|
Business Competitiveness Indicators |
||
|
The Department of Trade and Industry have been providing strategic indicators at the local area level since 1995. The Business Competitiveness Indicators (BCIs) were originally produced for (former) counties, Training and Enterprise Council (TEC) areas and Local Authorities. However the structure of the BCIs has evolved as geographic boundaries have changed and new boundaries have been introduced for the UK. The
indicator data provided in this edition of the BCIs reflect where
possible, the Unitary Authority and Local Authority District (UA / LAD)
boundaries drawn up between 1996 and 1998, and the Learning and Skills
Council (LSC) areas drawn up during 2000. The
estimates of Gross Value Added (GVA) included are based on Nomenclature of
Units for Territorial Statistics (NUTS) areas. These are a hierarchical
classification of geographies, which provide a breakdown of the European
Union's territory for producing regional statistics at levels that are
compatible across the Union. The
package of indicators, updated twice a year, is based on those provided in
the Regional Competitiveness and State of the Regions publication (RCSoR).
The RCSoR (www.dti.gov.uk/sd/rci) provide indicator data for the countries
and Government Office Regions of the UK, and are themselves published
biennially with hard copies available from the DTI Publications Unit
(email: [email protected]). The
indicators in this edition of the BCIs include:
With
the exception of GVA and business survival rates, the data included in the
BCIs are provided down to UA/LAD level. The UAs and LADs of Great Britain
vary widely in both physical size and resident population ranging from
Birmingham with a population of over 1 million to the Western Isles of
Scotland, with a total population of less than 30 thousand people. With
the exception of the claimant count, all of the indicators included in the
BCIs are based on sample surveys. As sample sizes are, in the main,
proportionate to population, the UA/LAD level estimates may vary widely in
their accuracy. In particular, care should be taken when interpreting year
- on - year changes since, due to relatively small sample sizes in certain
UAs and LADs, these would not be statistically significant in some
instances.
MEASURE The number of new VAT
registrations and de-registrations in the (calendar) year, and the number
of registrations as a proportion of the total number of VAT-registered
businesses at the start of the year. SOURCE These data are produced
by the Small Business Service (SBS) Statistics Unit www.sbs.gov.uk, based
on data collected from the Inter Departmental Business Register (IDBR). The IDBR is primarily a
sampling frame for a number of official business surveys. The Register
holds information on the location, industrial classification, turnover and
employment of over 1.9 million enterprises in the UK. TIMING Formation rates and
levels down to UA / LAD level are presently available to 2001. ISSUES New businesses with a
turnover below the mandatory VAT registration threshold (of £54,000
during 2001) may decide not to register and would not appear in the
estimates. The mandatory VAT
threshold (introduced in 1973) is normally reviewed on an annual basis.
Since December 1993, this threshold has increased by £1,000 per year.
Prior to 1993 the increases varied between £10,000 (in 1991) to £1,600
(in 1992). Due to this, there may be discontinuity in the VAT
registrations and de-registrations series prior to 1993. The registrations and
de-registrations data included in the BCIs do not equate to ‘actual’
business closures or start - ups. Firms can be removed from the VAT
register for a variety of reasons including; falling turnover, mergers,
takeovers and relocations as well as the business actually ceasing to
trade. However, VAT registrations and de-registrations are indicative of
the underlying rate of business ‘births’ and ‘deaths’. In Great Britain it is
estimated that the total number of business start-ups is around twice the
number of VAT registrations, with an estimated 530,000 new businesses
formed between the years 1995 and 1999. Comparisons of business
formation rates between areas should be interpreted carefully.
In areas where the stock of businesses is relatively low, the rate
of business formation will be artificially inflated. Also, the rate of business formation can increase at a time
when VAT registrations are falling if the stock of VAT registered
businesses is decreasing at a faster rate. QUESTIONS AND ANSWERS Q.
How many supposedly ‘new’ business formations are the result of
existing businesses relocating or expanding rather than new start-ups? A.
The estimates presented here refer to the number of businesses,
business registrations and business de-registrations within each UA/LAD
and LSC area. That is, the
area where the VAT registered address (often the address of the head
office) is located. For
example, a new factory or outlet owned by a business registered in another
local area or LSC would not appear as a new registration. Q.
