city dimensions of the great productivity slowdown...public policy issues today”. (andy haldane,...
TRANSCRIPT
City Dimensions of the Great Productivity Slowdown
Ron Martin, Peter Sunley, Ben Gardiner, Peter Tyler and Emil Evenhuis
‘Structural Transformation, Adaptability and City Economic Evolution’ An ESRC-Funded Project (ES/N0006135/1), University of Cambridge, UK
(www.cityevolutions.org.uk)
2
“Productivity isn't everything, but in the long run it is almost everything”. (Paul Krugman, 1994)
“Productivity is the challenge of our time…The UK has a long-term productivity problem, which has been made worse by the financial crisis. (UK
reasury, 2015)
“The productivity puzzle is among the most pressing public policy issues today”. (Andy Haldane, Bank of England, 2017)
The Productivity Problem
Why Productivity Matters
• Although not unproblematic (definition and measurements issues), productivity matters
• Productivity is not everything, but as Krugman stresses, it does shape much else in an economy
• Productivity growth influences:• Competitiveness in international markets• Real wage growth and hence living standards• Household incomes and consumption• Taxes and resources for public services
The Concern over Productivity
• Much concern and debate over ‘recent stagnation’ of productivity since the global financial crisis, especially in US and UK
• But in fact, ‘productivity problem’ long predates the crisis
• In most advanced economies, productivity growth seems to have peaked in the late-1960s and has been on a declining trend since, allowing for cyclical movements
The Productivity Problem, 1951-2016
5Source: Conference Board Total Data Base (Productivity converted to 2015 US$, 2011 PPP)
‐2
0
2
4
6
8
10
1951
1954
1957
1960
1963
1966
1969
1972
1975
1978
1981
1984
1987
1990
1993
1996
1999
2002
2005
2008
2011
2014
Annu
al Growth
Rate (Percent)
Labour Productivity Growth Rate ‐ Real Output per Person Employed2015 US$ (converted to 2015 price level with 2011 PPPs)
United KingdomGermanyJapanItalyFranceSwitzerlandCanadaUnited States
Trends Represented by Sixth Order Polynomial
The Productivity Problem, 1951-2016
6Source: Conference Board Total Data Base (Productivity converted to 2015 US$, 2011 PPP)
‐2
0
2
4
6
8
101951
1954
1957
1960
1963
1966
1969
1972
1975
1978
1981
1984
1987
1990
1993
1996
1999
2002
2005
2008
2011
2014
Annu
al Growth
Rate (Percent)
Labour Productivity Growth Rate ‐ Real Output per Hour Worked2015 US$ (converted to 2015 price level with 2011 PPPs)
United KingdomGermanyJapanItalyFranceSwitzerlandCanadaUnited States
Trend Represented by Sixth Order Polynomial
The UK Productivity ProblemAnnual Growth Rate of Real Output Per Worker, 1951-2016
7Source: Conference Board Total Data Base (Productivity converted to 2015 US$, 2011 PPP)
‐4
‐3
‐2
‐1
0
1
2
3
4
5
6
1950
1953
1956
1959
1962
1965
1968
1971
1974
1977
1980
1983
1986
1989
1992
1995
1998
2001
2004
2007
2010
2013
Annu
al Growth
in La
bour
Produ
ctivity
(GVA
pe
r Employed
Person), percent
Trends: Second Order PolynomialSixth Order Polynomial
The Great Slowdown in Productivity Growth
Average Annual Growth in Labour Productivity (percent per annum)
1950-1980 1980-2016
United States 2.49 1.93Canada 2.77 1.18Germany 8.35 1.61France 7.26 1.45Italy 9.82 0.78UK 2.81 2.02Switzerland 3.32 0.61
Source: Conference Board Total Data Base (Productivity converted to 2015 US$, 2011 PPP)
9
• Debate over causes• 1. Technological advances no longer show up in
measures of productivity • The ‘Solow Paradox’ – despite diffusion of computers and
ICT, these don’t seem to show up in productivity figures (Triplett, 1999; Crafts, 2002)
• ICT and Internet increase ‘consumer surplus’ rather than labour productivity?
