In the top 5 execution venues report, there is a Y/N flag to say whether in the previous year there was <1 trade on average per business day.
From the RTS “Where, for one or several classes of financial instruments, an investment firm only executes a very small number of orders, information on the top five execution venues would not be very meaningful nor representative of order execution arrangements.”
I interpreted this to mean that if we state that there is less than one trade per day on average, then we won’t fill in the rest of the template on the grounds that to do so would be at best unhelpful and at worst misleading.
However, the alternative view is that we should flag Y or N, and then fill in the template regardless. The presence of the flag allows the reader to assess the usefulness or not of the data which follows.
I would love to hear from anyone who has implemented this, or is in the process of doing so, to know what their interpretation is. (The follow up question is how did you do the calculation of the average, but don’t feel obliged to answer that one now.)
Many thanks,
David McKinney.
Surely I’m not the only one building these Rts 28 Top 5 reports? 
My question is: If, in the previous year, there was <1 trade on average per business day, do you fill in the rest of the table, or not?
Many thanks!
We consider it as out of scope. We do not pubhlish data on trade <1 /p. b. day executions.
Best Regards
Adis
Thanks Adis…that’s our position too…which is reassuring.
Has anyone found a tried-and-tested “machine readable” format to use for the RTS 28 publication? Somehow I don’t think Save As CSV is going to cut it.
Thanks!
Carl
Hi Carl, we are proposing an .xls for download. We decided against .csv as probably half of our clients will have their Excel set up to handle a comma separator while the other half expect a semi colon. XLS seemed liked a better option.
However were it not for that consideration, we would certainly have gone for a CSV, and indeed the format we propose will be entirely csv compatible / equivalent.
If you consider a single template, you basically have seven fields, five of which are repeated up to five times (for the top 5). Add to this additional fields to denote the Client Type / Venue Type / SFT / Year and you basically have a csv with around 10 fields. A ‘Rank’ column with a value from 1 to 5 means that each row of the csv uniquely identifies one row of a template report.
I should perhaps clarify that we propose to the user to download all the top 5 reports in one shot, so the csv will likely have a few hundred rows.
My take on ‘machine readable’ is that it’s best to provide in a format which allows the client to do their own sorting / filtering / slicing and dicing in their spreadsheet. This is one of the reasons we preferred csv to xml for example, which could be better were the data being passed to another system for consumption, but would not be very user friendly for the average client.
Those were my thought processes…interested to hear yours.
David.
I don’t think there is anything in the regulatory text that says the you should not publish a table where there is <1 trade on average per business day.
The fact that the “Notification if <1 average trade per business day in the previous year” field in the RTS 28 tables has a permitted value of “N” suggests that we should publish the table when there is <1 trade on average per business day. Otherwise a permitted value of “N” would be superfluous.
Carl / David.
Format has been discussed recently by the Best Ex working group. The consensus is that csv format is machine-readable, but xls/xlsx files are not.
In order to be machine-readable, a file must be in a non-proprietary format that can be read and processed by any software program (that is written to do so). I believe that the intention of this requirement is to enable electronic / programmatic consumption and processing of the information.
Regards,
Mark
[Re the display or non-display of data when avg < 1]
Thanks Mark…noted. We’ll reconsider for the 2018 report. For the 2017 report, we’ll not publish the rest of the table.
I still feel our position is justifiable, on the basis that it “would not be very meaningful nor representative of order execution arrangements.”
Cheers.
Thanks Mark,
We’ll propose a choice then between Excel and csv. For me, however, this download is for the client. If the intention was for something for automatic consumption, they would have specified a format. I can’t see anyone writing code for all the different versions…and if they did, I’m sure they could work out what to do with an Excel file.
But I can understand the reasons for the working group position.
Cheers,
David.