How much is the business formation rate a result of the mix of
industries within an LSC or UA/LAD area? A.
Some sectors are known to have higher start-up rates than others
(for example finance and business services) or tend to vary more than
average across the course of recent economic cycles (for example
construction and some manufacturing industries).
A high start-up rate can be due in part to concentrations of such
units. Q.
Do start-ups represent new business formations or simply a
redistribution of existing business stock? A.
An increase in start-ups may reflect genuine expansion.
Equally the trend for large firms to contract out many activities
to smaller businesses and the self-employed is also likely to stimulate
start-ups. Changes in ownership and / or location as well as mergers
reflect the redistribution of business stock. This is more common in
certain sectors (for example, retail and catering).
High levels of displacement and change of ownership or location
will inflate deregistration as well as registration rates differentially
across local areas. CONTACT: MEASURE The proportion of
businesses remaining registered for VAT at set periods after their initial
registration. SOURCE These estimates are
produced by the Small Business Service based on IDBR information. TIMING Estimates are normally
available in January each year. The estimates included in the October 2003
BCIs include both the one- and three–year survival rates of
VAT-registered business to 2000. ISSUES Refer to the issues
detailed in business formations. These also apply to business survival
rates. QUESTIONS AND ANSWERS Q.
How long do businesses with turnover below the VAT threshold
survive? A.
The limited evidence available suggests that those businesses that
are below the VAT threshold, of which there are around 2 million, have the
lowest survival rates. This
implies that VAT survival rate estimates could potentially over-estimate
the actual survival rates for all businesses in an area. Business
insolvency information, also published by DTI (www.dti.gov.uk/sd/insolv)
may provide a broader picture of this. Q.
To what degree is a high or low survival rate due to a particular
sector mix? A.
Certain industrial sectors are thought to have higher survival
rates than others, on average. A high or low survival rate within an LSC
or UA/LAD area could result from the industry composition of the area. It
is often informative to compare the mix of industries within an LSC or UA/LAD
area to that of England or the UK as a whole when interpreting these
survival rates. Q.
To what degree is a high or low survival rate related to the
‘size’ of VAT registered businesses in an area? A.
Among VAT-registered businesses, the tendency is for larger
businesses (typically those with 50 employees or more) to have higher
survival rates than smaller ones. Consequently,
if there are a high proportion of ‘larger’ businesses in an area it is
reasonable to expect that survival rates may be above the average. Again,
the size profile of the businesses within an LSC or UA/LAD area can be
compared to that for England or the UK as a whole when interpreting these
survival rates. CONTACT:
NOTE The estimates published
in this edition of the BCIs have been calculated on the basis of the
latest European System of Accounts (ESA95). Under this system the term
‘Gross Value Added’ (GVA) is now used to denote estimates that were
termed Gross Domestic Product (GDP) in previous editions. These estimates are
provided for Nomenclature of Units for Territorial Statistics (NUTS)
areas. These are a set of standardised geographic classifications used by
European Union member states, which were last revised in the UK during
1998. There are 133 NUTS level 3 areas in the UK. These areas are
equivalent to individual counties, UAs, or groups of UAs and LADs. SOURCE Regional GVA are drawn
from the regional economic accounts produced by Regional and Local
Division in ONS. TIMING The latest estimates
available run to 1998. ONS aim to release more recent estimates of GVA
below the regional level during late 2003, which will include detail for
individual industry sections. ISSUES The estimates of GVA
included in the BCIs are ‘workplace–based’. That is the earnings of
employees are allocated to the area of their workplace rather than the
area where they live in those cases where the two are different. The GVA provided are
valued at current prices, that is, the estimates are not adjusted to take
account of inflation in the prices of goods and services over time. ONS do not directly
produce estimates of GVA for LSC or LAD areas, although there are a number
of NUTS level 3 areas that correspond to UAs for which estimates are
produced. ONS produces two separate
indicators of economic activity below the regional level which fall under
the broad heading of ‘GVA’. The data included here are drawn from the
regional economic accounts that are calculated based on a wide range of
economic and labour market survey data. These data conform to the ESA95
framework for GVA agreed between the EU member states. ONS also produce a
separate GVA series based on a single survey, the Annual Business Inquiry
(ABI). The latter GVA estimates cover a far narrower range of components
than those covered in the former series and are also available for a far
shorter time series at the present time. QUESTIONS AND ANSWERS Q. What is the difference
between Gross Value Added (GVA) and Gross Domestic Product (GDP)? A.
