Modelling the Effects of Public Support to Small Firms in the UK - paradise Gained?
basic information on VAT registered businesses and PAYE returners. The period
covered by the study were the financial years 1993/94 to 1999/2000. Business Links
located and operational within the three TEC clusters ‘Industrial City’, ‘London and
Birmingham’ and ‘Rural areas’ were censured to provide the details of businesses that
matched the definition of ‘significant assistance’. Advice was sought from both an
academic panel and Business Link practitioners on what constituted significant
assistance. The definition agreed for the purpose of this study was:
Firms that had been assisted for the first time in the period January to
May 1996 who minimally had two visits/meetings with an advisor and had
a third visit by the November 1996.
On the basis of a feasibility study (unpublished) it was determined that each business
assisted needed a minimum of 12 matched comparators to ensure a viable sample of
assisted and unassisted firms at the end of study period. The matching was
undertaken on the basis of sector, employment size and geographical cluster.
It was expected there would be a high attrition rate due to comparators disappearing
as a result of normal business "events" (failure, takeover) and contamination
(receiving assistance from their local Business Link). Firms were matched on the
following basis by the Office for National Statistics and were subsequently checked in
the field as part of the verification of both assisted and comparator firms.
Enhanced Research Model
The SBRC, NIERC and Prism Research supplemented the basic dataset with survey
data to control for firm level factors likely to influence their growth projectory.
Without the additional in-company information obtained from the survey of assisted
and non-assisted businesses it would not have been possible to accurately isolate the
effects of 'selection' and 'assistance' on those assisted businesses that had grown faster
than non-assisted businesses. For example, where it is found that assisted small
businesses grew faster than non-assisted businesses it is not clear whether their faster
growth reflects:
• the benefits of assistance;
• a tendency for faster growing firms to be keener to apply for assistance;
• or, whether assistance was successfully targeted on faster growing firms.
The enhanced methodology examined the effect of business support on a range of
performance indicators and, using an approach (using selection models adopted from
Bates, 1995) to identify separately the ‘selection’ and ‘assistance’ elements of the
performance differential between assisted and non-assisted firms. Their proposed
approach endeavours to address the following two problems associated with the
evaluation of the desired impact on assisted firms of BL services:
• typically no differentiation is made between the types of assistance which
firms in the assisted group may have received. It is therefore not possible from
these studies to compare the relative benefits of different types of assistance.
Stephen Roper and Mark Hart