1. What future(s) for local economies

Abstract

This chapter discusses the impacts of the COVID-19 pandemic on local economies. It considers factors that may contribute to different impacts across local labour markets (e.g. share of jobs in sectors most at risk, teleworking potential, specialisation in tradable sectors). Disparities could also increase within local labour markets. Young people, the low-skilled and women are being hard hit by the economic impacts of COVID-19, and local SMEs and the self-employed face particular challenges. Even pre-COVID-19, rosy national labour market figures often hid significant disparities within countries, reflecting different patterns of resilience to the global financial crisis and adaptation to broader structural changes. Accordingly, some places were hit by COVID-19 when they were already struggling.

In Brief

An economic tsunami that will not hit all places equally

While managing the health impacts of COVID-19 is a first order concern, the pandemic has also put unprecedented pressure on local labour markets and economies. GDP has plummeted, the number of hours worked has drastically shrunk, and unemployment is spiking. The economic impacts of this health crisis dwarf any event in recent memory.

In a time of radical uncertainty, there are many unknowns for local jobs and development. What we do know is that the economic fallout of COVID-19 will be deep but not the same across communities. Where already available, initial data shows that unemployment is spiking unevenly across regions within countries. While early, some evidence indicates that big cities are taking particularly large hits. A number of factors may influence these local divides:

The share of jobs in sectors at risk due to the direct impacts from containment measures varies from less than 15% to more than 35% across regions, with large cities and tourist destinations typically having the highest share.
The rapid adoption of teleworking is an important means to preserve jobs when strict social distancing is needed. However, the share of jobs amenable to teleworking varies 15 percentage points across regions within countries, with urban areas typically having a higher share.
Temporary jobs are typically the first shed in downturns, and their share in total employment can vary by over 10 percentage points across regions in some countries. Temporary work is more common in regions with a lower-educated workforce, higher unemployment, and a smaller tradable sector.
The share of regional employment in tradable sectors can vary over two fold in some OECD countries, and regions with high shares may be more vulnerable in the short term to disruptions in supply chains and contractions in global trade. However, tradable sectors may also help regions bounce back more quickly once the recovery is underway, a trend seen in previous crises.
Localised outbreaks of the virus, the associated responses and changes in individual behaviours will impact economic activity in some places more than others.

Even before COVID-19 hit, the labour market was not as rosy as headline national figures suggested. While the overall OECD unemployment rate stood at 5.4%, national averages masked other issues such as stagnant wage growth and a shrinking middle class. They also hid the fact that some places continued to struggle with the legacies of the 2008 crisis.

Nearly half of regions had unemployment rates higher in 2008 than 2018. Only in one-third of OECD countries had unemployment recovered in all regions.
In over half of OECD countries, regional unemployment disparities were either growing, or shrinking for the wrong reasons (i.e. because of increasing unemployment rates in the best performing regions).
Jobs had also become more geographically concentrated in the past two decades in most countries, especially high-skilled jobs, suggesting growing divides in how places were adapting to longer-term structural changes.

If lessons from the global financial crisis hold true this time around, some places will be harder hit than others and could struggle for years to come. In roughly 80% of regions, employment levels (number of jobs) declined at some point following the global financial crisis, but the scale of this decline varied drastically. At their respective lowest points, employment declined by over 20% in some of the hardest hit regions in Spain and Greece, and by over 10% in some places in the United States, Denmark, Italy, Poland, Portugal, and Turkey, as well as Romania. The last crisis appears to have accentuated difficulties for places already struggling with other challenges – relatively high unemployment, a low-educated workforce, and low labour productivity.

COVID-19 could also lead to deepening divides within local labour markets. The low skilled, low-wage workers, and young people may be the most vulnerable to COVID-19-related job losses, and could face longer-term scarring effects. They are highly represented in the sectors most at risk, less likely to hold jobs that allow them to telecommute, and more likely to be on temporary contracts. Local SMEs and the self-employed also face large risks, and may have less reserves to survive the shock as well as face additional challenges accessing public supports.

Introduction

While managing the health impacts of COVID-19 is a first order concern, the pandemic has also put unprecedented pressure on local labour markets and economies, and generated radical uncertainty. GDP has plummeted, the number of hours worked has drastically shrunk, and unemployment is spiking. The economic impacts of this health crisis dwarf any event in recent memory. Yet, much remains unknown about how the COVID-19 pandemic will continue to impact our economies and societies:

From a health perspective, how will the virus spread in different places and seasons? How long will it take to develop and disseminate a vaccine, and what types of social distancing will be required until that point?
From an economic perspective, how many firms will go out of business permanently, and how will employers re-organise production processes? How will investment, demand and trade be impacted over the longer term?
From a policy perspective, what policy measures will governments use in the short and long term to mediate the impacts of the crisis? How will citizens’ expectations of governments change?
And finally, from a social perspective, how will people change their behaviours to adapt? Will the pandemic spark permanent changes to how and where people live, work, and learn?

What we do know is that the economic fallout of COVID-19 will be deep but not the same across communities. The question is therefore, not what future for our economies, but rather what future(s) for local economies. There are many different ways that COVID-19 and the associated economic downturn will impact the economy and jobs differently across places (see Table 1.1).

Table 1.1. How COVID-19 related job losses will hit some places harder than others
COVID-19 CONTAINMENT MEASURES AND CHANGES TO INDIVIDUAL BEHAVIOURS
Localised outbreaks	Share of jobs in sectors most impacted	Share of jobs amenable to teleworking
In response to specific local outbreaks, changes in individual behaviors and geographically-targeted containment measures will impact economic activity in some places more than others.	The share of jobs in sectors most directly impacted by containment measures varies from less than 15% to more than 35% across regions, with large cities and tourist destinations having a higher share of jobs at risk.	The share of jobs amenable to teleworking varies 15 percentage points across regions within countries. Urban areas can rely more on teleworking to help preserve certain jobs when stricter social distancing is needed.

ASSOCIATED ECONOMIC DOWNTURN BEYOND CONTAINMENT MEASURES
Local resilience to downturns	Pre-COVID-19 labour market health	Share of temporary jobs
Local economies have displayed very different patterns of resilience in past recessions. For example, at its lowest point, employment declined by over 20% in the regions hardest hit by the global financial crisis.	Many places were still struggling with the scars of the global financial crisis and other structural changes even prior to COVID-19 hitting. Half of regions still had unemployment rates higher in 2018 than in 2008, a full ten years after the crisis. In one-quarter of regions, unemployment exceeded 8%.	The share of temporary jobs, which are typically the first shed in downturns, varies by over 10 percentage points across regions in some countries. Temporary work is more common in regions with a lower-educated workforce, higher unemployment, and a smaller tradable sector.
Source: Author’s own elaboration

This chapter takes stock of local labour market1 health, particularly the impacts of COVID-19 and the associated economic downturn. While a spike in unemployment is likely across the board, some places will be more vulnerable to job losses than others based on sector specialisation and other factors, such as the share of jobs amenable to teleworking. The scale of local job losses during the global financial crisis shows that crises have very different impacts across territories, often accentuating existing labour market weaknesses. Local economies will also be impacted differently by an acceleration of longer-term structural changes, such as automation, an issue discussed further in Chapter 2. Even pre-COVID-19, the labour market picture was not as rosy as national figures suggested: unemployment rates and patterns of job creation and quality varied considerably across territories, reflecting the legacy of longer-term structural changes as well as different patterns of resistance and recovery from the global financial crisis. Finally, within local labour markets, COVID-19 could further entrench existing disadvantages for the low-skilled, young people and women, and have particularly negative impacts on SMEs and the self-employed.

The impact of COVID-19 on local labour markets

Unemployment is spiking unevenly across local labour markets

COVID-19 is causing unemployment to increase across the OECD, and some cities and regions are undoubtedly being harder hit than others. While unemployment is expected to increase in almost all OECD countries by the end of 2020, this surge came earlier for some countries than others. Countries that relied on expanded unemployment benefits or stimulus payments to support workers through job losses or reductions in working hours already saw unemployment significantly increase in the first half of 2020. In contrast, countries that made widespread use of job retention schemes, such as short-time work programmes which cover the wages of furloughed workers, staved off these initial increases in unemployment (OECD, 2020[1]). However, as these schemes are rolled back and businesses manage prolonged drops in demand, unemployment will tick up in many places.2

In countries where unemployment increased significantly in the first half of 2020 and with available data, regional divides are already apparent. For example, in the United States, the August 2020 unemployment rate ranged from 4.0% in Nebraska to 13.2% in Nevada. Unemployment increased by less than 1 percentage point in Nebraska compared to the previous year, while in Nevada, it increased by over 9 percentage points (U.S. Bureau of Labor Statistics, 2020[2]). Likewise, in Canada, regional patterns varied considerably. Unemployment increased over two-fold in British Colombia between January and July, but only by a magnitude of 1.3 in New Brunswick. In the United Kingdom and Norway, unemployment also rose in all regions, although the patterns were more similar across regions.

