10. The laws of attraction: Economic drivers of inter-regional mobility in selected OECD countries

Orsetta Causa
Michael Abendschein
Maria-Chiara Cavalleri
Nhung Luu

All of the authors are currently or formerly with the OECD Economics Department. This chapter is based on Causa, Abendschein and Cavelleri (2021[1]) and Cavelleri, Luu and Causa (2021[2]).

Inter-regional migration1 can spur economic growth, in particular by enhancing labour market dynamism and as such the efficient allocation of workers within a country. It can also enhance social mobility, in particular by allowing people from disadvantaged areas to move to areas, where they can find better opportunities. Inter-regional mobility is not always desirable: it can accentuate regional inequalities and depopulate areas that are left behind and that sometimes suffer from the closure of essential public services. In addition, inter-regional mobility, which is usually directed towards metropolitan areas (OECD, 2020[3]) can create congestion and hence contribute to environmental and health damages. There is no ideal level of inter-regional mobility and the extent to which policies should encourage people to move from one area to another will depend on country-specificities and social preferences. While inter-regional mobility is not an end in itself, policy settings should not create obstacle for individuals to move to places where they find better opportunities to fulfil their potential.

This chapter sheds light on inter-regional mobility and the role of policies. The underlying rationale for studying this topic is that the capacity of workers to move across regions in response to local economic shocks and conditions is one key dimension of labour market dynamism and also of resilience. The available literature in this area has so far been country-specific, with a large number of studies documenting a decline in labour and geographic mobility in the United States, and a more limited number of studies on European countries.2 The comparative perspective adopted in this chapter provides new evidence on the role of policies in shaping inter-regional migration flows. Two main stylised facts stand out:

  • OECD countries exhibit stark differences in inter-regional migration rates, with more than 4% of the population changing region each year in Hungary and Korea, around 3% changing state in the United States and less than 1% in Poland and Italy.

  • Trends in inter-regional migration also differ across countries. Since the early 2000s, inter-regional migration has declined in around half the OECD countries for which data are available, including North American and Asian countries as well as Spain, while it has increased in a number of Continental and Central European countries, including Austria, Germany and Hungary.

Inter-regional migration responds to local housing and economic conditions:

On average across OECD countries, regional GDP per capita is the strongest driver of regional migration inflows: an increase in regional income by 10% triggers a 5% increase in regional inflows. House prices at the regional level are the second most important driver. An increase in regional house price growth by 10% triggers a decline in regional inflows of around 2%. Unemployment at the regional level is also a significant economic driver of migration: an increase in regional unemployment by 10% triggers a decline in regional inflows by around 1.3%.

Policy settings are found to influence the responsiveness of inter-regional migration to local economic conditions:

Housing-related policies

  • Where housing supply is more flexible, inter-regional migration is more responsive to local economic conditions. Reducing policy-driven barriers in this area, for example by reforming the governance of land-use and planning policies, may facilitate moving towards better economic opportunities by reducing house price differences across regions.

  • Stricter rental regulations, both rent control and greater security of tenure, are associated with lower responsiveness of inter-regional migration to local labour market conditions. Striking the right balance between tenants’ and landlords’ interests, providing an adequate security of tenure and encouraging the supply of rental housing for all socio-economic groups is a difficult policy challenge.

  • Housing-related social transfers, both in-kind in the form of social housing and cash in the form of housing allowances, are associated with a lower responsiveness of inter-regional migration to regional economic conditions. This suggests that social housing may create “lock-in effects” whereby social tenants are reluctant to pursue better economic opportunities as they may lose access to social housing.

Labour market and social protection policies

  • Excessive job protection of regular contracts is associated with a lower responsiveness of inter-regional migration to regional economic conditions and may reduce regional labour mobility.

  • Higher levels of public spending on active labour market policies are associated with a lower responsiveness of inter-regional migration. Participation in active labour market programmes can create “lock-in” effects, for instance by reducing time for job search, especially outside the region of residence.

  • The effect of unemployment benefits varies across the unemployment spell: at the early stage of unemployment more generous benefits are associated with a lower responsiveness to regional GDP while at later stages with higher responsiveness.

  • Wider union coverage and centralisation of collective wage bargaining are associated with a lower responsiveness of inter-regional migration to regional economic conditions. Similar findings apply to higher minimum cost of labour and minimum wages. Such policies tend to narrow the wage dispersion, hence reducing incentives to move to another region with higher wages.

Regulatory policies

  • Policy barriers to business dynamism, such as barriers to entrepreneurship and administrative burdens, are found to reduce the pass-through from regional economic conditions to inter-regional migration.

  • Stringent regulations and occupational licensing for workers in professional and personal services are found to significantly reduce the responsiveness of inter-regional mobility to local economic conditions.

This chapter shows that structural policy settings at the country level have a significant effect on the responsiveness of inter-regional mobility to local economic conditions. Yet the extent to which policies should influence inter-regional mobility and the nature of appropriate interventions will vary depending on countries’ economic and social context. The responsiveness of internal migration to local economic factors varies across countries in terms of what factors matter the most and in terms of the magnitude of the response. This implies that policy requirements in this area will depend on country-specificities, challenges and social preferences. At the current juncture, there may be a case for helping both prospective movers and stayers: this can be achieved by supplementing structural policies with place-based policies, which seek to foster skills, economic and labour market dynamism at the local level, enhance the provision of public services where they are lacking, and that provide transport and digital infrastructure that allows connecting less developed to more developed areas.

The comparative analysis of inter-regional migration rates across OECD countries delivers the following main stylised facts (Figure ‎10.1):

  • OECD countries exhibit stark differences in country-level inter-regional migration rates, as conventionally measured by the proportion of the population within each national economy that changes the region of residence over one year. The migration rate ranges from more than 4% in Hungary and Korea to less than 1% in the Slovak Republic, Poland and Italy.

  • OECD countries also exhibit a wide dispersion of migration flows across regions, as measured by the differences between top and bottom decile regions ranked by inflow rates: for example, inter-regional migration is relatively equally distributed across regions in Korea and Hungary while less so in Mexico, Chile and Australia.

OECD countries have experienced very different developments in inter-regional migration over the last decades (Figure ‎10.2):

  • Less than half of the countries for which data are available since the mid-90s have experienced a trend decline in migration including the United States,3 Iceland, Korea and Japan. By contrast, some countries, in particular in Eastern Europe, have experienced a strong increase in inter-regional migration.

  • Since the mid-2000s, inter-regional migration has been on a downward trend in Spain and Australia, while it has been on an upward trend, rising by more than 30% between 2005 and 2017, in Lithuania, Austria and Germany.

