Metallgesellschaft's Case

[hsuwang]

Hello David,

I think I'm getting really confused about the different positions (long/short) in futures and forward contracts associated with the Metallgesellschaft's case. If the market is in backwardation (foward price lower than spot), wouldn't MG's short position in long-term fowards be making money (because it can later purchase the oil/gas at a cheaper price), and it's long position in short-term futures be losing money? However, in the case study, it seems like the other way around is true. I think what I'm not getting here is the term "backwardation" and "contango". In the videos you said in backwardation, spot price is increasing, but if it is, why would the curve be downward sloping? (spot price dropping would make more sense to me)

Also another confusing part is, from the Gallati text, it says:
"customers were given the option of exiting the contract if the nearest month futures price listed on the New York Mercantile Exchange (NYMEX) was greater than the fixed price defined in the contract."
- this part seems totally absurd to me because who would want to exit a contract and buy oil/gas on the market when it is cheaper to buy it at the lower contracted price?

I'm sorry if these should be really basic concepts. I probably would not be able to fall asleep tonight if I don't get this case cleared up! lol..

Thank you.

 

[hsuwang]

Hello David,
I've read about your article "Contango Vs. Normal Backwardation" on Investopedia - http://www.investopedia.com/articles/07/contango_backwardation.asp and it really helped me clear up my confusions. (I think it was because of your great articles on investopeida that I actually found out about Bionic Turtle, and became a customer!)

However, still here are some questions that I'm having about backwardation and contango:

-"As we approach contract maturity, the futures price must converge toward the spot price. The difference is called the basis. That's because, on the maturity date, the futures price must equal the spot price."
::: Here do you mean the spot price will converge toward the futures price? How I interpret this is as we approach the contract maturity, the contracted price will stay the same. eg. If you bought a 1-year futures contract at $90, upon maturity, you will pay $90, the price will not change.

- why are the curves you used in the slides different (for contango it slopes upward, backwardation slopes downward) as compared to the ones in your investopedia article?

- I'm still not sure about the 6-2 table on pg.447 of Gallati, as to why the overall futures price is dropping every month, but the "next-month futures price" is still higher than the "near-month futures price".

- and finally, if the market was in backwardation (prices rising), why would MG want to get involved in a foward contract?

Thanks!

 

[David Harper CFA FRM]

Hi Jack:

Really strong questions on a case that teaches me more each year. You may find this thread from last year helpful. Thanks for noticing the investopedia article, I'd forgotten about that

I think I’m getting really confused about the different positions (long/short) in futures and forward contracts associated with the Metallgesellschaft’s case. If the market is in backwardation (foward price lower than spot), wouldn’t MG’s short position in long-term fowards be making money (because it can later purchase the oil/gas at a cheaper price), and it’s long position in short-term futures be losing money? However, in the case study, it seems like the other way around is true. I think what I’m not getting here is the term “backwardation” and “contango”. In the videos you said in backwardation, spot price is increasing, but if it is, why would the curve be downward sloping? (spot price dropping would make more sense to me)

I think I said in the video (or I tried to say) that (the shift from backwardation to) contango resulted from spot prices dropping (more rapidly than forward prices dropped). So, I don't remember saying it, but if the spot price increases more than the forward, eventually backwardation ensues. IMO, the hard thing here is the time dimension. First, the backwardation that characterized MG's planned scenario could have (did initially) produce profits for both the short long-term forwards (because they were priced at a premium; so, this is to agree with this part of your statement. Second, the long short-term futures were "stack and roll" so, they are different due to their rolling over, and what overwhelmed their returns was the "roll return"

(and below you ask why my investopedia chart is different, that's because it is trying to show the time dimension: it is trying to show that normal backwardation [the green line] implies that the forward price is increasing as it converges to the spot - this is the postive roll return earned under backwardation. But having looked at it now freshly, i think it's technically correct but almost no help in understanding)

So, IMO (see this thread from last year, may be helpful), it is good to separate the two legs of the hedge:

1. the long term short forwards - I may be wrong, but my recollection is that these were never a problem. And the are not the essence of the case. The essence of the case, I think, is the epic basis risk created by the mistmatched hedge (namely, long unmarked forward positions hedged with marked short futures). Here is the thing which may cause confusion: the profitability of these, as you say, depends on the forward price (locked in more or less) and the future spot price (because they have to buy to supply). The shift from backwardation to contango can result from various causes:
* spot is stable but futures price increases
* spot increases but futures price increases more
* or, in their case (MG): future/foward price decreases but spot price decreases more! [important, as you suggest, this itself was profitable for the short forwards! The hedge ruined MG, not the underlying !]

2. the short-term futures. Here the contango hurts because of the negative roll return: if you think about the spot price as stable in the short term, and the forward price is higher (i.e., contango), as you go forward in time, the forward price is dropping to converge on the spot. The roll over loses here.

Also another confusing part is, from the Gallati text, it says: “customers were given the option of exiting the contract if the nearest month futures price listed on the New York Mercantile Exchange (NYMEX) was greater than the fixed price defined in the contract.” - this part seems totally absurd to me because who would want to exit a contract and buy oil/gas on the market when it is cheaper to buy it at the lower contracted price?

