DonIce wrote: ↑
Thu Feb 21, 2019 7:58 pm
However, when I look at the historical returns of the 10 year treasury note futures, for example here (data available from 1983):
https://www.investing.com/rates-bonds/u ... rical-data
and compare it to the historical returns of actually holding the 10 year treasury note, for example here:
http://pages.stern.nyu.edu/~adamodar/Ne ... retSP.html
I get an annual average annual return of about 1.3% for the futures (with a standard deviation of 7.5%), while holding the actual treasuries has an annual average annual return of about 7.4% (with a standard deviation of 9.5%) over the same time period.
Maybe you've figured out an answer to your question based on everyone's responses, but I didn't quite figure it out. I've still struggled to "see" where the yield of the underlying bond comes through on a future. I get everything in concept, but was missing something in the practical implementation.
What I've determined is that the future data that you are looking at (and what I've been looking at) removes the yield with the expressed purpose of focusing the futures price only on the price of the underlying bond (and other future things like embedded options and things I don't understand or think really make a big deal to us over the long term). I suspected this for a while, but haven't been able to put my finger on the cause until now.
I'll try to summarize the article that pointed this out to me, but it's linked below as I'll probably fail...
Non-Earth shattering background:
Future price = Spot - ([present value of the] yield - funding)
The (yield - funding) naturally decreases as time goes on in the bond market because there is less and less interest payments to be received as one gets closer to the delivery date. Therefore, if the underlying price of the bond doesn't change throughout the holding period, one should be able to buy a future for some amount and watch its price naturally increase as you get closer to delivery. (This the the yield we are trying to "see.")
When you roll this contract---still assuming the underlying bond hasn't changed in price---you'll sell the contract for a higher price than what you purchased it for and buy the new contract at that original lower price. There is the yield. It's call backwardation or roll yield (or roll up yield) in the biz.
Why doesn't the data show this?
The data, as mentioned, adjusts (or doesn't adjust, actually a little confused on this point) the opening price of that new contract to equal the closing price of the previous contract; thus, removing the effect of the yield. The data shows you opening a contract at 155 with no price change until you sell it at 155 and open a new contract at 155. But what actually happens is that you buy the contract at 145, and sell at 155 and buy a new contract at 145.
There might be very good reasons why future traders do this, but it isn't helpful for us buy-and-hold types because we care about the total return from rolling the contract. Based on price alone, the data would show an under performance as compared to the total return. The total return of the future should still under perform the bond itself by the funding rate. (This is where you take into account the yield earned on your cash, but for a leveraged portfolio, it will be of minimal help.)
Hopefully other gurus will chime in if this needs adjusting...
https://www.themacrotourist.com/posts/2 ... rn-lesson/
How do you adjust the data to find the total return?
I believe there must be a source for the data we're after, but it might not be available for free. The article uses Bloomberg to make that adjustment. Maybe Quandl does as well, but $20 only gets you the past 5 years of data to play with for a month. If one could find a source for the individual contracts, you could roll everything yourself, which is what Rob Bertam did in his back testing. He said he pulled the individual contracts from Quandl, but I haven't figured out how to do that yet.