[OpenSpending] Representing things in terms of other things?

Wed Apr 17 18:38:29 UTC 2013

On 17 April 2013 20:14, Jonathan Gray <jonathan.gray at okfn.org> wrote:
> This recent story [1] made me wonder whether we could do something for Open
> Spending (or even a separate little thing, related to OpenSpending) to show
> big spending items in terms of other smaller or similar spending items?

I have a major problem with that Guardian article, in that it seems
more to promote financial illiteracy, presumably to score some cheap
political points. It pretends to be adding clarity as to what £10m is,
but instead simply perpetuates (and strengthens) standard
misconceptions.

So, to take the first example, £10m wouldn't really "pay for" 322
nurses. The cost of employing a nurse (or indeed anyone) is
considerably more than their salary. A general rule of thumb is than
an employee will usually "cost" 2x their salary. Secondly, it ignores
the critical distinction between one-off costs and recurring costs.
Simply stating that it's "for one year" doesn't really address the
sleight of hand inherent in pretending that these costs are
equivalent.

But I'd still be concerned even if the information were presented
differently (e.g. "Could employ 20 nurses for 10 years"). This still
only gives the _impression_ of useful context, by cherry-picking a
single number (the cost of a nurse), and making that number look much
bigger than it really is. (And using an emotionally charged
example[1], to boot, but that's a slightly different rant.)

Most people will instinctively think of 20 extra nurses see it in a
hyper-local context — in terms of a hospital they've been at recently,
or with reference to someone they know who's recently lost their job,
or the like. But out of ~400,000 nurses employed by the NHS, 20 (or
even 322) makes very close to zero difference.

A more accurate way of expressing the same number in the same context
is that £10m could increase the median nurse salary from £31,095 to
£31,100 for five years. But as that suddenly makes the £10m seem
small, rather than large, I suspect it doesn't fit the narrative the
Guardian are trying to put across.

And that's only the problem with looking at £10m against the fraction
of government spending that is nurses' salaries. Out of the full
£720bn, £10m is even smaller still. And any attempt to 'translate' the
numbers into something that people can understand, but which doesn't
convey *that*, is hindering, rather than helping.

Lest anyone misunderstand — I very much agree with Jonathan's
underlying idea here. Contextualisation of information, and in
particular of big numbers, is hugely important. But it's so important
that it needs to be done right — and a lot of how it's being done
these days is very far from right.

I also have no objection to the Guardian having a narrative they want
to get across here. I probably even agree with what they're saying.
But I *do* object to manner in which they're doing it, and, more
crucially, I think the Open Knowledge Foundation, or those of us doing
related things in similar fields, need to be even more careful not to
fall into these same sorts of traps.

Tony

[1] Doubly so by the wider context being highly charged too.