Excellent post. Thank you for writing it. I've had a draft of a vaguely similar one for a while now, but I think yours is better.
Re immigration, I think a lot of people just aren't aware of the data on how much it's increased, so they think anyone who is worried about it or wants to reduce it must just be xenophobic.
(Also, 0.434 houses per capita, not 434. That... would definitely solve the housing crisis, but may create other problems.)
I think there is a similar phenomenon on the environment (thinking here of planning, not global warming). People are so used to the trope of 'evil company bulldozers nature reserve' they assume environmental groups are always right, without realising we've now got so many safeguards that we are building £100m bat tunnels and making it incredibly hard to build railways/reservoirs/etc.
One point - if you follow the links down, the fourth chart is of new construction, which is what most public discussion focuses on. But in some ways the relevant figure is for net additional dwellings. It is surprisingly hard to track down historical data for this - from the ONS, you can get figures back to the early 2000s easily, but not earlier. My memory, however, is that this makes the difference between the 1960s and the 1970s and 80s look less drastic - the exceptionally high figures achieved to 1968 are possible because they are knocking down a huge number of houses to build new ones. It is also worth keeping in mind that the population projections of the 70s and 80s were for flat or declining numbers, which must have played a role in thinking about housing supply.
Thank you! And yes, the difficulty in getting additional dwelling data (which I used in the final analysis) was why I only went back to 2000, as I couldn't find any data on it before that.
At the least, it allows you to see that while there is a downward slope from the 1960s to the 1990s, it isn't quite so sharp - you are looking (for England) at roughly 15.5% more dwellings from 1961-70, 11% for 1971-80, and then 9.6% for 1981-1990. From the 1990s onwards it is somewhere above 7% per decade.
What you really want, of course, is some measure that includes at least a proxy for housing quality - is the net additional dwelling a comfortable 1990s semi on an estate somewhere? Or a small 1960s flat in a tower block?
Not for the first time, I am fascinated by how relatively poor the easily available data for something so important are.
"This year we had an absolutely fantastic 381 entries"
Congrats!
(And that's tanked my score already - I was pessimistic on that question.)
"There is always a small number of people who write things such as ‘yes’ or ‘probably’ as answers, rather than giving percentages."
Would it be possible to make the entry boxes only accept numerals (or, alternatively, make them dropdown boxes offering the numbers 0 to 100)? Both these things are possible in HTML, but I don't know whether Google Forms exposes that functionality. ... ah, looks like it does: https://support.google.com/docs/answer/3378864?hl=en
I was wondering what it would make of the final two questions, given that it presumably doesn't have the context to dereference them. It seems to have decided the final question refers to the ACX prediction contest!
Thank you for doing this! And fascinating reading the reasoning. I don't agree in every case but it definitely comes across as a smart, thoughtful, person.
I'll link to this when I do the 'what did people think' post and add it to the spreadsheet for scoring. Which reminds me: I forgot to score the AI forecast that Rachael did last year, so will do that too!
This is a good look at the available data, though I think the interesting question is what happens to our society if you assume current trends continue - i.e. relatively low levels of house building and rapidly rising population. We are not far from the point where it will be essentially impossible for all but the very highest earners/couples to buy (or to be more precise, live in) a decent sized house in a prosperous place without substantial family support, or at least not without half a decade plus of intense saving. This is completely at odds with how British society functioned - or at least how its people believed it functioned - from (at latest) the 1930s through to the early 21st century.
So if that changes, what else changes? We've already seen that people stay living with their parents for much longer. And it seems obvious that family background will become more important to life chances. Does that change who people marry and why? Do class differences intensify? There are no doubt interesting geographical consequences too - perhaps we see less well off young people moving away from the South East rather than to it? Far too glib to say we're going back to the Victorian or Edwardian eras, but I do think we are moving away from the 20th century in more ways than one...
Very interesting and thoughtful article, I appreciated the objective approach, too often there's just an "immigrants bad" tone to these discussions about how to deal with our housing shortage. One thing that I wonder about the graph showing the post-Brexit "very high increase in net immigration" is - how did we measure net immigration at all accurately before we left the EU? As people from EU didn't need a visa to enter, and we weren't tracking them leaving? Anecdotally, I have had many EU students and teachers as lodgers over the years, and around 80% didn't settle in the UK in the end.
And that's a very good point: given that it was estimated that there were 3 million EU people in the country who might claim settled status, and the actual number was 5.7 million, it seems quite likely we were undercounting net migration for a good few years before that.
"At just 434 houses per capita..."
I think you mean 434 houses per 1000 people. 434 houses per capita would be nice!
I do! Will correct.
Excellent post. Thank you for writing it. I've had a draft of a vaguely similar one for a while now, but I think yours is better.
Re immigration, I think a lot of people just aren't aware of the data on how much it's increased, so they think anyone who is worried about it or wants to reduce it must just be xenophobic.
(Also, 0.434 houses per capita, not 434. That... would definitely solve the housing crisis, but may create other problems.)