In countries with widespread use of short-time work schemes, regional participation rates can provide an indication of where a high share of jobs were directly impacted by COVID-19 (see French and German examples in Figure 1.3). In France, for example, the Paris region (Île-de-France) had a higher share of workers on short-time work schemes than other regions. However, the degree to which this will translate to higher unemployment rates as these schemes are rolled back remains to be seen. Additionally, it is important to note that in a number of countries, these schemes were extended in the fall of 2020 in response to the second wave of the virus.

Figure 1.1. North America and Europe: regional unemployment divides are already showing up in national data

Figure 1.2. Latin America: regional unemployment divides are already showing up in national data

Figure 1.3. France and Germany: participation in short-time work schemes varied across regions

Job postings can provide another indication of local labour market health, as increases in unemployment during downturns typically result from both decreases in hiring and increases in job separations (OECD, 2009[3]). Across the 18 OECD countries with available data, online job postings decreased by an average of 35% on any given day between 1 February and 1 May 2020. “Public services” (i.e. services in education, health care and social work, or public administration and defence sectors), and business services, followed by trade and transportation, and the accommodation and food industries made the largest contributions to these declines (OECD, 2020[1]).

Regional trends in job postings suggest that hiring may be decreasing the most in large cities. Emerging evidence on the impact of COVID-19 on labour demand in the US shows that in the first half of 2020, online job postings contracted more and the recovery was slower than would have been expected in metropolitan areas that were larger, had a more educated workforce, and a more diverse industrial structure (Tsvetkova, Grabner and Vermeulen, 2020[4]). While this may indicate that patterns of resistance and recovery will be different this time around compared to the previous crisis, these initial results may also be influenced by differences in containment measures across metropolitan areas or other local considerations. However, similar trends can also be found in the other countries. In looking at job postings in the United Kingdom, postings were down more in London than the national average compared to 2019 levels (Office for National Statistics, 2020[5]). It is important to keep in mind, however, that online vacancy information provides only a partial picture of a labour markets, with a bias towards high-skilled occupations and sectors. Additionally, as the situation continues to rapidly evolve, it remains to be seen if these patterns hold true over time.

The structure of local economies may make some places more vulnerable to job losses than others

Some places may be more vulnerable to the direct impacts of COVID-19 than others. Sector specialisation, the share of jobs amenable to teleworking, and trade exposure may all impact local vulnerabilities. Of course, the likelihood that these risks materialise and for how long depends on a number of factors: the pace and scale of roll-backs of short-time work or other schemes to promote job retention; the rigidity of employment protection legislation; employer expectations about how long COVID-19 will impact their activities; and the degree to which firms go out of business, reduce or re-organise activities permanently.

Additionally, the scale of local job losses also depends significantly on local outbreaks of the virus and ensuing changes in individual behaviours and containment measures. Rolling waves of targeted containment measures in regions and cities will likely be a reality until a vaccine is found. This has already been in the case in many countries, where national containment measures were rolled back at different places across regions, or where stricter containment measures were re-introduced in response to local flare-ups. Accordingly, at the same time that economic activity in some places is restarting, in other places, it will essentially be re-frozen. This will undoubtedly have important impacts on local employment beyond what can be deduced based on local economic structure, but where and when cannot be predicted at this stage. However, at the time of this publication, a number of countries, particularly in Europe, were re-introducing stricter nationwide containment measures in response to a second wave of the virus.

Large cities and tourism destinations have a higher concentration of jobs in the sectors most at risk from strict containment measures

Across regions countries, the share of jobs in the sectors most impacted by strict containment measures represents less than 15% to more than 35% of local jobs (Figure 1.4).3 In one out of five of these regions, more than 30% of jobs are at risk. These figures are based on OECD estimates that jobs in manufacturing of transport equipment; construction; wholesale and retail trade; air transport, accommodation and food services; real estate services; professional service activities; and arts, entertainment and recreation are most at risk from strict containment measures (OECD, 2020[6]) (see Reader’s Guide for further information on the calculations of the share of jobs at risk). Within countries, the share of jobs at risk can vary by more than 20 percentage points across regions. In Greece, for example, they range from up to 55% in the South Aegean Islands to 22% in Central Greece. Regional differences are also particularly stark in the Slovak Republic, France, and Portugal as well as Romania.

Figure 1.4. Share of jobs in sectors most at risk from COVID-19 containment measures

Figure 1.5. Regions with the highest share of jobs at risk by country, TL2 regions

Tourist destinations, capitals and other large cities have the largest share of jobs in the sectors most at risk (Figure 1.5). The importance of tourism, local consumption, and services – including large retailers, general-purpose stores, and business in the hospitality industry, such as coffee shops and restaurants – partially explains these relatively high shares. The extent to which strict containment measures are active in tourism high seasons is an important determinant of the extent to which this risk is realised. In Europe, several major tourist destinations, such as Crete, the South Aegean and Ionian islands (Greece), Balearic and Canary Islands (Spain) as well as the Algarve region in Portugal have over 40% of jobs at risk. In Korea, the largest share of jobs at risk is in Jeju-do, a region where tourism represents an important pillar of the economy. For similar reasons in North America, Nevada (which includes Las Vegas) stands out as having the highest share of jobs at risk, followed by Hawaii. Indeed, unemployment in both Hawaii and Nevada spiked considerably in the first half of 2020 (see Figure 1.1).

In roughly one-quarter of countries, the capital region has the highest share of jobs at risk. This includes the Czech Republic, Denmark, Finland, France, Lithuania, Norway, Sweden, as well as Romania. Greece and Spain follow the same pattern if their island regions, which are highly exposed to the decline in tourism, are excluded. In most cases, the higher risk observed in capitals, or other large cities, reflects their specialisation in retail and wholesale trade. This is the case for Athens, Bucharest, Prague, Helsinki, Oslo, Stockholm, and Vilnius. On the other hand, large cities tend to have other protective factors – a more diverse economy, a more skilled labour force, a larger share of jobs compatible with teleworking – which can help them adapt to shocks and could facilitate the economic recovery.

Some of the sectors that have been particularly hard hit by containment measures are unlikely to recover quickly. For example, international tourism is anticipated to decrease by 80% in 2020, and is not expected to rebound quickly (OECD, 2020[8]). As a labour-intensive sector, the impacts on local employment in tourism destinations will be profound. Similarly, culture and creative industries will likely take a deep and prolonged hit. Social distancing brings ongoing challenges to venue-based activities such as theatres and museums, and organisations that rely heavily on public and philanthropic funding and visitor revenues may face greater financial challenges (see Box 1.1). Additionally, the high share of self-employed, freelancers and SMEs in the sector creates unique challenges that general public support schemes are not always well-tailored to address.

Box 1.1. The impact of COVID-19 on culture and creative sectors

Cultural and creative sectors are among the most affected by the current crisis, and account for less than 1 to over 5% of employment across OECD regions. Venue-based sectors (such as museums, performing arts, live music, festivals, cinema, etc.) are the hardest hit by social distancing measures. The abrupt drop in revenues puts their financial sustainability at risk and has resulted in reduced earnings and lay-offs for workers. It also has repercussions throughout their supplier networks, hitting suppliers in both creative and non-creative sectors. Some cultural and creative sectors, such as online content platforms, have seen an increase in demand for cultural content streaming during lockdowns, but the benefits from this extra demand have largely accrued to the largest firms in the industry.

The effects will be long lasting due to a combination of several factors. The impacts on distribution channels and the drop in investment will affect the production of cultural goods and services and their diversity in the months, if not years, to come. Over the medium term, the anticipated lower levels of international and domestic tourism, drop in general demand, and reductions of public and private funding for arts and culture, especially at the local level, could amplify this negative trend even further. In the absence of responsive public support and recovery strategies, the downsizing of cultural and creative sectors will have a negative impact on cities and regions in terms of jobs and revenues, innovation, citizen well-being and overall vibrancy and diversity.

Many of the broad supports to workers and firms rolled out in response to COVID-19 were not well suited to the peculiarities of the sector. Cultural and creative sectors largely consist of micro-firms, non-profit organisations and creative professionals, often operating on the margins of financial sustainability. Large public and private cultural institutions and businesses depend on this dynamic ecosystem for the provision of creative goods and services. Employment and income support measures are not always accessible or adapted to the new and non-standard forms of employment (freelance, intermittent, hybrid – e.g. combining salaried part-time work with freelance work) that tend to be more precarious and are more common in this sector. SME finance measures could also be better adapted to businesses with significant intangible assets. Similarly, innovation supports, largely catering to technological innovations, could be adapted to other forms of innovation more common in the sector, such as innovations in format and content, including through mixed use of different media. Such supports could also recognise that the sector generates innovation through creative skills, new ways of working, new business models, and new forms of co-production.

During lockdowns, many public and private providers moved content online for free to keep audiences engaged and satisfy the sharply increased demand for cultural content. While the provision of free and digitally mediated cultural content is not sustainable over time, it has opened the door to many future innovations. Massive digitalisation coupled with emerging technologies, such as virtual and augmented realities, can create new forms of cultural experience, dissemination and new business models with market potential. To capitalise on them, there is a need to address the digital skills shortages within the sector and improve digital access beyond large metropolitan areas, with the additional consideration that digital access does not replace a live cultural experience or all the jobs that go with it.

