Lies, damn lies & statistics

Going Postal, Godfrey BloomFor over forty years I have struggled in the financial world to interpret statistics. People in public life love to flaunt statistics, graphs, numbers, percentages, usually to make a case for a pre judged outcome. The treasury are adept at this, the government want to carry out a given policy they then task the Treasury or Bank of England to justify it with statistics to bamboozle an unwary electorate.
Brexit, climate change, national debt, NHS spending, university admissions, gender gap pay, stats come out of every department in spades. They are nearly always flawed. Stats can prove anything & often do. Yet the subject is at least a two year field of college endeavour.
As an ex pro let me run a few tricks of the trade past you. They illustrate academic concepts such as reversion to the mean & a delicious phenomenon known as the Simpson’s paradox.
No graphics just some anecdotes for the layman (just about every politician & bureaucrat).

Speed cameras have sprouted like mushrooms over the last decade, statistics from camera salesmen prove they save lives. But do they?
Five years ago on a dual carriage way into a major Lincolnshire town two maniac teenagers acting criminally dangerously raced their stolen hot hatch backs at speeds of over seventy MPH. Tragically they killed two pedestrians in this insane act for which they were justly sent to prison. A sharp suited salesman persuaded a not over bright chief constable & even less bright local councillors that this tragedy could have been prevented by speed cameras. The very cleverly offered two cameras free for a year to prove their case. Incidentally road deaths in the area were very rare historically. A year passed, there were no further accidents on this stretch of road, indeed there had been no accidents in the prior ten years. The salesmen argued that many lives could be saved across the whole county, look how well they had done in just one place.
Hundreds were then installed at vast expense. Much money was raised for the camera manufacturers & installers. But the effect on that dual carriage way would have been the same had balloons been tied to lampposts. Reversion to the mean means just that, back to normal,  criminal teenage lunatics are not thankfully a regular occurrence & certainly would not be stopped by cameras.

Gender statistics are notoriously misleading. Usually rolled out by pressure groups to make a case for budget increases.
Pick your preference, most radiographers, nurses, physiotherapists & mid wives are female. Most orthopaedic surgeons are male. Put up the fancy coloured graphs & prove somehow women are discriminated against in the NHS from surgical appointments in some way. But check out a graph to find what men & women actually want to do. Google Simpson’s Paradox for further information. Fascinating stuff.

Time weighted statistics is another epic wheeze to sway the laymen in the interpretation of numbers.
A personal anecdote for your amusement. Twenty years ago I worked for a leading investment bank which specialised in pension fund management. My stats team, Cambridge lady maths graduates, awesomely clever, were always feeding me splendid ammunition. At the time my fund was a front runner along with that of Schroder, a very competitive house which I knew was bidding to the trustees. I wanted to beat them with my presentation, stuffed with graphs. Not easy, as the late Bruce Forsyth would say.
Incidentally I was not supposed to know the Schroder hat was in the ring, I had a chum on the inside, “give me the last three years girls”, you could say girls in those days, we lost by a whisker, ” give me the last three calendar years”, lost again. Blast, “give me the last three fiscal  years”, bingo, a win. This kind of legitimate fiddling has kept the global warming alarmists going for twenty years. Although sooner or later truth will out.

My recent book, Magic of Banking illustrates dozens of these little tricks used by banks, central banks in particular.



© Godfrey Bloom 2018

Audio file