02Introduction
Over the long term, improving people’s earnings and standard of living depends on rising productivity. This is why the UK’s so-called ‘productivity puzzle’ – flatlining productivity after the financial crisis of 2008 – is so concerning.
Although the slowdown has been a common feature of developed economies during this period, it has been particularly profound in the UK.1 The country moved from having the second highest productivity growth rate (1.9 per cent a year) in the G7 before the financial crisis to the second lowest (0.7 per cent a year) between 2009 and 2019 (see Figure 1).2
Figure 1: Productivity slowdown has been particularly strong in the UK
Before the pandemic, this issue was a focus of successive governments. In 2015, the Treasury published its productivity plan to address the issue.3 Two years later, the next government announced its Industrial Strategy to ‘drive productivity growth across the country’.4 Both of these strategies, and the broader debate, have missed the geography of the observed slowdown in productivity growth, where London plays a disproportionate role.
London’s contribution to the national economy is well understood. It accounts for around a quarter of national output and is one of the most productive cities in Europe. This performance results from its ability to offer knowledge-based businesses access to skilled workers and a network of clients, collaborators and competitors, which makes them more productive. This is known as agglomeration.
However, the Capital’s role in national productivity flatlining is less well known. This paper focuses on changes in its productivity over the past two decades and the London-specific components of this slowdown by comparing it with a selection of European counterparts. It then makes policy recommendations aimed at improving not just the economic performance of London, but the UK economy overall.
Box 1: Methodology
Definition of London
For the purpose of this paper, the research focuses mainly on London. Due to the data availability, and to compare it with other cities, this report uses three different geographic definitions of London: Greater London (ITL definition); Primary Urban Area (PUA) and the OECD’s definition of a Metropolitan Area.5 Appendix 1 shows the geographical differences between these definitions.
Other cities and regions used
This paper compares London with other cities or regions in the UK or abroad, depending on data availability. In terms of international peers, the cities were selected using OECD metropolitan areas, and based on two criteria – being at least 15 per cent more productive than the national average and having at least 1 million residents.6 The group includes Paris, New York, Brussels, Stockholm and Milan.7 Due to data constraints, some of the analysis uses the closest NUTS3 regions instead (e.g. Metropolitan Paris and Île-de-France, respectively). For further details, see Appendix 1.
The definition of Greater London is used for all regional and most international comparisons. The PUA definition is used for comparisons with other British cities. Unless otherwise stated, the geographic definition refers to Greater London.
Potential limitations
Using slightly different geographies to analyse a city raises comparability issues. Reassuringly, the core of London’s population and economy is captured in the three geographies used in this paper (Figure 24, Appendix 1). More importantly, changes over time in key indicators like population and GDP are similar across different geographical definitions.
Data used for this research
This paper uses several publicly available datasets from the ONS, OECD and other sources. These include:
- Productivity data at the regional level from sub-regional productivity in the UK (ONS)
- National accounts by sector at the regional level from sub-regional GVA (ONS and OECD)
- Regional gross fixed capital formation (GFCF) estimates by asset type (ONS)
- R&D and investment (OECD)
- R&D spending by businesses and government (OECD)
- Nominal house prices at the regional level (OECD)
- Population by country of birth at the regional level (ONS)
- Regional equivalised disposable income, after housing costs (GLA Intelligence)
- Human capital estimates per head at the sub-national level (ONS)
- Commercial property data (the Valuation Office Agency)
- National Insurance number (NINo) applications from foreign-born residents at the regional level (the Department of Work and Pensions)
- International commercial property prices (CBRE)
As the paper focuses on the productivity slowdown after the financial crisis until the pandemic, data from 2020 onwards may not be analysed even if it is available.
Data relative to the regional GFCF by asset type are ONS experimental estimates.
Data from the ONS’ Secure Research Service
The data based on firm productivity percentile is estimated using the dataset from the ONS’ Regional Annual Business Survey, which includes the headquarters and branches of all businesses surveyed. The use of branch data sets this research apart from most other contributions to date, and while we recognise that the ONS apportioning output to branches within an enterprise has data limitations, it gives a better reflection of the business base at the sub-national level. The patterns shown in this report remain the same when looking at enterprise level and single plant firms.
The survey covers the non-financial business economy, excluding financial services and public administration. Publicly funded health and education activities have also been removed to look at the private sector only. The productivity figures from this specific dataset are shown in nominal terms as deflators at the firm level are not available.