Source: OECD (2020[9]).

Cities also host more high-skilled jobs that can be done remotely, which could help buffer the shock for some workers

Workers and firms rapidly and widely adopted teleworking during the periods with the strictest containment measures, with many governments providing financial supports and updates to legal frameworks to facilitate this transition. The OECD estimates that an average of 39% of workers teleworked in early 2020 during lockdowns, with significant differences across countries (OECD, 2020[1]). In early April 2020, up to half of American workers were working from home – more than double the amount who worked from home, at least occasionally, in 2017-18 (Guyot and Sawhill, 2020[10]). In France, an estimated 39% of employees were teleworking in May (ODOXA, 2020[11]), while the rate of employees working from home at least once a week was estimated at just 3% in 2017 (DARES, 2019[12]).

Yet the potential for remote working varies significantly across regions: on average, the share of jobs amenable to teleworking varies 15 percentage points across regions within countries (see Figure 1.6).4 This difference reaches more than 20 percentage points in the Czech Republic, France, Hungary, and the United States, driven by comparatively high levels of potential remote working in their capitals.

Figure 1.6. Regional differences in share of jobs amenable to teleworking are large

Box 1.2. How visits to workplaces have changed across regions

New data sources, such as anonymised geographic data from smart phones, can also provide insights into how containment measures have impacted mobility and activity in different regions. The charts below give examples of how visits to workplaces changed across regions compared to a baseline period in early 2020. These regional differences may reflect both how local economies were impacted differently by nationwide measures, as well as the impact of more geographically targeted containment measures. While it is impossible to tell from this data whether visits to workplaces reduced because of teleworking, employees being put on short-time work schemes, or lay-offs, it does show significant regional variations in how many people were travelling to workplaces at different phases of COVID-19 containment. However, these data should be interpreted with caution across countries and regions, as differences in how different types of locations are categorised across different types of regions (i.e. urban vs. rural) limits these comparisons.

Figure 1.7. Changes in visits to workplaces from Feb-August 2020

Cities and capital regions tend to have a higher share of jobs amenable to teleworking (OECD, 2020[13]). In Europe, the share of jobs amenable to teleworking in cities (above fifty thousand inhabitants) is 13 percentage points higher than in rural areas. In Croatia, Finland, Hungary and Luxembourg, the gap is larger than 17 percentage points. In towns and semi-dense areas, the potential for remote working is more similar to that of rural areas than that of cities. Unsurprisingly, there is also a strong correlation between the skills of the local workforce and the share of jobs amenable to teleworking. However, other research suggests that while cities have a higher share of jobs amenable to teleworking, this is at least partially compensated by the fact that non-metropolitan areas host other types of jobs that can be considered “safe”, i.e. those that are not amenable to teleworking but require a low level of physical proximity – such as in agriculture (Basso et al., 2020[15]). Additionally, the polarised nature of urban labour markets mean that they have both relatively high shares of high-skilled workers who can work remotely, and high shares of low-skilled workers, often in face-to-face service occupations, that are strongly impacted by COVID-19.

These geographic divides in teleworking have already appeared in the data. An April 2020 survey in France showed that 41% of the labour force was teleworking in Île-de-France, compared to 11% in Normandy (ODOXA, 2020[16]). Additionally, as described in Box 1.2, smartphone mobility data suggests that visits to workplaces changed differently across regions in the first half of 2020. However, this data does not allow for differentiation between reduced workplace visits due to increased teleworking or because people were furloughed or laid off, and therefore should be interpreted with caution.

Within regions, there are also important differences in terms of who can telework: as young people, the low-skilled, and low-wage workers are more likely to hold jobs requiring a physical presence. In May 2020, a French survey found that 89% of managers (cadres), 54% of “middle management” (professions intermédiaires), 26% of employees (employés) and only 3% of manual workers (ouvriers) teleworked during the lockdown period (ODOXA, 2020[11]). Other research has shown that higher-income workers are much more likely to be working from home during the pandemic and much less likely to be unable to work at all than lower-income workers (Reeves and Rothwell, 2020[17]). According to smartphone location data in the United States, lower-income workers were more likely to continue daily commuting during the early spring, while higher-paid workers were more likely to stay at home. Although people in all income groups were moving less than before the crisis, higher-income earners were limiting their movement the most, especially during the workweek. In nearly every state, they began doing so days before low-income earners. The differential was particularly high in metropolitan areas with large economic inequalities (Valentino-DeVries, Lu and Dance, 2020[18]). The higher share of young people in jobs requiring a physical presence may be linked to their overrepresentation in sectors such as wholesale and retail trade, and accommodation and food services (Brussevich, Dabla-Norris and Khalid, 2020[19]). Additionally, employees of large firms are more likely to have teleworking as an option compared to SMEs (OECD, 2020[20]).

Trade-exposed regions are likely to face higher short-term risks, but could also have longer-term protective factors

World trade sharply contracted in 2020, and supply chain disruptions impeded activity in a number of sectors. This scaling back of global trade has diverse effects on regions, with places more integrated in global trade potentially hit the hardest initially. Regions with higher shares of employment in tradable sectors (see Figure 1.8)5 may face higher risks due to disruptions in trade flows, although further study is needed. The longer global trade will take to return to before COVID-19 crisis levels, the harder the downturn could be for the more globalised regions, with potentially stronger rises in unemployment, at least in the short term. However, in the medium term, if global trade returned to pre-crisis levels, more globalised regions could recover faster, in line with trends from previous crises (OECD, 2018[21]).

Figure 1.8. Share of regional employment in tradable sectors

Divides could also deepen within local labour markets, as disadvantage becomes more entrenched

COVID-19 will likely not only exacerbate divides across local labour markets, but also divides within local labour markets. The low-skilled, low-wage workers, and young people may be the most vulnerable to COVID-19-related job losses (OECD, 2020[1]). They are in the sectors most at risk (Berube and Bateman, 2020[22]), they are less likely to hold jobs that allow them to telecommute (OECD, 2020[23]), and are more likely to be on temporary contracts (OECD, 2014[24]). These same groups are also more likely to hold jobs at higher risk of automation (Nedelkoska and Quintini, 2018[25]), a process that firms may accelerate in light of the pandemic (see Chapter 2). While the global financial crisis predominantly impacted male-dominated sectors and occupations, women are more at-risk from COVID-related job losses, as they are over represented in the sectors and occupations most at-risk (OECD, 2020[1]).

The impact of COVID-19 on these groups could persist for some time. Young people, particularly those facing multiple disadvantages, can face “scarring effects” from entering the workforce during periods of high unemployment, with persistent negative impacts for their career and wages, as well as other dimensions of well-being, over the long term (Scarpetta, Sonnet and Manfredi, 2010[26]). Many people from these groups could end up facing long-term unemployment, or dropping out of the labour market all together. In places where childcare and schools remain closed or with limited in-person activities, there may also be important increases in people dropping out of the labour force because of caring responsibilities, which disproportionately impacts women.

Box 1.3. Economic inactivity and discouraged workers in regions and cities

Pre-COVID-19, there were approximately 270 million adults who are not employed or looking for a job across the OECD (i.e. the “economically inactive”6) (OECD, 2019[27]). Young people, the low-skilled, and women are more likely to be economically inactive – the same groups most at risk from COVID-19-related job losses. The economic inactivity rate is 24 percentage points higher for people with low education levels (i.e. below upper secondary education) in comparison to those having attained tertiary education. Around one in three inactive individuals across the OECD is aged 15-24 years, and in all countries, women are more likely to be economically inactive than men (Barr, Magrini and Meghnagi, 2019[28]).

There are already important regional differences in economic inactivity rates, which COVID-19 could accentuate. Across OECD countries with more than one region, the average variation between regions with the highest and the lowest economic inactivity rates is 10.5 percentage points (see Figure 1.9). The variation is less than 5 percentage points in countries such as Slovenia, Denmark and Sweden, but it is above 20 percentage points in Chile, Israel, Italy and the United States. Chile and Italy are also among those with the highest gender gap in the inactivity rate. Evidence suggests the regional differences can be linked to prior job losses, particularly related to places with an industrial legacy, as has been found to be the case in the United Kingdom (Barr, Magrini and Meghnagi, 2019[28]).

Figure 1.9. Some countries have large regional gaps in economic inactivity rates

While the reasons for economic inactivity vary, from care responsibilities to disabilities to belief that there are no jobs available, at least some of this population could and would like to work. In 2017, the share of the economically inactive who were willing to work was on average 19% across the European Union, representing around 16.6 million people. This figure is above 30% in countries such as Denmark, Italy, Austria and Switzerland (Eurostat, 2019[29]). Pre-COVID-19, in the United States, 4.4 million people were out of the labour force but would like to work, just under 5% of the inactive population (BLS, 2020[30]). However, these official figures may actually undercount the share of people who could and would like to work if the right supports were available (child care, accommodating workplaces, etc.) Additionally, the rise of teleworking could open up employment possibilities for people with disabilities, for example by removing barriers to commuting and unsuitable workplaces (Ahrendt and Patrini, 2020[31]).

