Sticky prices and sh*tty platforms

Hi all,

Set your calendars for late November — and no, not just Radiohead’s first tour in seven years. It’s Rachel Reeves’s second Budget (see our chart of the week for more). Reeves is under pressure to prove she can deliver a British economy which is fitter, happier and more productive without making the sums look like 2+2=5. And while she hopes for no alarms and no surprises, the reality of running Britain’s finances is that there’s always a wolf at the door. Of course as we write she’s more likely to be playing Exit music (for a Deputy).

Hope back to school week hasn’t been too painful. I’d like to say a huge thanks to our guest editors Giles and Tera for providing us with some great summer reading and setting the bar high for my return. Read on for the latest on Jackson Hole, the AI takeover and why social media has got worse.

And if you fancy a job using your talents to help us lift living standards, apply here.

Have a great weekend,

Ruth
Chief Executive
Resolution Foundation

Robots return Fears of AI taking over jobs may be becoming a reality for some (dun, dun, dun…). This paper uses a large sample of payroll records to determine AI’s effect on the US job market. Employment for young US workers aged 22-25 in jobs most exposed to AI has declined since the launch of ChatGPT in late 2022 – a 6 per cent drop in employment – while older workers and other young workers saw growth. This is not explained by other shocks happening to hit firms with a concentration of AI-exposed young employees. Young software developers have been most affected, with employment falling by nearly a fifth. The authors suggest that entry-level jobs have declined where AI can automate work and replace workers, but not necessarily where it makes tasks easier. These findings chime with recent business survey evidence from the NY Fed suggesting a bigger impact on hiring than firing.

Down the drain We’ve all despaired at the declining quality of certain social media platforms. While there might be an Elon Musk-shaped numerous reasons for this, Cory Doctrow’s theory of Enshittification neatly captured the public’s growing discontent back in 2022. Paul Krugman’s recent blog spells out why he thinks this might be much more general, and becoming more common. Wherever a product is more attractive the more people use it, the profit-maximising strategy is to initially provide great value to customers, build up the base, then enshittify. The outcome? Not death but stagnation, with the monopolist increasing prices and/or reducing quality enough that the base stops growing, but not enough to drive it away. This is bad for all of us – including the tech bros running these companies, whose public approval has fallen off a cliff.

In a hole This year’s Jackson Hole Symposium, the annual jamboree for elite central bankers had an unusually RF-adjacent theme (and no, we haven’t taken it over… yet). The discussion pivoted around structural economic challenges, with a focus on labour, demographics and productivity. Fed Chair Jay Powell had to navigate hinting at interest rate cuts to come while being clear it had nothing to do with political pressure. But it’s worth delving into the other papers at the symposium too. A thoughtful look at the ability of central banks to look through inflationary shocks is presumably on the MPC’s summer reading list. Another paper finds that decline in US labour mobility has not been as dramatic as feared, but that policies to boost it would still be welcome.

Data dramas Last week we offered the ONS, if not the economy, some good news: the Labour Force Survey looks to have a decent handle on the unemployment rate. But more attention is needed on our earnings data too. Concerns have been raised for some time about the ONS’s main earnings survey, the Annual Survey of Hours and Earnings (ASHE). A team of researchers has been arguing that ASHE over-estimates earnings because its weights don’t fully account for the lower response rate among small businesses. Their latest paper suggests this means it’s underestimating the gender pay gap. This has real world impacts – an ONS methodology change last year led to an upward revision to its earnings estimates (rather than the downward one this evidence would suggest), which in turn led to a larger-than-expected rise in the National Living Wage this April.

Something for party conference season?

Parliament returned from summer recess on Monday but it’ll be off again soon for party conference season, which started today with Reform. To help you navigate through the next month of speeches, policy announcements and (for a lucky few) drinks receptions with warm white wine, we teamed up with UK in a Changing Europe to produce an evidence-based assessment of the key issues confronting our politicians. Here are some of my favourite nuggets:

• “The government’s strongest card, at least in the short term, is that of Mrs Thatcher in the early 1980s: there is no obvious or convincing alternative on offer.” Jonathan Portes
• Interventions to reduce the cost of certain essentials can only go so far, according to our very own Mike Brewer. If household bills were reduced by £300 a year that would result in someone on a median income having been 1% better off in 2025-26.
• Isacc Delestre of the IFS gives a cutting take on our tax system, “a needlessly opaque system, which imposes an unjustified penalty on income from employment, drags on growth and makes raising revenue more costly than it needs to be”.
• Paul Cheshire argues that the Government has not “grasped the reality” of how far the planning system needs reform. He points out that in the first quarter of 2025 planning applications were the lowest of any quarter in the last decade – not a great sign for future building.

Anita Charlesworth was optimistic at our event about the prospect of AI improving health outcomes. But she highlights in her essay the challenges of delivering reform. She says the 10 year plan has similar levels of ambition as that of the 2000s when funding increased by 6.8% a year. The equivalent figure today is 2.8% a year (both in real terms).

Chart of the week

And so to that Budget. As Chart of the Week shows, Chancellors tend to do almost all of their tax raising in the first Budget after an election, building space for goodies later in the Parliamentary term. Rachel Reeves did the first part – the £40-odd billion of tax rises were similar to Ken Clarke in 1993. But policy U-turns, higher gilt yields and a likely growth downgrade – coupled with the fact that spending totals were agreed only last June – means that further tax rises are needed. Join us on 23 September to discuss what the Government’s tax strategy should be. How unprecedented would a second round of tax rises be? Not completely. Ken Clarke came back for more in late 1993, taking tax rises in that year to over £60 billion. The second Budget of Boris Johnson’s term (admittedly Rishi’s first) also included over £30 billion of tax rises – though this came alongside unprecedented state support of Covid-19 shutdown Britain. A tough Budget beckons but maybe Radiohead can help her beat second album syndrome…