Thanks for your long reply, eklypse. My questions below. eklypse Wrote: ------------------------------------------------------- > hey sticky, i think i see where your gripe is. i > went back and read the CFA text on pg 300. It > says: Great. Let’s look at p.300 > “when the futures markets are in backwardation, a > positive return will be earned from a simple > buy-and-hold strategy”. > > the confusing point is that the wording says > “positive return”, when it should really say > “positive ROLL return”. when you initially read > it, the sentence implies that it is referring to > the TOTAL futures return. (side-comment: I think the text is trying to say it’s +ve roll return so it’s +ve total return — but this has little to do with our discussion) I doubt if backwardation means +ve roll return. Have a look of Exhibit 16 on same page. If the futures prices in columns (2) and (3) are swapped, then — a. futures prices are STILL going down as time goes by. YOU STILL HAVE BACKWARDATION b. all the roll returns in column (5) become -ve Can you still get +ve roll return from buy-and-hold? (in fact, with futures prices going down, I wonder why May prices are higher than April prices. That doesn’t sound right to me — though I can still do the calculations accordingly) > the total futures return will be affected by other > things, namely the underlying asset’s price. if > the underlying asset falls in value, the total > futures return in a buy-and-hold strategy could be > zero, or even negative, as you had hypothesized > previously. BUT it will exhibit a positive roll > return (simply because, as defined, a market in > backwardation has a downward sloping term > structure… so as time to maturity decreases, the > higher up the y-axis the futures contract will > go). > > so yea, total return = spot return + roll return + > etc etc… > if spot return is -ve and roll return is not > enough to counteract it, total return could > definitely be -ve. > > Hope that makes sense. Agree. In backwardation, I’d say (any) +ve roll return most likely can’t compensate a -ve spot return. Do you agree? If so, I would most likely NOT recommend a buy-and-hold strategy, as opposed to what’s mentioned in the text. Looking forward to light shed on this. - sticky
thoughtful insights, enjoying this healthy conversation. ok, on with the show… sticky Wrote: > I doubt if backwardation means +ve roll return. > Have a look of Exhibit 16 on same page. If the > futures prices in columns (2) and (3) are swapped, > then — > > a. futures prices are STILL going down as time > goes by. YOU STILL HAVE BACKWARDATION > > b. all the roll returns in column (5) become -ve > > Can you still get +ve roll return from > buy-and-hold? > > (in fact, with futures prices going down, I wonder > why May prices are higher than April prices. That > doesn’t sound right to me — though I can still > do the calculations accordingly) hmm… i think spot price needs to be isolated away from the roll return. if column (2) and column (3) swapped figures, then the decrease in futures price will be the result of the somehting other than backwardation. in backwardation, as TTM decreases, the futures price increases due to roll return. if the futures price drops, then either the underlying price had an effect on the futures price or something happened that changed the market to turn to contango. (side note: the curve doesn’t have to be a straight line. it can be upwards sloping and then downwards sloping, or vice-versa. it could be the case that it is in contango in the ST, but backwardation in the LT) the thing with roll return is that when you are calculating it, you must go under the assumption that spot is fixed. the change in spot will ultimately contribute to a change in total return, but if we are strictly only looking at roll return, then it must be isolated away from other factors, so it must have the assumption that spot is fixed. if spot is fixed, then the futures contract must increase to meet the spot, otherwise the market won’t be considered backwardation. the assumption that spot is fixed must hold. > Agree. In backwardation, I’d say (any) +ve roll > return most likely can’t compensate a -ve spot > return. Do you agree? If so, I would most likely > NOT recommend a buy-and-hold strategy, as opposed > to what’s mentioned in the text. > > Looking forward to light shed on this. > > - sticky not sure… i have not had the opportunity to work on a derivatives trading floor 
but the very nature of financial theory is that future outcomes cannot be predicted. +ve roll return could compensate a -ve spot return, or it could not. I would need more information before i can recommend a buy-and-hold strat, but i can tell my client that so long as: 1) the market is in backwardation and 2) spot price does not change, the futures contract will exhibit positive return from the roll return. (tangent: is this ceteris paribus?) side note: i did a google search and found an article talking about backwardation and roll return. hope it helps. just do a search for “roll return” http://www.marketthoughts.com/z20061022.html ok, i’m heaidng to bed, but will check in again tomrrow morning. take care,
sticky Wrote: ------------------------------------------------------- > JoeyDVivre Wrote: > > Futures prics are almost always higher than > E(S) > > because that’s the normal arbitrage > relationship > > (Futures price = price to finance, store, and > buy > > something now). Anyway, backwardation vs > contango > > gives you almost no information about whether > or > > not you should be long or short. It’s about > the > > same as saying “the stock pays a good dividend > do > > you want to be long or short?” > > (sorry miss this part) > > So why is the CFAI text saying “When the futures > markets are in contango, a +ve return will be > earned from a simple buy-and-hold strategy” (2nd > paragraph under exhibit 16, p.300, CFAI vol 4)? > > An example from CFAI just adds to my confusion. > Part 4, exhibit 18, p.304, CFAI vol 4. > > Question: (Given the backwardation situation in > parts 1,2) Recommend a futures strategy that will > provide a +ve return in this scenario. > > Answer: When futures markets are in backwardation, > a +ve return will be earned from a simple > buy-and-hold strategy. This occurs because as the > futures contract gets closer to maturity, its > price will rise to converge with the higher spot > price. This increase in value produces a +ve roll > return (as calculated in part 1). > > - sticky Those are remarkably stupid things to write in CFAI texts. Nothing like a zero-sum game that can be leveraged 10:1 and all you have to do is look in the newspaper and take CFA LIII and you’re on your way to riches.