Official statistics show that the share of discouraged workers tends to spike during crises, and can be one of the main drivers of increasing inactivity rates during crises. Discouraged workers are economically inactive people who report in labour force surveys that they would like to work but are not actively looking for a job because they believe none are available. The share of discouraged workers among the extended labour force (i.e. people employed, unemployed and discouraged) increased significantly following the 2008 crisis in some of the hardest hit countries (e.g. Spain, Portugal and Ireland) as well as Romania, and in some cases did not recede even as the overall economic situation improved. Emerging evidence suggests that the numbers of discouraged workers are likewise increasing as a result of COVID-19. For example, in Italy, following five years of decreases, the number of discouraged workers increased by 4.8 percent in Q2 2020 compared to Q1 2019 (Istat, 2020[32]).

Patterns in discouraged workers can vary significantly across regions. For example, in Italy, where the share of discouraged workers is relatively high, the share of discouraged workers was stable between 2008 and 2011 in most regions in the north of Italy, but increased by 2 percentage points or more in the south (Basilicata, Molise, Puglia). Among the seven regions in the south, the share of discouraged workers had returned to 2008 levels in one (Campania) by 2018. In two regions, it remained similar to the 2011 levels (Molise and Sicily), and in one region it had actually further increased (Sardinia).

Figure 1.10. The crisis caused a higher share of people to become discouraged workers in the south of Italy than the north

In some countries, relatively small changes in unemployment rates hide the fact that many formerly employed people have dropped out of the labour force all together. Pre-COVID-19, economic inactivity rates and shares of discouraged workers varied considerably across regions and changed differently as a result of the global financial crisis (see Box 1.3). In Italy, the number of inactive people grew by 5.5 percent between Q1 and Q2 2020, while the number of people officially counted as unemployed actually decreased (Istat, 2020[32]). In Poland, the number of inactive grew by over 200 000 in Q2 2020 compared to Q2 2019, accounting for most of the decreases in the number of people employed. Economic inactivity grew in particular for women and people living in urban areas (Statistics Poland, 2020[33]).

Within local economies, SMEs and the self-employed may face particular challenges

While mass layoffs at large firms make headlines, SMEs account for about 60% of employment and between 50% and 60% of value added across the OECD (OECD, 2019[34]). SMEs are overrepresented in sectors that have been highly impacted by COVID-19. On average across OECD countries, SMEs are estimated to account for 75% of employment in the most affected sectors (OECD, 2020[35]). In Ireland, for example, SMEs accounted for 79% of annual turnover in 2017 in highly affected sectors and 59% of annual turnover in highly and moderately affected sectors combined (in comparison, the share of SMEs in value added in the business economy in Ireland was 44% in 2016) (McGeever, McQuinn and Myers, 2020[36]; OECD, 2020[20]). SMEs are less equipped to manage these shocks since they have much lower equity and financial reserves to draw on than larger firms. According to surveys, more than half of SMEs faced severe losses in revenues as a result of COVID-19, with many having only a few months of reserves to withstand the crisis (OECD, 2020[20]).

On average across OECD countries, about 15% of working people are self-employed, and about one-third of these are employers. The way in which many of the self-employed engage with their customers, suppliers, staff and collaborators are being uprooted by the COVID-19 crisis. Many are losing clients, particularly where their businesses involve consumer or business services that are delivered face-to-face, fields in which the self-employed often dominate. Some of the self-employed are able to mitigate the adverse impacts by going online in terms of customer and staff interactions. However, they are often held back by low existing levels of digitalisation, for example an inability to operate through e-commerce, and emergency support measures are not reaching all self-employed people. Many do not qualify for the measures due to the nature or scale of their activities (see Chapter 3). The full impact on the COVID-19 crisis on the self-employed is not yet known as there are many uncertainties, concerning for example the duration and nature of restrictions on personal and commercial activities, the response of consumer demand and behaviours, bank liquidity supply and so on.

SMEs and the self-employed are particularly dependent on their local economies for demand and access to business support, but local economies and communities also depend on healthy SMEs. Beyond the jobs they provide, they are often active corporate citizens in their communities, and are an important component of dynamic and vital local communities. Thus, the impact of potential SME closures goes beyond just the economic activity and jobs they are directly responsible for.

Even before COVID-19 hit, the labour market picture was not as rosy as national figures suggested

Prior to COVID-19, headlines celebrated the relatively strong labour market position of many OECD countries. Just over a decade after the global financial crisis, the overall OECD unemployment rate stood at 5.4% before COVID-19 hit. This was one of the lowest rates in the last 40 years. However, even during this relatively boom time, these rosy figures masked other issues such as stagnant wage growth and a shrinking middle class. National averages also hid the fact that some places were still struggling with the legacies of the crisis when COVID-19, and as well as challenges in adjusting to ongoing structural changes.

Unemployment rates were still above 2008 levels in half of regions in 2018

Nearly half of regions still had higher unemployment rates in 2018 than in 2008 (44%). Only in one-third of countries had unemployment rates recovered in all regions, and in ten countries, no regions had yet returned to pre-crisis levels (see Figure 1.11). An even higher share of regions – two-thirds – had higher long-term unemployment rates in 2018 than 2008. In nearly one-third of regions, 40% or more of the unemployed have been out of work for 12 months more. Despite the fact that employment rates are now at record highs in most OECD countries 7 (pre-pandemic), about one-third of regions actually had 2018 employment rates below 2008 levels.

Figure 1.11. Half of regions had not recovered to 2008 unemployment levels by 2018

Regional disparities in unemployment remain stark, and are growing or shrinking for the wrong reasons in over half of OECD countries

In over half of OECD countries, there is a two-fold or more difference in unemployment rates between the best and worst performing regions (see Figure 1.12 and Annex Figure 1.A.1). Unsurprisingly, OECD countries with higher national unemployment rates tended to have the largest regional gaps.8 In Turkey and Italy, regional disparities between the best and worst performing regions were around 19 percentage points, while in Spain and Greece, they were around 14 percentage points. In contrast, Asian countries (Japan and Korea) and some Scandinavian countries (Denmark and Norway) have both relatively low unemployment rates and low regional disparities.

Accordingly, the same national unemployment rate at can actually hide very different regional patterns. For example, both Austria and Switzerland had an unemployment rate of 4.9% in 2018, but in Austria, unemployment actually varied over four-fold across regions, from 2.4% in Tyrol to 10.1% in Vienna. In Switzerland, the regional variation is still significant (over two-fold) but not nearly as stark.

Figure 1.12. Regional unemployment rates can vary by more than two-fold within some countries

Across countries, unemployment challenges concentrate in different types of regions. For example, in the Czech Republic and the Slovak Republic, the unemployment rate in the capital region was close to half of that of the national rate, while in Belgium and Austria, unemployment in the capital region was twice the national average. As described in Box 1.4, this may reflect the varying patterns of urban and rural unemployment across countries as a result of both economic and demographic characteristics.

However, in general, the best performing regions tend to stay on top, and the worst performers tend to stay on the bottom over time. In 15 countries, the region with the highest unemployment rate is the same in both 2008 and 2018. This aligns with previous OECD research that shows that employment challenges and successes tend to anchor in specific regions and spaces (OECD, 2005[37]).

Box 1.4. Cities drive growth, but can also concentrate unemployment

Large cities, capital regions and other more urbanised places consistently show the strongest performance along a number of economic indicators. Metropolitan areas (i.e., urban areas with population of greater than 500 000) account for 55% of the total OECD population, 59% of the employed, and around 60% of the total GDP in the OECD. Pre-pandemic, GDP growth was 32% higher in metropolitan areas than in the rest of the country since 2000. Capital regions host one in four firms in their countries and have a net firm creation rate over 60% higher compared to other regions (OECD, 2018[38]). Employment growth in more urbanised regions outpaced growth in rural or intermediate regions in most countries between 2008 and 2018, and capital regions specifically had the highest relative share of net employment growth in half of OECD countries with more than one region.

However, strong economic performance and growth does not always translate into lower unemployment rates. On average in the EU, unemployment is highest in urban areas, followed by towns and suburbs and then rural areas (8.1%, 7.1%, and 6.3% respectively) (Eurostat, 2020[39]). While cities and urban areas host a higher share of high-skilled and high-wage workers, they also concentrate inequalities and host many more vulnerable populations, such as low-skilled workers and immigrants. High levels of residential segregation in cities can also impede access to job opportunities for some populations, and be linked to discrimination in hiring and a lack of beneficial professional networks (OECD, 2018[40]).

However, across countries, there are different trends in terms of where unemployment is highest. For example, of limited countries with available data, in Korea, Japan, Hungary, and Switzerland, unemployment rates are highest in large metropolitan or metropolitan TL3 regions for the latest year data is available. In France, Norway, Spain, and Sweden, it is highest in non-metropolitan areas with access to a metro. In Denmark and Latvia, unemployment is highest in remote rural areas (OECD, 2020[41]).

Source: Eurostat (2020[39]); OECD (2020[41]); OECD (2018[38]); and OECD (2018[40]).

