games in the Middle East are in play making the future uncertain hence current high oil price volatility.
Oil price volatility may provide a crystal ball to the future direction of the oil price. Normally, after periods of high volatility, when the market has been undecided, prices have risen. But I do not believe that volatility has the sensitivity to guide the timing of investments. I would judge that if the bottom is in and the oil price rises from here then this may result in a lower for longer scenario.
If, on the other hand, the price goes sub $20 before the Summer, then we will see even more severe damage to global production capacity, followed by a very steep recovery in price. A sharp rise in volatility from current levels may herald that eventuality.