• 2. Innovation has stalled• Past three industrial-technological revolutions resulted in
peak of productivity growth in middle of 20thC, few remaining opportunities for improving pace of productivity advance (Gordon, 2012, 2016)
• Advanced economies have reached historical technological plateau (Cowen, 2011)
The Slowdown in Productivity
10
• Debate over causes• 3. Fall in business dynamism and ‘ageing
business structure’• Rate of new firm formation has stalled over recent
decades, and new small firms tend to have faster productivity growth than large firms (Brookings Institution, 2014; European Central Bank, 2016)
• 4. Over-regulation of markets • Increasing regulation of labour and product markets
from 1960s onwards slowed productivity advance• Deregulation in 1980s and 1990s led to a partial
resurgence of productivity but not sustained (Conway and Nicoletti, 2017)
The Slowdown in Productivity
11
• Debate over causes• 5. Monopolisation of ICT technologies
• Domination of ICT by handful of global corporations• Has limited competition in emergent e-economy, and
and restricted diffusion and application of technologies• 6. Lack of investment in public capital
• Public capital formation is key to economic growth• Widespread slowdown in investment in public capital
formation and infrastructural investment after 1970s (Aschauer, 1989; Munnell, 1990; Tatom, 1993)
The Slowdown in Productivity
12
• Debate over causes• 7. Shift to post-industrial economy
• Manufacturing main source of productivity advance• Many services have limited scope for productivity growth• These services have grown rapidly as share of national
economies (Baumol et al 1985; Williamson, 1991, Kim, 2006), hence slowed down overall productivity growth
• 8. The geographical dimension• Copious evidence that productivity varies between regions
and cities (OECD, USA)• Is it because the above causes are themselves
geographically uneven?• Are there also place-specific effects?
The Slowdown in Productivity
13
“While the pundits are right to debate the facts and causes of slowing productivity growth at the national level, they would do well also to explore the local dimension of the problem. After all, while many of the proposed causes of malaise—less competition in industries and fewer technological breakthroughs, among others—remain national, many of them may be distinctly local”. (Why Some US Cities Are So Much More Productive Than Others, Marc Muroand Joseph Parilla, 2017)
• Convergence of city productivity across US cities up to 2000, divergence since
Cities and the Productivity Problem
14
• No officially produced regular data on city productivity in UK
• Our ESRC project has constructed new unique yearly data set on 85 British cities over 1971-2015
• Cities defined as travel-to-work areas (2011 boundaries)
• Employment, gross value added (GVA), and GVA per employed worker (labour productivity), 2015 prices
• For 45, 82 and 249 (from 1981) sectors for each city• Also time series data for each city on population
(since 1977), 25 occupations (since 1981), patents (since 1981), and firm demographics (since 1997)
Constructing an Empirical Base for UK Cities
15
• Marked and persistent differences between southern and northern cities (correlation between labour productivity levels in 1971 and 2014 of R=0.686)
• Almost all northern cities have below national average productivity in both 1971 and 2014
• Majority of southern cities have above national average productivity in both 1971 and 2014
• With exception of London and Edinburgh, all other major cities have below national average productivity
• London has pulled away from other cities
The Productivity of British Cities
Labour Productivity across 85 British Cities, 1971 and 2014
16
25000
30000
35000
40000
45000
50000
55000
60000