In the decade following the global financial crisis, regional variation in unemployment rates shrank in most countries (19/32 OECD countries with more than one region and available data plus Romania). (Figure 1.13).9 The good news is that in most countries with a shrinking gap, gaps were closing for good reasons, i.e. because unemployment rates decreased more in regions where they were relatively high at the beginning of the period. However, in five countries (Canada, Finland, New Zealand, Portugal, and Slovenia), gaps were closing for the wrong reasons: shrinking gaps were mainly driven by increases in unemployment rates in the best performing regions. In countries where gaps were increasing, this was typically driven by a significant increase in the unemployment rates in the regions that were already the worst performing in 2008. In line with previous studies, these findings suggest that regions with low levels of unemployment have limited fluctuation over time whereas regions with higher unemployment tend to show more variation (Beyer and Stemmer, 2016[42]).

Figure 1.13. Regional unemployment gaps shrunk in just over half of countries in the ten years after the crisis, but not always for good reasons

Jobs are increasingly geographically concentrated in most OECD countries

As regions have displayed different capacities to attract and retain jobs and workers over time, employment opportunities have become increasingly geographically concentrated. Jobs (as measured by the number of people employed) still lagged behind 2008 levels in one-third of OECD regions in 2018. Looking at a longer time period (2000-2018), in most countries, jobs (as measured by the number of people employed), have become more geographically concentrated (in 14/27 OECD countries with available data plus Romania, concentration increased by 1% or more; see Figure 1.14). In most of these countries, the concentration of high-skilled jobs has increased even more than for jobs in general. While these patterns could reflect both economic and demographic trends, they suggest a shifting geography of opportunity in most OECD countries, with growing divides between leading and lagging places.

Figure 1.14. Jobs have become more concentrated in most countries

Looking at the past 10 years specifically, more urbanised regions tended to concentrate employment growth. Capital regions specifically saw the highest relative share of employment growth in about half of OECD countries with more than one region. Given that urban areas and capital regions already host an outsized share of employment in general, these trends help to explain why employment has become more concentrated over time.

Job quality is a growing concern, especially in places struggling with other labour market challenges

The health of local labour markets cannot be determined just by the number of jobs; the quality of local jobs also matters. While job quality can be measured in a variety of ways, one indicator is the incidence of non-standard work, including temporary and involuntary part-time work. In general, temporary work has increased somewhat across the OECD over the long term, albeit with some cross-country differences (OECD, 2016[43]; 2018[44]). Part-time work has also been generally increasing in recent decades. While the increase in part-time work in some cases can be considered a positive development, and may reflect an increase of female labour market performance and a trend towards more work-life balance, an increase in involuntary part-time employment is more worrying. Indeed, involuntary part-time employment (employees working 30 hours or less per week who report either that they could not find a full-time job or that they would like to work more hour) has increased in most OECD countries between 2006 and 2017, particularly in those countries places hit hardest by the crisis (OECD, 2019[45]).

Non-standard workers generally enjoy lower levels of job security and social protection compared to workers in standard employment relationships. Following the 2008 crisis, workers with temporary contracts were disproportionately affected by job losses, although employers also relied heavily on temporary contracts in hiring during the recovery period. Early evidence from the COVID-19 crisis likewise suggests that they are amongst the hardest hit. They are highly represented in some of the most impacted sectors, such as arts and entertainment and tourism; and employers may choose to not renew temporary contracts even when dismissal protection regulations prevent them from laying off permanent workers. Evidence from France, Italy and Canada suggest workers on temporary contracts were indeed among the first to lose their jobs in the spring (OECD, 2020[1]).

Temporary work is not evenly spread across territories, and is more common in regions with a lower-educated workforce, higher unemployment, and a smaller share of gross value added in tradable sectors (OECD, 2018[44]). In over half of European countries with more than one region, the share of temporary employment varies over 5 percentage points across regions, and in several, it varied over 10 percentage points. Overall, low-skilled workers are at higher risk of being in temporary work than the higher skilled, and that likelihood is even higher in rural areas than in cities (OECD, 2018[44]).10

Figure 1.15. Temporary employment patterns are not uniform within countries

Some places weathered the last storm better than others

Previous economic shocks have had very different impacts across geographies, and the same will likely be true for COVID-19, albeit some of the dynamics this time may be different. The global financial crisis caused employment to decrease in almost all regions, but the scale of these losses and the time it took employment to rebound varied considerably across territories. The hardest hit places lost 20% or more of their jobs at their respective lowest points, and in many places, employment levels have taken five years or more to recover. While the COVID-19 shock is of a different scale and nature than any other shock in recent history, patterns of local resilience to the last crisis suggest that the hardest hit places will again not bounce back quickly.

While local resilience can be defined and measured in a variety of ways (see Box 1.5), this analyses focuses on how resilient local employment was to the 2008 crisis, i.e. how the number of people employed evolved over the course of the crisis.11 More specifically, it considers how employment levels changed between 2008 and the respective local trough (i.e. the lowest point) during the crisis, and how long it took employment to bottom out and subsequently recover.

Box 1.5. The concept of local economic resilience

The term resilience was first used in engineering and ecology discourses in the 1970s, but it soon spread to psychology and the broader social sciences. Since then, a significant body of research has explored how the concept can be applied to local and regional economies, how it can be operationalised, and the normative assumptions that these definitions and methodologies imply. Resilience has been used to refer to the adaptive, absorptive or reactionary capacity of systems in response to both abrupt shocks and long-term threats, such as climate change. Resilience can be built in response to a range of economic, financial, social or natural shocks, from earthquakes to recessions.

While definitions of local economic resilience vary, it can generally be understood as the ability of a local economy to resist, recover, and adapt in the face of a shock. Various indicators have been used to operationalise and measure local resilience, from economic indicators (e.g. productivity and output) to labour market indicators (e.g. unemployment rates and employment levels) to social indicators (e.g. poverty rates). Likewise, researchers have explored a variety of factors that could influence local resilience. Factors typically considered include local economic and labour market structures and performance, levels of social capital and inclusion, and other place-based factors, such as local environmental factors or geography. Increasing attention has also been paid to how governance quality and arrangements, as well as international, national and subnational policies impact regional resilience differently across places.

Despite growing attention to this subject, there is no general consensus as to what makes regions resilient, or even a normative agreement on what a resilient region looks like. Can a region that bounces back quickly following a shock in terms of output but with high rates of poverty be considered resilient? If a region relies on large extractive sectors to resist declines in employment following a shock, can this be considered a resilient region over the long term? Accordingly, further research and debate is needed on the concept of local resilience, particularly as COVID-19 magnified and exposed fragilities in our economies and societies in new ways.

Source: Boschma, (2015[46]); Bristow and Healy (2020[47]); ESPON & Cardiff University (2014[48]); Martin et al. (2016[49]); OECD (2014[50]); and Sensier, Bristow and Healy (2016[51]).

Of course, the COVID-19 economic shock is of a scale and nature unseen in recent history, limiting the applicability of some of the lessons from the previous crisis. Not only will the challenges be larger, but the protective and risk factors could be different. For example, while evidence suggest that urban areas tended to fare better in the last crisis, there is an ongoing debate as to whether cities and denser areas are more vulnerable to the spread of the virus during this crisis. Additionally, many regions relied on tourism to pull themselves out of the last crisis (Psycharis, Kallioras and Pantazis, 2014[52]), while tourism dependent regions are likely more vulnerable to this shock. Indeed, even pre-COVID-19, there was a broader ongoing debate within the resilience research as to how static protective and risk factors are over time, across geographies, and in response to different types of shocks (Martin and Gardiner, 2019[53]). Despite these caveats, the experience of previous crisis as well as the early learnings from this crisis can give an indication of what is to come for local economies.

Employment decreased in four-fifths of regions, with some places losing 20% or more of their jobs at their respective lows

The global financial crisis caused wide scale employment losses: in roughly eighty percent of regions, the number of people employed fell at some point post-2008. Unsurprisingly, this largely reflects national trends: of the 20% of regions where employment did not decline, most were in countries where national employment did not decline or only declined marginally (i.e. Turkey, Mexico, Israel, and Luxembourg). Only a handful of regions were able resist any declines in employment, despite employment decreasing in their respective countries overall.

At their respective lowest points, employment declined by over 20% in some of the hardest hit regions in Spain and Greece, and by over 10% in some places in the US, Denmark, Italy, Poland, Portugal, and Turkey, as well as Romania. Within countries with more than one region, employment declined by 7 percent points more in the worst performing regions compared to the best performers on average.12 As shown in Figure 1.16, this difference exceeds 10 percent points in 7 OECD countries, as well as Romania. These large disparities can be seen both in countries that experienced large employment declines at the national level (e.g. Greece, Spain, Italy, as well as Romania), as well as countries that experienced relatively small or no declines nationally (e.g. Mexico and Turkey, where the best performing region actually never saw employment declines over this period).