The stark differences in levels and trends of inter-regional migration across advanced countries is likely to reflect a variety of non-economic factors that influence people’s choices and opportunities to move, embedded in history, culture and geography. This notwithstanding, economic theory and empirical evidence have modelled and identified the economic drivers of migration (Molloy, Smith and Wozniak, 2011[4]; Greenwood, 1997[5]; Treyz et al., 1993[6]): people move towards places that offer them better opportunities, in particular in terms of jobs, incomes and amenities, as well as lower living costs, in particular in terms of housing affordability. In theory, inter-regional migration should thus respond to differences in regional economic performance, and this in turn can raise welfare: at the micro-level as individuals move to better opportunities, and at the macro-level as labour market matching and labour market dynamism improve and regional imbalances decline. Moving from theory to practice, inter-regional migration does not seem to systematically respond to inter-regional differences4 in economic performance (Figure ‎10.3 and Figure ‎10.4):

  • In a number of OECD countries such as Australia, France, Korea and the United States, high-income5 regions tend to experience negative net migration, that is, less inflows than outflows from other regions. Interestingly, these countries have been identified among those experiencing a trend decline in migration (Figure ‎10.2).6 Migration appears more responsive to regional income disparities in lower income OECD countries (Figure ‎10.3).

  • There is no systematic link between the degree of unemployment dispersion between regions and that of net migration to low-unemployment regions (Figure ‎10.4): net migration to low-unemployment regions is for example negative in Austria and Türkiye, which feature relatively high levels of unemployment dispersion.

The responsiveness of inter-regional migration to local economic conditions is likely to also depend on living costs, especially housing costs. This has been put forward in a number of papers as one of the major explanations beyond the trend decline in internal migration in the United States, in particular for low-educated workers. The wage premium associated with a move has been shown to be too small to compensate for the rise in living costs due to local differences in house prices in high-productivity locations (Bayoumi and Barkema, 2019[7]; Ganong and Shoag, 2017[8]; Diamond, 2016[9]).7 Evidence linking inter-regional migration to regional house prices is much scarcer for other countries and inexistent in a cross-country perspective, given the lack of comparable data on regional house prices. This data gap has recently been addressed by the OECD as new harmonised regional house price indices have been produced and made publicly available. This allows to deliver insights on regional house price dynamics from a cross-country comparative perspective (Figure ‎10.5):

  • OECD countries have been experiencing a “great divergence” in regional house price dynamics.8 Between 2005 and 2017, median regional house prices grew by almost 60% in Norway and Sweden, while they declined by around 20% in Poland and Portugal. Growth was extremely unequal within all countries, with house price growth at the top of the distribution being around 30 percentage points higher than at the bottom.

  • There tends to be a positive cross-country correlation between median growth in regional house prices and inter-regional house price growth dispersion (Figure ‎10.5), which could suggest common underlying factors contributing to increasing house prices across many regions but also widening the gap between regions. However, there are outliers, as some countries experienced negative house price growth for the median region but marked differences between top and bottom regions: this is the case for the United Kingdom and the United States, where top regions saw house prices increase by around 20% while bottom regions saw house prices decline by around 20%.

Differences in housing affordability may act as a barrier to mobility for households seeking employment in parts of the country where labour demand is higher but they cannot afford to move due to differences in house prices. The extent to which regional house prices hinder inter-regional mobility and labour market adjustment is an empirical question that is formally addressed below.

The baseline analysis draws on the OECD Regional database, which provides a set of harmonised regional statistics and indicators for about 2 000 regions in 30 countries from 2000 to 2017. The advantage of this dataset is harmonisation, making it well suited for cross-country analyses. This is particularly true when it comes to regional classifications, because in any analytical study conducted at the sub-national level, the choice of the territorial unit is of prime importance. In this respect, the territorial grid applied by the OECD reflects the administrative organisation of countries. The regions are defined either at the territorial level 2 (TL2), which corresponds to the middle-tier of the sub-national government, for example, the Ontario Province in Canada, or at the territorial level 3 (TL3), which corresponds to local governments, with the exception of Australia, Canada and the United States. The empirical strategy, including the estimation of regression equations, is presented and explained in more detail in Causa et al. (2021[1]).

  • On average, regional GDP per capita is the strongest economic driver of regional migration inflows. For example, an increase in regional income by 10% triggers an increase in regional inflows by 5%. The strength of income as an in-migration factor likely reflects agglomeration effects that encourage individuals, and particularly, younger people, to move to cities or metropolitan areas which are characterised by income levels that are much higher than those of other areas.

  • House prices at the regional level are the second most important driver after income. The elasticity of regional migration inflows with respect to regional house prices implies that an increase in regional house price growth by 10% triggers a decline in regional inflows by around 2%. The finding that unaffordable housing has prevented people from moving has been put forward in the case of the United States (Bayoumi and Barkema, 2019[7]): the current results suggest that it applies also across a larger set of advanced economies.

  • The labour market situation at the regional level is also a highly significant economic driver of migration, in line with the literature (e.g. Bayoumi and Barkema (2019[7]), Liu (2018[10]). The estimated elasticity implies that an increase in regional unemployment by 10% triggers a decline in regional inflows by around 1.3%.

A number of other potentially influential variables (e.g., regional population, industrial structure, education, innovation, availability of public services and environmental quality) were tested in the estimation. However, only population was significant with a positive effect, reflecting well-documented agglomeration effects and the attraction of metropolitan areas.

In a second step, the empirical approach exploits cross-country time-series variation in policies and institutions to assess the role of policy settings in influencing the responsiveness of inter-regional migration to local economic conditions.9 While boosting inter-regional migration is not a policy objective in itself, making inter-regional migration responsive to economic conditions can be considered as a legitimate policy objective, because it enhances labour market dynamism, with benefits for economic and social resilience (e.g., individual and macro-level adjustment to local economic shocks), equality of opportunities (e.g., transitioning from joblessness to a job or towards higher quality jobs) and economic efficiency (e.g., matching between workers and jobs). For example, in the area of housing policy, the literature has shown that a less responsive housing supply reduces residential mobility (Causa and Pichelmann, 2020[11]; Caldera Sánchez and Andrews, 2011[12]). One policy question that is addressed in this chapter is whether a less responsive housing supply makes migration less responsive to local economic conditions.

The choice of the policies considered in the analysis draws on previous evidence on the effects of policies on internal migration, residential and labour mobility. The current analysis complements existing evidence with a novel angle on the effects of policies on the responsiveness of migration to local economic conditions. Against this background, the indicators included in the analysis cover three broad policy areas, namely housing-related policies, labour market and social protection policies, and regulatory policies:10

  • Housing-related policies include: rental market regulations covering both tenant-landlord regulation (rules regarding tenant eviction, tenure security and deposit requirements) and rent control (rules regarding the setting of rent levels and rent increases); housing supply elasticity, that is, the responsiveness of housing supply to price signals, which is partly policy-driven by e.g., land-use regulations; housing-related social transfers, both in kind (social housing) and cash allowances.

  • Labour market and social protection policies include: active labour market policies, job protection, unemployment benefits, minimum wages and collective bargaining institutions, as well as labour taxation. In addition, the analysis considers the effect of a set of labour-market related features which can be considered as partly policy-driven and may influence incentives and the possibility to move across regions: regional wage differences as well as labour force education and skills.

  • Regulatory policies include: various dimensions of product market regulation such as barriers to entrepreneurship, administrative requirements for limited liability companies and personally- owned enterprises, and regulations of professional services; and occupational entry restrictions, e.g., licensing procedures and administrative burdens applying to personal and professional services.

