Some oil price experts have pointed out an inverse correlation between the US dollar exchange rates against major currencies and the pattern of changes in the crude oil price. If this correlation is more than a statistical accident, it points to the following question: if the US dollar index trends downward in a steep slope over the coming months, how much upward pressure will that trend exert on the crude oil price index?
Why is this question important? Despite the strength of the opinion that international variation in developed-country interest rates and the US dollar’s safe-haven status might keep the US Dollar Index in the 94 to 99 range (or a higher one), other analysts cite an emerging realignment of trade relations and other factors that might push the Index into a lower range. Also, the chart of the Index’s values over the past year hints at the formation of a market ‘top’. Thus, given the inverse correlation cited above, the prospect of a falling dollar helping to push up the crude oil price seems to be gaining attention. Some details on the pattern of the inverse correlation are shown in Chart 1. There the blue curve represents the US dollar to euro exchange rate, the green curve represents the US dollar to yen exchange rate, and the red curve stands for index values for Crude Oil Futures. Time from May 2014 to early June 2016 runs along the horizontal axis. The left vertical axis is for the exchange rates (the US/JPY rate has been divided by 100 to facilitate charting).
A key message from this chart is that the strength of the correlation increases markedly when the exchange rates are in a distinct trend with a significant slope. Note the period from late May 2014 to late April 2015. During this period, there is a notable inverse association of the exchange rates with the Crude Oil Futures index. The chart fails to show a significant correlation beyond this period.
However, let us see what happens when we measure the multiple correlation of the two exchange rates (as predictors) with the oil price index, using a linear regression. Taking into account all the data points, the exchange rates explain (statistically, not causally) 85% of the variance in the Crude Oil Futures index. If we delete from the data set the points from 27 April 2014 to 5 May 2015 (to remove most of the surge in the US dollar) the explained variance falls to 79%.
These are impressive numbers. They suggest that in any model designed to forecast the crude oil price, the average exchange rate of the US dollar against major currencies needs to be in the model. However, before we break out the champagne over these results, we need to take into account the artificial boosting of the coefficient by the ‘time correlation’ between successive values in each of the series. Thus, it would be prudent to cut the just shown explained variance values by about one-half (say to about 40%) in making a rough-and-ready gauge of the strength of the inverse association net of the stated artificial boost.
Hence, investors need to use a complex prediction model in which an exchange rate variable is just one of the predictors, if they hope to gain satisfactory accuracy in understanding or forecasting oil price changes.
Disclaimer: This article is not designed to provide investment advice. Please use this article only as a source on information in setting up your own due diligence about investing. I have tried hard to avoid introducing inaccuracies in the data and inappropriate interpretations; however, I offer no guarantees of accuracy or fitness for use of materials in this article. Finally, to statistically oriented analysts who are unaware of the gold mine of data made public by Fusion Media, I recommend that you start ‘fishing around’ by going to the addresses shown in the footnote to Chart 1.