Tuesday, May 06, 2008

How Weak is this Oily Chain?

Today's Finanical Express carries an interesting column by Vikram Mehta, Chairman of the Shell Group in India.

The following passage caught my eye:

"Success in the exploration and production of oil & gas requires a company to overcome three interlocking sets of probabilities. The probability that a given geologic structure contains hydrocarbons [let's call this Event A]; the probability that hydrocarbons will be located [lets call this Event B], and the probability that once located, the find can be commercially exploited [let's call this Event C]."

Vikram's statement has vast practical implications for security analysts.

The market value of an asset is the present value of its expected future cash flows. Cash flows from an oil exploration company can be derived only out of hydrocarbons which can be commercially exploited. And for that to happen ALL of the above three events must happen.

Suppose that the probability of Event A happening is 40%, that of Event B happening is 20%, and that of Event C happening is 25%.

Then the probability of seeing cash flow which is valuable is 0.40 x 0.20 x 0.25 or 0.02. That comes to just 2%!

I wonder if the market participants think in those terms before valuing oil exploration stocks.

The man who said that "a chain is only as strong as its weakest link" was wrong.

He should have said "a chain has to be weaker than its weakest link."

6 comments:

Deepak Shenoy said...

What you should care about is really only event B and C. It is only of theoretical importance if hydrocarbons exist and can't be located. Because we cannot mentally make the difference; it doesn't exist until we find it.

I would actually put 50% for each - that there's a 50% chance of finding HCs (in all blocks owned by the company, and each block is pretty big) and 50% that it can be exploited. That gives you a ratio of 0.25 or a 1/4th chance of winning.

Now that's the "edge". What's the odds? Or the Win/Loss ratio? That is the amount you make compared to the amount you can lose. So say the average "find" can give you 5x profit. that is, 5 times the amount you would lose if you didn't find anything. Which is technically your investment in the find (after the cost of commercially exploiting it)

For each win you make 5x the loss. So if you win 1/4th of the time, your return on investment is 50%. (eg. if you bet Rs. 100 on four such bets, you lose Rs. 300 and make Rs. 500 - a Rs. 200 profit for a Rs. 400 investment)

So if the probabilities are correct, the return is good, no? Not necessarily, as a company can bet the entire capital on one bet, going bankrupt before the numbers work in its favour. So if the company is betting less than the % that will make it go bust - in this case, 10% of capital available for e&p (using kelly's formula).

If a company spends more per find than this amount (this is pre-commercialisation - just an explore and find gas funda) then it is likely to hit a wall. Otherwise they're ok.

But then one can argue about the probabilities to death and back.

If your probability set were true, the return must be of the order of 100x the loss, and the company shouldn't put more than 1% of available capital per bet (and there should be a lot of bets).

Here the chain is exactly as strong as the amount of weight it currently withstands.

Anonymous said...

I would actually put "50%" for each - that there's a "50%" chance of finding HCs (in all blocks owned by the company, and each block is pretty big) and 50% that it can be exploited
---
You're a pretty optimistic fella

What is interesting is that a company has to first pay for the rights to the field, and that price is paid based on a VERY approximate estimate arrived at by an "expert" team for each particular company which is involved in the bidding. There is NO common figure to go on per se, every bidder brings his own method of estimation to the party, the figure which is reported in the press is not what a bidder pays attention to in the actual estimation of the bid

You have to consider the fact that in a few cases - The probability that a given geologic structure "does not contain" hydrocarbons - will come true, I call this a fact because it is inevitable, more the exploration activity more the chances of this happening, just because we cannot mentally make a difference does not mean we should not consider it

Let's call this "hard luck"

Include the hard luck factor into your calculations and then see what difference it makes to the "Betting Strategy" - For each win you make 5x the loss. So if you win 1/4th of the time, your return on investment is 50%. (eg. if you bet Rs. 100 on four such bets, you lose Rs. 300 and make Rs. 500 - a Rs. 200 profit for a Rs. 400 investment) - The Math will look different

And if you actually study how the BIG oil companies abroad do it, they actually bet a little of their available bankroll, there is another reason to it, but I want to keep it simple

There is another pretty cute thing here, when a BIG oil company bids for a field using their calculations they have a price of oil in their calculations which is much lower than the prevailing price, this is the "margin of safety"

And the Prof was talking about Oil Exploration Companies, whose job is to explore, not refine, not market, just explore and the interesting bit is how big a divergence there is in the bidding strategy of the oil exploration company and an integrated Oil company

Still wondering from where you got the 5x figure

The application of Kelly was great only if the underlying assumptions were sound !!!

Looking forward to replies !

Luv and kisses
Rana Pratap

Anonymous said...

I will go with Deepak's arguement that event A doesn't exist. One might choose to assign a lower probability to event B to compensate for the same. Factoring in event A would tantamount to double counting.

Coming to the probabilities assigned of 50% and payoffs of 5X, I dont have the faintest idea or clue on how these numbers can be arrived at.

There is a different angle to this calculation. Valuations also depend on the amount of reserves that have been discovered. The trick is not in discovering oil but how much oil.

This is a important variable cause this affects not just the valuation of the company but also the price of the commodity. Lets assume I tomorrow discover 10% of the existing oil reserves in a new discovery. The announcement will lead to my stock rerated upwards but oil prices going down and hence the subsequent derating.

This is the dichotomy and hence i m not sure about the validity of the data brandished by oil exploration companies with so called independent assessment.

From a value investors perceptive I would let this one pass.

Ninad Kunder

Deepak Shenoy said...

rana pratap: Perhaps I'm wrong in the probabilities but also in the payoff (i.e. if 50% is wrong, 5x is also too less, perhaps it's 50x!)

The x here is the total cost of exploration. that's the price they pay for the block plus the cost to explore.

"You have to consider the fact that in a few cases - The probability that a given geologic structure "does not contain" hydrocarbons - will come true"

This is what I call 50% for B. It's not "in a few cases" but in 50% of the cases, as per my assumption. There is no need of a hard luck factor again - this is the hard luck factor!

If companies are betting say 5% of their bankroll per block, then it is likely that a) the payoff is higher and b) the probabilities are smaller.

I can't think of a reason not to apply Kelly - it's based on presumptions of probabilities anyhow.

Ninad: I agree with you there - the quantity matters. Also the quality (heavy sells for less).

It's not been profitable so far to explore areas where quality is low (heavy, bitter, too far below) but as oil approaches $200 these are likely to be reopened - as you said, supply concerns may ease.

Rohit Chauhan said...

Hi
this sounds similar to the drug discovery process. If i remember correctly, buffett noted that it is diffcult to predict the sucess rate for individual companies, but as a group you can bet that some one will discover a blockbuster.

also in one of the interviews he mentioned that he would prefer not to pay for the possible blockbuster.

maybe we can look at the oil exploration companies in a similar way. buy if the valuations dont assume major discoveries ..if one happens then that is the bonus

regards
rohit

Aman Thakur said...

Mathematically that is not completely correct.

There is also the case that if Event A does not happen then no funds are allocated to see the outcome of Event B and if Event B doesnt happen then Event C has the same fate.

The true probability is dependent on rolling back the decision tree but I think your post was highly instructive and served the purpose of highlight analyst biases