That marketwatch paper is a fine one to read. I especially like Bernstein’s comments “How can you not own these things [commodities]? Easy. The planet described by Gorton and Rouwenhorst is not the one you and I live on.” I think that sums up the problem with the CFAI text as well. Trading futures contracts is hard because it’s zero sum. There just aren’t simple strategies that give you positive retuns.
thanks everybody for their input. I found myself a bit mixed between roll return and spot return before, not realizing that spot price is assumed fixed while considering the roll return. Having said that, my conclusions are not really changed that much: 1. When backwardation, roll return is +ve (since assumption being that spot is unchanged). Having said that, if you long futures, whether you will earn +ve total return is another story … 2. When backwardation, I would NOT try to recommend a buy-and-hold strategy for futures contract. (If I were to choose, a -ve total return is probably of higher chance!) Please let me know if you think different. Thanks! - sticky
eklypse Wrote: > I would need more information before i can > recommend a buy-and-hold strat, but i can tell my > client that so long as: 1) the market is in > backwardation and 2) spot price does not change, > the futures contract will exhibit positive return > from the roll return. (tangent: is this ceteris > paribus?) YES … potentially 
(you have not been found wrong yet, so no defend necessary yet) - sticky
sticky Wrote: ------------------------------------------------------- > thanks everybody for their input. I found myself > a bit mixed between roll return and spot return > before, not realizing that spot price is assumed > fixed while considering the roll return. Having > said that, my conclusions are not really changed > that much: > > 1. When backwardation, roll return is +ve (since > assumption being that spot is unchanged). Having > said that, if you long futures, whether you will > earn +ve total return is another story … > > 2. When backwardation, I would NOT try to > recommend a buy-and-hold strategy for futures > contract. (If I were to choose, a -ve total > return is probably of higher chance!) > > Please let me know if you think different. > Thanks! > > - sticky Looks good to me. So to throw a monkey in the wrench… Backwardation and contango really refer to the relationship between the futures price and the expected spot price. Since the expected spot price is not knowable, it’s not really possible to tell by looking at prices whether a market is in contango or backwardation. A market in which futures prices are greater than spot prices is a normala market and one with decreasing futures prices is inverted. Of course, backwardation and contango are much cooler words so people use them for describing a series of prices seen in the newspaper. That creates a problem, because when someone says that a futures market is in backwardation so you can have a positive expected return by going long that is either completely true (if F > E(S) then E(F - S) = F - E(S) > 0) or probably not true if you are talking about declining futures prices as “backwardated”. Since people use the term backwardated to mean both things, there is lots of confusion. As with most markets in finance, futures are pretty efficient. As a general principle, futures prices are either a) no-arbitrage functions of spot prices or, if that’s not a possibiity, b) close to unbiased estimates of expected spot prices. “Roll return” might be more interesting than above because roll return has skew. That is, in general you shouldn’t expect a roll return more than is suggested by pricing above but delivery problems always cause the the spot price to increase and there is not such a thing as “unexpectedly easy delivery” or something. You can try to play that but it just causes you trouble in life unless you are really on top of things. Otherwise you end up getting nailed by delivery options, taking delivery inadvertantly, closing out your position at odd prices because you don’t want to take delivery and everyone else knows that, etc…