Some of the comment in the blog seems to assume that a good proportion of Airbnb lets are driven by economic necessity - which I take to be people letting a spare room or two in the house where they are living. I'd assumed almost all lets in Central London are not like this - they are buy to let properties. Not sure who's right. But the answer would give an insight into the effectiveness of regulation, in that it's hard to believe someone letting out a buy to let property on Airbnb would survive letting it for only 90 days a year, so are quite likely to be letting via other sites too.
1. Until the overnight visitor levy comes in (if mayors choose to give councils a share), councils themselves don't really benefit from tourism spend financially. No council tax, very minor use of services paid with fees, I suppose indirectly through business rates growth but very minor and indirect given it's based on premises size. It's purely a cost to them.
2. Sure you thought about this but at first glance, aren't the hot spots just tourism demand driven by good connectivity/centrality/fashionableness incl on TikTok, rather than rent-income? Know slightly overdetermined but feels to me like demand is driving spatial distribution. Interesting caveat that high density requires either low social housing proportion (legally, unless now private through RtB) and will be boosted by absentee landlordism.
Really important point and one I didn't address - the fiscal asymmetry is stark. Councils bear the externalities (enforcement, housing pressure, infrastructure wear) but capture almost none of the upside. The visitor levy might help but as you say, it's optional and the revenue share to boroughs is uncertain. There's something perverse about a system where local authorities are essentially subsidising a platform's business model. Do you know if there's any breakdown yet of how the levy revenue would be distributed - what share goes to the GLA vs boroughs, and whether it's earmarked for anything specific?
You're right that demand is doing a lot of the work spatially - connectivity, centrality, TikTok-ability. I probably undersold that. My hunch is it's mutually reinforcing though: tourism demand creates the opportunity, but the rent-gap (what you could earn on Airbnb vs long-term) determines whether it gets exploited. High-demand areas with strong tenant protections or high owner-occupancy might not convert the same way. But yes, the causality is overdetermined and I don't have the data to disentangle it cleanly - your framing is probably closer to the truth than mine implies.
On 1, the consultation is open on whether to have a borough share and it'll be the mayor's choice as you say, so no set share. Broadly for economic growth activities on the mayoral side but leaving it fairly open.
On 2, very fair! Definitely both, I just wonder whether rent to income gaps explain the behaviour of what will increasingly be mortgage-less owners. Look forward to seeing more of your work!
The best possible compliment - thank you! Though I suspect he'd have wanted cleaner identification than "here's a map, isn't it interesting. The tipping point dynamics do feel very Schelling-esque: individual decisions, collective patterns no one intended.
Hi, fair play in the socio-economic angle and your not wrong. However from a stylistic perspective this reads like the raw output of an LLM. I assume you had AI help, who doesn't, but when the writing reads like literal ai content it's a big turn off.
Uber shocked many complacent if not corrupt taxi companies into shape. Witness G7 in France. Airbnb is doing the same to the hotel industry in some places if allowed to.
I think it is important that we reflect on the best ways to regulate things that are not working and to promote housing rights and equality. And I agree that looking at the data can offer us precious indications about policy.
However, I found the reasoning and logical arguments of this post difficult to follow and in some points actually derimental to the efforts of many grassroots groups that are living the reality of what happens when cities are turned into Airbnbs. Or dangerous to the efforts of policy makers who have been fighting for years to find ways to come up with a legislation that, while not perfect, is actually working. Or oblivious to the academic research assessing that, while we still have a problem, regulation is working in some aspects. But there are also concrete examples of arguments that I found weak:
- "Airbnb is endogenous: ban it and the pressure finds another outlet". This claim is logically very weak: there is no evidence to support it and it is framed so broadly that it cannot be demostrated ("what other outlet" could it be? and how do we measure and test whether pressure will find it?"). It is an unfalsifiable hypothesis, a rhetorical phrase rather than an analytical conclusion, whil you are framing the piece as an analytical piece based on data. I personally agree that bans are problematic, mostly because they are very hard to enforce in realty. But to claim that Airbnb is endogenous and thus bans are useless or misguided is a very large leap to take without any data or counterfactual analysis. Moreover, this framing avoids engaging with a deeper possible argument for bans: that housing is a place to live, not a commercial asset to be systematically converted into tourist infrastructure. This is not an empirical claim, but a political and ethical one, and dismissing bans solely on efficiency grounds sidesteps that debate rather than addressing it.
