Danny Walker
Many people expect the rise in interest rates over the past 18 months to lead house prices to fall. Average prices have already fallen by 1–2% in the UK and by more in the US. In this post I show that historically there have been large differences in how an interest rate shock affects prices in different areas of the country, even though interest rates are determined nationally. House prices respond more to interest rates in areas with more restrictive housing supply, like London and the South East of England. These are also the areas where price growth has been strongest in recent decades.
Some commentators – including on this blog and elsewhere – have argued that the large increase in UK house prices over recent decades was driven by a fall in interest rates. But I showed in a previous post that this could not be the only reason: house price changes have varied substantially across the UK, even though interest rates do not. There has been a quintupling in prices since 2000 in parts of London, and less than half that in parts of Northern England and Scotland.
Now interest rates are rising. Existing estimates based on a range of housing markets globally imply that a 1 percentage point increase in interest rates could reduce average house prices by between 2% and 20% in a couple of years – a wide range. Where we are in that range depends on whether you look at narrow windows around monetary policy announcements or more comprehensive theoretical frameworks, like those used in previous posts. There is some evidence that prices in the UK are more responsive to rates than in other countries.
There are few estimates of whether the impact of interest rates on house prices is bigger in some areas than others
A key factor in determining how a demand shock – like a rise in interest rates – affects house prices is the elasticity of housing supply. More elastic supply leads to smaller price fluctuations in the face of a demand shock, and vice versa for less elastic supply.
There are widely-used estimates of housing supply elasticity for regions of the US. There is also evidence that interest rate shocks have bigger impacts on house prices in the US when supply elasticity is low. But there is little evidence on the impact of interest rates in different regions within the US.
The evidence for the UK is even less extensive. One study produced elasticity estimates as an input to analyse loan collateral. Another (Hilber and Vermeulen (2014)) documented that income shocks have historically had larger impacts on house prices in areas of England with more binding regulatory and physical constraints on housing supply. I re-use the constraints from that study in this post and apply the analysis to interest rate shocks in England (data not available for other nations of the UK).
Interest rate shocks have historically had very different impacts in different areas of England
I run a local projections regression, using data on house prices in local authorities in England – around 300 local government areas with populations of around 150,000 each – combined with estimated shocks to the 10-year UK gilt yield around monetary policy announcements since the 1990s.
The coefficient I am interested in is an interaction term. It is an estimate of the relative impact of interest rate shocks on house prices in areas with different regulatory and physical housing supply constraints.
The constraints are proxied based on the historical local government refusal rate for development projects – covering the period before the crisis from 1979 to 2008 – and the share of land developed in the area in 1990. Both of these variables are lagged by several years to help deal with endogeneity i.e. the possibility that prices determine constraints rather than vice versa.
The controls include population and mortgage market differences across regions and over time: total credit as well as the share obtained by first-time buyers and riskier borrowers. I include time fixed effects to control for macroeconomic and demographic changes, which also absorbs the average effect of interest rates on prices, given it varies over time but not across local authorities. I also include a set of local authority fixed effects to capture persistent differences in the characteristics of regions.
Areas of England with greater regulatory and physical constraints on housing supply face much bigger falls in house prices when interest rates rise.
My estimates suggest a 1 standard deviation higher supply refusal rate is associated with falls in house prices that are 12 percentage points higher when interest rates rise by 1 percentage point, a year after the rate rise. A share of developed land that is 1 standard deviation higher is associated with falls in house prices that are 9 percentage points higher.
These differences are statistically significant and very large. At face value they imply that St Albans – a local authority with a very high refusal rate of 45% – would face a fall in house prices that is around 40 percentage points greater than Middlesbrough – a low rate of 7% – when rates rise by 1 percentage point. Islington – where more than 96% of developable land was already developed – would face a fall in house prices that is 30 percentage points greater than Northumberland, where only 1.4% of land was developed.
Chart 1 shows the estimated difference in the responsiveness of house prices to a 1 percentage point interest rate shock in the average local authority versus the least responsive decile of local authorities, at different horizons following the shock. The difference suggests the average price fall is around 15 percentage points bigger than in the least responsive decile.
