Modeling the price of crude oil and motor fuel: a five-year revision

I've published a new research paper as based on many posts in this blog. This is a five-year revision of our model predicting oil and energy price. The paper is available on RePEc:

Modeling the price of crude oil and motor fuel: a five-year revision

We present a five-year revision of an empirical study started in 2007. Seven years ago, we found two three distinct periods characterized by sustainable linear trends in the difference between the headline consumer price index (CPI) and the core CPI in the USA. Then we revealed similar behavior in the differences between the CPI and indices of various consumer expenditure categories. We estimated the duration of these trends which varies in a wide range from 5 years to more than 20 years. The transition periods to new trends span shorter intervals of 2 to 5 years. The transition is characterized by a higher level of volatility in the studied CPI differences. In April 2009, we introduced a simple quantitative model representing the evolution of motor fuel price (a subcategory of the consumer price index of transportation) relative to the core CPI as a linear function of time. Under our framework, all price deviations from this linear trend are transient and the price must return to the sustainable trend. The model predicted that oil price would fall to $30-$60 per barrel in 2016, which is very close to the current price. The behavior of actual oil and motor fuel price since 2010 has shown that this prediction is accurate in both amplitude and trajectory shape – a good support for the credibility of our empirical mode. We conclude that the concept of price decomposition into a short-term (oscillating) and long-term (linear trend) components deserves a deeper theoretical consideration of the driving forces behind linear time trends and can be used as a workhorse for a wide spectrum of commodity investors. According to the model, the price of crude oil will be falling to the level of $30 per barrel during the next 6 years and motor fuel will follow up the oil price. Moreover, the periodicity of the related normalized difference indicates that this low-price level may extend into the second half of the 2020s. The secular fall in energy prices may induce a lengthy period of very low inflation.


The era of low energy price will last 10 years with oil at $20

We have been studying the long-term evolution of energy prices in the USA since 2007 and reported several important observations useful for profitable and safe investments. The simplest fact we have discussed is the effect of energy/oil price on stock prices. Share price of Exxon Mobil and ConocoPhillips does depend on oil price. The link is direct and positive – oil price increase is reflected in share price growth. We have calculated regression coefficients in order to understand which company is most sensitive.
In 2008, we reported that several major categories of consumer prices (various goods and services) have sustainable and quasi-linear in time trends relative to energy prices. Similar behavior was reported for producer prices and oil. One can easily imagine what a great opportunity arises for sound investments. Moreover, our observations demonstrate that many of these sustainable trends have clear turning points which provide investors with invaluable information on effective buy/sell decision. In this article, we revisit our previous results related to energy (oil) in order to demonstrate that the fall in the energy price in 2014 was well predicted. Then we extend our investigation of the future evolution of the consumer price index (CPI) of energy with new data obtained since the beginning of 2013 when the fall prediction was done.  Currently, one is interested to know how deep the price will fall and when it will come back to $100.
Our theoretical approach was first published more than six years ago in a paper on the presence of long-term sustainable trends in the differences between various components of the CPI in the USA. We started with the difference between the core CPI (i.e., the headline CPI less food and energy) and the CPI. Then the consumer price index of energy, which gives approximately 9% of the CPI, was analyzed. In the beginning of 2008, we successfully identified and predicted that difference between the CPI and the energy index was approaching a turning point (actually observed in the summer of 2008) and forecasted the energy prices to fall relative to the core CPI through the first half of the 2010s. Based on the general approach, we also estimated that after the turning point in 2008/2009 oil price would go down from $100 to $50 per barrel in 2015 and $30-60 in 2016.  As one may know this prediction of the turning point timing and the level of oil price and the duration of higher oil prices is relatively accurate. 

Here we revisit the relative evolution of the core consumer price index (CPI) and the CPI of energy. Figure 1 displays the difference between the core CPI and the index for energy for the period between 1960 and 2014. All CPIs are seasonally adjusted and borrowed from the BLS.  Before 1980, these two indices had been growing almost in sync with fluctuation around 10 units of price index. Between 1981 and 1999, the difference grew from -10 to almost 80 units. Between 2001 and 2008, a period of intensive growth in the energy index was observed. Qualitatively, one can distinguish three periods of linear trend and three turning periods with a higher volatility. The last turning point was in 2008 and the index of energy is now on decline relative to the core CPI.  However, the extremely high volatility masked the new trend in the difference before 2011. Currently, the new trend is clear and shows the expected behavior - energy price goes down.

The question is - when will it reach its bottom? And when the price will reach the hard bottom?
Six years ago we expected that oil price might go down to $22. Is this the case now?

Figure 1. The difference between the core CPI and the index for energy between 1960 and 2014. There are three periods of linear trend and three turning periods. The most recent turning point was in 2008 and the new trend emerged more of less clearly in 2011.