- "The city is enforcing the law perfectly – it’s just enforcing the wrong law.". Is this an unchecked assumption or do you have ecidence that it's been enforced perfectly? I'd be curious as often the problem with regulation is precisely the enormous burden needed for good enforcements. Yet declaring enforcement “perfect” rhetorically shifts blame from compliance failure to policy design, all based on an assumption that doesn't seem proven. This is the same criticism I have for the claim "If the 90-day rule were meaningfully shaping behaviour… The map shows clusters there anyway." You cannot say that, as no comparison is made to pre-2015 London or to an unregulated control city to really prove that the 90 day didn't mitigated the phenomenon. The argument assumes regulation must eliminate clustering to be effective, which is unrealistic given the stage of Airbnb diffusion we have now reached..
- “A listing where wages don’t cover rent is more likely a survival strategy.” But was this actually tested? Were host-level income or ownership structures examined? Or is this inference based on averages and area-level indicators which might not be representative of hosts? This framing risks obscuring the well-documented role of corporate landlords, professional operators, and multi-property hosts, who are often most active precisely in high-pressure areas.
- “Airbnb didn’t create the problem; it just found the gaps.” This conclusion is not proven by the spatial clustering alone you provide with the map, which is a descriptive visualization of where Airbnb concentrates now, not an analysis of whether it created, amplified, accelerated, distributed or locked in housing harm. It may well be that Airbnb is not the root cause of housing scarcity and that wage stagnation, planning failures, and underbuilding are central. But it doesn't mean that it also doesn't have a role in this dynamic.
- Your examples of other cities are again flawed as there is no counterfactual analysis. Also some of them actually do take into account the spatial component you have suggested as central to regulation (amsterdam and florence at least, as far as I know).
- You present time regulation / space regulation / bans as competing approaches from which to choose the best. But this is not a war between measures: housing systems are complex and I doubt any of them alone would work. And could work everywhere anytime. You criticize bans and time regulation, but we could make similar arguments also against spacial regulation. A proper regulation isn't focused on one of these items alone operating in a vacuum, we need a policy that creates incentives through time regulation, sets limits on numbers and geography to reduce density, protects cities from corporate landlords running Airbnbs... and it could very well be that after an initial phase of solid regulation and a more solid general housing policy plan the ultimate goal becomes an Airbnb ban. These things don't need to be proposed as alternatives, but should all be assessed together (together with many more) to decide the specifics: what time limit would actually work to be a disincentive? what number cap would allow residential housing to breathe? which areas do we need to prioritize and which are still doing good but we need to protect as they could soon be at risk? what licence system could discourage coroprate landlords? how can we assure enforcement? etc.
I raise these points in the hope that future iterations of this work (if planned) will engage more carefully with existing research, activist knowledge, and the lived realities of communities that have been documenting and resisting the impacts of short-term rentals for years.
As far as recommendations are concerned though, density cap seems politically impossible. You would need to tell people wanting to rent out a room, Oops, your neighbors already filled the quota, you’re SOL.
Regulation of time is at least somewhat equitable, regulation of space is intrinsically arbitrary. That includes lotteries.
A question occurs to me - can we really infer from the income-pressure linkthat this really about individuals letting their houses to cover income gaps? Anecdotally, plenty of Airbnbs seem to be part of larger operations run as pseudo-hotels rather than the more circumstance-driven uses (2nd homes after moving in with a partner, periods away from the house etc.).
Another explanation might be that being a traditional landlord has higher risk in areas where people are underpaid, so wealthy owners let them out as pseudo-hotels to rich tourists instead of renting to locals.
Some of the comment in the blog seems to assume that a good proportion of Airbnb lets are driven by economic necessity - which I take to be people letting a spare room or two in the house where they are living. I'd assumed almost all lets in Central London are not like this - they are buy to let properties. Not sure who's right. But the answer would give an insight into the effectiveness of regulation, in that it's hard to believe someone letting out a buy to let property on Airbnb would survive letting it for only 90 days a year, so are quite likely to be letting via other sites too.