Housing-related policies and institutions are found to significantly influence the responsiveness of inter-regional migration to regional economic conditions, in particular with respect to labour market conditions:

  • Where housing supply is more responsive to housing demand, inter-regional migration is found to be more responsive to both regional GDP per capita and regional unemployment. This result is in line with studies finding a direct positive effect of the housing supply elasticity on residential mobility (e.g. Causa and Pichelmann (2020[11]), Andrews et al. (2011[13])) and with studies finding that low supply responsiveness implies that house prices rise more following stronger demand, which contributes to a rising regional dispersion of house prices, typically between higher-income cities and lower-income rural areas (e.g. OECD (2017[14])). Taken together, this evidence suggests that when housing supply is weakly responsive to demand, inter-regional migration is relatively less responsive to regional economic conditions because expected income gains from moving are more than offset by increases in living costs due to large differences in regional house prices.

  • Stricter rental market regulations, both rent controls and landlord-tenant regulations, are associated with a lower pass-through from regional labour market conditions to inter-regional migration. This result is in line with studies finding a direct negative effect of rental market regulations on residential mobility (e.g. Causa and Pichelmann (2020[11]), World Bank (2018[15]), Caldera Sánchez and Andrews (2011[12])). Tenants in rent-controlled dwellings may be reluctant to move and give up their below-market rents. This result could also reflect an indirect channel going from rental market regulations to housing supply and the dispersion of regional house prices: strong de-linking of rents from housing market conditions have been found to curtail the size of rental markets by reducing supply (Cavalleri, Cournède and Özsöğüt, 2019[16]) with negative repercussions for affordability. Too strict rent control could then make tenants in rent-controlled dwellings less responsive to move towards places with better labour market opportunities and, also, make it unattractive to do so because of unaffordable housing in such places.

  • High housing transaction costs in terms of notary and legal fees associated with buying or selling a property are found to reduce the pass-through elasticity from regional labour market conditions to inter-regional migration. This result is consistent with previous studies finding a negative effect of housing transaction costs on residential mobility, especially among young households (e.g. Causa and Pichelmann (2020[11]), World Bank (2018[15]), Hilber and Lyytikäinen (2017[17]), Caldera Sánchez and Andrews (2011[12])). High levels of notary fees associated with housing transactions may increase relocation costs and thereby reduce incentives to migrate for labour-related reasons among prospective buyers, most often relatively young mobile households.

  • Housing-related social transfers, both in-kind in the form of social housing and in cash in the form of housing allowances are associated with a lower responsiveness of inter-regional migration to regional income and, for social housing, to labour market conditions. This suggests that social housing may, in the presence of constraints to the portability of benefits, create “lock-in effects” whereby social tenants have lower incentives to move for better economic opportunities as they may lose access to social housing. This result is in line with various micro-studies of the decision to move finding that social tenants and tenants in subsidised housing are less likely to move than private tenants and owners (e.g. Causa and Pichelmann (2020[11]), World Bank (2018[15]), Caldera Sánchez and Andrews (2011[12])). While housing allowances are in principle more mobility-friendly, the current results tend to suggest that they may also create disincentives to move for economic reasons, which may reflect weak portability. At the same time, not all evidence goes in the same direction as Causa and Pichelmann (2020[11]), who found that social spending on housing (the same variables), which includes both cash and in-kind transfers, is associated with higher residential mobility.

Labour market and social protection policies shape the pass-through of regional economic conditions, in particular regional GDP per capita, to inter-regional migration:

  • Strong job protection of regular contracts is found to significantly reduce the pass-through elasticity from both regional income and regional unemployment to inter-regional migration. Workers enjoying strong protection have little incentive to move to another region, even if this would be associated with a better job match or a higher wage. This result is in line with previous empirical evidence on: i) the negative effect of job protection of regular contracts on workers’ reallocation, in particular on job-to-job transitions (Bassanini and Garnero, 2012[18]) and, ii) the negative effect of job protection of regular contracts on residential mobility, especially among youth and low-educated individuals (Causa and Pichelmann, 2020[11]). This result is also in line with findings in Adalet McGowan and Andrews (2015[19]) showing that less stringent job protection is associated with lower mismatch amongst youth, since it provides scope to improve the quality of job-worker matching, which in turn, is associated with higher residential mobility.

  • Spending on active labour market policies is associated with lower migration responsiveness with respect to both regional GDP per capita and regional unemployment, and this result applies in particular to the spending categories of sheltered and supported employment, and public employment services. This suggests that the design or delivery of active labour market policies may provide little incentive for jobseekers to look for a job in another region. One reason could be that when jobseekers are engaged in a local programme, they have little time to seek better opportunities elsewhere and incentives to engage in an intense job search decrease with the length of the programme, as found in the literature on “lock-in effects” associated with programme participation (Wunsch, 2016[20]). This interpretation is confirmed by the significant effect found for spending on sheltered and supported employment as jobseekers benefitting from such public work programmes in their region may miss better work opportunities in another region. Another explanation could be a lack of co-ordination between local agencies in different regions as counselling services are delivered at the local level (OECD, 2020[21]), as well as little incentive or possibilities for workers in such agencies to counsel the unemployed on job opportunities in other regions. These results indicate that active labour market policies, do not seem to successfully encourage labour market reallocation and labour market dynamism.

  • Unemployment benefits influence migration with respect to regional GDP per capita. The effect varies depending on the duration of unemployment: benefits tend to weaken responsiveness at the early stage of unemployment (after 6 months), especially for lone parents, while they tend to increase responsiveness at a later stage of unemployment (after 12 months). On the one hand, adequate income support during the unemployment spell is essential to help jobseekers to find a job. On the other hand, too generous income support may reduce jobseekers’ incentives to search for a job, including by moving region. Some studies have found more generous benefits to be associated with higher residential mobility (Causa and Pichelmann, 2020[11]; Caldera Sánchez and Andrews, 2011[12]), while others have found that more generous benefits reduce the probability of finding a job in another geographical area more than it reduces the probability of finding a job locally (Kristoffersen, 2016[22]; Antolin and Bover, 1997[23]). The current finding that higher replacement rates dampen migration elasticities in the short run, but increase them in the medium-run may indicate that unemployment benefit systems tend to balance the objective of protecting jobseekers from potentially disruptive short-term relocation following temporary shocks and that of helping them coping with medium-term relocation following shocks that turn out be of a more permanent nature.

  • Personal income taxes and cash transfers have a weak, significant positive effect on the pass-through of regional labour market conditions to inter-regional migration. This may indicate that income support provided by the tax and transfer system can help low-income households and workers to move towards better opportunities, and ultimately contribute to better spatial labour reallocation.