Figure 2 illustrates the most recent period. Since 2001, we observe a slow decrease in energy price relative to the core CPI. Their difference in Figure 2 is below the green line representing the predicted trend for the period after 2010 - the slope is as between 2001 and 2008 but with an opposite sign. Since summer 2014, the energy index has been falling much faster than before and the (green line) trend was intercepted in November 2014. Linear regression gives a slope of +6 for the difference curve since 2011 which is still much lower that the absolute value of the slope between 2001 and 2008 (-14). The energy price had extremely high volatility between 2005 and 2011. Then the price calmed down and demonstrated only short term and low amplitude fluctuations. The level of volatility is again high since summer 2015. One might expect a few years of larger volatility from deep falls to sky heights.
Data between 2001 and 2014 give clear indication of the new trend direction and slope. Following this trend, the difference of core CPI and consumer price index of energy will likely reach its peak value in 2016-2018. We expect no further decrease in oil price beyond 2018. The question is when the trend will change to the opposite like that was observed in 2001 and 2010? Historic evidences are unbreakable – cycles rule this world. There are many explanations of the current energy price behavior. Various actors are suggested as drivers with own motives open and hidden. We do not consider them as wise and explanatory. To say “cycle” does not mean to explain something as well. This is just a set of observation in favor of future evolution.

Figure 2. Same as in Figure 1, for the period after 2002. Linear trends are shown.

We used only absolute difference so far. It is instructive to analyze the difference in relative terms and we have normalized the difference to the core CPI. Figure 3 illustrates the new pattern. In contrast to Figure 1, the amplitudes and periods of long term fluctuations are similar and the overall evolution seems to be repeatable, i.e. cyclic. Figure 4 exercises the assumption of repeatability. We have shifted the original curve by 27 years ahead and obtained a striking similarity in the amplitude and timing of the energy price falls and rises.

Figure 3. The difference between the core and energy CPIs normalized to the core CPI.

Figure 4. Same as in Figure 3 with red curve representing the original (black) curve shifted 27 years ahead
Finally, Figure 5 shows a detailed picture which could be helpful in the bottom price estimation. The current difference presented by black curve is expected to reach red line in during this phase of energy price fall. The intercept price is about $20 per barrel. Two years ago we foresaw a cliff in energy price and it is here today. When the black line touch the red one - the energy index returns to the long term sustainable trend stretching into the 2020s. The era of low energy prices has come and the trend will not change to the opposite before, say, 2025.    

Figure 5. An energy cliff is not over.

We have three general conclusions:
-        The energy price will be falling another two years with the bottom of $20.
-        The change in energy price will be characterized by an elevated level of volatility in the next two to three years.

-        The era of low energy price will extend into the 2020s.



nice fantasy. The next movie will bring a brave woman to the previous universe through the singularity (before the Big Bang) in order to save the whole banch of universes before and after us. 


Real GDP in US - mediocre growth /slow down. As in the next ten years.

The Bureau of Economic Analysis reports that real GDP in the USA has been growing at almost the same rate since 2010 - around 2.2 % per year. In 2014, the rate of growth was 2.4% per year slightly above the 2013 value of 2.2% per year. Figure 1 illustrates the history (recession in 2008 and 2009). By all means this is a mediocre growth rate. It's even worse when real GDP per capita is considered - the growth in 2014 was  smaller than that in 2013 - 1.8% vs 2.4% per year, respectively. The economy is slowing down. Some optimists expected healthy growth.  Not in the next ten years.

Figure. Real GDP growth rate % per year


word is lethal weapon

When considering freedom of speech one should remember that word is recognized (in common sense as well as in criminal legislation) lethal weapon. Words in various forms stay behind almost all actions - positive and negative. So, be careful when supporting all speeches. 

Слово - смертельное оружие

Понимание свободы слова порой расширяется до невозможных пределов вседозволенности, хотя всем известно, что слово опаснейшее оружие, которое может убить непосредственно  (ст. 110 УК РФ) или опосредовано - через пропаганду религиозной, национальной, родовой исключительности. Оскорбление словом, то есть "унижение чести и достоинства другого лица, выраженное в неприличной форме"  (ст. 130), также карается законом. Может так и надо бороться с проявлениями "свободы слова"?


global wars on market shares

Oil prices have been falling since the last summer. Saudis make it clear that they have deep interests in retaining (increasing) their market share as long as possible. Reasonable explanation of the current market strategy. Timing is perfect and shale/tar/renewables over invested during the last decade seem having no chances at a one-two year horizon. European countries with windmills producing more than 25% of energy will be interesting case to observe.    
However, this is just the first step into the global war for market shares. Among many negative consequences, falling oil prices can make a few good things for oil producing countries. Transportation becomes cheaper. For example, Russian coal gets a $40 advantage and becomes more and more attractive for many consumers. Moreover, dramatic rouble  weakening adds  competitive power to Russian export while import of oil-related machines from the Netherlands and Germany becomes irrelevant because projects of oil extraction in Arctic and other hard conditions must be delayed or cancelled (same in Canada). Germany, France and the Netherlands will soon meet export fall.  Other commodities, goods/services, countries, regions, the world itself and its economy will be subject to changes through usual mechanisms of trade, cold and hot wars.  So, interesting and dangerous time is coming.