Rethinking Rent Indices
by Ken Gibb
Private renting in the UK has been the primary beneficiary of the home ownership affordability crisis that predates the financial problems that occurred after 2007. The Buy to Let model allowed small landlords to grow quickly and the BTL formula responded well in the aftermath of the housing market downturn. Private renting has grown strongly from 8.7% in 1991 to 17.8% in 2012 (GB).
However, the sector has not been short of critics. Bad practice in some quarters had led to call for stronger regulation alongside growing interest in establishing longer more secure tenancies. This is now happening in Scotland through legislation currently at Bill stage. The Bill also includes limited controls on rent increase after a tenant and landlord agree the initial rent, provided the local area is deemed to be high pressure. At the same time, and perhaps linked to the belief coming out of the Bank of England that there may be a bubble in the Buy to Let sector, the Chancellor has been explicitly moving against the Buy to Let sector. First of all, he removed higher levels of tax relief on mortgage interest payments and then, second, in the recent Spending Review, he announced adding 3% to all levels of Stamp Duty Land Tax on second homes and rental investment properties.
So, there is a lot of policy action currently concerning the rental market. In this light it is striking how impoverished our evidence base remains on the rental market. A long standing view among researchers is that the sector is highly regionally varied, it is made up of quite distinct segments and sub-markets; it has problematic aspects undoubtedly but it also has areas that appear to be working well. We know relatively little about the behavioural impulses and motivations of the large number of small scale (often recent) landlords who dominate the sector. At the same time, we assume that most tenants are, in Peter Kemp’s nice phrase, conscripts rather than volunteers i.e. it is a constrained choice because of a lack of affordable alternative tenures especially home ownership that puts people in private renting, rather than a direct preference to rent. But is that really true and is it still as true as say ten years ago?
All of this heterogeneity is not reflected in much of the policy discourse about the sector. One area of evidence that is remarkably absent is good consistent data on rents. ONS have an experimental rent index, which suggests Scottish rents have risen 2.1% in the year to June 2015 and 2.5% in the UK. There was a good index developed in Northern Ireland and there are a few other significant commercial data providers. However, compared to the high quality house price indices available there is a serious lack of consistent local market transparency, despite the small number of honourable exceptions noted.
This is all the more important because, for one thing, the local housing allowance system relies on robust measures of the distribution of private rents by size across so-called broad rental market areas (something that assumes that these are in all places well-defined unitary market areas).
Why am I writing this just now? The main reason is that, and given this context, I saw today an interesting article in the new issue of the Review of Economics and Statistics by Brent Ambrose, Edward Coulson and Jiro Yoshida (‘The Repeat Rent Index’, Volume 97 (5) pp. 939-50). In an American context they point to the lack of data on housing flows (as opposed to the stock prices transacted as assets for home ownership and investment).
Interestingly, the authors seek to construct a repeat rent index modelled on the famous repeat sales index developed by Case and Shiller and now widely adopted, particularly in America. The essence of the idea is to link sales or, in this case, new leases, when the same properties come back on the market. One can measure price/rent change on the same property directly, as, if suitably controlled, you have the same broad quality of property and hence can overcome quality variation problems that bedevil house price data. There are well known problems in the repeat sales index. Bias in the type of properties that trade more frequently e.g. starter homes and also heterogeneity in terms of the time gap between two sales points for different properties – but this is arguably different and less serious in the rental market. Based on national rent contract data held by a major data provider, the authors worked with observations from 1998-2010 involving 1.4 million leases. Analysis was then conducted on cleaned and filtered data on metropolitan housing markets and the research also included robustness checks and careful supplementary empirical analysis.
The actual results are less important in a sense than the filling of a significant evidence gap – providing robust rental data for American cities to a level that simply did not exist before. The authors did find considerable spatial variation across the cities and over time. While this is new research, and by no means the final word, it is an indication of what is possible.
While repeat sales or rent models are not without their problems, it does seem to me that with sufficient observations, some of the constant quality assumptions that are important for sales indices are arguably less problematic for rental market housing. This is in itself a reason to seriously consider such indices as a useful way of adding consistency, transparency and policy relevant knowledge about how our rental markets actually behave.