[wdmmg-discuss] "Unknowns" and Why is PESA better than CRA?

[David Jones]

Lisa, I think most of this is review, but the last half of the e-mail
could be worth bringing up with contacts at Treasury.

Moving forward from my earlier e-mail about COFOG classifiers for
education ( http://lists.okfn.org/pipermail/wdmmg-discuss/2010-June/000286.html
), and summarising the investigation I mostly did in ticket 402:
http://knowledgeforge.net/okfn/tasks/ticket/402

The current dashboard has some large "Unknown" bubbles (For example,
underneath Education (82bn) there is a 44bn Unknown (for 2008/09 tax
year)).  Pretty much most of these large Unknowns are the "LA data"
problem described in Alistair's earlier e-mail:
http://lists.okfn.org/pipermail/wdmmg-discuss/2010-April/000165.html
).

There are 2 key observations:

1) This data is better in PESA (than CRA);
2) The "Unknowns" disappear in more recent tax years (because they are
projections?  if so, why is that?).

A spreadsheet compares the "Social Protection" spending described in
the PESA and the CRA:
http://spreadsheets.google.com/ccc?key=tRWEg5jwKn88y91Efkpk_JQ&hl=en#gid=0

>From this we can see that the spending totals are roughly similar, but
the granularity of the CRA is less good than that of PESA.  Question
to ask the Treasury: Why is that? (and can Treasury make it better).

A specific example is Housing:

Housing under the Social Protection bubble is 20bn in the PESA
analysis (for 2008-09; this can be viewed at
http://www.wheredoesmymoneygo.org/_prototype/ ), and 3.097bn in the
CRA analysis ( http://www.wheredoesmymoneygo.org/dashboard/#/uk-bubble-chart/focus=10&year=2008-09
).  The difference, 17bn, is somewhere in the "Unknown" bubble, but
this is all the data marked as "LA data subfunction" in the CRA.

drj