GDP is equal to GVA plus taxes (net of subsidies) on products, with
Value Added Tax (V.A.T.) the major component of these. For the UK as a
whole, taxes (less subsidies) on products amounted to around £110 billion
during 2001, just less than 12 per cent of total GDP. Q.
How does commuting affect this form of GVA per head in NUTS level 3
areas? A.
Workplace–based estimates of GVA per head are calculated by
dividing the estimate of workplace–based GVA for a NUTS level 3 area by
the resident population within that particular area. Areas with a low
resident population and high levels of inward commuting (such as Inner
London–West) will have substantively higher levels of GVA per head.
Conversely, areas with high levels of outward–commuting of workers and a
high residential population (such as Berkshire) may have reduced estimates
of GVA per head. Q.
How is GVA estimated below national level? A. At the regional level
the Office for National Statistics produces income-based estimates of
regional and sub-regional GVA. The income approach measures the income of
the area, defined as the sum of all incomes earned from productive
activity carried out within that area. The figures included in this
publication are consistent with the national estimates of GVA published in
the United Kingdom National Accounts (Blue Book). Q.
What are the key components of income-based GVA? A.
GVA at both the national and sub-national level is comprised of a
series both actual and theoretical (or imputed) components. The three
largest components are: Compensation of Employees (basically all employee
earnings), the income from both labour and profits of sole trader
businesses (termed mixed income), and the gross operating surplus of
incorporated enterprises (basically the surplus, or profit, arising from
the production of goods and / or services). Q.
Does GVA per head measure the wealth of the residents within an
area? A.
The GVA included in the BCIs is workplace-based and, as such, says
little about the areas to which the income that is generated flows. Due to
this, it is possible for an area to have high GVA per head but conversely,
a low level of residential income. At
sub-national level, it is more useful to use the household sector income
estimates produced by ONS when investigating the income received by the
residents of an area. CONTACT:
MEASURE The BCIs provide
employment data on the basis of the areas where workers live and where
they work (in cases where the two are different). Three separate sets of
estimates are provided for UA/LADs and LSCs, drawn from the quarterly and
annual Labour Force Survey (LFS), and the Annual Business Inquiry (ABI)
employee jobs series. SOURCE AND TIMING Two sets of jobs data on
a residential basis are provided in the BCIs. These are taken from the
quarterly and annual LFS series respectively. Quarterly results are (under
normal circumstances) published four months after the end of the quarter
to which they refer, and are included up to winter 2001/02. Annual LFS
data are taken from the Local Area Database and cover years to 2001. The workplace-based
employee jobs estimates are taken from the ABI. The ABI measure contains a
number of structural and methodological improvements over its predecessor,
the Annual Employment Survey (AES). Nationally, the number of employee
jobs as measured by the ABI was around 1 million higher in 1998 than the
AES level because of some undercounting and estimation problems with the
AES. ISSUES The LFS and population
estimates used in this edition of the BCIs are not yet consistent with the
latest 2001 Census results. LFS estimates for all available years will be
reweighted to the 2001 Census population level by the Office for National
Statistics in due course. The LFS measures
employment among residents within each UA/LAD or LSC. This is appropriate
if a high proportion of residents work within the area, but not so
appropriate if there is a high level of commuting to work outside of that
area (such as in many of the UA/LAD areas which surround London). The ABI/1 measures the
number of jobs within a given area (as distinct from the number of
workers). If a person holds more than one PAYE job and earns wages above
the taxation threshold in each, then they could be counted more than once
in the ABI/1 estimates. From 2000 onwards, the
annual LFS estimates included in the BCIs incorporate data from the
English Local Labour Force Survey (ELLFS). The ELLFS is a partnership
project between the ONS, DfES, and the DWP aimed at enhancing the sample
size of the existing LFS in England. When combined with annual LFS results
(for 2000), the inclusion of ELLFS results has led to a boost of around 40
per cent in the number of households sampled. This means that the annual
LFS estimates supplied in the BCIs should be interpreted with care as the
increase in sample size arising from the inclusion of the ELLFS for 2000
and some adjustments to the estimation methodology for the latest year
does mean that discontinuities could exist in the annual data between 2000
and earlier years. The LFS annual and
quarterly series include both employees and self-employed persons. The ABI/1
employee jobs series does not include self-employed people. Other factors may also
lead to differences between the LFS and ABI, for example, commuting
effects or the recording of workers with more than one job. The LFS is a quarterly
sample survey of around 60,000 households across the UK. In some cases the
