The state of UK cities

The Centre for Cities produced a very interesting report this week, the Cities Outlook 2014, which has been all over the media. The BBC led with stories about how "the economic gap between London and the rest of the UK is widening because other cities are punching below their weight" and the Guardian described the report as "The great migration south".

The report's headlines are that 80% of new private sector jobs over 2011-12 are in London, and that the migration of the young and skilled to London can indeed be broadly characterised as "London sucks in all of the talent".

The report is consistent with the case I've made previously that the UK apparently suffers, relative to the rest of Europe, from poor regional policy. It is also consistent with other studies of UK cities, e.g. LSE's Henry Overman says: "... an empirical regularity called Zipf’s law. Applied to cities, Zipf’s law suggests that the second largest city should be half the size of the largest, the third largest city should be a third the size of the largest, and so on. To a reasonable approximation, this law holds for the relative sizes of cities in most countries. But it does not hold in Britain where our second-tier cities appear to be too small." And it is important because economic theory (e.g Krugman: The new economic geography, now middle aged) suggests that "'Cities [are] evidence that increasing returns and positive external economies actually play [an] important economic role".

If economies of scale are important, and if most of our cities are too small, then the dominance of London and its hinterland is a big deal. Scale economies are the pre-eminent example of circular, but correct, reasoning: economic activity goes to London because that's where it's most efficient, and it's most efficient because economic activity goes to London. The impact on the rest of us is that economic activity here cannot be justified because it can be better done in London, and it can be better done in London only because not enough is economic activity is being done here. George Rosie describes the loss of company headquarters from Scotland which is an aspect of this process. John Kay speculates on the effectiveness of independence as a solution for this: "How much ... is due to Scotland’s membership of the United Kingdom and how much would have happened anyway in a world in which London is a major financial and business centre is an open question. But certainly, Scotland has not been able to adopt policies of retaining corporate headquarters within its boundaries in ways that it might have been able to do as an independent country."

I've had a quick go at looking at the growth rate of private sector employment and the share of total change in private sector employment of all the cities in the report (done both including and excluding London), and trying to explain these using log population size, dummy variable for parliamentary powers (i.e. 1 for London, Edinburgh or Cardiff, 0 otherwise), travel time to London (drive time from google maps), and the share of the population employed in the public sector. The results are far from conclusive but they suggest all the things that we might expect: negative dependence on travel time to London, positive dependence on parliamentary powers, positive dependence on population, negative dependence on public sector share of employment. The most significant explanatory variable, even if we exclude London, is the dummy for parliamentary powers.

The Centre for Cities finishes with it's 'Key Messages':
   - Cities are not islands
   - Talk of constraining London is misplaced
   - The next largest cities punch well below their weight.
   - The policy privileges afforded to London should be extended to other cities
These are fine: essentially more devolution of power and local autonomy - which is consistent with the results of the above table. But their message about not constraining London is weasel-ly worded. No-one is suggesting "constraining" London - but the implicit subsidies of London should certainly be ended.

Public investment is distorted in London's favour (see e.g. Left Foot Forward's article on the New national infrastructure plan, which featured the following chart). And even though overall public sector employment as a share of population is not especially high in London, its share of central government employment is massive, and this (especially given the talent, status and influence of senior civil servants) will have associated positive spillover effects for the London economy.

The rest of the country also subsidies London wages and housing costs through the London weighting on public sector salaries and housing benefit payments that are linked to local rather than national rental rates. If it is too expensive for the government to employ people and to operate out of London, then it should move its operations, not provide wage subsidies not available in poorer regions. If housing costs are too high and so real wages too low, then people would be discouraged from migrating to London. That would be the natural operation of the market.

As policy stands currently, the rest of the UK subsidises the operation of both the public and the private sectors in London through massive public investment, the placement of a stock (which provides for a "thick labour market") of talented and influential public employees(*), and through wage and cost of living subsidies. In return the rest of the UK sees the loss of its human capital and the elimination of possible economies of scale in other locations. A proposal to spread out central government employment and to eliminate central government funding of this level of public investment or these wage and cost of living subsidies is not a right wing proposal: a local government in London could continue to invest or pay very high levels of housing benefit if it chose to raise the money locally - incomes are high enough to support this. But it must be London that bears the cost of subsidising London, not the rest of the UK. This is not a policy of constraining London, merely a cessation of the policy of favouring London.

Other links:
   - Ron Martin, Ben Gardiner, Peter Sunley and Peter Tyler at the LSE Blog
   - Larry Elliot in the Guardian
   - My article on Concentrating economic activity in London
(*) Business for Scotland has more on relative numbers of civil servants in Scotland, London & UK as whole

What a week!

