by Soft_Landing » Fri 24 Sep 2004, 16:16:32
$this->bbcode_second_pass_quote('', 'T')hat would lead naturally lead to make another kind of chart...
Charting spare capacity....vs. prices.. (over time..like the chart above to determine linear relationship
Maybe the chart's available somewhere...or we could make it...
We could we get the data?
This is trickier than it sounds. The relationship plotted can be meaningfully because price and US stocks are pretty much known quantities on a weekly basis. As far as spare capacity is concerned, well, it's an estimation based upon state secrets. In any case, that estimation doesn't change on a weekly basis, but is probably more relevant on a yearly scale. Price and stocks obviously change a lot over the course of a year.
If trading was changing based upon spare capacity, what you might expect is something like this:
In this model, there is a different relationship between stocks and price for each given spare capacity regime (indicated by the green lines).
Although we can see at once that this kind of idea isn't going to work in any simple way. After all, we know that in 2003, there wasn't a great deal of spare capacity (this is right, isn't it?) and yet, in 2003 (red squares), we can see the price vs stocks points seem to be on the same stocks X price curve. What we do notice, though, is that all the 2003 data points are on the left of the scatter, indicating that stocks were on the low side historically for the entire year. Given what we know about US refining capacity, this is probably just explained by an unexpected increase in demand.
Of course, this is a problem for the 'spare capacity as a variable' model. The way you'd want to alter it, in my opinion, is factor in an expectation of future spare capacity, or something like that. When you have a relationship like this, i believe it tends to get well known among traders in the commodity. Regular traders are likely to decide when to buy based upon curve in relation to the typical variables. This makes the old model something of a self fulfilling prophesy. It takes a while for the traders to start acknowledging new information.
One reason spare capacity is interesting in a model is because it eliminates downside risk for speculators. If there is little spare capacity, then price can hardly come down too much - there'd be no more product for the market to sell. Thus, the lower speculator's (i'm taking long term here) believe spare capacity is, the more that oil becomes a safehaven of sorts. That was my reasoning behind proposing spare capacity as a variable.
But if spare capacity is going to work this way, then that means the speculators first have to believe that spare capacity is low, and that it's not going up any time soon.
Now, if you look back at the graph, you can see that the 2004 datapoints seem to cover a range between the mainstream correlation, and a second accumulation that is off the main curve (circled region). If you were to use your imagination, you could imagine traders figuring out that spare capacity was low near the beginning of 2004, and as a new trend forms, data points drift right toward the new accumulation.
This could be tested if we had dates for each of those 2004 data points. They're probably on the IEA website or somewhere. I'll see if I can find them and post the results...
I've just realised one other thing worth mentioning... The vertical axis is dollars in nominal terms, as far as I can see. If inventories have changed significantly since 1993 toward more of a just in time schedule, then you'd expect inventories to get smaller as price goes up (ie time). Make sense? This just suggests an artifact that could contribute to the appearance of the curve. Ideally, the vertical axis should really be in constant dollars.