Today I wish to look at an area I have been involved in formally since 2012 and yes the problems with UK official statistics do go back to then. In fact the problems then were driven by something they did in 2010 and then tried to cover up and dissemble about. Those themes have been on repeat in the intervening years. But let me first bring you up to date on a different area (back then it was inflation) where to quote Taylor Swift there is “trouble,trouble,trouble”.
The UK statistics agency has drawn up contingency plans to delay the launch of its new labor market survey by six months, raising the possibility of a further holdup as it battles to restore the credibility of the country’s economic data. (Bloomberg)
For those unaware UK labour market statistics have been in trouble for some years and in the end even the Office for National Statistics had to publicly admit this after doing its best to cover it up. Or as a 2019 parliamentary inquiry by PACAC put it there is a tendency to “bury bad news”. That poses its own problem in that it is now 2026 so the previous PACAC inquiry was ignored. Now back to present difficulties.
While the Office for National Statistics wants to make a transition in November as scheduled, another scenario where the survey is launched in May 2027 is also an option under consideration, according to people familiar with the matter. One added there are still a number of technical difficulties that need to be ironed out that could lead to another delay.
The ONS plans to decide in the summer whether to stick to the November roll-out date.
The emphasis is mine and I have done so to show the scale of the mess here. After quite some time they still have “technical difficulties”.
The ONS accelerated work on a new “Transformed Labour Force Survey” after a collapse in responses to its current version forced it to temporarily suspend the publication of unemployment, employment and economic inactivity estimates in late 2023.
Or if you prefer the work was obviously not accelerated enough as some years later there are still problems. Or as Bloomberg put it.
It would just be the latest delay to the plans to switch to a shorter and online-focused questionnaire that the ONS hopes will revive its key labor market statistics. If it was launched in May 2027, it would mark over 3 1/2 years since the problems were first publicly flagged by the ONS.
The essential problem is that the response rate collapsed and in general fell to just 14% with one area as low as 10%. This is what really made the UK stand out as engagement was already poor and fell heavily. Those who follow the comments section will know I suggested a generic issue with labour force surveys when discussing Canada. But the UK is in a worse place due to an awful response rate and the attempt to hide it. for example I believe that Better Statistics raised the issue in 2020.
Moving onto why this matters? There are quite a few reasons but here is a big one.
The lack of credible labor statistics has been a thorn in the side of BOE officials who have voiced their criticism about the ONS’s response to the crisis………Policymakers have been forced to rely more heavily on private-sector surveys of the jobs market, though these shed little light on the millions of so-called economically inactive people who are neither in work nor looking for employment.
These things can particularly matter as if we look back to 2013 it was the labour market that showed an economic upturn before the GDP data.
Inflation Measurement
Whilst the labour market is the present enfant terrible so much of this has been true in the inflation arena and it too came up in the comments section yesterday as JW challenged the use of what we call Imputed Rents (OER in US parlance). The recent machninations in US inflation measurement would be a joke if it were not so serious.
The issue for the UK started in 2010 when the ONS changed the way they looked at clothing inflation and got numbers so different they effectively blew up the series. The Treasury saw an opportunity here as it has never liked the RPI inflation measure due to the fact that it nearly always provides a higher reading that CPI. So there was a debate and a type of vote and of the 440 or so replies around 400 were in favour of the RPI and some were spoilt papers.
But as we return to arrogance the establishment ignored this. They were delayed as companies like the National Grid pointed out they had tens of billions of RPI linked debt. But a couple of years later the establishment and the ONS drove through the Imputed Rent driven CPIH. In 2015 this was further pushed by Paul Johnson of the Institute for Fiscal Studies, something no doubt he is hoping will be forgotten. My argument against has been than Imputed Rents are fantasy numbers as you buy a property to not pay rent! This leads to the risk that they can be made up as appears to have happened recently in the United States due to the shut down of the Bureau of Labor Statistics in the autumn.
There is another problem which I highlighted in my evidence to the current parliamentary inquiry.
The official view is that rents need a stock of 14 months. The problem with that is that your measure is some time behind reality. I have long thought that the lag may be of the order of 18 months but the period since the worst of the cost of living crisis suggests it may be as long as 2 years.
Which meant that the frightfully clever Imputed Rent methodology produced by so many frightfully clever people confused up with down.
Apart from being misleading in itself as we came into 2025 this methodology really backfired as we were told we had record levels of rental inflation when some more up to date measures had UK rents falling. Yes they really did publish figures saying rent rises were at or near record levels when in reality according to more concurrent measures they were somewhere between low and falling.
There always was a theoretical issue when you produce figures for say June and they are not for then, but that was ignored because everyone was too clever for that.
Looking ahead there are plans to replace the RPI with CPIH. After what has happened in America that looks awful but they are hoping that you will not know that.Tobe specific it will continue to be called RPI but actual prices ( house prices and mortgage costs) will be replaced by Imputed Rents.
Comment
There are other areas I have challenged and I will add my evidence to PACAC at the bottom of this as it has more detail. But stepping back the themes are of arrogance, Treasury involvement, unqualified people being in charge, and cover-ups. For example how did someone who as far as I can tell was qualified in nursing rather than statistics become National Statistician?
Acting national statistician Emma Rourke is to leave post to take up a new job, the UK Statistics Authority has announced.
UKSA said Rourke would depart later this month ahead of starting work on the Whitehall in Industry scheme. (Civilserviceworld)
There is now new management at the ONS but things need to go further as the UK Statistics Authority and the Office for Statistics Regulation have been failures as well.
Let me finish with another establishment error which was sending the ONS to Newport in Wales. Whatever one might think of that it has struggled ever since to get qualified staff.