estimates for certain UA/LAD areas may be based on relatively small sample
sizes and therefore contain higher proportions of sampling error than
those based on larger sample sizes. In particular, caution should be taken
when interpreting changes between either quarters or years at UA/LAD
level, as in some cases changes would need to be around 8 to 10 per cent
of the original estimate in order to be deemed to be statistically
significant. The quarterly LFS data in
the BCIs are not adjusted to account for seasonal variations in
employment. Caution should be used when comparing consecutive quarters as
the effects of seasonal variations on employment can often obscure changes
caused by more fundamental factors (such as changes in the structure of
the industry within an area over time). ONS recognises that in
certain years, the AES undercounted the level of employee jobs because of
issues with both the data collection and estimation procedures. The total
undercounting was thought to be around 1 million jobs in 1998.
Furthermore, it is believed that this undercounting affected some
industries more than others. With the move to the full ABI/1 survey for
1998 onwards, the 1995 to 1997 AES levels were adjusted upwards to
compensate for this undercounting using industry-specific adjustment
factors. In certain years
sub-regional areas may not sum to the respective regional totals. This is
because estimates of employment in the agricultural industries (Standard
Industrial Classification 1992, Division 01) are suppressed in certain
areas due to disclosure issues. QUESTIONS AND ANSWERS Q.
What are the differences between workplace- and residence-based
estimates? A.
The estimates provided for UA/LAD and LSC areas from the LFS are
residence-based which is appropriate if most people work within the area
the area that they reside. However, at UA/LAD level this is often not the
case. The ABI/1 estimates are
workplace-based and indicate the number employee jobs within an area,
irrespective of where the people who do them live. Q.
Are the employee jobs indicators included in the BCIs influenced by
the time of year? A.
The quarterly LFS employee jobs indicator provided in the BCIs are
not seasonally adjusted (i.e. there are no adjustments made to the
estimates to compensate for ebbs and flows in industries such as tourism
which are affected by the time of year). The ABI/1 estimates included in
the BCIs now refer to December of each year and because of this, there may
be seasonal variations in the number of jobs between certain areas,
especially below the regional level. For example, a UA/LAD areas’ share
of total regional jobs may vary throughout the course of the year with
fluctuations in certain seasonal industries. Q.
How accurate are the estimates of employment? A.
Both the ABI and LFS are sampl –based surveys rather than a full
census of either businesses or households and will be subject to sampling
error. In other words, had different selections of business or households
been used, the results would almost certainly have been different. CONTACT: MEASURE Average hourly earnings,
(including overtime and premium pay), for full-time employees only. SOURCE AND TIMING The New Earnings Survey (NES)
is sampled during April each year. Initial results are normally available
the following November, with the inclusion of late returns around 4 to 6
months after this. Provisional UA/LAD and LSC estimates for 2001 are
included in this edition of the BCIs. ISSUES Businesses in NES are
sampled directly and the estimates are workplace-based. The boundaries of the
geographies for which NES earnings results are presented in the BCIs may
change slightly over time as the boundaries of the electoral wards, from
which these geographies are comprised, are amended. Since 1991, ward
boundaries have been revised for the years 1993, 1995, 1997 and 1998. As
such, the estimates provide a ‘snapshot’ of the state of a particular
area as it stood in each year, rather than providing a strictly comparable
geographic series over time. Comparisons of the value
of hourly earnings both between areas and across time should be
interpreted with caution. NES estimates of earnings do not take account of
variations between areas in the cost of living, and as such do not
represent the ‘actual’ buying power of the earnings. No Retail Price
Indices are produced below the national level in the UK. In some instances the NES
estimates for certain UA/LAD areas may be based on relatively small sample
sizes and could contain a high proportion of sampling error. Care should
be taken when interpreting estimates of average earnings and in
particular, changes between years at this level. The estimates of average
hourly earnings for 2000 included in this edition of the BCIs have been
revised since the last release in December 2001. When provisional earnings
estimates for 2000 were published in October 2001, there were a higher
than usual number of responses not included in the survey as the returns
were awaiting validation. These have now been incorporated into the
revised 2000 results, with the most notable amendments arising from the
inclusion of data for high earners in the financial, business services,
media and marketing sectors. This has led to revisions in average pay, in
particular, for areas within London and the South East Government Office
Regions where there are concentrations of jobs in these industries. QUESTIONS AND ANSWERS Q.