10000 15000 20000 25000 30000 35000
Prod
uctiv
ity (G
VA per
Employed
Worker)
in
2014
Productivity (GVA per Employed Worker) in 1971
London
Reading
Milton Keynes
OxfordSlough
Basingstoke
SunderlandBlythNewcastle
Sheffield Birmingham
Leeds
Swindon
R=0.69
GB Average
GB Average
Liverpool
Manchester
Aberdeen
Northern CitiesSouthern Cities
Edinburgh
Labour Productivity in London and Major ‘Northern Powerhouse’ Cities, 1971-2014 (GB=100)
17
70
80
90
100
110
120
130
140
150
1971
1973
1975
1977
1979
1981
1983
1985
1987
1989
1991
1993
1995
1997
1999
2001
2003
2005
2007
2009
2011
2013
Relative Labour
Productivity
(Gross
Value
Added
per E
mployed
Worker),
GB=
100
London
LeedsManchesterLiverpoolNewcastleSheffield
GB=100
18
• However, growth over two subperiods, 1971-1991 and 1991-2014 suggests a change in dynamics occurred around 1990
• Convergence over 1971-1991 (Higher productivity growth rates in Northern cities – productivity catching up with Southern cities)
• But divergence over 1991-2014 (Southern cities grew faster – productivity pulling ahead ofNorthern cities)
• And productivity growth in Northern cities in 1991-2014 period also slower than in 1971-1991 period
The Productivity of British Cities
Northern Cities Grow Faster in Productivity, 1971-1991
19
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
5
10000 15000 20000 25000 30000 35000
Average An
nual
Growth
of Lab
our
Prod
uctiv
ity , 1
971‐1991
(per
cent)
Labour Productivity, 1971
Period 1971‐1991
Oxford
Reading
MedwayLeamington
High Wycombe
Milford Keynes
London
Sunderland
BlythMansfield
Northern Cities
York
Merthyr Tydfil
Southern Cities
R=‐0.59
Southern Cities Grow Faster in Productivity, 1991-2014
20
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
5
20000 25000 30000 35000 40000
Average An
nual
Growth
of Lab
our
Prod
uctiv
ity, 1991‐2014
(Percent
)
Labour Productivity, 1991
Period 1991‐2014Northern CitiesSouthern Cities
London
Reading
Tunbridge Wells
OxfordGuilford
Stevenage
Preston York
Swansea
BlackpoolExeter
CardiffR=0.20
A Shift in Geographical Dynamics? Relative Productivity Paths of Selected Cities, 1971-2014,
GB=100
21
50
60
70
80
90
100
110
120
130
140
150
1971
1973
1975
1977
1979
1981
1983
1985
1987
1989
1991
1993
1995
1997
1999
2001
2003
2005
2007
2009
2011
2013
Relative GV
A per W
orker, GB
=100
LondonReading
Oxford
BasingstokeStevenageMedwaySunderlandLiverpoolMerthyr TydfilDoncaster
Convergence Divergence
Change in Dynamics?
22
Labour Productivity Growth in Northern and Southern Cities(Average Annual Growth, percent per annum)
1971-1991 and 1991-2014
The Geographical Switch in Productivity Growth across British Cities
1971-1991 1991-2014
Southern CitiesNorthern CitiesLondonManchesterGreat Britain
1.842.282.992.45
2.08
2.051.512.751.63
1.69
23
• Various explanations advanced:• Economic Structure - (export) specialisation or
diversity? Both argued to increase productivity• Agglomeration - ‘mass’ (size/density) confers
positive externalities and increasing returns effects that raise productivity
• Connectivity – linkages and accessibility boosts labour markets, product markets and knowledge transfers; idea of ‘clusters’ of cities, and integrated regional systems of cities
What Determines Productivity Growth in Cities?
24
• Innovation – local innovation systems, close university-business interactions promote technological advance, thence productivity
• Human Capital - production, attraction and retention of educated and skilled workers is key
• Local Infrastructure – modern and efficient physical, informational, cultural and social capital
• Governance – coherent system of local institutions with common strategic purpose in promoting business, and fiscal powers over revenue and spending on housing, skills, infrastructure, etc
What Determines Productivity Growth in Cities?
