02 Jun
With Brexit (now a noun in the Oxford Dictionary) high on everyone’s agenda and conjecture commonplace, we decided to look at the market from a purely quantitative point of view in order to understand what has happened and what is likely to happen in the near future.
Prior to the referendum, the residential property market was relatively strong—although in some areas such as central London, it was already starting to weaken due to a combination of high prices and new regulations.
The market has since transitioned to a more nuanced picture. We look at the past eight months since the referendum, mapping performance. According to the results, the market as a whole is split: central London is cooler than the rest of the country. The positive momentum prior to the vote has continued to ripple out into more affordable areas, especially eastwards from the capital.
The table below shows the performance of our repeat sales indices over the past eight months, showing the highest and lowest performing areas of the UK.
Our analysis starts with Land Registry Price Paid data. We use a factor model to characterise behaviour and discern the trends in relative performance. We have estimated postcode index returns and fed them into a three-factor risk model to identify the systematic forces at play. Such factor models form the underpinning for risk management of large institutional security portfolios, and they can similarly be used to identify the drivers at play in the residential property market. A factor model recognises that each asset is composed of a blend of underlying factors which each belong to one of three categories, and the asset’s net behaviour in a portfolio context depends on details of the mix. The three groupings are characterised by their correlation:
- Correlation = 1 Perfectly correlated assets all move in sync, so portfolio-based risk management is limited. This is the case with Factor 1, the market directional factor, or with Factors 2 and 3 when postcodes have the same sign of exposure, which often happens when they are in close proximity.
- Correlation = 0 Diversification is now possible and is most effective for portfolios composed of many assets. This is the case with risk specific to one asset or one postcode.
- Correlation = -1 Risk can be hedged through outright cancellation. This is possible when postcodes have exposures of opposite signs, which happens with Factors 2 and 3 when looking at postcodes from different regions.
The pie chart shows the breakdown of risk for a typical postcode, demonstrating that systematic risk is key and that directional risk is dominant.
NOTE:
- Factor 1 is the directional factor that moves all postcodes in proportion to their ‘beta.’
- Factors 2 and 3 drive differential performance between regions.
- Specific risk is uncorrelated between postcodes and is not systematic.
You can get the full story by looking the correlation of each postcode with all the others, but we focus here on the correlation with a single key asset: the benchmark portfolio. For many investors, especially those overseas, the benchmark has been Prime Central London (PCL) in recent years. From the risk exposures of areas to Factors 2 and 3, we created four groupings: Central, London, Commuter, Other parts. In essence, we have divided the country using a framework based on the risk factor exposures. The map shows that these groupings roughly form concentric regions around central London. The table demonstrates that the correlation drops as we move radially outward. We group these into portfolios of assets, revaluing each property to the present and weighing them accordingly in the portfolio.
As demonstrated by the table above, Central and London have been the most volatile and have had the worst returns in the eight months since the referendum.
Much of the country gets lumped into Other Parts because so many regions have positive exposure to Factor 2 and negative exposure to Factor 3. In order to be more balanced across regions, we algorithmically split postcodes into groupings with similar factor exposures, essentially ‘neighbours’ in a three-dimensional factor space. The map below shows how these empirical ‘regions’ look, and the table shows their characteristics, including correlation.
From these two different representations of the data, we can draw one conclusion: post-referendum performance has gone in different directions, with London as the underperformer. The risk pie chart emphasized the importance of the directional factor (Factor 1), but right now, the other systematic and local factors (Factors 2 and 3) are driving the patterns of post-Brexit differential return. Factor 1, meanwhile, is just coasting in neutral.
So far the emphasis has been on risk and returns over the longer term, so we now turn to two specific events with heavy impacts on the market:
- 3rd December 2014: Chancellor George Osborne changed the system of land transaction tax (stamp duty).
- 23rd June 2016: The European political landscape suffered the unexpected shock of Brexit. On the night of the referendum polling, the odds were 0.92 for Remain. By 7:00 a.m. the next morning, the odds were at 0.
The charts below show the impact of these two events on our suite of property portfolios. The tax change put a sudden brake on a very strong market in London, but had little effect in outer areas. The impact of Brexit further exacerbated this effect, sending the London market into the negatives. Whilst Factor 1 remains in neutral, returns move in opposite directions according to the exposures to Factors 2 and 3.
Both the cluster and radial models indicate that it’s time to look outside prime central London to where affordable areas are playing catch-up.
It seems that the results of the referendum may ultimately diminish the wealth dispersion, which had, in some measure, contributed to the result. Take a look at our previous article on the Politics of Property: EU Referendum.
Only time will tell if the subdued returns in the market will persist and, if so, for how long. Our forecast model is currently bullish, but more on this at another time.
Like what you’ve read? Sign-up for our weekly newsletter by clicking here!