Are high average earnings a good thing? A. Not always.
Interpretation of earnings data is quite complex and requires a
comprehensive knowledge of the local economy.
High earnings can be indicative of prosperity, and that an area’s
economy is growing. Alternatively, they could be a result of skill shortages or
high living costs both of which could undermine the competitiveness or the
attractiveness of an area to outside investors. Q.
Why are the earnings of part–time employees not included in the
BCI estimates? A.
Due to the sampling design of the NES, in certain UAs and LADs the
proportion of part–time employees can vary quite substantially between
years. In addition to this, the NES can pick up very high-earning
part-time employees (for example, non–executive company directors). In
areas where sample sizes are relatively low, both of these factors can
make the earnings of all employees very volatile and less representative
of the area as a whole than the information for full-time employees alone. Q.
Would the hourly earnings estimates be affected by the time of year
that they are collected? A.
Yes. NES results are collected in April of each year. Although
information gathered at this point in time may be influenced less by
fluctuations in earnings within the known seasonal industries (i.e. retail
and tourism), for many businesses the financial year ends in April. This
could affect some of the components of employee earnings, particularly in
the case of the earnings from overtime that are available. CONTACT: MEASURE Unemployment rates are
based on the International Labour Organisation
(ILO) definition and are drawn from the Labour Force Survey. SOURCE AND TIMING The ILO definition of
unemployment includes as unemployed; all persons out of work, who want a
job, have actively sought work in the last four weeks and are available to
begin employment in the next two weeks. Unemployment rates for UA/LADs
are modelled estimates based on (annual) Labour Force Survey information
in addition to a range of other economic and demographic variables. ISSUES Unemployment rates are
residence-based. They are calculated as the number of people resident in
an area who are classified as unemployed divided by the number of
economically active people resident in that area. The
LFS and used in this edition of the BCIs are not yet consistent with the
latest 2001 Census results. LFS estimates for all available years will be
reweighted to the 2001 Census population level by the Office for National
Statistics in due course. The ONS have not produced
modelled estimates of ILO unemployment for LSC areas or regions. This
publication only contains UA/LAD level estimates. CONTACT:
The Claimant Count is the
number of people claiming Job Seeker’s Allowance and is taken from
monthly records. SOURCE AND TIMING The claimant count (CC)
is an administrative count of people claiming Job Seekers Allowance at an
Employment Service office. QUESTIONS AND ANSWERS Q.
Are claimant count rates workplace- or residence-based based? A.
Following a recent labour market review, the denominator for
calculating claimant count rates at a sub-regional level has moved from
being workplace-based (i.e. expressed as a proportion of the number of
workforce jobs plus claimants in that area), to being residence-based (the
number of people living in an area). CONTACT:
MEASURE The number of people
within families that are dependant on means-tested Income Support (IS)
benefits by UA/LAD and LSC areas. SOURCE AND TIMING The data used are counts
of IS dependent persons originally supplied by the Department of Work and
Pensions (DWP) as well as estimates of the population as at midyear (1998)
developed by Oxford University. The numbers of IS
dependants have been drawn from the Income Strand of the ward level
Indices of Deprivation and the Index of Multiple Deprivation 2000 ('IMD
2000'). These contain six separate domains: -
Income ISSUES These analyses deal with
the percentage of people dependant on IS benefits and not the value of the
IS benefits claimed in each area. Whilst
IS dependants may occur with some frequency in the majority of the wards,
it may well be that the average value claimed in the most deprived wards
is higher than in the less deprived. This could mean that the difference
between areas may be greater than is indicated here. The
Indices of Deprivation although updated periodically (for example the last
two updates were for 1998 and subsequently during 2000), they are not
necessarily consistent over time. With every update of the Indices of
Deprivation, new indicators are added and the latest geographical
classification of wards are used QUESTIONS AND ANSWERS Q:
How often are the Indices of Deprivation updated? A: The Indices of
Deprivation (ID) are a periodic series that are generally published every
2 to 3 years after a series of user consultations. Development of the ID
2000 began in 1998 with estimates published during 2000. Consultation
began during 2002 to update ID 2000. Q.