With the press release of our paper on 'Constitutional change and inequality in Scotland', last week was hectic. Apart from media appearances (e.g. GMS), which are a fairly new experience, the most stressful aspect was the concern that one side or other in the referendum campaign would claim the report as vindicating their position, and that we'd be seen as biased and partial in our analysis. For an example of how political our paper turned out to be, see the last 5 minutes of FMQs last week, which end with Baillie and Salmond arguing over who had read the report most closely!

I started to get worried when I saw the Scotsman article on the paper, which inserted the word "independent" every time our press release mentioned only "Scotland" which gave the impression that independence was somehow worse than the other constitutional options at tackling inequality. I then saw conversations on twitter accusing us of producing research to order for the No campaign. I joined one of these twitter conversations with a link to the press release accompanied by the words "needs a lot of spin to be anti-Indy". When this was repeatedly retweeted I thought maybe I'd gone too far the other way...

Anyway, what did our research say? And what comfort could it give to either campaign?

Our research quantified the impact on the level of inequality in Scotland of small changes in various tax and benefit levels. Our model included behavioural responses to policy changes, including a migration response. The conclusion of the research was that these small changes would have only a minute impact upon inequality levels - to achieve Scandinavian levels of inequality using on fiscal policy would require massive changes to current fiscal policy, which would likely be accompanied by adverse behavioural responses (read: decent proportion of tax base could flee the country).

If we assume that achieving Scandinavian levels of inequality is the aim, then the message of our research is that policy other than simply fiscal policy is likely required to achieve this aim. Changes in labour market policy, and in the levels of public service provision (such as improved childcare provision), that serve to bring down the level of wage inequality before the impact of taxes and benefits must be implemented - perhaps alongside small changes in fiscal policy. Over the course of a generation perhaps, the expectations of the population and the social norms that they operate under, will evolve to permit a more equal, Scandinavian style economy. All very consistent with the policy prospectus underlying the Common Weal project.

Conversely, Taxpayer Scotland also welcomed the report, especially its inclusion of behavioural responses, and they do have a certain logic: the report shows that reducing inequality to Scandinavian levels will be very difficult using only fiscal policy. If actually achieving Scandinavian levels of inequality will be slow and will require the commitment of the population over timescales of a generation or more, then Taxpayer Scotland suggest that the aim of reducing inequality should perhaps be reconsidered.

But what can the campaigns take from this? Well, The No side can claim that as an independent country, the UK already has all the levers that an independent Scotland would gain in order to tackle inequality. Indeed, the lower behavioural responses induced by the whole UK acting in concert, compared to a Scotland acting independently (with the rest of the UK functioning as an outside option for the fleeing Scottish tax base), means that, if the UK was serious about tackling inequality, then it would be easier to tackle as a 'whole of the UK' issue. The No side can also question the commitment of the Scottish population to actually implement policy over a generation to tackle inequality (especially given polling evidence like this).

The Yes side can point to the numbers in the report that show the UK being relatively unequal in OECD terms, markedly unequal in European terms, and extremely unequal when compared with the Nordics. They can point to the rationality of this situation from the point of view of what the UK has to lose in terms of the large contribution to this inequality from the City of London: the UK is not doing much to tackle inequality, and this may be its optimal policy; the UK is unlikely to ever do much to tackle inequality because it has to look after London. They can also point to the fact that fiscal policy is not nothing: power to do something here, even if relatively ineffectual, is perhaps better than the nothing that the UK is doing (in fact the UK may even be going backwards, given future planned welfare cuts). Finally, the Yes side can point to the potential of additional policy levers that do not come under the heading of fiscal policy, and which are unlikely to be included in any enhanced devolution settlement.

As one of the authors, I don’t think the report necessarily strengthens the case for or against independence, but it does highlight that there are real difficulties in using fiscal policy to tackle inequality in either case.

New Year arguments for Yes

Whilst ranting/discussing on hogmanay about why a Yes vote isn't obviously a bad economic choice and quite possibly a good economic choice, I had the flow of my argument disrupted somewhat. It wasn't just the drink: my problem was that I was actually making two separate arguments, and the logic of one does not translate to the other.

The two arguments were:

1. Small countries do as well as large countries, if not better
2. The UK is run particularly badly as far as its peripheral regions are concerned

Obviously saying Germany's great and that the UK does badly and that this is because small countries function better is nonsensical. But I think the two points both stand - it's just that they stand independently (so to speak) of each other.

Consider those countries of Northern Europe which have never been part of the communist block, as the set of suitable comparators: Iceland, Norway, Sweden, Finland, Ireland, UK, Denmark, France, Belgium, Netherlands, Luxembourg, Germany, Switzerland and Austria. All data is 2010 GDP per capita at Purchasing Power Standard from Eurostat.