Labour market institutions affecting wage dispersion are found to affect the responsiveness of migration, in particular with respect to regional labour market conditions

  • Countries with higher collective wage-bargaining coverage tend to display less responsive migration with respect to inter-regional GDP differences. Similarly, higher labour costs at the bottom of the distribution and minimum wages are found to reduce migration responsiveness to regional unemployment. This could reflect a relatively compressed wage distribution across industries and regions, which may reduce workers’ incentives to move, as well as a downward wage rigidity that may slow down regional adjustment following local labour market shocks. This is corroborated by the finding that more centralised wage-bargaining, which is typically associated with lower levels of wage dispersion,11 also reduce the pass-through of regional GDP and unemployment to regional migration. This result is line with Poghosyan (2018[24]), who argues that in the case of Finland the wage bargaining system promotes wage compression which tends to reduce inter-regional migration. It also echoes recent findings by Boeri et al. (2019[25]) who find that the more centralised wage bargaining in Italy compared to the more decentralised in Germany tends to reduce spatial reallocation. Finally, these findings are also consistent with recent OECD work on wage premia documenting that the pass-through from firm productivity to wages, and therefore wage dispersion, is lower in countries characterised by highly centralised bargaining systems and higher minimum-to-median wages (Adrjan et al., 2021[26]).

  • Wider labour tax wedges, reflecting both employers’ and employees’ social security contributions, are associated with lower responsiveness of migration. This may arise because of the potential disincentive effects from higher taxation on labour supply, both at the extensive (moving from jobless to job) and at the intensive margin (increasing hours worked). Evidence shows that such effects are particularly strong among the low-skilled but also among people at the early stages of their career (Blundell, 2014[27]) which are likely to be the most geographically mobile to start with.

The argument that a lower wage dispersion may reduce incentives for inter-regional labour market mobility, especially among low-wage earners, is also supported by the estimated negative correlation between earnings inequalities and the responsiveness of migration to regional economic conditions. The estimates suggest that the “overall” inequality effect (D9/D1 ratio) may be driven by a “lower-tail” effect (D5/D1 ratio and incidence of low pay). Moving to the impact of skills, the results indicate that a better educated workforce is more responsive to regional economic dispersion and shocks: where the share of the working-age population with below upper-secondary education and skill shortages are higher,12 inter-regional migration is less responsive with respect to both GDP and unemployment. This result is in line with micro-based evidence finding that the probability to change residence rises with the education level (Causa and Pichelmann, 2020[11]; Caldera Sánchez and Andrews, 2011[12]).13

Policy barriers to business dynamism can affect labour market dynamism and labour mobility:

  • Policy barriers to business dynamism, such as barriers to entrepreneurship and administrative burdens, are found to reduce the pass-through from regional economic conditions on inter-regional migration. This result is in line with results in Adalet McGowan and Andrews (2015[19]) showing that less stringent product market regulations are associated with lower skill mismatch, which in turn, is associated with higher residential mobility.

  • Stringent regulations of professional services (e.g., lawyers, accountants, engineers and architects) are found to dampen the dynamism of regional migration, in particular with respect to regional GDP.

  • Stringent occupational licensing for workers in professional and personal services is found to significantly reduce the responsiveness of inter-regional migration to GDP and, to a lesser extent, labour market conditions. This result is coherent with the previous one on regulations of professional services, as such regulations dampen opportunities to move to another region to start a small business. This result is also fully consistent with the evidence for the United States, documenting a strong negative effect of stringent state-level occupational licensing on inter-state mobility, job-to-job-mobility and labour market dynamism (OECD, 2020[28]; Hermansen, 2019[29]; Johnson and Kleiner, 2017[30]).14 The current findings suggest that the dampening effect of overly stringent occupational licensing on labour market dynamism is present in a wider set of countries beyond the United States. This is also in line with recent cross-country evidence on the detrimental effects of strict country-level occupational licensing on labour reallocation-driven productivity (Bambalaite, Nicoletti and von Rueden, 2020[31]).

In order to provide an order of magnitude of the estimated policy effects, the empirical results are used to run some illustrative policy simulations. The direction of the policy change is chosen to enhance labour market dynamism by making internal migration more responsive to regional economic conditions. The simulations are reported in various figures showing how different policies influence the pass-through of GDP and unemployment to migration. Each dot is the estimated pass-through evaluated at the cross-country policy average, taking the latest available data point for the policy indicator. The distance between the cross-country minimum/maximum and the average is the change in the pass-through associated with a policy change from average to minimum/maximum. Since the simulations are based on estimated interaction effects between country-level policies and regional GDP (unemployment), they allow to identify cases where, below or above a certain policy threshold, the pass-through is no longer statistically significant.

This illustrative quantification exercise delivers the following results:

  • When housing supply is weakly responsive to housing demand, internal migration is unresponsive to regional economic shocks (Figure ‎10.6, Panels A and B). Moving from the average housing supply responsiveness to the maximum would be associated with an increase in the pass-through from unemployment to internal migration from less than 1% to around 2% (Figure ‎10.6, Panel B).

  • Relaxing rental market regulations, especially rent control, would contribute to make internal migration more responsive to regional unemployment shocks (Figure ‎10.6, Panel B). According to the estimates, moving from the average rent control to the minimum would be associated with an increase in the pass-through from unemployment to internal migration from around 1% to around 3%. Moving from the maximum to average tenant protection would be associated with an increase in the pass-through from unemployment to internal migration from around 1.7% to close to 2%, a statistically significant but economically negligible impact.

  • When job protection is very restrictive for workers on regular contracts, internal migration is found to be unresponsive to both regional unemployment and GDP (Figure ‎10.7, Panels A and B). Reducing job protection of regular contracts from the average to the minimum level would be associated with an increase in the pass-through from unemployment to internal migration from around 0.5% to close to 2%. Reforms to ease occupational licensing restrictions in service sectors would also contribute to make internal migration more responsive to regional GDP: moving from the lowest level of restriction to the average is estimated to increase the income-migration pass-through from about 2% to around 5%.

  • Product market reforms aimed at easing overly restrictive administrative burdens for business and barriers to entrepreneurship would facilitate internal migration in response to regional unemployment and GDP (Figure ‎10.8, Panels A and B). According to the estimates, reducing barriers to entrepreneurship from the maximum to the average level would move the pass-through from unemployment to migration from statistically insignificant to close to 1.7%. The same result applies to regulations of professional services. Reforms that relax barriers to entry in accounting services are reported as an example, with an estimated order of magnitude similar to the one reported for overall barriers to entrepreneurship (Figure ‎10.8).

The cross-country analysis presented above is complemented by a country-by-country study based on Cavalleri et al. (2021[2]). A gravity model is used to detect and exploit country-specific differences and characteristics in terms of drivers of inter-regional migration. Often, the results from the cross-country exercise are confirmed, yet the relevance of these drivers differs across countries. Housing affordability is an important barrier to migration in countries having experienced strong increases in the level and cross-regional dispersion of house prices, while less so in countries where commuting is an alternative to migration. Table ‎10.1 summarises the country-specific results:

The results reveal that the perspective of enjoying higher living standards in the destination region tends to be the strongest driver of mobility in a majority of countries. On average across the countries covered, a 10% rise in GDP per capita in the destination region increases migration by about 5%. This effect is more pronounced in countries characterised by relatively large regional income disparities, such as Italy and Canada.

