How to (possibly) stop worrying about London rental prices

Stay informed with free updates

Simply sign up to the UK property myFT Digest — delivered directly to your inbox.

MainFT yesterday:

UK rent growth accelerated in October with London registering the largest increase, according to official data that suggests more pain for tenants after two years of rising costs.

Here, via the Office for National Statistics, is the relevant chart:

Rental inflation is especially important right now, not least because another ONS release on Wednesday — the UKâ€s headline consumer price inflation figure — found rising rents to represent almost exactly a quarter of the overall annual increase in price levels:

For obvious reasons, the headline CPI figure matters greatly to the UKâ€s overall macroeconomic trajectory.

But for reasons that weâ€ll get into shortly, the ONS rent figures (supplied by the Valuation Office Agency) rarely lined up neatly with those produced by other indices, such as HomeLet, Zoopla (PDF link), Rightmove, Spareroom and Goodlord.

At the moment, all of these agencies are producing estimates of London rent increases that are radically different from the ONSâ€s:

The basic dynamic occurring here is explained through something called the bathtub analogy. Hereâ€s the ONSâ€s explanation of this from a blog a couple of years ago:

A useful analogy, as provided by a colleague from the lettingâ€s [sic] specialist Countrywide several years back, is to think of the PRS as a bathtub of water. The temperature of the water in the tub represents the stock of rents (all rents), while the water entering the tub from the taps represents the flow of rents (advertised rents). The flow of water entering the tub is hot when rents are increasing (or cold when falling), as this is reflecting the newly advertised, market driven rental prices. However, it will take time for this flow of water to change the overall temperature of the water in the tub.  

As a result of the different methods used, it is correct that the IPHRP [Index of Private Housing Rental Prices] usually shows lower growth in rental prices than the private sector measures. This is not an attempt to ‘smooth†prices , but to reflect the fact that IPHRP is intended to cover all rents – the majority of which will not see price change month-on-month.  

So: ONS figures = water in the bath (existing rents), agency figures = water coming from the tap (advertised rents). The area directly underneath the tap is basically renters who have started paying a higher advertised rate, while distance from the tab roughly equates to tenancy length. It takes time for the tap to feed though to the whole bath.

Itâ€s an OK analogy. The liquid and thermal dynamics involved can get a little confusing if you think about them too much…

— a bath with the tap running fills up
— a warm bath loses heat
— a bathtub typically doesnâ€t change size in response to supply / demand
— tap water has no capacity to negotiate the temperature it will be when it enters the tub
— evaporation?????
— this analogy has to include parts of the bath that for some reason will never change temperature
— head hurts

…so letâ€s not dwell.

One effect that ought to be present in both in the analogy and the statistical reality is that (assuming the data is perfect, which it never is) during periods of rental inflation, the temperature of the bathtub should never be higher than that of the water coming from the tap. As an extension of that, as mentioned above, the bathtub analogy usually creates a bias towards higher inflation in advertised measures, which is categorically not the case at the moment.

So the key issue is one of delay. As the ONS wrote in an earlier blog, in 2020:

Due to the difference in how the various rental indices are measured, the IPHRP tends to lag the private measures by around 6 months – so using our analogy, it takes approximately 6 months for the overall temperature of the tub to reflect the heat of the hot water flow.

And, right now, a six-month lag looks somewhat encouraging (from the Bank of Englandâ€s perspective). Hereâ€s the ONS measure versus six-month delayed figures from HomeLet:

Less optimistically, one might note that the gap between the peaks appears much longer — well over a year.

The ONS seems upbeat. A spokesperson told us:

As usual with London, things are always a bit more extreme given the nature of the market, however, Homelet were showing London at around 9% in April. So we would expect the slowing seen in private measures to start feeding into PIPR soon.

When we asked whether they were concerned the lags have increased, they added:

[We] do not have concerns on the lag. There is not direct pass through from the advertised rents (which are just asking prices reflecting what landlords/agents hope they can get at a specific point), and the stock, as the stock is dependent on the length of tenancy which can differ significantly.

Average tenancy length tends to be around 9 – 11 months due to a mix of 6m and 12m fixed period tenancies during which time rents donâ€t change. So we would expect it to take around 12 months before our stock measure would be expected to capture the effect that new lets price changes from 6-12 months ago have on the average of the rental stock.

Even when a fixed period tenancy ends, there may be a renewal, and those tend to see a lower price increase than a new let price increase, so that would reduce the impact of new let price changes on the stock (PIPR) measure and hence would dampen and extend the time taken for PIPR to respond to changes in new lets.

Rents have seen record-high growth in recent years which may have led to a change in tenant behaviour. For example, a tenant faces contract expiry but new lets are much higher than the existing let price, a tenant may be more likely to renew their contract than they may have been a couple of years ago, to reduce the price increase they face. Changing tenant behaviours may lead to a change in the relationship between the flow and the stock, and hence increase the time taken and/or dampen the influence of changes in new lets to impact PIPR. It is difficult to directly measure this.

Of course, alternative lags are available. Zoopla, another rental index provider, prefers a (huge) rolling 33-month rolling average, saying:

The average length of a tenancy varies depending on different data sources and ranges between 30 and 40+ months. 

Taking the average monthly growth in our new lets index over 33 months – and then expressing this on an annualised basis – delivers an annual growth rate that is very similar to the ONS index. 

Hereâ€s how that looks:

This may feel a bit like an opportunity to choose to your own adventure in data and, well, yeah. It sure looks like a slowdown is coming, but there could still be a bumpy ride first.

Further reading:
— 904 (xkcd)