Left Watch

« The Lib Dems draw their fantasy constituency map for a general election fought under Single Transferable Vote | Main | Jon Cruddas and the Labour leadership race »

The real statistical crime is that Labour is making it harder and harder for anyone to compare government performance over time

N691659668_6651 Over on ToryDiary Jonathan Isaby writes about the attacks on Chris Grayling's use of crime statistics. In this piece for LeftWatch, Natalie Elphicke suggests there is a wider problem of statistical collection methods being changed so that a government's performance over time cannot be measured.  We already know that many statistics on marriage are no longer being collected as the state attempts to hide the importance of marriage to child welfare. Treasury forecasts and data standards have also become highly questionable. Natalie Elphicke, a member of the Candidate Lists, housing expert and CPS author, asks:

What are the national statistics for?

Surely we must be able to assess the performance of the country and the Government over time?

The furore over the Conservative Party’s use of crime statistics demonstrates that national statistics are not fit for purpose.

It is not just in crime that statistical changes have occurred so as to make it difficult to compare "apples with apples".   The statistics are not there to present a historic record or for an interesting academic exercise.  They are collated for the use and information of government, and its accountability to the people.

Take my own area of research in housing repossessions.  In February 2009 I published a paper with the Centre for Policy Studies, "Saving 100,000 Homes".  It analysed the 1990s recession and recommended individual solutions to arrears and repossessions. As opposed to expensive and ineffective grandstanding state schemes of the type Labour are implementing. By August 2009 the statistics had been changed to make a direct comparison to the 1990s recession statistically impossible.

The effect of this change?

ConHome readers can see for themselve in the Second Statistical Bulletin of 2009 from which I take this extract (my underlining):

"The approximate numeric effect of changing the definition of the published ‘orders’ statistics is as follows.
  • Between 1999 and 2001, the new measure was around 5% below the old measure for mortgage orders and around 2% below the old measure for landlord possession orders.
  • Between 2002 and 2006, the new measure was around 3% below the old measure for both mortgage and landlord possession orders.
  • From 2007 onwards (following the roll-out of PCOL and associated changes in court administrative procedures), the new measure has been between 4% and 5% below the old measure for mortgage possession orders, and between 7% and 13% below the old measure for landlord possession orders.
Prior to 1999 the possession orders count is provided on the old basis as it cannot be calculated using the new measure."

What does the change in the statistics mean?
(1) The Government looks a lot better.  By sheer numerical chance it seems the largest downward adjustment to the figures takes place just when the repossessions numbers get proper scrutiny.  So just how independent are government statistics and how much can we trust them?  Particularly as, in the case of these crime figures, as Shadow Home Secretary Chris Grayling MP points out, they have been used for years and years by all parties. Isn't it strange that Sir Michael Scholar chooses this moment, and this key issue of public concern, to make a fuss.

(2) Governments are accountable to the people for what works, and what doesn't.    That is part of leadership.  Accepting responsibility for what goes wrong, not just taking the glory for what goes right.  When Conservatives get into Government, the temptation of statistical manipulation is one which must be resisted. If we are serious about changing our broken politics and about accountability we do need national statistics to be independent and reliable.  This is an essential part of the need to rebuild trust and honesty.