Wage bargaining systems can influence economic incentives to move as well. Centralised (decentralised) settings reduce (increase) the dispersion of wages across firms and locations, but also the link between local productivity and local wages. When the regional dispersion in wages is low due to centralised wage bargaining, workers may opt to remain in a low-productivity region where housing costs are lower because the salary gains from moving into a high-productivity region are smaller in real terms and may not compensate for the higher housing costs. Empirical evidence based on the comparison between Italy and Germany suggests that centralised bargaining may create barriers to mobility, resulting in an inefficient spatial allocation of labour and persistent regional inequalities (Boeri et al., 2019[25]). Consistent with this, in this study, internal migration appears to be more responsive to regional GDP per capita in countries where wage bargaining systems are relatively decentralised (e.g., Canada, Poland, Switzerland, the Netherlands and the United Kingdom) and less so in countries where wage bargaining systems are relatively centralised (e.g. Finland and Spain).15

Local labour market conditions are significant drivers of inter-regional migration in several countries. Yet, the magnitude of this effect is smaller than that of income and house prices: on average across the countries covered, a 10% rise in unemployment in the destination region is found to reduce migration by about 1.6%. This is consistent with evidence that employment opportunities are not the main reason for changing residence (Causa and Pichelmann, 2020[11]).

Still, labour market conditions matter more in countries suffering from relatively high unemployment (e.g., Italy and Spain), or cross-regional dispersion (e.g., Canada, Switzerland and Finland) (Figure ‎10.5). This may indicate that inter-regional migration can act as a labour market adjustment mechanism.16 Labour market conditions have a consistently smaller effect on migration in countries where labour market adjustments are less needed, either because of low average unemployment (Korea, Japan) or because of a low cross-regional dispersion in unemployment (Denmark).

Cross-country differences in labour market policies may explain cross-country differences in the responsiveness of migration to local labour market shocks. For example, strong job protection of regular contracts, high barriers to entrepreneurship and strict occupational licensing tend to reduce the responsiveness of internal migration to local labour market shocks. By contrast, policies that favour the portability of social protection and risk-taking at the individual level can encourage geographic and labour mobility. Such is the case of well-designed portable housing allowances as well as active and passive income support for the unemployed that do not discourage reallocation. In addition, the literature has pointed to the risk that social housing may unintentionally create lock-in effects (Salvi del Pero et al., 2016[32]) This can arise when households are not willing to move to areas offering better labour market opportunities because of fear of losing the entitlement to social housing. This could explain, for example, the low responsiveness of migration to labour market incentives observed in countries with large social housing sectors, such as Denmark, the Netherlands and the United Kingdom. However, while housing allowances are in principle more mobility-friendly than direct housing support, they may also create disincentives to cross-regional mobility, especially when their portability is limited due to the regional provision of such benefits. Overall, the extent of barriers to geographic mobility that housing-related social benefits can create depend crucially on the design of such schemes.

Housing costs matter for mobility decisions in almost all countries covered by this study. On average, a 10% increase in house prices in the destination region reduces inward migration by more than 3%. Yet, this average estimate masks substantial heterogeneity across countries. House prices have a strong impact on internal mobility in countries such as (Sweden, Australia and Canada, where house prices have strongly increased over the last decade. In other countries where the impact of house prices is estimated to be more muted, a modest decline in house prices for the median region hides widening regional house price dynamics (e.g., the United States and the United Kingdom), signalling that some areas are growing increasingly unaffordable relative to others. Rising cross-regional differences in house prices may have important consequences for the level and composition of inter-regional migration flows, for instance by creating barriers to the mobility of low-skilled workers from lagging regions to metropolitan areas, as shown for the United States (Autor, 2019[33]).

Rising house prices or rents may not be particularly concerning if income rises at the same pace and housing affordability across all socioeconomic groups is maintained. Yet, empirical evidence suggests that this is not the case. Estimates by Bricongne et al. (2019[34]) show that over the course of a generation, a number of countries covered in this study, especially Sweden, the United Kingdom and Australia, have experienced a sharp increase in the years of average household disposable income required to buy a home. This is in line with the finding that across many OECD countries, housing costs have risen faster than median income and overall inflation, contributing to eroding the purchasing power of the middle class (OECD, 2019[35]).

Housing costs can also alter the nature of labour mobility to the extent that people choose to reside in a region and commute to work in another. When house prices tend to increase in urban areas and the transportation system works well, households can choose to live in suburbs around major cities. In fact, the share of workers that commute for work to another region has increased in the past ten years across European countries and it is very high in some of the countries covered by this study, that is, the United Kingdom, Switzerland, Denmark and the Netherlands. The rise in teleworking since the Covid pandemic may accentuate this phenomenon, potentially increasing the distance between workplace and residence (Davis, Ghent and Gregory, 2021[36]). This would also make housing-related factors even more relevant for migration decisions.

GDP per capita and house prices in the destination region tend to be the strongest drivers of migration in most countries, while labour market conditions seem to play a secondary role. At the same time, the results also point to large cross-country differences17 in the estimated responsiveness of internal migration to local economic conditions and living costs.

Drawing on the empirical results reported in this paper, policy interventions to support inter-regional migration can be subsumed under two broad categories: addressing policy-driven lock-in effects and removing policy obstacles to mobility. Yet policy interventions may be needed not only for movers but also for stayers. There is a case for policies that create opportunities in places where those are currently lacking, which could, if successful, encourage locals to stay instead of migrating; and attract migrants coming from other regions.

Addressing policy-driven lock-in effects requires policy interventions in the area of social benefits and active labour market and training policies, including:

  • Activation and training programmes providing adequate cash benefits that help unemployed people searching for and finding quality jobs including outside their region of residence.

    To achieve this, information sharing and co-operation between local public employment services in different regions should be encouraged, so as to inform jobless people about job availability in other regions (OECD, 2020[37]).

  • Jobseekers claiming benefits are in some cases expected to commute or to move to a new location where suitable employment is available, albeit within certain limits. A small number of countries can even require moves to a different region as part of the availability requirement and suitable work criteria (Immervoll and Knotz, 2018[38]). This requirement may speed-up reallocation but risks lowering the quality of matching and jobs.

  • Social housing eligibility rules should support mobility, alongside with fully portable housing allowances. Policy design is key to help residential and labour mobility among social housing tenants and incentivise employment, so as to ensure that vulnerable households have access to affordable housing options in other and potentially distant labour markets that offer better employment opportunities (OECD, 2020[21]). This can be achieved by removing queuing or residency requirements in the case of employment take-up, such as the “Right to Move” policy implemented in English housing associations in 2015. This may also require reinforcing institutional support, as residential mobility of the most vulnerable may be dampened by informational barriers and a lack of support in housing search and application processes. One approach recently introduced in the Paris region is an online platform, echangerhabiter.fr, that collects information from 24 major social housing providers (representing around 60% of the regional social housing stock) to enable social housing tenants to exchange their dwellings. In addition, mobility barriers and lock-in effects for lower-income social housing tenants can be reduced by gradually phasing out social rent benefits at higher income levels, as with the income-dependent rent increases introduced in the Netherlands or in France. Such measures can reduce waiting lists for social housing units, which in turn would make residential moves within the social housing system easier.