This is utterly fantastic. Thanks for the hard work.
Great stuff! Why isn't there a big tech company trying to recruit you right now?
Thanks! The algorithm hasn't found me yet, I'm afraid.
Such an interesting piece, really good analysis.
Two things that came to mind when reading:
1. Until the overnight visitor levy comes in (if mayors choose to give councils a share), councils themselves don't really benefit from tourism spend financially. No council tax, very minor use of services paid with fees, I suppose indirectly through business rates growth but very minor and indirect given it's based on premises size. It's purely a cost to them.
2. Sure you thought about this but at first glance, aren't the hot spots just tourism demand driven by good connectivity/centrality/fashionableness incl on TikTok, rather than rent-income? Know slightly overdetermined but feels to me like demand is driving spatial distribution. Interesting caveat that high density requires either low social housing proportion (legally, unless now private through RtB) and will be boosted by absentee landlordism.
Thank you! On your two points:
Really important point and one I didn't address - the fiscal asymmetry is stark. Councils bear the externalities (enforcement, housing pressure, infrastructure wear) but capture almost none of the upside. The visitor levy might help but as you say, it's optional and the revenue share to boroughs is uncertain. There's something perverse about a system where local authorities are essentially subsidising a platform's business model. Do you know if there's any breakdown yet of how the levy revenue would be distributed - what share goes to the GLA vs boroughs, and whether it's earmarked for anything specific?
You're right that demand is doing a lot of the work spatially - connectivity, centrality, TikTok-ability. I probably undersold that. My hunch is it's mutually reinforcing though: tourism demand creates the opportunity, but the rent-gap (what you could earn on Airbnb vs long-term) determines whether it gets exploited. High-demand areas with strong tenant protections or high owner-occupancy might not convert the same way. But yes, the causality is overdetermined and I don't have the data to disentangle it cleanly - your framing is probably closer to the truth than mine implies.
On 1, the consultation is open on whether to have a borough share and it'll be the mayor's choice as you say, so no set share. Broadly for economic growth activities on the mayoral side but leaving it fairly open.
On 2, very fair! Definitely both, I just wonder whether rent to income gaps explain the behaviour of what will increasingly be mortgage-less owners. Look forward to seeing more of your work!
Thomas Schelling would have loved this 😁
The best possible compliment - thank you! Though I suspect he'd have wanted cleaner identification than "here's a map, isn't it interesting. The tipping point dynamics do feel very Schelling-esque: individual decisions, collective patterns no one intended.
Hi, fair play in the socio-economic angle and your not wrong. However from a stylistic perspective this reads like the raw output of an LLM. I assume you had AI help, who doesn't, but when the writing reads like literal ai content it's a big turn off.
Hi - just wondering whether you scraped the data (like you said above: "I scraped every listing"), or downloaded "Inside Airbnb" data (which your github code references: --url "http://data.insideairbnb.com/united-kingdom/england/london/2025-01-01/visualisations/listings.csv")?
Uber shocked many complacent if not corrupt taxi companies into shape. Witness G7 in France. Airbnb is doing the same to the hotel industry in some places if allowed to.
Great point. I’m still mad at Airbnb.
https://fowlerbrand.substack.com/p/your-brand-is-a-lie-airbnb?r=18uoy&utm_medium=ios
Amazing work! I would love to see that applied to Paris!
I think it is important that we reflect on the best ways to regulate things that are not working and to promote housing rights and equality. And I agree that looking at the data can offer us precious indications about policy.
However, I found the reasoning and logical arguments of this post difficult to follow and in some points actually derimental to the efforts of many grassroots groups that are living the reality of what happens when cities are turned into Airbnbs. Or dangerous to the efforts of policy makers who have been fighting for years to find ways to come up with a legislation that, while not perfect, is actually working. Or oblivious to the academic research assessing that, while we still have a problem, regulation is working in some aspects. But there are also concrete examples of arguments that I found weak:
- "Airbnb is endogenous: ban it and the pressure finds another outlet". This claim is logically very weak: there is no evidence to support it and it is framed so broadly that it cannot be demostrated ("what other outlet" could it be? and how do we measure and test whether pressure will find it?"). It is an unfalsifiable hypothesis, a rhetorical phrase rather than an analytical conclusion, whil you are framing the piece as an analytical piece based on data. I personally agree that bans are problematic, mostly because they are very hard to enforce in realty. But to claim that Airbnb is endogenous and thus bans are useless or misguided is a very large leap to take without any data or counterfactual analysis. Moreover, this framing avoids engaging with a deeper possible argument for bans: that housing is a place to live, not a commercial asset to be systematically converted into tourist infrastructure. This is not an empirical claim, but a political and ethical one, and dismissing bans solely on efficiency grounds sidesteps that debate rather than addressing it.
