[wdmmg-discuss] Local Spending Report

[Alistair Turnbull]

I've been getting to grips with the Local Spending Report:

 	http://ckan.net/package/uk-local-spending-report

It's a lot harder to understand than the CRA, even though (and partly 
because) it only covers about a quarter of government spending, for one 
year.

The first problem is the large number of organisations that do the 
spending. They are split up geographically (obviously) but also by 
function. For example, about 3/4 of local spending is health spending, and 
that all goes through various bits off the NHS. Of the rest, quite a lot 
is spent by the other emergency services. Each of these services (NHS 
Trusts, Ambulance Trusts, Police and Fire authorities) is split into local 
authorities in a different way.

The second problem is that some of the spending bodies do nothing but 
manage and fund the other spending bodies. This is especially true of 
health care. For example, the new-fangled Strategic Health Authorities 
have a purely managerial role, and the Primary Care Trusts are all about 
procurement from other authorities.

In theory we have a good data model to represent these complex flows of 
money. We can make accounts to represent each authority, and we can make 
transactions that transfer money between them. We will need further 
accounts and transactions (far more!) to represent the money actually 
spent on services. We should be able to see how much money is lost (spent 
on paper clips, window cleaning, etc) at each level of government.

However, it is difficult to work out who pays money to whom. The general 
pattern is as follows, I think (but I'm not very confident):

  - Central government pays Upper-tier local authorities.
  - Central government pays Strategic Health Authorities.
  - Central government pays Primary Care Trusts.
  - Upper-tier local authorities pay police and fire and waste services.
  - Primary Care Trusts pay Ambulance Trusts.
  - Primary Care Trusts pay NHS Trusts.
  - Department of Work and Pensions... dunno. Independent?

There is a table in the report for each of these seven classes recipients. 
There is also a mapping table which gives the geographical relationships 
between them. In most cases it is either one-to-one, or a hierarchy, but 
there are a small number (about 20%) of exceptions. If anybody can work 
out how the money flows in the hierarchical and the many-to-many cases I 
would be grateful.

There are also lower-tier local authorities, which I have not understood 
yet. I think these may be in table 1 along with the upper-tier ones. For 
example, I can find both Fenland and Cambridgeshire in table 1, despite 
the fact that Fenland is a lower-tier authority inside the Cambridgeshire 
upper-tier authority. In that case even table 1 alone suffers from 
double-counting; we need to separate it into the two separate levels.

Finally, I could do with some help working out how the various authorities 
are coded. I think many of the authorities are identified by standard 
codes, but I don't know what the coding systems are called or where I can 
get the complete lists of codes. If anybody happens to know those answers, 
I would be grateful for them.

I'm pleased to say that this data set has good documentation and cites its 
sources properly. I hope that by digging a bit deeper we can answer all 
these questions.

I've written up my notes so far here:

 	http://knowledgeforge.net/okfn/tasks/ticket/269

Please feel free to help.

 	Alistair