Modelling the Effects of Public Support to Small Firms in the UK - paradise Gained?
Sections (a) to (e) were included primarily as factors, which might influence business
performance. Section (f) related primarily to the probability that firms would have
received BL assistance. Section (g) relating company performance was collected
largely to verify the information available on the original Impact Indicators database.
Methodological Considerations
There are a number of key methodological issues that need to be investigated before
moving to the modelling of 'selection' and 'assistance' effects using both the Impact
Indicators database and the additional survey of assisted and non-assisted businesses.
The most important is the issue of the attrition of the original numbers of assisted and
non-assisted businesses which may yield insufficient numbers in the final database to
undertake the econometric analysis. The original study aimed to find 12 matched non-
assisted businesses per BL assisted businesses, on the basis of the matching criteria
used, in order to have available, after attrition effects, at least 3 non-assisted
businesses per assisted business by the time of the current study. Second, to what
extent the remaining businesses in the database are valid in terms of their compliance
to the original definitions of assisted and non-assisted comparators.
Attrition Effects
On the basis of the a preceding feasibility study commissioned by the DTI it was
determined that sample attrition would result from the following factors:
• Increased penetration by BLs of their target market - comparator non-assisted
business at the time of initial checking could be lost through BLs providing
significant assistance to them.
• Inability to match controls on BL databases The comparator group was
therefore to be limited to those businesses which could be located in both
databases
• Closure/acquisition of either Assisted or Comparator non-assisted businesses
- it was recognized that losses of both assisted and non-assisted businesses
could arise from the death of or acquisition of businesses in either group.
• Indirect losses due to assisted or the non-assisted comparator business
dropping out, leading to losses of their associated matched business.
Stephen Roper and Mark Hart