Thanks to everyone who attended my LSE works public lecture last Thursday.
Podcast and slides are available here:
The starting point for the lecture was the recent performance of British cities. I think it's easy to forget that growth in our cities in the last 15-20 years represented a significant turn around from what we had seen in the period following the second world war. London, for example, lost a couple of million residents in the period after the war, only to see that loss completely reversed more recently. Positive population growth in Birmingham, Manchester and Newcastle during the 2000's represented an even more recent turn around - those cities had all seen declining populations in the 1990s. It's also important to note that this turn around wasn't uniformly experienced - Liverpool and Sunderland, to take just two examples continued to see population static or declining. It's also not clear the extent to which this (partial) resurgence will be hit by the continued economic downturn.
How do we understand this recent economic performance. It's essentially driven by two inter-connected phenomena. First, the structure of the economy has shifted towards activities that tend to benefit more from urban locations. This is true both of the general shift from manufacturing to services, but also of the shift within those broader areas. Second, at the same time as cities became more important as places of production, they have also become more important as places of consumption. The fact that the structural shift has also tended to benefit more skilled workers means that these two changes have tended to reinforce one another. Successful cities have attracted higher skilled workers, who in turn have made those places more successful. In fact, I argued in the lecture that these two factors explain resurgence, the resilience of places to the downturn and the likely economic future of British cities.
Of course, such stories are easy to tell, but a lot of work we do at SERC is about trying to gather much more systematic evidence on what drives the underlying performance of cities. It turns out that this more rigorous evidence is pretty consistent with the story I have just told. For example, research that I did with Steve Gibbons and Panu Pelkonen followed hundreds of thousands of workers as they moved across areas in Britain. This allowed us to decompose area differences in wages in to an 'area effect' driven by where people live and a 'people effect' driven by the kind of people that live in an area. We found that people effects, i.e. the composition of the area, were the most important factor underpinning area differences in wages. Area effects do, however, play a role and SERC research tries to understand what causes those area effects. Why is it that workers are more productive when they locate in some places rather than others?
The evidence (not just from SERC) suggests that pure size of the place plays the most important role in understanding area effects. Bigger places makes it easier to interact with other firms and people (be that buying and selling from them, or alternatively exchanging ideas with them). Such agglomeration economies - benefits from size - tend to be self reinforcing. As a place becomes more productive it attracts more people and firms which, in turn, makes it more productive. If a place itself isn't big, the next best thing is often to be well connected to places that are big and successful (think commuter towns around London). It also helps to have lots of skilled workers because composition appears to have an additional affect on productivity over and above the fact that higher skilled people earn more. Finally, it helps if an area happens to have one of those industries that do well in cities, or is doing well nationally. These issues are discussed more extensively in our work for the Manchester Independent Economic Review.
If these benefits tend to reinforce one another, how come we don't just have one or two very large cities. The answer, of course is that the costs of living and working in a city also rise as cities grow. These costs play an important part in the second half of the lecture.
With a better understanding of what drives the economic performance of cities, we are better placed to start thinking about the impact of different urban policies. In the lecture, I drew on some of our work in progress to make the case that it's very difficult for policy to directly target and improve area productivity. Often, area based initiatives like the Single Regeneration Budget and Local Enterprise Growth Initiative appear to simply shift employment around relatively small areas (most economists don't find this terribly surprising). Giving money directly to firms (e.g. through Regional Selective Assistance) does seem to be able to create some additional employment at larger spatial scales but doesn't seem to have any measurable effect on productivity. That's troubling from a productivity point of view, because increasing employment in smaller low productivity firms will tend to drive down area level productivity. Improving road links has similar impact - it clearly effects the location of employment, but in our research to date, measurable effects on productivity have proved allusive.
If the evidence that we have suggests it's very hard to affect area level productivity, are there other things that urban policy can do to help improve the economic performance of British cities and respond to the challenges they face? In the second half of the lecture I argued that this might be achieved by paying far greater attention to the costs of living and doing business in our more successful cities. I'll outline my arguments on those issues in more detail in tomorrow's post. If you can't wait till then, you'll have to listen to the podcast.