- "The city is enforcing the law perfectly – it’s just enforcing the wrong law.". Is this an unchecked assumption or do you have ecidence that it's been enforced perfectly? I'd be curious as often the problem with regulation is precisely the enormous burden needed for good enforcements. Yet declaring enforcement “perfect” rhetorically shifts blame from compliance failure to policy design, all based on an assumption that doesn't seem proven. This is the same criticism I have for the claim "If the 90-day rule were meaningfully shaping behaviour… The map shows clusters there anyway." You cannot say that, as no comparison is made to pre-2015 London or to an unregulated control city to really prove that the 90 day didn't mitigated the phenomenon. The argument assumes regulation must eliminate clustering to be effective, which is unrealistic given the stage of Airbnb diffusion we have now reached..
- “A listing where wages don’t cover rent is more likely a survival strategy.” But was this actually tested? Were host-level income or ownership structures examined? Or is this inference based on averages and area-level indicators which might not be representative of hosts? This framing risks obscuring the well-documented role of corporate landlords, professional operators, and multi-property hosts, who are often most active precisely in high-pressure areas.
- “Airbnb didn’t create the problem; it just found the gaps.” This conclusion is not proven by the spatial clustering alone you provide with the map, which is a descriptive visualization of where Airbnb concentrates now, not an analysis of whether it created, amplified, accelerated, distributed or locked in housing harm. It may well be that Airbnb is not the root cause of housing scarcity and that wage stagnation, planning failures, and underbuilding are central. But it doesn't mean that it also doesn't have a role in this dynamic.
- Your examples of other cities are again flawed as there is no counterfactual analysis. Also some of them actually do take into account the spatial component you have suggested as central to regulation (amsterdam and florence at least, as far as I know).
- You present time regulation / space regulation / bans as competing approaches from which to choose the best. But this is not a war between measures: housing systems are complex and I doubt any of them alone would work. And could work everywhere anytime. You criticize bans and time regulation, but we could make similar arguments also against spacial regulation. A proper regulation isn't focused on one of these items alone operating in a vacuum, we need a policy that creates incentives through time regulation, sets limits on numbers and geography to reduce density, protects cities from corporate landlords running Airbnbs... and it could very well be that after an initial phase of solid regulation and a more solid general housing policy plan the ultimate goal becomes an Airbnb ban. These things don't need to be proposed as alternatives, but should all be assessed together (together with many more) to decide the specifics: what time limit would actually work to be a disincentive? what number cap would allow residential housing to breathe? which areas do we need to prioritize and which are still doing good but we need to protect as they could soon be at risk? what licence system could discourage coroprate landlords? how can we assure enforcement? etc.
I raise these points in the hope that future iterations of this work (if planned) will engage more carefully with existing research, activist knowledge, and the lived realities of communities that have been documenting and resisting the impacts of short-term rentals for years.
Very interesting analysis, thank you!
As far as recommendations are concerned though, density cap seems politically impossible. You would need to tell people wanting to rent out a room, Oops, your neighbors already filled the quota, you’re SOL.
Regulation of time is at least somewhat equitable, regulation of space is intrinsically arbitrary. That includes lotteries.
A question occurs to me - can we really infer from the income-pressure linkthat this really about individuals letting their houses to cover income gaps? Anecdotally, plenty of Airbnbs seem to be part of larger operations run as pseudo-hotels rather than the more circumstance-driven uses (2nd homes after moving in with a partner, periods away from the house etc.).
Another explanation might be that being a traditional landlord has higher risk in areas where people are underpaid, so wealthy owners let them out as pseudo-hotels to rich tourists instead of renting to locals.