Britain's fraying social fabric exposed
Britain's fraying social fabric exposed
A system that fails to tax wealth fairly, or to tackle increasing poverty or economic and social inequality, isn't working. And technological progress might make things worse rather than better.
Last Monday I wrote about the state of our social fabric (and of the social contract). In the days that followed, a few superficially unrelated events showed how thin the fabric of fairness is wearing in Britain.
A failure to tax income from wealth fairly…
Firstly (and this is not in chronological order), there were the tax returns published by Rishi Sunak and Sir Keir Starmer. It was widely reported that Sunak’s effective tax rate last year was 22%, whereas Starmer’s was 32%. This despite the fact that Sunak’s income last year was almost £2m, compared to just over £210,000 for Starmer. How could this be? It’s because £1.6m of Sunak’s income last year was in the form of capital gains, which was taxed at just over 20%, much less than income from work (as it happens, £85k of Starmer’s income was also from capital gains).
In October’s Fairness Index we included an indicator on tax, based on research by the University of Warwick and LSE showing that some people who earn £10m pay an effective tax rate of just 21%, a similar rate to someone on a median UK salary of £30,000. This pointed to the same cause: some largely ‘unearned’ sources of income (investments, capital gains and so on) are taxed at much lower rates than earnings. Overall, the exchequer loses an estimated £20bn per year because of the discrepancy between headline tax rates and effective tax rates.
There’s growing momentum to reform taxes on wealth (and a majority of Britons now support an annual wealth tax, which is much more ambitious than simply taxing income from capital gains at the same rate as income from work). But public attitudes to wealth are complicated (an issue that we intend to explore further in the coming months), which is perhaps why Labour disappointingly stated on Thursday that they have “no plans to raise capital gains tax.” As Ben Ansell suggested in his post last week, one way to build public support (or reduce political risks) around such a policy, especially for floating voters, is to outline the arguments in terms of fairness: “hammering on about it, without making a broader case in language people understand about, say, fairness, is going to be pretty meaningless to these people”. This policy appeals to fairness on two levels: fair treatment (same rates for everyone) and fairer outcomes (helping to reduce inequality). So it should be sellable on the doorstep, if the right language is used.
… Is a missed opportunity to tackle unfairness in society
Meanwhile, more and more evidence is produced to show that we live in a society that is not only unequal, but unfairly unequal.
On Thursday the government published its annual Households below average income (HBAI) statistics, which showed that the number of children living in poverty in the UK has increased, to 29% of children (4.2 million in all) in 2021/22. The figures (via the End Child Poverty Coalition) also revealed that 800,000 children lived in households that needed to use a food bank, 71% of poor children were living in working families, and 42% of children in families with three or more children were living in poverty. Rates of child poverty were higher in Black and Asian families (53% and 47% respectively) than in White families (25%).
On the subject of racial inequalities, the Office for National Statistics published some analysis earlier this month of ethnic group differences in health, employment, education and housing from the 2021 Census. The data showed that some groups (people who identified as White Gypsy or Irish Traveller, and Bangladeshi) reported poorer than average health outcomes despite being younger than than average. The analysis found that people in ethnic groups with higher rates of disability and poorer health typically also provided more unpaid care. And people who identified as African, Caribbean, White and Black African, and White and Black Caribbean were just over twice as likely to be unemployed as the average Briton, while Black households were more than twice as likely to be living in social rented housing than the average household. However, people who identified as Chinese or Indian had the highest levels of education of the 19 ethnic groups analysed, closely followed by people who identified as African.
And speaking of health inequalities, a new campaign has been launched to raise awareness of them in the UK and to ‘improve health opportunities’ for everyone. Health Equals is a group of organisations and voices across different sectors, including employment, housing, education and the environment, who all want to make a positive difference to society’s health and wellbeing. We’re proud to be a member. The launch campaign, #LivesCutShort, is designed to raise awareness of the differences in life expectancy across different parts of the UK, and the factors that contribute to those differences. As we tweeted last week, it's not fair that the place and circumstances you are born into affects your health, wellbeing and life expectancy.
Meanwhile, we’re largely ignoring how technology might exacerbate unfairness
Could society be about to become a lot less fair?
Iason Gabriel, an ethics researcher at DeepMind, commented last week on a revealing paragraph in a paper that OpenAI published about the ‘safety’ challenges and processes linked to the recent deployment of GPT-4, the latest incarnation of the AI chatbot. The paper makes for alarming reading, even when it focuses on the work that has gone into making it safer; there is an eye-popping section on various nefarious use cases for the tool that have been anticipated and blocked. But perhaps most worrying of all is this paragraph, which Gabriel highlights:
“AI systems will have even greater potential to reinforce entire ideologies, worldviews, truths and untruths, and to cement them or lock them in, foreclosing future contestation, reflection, and improvement. In fact, we should expect AI systems to do so in the absence of anticipatory work to address how best to govern these systems, how to fairly distribute the benefits they generate, and how to fairly share access.”
Where is the public conversation about exactly this topic? It needs to happen now, before it’s too late. And articles about whether ChatGPT would make a good prime minister don’t count.