25
• Much is made in regional studies literature about key role of economic structure – specialised v. diversified v. related variety
• Most British cities have become less specialisedover past 40 years – convergence of economic structures
• Northern cities have lost their former manufacturing tradable (export) base
• So has London, but…• London has replaced it with high-value professional,
business and technology related KIBS, while northern cities have lagged in this respect (and have more lower-value services)
The Case of Structure
Labour Productivity in Selected Sectors of UK Economy, 1971-2014
26Source: ONS
0
10000
20000
30000
40000
50000
60000
70000
80000
90000
1971
19
73
1975
19
77
1979
19
81
1983
19
85
1987
19
89
1991
19
93
1995
19
97
1999
20
01
2003
20
05
2007
20
09
2011
20
13
Gross
Value
Add
ed per
Worker (£,
200
5 Prices)
Advanced and High Tech Manufacturing
Knowledge Intensive Business Services
Light Manufacturing
Retail and Personal Services
Fastest Employment Growth has been in Low Productivity Services, 1971-2014
‐1
0
1
2
3
4
5
6
7
‐12 ‐10 ‐8 ‐6 ‐4 ‐2 0 2 4 6 8
Average An
nual
Rate of
Growth
of Labour
Productivity
Average Annual Rate of Employment Growth (Percent)
ManufacturingPrivate ServicesPrimaryPublic ServicesUtilities
R=‐0.366Pharmaceuticals
Transport equipment
Textiles
Coal mining
Metal ores
ChemicalsInformation services
Computer software
Head office/Management con
Computer equipMotor vehicles
Finance
27
High productivity growth but falling employment
Low productivity growth but rising employment
Sectoral Convergence in Output Structures across Cities, Krugman Index, 1971-2014
28
‐0.4
‐0.3
‐0.2
‐0.1
0
0.1
0.2
0.3
0.2 0.3 0.4 0.5 0.6 0.7 0.8
Chan
ge in
Krugm
an Inde
x, 197
1‐20
14
Krugman Index 1971
Oxford
Kettering
Wolverhampton
Trowbridge
AberdeenLondon
DoncasterReading
Manchester
Glasgow
CardiffEdinburgh
Bristol
R=0.379
Initially most-specialised citieshave seen greatest decline in specialisation
Only a few cities have become more specialised
Convergence in Export Intensity, 1971-2014
29
‐0.6
‐0.4
‐0.2
0.0
0.2
0.4
0.6
0.8
0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8
Chan
ge in
LQ (E
xport‐Inten
sive Sectors )
1971
‐2014
LQ (Export‐Intensive Sectors) 1971
Dudley
Wolverhampton
Birmingham
Middlesbrough
Swansea
Exeter
Colchester
Trowbridge
Oxford
Northern CitiesSouthern Cities
Most Southern Cities have become more export intensive
Most Northern Cities have become less export intensive
Decline in Output Specialisation (Herfindahl-Hirschman Index): Selected Cities, 1971-2014
30
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0.16 19
71
1973
19
75
1977
19
79
1981
19
83
1985
19
87
1989
19
91
1993
19
95
1997
19
99
2001
20
03
2005
20
07
2009
20
11
2013
HH In
dex
Oxford
Trowbridge
Plymouth
Gloucester
Portsmouth
Basingstoke
Reading
Exeter
Swansea
Medway
Export Base Employment in London and Major Northern Cities, 1971-2014 (1971=100)
31Source of data: Martin et al (2016, 2017)
Decline in Manufacturing export base
Growth in KIBS export sectors
Contribution of Structural Change to City Productivity Growth
32
Growth due to Structural Change
Total Growth Within Sector Growth
Within-Sector and Structural-Change Contributions to City Productivity
Growth
33‐60‐40‐200
20406080
100120140
0 20 40 60 80 100
Contrib
ution to
Totl Change
(Percentage Points)
Total Productivity Change (Percent)
1991‐2014
Within Sector Structural
‐60‐40‐200
20406080
100120140
‐20 0 20 40 60 80 100
Contrib
ution to
Total
Change
(Percentage Points
Total Productivity Change (Percent)
1971‐1991Within Sector Structural
Structural change contribution mainly negative - cities have shifted to lower productivity growth sectors - but within-sector contribution dominatesin both periods
34
• Over 1971-1991, productivity growth in northern cities faster than in southern cities
• Productivity growth slowed almost everywhere in between 1971-1991 and 1991-2014
• Especially in northern cities• By beginning of 1990s,these cities had lost their
historic manufacturing export base• Shift to services everywhere has led to structural
convergence, but has favoured southern cities, • However, many service sectors have lower
productivity growth than manufacturing
Summary of Findings
35
• Negative structural change component of total city productivity growth indicates that, overall, cities have shifted their employment structures to slower productivity growth sectors – the ‘post-industrial argument
• But ....Within-sector component of city productivity growth outweighs structural change component
• Differences across cities in this within-sector component accounts for most of city differences in total productivity growth
• The key issue is that structure is not main determinant of city differences in productivity growth
Summary of Findings
36
• Indicates that sector specific productivity growth varies across cities – Why?
