Mathematical Solutions for Optimal BidsWed, 01/18/2017 - 14:29
Q: What is your main take Mexico's licensing rounds?
A: In R1.1, the Mexican government did not disclose the reserve price to potential bidders, which was a huge mistake. In my opinion, the best auction design so far was R1.2. Bear in mind though that sufficient oil and gas reserves had already been found in all the blocks offered for sale in R1.2. This explains why bids were so high in that round. In R1.3, 25 different blocks were offered for sale. Bidders had to submit all their bids before the first block was awarded and, after the auction, many firms regretted their bidding strategy.
Q: What are the main challenges in the bidding processes?
A: The main risks companies face when bidding in oil and gas auctions is the lack of certainty regarding the extent of the oil reserves in any given block. The name of the game is to avoid paying too much by accurately assessing the hydrocarbons potential and risks involved. Round 1.2, for example, saw higher bids from companies because it was already ascertained that there was oil there. The geological risk was lower and companies knew the necessary reserves were present to make extraction economically viable. The same situation arose with the Trion farm-out. The only issue with Trion is the high cost of deepwater drilling and development.
Q: What programs are you developing that could eventually be applied to the oil and gas industry?
A: Athena could help oil and gas companies compute their optimal bids. One could use available seismic data to compute the implicit ROI present in each firm’s bid. Next, one could regress those implicit returns on different firm characteristics. This allows us to classify different firms in different clusters. IOCs have many global options for investments and this should be reflected in their bids. Mexican firms or service providers, on the other hand, may face a lower opportunity cost, which should also be reflected in their bids. One would thus expect both types of firms to end up in different clusters.