Geography lesson

This article was first published in Public Finance Perspectives and is reproduced here with kind permission.

The UK government has committed to an ambitious program of ‘levelling-up’ poorer and ‘left-behind’ areas of the country to reduce inequalities between places. This ambition is not a new one: efforts to reduce the UK’s spatial disparities date back at least as far as the 1934 Special Areas Act, and the Treasury was using the language of ‘levelling-up’ long before Boris Johnson entered Downing Street. Yet geographic inequalities in the UK remain substantial, and this generation of politicians and policymakers must grapple with contentious questions over where, how and what to level up – questions that, so far, remain largely unanswered. Should the aim be to narrow gaps in incomes, productivity or employment outcomes between areas, or should it be to equalise opportunities and enable those living in ‘left-behind’ areas to move to more prosperous places? Should the government assist broad regions, potentially providing funds to some already prosperous areas, or target support more narrowly at individual neighbourhoods or towns?

A number of factors make these questions particularly tricky.  The first is that, despite all the talk of a North-South divide, on many important metrics there is greater inequality within regions than between them. Consider the case of productivity. We look here at a common measure of productivity, gross value added (GVA) per head using the most recently available data (from 2015). If we compare broad regions, productivity in the most productive region, London, is 1.83 times that of the least productive region, Wales. (Note that we have excluded Westminster and the City of London from London-wide figures, because low populations and a concentration of economic activity in these places skew the figures. If those areas are left in, the ratio rises from 1.83 to 2.42). Similarly, productivity per person in the South-East of England is 1.55 times higher than in Wales. These differences are significant and important. But if we look within these broad regions, the gaps are even greater. The ratio between productivity of the most- and least productive boroughs of London is 6.65 (excluding Westminster and the City of London – the figure rises to 13.4 if we include Westminster, or more than 300 if we include the City of London). Within the South East, the ratio between most- and least-productive local authority is 3.36, and it is 2.30 and 1.83 within Wales and the North East, respectively.

There are substantial differences in productivity between the larger regions of the UK, and particularly large gaps when comparing London and its surrounding areas to the rest of the country. Yet, focusing just on these differences misses the fact that differences in productivity within regions are often larger, and potentially just as important.

While on average there is a difference between London and ‘the North’, the differences between a more prosperous city like Leeds or Manchester and isolated, so called ‘left-behind’ towns such as Burnley or Wakefield in their hinterland can be just as stark. At the same time, despite its high average levels of productivity and income, London contains many pockets of deprivation. For example, Camden is the local authority with the third highest GVA per head of the UK. Yet, over half of neighbourhoods (defined in the jargon as Lower Super Output Areas) within Camden are in the most deprived half of neighbourhoods nationally (on the government’s index of multiple deprivation, which captures economic and social factors related to poverty and a lack of opportunity). In fact, the three most deprived local authorities by this measure are all within Greater London.

This matters for the design and targeting of levelling-up policy. To consider why, we can envisage a spectrum of policies, targeted at different levels of geographies – bearing in mind that the ultimate aim of policy is to help people within places.

At one end of the spectrum, support can be directly targeted at disadvantaged individuals, wherever they are found, through policies such as training, job subsidies or direct cash transfers. Places with more disadvantaged people would indirectly receive more support, but these policies would not be ‘place-based’ in the way levelling-up clearly intends to be.

Moving along the spectrum, support could be targeted at a town or community, rather than at individuals. Some people living in these places will not be disadvantaged, even if many are. Targeting support at places rather than people means focusing support less directly at the disadvantaged. The justification for doing so comes from the existence of economic spillovers within an area. For example, creating jobs within an area might lead to further spending on local businesses. The UK’s high costs of housing impose barriers to moving – plus, many people feel a keen sense of attachment to the place in which they live, and so would have little interest in moving even if those barriers were reduced. Helping the place in which those people live, rather than making it easier for them to move to areas with more opportunities, can therefore be an effective way to make them better off. Moving further up this spectrum, some policies would seek to leverage spillovers across wider regions, for instance by linking two towns together with improved rail links. Such a policy could potentially benefit everyone living in and around those two towns, not just the disadvantaged.

