Seasonal Price Patterns

Originally published in April 2019’s issue of Energy & Metals Consensus Forecasts. The unabridged article and a recent sample issue are available to industry professionals. Simply send us an email at editors@consensuseconomics.com

Like many other economic variables, energy and metals prices are sometimes judged to be subject to seasonal influences. These could arise from a number of sources, including dependence on industries which are subject to seasonal swings in demand and the annual weather cycle. In order to identify such patterns (or lack thereof), we have adopted a widely-used statistical approach, known as the X12 ARIMA (Auto Regressive Integrated Moving Average) model. This statistical analysis removes the systematic calendar-related trend of an individual series and then calculates the residual seasonal effect. Examining end-month spot price data from the period January 1999 to December 2018, we calculated seasonal factors for fifteen major commodities. The results measure the average deviation of a commodity from its trend rate at the end of each month over the past 20 years.

The charts below illustrate some of the results of our X12 ARIMA study above, with the solid blue line representing the long term 20 year average for the period 1999-2018 and the dotted black line a shorter 9-year average 2010-2018. The recent but perhaps more volatile 9-year average was added following the longer-term study in an effort to remove the influence of the 2008 financial crisis on traditional patterns of seasonality. Certainly, swings in investor demand and hedge fund activity during the earlier boom to bust cycle are likely to have affected the average deviation of some commodities from their underlying trend rates. From a macro perspective, some of the obvious seasonal effects fit closely with expected patterns in both the short and long-term ARIMA results. The upward autumn pressure on Natural Gas spot prices ahead of the winter heating season in North America seems to be apparent (for both the 9- and 20-year measures), as is the climb in RBOB gasoline prices during the summer driving period. Futures prices for the latter, of course, show a similar cylical pattern of expectations, albeit not for Heating Oil, which appears relatively flat. For Nickel, both the shorter- and longer-term X12 ARIMA results suggest a peak in price strength in the month of April, which declines during the remainder of the year. As a measure of the consistency of the X12 ARIMA average results, we calculate and show below the graphs the proportion of months which have absolute rises or declines in the spot price during the 20 year analysis period.

 

 

 

 

DOWNLOAD A RECENT COPY OF ENERGY & METALS CONSENSUS FORECASTS

Receive a full sample copy of Energy & Metals Consensus Forecasts with no pressure or commitment to purchase by completing the registration form.

Your personal information will be kept strictly confidential and only used for occasional news and information we believed may be of interest to you. We do not send spam and you can opt out of our email communications by simply replying UNSUBSCRIBE in the subject line at any time.

To subscribe to Energy & Metals Consensus Forecasts click here.