Removing policy obstacles to mobility requires policy interventions in the area of housing, labour and product markets, including:

  • Removing poorly-designed land-use regulations contributing to rigid housing supply and therefore to housing unaffordability and house price divergences.

    Increasing the responsiveness of housing supply to demand would contribute to reduce living costs in attractive metropolitan areas, making it possible for prospective low-income movers to move there to enjoy better opportunities. Reforms in this area often imply revising the design of land-use governance arrangements to avoid overlap in the allocation of housing policy functions across the different levels of administration and to favour planning at the metropolitan level rather than lower levels of government. This can facilitate the matching of supply and demand within broader catchment areas and therefore increase the responsiveness of supply to evolving demand, mitigating upward pressure on prices and making housing more affordable.

  • Reviewing local rental market regulations in places where evidence suggests that they curtail the size of the rental market.

    Reforms to make rental market regulations such as rent control and tenure security more flexible have the potential to contribute to reducing obstacles to mobility as well as making housing markets more efficient and affordable in the long term. Still, they could undermine affordability for some households in the short term, especially for incumbents. There is a case for providing tenants with reasonable security over tenure and rent levels: a compromise can be a system of rent stabilisation, whereby rents can be varied for new contracts and renewals but regulated in line with market developments during the duration of the contract.

  • Reforming labour and product market regulations where such policy settings tend to favour insiders over outsiders.

    Job protection reforms can contribute to reducing barriers to job mobility and are especially relevant in countries characterised by labour market duality. In such cases, reforms are likely to reduce spatial misallocation and make labour markets more inclusive by better integrating outsiders, often the jobless, less-qualified, women and young people. Policy action to reduce labour market duality also involves aligning social contributions and working conditions between temporary and regular contracts.18

  • Reducing barriers to firm entry and entrepreneurship, including by reviewing occupational entry regulations, may reduce obstacles to job mobility along with promoting labour and business dynamism. Empirical evidence by Bambalaite et al. (2020[31]) suggest that many countries have ample scope for achieving public goals in terms of safety and consumer satisfaction with lighter occupational entry requirements. In particular, easing regulations concerning qualification requirements in personal services would eliminate mobility restrictions that create unnecessary labour market rigidities, with disproportionate benefits for low or middle-income workers such as aestheticians, hairdressers, nurses, painters, plumbers and taxi drivers. Reforms in this area would thus achieve both productivity and inclusiveness objectives.

Creating opportunities does not necessarily imply moving individuals out of less developed regions. It can be deploying quality infrastructure and amenities in such regions, for instance to allow individuals to live there and work elsewhere, especially in a context of rising digitalisation and teleworking. This is about helping stayers, which could contribute to achieve several objectives: i) reducing regional labour market imbalances and raising productivity growth; ii) reinvigorating and rejuvenating left-behind places;19 iii) reducing congestion and air pollution in metropolitan areas; and iv) making housing more affordable in cities and thus reducing regional divergences in house prices. Another argument in favour of this approach is based on evidence that falling migration rates have often been associated with limited migration from struggling to thriving places (Figure ‎10.2 and Figure ‎10.3). While for some countries this is likely to be a policy concern in itself, it may raise the returns to local interventions, making it less likely that the benefits from such interventions are captured by those who initially live outside the target location, or by landowners in the struggling region.

Place-based policies to support stayers in lagging-behind regions require investing in quality infrastructure, transport and public amenities:

  • Hospital and medical facilities

  • Quality childcare, schools, vocational training and universities20

  • Digital coverage and connectivity

  • Well-functioning public transportation infrastructure, for instance to improve access to urban areas.

Place-based policies have recently regained prominence in many countries and international organisations (OECD, 2020[37]; OECD, 2020[39]; OECD, 2019[40]; Iammarino, Rodriguez-Pose and Storper, 2018[41]; Shambaugh and Nunn, 2018[42]). However, they often continue to be associated with spatial subsidies and compensatory policies so as ex-post redistribution interventions instead of ex-ante policy interventions that would exploit the growth potential of lagging regions. Place-based policies go beyond direct support for lagging regions, to include recognition of and adaptation to specific territorial assets, investment strategies, involvement of stakeholders, the search for complementarities across different sectoral policy lines and the implementation of an effective multi-level governance system (OECD, 2019[40]).

Social policies to promote inclusiveness and crisis resilience are very likely to require articulating “spatially-blind” with “spatially-aware” measures. One relevant area of policy intervention is that of jobseekers’ support. Unemployment benefits are usually established at the national level yet lessons from crisis episodes suggest that allowing them to vary in response to local labour market shocks can promote resilience. For example, as part of the COVID-19 crisis, Canada has allowed an automatic extension of maximum duration according to the regional unemployment rate (Box 2.6 in OECD (2020[39])). Adequate, potentially state-contingent, income support needs to be complemented with locally-provided activation and training policies. These could involve local employers and take into account the local context in terms of unemployment level and persistence, socioeconomic composition and skills of the workforce, availability of jobs or increasingly demanded jobs, as well as sectoral specialisation in declining/expanding sectors.

Finally, place-based policies may imply to direct more investment funds towards disadvantaged regions, where the marginal value of public spending could be highest. In the context of the COVID-19 crisis, a number of OECD countries have been taking action to bridge the digital divide across regions (OECD, 2020[43]). For example, in Portugal, in October 2020, the European Commission approved the reallocation of EUR 1 billion from EU Cohesion policy funds to support seven Portuguese regions. Funds will also support the digitalisation of schools, SMEs and the tourism sector. In the United States, several states have adopted measures to bridge the digital divide. For example, the City of Los Angeles is partnering with the private sector to provide options for low-cost internet, access to computer and digital literacy services, as well as device and digital training resources to its residents through its ‘Get Connected’ programme.


[19] Adalet McGowan, M. and D. Andrews (2015), “Skill Mismatch and Public Policy in OECD Countries”, OECD Economics Department Working Papers, No. 1210, OECD Publishing, Paris, https://doi.org/10.1787/5js1pzw9lnwk-en.

[26] Adrjan, P. et al. (2021), “Will it stay or will it go? Analysing developments in telework during COVID-19 using online job postings data”, OECD Productivity Working Papers, No. 30, OECD Publishing, Paris, https://doi.org/10.1787/aed3816e-en.

[48] Andrews, D. and A. Caldera Sánchez (2011), “The Evolution of Homeownership Rates in Selected OECD Countries: Demographic and Public Policy Influences”, OECD Journal: Economic Studies, https://doi.org/10.1787/eco_studies-2011-5kg0vswqpmg2.

[13] Andrews, D., A. Caldera Sánchez and Å. Johansson (2011), “Housing Markets and Structural Policies in OECD Countries”, OECD Economics Department Working Papers, No. 836, OECD Publishing, Paris, https://doi.org/10.1787/5kgk8t2k9vf3-en.

