This page outlines the methodology used for producing the 2000 car
theft index. The methodology was designed by the Home Office Policing
and Reducing Crime Unit, who contracted the data processing to the
Transport Research Laboratory.
The car theft index is based on thefts of cars and excludes
attempted thefts and thefts from cars.
The index is based on cars reported stolen in England, Scotland
and Wales which are recorded on the Police National Computer (PNC).
The PNC only contains details of unrecovered vehicles because once a
stolen vehicle is recovered it is deleted from the system. The data
used in the theft index were obtained from Experian Ltd who receive
PNC stolen vehicle data but do not delete records when vehicles are
recovered. An additional weighting factor was also built into the
theft figures to account for those recovered vehicles which were not
picked up by the data-set provided by Experian.
The time-frame for the stolen vehicle information was 1st January
1999 to 31st December 1999.
The vehicle parc (population of vehicles on the road) data were
collected by the Driver and Vehicle Licensing Agency (DVLA) and
obtained through the Society for Motor Manufacturers and Traders (SMMT).
This consisted of all vehicles licensed with DVLA on 31st December
Calculating the index
Number of vehicles reported stolen
X 1000 = Theft rate
vehicles licensed with DVLA
Using this formula, a theft rate was calculated for each model of
car by each year of registration between 1985 and 1999.
In order to produce a more readable document, the years of
registration were grouped into bands of three years from 1985
onwards, creating six age bands (Pre-1985, 1985-1987, 1988-1990
etc.). It should be noted that the year of registration is the year
of first registration where this information is available.
However, in a few cases vehicles have been re-registered and this
results in some vehicles which have not been produced for a number
of years having a recent date of registration.
Individual theft rates were grouped into three broad bands
indicating those which were �most at risk�, �medium risk�
and �lower risk�. These bands were produced by taking the 20th
and 80th percentile in the rates of theft. This means that 20% of
cars will be classified as �lower risk�, 60% as �medium risk�
and 20% as �most at risk�.
Using the 20th and 80th percentiles to distinguish between bands
produced the following cut off points in the theft rates for both
the 1998 and 1999 indices:
Higher risk = more than 21 cars stolen in every 1000 on the road.
Medium risk = between 5 and 21 cars stolen in every 1000 on the
Lower risk = up to 4 cars stolen in every 1000 on the road.
About the spreadsheets
The attached spreadsheets contain the data underlying the 2000 car
theft index. This includes the following information.
These spreadsheets are written in Microsoft Excel version 97 and
can be downloaded by clicking on the links above.