How reliable is personal experience? This a question raised by Janan Ganesh in the FT. He writes:

Modern London is liberalism...in excelsis. It has taken the free movement of people, goods, services, capital and ideas to anarchic extremes that might have no precedent.
Anyone can talk a good game about freedom and diversity; living here is a practical test of one's commitment to these things

I lived in London for over 20 years, and this struck me as nonsense. For one thing, I saw little diversity. I lived in Belsize Park and worked in the FT building, both of which seemed no more ethnically diverse than a Ukip party conference*; some of London's supporters seem to forget that ethnic diversity exists in other cities.

Nor did I experience any "free movement". Perhaps my biggest reason for leaving London was that I wasted hours travelling; one of the biggest adjustments I had to make when I moved to Rutland was realizing that it is quicker and more pleasant to travel the 25 miles into Leicester for an evening than it was to get from the office to the West End.

Like millions of others, I was just a mindless drone moving from silo to silo. The notion of the city as a serendipity engine in which innovation and creativity is sparked by chance meetings always struck me as romantic drivel.

My experience, however, runs into a problem. The fact is that London is far more productive than the rest of the country - 29.2 per cent more so, according to ONS data**. How can this be?

The obvious possibility is that my experience is wrong and cities do indeed benefit (pdf) from agglomeration effects: people learn from living and working alongside each other. As Ed Glaeser has said:

We are a social species that gets smarter by being around other smart people, and that's why cities thrive. That's why Wall Street traders still have trading floors where incredibly wealthy people are all willing to sit right on top of each other.

But there's a problem here. If this were the case, we'd expect that cities generally would have higher productivity than less densely populated areas. But in the UK, this is not so. Researchers at the University of Cambridge point out (pdf) that our other major cities such as Birmingham and Manchester have below-average productivity.

This is consistent with another theory - that London's wealth is due not to benevolent agglomeration effects but to parasitism. It might be that big legal (pdf) and financial industries actually depress economic activity elsewhere, in part by sucking talent away from other uses. Professor Glaeser might be right to say that wealthy traders want to sit on top of each other - but this could be because doing allows them to benefit from insider dealing, front-running and rigging markets rather than because of genuine productivity improvements.

I offer this as just a hypothesis. To reject it one must answer the question: why is it that London is one of our few cities to benefit from agglomeration effects?

* You might reply that this is because I lived and worked in affluent areas - to which I reply that the richer parts of Leicester, for example, have a large Asian population.

** This raises another paradox. Higher productivity doesn't translate into higher subjective well-being: Londoners are actually slightly less happy than the national average.