Sunday, January 11, 2015

Gas price hysteresis, Spain 2014

In 2008 we did a hysteresis analysis to try to understand how variations in the price of oil translate to retail prices of gasoline and gasoil (diesel) in Spain. As crude oil prices have plummeted in the last few months, spurring some interest in whether consumers have benefited from this as much as they should, we redo here the analysis with 2014 data:
The figure shows the weekly evolution during 2014 of prices of Brent oil and average retail prices without taxes of 95 octane gas and gasoil in Spain, all in c€ per liter.
Below is the scatter plot of Δ(gasoline price before taxes) against Δ(Brent price),
with linear regressions for the semiplanes Δ(Brent price) ≥ 0 and ≤ 0, given by
ΔBrent ≥ 0 → y = f+(x) = b+ + m+x = 0.0734 − 0.3685x,
ΔBrent ≤ 0 → y = f(x) = b + mx = 0.2338 + 0.5447x,
respectively (shown in the plot as dotted red lines). This is entirely unexpected: retail gas prices tend to reduce even for positive variations of oil. The overall regression line for the scatter plot (dotted black line) is
y = f(x) = b + mx = −0.2489 + 0.4208x.
As for gasoil, we have
and
ΔBrent ≥ 0 → y = f+(x) = b+ + m+x = −0.1965 − 0.1115x,
ΔBrent ≤ 0 → y = f(x) = b + mx = 0.16 + 0.4498x,
 overall → y = f(x) = b + mx = −0.2169 + 0.5039x,
which is qualitatively similar to the behavior for gasoline. In both cases, retail prices under constant oil prices (that is, when ΔBrent is small) have reduced around 0.2c€/l per week.
So, oil's fall during the second half of 2014 has of course resulted in a general reduction of gas prices, but these were are also showing a correction trend during the first half of the year, when oil rises were translated less strongly than drops. At least for 2014, we haven't found evidence of the "rocket and feather" effect fuel companies are usually suspected of.