Figure 1.16. Employment declined by over 10% in the hardest hit regions, while in others, it never dropped below 2008 levels

It is important to note that employment hit its low point before starting to rebound at different times across regions. National exposure and vulnerabilities to different waves of the crisis can help to explain cross-country differences in terms of when employment reached its respective low (e.g. the collapse of the subprime mortgage industry in the US vs. the Eurozone debt crisis). However, variations within countries also suggests that there were different vulnerabilities across any given country’s regions. One underlying factor may be local sectoral specialisation, both in terms of the sensitivity of local sectors to the business cycle, and how sectors are impacted differently over time by different waves of the crisis. Sectors such as construction, durable manufacturing and business services tend to be most sensitive to the business cycle. Following the bursting of the housing bubble, the construction industry was immediately impacted in a number of countries, and job losses then spread to manufacturing and business services (OECD, 2009[3]). In Europe, high shares of local public sector employment was initially a protective factor against job losses, but later likely became more of a risk factor in countries that implemented large austerity measures (ESPON & Cardiff University, 2014[48]). For example, in the Czech Republic, unemployment increased more in rural regions with export-oriented economies over the period of 2008-2010, while larger cities were hit harder in 2012-2013 following the implementation of austerity measures (Ženka, Slach and Pavlík, 2019[54]).

The crisis magnified underlying local weaknesses

Places that experienced larger employment losses tended to already be struggling with other labour market challenges. Relative to national values, evidence suggests that larger employment losses were associated with having higher unemployment rates, a less educated workforce, and lower labour productivity in 2008 (Annex Figure 1.A.3). While further study is needed to confirm these relationships, they do align with other research on regional resilience that suggests that downturns accentuate local weaknesses and reward local strengths. For example, other research has found a positive relationship between having a highly skilled workforce and resilience in European regions (ESPON & Cardiff University, 2014[48]) and UK local authorities (Bristow, Healy and Kitsos, 2020[55]), Other work in the United States has shown specific types of skills (such as people or cognitive skills) as being especially important for a quicker local recovery (Weinstein and Patrick, 2020[56]).

However, the broader local development pathway may have been as, if not more important, than any static measure of labour market health. In particular, the shock may have exposed fragility in regional growth models, regardless of performance on labour market indicators at any single point in time. Previous OECD research found that the places that lost more jobs between 2008 and 2009 tended to experience faster GDP growth and larger reductions in unemployment from 1999 to 2007. (OECD, 2011[57]). Likewise, European regions that experienced high levels of employment growth prior to the 2008 crisis demonstrated lower levels of resiliency (ESPON & Cardiff University, 2014[48]), and having a more stable growth pattern in the lead up to the last crisis was associated with greater resilience (Webber, Healy and Bristow, 2018[58]). Similar results have been found for the response of local GDP to the crisis (OECD, 2018[21]). However, these patterns may be specifically related to the unsustainable growth patterns leading up to the global financial crisis rather than a dynamic underlying regional resilience to crises more generally.

There is also evidence that a more diversified, rather than specialised, economic structure promotes resilience. Regions vary considerably in terms of the degree of local economic diversification and specialisation. The largest tradable cluster accounts for less than 5% of the workforce in some European regions, whereas in others, it accounts for more than 40% of the workforce (OECD, 2018[21]). While hosting a diversity of sectors may make a region more vulnerable to taking some type of hit from any given shock, it minimises the risk that any given shock will have a large negative impact on the local economy overall. In particular, having a variety of skill-related industries that have few input-output relationships but are of a related variety is thought to enhance regional resilience over the longer term (Boschma, 2015[46]). Indeed, new OECD research on the resilience of U.S. counties shows that the ability of workers to move between local sectors and occupations as being an important factor for local resilience, particularly in rural areas and places with relatively poor performance (Box 1.6). However, the relationship between economic diversity and regional performance is not straightforward – the added value of a more diverse economic structure can vary at different stages of development (OECD, 2018[21]) and may contribute to better performance more during times of shocks than when the economy is relatively strong (Brown and Greenbaum, 2017[59]).

Box 1.6. Local “rewiring” in the United States

New OECD research suggests that factors associated with greater employment growth were different for growing and stagnating (or declining) counties. Looking at the period before and after the global financial crisis, for more well-off places (i.e. those in the middle and at the higher end of economic performance distribution), a local industrial structure concentrated in industries growing nationally (a positive demand shock) helps to boost employment growth and to cut poverty rates. Less well-off counties appear to be unable to benefit from these national growth processes, falling further behind.

Both rural and lagging places performed significantly better in terms of employment growth post-recession if they had an industrial composition that facilitated greater inter-sectoral worker flows (e.g., workers from one sector were able to move into another) and if they enjoyed larger changes in occupational structure, with relatively more people moving from one occupation to another.

These findings suggest that growth of local economies increasingly depends on their ability to “rewire” and adjust to changing labour market realities. Local “rewiring” appears to work particularly well for rural and weaker-performing counties in the United States. Accordingly, encouraging labour flows within the region, ideally from lower- to better-performing sectors, industries, firms and occupations, may be particularly important for lagging regions.

Source: Partridge and Tsvetkova (2020[60])

Cities and capital regions were generally more resilient on average, but not across the board

On average, capital regions and other more urbanised regions saw smaller decreases in employment at their respective lows, although patterns differed significantly across countries. In Austria, Belgium, Sweden, and Switzerland, as well as Romania, employment in capital regions never fell below 2008 levels, despite national losses at some point. However, this pattern does not hold true across the board, particularly in some of the hardest hit countries. In Portugal and Greece, employment declined relatively more in the capital region than in most other regions.

These findings align with previous research that shows considerable variation in resilience across cities and urban regions. Urban regions showed considerable variation in job losses immediately following the 2008 shock, particularly when compared to the pre-crisis period (OECD, 2011[57]). Likewise, other research has shown that patterns of resilience can vary across types of urban areas. For example, in Europe, the presence of a second-tier city in a region made a particularly positive difference (ESPON & Cardiff University, 2014[48]). The United Kingdom is a particularly striking case in point. In studying the resilience of UK cities over four major recessions since 1970, Martin and Gardiner (2019[53]) found varying patterns of resiliency between cities in the north and south over time, with London demonstrating increasingly strong resilience over time. For the two earlier recessions, cities that resisted larger employment shocks also recovered more quickly, while for the last two recessions, this relationship disappeared and even showed a slightly negative pattern.

The hardest hit places have taken years to recover, if at all

Employment recovered more quickly in same places than others. In about half of regions where employment declined in the six years following the initial crisis, the recovery took three years or more, or has not yet happened as of 2018. Unsurprisingly, those places that took smaller employment hits recovered more quickly, while it took longer for places that took larger hits to rebound. This suggests that the negative impacts of shocks can linger for years in the hardest hit places. Other research looking at longer time frames has found that the negative effects can persist for even longer than the time period covered in this analysis. Looking back across the previous five recessions in the United States, the most affected local labour markets experienced employment, population and wage losses that persisted for at least a decade (Hershbein and Stuart, 2020[61]).

Figure 1.17. Employment recovered at a different pace across regions

The local persistence of economic distress can result from a combination of factors. For one, many job losses during recessions are not cyclical, but rather reflect an acceleration of structural changes. Accordingly, these jobs are unlikely to recover even when the economic situation improves. This can be especially problematic for local economies where concentrated job losses in specific sectors can have negative spillovers for jobs in the local economy more generally (see Chapter 2). Poor labour market outcomes, such as unemployment and low wages, can be associated with a broader range of quality of life challenges at the individual and community level, from poor mental and physical health to drug abuse to crime. Likewise, local downturns can put significant pressure on local public budgets, impacting local quality of life and public services such as education and infrastructure. In the short term, this can make it hard to attract new residents and businesses, and over the longer-term, affect intergenerational education and labour market outcomes. Many of these factors will be relevant for the COVID-19 recovery, and perhaps even magnified.

Conclusion

All local economies will feel the impacts of COVID-19: large cities where polarised labour markets means strong divides between high-skilled workers with relatively secure jobs and low-skilled workers in face-to-serve service and retail jobs at risk; tourist destinations struggling with historically low visitor numbers; manufacturing regions dealing with supply chain interruptions. Depending on the spread of the virus and the response of consumers, businesses, and investors, unemployment will spike to different levels and at different times across places. But if past patterns hold true, the hardest hit places could struggle for years to come. Even as national economies eventually turn around, targeted actions will be needed to ensure that some places are not left even further behind.

References

[31] Ahrendt, D. and V. Patrini (2020), How to use the surge in teleworking as a real chance to include people with disabilities, Eurofound, https://www.eurofound.europa.eu/fr/publications/blog/how-to-use-the-surge-in-teleworking-as-a-real-chance-to-include-people-with-disabilities (accessed on 16 September 2020).

[28] Barr, J., E. Magrini and M. Meghnagi (2019), “Trends in economic inactivity across the OECD: The importance of the local dimension and a spotlight on the United Kingdom”, OECD Local Economic and Employment Development (LEED) Working Papers, No. 2019/09, OECD Publishing, Paris, https://dx.doi.org/10.1787/cd51acab-en.

[15] Basso, G. et al. (2020), “The new hazardous jobs and worker reallocation”, OECD Social, Employment and Migration Working Papers, No. 247, OECD Publishing, Paris, https://dx.doi.org/10.1787/400cf397-en.

[22] Berube, A. and N. Bateman (2020), Who are the workers already impacted by the COVID-19 recession?, Brookings Institution, https://www.brookings.edu/research/who-are-the-workers-already-impacted-by-the-covid-19-recession/ (accessed on 27 April 2020).

