A Regional Analysis of Productivity in Business Services: Implications for the UK Knowledge Economy Joanne Roberts and Andrew Hunt
Jan 17, 2016
A Regional Analysis of Productivity in Business Services: Implications for the UK Knowledge Economy
Joanne Roberts and Andrew Hunt
Background and Purpose of the paper• Exploratory in nature
• Reporting findings of an initial analysis
• Investigates difference in productivity levels in the UK business services sector by region and sub-sector.
• Reviews GVA data and GVA per head
• Employs the econometric technique of stochastic production functions to examine efficiency in selected business service sectors.
• Identifies directions for further development of the research
• Broad aim to contribute to a deeper appreciation of UK business services
• In so doing, to offer insights into an important component of national and regional knowledge economies.
Table 1. UK Standard Industrial Classification of Economic Activities 2003, Section K: real estate, renting and business activities
• Division Group Description• 70 Real estate activities•• 71 Renting of machinery and equipment without operator and of personal and household
goods•• 72 Computer and related activities• 72.1 Hardware consultancy• 72.2 Software consultancy and supply• 72.3 Data processing• 72.4 Database activities• 72.5 Maintenance and repair of office, accounting and computing machinery• 72.6 Other computer related activities•• 73 Research and development• 73.1 Research and experimental development on natural sciences and engineering• 73.2 Research and experimental development on social sciences and humanities•• 74 Other business activities• 74.1 Legal, accounting, book-keeping and auditing activities; tax consultancy; market research and
public opinion polling; business and management consultancy; holdings• 74.2 Architectural and engineering activities and related technical consultancy• 74.3 Technical testing and analysis• 74.4 Advertising• 74.5 Labour recruitment and provision of personnel• 74.6 Investigation and security activities• 74.7 Industrial cleaning• 74.8 Miscellaneous business activities not elsewhere classified
Source: Extract from UK SIC 2003, available from the Office of National Statistics,
http://www.statistics.gov.uk/methods_quality/sic/contents.asp (assessed 15th July 2008).
Significance of Business Services• Engage in the creation, dissemination and application of
knowledge both within and between firms at the level of the region, the nation and internationally (Miles et al., 1995; Antonelli, 1999; Andersen et al., 2000).
• Contribute to the competitiveness of enterprises generally - an important factor driving long-term growth (OECD, 1999; CEC, 1998, 2001).
• Among the EU 15 employment in the sector has grown at a rate of 4.4 per cent per annum between 1979 and 2003 (Kox and Rubalcaba, 2007 p. 6 & 7)
• Total employment level of over 19 million or 11.4 per cent of total EU 15 employment by 2003 (ibid)
• Over the same period value added grew by 4.2 per cent per annum amounting to 11.2 per cent of total value added in the EU 15 in 2003 (ibid)
BS Contribute to:
• Client performance (O’Farrell et al., 1995; Dawson, 2000)
• Economic performance across the economy (Tomlinson, 2000; Antonelli, 1999)
• Regional development (Marshall et al., 1988; Perry, 1991; Kebble et al., 1991)
• Innovation (Howells, 2006; Metcalfe and Miles, 2000; Wood, 2002)
Productivity in Services and BS• Traditionally associated with low levels of
productivity because of their labour intensive nature (Baumol, 1967; Petit, 1986)
• Growing interest in understanding productivity in the service sector (Wolfl, 2003; Gronroos and Ojasalo, 2004; OECD, 2005; Crespi, Criscuolo, Haskel and Hawkes, 2006; Griffith, Harrison, Haskel and Sako, 2003; inter alia.)
• Productivity levels vary across service sub-sectors. • Studies of the business service sector highlight
variability in the levels of productivity by sub-sector and location (Rubalcaba-Bermejo, 1999; Kox and Rubalcaba, 2007)
• Measuring productivity in business services
Regional Productivity Levels in BS
• To explore productivity we begin by looking at Gross Value Added (GVA) by sector and region
• Productivity can be measured by GVA per head
• Such data gives an indication of levels of productivity
• What factors account for the variations in levels of productivity?