• Do best performing cities have certain sectors that outperform their counterparts elsewhere? This might suggest functional differences in given sector across cities, or different supply chain roles, etc (Baldwin, 2016)
• Are best performing cities those that have higher productivity growth across all of their sectors? This suggests role for city-specific factors, such as agglomeration, skills, infrastructure, linkages, etc (Glaeser, 2010; Simon and Nardinelli, 1996, 2002)
What Determines this Inter-City Variation in Within-Sector Productivity Growth?
37
• City dimension has a role in explaining slowdown in productivity growth
• Research project has attracted close attention of UK Government and cities themselves (Core Cities, Key Cities, Greater London Authority)
• Our work reinforces new focus on ‘place’ in UKGovernment’s emerging New Industrial Strategy
• Also relevant for new city-region devolution deals
• But more analysis is sorely needed to inform policy• Currently exploring different models of city
productivity growth
Policy Relevance
38
Labour Productivity Levels (Output per Worker) across US Cities, 2015
39
Unequal and Divergent Growth
• British cities have exhibited marked differences in economic growth (employment, output) over past 45 years
• Cumulative growth gaps are substantial• Most southern cities have grown consistently faster
than northern cities• A key dimension of the ‘spatial imbalance’ problem
that characterises the British economy• Up to early-1990s, London part of the ‘north’ –
turnaround since• Largest, core, cities have lagged national growth
41
• City productivity growth can be decomposed into three components (eg Kruger, 2008; Rodrik, 2011):
• Within-sector growth (firm dynamics)• Between-sector growth (shifts into higher or lower
productivity sectors)• Interaction (Covariance) of Within and Between
effects• Results indicate that within-sector effects have
dominated –• Same sector performs differently across
different cities
Structural Change and City Productivity Growth
42
Sources of Within- and Between-Sector City Productivity Growth
Within sector developments (firm dynamics)
Between sector developments (cross sectoral structural shifts and reallocations)
Death of low productivity firms and birth of new higher-productivity firms
Broad switch from Primary Industry to Manufacturing to Services
Adoption of new organizational and operational methods and technological innovations
Decline of low productivity activities and birth and development of high productivity activities
Production of new higher-value products and services and switch into new markets
Industry restructuring – mixing and new combinations with other sectors to produce higher productivity industries
43
• ‘Functional’ structures may matter more than sectoral structures
• Occupational mix (Simon and Nardinelli, 1996, 2002)• Supply chains and networks may differ between cities
• Agglomeration effects• Early results indicate that size/density not significant (especially once London is excluded)
• Export orientation/base (Rowthorn, 2010; Moretti, 2013)• Found some evidence for this
• Market accessibility/connectivity• ‘City clusters’ and city linkages
What Determines this Inter-City Variation in Within-Sector productivity Growth?