[42] Beyer, R. and M. Stemmer (2016), “Polarization or convergence? An analysis of regional unemployment disparities in Europe over time”, Economic Modelling, Vol. 55, pp. 373-381, https://doi.org/10.1016/j.econmod.2016.02.027.

[30] BLS (2020), Table A-16. Persons not in the labor force and multiple jobholders by sex, not seasonally adjusted, https://www.bls.gov/news.release/empsit.t16.htm (accessed on 31 March 2020).

[46] Boschma, R. (2015), “Towards an Evolutionary Perspective on Regional Resilience”, Regional Studies, Vol. 49/5, https://doi.org/10.1080/00343404.2014.959481.

[47] Bristow, G. and A. Healy (2020), Handbook on Regional Economic Resilience, Edward Elgar Publishing, https://doi.org/10.4337/9781785360862.

[55] Bristow, G., A. Healy and T. Kitsos (2020), “Economic resilience in Great Britain: an empirical analysis at the local authority district level”, in Handbook on Regional Economic Resilience, Edward Elgar Publishing, https://doi.org/10.4337/9781785360862.00017.

[59] Brown, L. and R. Greenbaum (2017), “The role of industrial diversity in economic resilience: An empirical examination across 35 years”, Urban Studies, Vol. 54/6, pp. 1347-1366, https://doi.org/10.1177/0042098015624870.

[19] Brussevich, M., E. Dabla-Norris and S. Khalid (2020), “Who will Bear the Brunt of Lockdown Policies? Evidence from Tele-workability Measures Across Countries”, IMF Working Paper, No. 20/88, IMF, https://www.imf.org/en/Publications/WP/Issues/2020/06/12/Who-will-Bear-the-Brunt-of-Lockdown-Policies-Evidence-from-Tele-workability-Measures-Across-49479 (accessed on 8 September 2020).

[12] DARES (2019), Quels sont les salariés concernés par le télétravail ?, https://dares.travail-emploi.gouv.fr/IMG/pdf/dares_analyses_salaries_teletravail.pdf (accessed on 3 June 2020).

[48] ESPON & Cardiff University (2014), ECR2 - Economic Crisis: Resilience of Regions, https://www.espon.eu/programme/projects/espon-2013/applied-research/ecr2-economic-crisis-resilience-regions (accessed on 14 August 2020).

[39] Eurostat (2020), Urban and rural living in the EU, https://ec.europa.eu/eurostat/en/web/products-eurostat-news/-/EDN-20200207-1 (accessed on 15 September 2020).

[29] Eurostat (2019), People outside the labour market - Statistics Explained, https://ec.europa.eu/eurostat/statistics-explained/index.php/People_outside_the_labour_market#C2.A0.25_of_the_economically_inactive_population_are_interested_to_work (accessed on 28 October 2019).

[63] Google LLC (2020), COVID-19 Community Mobility Reports, https://www.google.com/covid19/mobility/ (accessed on 9 September 2020).

[10] Guyot, K. and I. Sawhill (2020), Telecommuting will likely continue long after the pandemic, Brookings Institution, https://www.brookings.edu/blog/up-front/2020/04/06/telecommuting-will-likely-continue-long-after-the-pandemic/ (accessed on 26 May 2020).

[61] Hershbein, B. and B. Stuart (2020), Recessions and Local Labor Market Hysteresis, W.E. Upjohn Institute, https://doi.org/10.17848/wp20-325.

[32] Istat (2020), Il mercato del lavoro, https://www.istat.it/it/archivio/247045 (accessed on 15 September 2020).

[53] Martin, R. and B. Gardiner (2019), “The resilience of cities to economic shocks: A tale of four recessions (and the challenge of Brexit)”, Papers in Regional Science, Vol. 98/4, pp. 1801-1832, https://doi.org/10.1111/pirs.12430.

[49] Martin, R. et al. (2016), “How Regions React to Recessions: Resilience and the Role of Economic Structure”, Regional Studies, Vol. 50/4, pp. 561-585, https://doi.org/10.1080/00343404.2015.1136410.

[36] McGeever, N., J. McQuinn and S. Myers (2020), “SME liquidity needs during the COVID-19 shock”, Financial Stability Notes, https://www.centralbank.ie/docs/default-source/publications/financial-stability-notes/no-2-sme-liquidity-needs-during-the-covid-19-shock-(mcgeever-mcquinn-and-myers).pdf?sfvrsn=6.

[25] Nedelkoska, L. and G. Quintini (2018), “Automation, skills use and training”, OECD Social, Employment and Migration Working Papers, No. 202, OECD Publishing, Paris, https://dx.doi.org/10.1787/2e2f4eea-en.

[16] ODOXA (2020), Le Covid-19 bouleversera durablement le rapport au travail des Français, http://www.odoxa.fr/sondage/covid-19-bouleverse-deja-modifiera-durablement-rapport-francais-travail/ (accessed on 3 June 2020).

[11] ODOXA (2020), Télétravail : les Français tirent un premier bilan très positif malgré des problèmes d’équipements et des craintes sur la vie personnelle, http://www.odoxa.fr/sondage/teletravail-francais-tirent-premier-bilan-tres-positif-malgre-problemes-dequipements-craintes-vie-personnelle/ (accessed on 3 June 2020).

[13] OECD (2020), “Capacity for remote working can affect lockdown costs differently across places”, OECD Policy Responses to Coronavirus (COVID-19), http://www.oecd.org/coronavirus/policy-responses/capacity-for-remote-working-can-affect-lockdown-costs-differently-across-places-0e85740e/ (accessed on 20 August 2020).

[23] OECD (2020), “COVID-19: Protecting people and societies”, OECD Policy Responses to Coronavirus (COVID-19, https://read.oecd-ilibrary.org/view/?ref=126_126985-nv145m3l96&title=COVID-19-Protecting-people-and-societies (accessed on 27 April 2020).

[9] OECD (2020), “Culture shock: COVID-19 and the cultural and creative sectors”, OECD Policy Responses to Coronavirus (COVID-19), https://www.oecd.org/coronavirus/policy-responses/culture-shock-covid-19-and-the-cultural-and-creative-sectors-08da9e0e/ (accessed on 13 September 2020).

[62] OECD (2020), Delineating Functional Areas in All Territories, OECD Territorial Reviews, OECD Publishing, Paris, https://dx.doi.org/10.1787/07970966-en.

[6] OECD (2020), “Evaluating the initial impact of COVID-19 containment measures on economic activity”, OECD Policy Responses to Coronavirus (Covid-19), https://www.oecd.org/coronavirus/policy-responses/evaluating-the-initial-impact-of-covid-19-containment-measures-on-economic-activity/ (accessed on 23 April 2020).

[7] OECD (2020), “From pandemic to recovery: Local employment and economic development”, OECD Policy Responses to Coronavirus (Covid-19), http://www.oecd.org/coronavirus/policy-responses/from-pandemic-to-recovery-local-employment-and-economic-development-879d2913/.

[1] OECD (2020), OECD Employment Outlook 2020: Worker Security and the COVID-19 Crisis, OECD Publishing, Paris, https://dx.doi.org/10.1787/1686c758-en.

[14] OECD (2020), OECD Regions and Cities at a Glance 2020, OECD Publishing, Paris, https://dx.doi.org/10.1787/959d5ba0-en.

[41] OECD (2020), “Regional labour markets”, OECD Regional Statistics (database), https://dx.doi.org/10.1787/f7445d96-en (accessed on 15 September 2020).

[20] OECD (2020), “SME Policy Responses”, OECD Policy Responses to Coronavirus (Covid-19), https://read.oecd-ilibrary.org/view/?ref=119_119680-di6h3qgi4x&title=Covid-19_SME_Policy_Responses (accessed on 27 April 2020).

[35] OECD (2020), Statistical Insights: Small, Medium and Vulnerable, https://www.oecd.org/sdd/business-stats/statistical-insights-small-medium-and-vulnerable.htm.

[8] OECD (2020), “Tourism Policy Responses to the coronavirus (COVID-19)”, OECD Policy Responses to Coronavirus (COVID-19), https://www.oecd.org/coronavirus/policy-responses/tourism-policy-responses-to-the-coronavirus-covid-19-6466aa20/ (accessed on 6 August 2020).

[27] OECD (2019), “Labour Market Statistics: Labour force statistics by sex and age (Edition 2019)”, OECD Employment and Labour Market Statistics (database), https://dx.doi.org/10.1787/929f18c4-en (accessed on 15 September 2020).

[45] OECD (2019), OECD Employment Outlook 2019: The Future of Work, OECD Publishing, Paris, https://dx.doi.org/10.1787/9ee00155-en.

[34] OECD (2019), OECD SME and Entrepreneurship Outlook 2019, OECD Publishing, Paris, https://dx.doi.org/10.1787/34907e9c-en.

[40] OECD (2018), Divided Cities: Understanding Intra-urban Inequalities, OECD Publishing, Paris, https://dx.doi.org/10.1787/9789264300385-en.