• To address this question we turn to an examination of efficiency levels
SIC London
South East
North West
East West Midlands
East Midlands
Scotland
South West
Yorkshire and The Humber
North East
Wales Grand Total
Grand Total £’m
721 : Hardware consultancy 19% 24% 8% 11% 7% 10% 6% 4% 7% 3% 1% 100% 745
722 : Software consultancy etc
25% 32% 7% 9% 6% 5% 4% 5% 4% 1% 1% 100% 16260
723 : Data processing 19% 27% 12% 7% 11% 9% 3% 4% 3% 4% 1% 100% 2743
724 : Data base activities 47% 20% 7% 6% 2% 6% 2% 5% 2% 2% 1% 100% 508
725 : Maintenance and repair of office, & computing machinery
18% 22% 7% 9% 4% 8% 13% 7% 6% 3% 2% 100% 1192
726 : Other computer activities
28% 25% 11% 8% 4% 6% 5% 7% 4% 1% 2% 100% 4902
731 : R & D on natural sciences and engineering
14% 24% 6% 12% 7% 12% 5% 8% 5% 4% 1% 100% 3056
732 : R & D on social sciences and humanities
37% 26% 4% 5% 2% 9% 0% 12%
2% 1% 2% 100% 130
741 : Legal, accounting, book-keeping, auditing activities; etc
47% 13% 7% 5% 5% 5% 5% 5% 5% 2% 2% 100% 38975
742 : Architectural and engineering activities & related
27% 18% 9% 7% 6% 5% 11% 6% 5% 3% 2% 100% 13345
743 : Technical testing & analysis
11% 30% 11% 8% 10% 6% 11% 2% 4% 5% 2% 100% 1239
744 : Advertising 58% 14% 8% 3% 4% 3% 3% 4% 3% 1% 0% 100% 4766
745 : Labour recruitment and provision of personnel
34% 16% 10% 7% 7% 6% 6% 6% 5% 2% 2% 100% 15882
746 : Investigation and security activities
30% 13% 11% 7% 7% 7% 7% 6% 7% 4% 2% 100% 3502
748 : Miscellaneous business activities not elsewhere classified
40% 16% 8% 6% 5% 6% 4% 5% 5% 2% 2% 100% 16419
Grand Total 36% 18% 8% 7% 6% 6% 5% 5% 5% 2% 2% 100% 123,664
Regions Share of GB total GVA 17% 16% 10% 10% 8% 7% 8% 8% 8% 3% 4% 100% 898,446
Locations of regional sectoral GVA, proportions averaged over 2001-3
SIC 724 : Data base
activities
731 : R & D on natural
sciences and engineering
741 : Legal, accounting,
book-keeping, auditing
activities; etc
744 : Advertising
Grand Total Regions Share of GB total GVA
London 47% 14% 47% 58% 36% 17%South East 20% 24% 13% 14% 18% 16%North West 7% 6% 7% 8% 8% 10%East 6% 12% 5% 3% 7% 10%West Midlands
2% 7% 5% 4% 6% 8%
East Midlands
6% 12% 5% 3% 6% 7%
Scotland 2% 5% 5% 3% 5% 8%South West 5% 8% 5% 4% 5% 8%Yorkshire and The Humber
2% 5% 5% 3% 5% 8%
North East 2% 4% 2% 1% 2% 3%Wales 1% 1% 2% 0% 2% 4%Grand Total 100% 100% 100% 100% 100% 100%Grand Total £’m
508 3056 38975 4766 123,664 898,446
Locations of regional and selected sectoral GVA, proportions averaged over 2001-3
SIC London South East
North West
East West Midlands
East Midlands
Scotland
South West
Yshire & Humber
North East
Wales Grand Total
721 : Hardware consultancy
50,000 41,000 38,000
39,000 33,000 32,000 41,000
29,000
40,000 31,000
28,000
37,000
722 : Software consultancy etc
60,000 57,000 49,000
55,000 45,000 39,000 44,000
38,000
46,000 46,000
35,000
47,000
723 : Data processing 72,000 71,000 64,000
55,000 69,000 60,000 52,000
53,000
42,000 56,000
43,000
58,000
724 : Data base activities 66,000 42,000 29,000
36,000 28,000 29,000 36,000
42,000
34,000 21,000
28,000
36,000
725 : Maintenance and repair of office, & computing machinery
63,000 55,000 48,000
44,000 40,000 41,000 62,000
48,000
44,000 45,000
38,000
48,000
726 : Other computer activities
50,000 49,000 40,000
39,000 32,000 35,000 43,000
34,000
36,000 25,000
28,000
38,000
731 : R & D on natural sciences and engineering
34,000 29,000 43,000
45,000 40,000 35,000 25,000
39,000
41,000 67,000
31,000
40,000
732 : R & D on social sciences and humanities
39,000 35,000 35,000
33,000 28,000 29,000 - 28,000
32,000 31,000
16,000
31,000
741 : Legal, accounting, book-keeping, auditing activities; etc
69,000 39,000 33,000
36,000 33,000 33,000 33,000
33,000
37,000 34,000
28,000
37,000
742 : Architectural and engineering activities & related
63,000 44,000 34,000
34,000 35,000 30,000 42,000
35,000
35,000 37,000
30,000
38,000
743 : Technical testing & analysis
47,000 70,000 34,000
34,000 44,000 37,000 43,000
28,000
32,000 37,000
27,000
39,000
744 : Advertising 73,000 40,000 46,000
30,000 33,000 29,000 35,000
34,000
34,000 27,000
24,000
37,000
745 : Labour recruitment and provision of personnel
35,000 22,000 21,000
20,000 18,000 17,000 21,000
19,000
19,000 19,000
14,000
20,000
746 : Investigation and security activities
27,000 24,000 20,000
24,000 23,000 22,000 19,000
35,000
22,000 22,000
17,000
23,000
748 : Miscellaneous business activities not elsewhere classified
60,000 35,000 31,000
33,000 30,000 31,000 31,000
30,000
29,000 33,000
27,000
34,000
GrandTotal 54,000 44,000 38,000
37,000 35,000 33,000 38,000
35,000
35,000 36,000
27,000
37,000
GVA per head average over 2001-3
SIC 724 : Data base
activities
731 : R & D on natural
sciences and engineering
732 : R & D on social
sciences and humanities
741 : Legal, accounting,
book-keeping, auditing
activities; etc
744 : Advertising
GrandTotal
London 66,000 34,000 39,000 69,000 73,000 54,000
South East 42,000 29,000 35,000 39,000 40,000 44,000
North West 29,000 43,000 35,000 33,000 46,000 38,000East 36,000 45,000 33,000 36,000 30,000 37,000
West Midlands
28,000 40,000 28,000 33,000 33,000 35,000
East Midlands
29,000 35,000 29,000 33,000 29,000 33,000
Scotland 36,000 25,000 - 33,000 35,000 38,000
South West 42,000 39,000 28,000 33,000 34,000 35,000
Yshire & Humber
34,000 41,000 32,000 37,000 34,000 35,000
North East 21,000 67,000 31,000 34,000 27,000 36,000
Wales 28,000 31,000 16,000 28,000 24,000 27,000
Grand Total 36,000 40,000 31,000 37,000 37,000 37,000
GVA per head selected averages over 2001-3
Stochastic Production Frontiers
Input 1
Input 2
f(xi,β)
Standard frontier
Stochasticfrontier
yi≤f(xi,β)
Estimated function
v is not a measure of efficiency but rather an approximation of inefficiency
A trans-log production function was estimated, in natural log form...