[23] Antolin, P. and O. Bover (1997), “Regional Migration in Spain: The Effect of Personal Characteristics and of Unemployment, Wage and House Price Differentials Using Pooled Cross‐Sections”, Oxford Bulletin of Economics and Statistics, Vol. 59/2, pp. 215-235, https://doi.org/10.1111/1468-0084.00061.

[33] Autor, D. (2019), “Work of the Past, Work of the Future”, AEA Papers and Proceedings, Vol. 109, pp. 1-32, https://doi.org/10.1257/pandp.20191110.

[31] Bambalaite, I., G. Nicoletti and C. von Rueden (2020), “Occupational entry regulations and their effects on productivity in services: Firm-level evidence”, OECD Economics Department Working Papers, No. 1605, OECD Publishing, Paris, https://doi.org/10.1787/c8b88d8b-en.

[18] Bassanini, A. and A. Garnero (2012), Dismissal Protection and Worker Flows in OECD Countries: Evidence from Cross-Country/Cross-Industry Data, https://sites.google.com/site/bassaxsite/home/files/.

[7] Bayoumi, T. and J. Barkema (2019), Stranded! How Rising Inequality Suppressed US Migration and Hurt Those Left Behind, WP/19/122, June 2019.

[46] Ben-Shahar, D., S. Gabriel and R. Golan (2020), “Can’t get there from here: Affordability distance to a superstar city”, Regional Science and Urban Economics, Vol. 80, p. 103357, https://doi.org/10.1016/j.regsciurbeco.2018.04.006.

[27] Blundell, R. (2014), “How responsive is the labor market to tax policy?”, IZA World of Labor, https://doi.org/10.15185/izawol.2.

[25] Boeri, T. et al. (2019), “Wage Equalization and Regional Misallocation: Evidence from Italian and German Provinces”, NBER Working Paper N. 25612.

[34] Bricongne, J., A. Turrini and P. Pontuch (2019), “Assessing House Prices: Insights from “Houselev”, a Dataset of Price Level Estimates | European Commission”, European Economy Discussion Papers, Vol. 101.

[12] Caldera Sánchez, A. and D. Andrews (2011), “To Move or not to Move: What Drives Residential Mobility Rates in the OECD?”, OECD Economics Department Working Papers, No. 846, OECD Publishing, Paris, https://doi.org/10.1787/5kghtc7kzx21-en.

[1] Causa, O., M. Abendschein and M. Cavalleri (2021), “The laws of attraction: economic drivers of inter-regional migration, housing costs and the role of policies”, OECD Economics Department Working Papers No. 1679.

[11] Causa, O. and J. Pichelmann (2020), “Should I stay or should I go? Housing and residential mobility across OECD countries”, OECD Economics Department Working Papers, No. 1626, OECD Publishing, Paris, https://doi.org/10.1787/d91329c2-en.

[16] Cavalleri, M., B. Cournède and E. Özsöğüt (2019), “How responsive are housing markets in the OECD? National level estimates”, OECD Economics Department Working Papers, No. 1589, OECD Publishing, Paris, https://doi.org/10.1787/4777e29a-en.

[2] Cavalleri, M., N. Luu and O. Causa (2021), “Migration, housing and regional disparities: A gravity model of inter-regional migration with an application to selected OECD countries”, OECD Economics Department Working Papers, No. 1691, OECD Publishing, Paris, https://doi.org/10.1787/421bf4aa-en.

[45] Ciani, E., F. David and G. de Blasio (2019), “Local responses to labor demand shocks: A Re-assessment of the case of Italy”, Regional Science and Urban Economics, Vol. 75, pp. 1-21, https://doi.org/10.1016/j.regsciurbeco.2018.12.003.

[36] Davis, M., A. Ghent and J. Gregory (2021), “The work-at-home technology boon and its consequences”, NBER Working Paper N. 28461, https://doi.org/10.3386/w28461.

[9] Diamond, R. (2016), The determinants and welfare implications of US Workers’ diverging location choices by skill: 1980-2000, American Economic Association, https://doi.org/10.1257/aer.20131706.

[8] Ganong, P. and D. Shoag (2017), “Why has regional income convergence in the U.S. declined?”, Journal of Urban Economics, Vol. 102, pp. 76-90, https://doi.org/10.1016/j.jue.2017.07.002.

[5] Greenwood, M. (1997), “Chapter 12 Internal migration in developed countries”, in Handbook of Population and Family Economics, Handbook of Population and Family Economics Volume 1, Elsevier, https://doi.org/10.1016/s1574-003x(97)80004-9.

[29] Hermansen, M. (2019), “Occupational licensing and job mobility in the United States”, OECD Economics Department Working Papers, No. 1585, OECD Publishing, Paris, https://doi.org/10.1787/4cc19056-en.

[17] Hilber, C. and T. Lyytikäinen (2017), “Transfer taxes and household mobility: Distortion on the housing or labor market?”, Journal of Urban Economics, Vol. 101, pp. 57-73, https://doi.org/10.1016/j.jue.2017.06.002.

[41] Iammarino, S., A. Rodriguez-Pose and M. Storper (2018), “Regional inequality in Europe: evidence, theory and policy implications”, Journal of Economic Geography, Vol. 19/2, pp. 273-298, https://doi.org/10.1093/jeg/lby021.

[38] Immervoll, H. and C. Knotz (2018), “How demanding are activation requirements for jobseekers”, OECD Social, Employment and Migration Working Papers, No. 215, OECD Publishing, Paris, https://doi.org/10.1787/2bdfecca-en.

[30] Johnson, J. and M. Kleiner (2017), Is Occupational Licensing a Barrier to Interstate Migration?, National Bureau of Economic Research, Cambridge, MA, https://doi.org/10.3386/w24107.

[49] Kaplan, G. and S. Schulhofer‐Wohl (2017), “Understanding the Long-run Decline in Interstate Migraion”, International Economic Review, Vol. 58/1, pp. 57-94, https://doi.org/10.1111/iere.12209.

[22] Kristoffersen, M. (2016), Geographical Job Mobility and Wage Flexibility, Danmarks Nationalbank.

[10] Liu, L. (2018), Regional Labor Mobility in Spain, WP/18/282, December 2018.

[44] Molloy, R. and C. Smith (2019), “U.S. Internal Migration: Recent Patterns and Outstanding Puzzles”, https://www.hamiltonproject.org/blog/americans_arent_moving_to_economic_opportunity.

[4] Molloy, R., C. Smith and A. Wozniak (2011), “Internal migration in the United States”, Journal of Economic Perspectives, Vol. 25/3, pp. 173-196, https://doi.org/10.1257/jep.25.3.173.

[39] OECD (2020), Capacity for remote working can affect lockdown costs differently across places, https://www.oecd.org/coronavirus/policy-responses/capacity-for-remote-working-can-affect-lockdown-costs-differently-across-places-0e85740e/.

[43] OECD (2020), Cities policy responses, https://www.oecd.org/coronavirus/policy-responses/cities-policy-responses-fd1053ff/#part-d1e211.

[28] OECD (2020), OECD Economic Surveys: United States 2020, OECD Publishing, Paris, https://doi.org/10.1787/12323be9-en.

