My Daughter

My Daughter
Remember when you learned how to do this?

Monday, June 30, 2008

Weekly US Petroleum and NG Supply Reports

Unleaded Prediction 9-May
Beginning Inv mbbl 211.9
Imports Wk/Day 9.8 1.4
Production Wk/Day 61.6 8.8
Available 283.3
Balance Wk/Day 72.1 10.3
Ending Inv Mbbl 211.20
Prod Supplied 9.3
Predicted Change -0.7

Distillates Prediction 9-May
Beginning Inv mbbl 105.7
Imports Wk/Day 1.4 0.2
Production Wk/Day 29.4 4.2
Available 136.5
Balance Wk/Day 30.8 4.4
Ending Inv Mbbl 105.7
Prod Supplied 4.3
Predicted Change 0.0

Crude Oil Prediction 9-May
Beginning Inventory 325.6
Domestic Prod 35.679 5.097
Imports 72.8 10.4
Total Available 434.079
Provided to Refineries 104.3 14.9
Ending Inventory 329.779
Predicted Change 4.179
Ref Utilization 86

I am not smart enough to predict that this week will be any different from last week. The last couple of weeks we have seen a flood of imports in both crude oil and unleaded (not distillates) and lower than normal refinery utilization, because of the high crude oil prices and relatively low refinery margins.

The unleaded inventory will be especially interesting. We should be in the time of year when this inventory is built up to take care of summer demand, and the numbers themselves suggest that there is plenty of gas around, but even with fairly strong imports of 1.4 mbpd, the demand lately has been such that with the refinery system running the way it was last week, at about 85% capacity, we will still see a slight draw down in unleaded inventory. The longer this goes on, the more serious it is going to get.

In crude oil, I have assumed above the same low inputs to refineries, high imports (despite the pricing) and the same domestic production we always have, and the result is a pretty strong build.

In distillates, we saw a little bump in demand last week, with spring planting just underway, and very low net imports, and if that same thing happens this week, with typical production levels of 4.2 or so, we will see this inventory break just about even.
Closer to Reality

pup55 28
Analysts 26
Tie 0
Avg 0.518518519

Directional Correctness
pup55 36
analysts 41
ttest 0.028524

Sum of Weekly Predictions
pup55 Analysts Actual
Unleaded 12.187 -0.125 4.1
Distilla -19.537 -13.820 -21.5
Crude 34.839 32.660 36

Avg Difference from Reality
pup55 analysts t-prob
Unleaded -0.426 0.222 0.07
Distilla -0.103 -0.404 0.22
Crude 0.061 0.176 0.34

Deviation from Reality (Abs Value)
pup55 analysts t-prob
Unleaded 1.601 1.536 0.41
Distilla 1.168 1.180 0.48
Crude 3.844 3.021 0.02

pup55 Forecast Correctness
Avg Exact
Unleaded 0.019 4
Distilla -0.013 5
Crude 0.007 18

Avg Exact
Unleaded 0.025 5
Distilla 0.000 7
Crude 0.026 1

Avg Exact
Unleaded 0.104 2
Distilla 0.047 3
Crude -0.061 5

Here are the stats for the year-to-date. Despite being slightly ahead of the analysts on closeness to reality (over 50%, which as we all know was my long term goal) the pesky analysts are beating me on average deviation from reality in crude oil and unleaded, by a statistically significant margin.

The reason for this is that my little unleaded demand model underestimated demand consistently in Jan-March (it has been pretty close the last couple of weeks) and also, the lack of ability to predict crude oil and unleaded imports. Of the first roughly six months of the year, I have only managed to correctly predict crude oil imports once.

Other than that, pretty good. Now that I understand the seasonality a little bit better, it should be possible to refine this demand model some.

The crude oil imports are up to the Saudis, Mexicans, and weather conditions in the Gulf, so I do not feel too bad about that.

Thanks for your help as we continue to try to understand this important data a little better, to keep with the original goal of the thread, which is to use these reports as a potential indicator that the effects of PO are being felt in the US.

No comments: