Stay informed with free updates
Simply sign up to the Artificial intelligence myFT Digest — delivered directly to your inbox.
MainFT, yesterday:
Chancellor Rachel Reeves has been privately warned by the UK fiscal watchdog that its estimates for productivity are likely to be downgraded ahead of the Budget, making big tax rises more probable.
[…]
People briefed on Budget preparations have warned the total fiscal hole could amount to “tens of billions” of pounds — perhaps to £30bn. The OBR downgrade would make up a large portion of the gap.
The Office for Budget Responsibility’s productivity guesstimates have for a long time been one of the most Tobias Fünke-able parts of the UK macroeconomic framework…
…as highlighted by this chart (stolen from Main’s excellent deep-dive into this topic):
Some content could not load. Check your internet connection or browser settings.
There are two black box dynamics at work here: firstly, whether the OBR is sufficiently equipped with data to make these kinds of estimates (given everything not going on at the Office for National Statistics); secondly, whether they are equipped with the models needed to produce good estimates from whatever data they do end up using.
Given the context of persistent over-optimism, a downgrade seems reasonable.
But at the risk of Fünkeing ourselves… what if this time actually is different?
FT Alphaville is certainly sceptical of artificial intelligence. But it’s worth acknowledging that journalism may be — for a job that involves so much writing — unusually AI-resistant. After all, it ideally relies on relaying new and accurate information in a way that is antithetical to an LLM.
Not all knowledge/content-based jobs are like that. After all, Britain is (among other things) a consultation nation, and that’s a potentially fertile area for AI — it’s reductive to say consultancy just involves producing briefing reports and PowerPoint presentations, but it doesn’t not involve those things.
In that context, it’s interesting that a period of apparently slow-but-steady economic growth (especially once inflation is discounted) has coincided with a mushy labour market. It’s a differential highlighted nicely by this chart, published by Morgan Stanley last month, showing the spread between the S&P purchasing managers’ index readings for composite activity and composite employment:
MS’s Bruna Skarica writes:
taken at face value, the uptick in the headline PMIs paired with a consistently soft hiring index points to a rebound in productivity.
Relatedly, here’s a bit from the latest UK services PMI report:
Many firms noted either hiring freezes, the non-replacement of voluntary leavers or ongoing redundancies. In some cases, service providers also reported a focus on automation and investments in enhanced productivity to help offset rising wage bills.
Is the UK undergoing a quiet supply-side improvement? There are plenty of reasons to be sceptical of PMIs, but equally, there’s plenty of reason to be sceptical about much of the UK’s macroeconomic data at the moment.
The OBR is not ignoring AI. Its July ‘Fiscal risks and sustainability’ report says:
The rapid development and dissemination of artificial intelligence could be one driver of upside risk to future productivity growth, although the magnitude and timing of the possible boost to productivity remains highly uncertain.
But it’s hard to tell how, if such a shift began to occur, the OBR would initially spot it. Maybe it’s already happening.
(There’s also, obviously, the possibility of productivity improvements that have little to do with AI, such as the pick-up in NHS productivity identified last year by the Institute for Fiscal Studies.)
Clearly, a downgrade looms — putting the OBR’s stance more in line with a sceptical Bank of England — but if there is any merit in nascent indications of a productivity pick-up, it raises a couple of questions.
The first is well-worn: if fiscal projections are to be based on such conflicted and (sometimes) flimsy data, does it make sense to tie spending decisions to them so closely?
The second is possibly trickier: if the OBR doesn’t expect AI to improve productivity, then what exactly does it expect the end result of all that capex to be?
Further reading:
— Britain’s statistics scandal means it cannot answer its most pressing questions (MainFT)