Results - Sector effects
COLSHalf normal
Exponential Truncated normal
ONE 2.63 2.75 2.70 2.70S721 0.49 0.51 0.54 0.54S722 0.32 0.33 0.34 0.34S723 0.87 0.88 0.88 0.88S724 0.54 0.58 0.64 0.64S725 0.59 0.61 0.62 0.62S726 0.39 0.41 0.44 0.44S731 0.21 0.22 0.25 0.25S732 0.45 0.48 0.52 0.52S742 0.10 0.11 0.12 0.12S743 0.45 0.47 0.49 0.49S744 0.47 0.48 0.50 0.50S745 -0.38 -0.38 -0.41 -0.41S746 -0.23 -0.23 -0.25 -0.25S748 0.03 0.04 0.04 0.04
Results – Labour effects COLS Half
normalExponential
Truncated normal
ln(w p) -0.54 -0.53 -0.54 -0.54ln(ft e) 1.34 1.35 1.37 1.37ln(pt e) 0.03 0.01 -0.03 -0.02ln (wp * wp)
-0.10 -0.10 -0.09 -0.09
ln(ft e * ft e)
-0.09 -0.10 -0.10 -0.10
ln(pt e * pt e)
-0.02 -0.03 -0.05 -0.05
ln (wp * ft e)
0.21 0.20 0.18 0.18
ln (wp * pt e)
-0.01 0.00 0.01 0.01
ln (ft e * pt e)
0.03 0.04 0.07 0.07
Results – Error terms etc COLS Half
normalExponential
Truncated normal
TIME 0.09 0.09 0.09 0.09Lambda 1.04 15.13Sigma 0.23 2.24Theta 9.00
Sigmav 0.15
Mu -44.62
R2 0.99
LnL 117.83 125.38 125.30
Observations
449.00 449.00 449.00 449.00
Efficiency Mean Std.Dev. Minimu
mMaximum
EFFTRU 0.90 0.07 0.44 0.98EFFEXP 0.90 0.07 0.44 0.98EFFHAL 0.88 0.05 0.61 0.97
EFFTRU
1.79
3.59
5.38
7.18
8.97
.00.20 .40 .60 .80 1.00.00
Kernel density estimate for EFFTRU
Den
sit
y
Truncated Normal Efficiency Kernel
The efficiency equationVariable CoefficientONE ***0.896HIGHSKIL 0.0246GVAUNIT ***3.77e-005S723 ***-0.0405
S724 ***-0.0527
S725 **-0.0256
S731 ***-0.0675
S732 **-0.0325
R7 *0.0194PERWP **-0.0216
R2 9.56E-02N 449
Key Points• Consistent conclusions across specifications• The broad business service sector is estimated as just under
90% efficient• The maximum observed efficiency is 97-8%• Certain sectors appear to be less efficient, namely those
among computer and related activity and research and development
• Increases in the proportion of working proprietors was associated with a decline in efficiency – possibly correlated to sectors
• Significant ‘London effect’ - higher efficiency– Differences in skill level does not appear to fully account
for the gap between London and other regions • Further work needed to unpick relationship between skill
levels, sector and efficiency
Policy Implication
• Scope to improve BS efficiency levels• The role of working proprietors• Skills development is important, but not
sufficient• Support further research into the business
service sector at both a national and regional level
Next Steps
• Further exploration of interrelationships between working proprietors, sectors, skill levels
• Adding in 2004-5 data• Adding in variables relating to the condition of
regional economies and sector specific variables (output growth?)
• Incorporating additional factors, e.g.:• Innovation• International trade