[wdmmg-discuss] CRA 2010: description and questions

Thu Aug 12 12:12:04 UTC 2010

On Thu, 12 Aug 2010, Anna Powell-Smith wrote:

> By '2009 CRA database dump' you mean the CSV file hosted on the
> Treasury site, and used by the bitbucket package, right?
> http://www.hm-treasury.gov.uk/d/cra_2009_db.csv

That's exactly right.

> I wonder why it's not linked to from anywhere on the Treasury site,
> including the official PESA 2009 pages.

It is linked to on the Treasury site here:
http://www.hm-treasury.gov.uk/pesp_cra.htm#data 
they call it 'CRA Database (CSV file, 5.4MB)'

>> This year the Treasury have given us two well defined data sets:
>> - one for the data behind PESA table 9
>> and
>> - one for the data behind PESA table 10.
>
> These data sets are also available in 2009 - would you just have faced
> the same problems as with the data dump? or were they not available
> when you started work? http://www.hm-treasury.gov.uk/pespub_pesa09.htm

The raw data behind the 2009 PESA tables 9 and 10 was not avilable in any 
form other than the database dump.

>> We can join the data together ourselves in a controlled and well defined
>> way.
>
> Ahh! I'm starting to see how we might do this. We can use POG IDs,
> because these are in both spreadsheets, conveniently.
>
> So if you filter the 'POG Alias' column by e.g. "P08 S081305 Fire
> Superannuation" (fire service pensions, presumably), then in Table 9
> you see nine rows with spending in 2009-10, one for each region,
> classified as e.g. England > East Midlands > Social protection.
>
> And conveniently, if you do the same in Table 10, you see nine rows
> with the same sums, all of which are classified as England > Social
> protection > Old age (I assumed they would have collapsed all the
> regions into a single row for England, but not so).
>
> We can merge them together - well, assuming that they are consistent anyway.
>
> Lisa, is this what you had in mind all along? You were one step ahead of me :)

That is what I had in mind.

>> (Alistair, the 2009 CRA database dump is not a one off either, we have
>> database dumps for every year going back to 2005).
>
> Did these all come out as part of the same FOI request?

I had to request the database dumps for 2008 to 2005, they arn't published 
on the Treasury website.