Time to Really Do Contango Homework

There are recently a boatload of misleading, misinformed and inaccurate (repetitive, I know) posts by unnamed others about doing one's "contango homework", about looking at the shape of the futures curve to somehow make some conclusion about the prospective gains or losses to investors in said "contango" or "backwardated" commodities.

So, ok, let's do the homework.  Below are the results of 33 years rolling positions in the VistaCTA commodity basket of 15 major commodities (crude, heat, natgas, rbob/gasoline; corn, wheat, beans; gold, silver, copper; coffee, sugar, cocoa; cotton and frozen concentrated orange juice).  These are the highly select well-formed futures contracts included in the VistaCTA basket.  The results of rolling these contracts since 1980 are shown below.

Per the VistaCTA rules, the basket is rolled once per year, held for a year, sold and replaced with a new basket.  The sell and replacement is the "roll".  Note rolling does not create a return event, only the liquidation (vb sell) of the existing basket versus the original basis (vb buy) creates return.  The vb's shown are each year's new basis.

yr          vb                 cb       %cb     r

1980 613,794.44 contango 5% -26%

1981 497,085.43 contango 9% -16%

1982 431,779.92 contango 3% 6%

1983 459,955.46 contango 0% -11%

1984 424,249.76 contango 4% -11%

1985 375,251.52 backwardation -1% -10%

1986 348,666.38 contango 3% 2%

1987 367,878.51 contango 3% -2%

1988 349,500.62 backwardation -3% 0%

1989 345,122.92 backwardation -1% 19%

1990 395,668.58 backwardation -4% -14%

1991 343,179.91 contango 1% -8%

1992 321,498.46 contango 2% 0%

1993 336,825.07 contango 4% 16%

1994 406,335.85 contango 4% -7%

1995 368,142.81 backwardation -3% 4%

1996 365,516.05 backwardation -4% 6%

1997 375,650.90 backwardation -3% -15%

1998 338,576.80 contango 6% -4%

1999 325,407.08 backwardation 0% 14%

2000 369,313.10 backwardation -1% -22%

2001 307,072.03 contango 7% 17%

2002 348,316.05 backwardation -3% 11%

2003 376,951.43 backwardation -2% 31%

2004 476,701.93 backwardation -4% 41%

2005 608,361.51 backwardation -10% 10%

2006 715,832.95 contango 7% 14%

2007 805,308.50 backwardation -1% 8%

2008 911,174.50 contango 5% -14%

2009 815,403.10 contango 5% 16%

2010 957,105.35 contango 1% 11%

2011 1,065,868.55 contango 0% 8%

2012 1,130,263.50 backwardation -2% -11%

correl correl (0.20)

yr = roll year

vb = VistaCTA basket value when purchased on the roll date.

cb = contango (%cb>0) or backwardated roll (%cb<0)

%cb = % contango or -%cb for backwardation = (vb new)/(vb old) -1

r = subsequent rate of return of the basket after the roll = (vb sell)/(vb buy) -1

Clearly, there is no relationship between contango/backwardation and futures returns.  How sad that billions apparently are committed to a false concept, a relationship that does not exist.  A recent academic paper demonstrates that "roll yield" doesn't even exist, it is an error in the construction and calculation of legacy commodity indices.