When trying to estimate inequality in Britain, it sort of depends what you’re looking at.
While UK income inequality rose significantly in the 1980s, it has stayed pretty steady since the early 1990s. Wealth inequality is greater by an order of magnitude, and wealth in the UK means private pension wealth (42% of household wealth) and property (35% wealth), with physical and financial wealth making up the other 24%. I have blogged before about wealth taxes, which I think are a good idea in principle. They may even feature as a key issue in the US Presidential election, depending on who gets the Democratic nomination. But it seems likely that the topic of wealth taxation will be off the agenda for some time in the UK.
The ONS calculates splits UK wealth inequality (with a Gini coefficient^ of 0.63) into net financial wealth (0.91), private pension wealth (0.72), and net property wealth (0.66). I’m guessing that with each of these having a higher Gini than the Gini for total wealth, some households have more pension but less property, some less property and more financial wealth etc.
Gini coefficients for UK wealth and income
The biggest driver of private pension wealth is whether you are in a pension scheme at all. Only 53% of individuals (aged 16-State Pension age) are in an active scheme, and this is after the +10% bump up that came from auto-enrollment .
If you are in a scheme, the biggest driver of your pension wealth is whether you are in a Defined Benefit scheme or a Defined Contribution scheme. I wrote a thread on the latest data here , and leave you with this cool ONS chart from the threads that shows pension wealth by age and scheme type. This is a big deal, and I would expect it to become an increasingly big deal.
But, as Arlo Guthrie said, that’s not what I’m here to tell you about. I’m here to talk about property inequality. Or rather, I want to know what level of wealth inequality is embedded in the physical stock of UK property wealth.
The seed of this question grew in the run-up to the General Election. Basically. if you’re going to apply wealth taxes to property (as an asset that can’t flee the country), then given my assumption that wealth inequality is way higher than income inequality (see first chart above), doesn’t this mean that people with top-quartile incomes aren’t going to be able to afford to live in their homes? (Yes, I know, tiny violin time.)
In order to make some headway I asked @resi_analyst for advice, and he kindly pointed me to annual transaction data files. With around 4-5% of residential properties changing hands each year it is maybe not a stretch to infer that they are a reasonable sample of the UK residential stock. I like this a lot more than looking at the (excellent) ONS Wealth & Assets Survey data because the WAS is a survey, while transactions represent a much bigger dataset.
And this is what I found, looking at easily-available from the last six years. The chart is log-scaled and shows the proportion of properties sold by price-point. So in the first 10 months of 2019 the median residential property transaction value was £235k. A quarter of properties changed hands for £355k or more, and 1% of properties changed hands for £1.26m or more. The stuff in the very top one percent is a bit whacky, and is driven by a very small number of super-high property transactions.
UK residential transactions by percentile price point, 2013-Oct 2019
This top-tail is interesting, but very small. I could imagine some of these being split up into flats (although a fair few are Belgravia flats already), and as such the structure of inequality in the property stock could be amended pretty easily. This would all require building work to do and my question is around the physical stock as it stands today.
If we remove the top-priced property from the distribution and examine only the bottom 99% on transaction values, things look a bit different (still requiring a log-scale, mind).
I applied a bit of VBA code kindly left hereto then calculate the Gini coefficients of the transaction values.
I was quite surprised by what I found.
Firstly, in contrast to the net property wealth data from the ONS, the level of inequality embedded in the value of physical private residential stock is actually not too far away from the level of income inequality in the UK. Taking out the top and bottom percentile of transactions (of course) makes this even more the case:
Secondly, if we were to take transaction values as a good proxy for the value of the stock, we have seen an increasing convergence between income inequality (dark blue line) and the embedded inequality attached to the physical stock of UK residential property (red line). Moreover, this isn’t a function of those whacky idiosyncrasies at the very top of the distribution of property transactions. We can see this to be true by looking at the Gini coefficient of the 1-99%ile of transaction values (light blue line).
The data exists to calculate this going back to 1995, but sadly the ONS data file is too monstrous for me to deal with in simple Excel.
Why is my estimate of the stock of private property stock value inequality so different from the net property wealth measure of inequality from the ONS?
- Firstly, the whole point of my inquiry was to see how equal property wealth could be, given the current physical stock of property that is in place today. This is very different to measuring the current level of inequality that resides in property assets.
- Secondly, the ONS net property wealth is calculated by netting mortgage debts off against the value of residential property owned. So If I buy a house for £235k with a mortgage of £200k, my net property wealth is £35k. That’s what net means. I care about the £235k while the ONS cares about the £35k.
- Third, there are only 20m private sector residential dwellings (out of a total 24.2m dwellings), and I am looking at value-inequality within this dataset. The ONS reckons that there are 27.6m households. It’s not immediately clear to me how households are defined, as they appear to be defined in a way that is detached from homes. Regardless of this quirk, we know that there are a *lot* of households who don’t own.
- Lastly, there are lots of wild assumptions in my data: that transactions are a decent proxy for the evolution of stock values; that things observed in England & Wales are true for the UK as a whole; that excluding BTL and offshore purchases doesn’t screw things up. Each of these could be wrong.
Maybe I shouldn’t have been so surprised. Somewhere between ‘a lot’ and ‘almost all of’ the value of a residential property is tied up in its locational value, and this locational value is significantly a function of the proximity to cash flows that flow from employment. If high paying employers set up in a given locality it’s going to boost the locational value of the area. Simples.
^ To quote George Banham of the Resolution Foundation, the Gini coefficient measures inequality on a scale from 0 to 1, where 0 is perfect equality and 1 is perfect inequality (a situation where one person has all the wealth).