There has been much discussion in the media recently concerning the number of vacant houses in London. The Daily Mail reports that "Only two thirds of London property buyers intend to live and work in London, which has led to hundreds of unoccupied properties." The New York Times contends that an influx of foreign buyers "has made parts of London more international, more expensive and more empty." Similarly, the London Evening Standard notes that, in the London borough of Chelsea and Kensington, there has been a dramatic increase in the number of unoccupied properties since 2012. On the other hand, Allister Heath at City A.M. is skeptical, arguing that the situation has been getting better not worse. In this post, I will examine the situation myself, using data from the Department of Communities and Local Government.
The first chart plots the proportion of vacant and long-term vacant dwellings in London between 2004 and 2013. Long-term vacant dwellings are those that have been unoccupied and substantially unfurnished for more than 6 months. Consistent with Heath, the stocks of both vacant and long-term vacant dwellings have declined monotonically since 2004; the decline in the stock of vacant dwellings appears to have been particularly rapid since 2009. Do we see the same trend for Chelsea and Kensington in particular?
The second chart plots the proportion of vacant and long-term vacant dwellings in Chelsea and Kensington. Consistent with the Standard, the stock of long-term vacant dwellings has increased since 2010. However, there doesn't seem to be any evidence of runaway vacancy. The stock of vacant dwellings is in fact at its lowest point since 2004. A possible caveat is that the properties which have recently gone unoccupied for more than six months are larger than average, meaning that if vacancy were plotted as a percentage of total square-footage in the borough, the upward trend since 2010 would be steeper.
A couple of notes on methodology: the numbers on vacancy don't go back further than 2004, and the charts look more-or-less identical if absolute numbers are plotted rather than percentages.