The key trade-off to consider here is that between targeting and spillovers. As you offer support to a larger area, this support is targeted less directly at the most disadvantaged, but potentially takes better advantage of spillovers between places. The extent to which targeting worsens as you aim funding at larger areas depends on the extent to which inequalities within regions are larger than those between them at different levels of geographic aggregation (i.e. as you move along the spectrum).

One complicating factor is that this calculation will vary depending on exactly what is being measured. When looking at inequalities in productivity and pay, a great deal depends on whether we use measures based on where people live, or where people work – particularly when looking at London and the South East. When looking at pay, it also matters whether we look at the median or the mean: the largest differences between areas within a broad region tend to be at the top of the earnings distribution (i.e. in the number of high-earners), which affects the mean much more than the median. Within-region inequalities in mean pay, relative to between-region inequalities, are therefore greater than for median pay. Or, if instead of pay we are interested in incomes, it matters whether we look at incomes before or after housing costs. Median income before housing costs in London is 14% higher than the UK average; if we look at median income after housing costs, it is just 1% higher.

The extent of within-region inequalities can therefore matter more or less, depending on exactly who and what the government is trying to ‘level-up’: something that hasn’t yet been made entirely clear. The government certainly needs to account for these issues when designing policy. If inequalities within regions are very significant for a measure the government wishes to target, this might point towards the need for a more targeted approach.

However, the other side of the trade-off – the fact that policy may take better advantage of spillover effects when aimed at larger regions, even if it is less well targeted as a result – must also be considered. As explored earlier, some areas can appear relatively prosperous but still contain significant pockets of deprivation, with London a prime example. However, the best way to help disadvantaged people in London may not be to create jobs in their neighbourhood, but instead to equip them with the skills to work in the many good jobs London already creates, and to improve transport links so they can get to these good jobs. This means that whilst you might target skills training at smaller areas, the spillover benefits from transport links and job creation might make these policies better targeted at a wider area. In some cases – like direct public transport provision – policy may only be practicable when targeted at a larger area.

There are other policies that might also be better targeted at a broader region, rather than directly at just the most disadvantaged individuals and communities. A large part of the government’s ‘levelling-up’ strategy appears to revolve around creating new industries in green technology and research. Yet the tendency of such industries to cluster around a hub means that it is unlikely that every ‘left-behind’ town in the country can be home to these new jobs and companies. Instead, if the government is to provide subsidies or support in this space, the best strategy might be to focus job creation on a smaller number of larger hubs, whilst also providing improved education, training and transport to allow those in surrounding disadvantaged towns to benefit from the jobs in these new industries. 

An optimal levelling-up policy framework would seek to balance these competing factors: the need for greater targeting by focusing on smaller areas, and the benefits of agglomeration and spillovers that might push the dial in the opposite direction. These trade-offs themselves depend on which inequalities the government is trying to reduce, and on which policies they intend to use to achieve this aim, but they certainly exist.

This brings us to the tension at the heart of the government’s ‘levelling-up’ strategy. Much of the policy focus so far has been on centralised pots of money allocated to areas deemed in need of support. There are now a large number of such place-based spending schemes, which allocate funding to different bodies and levels of governance. The Levelling Up Fund allocates money to local authorities, for instance; Local Growth Deals allocate funding to local enterprise partnerships; and the Transforming Cities Fund provides funding to combined authorities (which are themselves groups of local authorities). To a degree, this makes sense: as we have argued here, the appropriate level of geography will likely vary from policy to policy, and from scheme to scheme. But it is not always obvious what each of these schemes is trying to achieve, and thus whether the chosen level of geography is appropriate. The complex, overlapping patchwork of funding can also cause problems for local governments seeking to make use of them. Some rationalisation of existing schemes could be beneficial. At the very least, a clear and coherent strategy for what each of these schemes is trying to achieve, and how they link together, would help encourage joined up local development plans. For levelling-up to truly succeed, this coordination, alongside careful consideration of the trade-offs between targeting and spillover benefits across places, is key.