Why Futures Basis Divergences Are Not Directional Signals

# Do futures have predictive value?


# Do futures have predictive value?












Futures closely mirror their underlying, as can be seen in the charts below. Eventually, at expiration, they reach the value of the underlying. However, they seem to show no extra information about the underlying's future; i.e. they don't trend any closer to the underlying's value at expiration than the underlying itself does.

Below is an underlying, SPY.

And below is the ES futures Dec 2012 contract.

However, it is possible to calculate how often a given future has been directionally correct. For instance, SPY closed at 142.79 on Dec 21, 2012. To determine the directional predictive power of the ES future, one would run the following pseudocode in a backtesting system:

```
$SPY_close = 142.79
$future_predicted_direction = 0
$days = 0
FOR $day (2012-01-01 .. 2012-12-21) {
  $days++
  IF $EOD_quotes["SPY"][$day] > $SPY_close
      THEN $future_predicted_direction += $EOD_quotes["ES"][$day] <= $EOD_quotes["SPY"][$day]
      ELSE $future_predicted_direction += $EOD_quotes["ES"][$day] >= $EOD_quotes["SPY"][$day]
}

print "ES futures correctly predicted direction $future_predicted_direction out of $days days, that is, "
+ int($future_predicted_direction / $days * 100) + "% of the time."
```

Are there any studies on this? Do futures have any predictive power?

## Answer by Matt Wolf (score 7)

https://quant.stackexchange.com/a/7371

No I believe there is no directional predictive value derived from looking at divergences between futures and their underlying price value. The reason for divergences are of the no-arbitrage argument type. Futures could be arbitraged (and are immediately if such arbitrage opportunities surface, even those opportunities may only fill the stomach of a single person, not a whole family.)

Now, one may argue that there is reason to believe that when futures trade in contango or normal backwardation that this tells something where the market believes price will tend to trade. That is absolutely true but I would argue that there is no alpha to be generated from such knowledge. How farther out futures trade vs near months is something everyone knows and it comes down to the same old risk/reward equation since mankind was created: do you believe an asset is overvalued for a reason and thus go against what most market participants believe will happen or do you sell tops and buy bottoms , which is another way of saying that you want to be a value investor. The choice is yours, in fact if done right both approaches can work. There are a few highly successful traders/investors in either camp. So it comes down, IMHO, to not whether prices will converge to what you define their long term mean and thus whether futures lead their underlyinging asset prices but it comes down how you time your own trades, and particularly how you manage your capital and positions. I believe an understanding of futures pricing is similar to you reading in the newspaper how to hold a tennis racket, you don't win Wimbledon with that. It's a basic ingredient but not a recipe for success.

I would not invest too much time in searching for academic studies in this realm, after all have you ever seen a single academician who published a study in this particular field who struck it rich? I have not.