Its the midpoint of the closing range (the last 30 seconds from 4:14:30 to 4:15:00 PM EST... The settlement is the price used by all the brokers and accountants for end of day tracking, margin, etc.
Thank you, I am aware of this definition. If I look at the close of the large S&P, it was at 1052.10. The settlement was at 1048.30. Assuming that 1048.30 is the midpoint, this would imply that the low of this range was smaller or equal than 1044.50.
Clearly, if I look at an intraday chart of the large S&P, it did not trade around 1044.50 prior to the close and by no means there was a range of 8 points during the last 30 seconds!
So your answer is correct, but does not allow me to solve the puzzle.
As well, it seems to be a bit of a relic that the big contract still determines settlement price. When you see everyone at the CME floor watching the ES, it's clear the big contract volume is not what it used to be.
Its based on pit activity alone, no? We can't see that from our charts... Isn't that why we have to rely on the CME to publish the settlement prices after trading hours.
No, we cannot see that from our charts. But CME publishes charts of the large S&P as well, and as you can see, the large contract did not trade anywhere near 1044.50 during the time prior to close.
I doubt that it traded at 1044.50 in the pits during the last 30 seconds.
Because nobody answered my question, I will try to do so myself. After searching the web, I have found an interesting document that might explain, why Tuesday's settlement price was different from the midpoint of the closing range.
CME seems to use a different method to establish the settlement prices for index futures on the last business day of the month. The settlement price is derived from the cash index on a fair value basis, adding finance charges and substracting dividends. The cash index used as a reference is the cash index at the NYSE close, which is 3:00 PM.
In short, the daily settlement time for index futures on the last business day of the month is not 15:14:30 - 15:15:00 CT, but simply the close at 15:00 CT.
If you look at the chart below, you will see this explanation confirmed.
Thanks everybody who contributed, the answer is here. Thanks also to Yvan, who first noticed the abnormally low settlement value of ES for yesterday.
The following 2 users say Thank You to Fat Tails for this post:
The market was moving fast and either your or my data has inaccurate time stamps. If the time stamps are a few seconds off, this can make up for the difference.
Here it is again, i reloaded my data. On the included chart you'll see the 1min candlestick at 16:00 Eastern Time. The middle of this candle is exactly at 1048.25, ie, the settlement price as found on the CME web site. I also checked last month too and my settlement price is correct using the middle of range for the candle at 16:00. For July 30 i have 1098.25 at the middle of the range of the 1 min candle at 16:00.
Again, my chart shows different closes than yours (Tuesday 31, August) for the 1minute candlestick chart. I'd be curious to see other people 1min. chart for Tuesday 31, August. So Fat i think you have a problem on your end.
Agree with the settlement prices. Still I do not know, whether they are based on
- the last calculated price of the day for the S&P 500 index
- or the midpoint of the calculated price for the S&P 500 index for the 30 seconds prior to close
Your data is different from mine. Same applies to July 30, where my close is a few ticks off your close.
So, at least one of our data sources is inaccurate. Need somebody with a different data feed to check. I would not be astonished, if Interactive Brokers is a few seconds late.
The following user says Thank You to Fat Tails for this post:
You solved this yesterday - end of month settlement is FV of cash. (plus settlement is never the mid of the last min, it's the last 30 sec)
But to deal with the current discussion, iqfeed has the 4pm last trade at 1048.50
Looks like the exchange does not provide for time stamps. So every data feed provider uses his own time stamps. Interactive Brokers possibly gets the data a bit delayed, so what they stamp is the data of the a past which is already a few seconds ago. This applies to their historical data, so no conclusions possible as to real-time data.
What is the difference between three datafeeds and three lawyers?
The datafeeds will give you only three different versions, the lawyers at least four.