[44] OECD (2018), Job Creation and Local Economic Development 2018: Preparing for the Future of Work, OECD Publishing, Paris, https://dx.doi.org/10.1787/9789264305342-en.

[38] OECD (2018), OECD Regions and Cities at a Glance 2018, OECD Publishing, Paris, https://dx.doi.org/10.1787/reg_cit_glance-2018-en.

[21] OECD (2018), Productivity and Jobs in a Globalised World: (How) Can All Regions Benefit?, OECD Publishing, Paris, https://dx.doi.org/10.1787/9789264293137-en.

[43] OECD (2016), OECD Employment Outlook 2016, OECD Publishing, Paris, https://dx.doi.org/10.1787/empl_outlook-2016-en.

[24] OECD (2014), OECD Employment Outlook 2014, OECD Publishing, Paris, https://dx.doi.org/10.1787/empl_outlook-2014-en.

[50] OECD (2014), OECD Regional Outlook 2014: Regions and Cities: Where Policies and People Meet, OECD Publishing, Paris, https://dx.doi.org/10.1787/9789264201415-en.

[57] OECD (2011), OECD Regions at a Glance 2011, OECD Publishing, Paris, https://dx.doi.org/10.1787/reg_glance-2011-en.

[3] OECD (2009), OECD Employment Outlook 2009: Tackling the Jobs Crisis, OECD Publishing, Paris, https://dx.doi.org/10.1787/empl_outlook-2009-en.

[37] OECD (2005), OECD Employment Outlook 2005, OECD Publishing, Paris, https://dx.doi.org/10.1787/empl_outlook-2005-en.

[5] Office for National Statistics (2020), Coronavirus and the latest indicators for the UK economy and society: 5 November, https://www.ons.gov.uk/peoplepopulationandcommunity/healthandsocialcare/conditionsanddiseases/bulletins/coronavirustheukeconomyandsocietyfasterindicators/latest (accessed on 7 November 2020).

[60] Partridge, M. and A. Tsvetkova (2020), “Local ability to rewire and socioeconomic performance: Evidence from US counties before and after the Great Recession”, OECD Local Economic and Employment Development (LEED) Papers, No. 2020/04, OECD Publishing, Paris, https://dx.doi.org/10.1787/31b980f6-en.

[52] Psycharis, Y., D. Kallioras and P. Pantazis (2014), “Economic crisis and regional resilience: detecting the ‘geographical footprint’ of economic crisis in Greece”, Regional Science Policy & Practice, Vol. 6/2, pp. 121-141, https://doi.org/10.1111/rsp3.12032.

[17] Reeves, R. and J. Rothwell (2020), Class and COVID: How the less affluent face double risks, The Brookings Institution, https://www.brookings.edu/blog/up-front/2020/03/27/class-and-covid-how-the-less-affluent-face-double-risks/ (accessed on 8 June 2020).

[26] Scarpetta, S., A. Sonnet and T. Manfredi (2010), “Rising Youth Unemployment During The Crisis: How to Prevent Negative Long-term Consequences on a Generation?”, OECD Social, Employment and Migration Working Papers, No. 106, OECD Publishing, Paris, https://dx.doi.org/10.1787/5kmh79zb2mmv-en.

[51] Sensier, M., G. Bristow and A. Healy (2016), “Measuring Regional Economic Resilience across Europe: Operationalizing a complex concept”, Spatial Economic Analysis, Vol. 11/2, pp. 128-151, https://doi.org/10.1080/17421772.2016.1129435.

[33] Statistics Poland (2020), Information regarding the labour market in the second quarter of 2020 (preliminary data), https://stat.gov.pl/en/topics/labour-market/working-unemployed-economically-inactive-by-lfs/information-regarding-the-labour-market-in-the-second-quarter-of-2020-preliminary-data,8,35.html (accessed on 15 September 2020).

[4] Tsvetkova, A., S. Grabner and W. Vermeulen (2020), “Labour demand weakening during the COVID-19 pandemic in US cities: Stylised facts and factors related to regional resilience”, OECD Local Economic and Employment Development (LEED) Papers, No. 2020/06, OECD Publishing, Paris, https://dx.doi.org/10.1787/700d91ba-en.

[2] U.S. Bureau of Labor Statistics (2020), Over-the-Year Change in Unemployment Rates for States, https://www.bls.gov/web/laus/laumstch.htm (accessed on 5 August 2020).

[18] Valentino-DeVries, J., D. Lu and G. Dance (2020), Location Data Says It All: Staying at Home During Coronavirus Is a Luxury, The New York Times, https://www.nytimes.com/interactive/2020/04/03/us/coronavirus-stay-home-rich-poor.html (accessed on 8 June 2020).

[58] Webber, D., A. Healy and G. Bristow (2018), “Regional Growth Paths and Resilience: A European Analysis”, Economic Geography, Vol. 94/4, pp. 355-375, https://doi.org/10.1080/00130095.2017.1419057.

[56] Weinstein, A. and C. Patrick (2020), “Recession‐proof skills, cities, and resilience in economic downturns”, Journal of Regional Science, Vol. 60/2, pp. 348-373, https://doi.org/10.1111/jors.12446.

[54] Ženka, J., O. Slach and A. Pavlík (2019), “Economic resilience of metropolitan, old industrial, and rural regions in two subsequent recessionary shocks”, European Planning Studies, Vol. 27/11, pp. 2288-2311, https://doi.org/10.1080/09654313.2019.1638346.

Annex 1.A. Additional figures

Annex Figure 1.A.1. Regional variation in unemployment rates, 2018

Annex Figure 1.A.2. Regional range in long-term unemployment rates

Annex Figure 1.A.3. Regions struggling with other labour market challenges tended to lose higher shares of employment

Notes

← 1. Local labour markets vary in size and shape and often do not correspond to administrative boundaries, making it difficult to collect internationally comparable data that correspond to travel-to-work or functional areas. Often, functional local labour markets can operate on a scale smaller than the OECD’s TL2 regional classification, but span several TL3 regions. This publication predominantly uses TL2 data to ensure as broad a coverage as possible, as data availability is limited across countries and time for TL3 regions. For many analyses, the regional variation at the TL2 level within a country should be considered the lower bound of the actual variation across local labour markets. For more information, see (OECD, 2018[44]) and (OECD, 2020[62]).

← 2. Differences in unemployment rates between countries should be interpreted with caution, particularly in relation to COVID-19. They are influenced by methodological differences in how workers are classified in official surveys, such as those on temporary layoffs or short-time work schemes, and preliminary figures may be revised as further data becomes available.

← 3. These estimates are based on an analysis of jobs at risk during the first wave of containment measures in spring 2020. These results were first presented in OECD (2020), “From pandemic to recovery: Local employment and economic development”, OECD Policy Responses to Coronavirus (Covid-19).

← 4. This analysis was first presented in OECD (2020), “Capacity for remote working can affect lockdown costs differently across places”, OECD Policy Responses to Coronavirus (Covid-19). Further information is drawn from OECD (2020), “Exploring policy options on teleworking: Steering local economic and employment development in the time of remote work”, OECD Local Economic and Employment Development (LEED) Papers, as well as OECD (2020), Regions and Cities at a Glance 2020.

← 5. The definition of tradable activities in this report allows for comparison across regions in most OECD countries. As disaggregated data is not universally available, harmonisation requires sectoral aggregation. National estimates of tradable activities can therefore differ and offer more precise estimates for individual countries. For example, in logistics hubs, these figures may understate the share of employment in tradeable sectors, as the Transport, Retail and Hospitality group (GHI) combines jobs in both tradeable and non-tradeable sectors, but has been classified as non-tradeable for the purposes of these estimates. Additionally, they are not intended to show how tradeable sectors contribute to regional and national GVA, as there are important productivity differences across regions and countries.

← 6. People who are not employed or looking for a job are generally defined as economically inactive.

← 7. The United States is also a notable exception to the longer term trend of increasing employment rates – employment rates remain below their early 2000 peak.

← 8. The strength of the relationship varies based on the measure of regional variation used (i.e. range, coefficient of variation and 80/20 range) but is always positive.

← 9. Robustness checks using the coefficient of variation and the 80/20 range as alternative measures of regional variation over time were conducted. For all countries except for Colombia, Korea, and Poland, the direction of the trend shown by the range matches at least one of these other indicators. For these three countries, both the coefficient of variation and the 80/20 range indicate that the regional variation has gone in the opposite direction than indicated by the change in the range.

← 10. Estimates for involuntary part-time work are limited at the regional level due to survey sample sizes.

← 11. As this analysis considers just the number of people employed, it does not account for the quality of employment, e.g. the share of people working part-time work or on temporary contracts.

← 12. This includes differences in countries where employment in the best performing region never declined below 2008 levels.

╳

Metadata, Legal and Rights

This document, as well as any data and map included herein, are without prejudice to the status of or sovereignty over any territory, to the delimitation of international frontiers and boundaries and to the name of any territory, city or area. Extracts from publications may be subject to additional disclaimers, which are set out in the complete version of the publication, available at the link provided.

https://doi.org/10.1787/b02b2f39-en

The use of this work, whether digital or print, is governed by the Terms and Conditions to be found at http://www.oecd.org/termsandconditions.