Today’s polarising politics are characterised by deep geographical ruptures, clearly visible in the Brexit referendum, as well as in national elections in the US, France, Italy, Germany and elsewhere. This has shone a spotlight on the severe and persistent regional disparities in economic opportunity between former industrial heartlands and service-based cities, which afflict many countries in the developed world.

But what explains the persistence of these disparities? In this research, we focus on the case of Great Britain. Figure 1 shows that a Travel-To-Work-Area’s (TTWA) employment-population ratio (among working-age individuals) in 1980 is a strong predictor of its ratio in 2010 (with a correlation of over 50 per cent). There is clear evidence of a North-South divide. And remarkably, these disparities cannot at all be explained by differences in local demographics – whether age, education or ethnicity. Instead, we interpret these disparities as representing genuine differences in economic opportunity. The patterns of persistence look very similar to those we observe in the US: see our earlier work.

Figure 1. Persistence of local employment-population ratios

In principle, these disparities should be eliminated through labour mobility – whether through migration or changes in long-distance commuting patterns. One might expect that commuting is particularly important for a densely populated country such as Britain. (Indeed, our focus on commuting is what distinguishes this research from our earlier work on the US.) For example, a former coal miner in South Wales might choose to move to a more prosperous part of the country, but he could also commute to Cardiff in search of employment.

Building on these ideas, a common view is that persistent disparities are a consequence of a sudden collapse of manufacturing industries in the 1970s and 1980s, coupled with an immobile labour force. This appears to have been Norman Tebbit’s diagnosis of the problem, when he famously recalled how his father “got on his bike and looked for work and … kept looking ’til he found it”.

But our analysis suggests otherwise. We identify large (though by no means instantaneous) population responses to local disparities in economic opportunity: see Figure 2. Interestingly, the population responses are just as large as those we estimate in the US – despite common concerns over British immobility. And comparing 9,975 very local geographical units (known as “wards”), we also find that local shocks elicit large inflows of commuters from elsewhere. In quantitative terms, these adjustment mechanisms are sufficient to (mostly) eliminate the impact of a local shock to labour demand within two decades. And therefore, any sluggishness in these responses cannot account for the sheer scale of the persistence apparent in Figure 1.

Figure 2. Response of local population growth

Our argument is that the effectiveness of these population and commuting responses are severely hampered by the patterns of the shocks themselves. First, those areas that suffered weak employment growth in the 1970s and 1980s continue to do so today (see Figure 3). This generates a “race” between jobs and population: though population does respond strongly, it cannot keep up with the pace of local job loss. As a result, local employment-population ratios in the affected areas appear persistently low – despite large underlying employment and population movements.

Figure 3. Persistence of local employment growth

The second problem is that local shocks are very correlated across space. Those wards which shed many manufacturing jobs are likely to neighbour other wards with similar experiences. And given that commuting is very costly (50 per cent of workers commute less than 5km, and 90 per cent less than 30km) the opportunity to commute offers little insurance against local shocks. Indeed, we find that workers will often change their residence within so-called TTWAs (in response to very local shocks), rather than simply commuting longer distances.

To summarise, we find that workers do respond to local job losses both through migration and changing commuting patterns. But the effectiveness of these responses is impeded by a tight correlation in these job losses both over time and across space. In this context, based on a simulation of our model, we find that a policy that universally reduces commuting costs by 10 per cent would contract local disparities in employment-population ratios by just 4 per cent. But a policy which expands the migratory response to local shocks by 10 per cent would have a larger effect – reducing disparities by 24 per cent.

Of course, in practice, a 10 per cent increase in residential mobility may be very difficult to achieve. Instead, the evidence here may lend support to heavier investment in transport infrastructure which facilitates longer-distance commuting (such as “HS3” connections between Northern cities). Also, given that the fundamental problem appears to stem from persistent local shocks (rather than labour mobility), one might conclude that the government should intervene more heavily on the demand-side in regional policy. But to date, much of the evidence on such interventions had been dispiriting: see, for example, the research output of the LSE’s What Works Centre.

♣♣♣

Notes:


Michael Amior is an assistant professor of economics at Hebrew University and a research associate at the LSE’s Centre for Economic Performance (CEP). His research mostly focuses on the labour market causes and consequences of migration, both within and across countries. Michael has a PhD in economics from University College London (UCL).

 

 

alan-manningAlan Manning is professor of economics in the department of economics and director of the community programme at CEP. His research generally covers labour markets, with a focus on imperfect competition (monopsony), minimum wages, job polarisation, immigration, and gender. On immigration, his interests expand beyond the economy to issues such as social housing, minority groups, and identity. Alan holds a DPhil in Economics from Oxford University. For more on his work, visit his personal website.