Chart 1: House prices are much more responsive to a 1 percentage point interest rate shock in some parts of England versus others
The areas where house prices are most responsive to interest rates tend to be in London and the South East
I use the regression estimates described above to estimate the responsiveness of house prices to interest rate shocks in different areas. Technically that means estimating the responsiveness as a linear function of the regression coefficients and the measures of regulatory and physical constraints on housing supply for each local authority. Authorities with greater regulatory and physical constraints to housing supply have higher responsiveness estimates.
Chart 2 plots percentiles of these responsiveness estimates, showing that the most responsive areas are in London and the South East, and the least responsive are in Northern England.
Chart 2: House prices appear to be most responsive to interest rates in London and the South East of England
Chart 3 shows that the responsiveness estimates are positively correlated with an index that measures barriers to housing access, where higher numbers reflect higher barriers. House prices are more responsive to interest rates in areas that are deemed to have poor access to housing, including many areas in London and the South East. Low supply elasticity could be behind both of those outcomes.
Chart 3: Barriers to accessing housing tend to be seen as higher in the parts of England with the most responsive house prices to interest rates
The areas where house prices are most responsive to interest rates tend to have had higher house price growth since the global financial crisis
Chart 4 shows a positive correlation between the responsiveness estimates and the cumulative change in house prices in a given area since the start of 2009. This may be because the large fall in interest rates during and after the financial crisis – much of which was unexpected – translated into a bigger rise in house prices in areas with low elasticity of housing supply.
Chart 4: House price growth since 2009 Q1 tends to have been higher in the parts of England with the most responsive house prices to interest rates
There is some evidence that the relative responsiveness of house prices to interest rates is higher during rate hiking cycles
We can also produce separate estimates for periods where Bank Rate was rising and when it was not. This shows that the relative responsiveness of prices is higher during hiking cycles. This could be consistent with theories that suggest housing supply is less responsive to reductions in demand than increases in demand (the supply curve is kinked), although that arguably wouldn’t explain a wider difference between regions.
These results could in theory be driven by a couple of different channels
On the one hand, one would intuitively expect regions with lower housing supply elasticity to see changes in housing demand translate more into higher prices than into higher housing supply (eg new housebuilding). That probably applies to London and the South East. On the other hand, recent evidence suggests that the risk premium might vary in different housing markets, leading interest rates to affect house prices more in areas with lower rent to price ratios, which might include major cities such as London. Both of these factors may be at play.
Summing up
This post suggests that constraints on housing supply increase the price impact of a given change in housing demand. So the recent rapid rise in interest rates could put more downwards pressure on house prices in areas of the country with more supply constraints. There is already tentative evidence that prices in London are weaker than elsewhere. For the Bank of England, this means a key part of the monetary policy transmission mechanism varies across the country and might also vary over time as supply constraints evolve. And as mortgages are the largest loan exposure of banks, the evolution of house prices matters for financial stability and prudential regulation too. For other policymakers, this evidence points to the importance of housing supply in influencing the level and volatility of house prices.
Danny Walker works in the Bank’s Deputy Governor’s office.
If you want to get in touch, please email us at bankunderground@bankofengland.co.uk or leave a comment below.
Comments will only appear once approved by a moderator, and are only published where a full name is supplied. Bank Underground is a blog for Bank of England staff to share views that challenge – or support – prevailing policy orthodoxies. The views expressed here are those of the authors, and are not necessarily those of the Bank of England, or its policy committees.
Danny, I think this an interesting research piece, but I suspect three aspects are critical:
1. Markets need sellers for prices to fall; I suspect there is a higher percentage of mortgage free homeowners in the areas you see at greatest risk, so less selling pressure.
2. Fixed rate mortgages are far more prevalent than pre-GFC, so interest rate increases will take longer to force mortgaged homeowners to sell.
3. One might need a major fall in employment in the areas you see as most at risk, yet many work in the Finance, Tech and Service sectors, who are largely dependent on international economies.
Charles