What geographic level is the ID 2000 published for? A.
ID 2000 is published for electoral wards with boundaries as at
1998. For the purpose of the analyses in the BCIs these wards have been
aggregated to current Unitary / Local Authority as well as Learning and
Skills Council areas. Q.
What does the information taken from the Income Strand of ID 2000
measure? A.
The information used in the BCIs measures proportion of people
within families that are dependent on IS benefits (for example, because
the principal wage earner receives a very low wage or is unable to work
because of disability) As such, this only reflects one facet of
deprivation. It is preferable to interpret these estimates alongside other
sources of economic and demographic information such as: unemployment
rates, average earnings, vacancies and skills, household income, mortality
and illness breakdowns, homelessness, and housing condition information. CONTACT:
MEASURE The percentage of the
working age population (16 to 64/59) with qualifications to either
National Vocational Qualification (NVQ) level 1 or 2/equivalent, NVQ level
3 or 4 or a trade apprenticeship or with no formal qualifications. SOURCE AND TIMING Both the skills and working age population estimates
have been drawn from the Annual Local Area Labour Force Survey (LALFS).
The LALFS sample is taken from the 1st and 5th waves of quarterly
Labour Force Survey to provide the largest non-overlapping sample. For
1999/2000 LALFS, the non-overlapping sample amounted to approximately
95,000 households. More detailed breakdowns
of these skills information are available from this source. However, due
to relatively small sample sizes in some UA/LAD areas the estimates in the
BCIs have been combined into 3 broad groups. This is to ensure that
results meet the National Statistics LFS publication thresholds for the
majority of UA/LAD areas. ISSUES For 2000/2001 onwards the
existing LALFS survey sample in England was boosted to ensure that the
total sample size for local education authorities outside of London would
reach a minimum sample size in each year of 875 economically active
adults. This has led to a boost in England of around 40,000 economically
active adults from 2000 onwards. This boost means that the skills
estimates for the years 2000 and onwards may not be comparable with the
annual results for previous years. The boosted LALFS samples for 2000 and
onwards are weighted differently to previous years and will have different
levels of precision to the estimates for earlier years. Caution should
therefore be used when comparing the estimates for 2000 or later with
earlier years. The LFS and population
estimates used in this edition of the BCIs are not yet consistent with the
latest 2001 Census results. LFS estimates for all available years will be
reweighted to the 2001 Census population level by the Office for National
Statistics in due course. QUESTIONS AND ANSWERS Q. What does the
breakdown of skills within a particular area tell us about the
characteristics of jobs there? A.
Low skill levels among the residents of an area can in some cases
reveal a general low skill requirement /low quality profile of jobs in
that area. However, it can also be indicative of a mismatch between the
skills of the residents and the requirements of the jobs within that area.
For example, some UA/LAD areas (and in particular those that encompass
central or outer – central city areas) can have a relatively low
proportion of qualified residents whilst also having a high percentage of
jobs with high skill requirements located there. In such cases, these jobs
are filled by more highly skilled workers who commute in from neighbouring
areas. Further information covering commuting patterns and skills will be
available from the 2001 Census during late 2003 or early 2004. Q.
What else does a low proportion of adults with qualifications
within an area suggest? A.
A low percentage of adults with formal qualifications living within
a particular area can be indicative of generally low skill levels. As
mentioned, it can also be indicative of low skill requirements of jobs in
that area. However, an area with a low proportion of adults with formal
qualifications does not necessarily mean that the standard of education in
that area is poor. Low proportions of adults with qualifications living in
an area can often say as much about the economic ‘pull’ on people with
qualifications (for example through higher earnings) of neighbouring areas
as it does about the standard of education in that area. CONTACT: |
||