The trend (solid line) for income per capita against population size is decreasing, hence the first argument. The vertical dashed line shows Scotland's population and the horizontal dashed line shows average income per capita across these countries. The red data point is the UK - slightly below the average for this group of countries, but not by much.

This is not a very strong argument for independence: it is suggestive of small countries doing slightly better than large countries, but tells us nothing about the direction of causality. Small wealthy regions may choose independence rather than independence leading to wealth. It remains the case though that we have circumstantial evidence in favour of the first argument: the small countries of North West Europe appear to do a bit better than their larger neighbours, and this bodes well for the prospects of an independent Scotland.

The second argument is that the UK's regional policy is particularly poor and that the UK is a particularly bad place to choose to be a peripheral region. We have seen that per capita income in the UK is slightly below, but very close to, the average for this group of northern European countries - but how is this income distributed within the country? I rank the NUTS-2 Regions within this group of countries from lowest to highest. If the distribution within the UK was similar to that within the other countries of this group then we should see the regions of the UK (red columns) clustered around the middle of this ranking.

We see that rather than clustering around the middle, the red bars are disproportionately in the bottom half of the distribution, with very few in the upper half. But the UK is not a poor country: it's an average country within this group, so why are almost all its regions in the lower half of the income distribution? The answer is that it has the most skewed regional distribution and London is the richest NUTS-2 region in Europe. The following chart shows the same data as the previous but as deviation in income per capita from the NUTS-2 region with the national capital. The regions of the UK occupy ALL of the bottom spots.

So I think there is strong evidence in favour of argument 2: regional policy in the UK is terrible and if you are not the capital then independence is not a bad idea. Again this evidence is not overwhelming: it could be the case that the UK is fundamentally a very poor place compared with everywhere around it, and the wealth of London is pulling us all up. I don't believe that this is the case though. I think it is likely that policy is such that human and physical capital investment which in other countries is spread around, in the UK is instead directed towards London, and almost every single region of the UK is losing out as a result.

UPDATE: What I've done here repeats some reporting from last summer - http://www.scotsman.com/news/bill-jamieson-data-shows-why-uk-is-truly-divided-1-2944789
- http://www.neweconomics.org/blog/entry/Breaking-free-of-London-focused-growth
- http://www.heraldscotland.com/politics/referendum-news/london-calling-the-shots.21228380

End December Links

# New national infrastructure plan – old North/South divide including Garden bridges to nowhere

# In Two observations on the Autumn Statement, Simon Wren-Lewis points out that the OBR’s accompanying forecast notes that government plans imply that “government consumption of goods and services falls from 23.2 per cent of nominal GDP in 2009 to 16.1 per cent by the end of the forecast period, its lowest on record in data back to 1948.” Even if the electorate had expressed a preference for exceptionally low levels of public services (and I don't think, despite the Tory's being elected as the largest party in 2010, that this was the preference expressed), you would not choose a period in which we were flirting with depression to make this adjustment. Ridiculously pro-cyclical fiscal policy.

# Interfluidity is great. He makes another fantastic point in Standards of evidence: we don't know what the unconditional expectation is of the impact of tax rises or reductions in equality (or for that matter, of the imposition of borders); but that doesn't stop those whose main policy objective is low taxes arguing that taxes are bad for growth. Why then are those whose main objective is the reduction in inequality so coy about claiming that inequality is bad for growth? There are plenty of models in which the conditional (ie. all other things equal) impact of unequal income distribution is bad for growth...

# Big banks versus small banks: size doesn't matter: "interconnectedness, not size, [is] principle cause of systemic risk". This sounds right to me, but I think sharing a company so that you are one big bank is likely to be an 'interconnectedness enabling technology', so average size of banks and degree of overall interconnectedness in the sector is likely to be correlated.

# Interesting stats: Storify: Regional GVA stats & Mapping the recovery

# If negative interest rates are a tax then positive interest rates are a subsidy from the future

# Much is talked about Scotland taking a share of the UK debt, but much less is said about the asset side. Supposedly they are similar values: Independence will generate a £109,000,000,000 asset windfall for Scotland via Debt equals assets. I need to dig into this.

# Great column from the KrugTron, Bits and Barbarism: "Money, [Adam Smith] understood, was a way to facilitate commerce, not a source of national prosperity — and paper money, he argued, allowed commerce to proceed without tying up much of a nation’s wealth in a “dead stock” of silver and gold. So why are we tearing up the highlands of Papua New Guinea to add to our dead stock of gold and, even more bizarrely, running powerful computers 24/7 to add [bitcoins] to a dead stock of digits?" Charlie Stross on this too: Why I want Bitcoin to die in a fire