ABSTRACT. This paper utilises some data from a survey of
recent VAT registrations in the North of England to examine
two issues: first, the extent to which registrants are involved
in setting up entirely new businesses; and secondly, the
relationship between the date of registration and the start
of trading. The paper suggests that some caution should be
exercised in the use of VAT registration statistics in the
analysis of firm births.
Over the past decade or so numerous studies have
examined the determinants of variations in U.K.
firm birth rates across geographical areas, indus-
trial sectors and over time. (For a good review of
these studies and for recent empirical work, see
et al., 1993; see also Storey, 1994, pp.
49–77.) Work is now also being done on the effects
of firm births on employment and other economic
indicators (eg Ashcroft and Love, 1994; Johnson
and Parker, 1994, 1996).
An increasingly utilised data source on firms
births in the U.K. is now the VAT registration data
produced by Customs and Excise and analysed by
the Department of Trade and Industry.
time series analyses of these data include those by
Black et al. (1992); Keeble et al. (1993); and
Robson (1994). Spatial variations in VAT regis-
trations have been examined by Ashcroft et al.
(1991); Westhead and Moyes (1992); Keeble et al.
(1993); Hart and Gudgin (1994); and Johnson and
Parker (1994, 1996). These registration data have
the advantage that they are readily available and
provide comprehensive spatial and industrial
coverage of registrations. Consistent data are
available back to 1980.
The disadvantages of VAT registration statistics
– which result from a taxation requirement
imposed on businesses rather than from the needs
of applied economic research – as a measure of
firm births are well rehearsed (see for example
Daley, 1990; Storey, 1994, pp. 50–51). Firms are
not required to register, although they may do so,
until they reach the threshold level of annual
turnover, currently £46,000. Thus many very small
firms are excluded from the VAT data: Bannock
and Partners (1989) estimated that in 1986, only
60 per cent of all firms were registered for VAT.
The VAT threshold has changed over time,
although during the 1980s, it moved broadly in
line with inflation.
Registration may sometimes
result from a business reorganisation or a change
of ownership. For all these reasons, Storey has
rightly concluded (1994, p. 51) that ‘In only the
broadest sense . . . can the number of businesses
which are newly registered for VAT . . . be
regarded as an indicator of the number of new
business starts in any particular year’. Never-
theless, Keeble and Walker (1994) have argued
that, at least in respect of spatial analyses, the data
‘. . . represents the most up-to-date, comprehen-
sive, reasonably long term and spatially disaggre-
gated data source currently available . . .’.
Investigators using the VAT registration data
have usually been careful to point out their limi-
tations, but have been unable to indicate how
serious they are. To do so would require direct
contact with the registrants involved. Such contact
is extremely costly in terms of research resources.
This paper sheds further light on the data by
utilising some data derived from a survey of VAT
registrations carried out by the authors in the
How Good Are the U.K. VAT Registration
Data at Measuring Firm Births?
Small Business Economics 9: 403–409, 1997.
1997 Kluwer Academic Publishers. Printed in the Netherlands.
Final version accepted on November 11, 1995
Department of Economics
University of Durham
23–26 Old Elvet