The point of this blog is to follow the path of our money from the point it is handed to the government as tax, to the point it leaves their hands to give benefit to somewhere.
It is that ‘somewhere’ I have recently been looking into — I’ve been investigating the stage where the money leaves the government hands. Who benefits and in what way?
The Treasury devoted some resource to answering this exact question.
The Treasury Data
The data set has a name, it is the Country Regional Analysis (CRA). I’ve collected all years of this report back to 2005 in a CKAN package.
How the data is gathered
You can see that the Treasury fill the spreadsheet with the department’s spending codes (programme object group) from COINS. The department is then asked to work out which regions, in the country, benefited from each of their spending codes.
The regions are quite crude, the whole of England is split into 9 regions and then there is the option of Northern Ireland, Scotland and Wales.
Some of the ways of tracing the funds are easy, say, if there is a spending code for a local initiative that people from only one region are entitled.
A more complex analysis would be something like if there is a spending code for a museum that people travel from all over the country to see, then a statistician is called upon, and they will determine which parts of the country benefit from this service probably based on visitor surveys etc.
How to get something useful out of the CRA
The 2009 CRA data had lots of inconsistent use of classification of the spending. We wrote scripts to fix many of these problems.
The 2010 CRA is much better but still requires some work.
The 2010 CRA comes in two sets. Both sets describe the same spending, it is just one (called table 9) describes the *areas* the money was spent in more detail, and the other data set (called table 10) describes the *type* of spending in more detail.
Table 9 and 10 have to following fields in common:
- Department code
- Department name
- COFOG level 1
- HMT Functional Classification
- Programme Object Groups
- Programme Object Group alias
- ID and non ID
- CAP or CUR
- spending for dates 2004-05 to 2009-10
when we find a match between table 9 data and table 10 data for the fields above then we need to give it the NUTS 1 region code and spending for 2010-11 from table 9 and the ‘CG LG or PC’, COFOG 2 and HMT subfunction class from value from table 10
Then the full detail of the spending in one line.
When this join is done, there is one more problem to fix with the data.
This issue all stems from the fact that local authorities do not have programme object groups.
We think the best solution is to replace the ‘dummy sprog …’ programme object group that is used in the CRA for local authority spending, with something like ‘Local Authority Spending’ rather than using ‘Unknown’.