Thank you! That's very kind. :-)
I think there is a similar phenomenon on the environment (thinking here of planning, not global warming). People are so used to the trope of 'evil company bulldozers nature reserve' they assume environmental groups are always right, without realising we've now got so many safeguards that we are building £100m bat tunnels and making it incredibly hard to build railways/reservoirs/etc.
This is very interesting and clearly put.
One point - if you follow the links down, the fourth chart is of new construction, which is what most public discussion focuses on. But in some ways the relevant figure is for net additional dwellings. It is surprisingly hard to track down historical data for this - from the ONS, you can get figures back to the early 2000s easily, but not earlier. My memory, however, is that this makes the difference between the 1960s and the 1970s and 80s look less drastic - the exceptionally high figures achieved to 1968 are possible because they are knocking down a huge number of houses to build new ones. It is also worth keeping in mind that the population projections of the 70s and 80s were for flat or declining numbers, which must have played a role in thinking about housing supply.
Thank you! And yes, the difficulty in getting additional dwelling data (which I used in the final analysis) was why I only went back to 2000, as I couldn't find any data on it before that.
So, this motivated me to go off and do a bit more hunting - this isn't the data set I remember, but the various discontinued series by tenure I think allow you to do the same work: https://www.gov.uk/government/statistical-data-sets/live-tables-on-dwelling-stock-including-vacants
At the least, it allows you to see that while there is a downward slope from the 1960s to the 1990s, it isn't quite so sharp - you are looking (for England) at roughly 15.5% more dwellings from 1961-70, 11% for 1971-80, and then 9.6% for 1981-1990. From the 1990s onwards it is somewhere above 7% per decade.
What you really want, of course, is some measure that includes at least a proxy for housing quality - is the net additional dwelling a comfortable 1990s semi on an estate somewhere? Or a small 1960s flat in a tower block?
Not for the first time, I am fascinated by how relatively poor the easily available data for something so important are.
"This year we had an absolutely fantastic 381 entries"
Congrats!
(And that's tanked my score already - I was pessimistic on that question.)
"There is always a small number of people who write things such as ‘yes’ or ‘probably’ as answers, rather than giving percentages."
Would it be possible to make the entry boxes only accept numerals (or, alternatively, make them dropdown boxes offering the numbers 0 to 100)? Both these things are possible in HTML, but I don't know whether Google Forms exposes that functionality. ... ah, looks like it does: https://support.google.com/docs/answer/3378864?hl=en
Ooh, thank you! Will use that next year.
Can I add a late forecast from Perplexity+Deepseek ? No doubt you may have done this yourself.
The reasoning thread is here
https://www.perplexity.ai/search/think-deeply-about-making-pred-L8m2DogESsqRFL2hEx3Mrg
85
80
70
85
75
75
25
90
65
55
60
70
20
65
40
70
30
45
85
35
10
40
15
5
50
60
20
35
70
20
85
60
90
80
65
65
40
80
55
10
75
60
75
60
90
Yay, I was hoping someone would do that. I did it last year but hadn't got around to it this year.
Really interesting reading the reasoning thread!
I was wondering what it would make of the final two questions, given that it presumably doesn't have the context to dereference them. It seems to have decided the final question refers to the ACX prediction contest!
Thank you for doing this! And fascinating reading the reasoning. I don't agree in every case but it definitely comes across as a smart, thoughtful, person.
I'll link to this when I do the 'what did people think' post and add it to the spreadsheet for scoring. Which reminds me: I forgot to score the AI forecast that Rachael did last year, so will do that too!
This is a good look at the available data, though I think the interesting question is what happens to our society if you assume current trends continue - i.e. relatively low levels of house building and rapidly rising population. We are not far from the point where it will be essentially impossible for all but the very highest earners/couples to buy (or to be more precise, live in) a decent sized house in a prosperous place without substantial family support, or at least not without half a decade plus of intense saving. This is completely at odds with how British society functioned - or at least how its people believed it functioned - from (at latest) the 1930s through to the early 21st century.
So if that changes, what else changes? We've already seen that people stay living with their parents for much longer. And it seems obvious that family background will become more important to life chances. Does that change who people marry and why? Do class differences intensify? There are no doubt interesting geographical consequences too - perhaps we see less well off young people moving away from the South East rather than to it? Far too glib to say we're going back to the Victorian or Edwardian eras, but I do think we are moving away from the 20th century in more ways than one...
All good, if depressing, questions.
Very interesting and thoughtful article, I appreciated the objective approach, too often there's just an "immigrants bad" tone to these discussions about how to deal with our housing shortage. One thing that I wonder about the graph showing the post-Brexit "very high increase in net immigration" is - how did we measure net immigration at all accurately before we left the EU? As people from EU didn't need a visa to enter, and we weren't tracking them leaving? Anecdotally, I have had many EU students and teachers as lodgers over the years, and around 80% didn't settle in the UK in the end.
Thank you!
And that's a very good point: given that it was estimated that there were 3 million EU people in the country who might claim settled status, and the actual number was 5.7 million, it seems quite likely we were undercounting net migration for a good few years before that.