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

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

[21] OECD (2020), Social housing: A key part of past and future housing policies, https://oe.cd/social-housing-2020.

[47] OECD (2020), TALIS 2018 Results (Volume II): Teachers and School Leaders as Valued Professionals, TALIS, OECD Publishing, Paris, https://doi.org/10.1787/19cf08df-en.

[40] OECD (2019), OECD Regional Outlook 2019: Leveraging Megatrends for Cities and Rural Areas, OECD Publishing, Paris, https://doi.org/10.1787/9789264312838-en.

[35] OECD (2019), Under Pressure: The Squeezed Middle Class, OECD Publishing, Paris, https://doi.org/10.1787/689afed1-en.

[14] OECD (2017), The Governance of Land Use in OECD Countries: Policy Analysis and Recommendations, OECD Regional Development Studies, OECD Publishing, Paris, https://doi.org/10.1787/9789264268609-en.

[24] Poghosyan, T. (2018), Regional Labor Mobility in Finland, WP/18/252, November 2018.

[32] Salvi del Pero, A. et al. (2016), “Policies to promote access to good-quality affordable housing in OECD countries”, OECD Social, Employment and Migration Working Papers No. 176, https://doi.org/10.1787/5jm3p5gl4djd-en.

[42] Shambaugh, J. and R. Nunn (eds.) (2018), Place-Based Policies for Shared Economic Growth - The Hamilton Project, Brookings Institution.

[6] Treyz, G. et al. (1993), “The Dynamics of U.S. Internal Migration”, The Review of Economics and Statistics, Vol. 75/2, p. 209, https://doi.org/10.2307/2109425.

[15] World Bank (2018), Living and Leaving: Housing, Mobility and Welfare in the European Union.

[20] Wunsch, C. (2016), “How to minimize lock-in effects of programs for unemployed workers”, IZA World of Labor, https://doi.org/10.15185/izawol.288.


← 1. In this chapter, inter-regional migration refers to movements of the population from one region to another within the same country. The focus is on internal as opposed to international migration. Migration flows across regions are sourced from the OECD Regional database.

← 2. On the United States, references include: Bayoumi and Barkema (2019[7]), Molloy and Smith (2019[44]), Ganong and Shoag (2017[8]), Kaplan and Schulhofer‐Wohl (2017[49]) and Molloy et al. (2011[4]). On Europe, see Ciani et al. (2019[45]) on Italy, Liu (2018[10]) on Spain, Poghosyan (2018[24]) on Finland; also Ben-Shahar et al. (2020[46]) on Israel.

← 3. Inter-state migration in the United States has been mostly measured on the basis of two sources: the Current Population Survey (CPS), which is a standard micro-based survey that allows to cover a long time period, and the Internal Revenue Statistics (IRS), which is based on tax declarations and is considered as a superior source to track migration, but is only available for the latest decade. Panel B of Figure ‎10.2 presents changes since the mid-2000s on the basis of both CPS and IRS. The CPS tends to overstate the decline in migration. See Molloy and Smith (2019[44]) for a discussion.

← 4. It is important to recognise that the level of regional disaggregation used in this chapter, which is dictated by data availability, may fail to fully capture regional dispersion. OECD (2020[3]) provides evidence that regional inequalities increase with the level of the regional disaggregation. This is driven by high levels of inequalities between cities (or metropolitan areas) and rural areas within granularly-defined regions.

← 5. Income is measured by real GDP per capita, consistent with the regression analysis. Household disposable income cannot be used in the regression analysis because of data availability issues, especially in the time series dimension. While household disposable income is a better measure of living standards relative to GDP per capita, GDP per capita is more likely to capture destination factors associated with higher wages and agglomeration effects. In addition, household disposable income includes income redistribution through country-level income taxes and cash transfers, which may influence international more than internal migration.

← 6. This is in line with more granular US evidence, e.g. Shambaugh and Nunn (2018[42]).

← 7. By contrast, Molloy and Smith (2019[44]) and Kaplan and Schulhofer‐Wohl (2017[49]) argue that regional house price divergence cannot explain the decline in migration in the United States.

← 8. These findings are in line with OECD Statistical insights available here http://www.oecd.org/sdd/prices-ppp/statistical-insights-location-location-location-house-price-developments-across-and-within-oecd-countries.htm.

← 9. Regional-level policies cannot be considered in the analysis, mainly because of data availability issues.

← 10. Some inevitable arbitrariness in the assignment of each policy under each area along with some overlap across areas for given policies needs to be acknowledged. For example, occupational licensing can be considered as both a labour market policy and a regulatory policy.

← 11. The correlation between collective bargaining coverage and wage centralization, with these indicators averaged over the period 2000-2015, is -0.8879 and significant at the 1% level.

← 12. The variable “skill needs” from the OECD’s new Skills for Jobs Indicators database measures the shortage or surplus of technical skills: positive values indicate skill shortage while negative values point to skill surplus. The larger the absolute value, the larger the imbalance.

← 13. The finding of an opposite effect between the dispersion of wages in the workforce and the level of education of the workforce may tentatively reflect the equalising effect of education.

← 14. The indicators of occupational licensing in this chapter are at the country-level as region-level indicators are not yet available on a cross-country basis.

← 15. However, Italy and Sweden are exceptions. In Sweden, the wage bargaining system is more decentralised but the gravity model does not yield significant estimates of the wage/income term. On the contrary, in Italy, the wage bargaining system is more centralised, but the income elasticity of migration is very high. Other factors than the wage setting system likely explain migration decisions in these countries, including: (i) regional differences in disposable income (relatively high in Italy and low in Sweden); and (ii) the prevalence of homeownership (relatively high in Italy and relatively low in Sweden), which is negatively associated with housing mobility (Causa and Pichelmann, 2020[11]; Andrews and Caldera Sánchez, 2011[48]).

← 16. Sweden is an exception to this rule, as the gravity regression reports a high elasticity of migration to local labour market effects despite small regional differences in unemployment rates.

← 17. Cross-country differences in estimated elasticities could partly reflect differences in the definition of regions and on their size. Most country estimates are based on the TL2 classification, with the exception of Korea, Japan, Finland, and Denmark (TL3). It is in principle possible to estimate gravity models at the TL2 level for these four countries. However, this is not done because: (i) aggregating the data from TL3 to TL2 triggers information losses and (ii) Denmark and Finland would need to be excluded from the empirical study due to an insufficient number of observations, as these countries have very few TL2 regions.

← 18. See Chapter 3 in OECD (2020[37]) for a focus on developments in job protection legislation.

← 19. See OECD (2020[43]; 2019[40]) for comprehensive data analysis and discussion of regional inequities with respect to megatrends such as ageing and automation.

← 20. Evidence suggests large cross-regional disparities in spending and quality. For example in France, there are large geographical variations in spending per student, especially in primary and secondary. In a recent survey (OECD, 2020[47]), two out of five school directors in France complained of insufficient internet access in school which hampers the schools’ capacity to provide quality education. Close to 60% also lament lack of computer hardware and software. These gaps appear mostly in disadvantaged zones, rather than in big cities.

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.

© OECD/KIPF 2023

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