I am not an oil economist, but on a whim I grabbed active oil and natural gas rig counts from Baker Hughes and end-of-month spot oil prices from the EIA. I limited the sample to January 2002 through April 2005 and estimated (simple) price elasticities of active rigs: U.S. about 0.56, Canada 0.51, Middle East 0.27, South America 0.49, Far East 0.28, Africa -0.17, and Europe -0.49.

What's up with Europe and Africa?

Here is Europe's rig count plotted against end of month spot oil prices from January 2002 through April 2005:

Here is Europe's rig count and the end-of-month spot price trend lines from January 2002 through April 2005:

For comparison, here are the similar pictures for:
Rigs versus spot prices: World, USA, and Canada
Rigs versus spot prices: Middle East (different scale)
Rigs versus spot prices: USA

My initial thoughts:

1. Perhaps the US (and most of the rest of the world?) is firing up more natural gas rigs over the sample period thereby giving spurious results. If Europe and Africa do not have natural gas reserves, then the Europe/Africa results might be spurious - but I don't buy that.
2. Perhaps if I expand the sample back into the 1980s or 1970s there will be a different overall relationship, but this still wouldn't explain the negative relationship post Iraq invasion.
3. Perhaps there is a political (oil for food?) explanation?
4. Perhaps the US (and most of the rest of the world?) has more marginal oil rigs that activate when price increases? (this picture at the Dallas Fed might support the marginal-rig claim)
   
5. African and European rigs are running dry just at the time price is increasing - dumb luck that.
   
6. African and European rigs are owned by the government and therefore the owners are anti-profit
   
7. There is some odd optimal extraction rate algorithm being used in Europe and Africa
   
8. Africa and Europe operate on target revenue rather than profit maximization

If someone has an explanation, I'm listening. It's kinda late over here, so maybe I am missing something obvious.

For those interested - here's the STATA data file.