Wednesday, 30 September 2009

Bad mathematics

I noticed some bad mathematics creeping into the States debate on 9 September 2009 (yes I do actually wade through them from time to time!) - see below.
This was a question and answer:

2.3.6  Deputy S. Pitman of St. Helier:

The Minister said that this rise in rents will not be affecting the poorest.  Could he clarify this for me because, as I understand it, those on income support have been given a 2.5 per cent rise and the next thing they hear is that their rent has gone up by 2.5 per cent.  These people are people who, through no fault of their own through illness, et cetera, are having to live off benefits.

Senator T.J. Le Main:
The housing component of low income and employment and social security has been increased by 2.5 per cent specifically to cover the increase of 2.5 per cent imposed by this Assembly upon the Housing Department, so those on low income should not be affected whatsoever in regard to this rental increase.

Now I have no idea what the housing rent is, or the income support component, but unless they are equal, this will lead to an increase in the difference between rent and income support because the same percentages do not automatically equate to the same value. The fact that the percentages are the same does not mean the absolute values will be.

To see this, suppose, for example, a rent of £100 per week, and an income support component of £50 per week. This becomes respectively £102.50 and £51.25, so that the difference has increased by £1.25 from £50 to £51.25.

On another mathematical matter, I heard a clip of Sean Power on BBC Radio Jersey about the need for more inspectors for non-States rental housing, saying there was only one part-time inspector, which seems a very bad situation.
I did a little calculation, and if an inspection takes an hour (including paperwork) - and probably more, and say an inspector works 4 hours a day, 5 days a week, that makes 20 properties a week, and allowing 50 weeks (probably less because of holidays), 1,000 properties inspected in a year. I don't know what the actual figures are, or the number of properties to be inspected, but that sounds very few.
Lastly, as I mentioned briefly in the "Guernsey Watch" posting, the figures on "maverick" bank staff giving "advice" about legal schemes to avoid paying tax are preposterous. The repeated mantra is that 99% of bank staff (or even 99.9%!) would not have given such advice to the Panorama journalist, because Jersey is well regulated and had a glowing report from the IMF.
Now I am all against baseless accusations, as for instance, about "hidden billions" in offshore bank accounts, for which there is about as much evidence as there is for the existence of the Higgs boson; in other words, theoretically possible, but evidentially weak. But in the case of Panorama, to believe a single reporter could go into a bank (of many), and find one employee (of many) who was a maverick, beggars belief.

Given the statistics of 99.9%, this is akin to picking a number in roulette wheel, and that number coming up. It does happen. But to happen in Guernsey (with the non-trading company dodge) and Jersey (with the Hong Kong paper transactions), is like picking a number and getting it right twice. For its reputation to be restored, Jersey needs not just training of staff to spot money laundering and suspicious transactions, but also courses in ethical fund management, so that the mindset of the employees who can come up with these schemes ("brainstorm" in the Jersey case) is significantly changed.

As Nick Spencer notes in the excellent booklet "Rebuilding Trust in Business":

A business system of enforced transparency, precisely articulated accountability, ongoing and detailed auditing, abundant legal safeguards, and slick PR campaigns is a valuable but expensive and ultimately ineffective substitute for the honesty, directness and integrity that breed and maintain trust. 'If we want to restore trust we need to reduce deception and lies rather than secrecy.

The task is not, however, to abandon these measures, which can and do contribute to a society's 'trust capital,' but rather to place them in their rightful context, that is, within a properly functioning, trust-building system. Current measures consist of more and more regulations to make a system work better; they are not measures to make a system work in the first place.

No comments: