Charles Knox-Vydmanov
@knoxvydmanov.bsky.social
Researcher, trainer, policy analyst. Social protection, disability, public finance, pensions. Independent consultant, previously ILO and HelpAge International.
Doesn't this also mean that defence and public servant pensions are included?
October 23, 2025 at 10:51 PM
Doesn't this also mean that defence and public servant pensions are included?
Quite. I just wonder what happens when things people like actually stop working.
Very interesting about the SSA data!
Very interesting about the SSA data!
March 3, 2025 at 11:14 AM
Quite. I just wonder what happens when things people like actually stop working.
Very interesting about the SSA data!
Very interesting about the SSA data!
Just to clarify, I'm not underplaying the huge damage that he/they can do, just that no one exists purely outside these dynamics of public opinion.
March 3, 2025 at 10:30 AM
Just to clarify, I'm not underplaying the huge damage that he/they can do, just that no one exists purely outside these dynamics of public opinion.
In a way, the Friedman example is instructive on this. Despite his huge intellectual influence through the 80s and 90s, the global outcome was not to discard social insurance systems.
In the end, whatever Elon says will have to rub up against the political economy of this.
In the end, whatever Elon says will have to rub up against the political economy of this.
March 3, 2025 at 10:30 AM
In a way, the Friedman example is instructive on this. Despite his huge intellectual influence through the 80s and 90s, the global outcome was not to discard social insurance systems.
In the end, whatever Elon says will have to rub up against the political economy of this.
In the end, whatever Elon says will have to rub up against the political economy of this.
I'd add, people of younger ages also often provide financial support to their older parents (or would in the absence of adequate pension systems). So they have a direct stake in questions of pension adequacy.
January 17, 2025 at 2:05 PM
I'd add, people of younger ages also often provide financial support to their older parents (or would in the absence of adequate pension systems). So they have a direct stake in questions of pension adequacy.
I really like this! My immediate reaction was that *even* for regular migrants, factors such as informal employment and social assistance limited to citizens means few may have access to social protection. But I guess that is captured under the "several barriers"!
January 7, 2025 at 11:08 AM
I really like this! My immediate reaction was that *even* for regular migrants, factors such as informal employment and social assistance limited to citizens means few may have access to social protection. But I guess that is captured under the "several barriers"!
Florian and I had a bit of a debate about this!
The article is interesting, although I do think there is a danger sometimes of conflating questions of AI (e.g. fraud detection), eligibility criteria and benefit adequacy.
bsky.app/profile/knox...
The article is interesting, although I do think there is a danger sometimes of conflating questions of AI (e.g. fraud detection), eligibility criteria and benefit adequacy.
bsky.app/profile/knox...
I think we need to be careful here. It's not obvious to me that the "disparities" described in the article and accompanying report result in discrimination.
Indeed, the report reads like one of an organisation taking these issues seriously (or perhaps I am naïve).
Indeed, the report reads like one of an organisation taking these issues seriously (or perhaps I am naïve).
December 13, 2024 at 9:28 PM
Florian and I had a bit of a debate about this!
The article is interesting, although I do think there is a danger sometimes of conflating questions of AI (e.g. fraud detection), eligibility criteria and benefit adequacy.
bsky.app/profile/knox...
The article is interesting, although I do think there is a danger sometimes of conflating questions of AI (e.g. fraud detection), eligibility criteria and benefit adequacy.
bsky.app/profile/knox...
Agree!
Just on the question of the AI being perceived as more "scientific", in this case the human reviewers don't know whether the review is from the AI, or randomly selected for review.
Just on the question of the AI being perceived as more "scientific", in this case the human reviewers don't know whether the review is from the AI, or randomly selected for review.
December 6, 2024 at 6:19 PM
Agree!
Just on the question of the AI being perceived as more "scientific", in this case the human reviewers don't know whether the review is from the AI, or randomly selected for review.
Just on the question of the AI being perceived as more "scientific", in this case the human reviewers don't know whether the review is from the AI, or randomly selected for review.
2. Fraud detection: An important/necessary function of implementing agencies. Is use of AI as a filter inherently a problem where review/decision is by human? I'm not sure. I guess as the article (and you) say it comes down to the how, and to the process being transparent.
December 6, 2024 at 2:34 PM
2. Fraud detection: An important/necessary function of implementing agencies. Is use of AI as a filter inherently a problem where review/decision is by human? I'm not sure. I guess as the article (and you) say it comes down to the how, and to the process being transparent.
Agree! My basic sense is we need to distinguish between different types of use. E.g.
1. Algorithmic/black box decision-making in eligibility determination: problematic by its nature (e.g. proxy means testing) and can only get worse with AI. The key issue for me is inherent lack of transparency.
1. Algorithmic/black box decision-making in eligibility determination: problematic by its nature (e.g. proxy means testing) and can only get worse with AI. The key issue for me is inherent lack of transparency.
December 6, 2024 at 2:34 PM
Agree! My basic sense is we need to distinguish between different types of use. E.g.
1. Algorithmic/black box decision-making in eligibility determination: problematic by its nature (e.g. proxy means testing) and can only get worse with AI. The key issue for me is inherent lack of transparency.
1. Algorithmic/black box decision-making in eligibility determination: problematic by its nature (e.g. proxy means testing) and can only get worse with AI. The key issue for me is inherent lack of transparency.
I can see that, and the point on transparency/scrutiny I think is really key as you say. Which is the case of the Netherlands that you have in mind?
December 6, 2024 at 2:07 PM
I can see that, and the point on transparency/scrutiny I think is really key as you say. Which is the case of the Netherlands that you have in mind?
Fascinating! These are great resources and charts. Thank you so much!
In case it's not already on your radar, this is another nice resource for OECD level comparison: www.oecd-ilibrary.org/employment/d...
Seems to roughly match the UK and Sweden data (reassuringly)
In case it's not already on your radar, this is another nice resource for OECD level comparison: www.oecd-ilibrary.org/employment/d...
Seems to roughly match the UK and Sweden data (reassuringly)
December 6, 2024 at 1:31 PM
Fascinating! These are great resources and charts. Thank you so much!
In case it's not already on your radar, this is another nice resource for OECD level comparison: www.oecd-ilibrary.org/employment/d...
Seems to roughly match the UK and Sweden data (reassuringly)
In case it's not already on your radar, this is another nice resource for OECD level comparison: www.oecd-ilibrary.org/employment/d...
Seems to roughly match the UK and Sweden data (reassuringly)
OK... accessed the (great) article and can see form underlying data that this doesn't include PiP. So I have the same question as you, Sam!
By the way, lovely chart from the ONS data on the changing composition of incapacity benefit caseload.
By the way, lovely chart from the ONS data on the changing composition of incapacity benefit caseload.
December 6, 2024 at 12:39 PM
OK... accessed the (great) article and can see form underlying data that this doesn't include PiP. So I have the same question as you, Sam!
By the way, lovely chart from the ONS data on the changing composition of incapacity benefit caseload.
By the way, lovely chart from the ONS data on the changing composition of incapacity benefit caseload.