See next: "When this option was exercised, Metallgesellschaft made a cash payment to the company for half the difference between the futures price and the fixed price. A company might choose to exercise this option in the event of financial difficulties or if it did not need the product"

-"As we approach contract maturity, the futures price must converge toward the spot price. The difference is called the basis. That's because, on the maturity date, the futures price must equal the spot price."
::: Here do you mean the spot price will converge toward the futures price? How I interpret this is as we approach the contract maturity, the contracted price will stay the same. eg. If you bought a 1-year futures contract at $90, upon maturity, you will pay $90, the price will not change.

Right, agreed, but this convergence refers not to the contract you entered into but to the market (quoted) futures price. Say you purchased 1-year $90 on Jan 2009, when the future/forward price was $90 (F0), maybe the spot was $80 (S0). Now forward in time to May 2009, both the spot (S1) and future price (F1; i.e., a six-month future maturing on Jan 2010) are different. Basis convergence refers to, as approach Jan 2009, Sn converges with Fn (in theory)

- why are the curves you used in the slides different (for contango it slopes upward, backwardation slopes downward) as compared to the ones in your investopedia article?
As above, trying to illustrate why roll return is positive under backwardation

- I'm still not sure about the 6-2 table on pg.447 of Gallati, as to why the overall futures price is dropping every month, but the "next-month futures price" is still higher than the "near-month futures price".

Right, that's the hard part and you (IMO) found the essence of the whole case! The near month < next month is the contango, and the monthly losses under future stack (I just noticed an error in my print of Galletti's case: p447, third column from right [futures monthly settlements] should be all negatives, those positives at bottom of column are wrong), that is the rollover loss as, e.g., the November 17.06 rolls over into a $234 million loss as the future *drops* to 14.41 in December (14.41-17.06). See how the forward is converging (chasing down) to the spot price; that is the rollover loss created by the contango. (and, finally, these losses were realized but the offset forward profits were not; the paper problem was worse than the actual - ironically, the underlying was "hedging" the hedge at this point!)

- and finally, if the market was in backwardation (prices rising), why would MG want to get involved in a foward contract?
right, it's a good question but hopefully the above helps to explain why this commingles two dynamics. Backwardation does not necessarily imply rising spot prices. Rather it implies that the future price (i.e. not contracted, but future price on same expiration data with decreasing maturity over time) will increase more, or decrease less, than the spot as maturity nears.

David

 

[hsuwang]

Hello David,

I think I got it! It took me a while to get the dots connected. It sure is a complicated case! I will still have to spend a bit of time on it, but thank you so much for the clarification! I feel bad for asking long questions like this.. thank you for taking the time & effort!

Jack

 

[hsuwang]

Hello David,
I read the thread you suggested and it was really helpful. However, you mentioned

"in MG’s case the spot price declined yet the backwardation shifted to contango (i.e., the forward declined more) and these forward-contracts-with-customers are still okay (temporarily)."

shouldn't it be the spot price declined more than the forward? (to make it contango) I think what I'm not getting here is "the forward declined more" part.

Thanks!

 

[David Harper CFA FRM]

Hi Jack,

sorry for delay, I missed this question. Right, the prior thread had a mistake (I just corrected). Of course (I think it's pretty obvious) I didn't mean that, as reflected by all of the above: if forward is falling, backwardation can only shift to contango if the drop in spot is exceeding the drop in forward. Thanks, David

 

[hsuwang]

Hello David,

I was just looking over the MG case again and was wondering since MG was hedging with long position in short-term futures, couldn't they just stopped hedging when their hedge started to loose money? ie. since the hedge was short-termed, as soon as they see signs of contango, they could've stopped purchasing futures. Or, was there a term in the contract where they weren't able to do so? Thanks!

 

[David Harper CFA FRM]

Hi Jack

It's a good question to which I don't know the answer. If you look at Table 6-2, the futures stack was losing over $100 million per month from May to December, except only $5 in August (perhaps that month gave them false hope?). Therefore, given they were rolling short-term contracts, your question seems good; presumably they did not need to maintain this hedge strategy in light of the contango. Now that you raise this issue, which I did not consider, b/c i didn't really think about the fact the strategy *appears* to be in obvious deterioriation for fully 8 month (!?) rather than abrubtly in a single month.

The hedge was a stack and roll, so that means the short term futures positions were *huge* but that doesn't explain why they persisted over several months. Of course, internally, they viewed these stacked futures as hedges that associated with the long forwards, so perhaps they weren't able to disassociate the hedge from the underlying (but that doesn't feel convincing, though; in May, when they lose 137 million, that would seem to be a wake-up call that the hedge has, er, issues?!). Good question...

David

 

[David Harper CFA FRM]

David

....although, on reflection, I think I've overstated the immediate failure of the hedge: I don't think Table 6-2 reflects the paper gains on the short forwards associated with the falling spot prices (I *think* the small positive, supply contracts are only current realized). If that's true, a possible explanation, consistent with a theme of the case (i.e., that futures settled daily and required cash but forwards did not) is that: initially the hedges were basically working from an accrual standpoint and they underestimated the cash flow problem. As the case says somewhere, in the long run, they probably were in good shape...David