131. Why The Stats Behind The Budget Might Be Wrong

0:11Hello and welcome to The Rest is Money with me Steph McGovern. I'm with me, Robert Paston. So today with us is a man, I know you know very well, Robert, because you work together at the FT, but he's a man I've interviewed quite a few times as well. He's had loads of amazing jobs. He's currently chair of the UK Statistics Authority. And these are the guys who basically kick off when particularly politicians misuse statistics. The person we're talking about is Robert Chost, but before that, he was the boss of the OBR, which we have talked about a lot on this show, haven't we, Robert? We have the Office of Budget Responsibility set up in 2010 by George Osborne.

0:47It is an independent forecaster of growth, inflation, but more importantly, when it comes to decisions taken by chancellors, an independent forecaster of the state of the public finances. It's become a very controversial organisation because, you know, Liz Truss, for example, she's prime minister, highly critical of it. The Tory right don't like it. So one of the things obviously we're going to talk to him a bit about is, has the OBR made a positive contribution to decision making by governments? And I mean, I should point out, I'm incredibly fond of Robert. We were in healthy competition back in the 90s when he was economics editor at the Financial Times, I was political editor, and we have remained mates since then.

1:36So I probably won't be putting the boot in quite as aggressively to Robert as to some of our guests, but I hope you'll forgive me for that, Steph. Yeah, here's our interview with Robert Chott. Robert, really great to see you. You're now the man who keeps, allegedly, as you're trying to keep all our politicians and others, honest about use of data and statistics. But you were the founding boss of the Office for Budget Responsibility. You ran it for 10 years. In recent years, the OBR has become quite a controversial organisation. I mean, I remember when it was set up, I used to think, God, dull as ditch water, you know, independent.

2:18That was the aim. Anyway, you know, you are absolutely, not you, but the OBR is now absolutely at the centre of all sorts of political arguments. I just want to focus on one aspect at the moment, which is that on all sensible analysis, because interest rates have gone up, the government has run out of what's known as headroom under its fiscal rules. It's another way of just saying it's run out of money if it's going to meet these self-imposed rules for keeping the nation's debt under some kind of control. Now, in your view, when there are these ups and downs at any particular moment in time in respect of meeting the fiscal rules, should the government go into emergency mode?

3:14mode. It's run out of money. So should it now cut its spending plans, put up taxes? What's the rational approach, the rational use of these rules? Well, as you say, I mean, the tricky position to start with is that the interest rates at which government can borrow have gone up, not just in the UK, but in most countries. And that's largely down to expectations that, you know, tariff policy, Donald Trump's tariff policy is likely to push up inflationary pressure. So that's that pushes up the amount that it it costs all governments to borrow. And in the UK, that seems to have gone up a bit more than it has done in most other places.

3:50So when the OBR is putting together its forecast, all other things being equal, and of course all other things aren't equal, they'll be needing to pencil in a higher debt interest bill in future years. And possibly lower growth, since growth in this country is also pretty disappointing at the moment. That's true. Typically, if you assume slower growth in the short term, the OBR will assume that that gives you space for higher growth over the longer term. So by the time you get to the fourth, fifth year of their forecast, that probably makes less difference than the direct increase in the cost of borrowing.

4:24When Rachel Reeves set out her first set of fiscal rules, she left a bit of wiggle room in that against one of the targets of the difference between revenue and spending to simplify somewhat. And as you say, we will discover how much of that has been eaten up when the OBR gets to do the forecast. But let's assume it's all been eaten up. Well, the issue there, I think, is that the OBR in normal circumstances is producing a forecast twice a year to accompany the budget and the spring statement. And there's always going to be volatility and noise as you move from one forecast to the next. Because, for example, financial markets move.

4:58There may also be judgments. You may get new news about how quickly particular tax receipts are coming in and out. So I think there's a genuine question as to do you really want to be fine-tuning the tax rates you're expecting to impose on people, the amount of money you want to devote to public services, when these things are going to bob around to some degree from one forecast to the next. But if markets are saying you've bust your rules, that's a problem, and you don't react as a government, you could find yourself in the position where investors punish you even more. I think there's an interesting question as to how fixated investors are on particular fiscal rules.

5:35I think investors will realise, like the rest of us, that governments over the last 10 to 15 years have embraced and abandoned more fiscal rules than Henry VIII did wives. So the idea of getting too invested in a particular precise target is dangerous. And obviously, if as a Chancellor or as a Prime Minister, you say we're absolutely committed to ensuring that this is met at every given moment, you're setting yourself a pretty hard task. So do we ditch them? Do you ditch the rules? I think rules are still helpful. You want governments to give people a sense of what they think good management of the public finances looks like.

6:11But one way, for example, to approach it would be to say, look, we're now moving to a world in which we're going to essentially set out three year spending plans every couple of years in what are called spending reviews. So maybe those are the points at which you want to say, look, we're now making some serious concrete decisions over the next couple of years. How does our revenue look? How do the fiscal targets look on that basis? Rather than twiddling them in the intervening periods. We are getting a new OBR assessment on March the 26th, I think. We are getting a new spending review, the three-year plans for individual departments.

6:53I think it'll be June the 10th or there or thereabouts, something like that. Now, I was talking to a very senior member of the government yesterday who said that if at the end of March, the OBR says that 10 billion of headroom of cushion in the public finances has gone negative. This person was saying the Chancellor will have to tell the markets that there will be spending cuts. They may not announce them, but they will have to be spending cuts. And it will accelerate moves to reform the welfare budget and accelerate moves to change what are often regarded as the incentives for people to go on to sickness and disability benefit rather than if they lose their jobs to stay on, you know, essentially unemployment support.

7:51because this is a government, a bit like the last government, that seems to believe, and I don't know in the end how much evidence there is for this, that there are some people who are saying they're disabled who could work but choose not to because they're better off if they basically sign themselves off work almost permanently because they're classified as disabled. Now, that's a really, let's be clear, that whole welfare reform, it's a very radical welfare reform. It'll be very controversial. But the question for me is, should all those moves to reform be accelerated simply because at the end of March, it looks as though there's a bit less money than people thought there was?

8:30Well, I think welfare reform on that scheme and those sorts of issues, successive secretaries of state for work and pensions have wrestled with those for years and how you design the system. You want to get that right and you want to get it right for the medium term. So the idea of sort of speeding all that up, if the idea was to speed it up. do it in a way that you're less confident is going to work just because financial market expectations of the inflationary impact of Donald Trump's presidency have moved as they might move in either direction between the March forecast and one in the autumn doesn't seem to me to be a terribly sensible idea.

9:04You want to clearly be looking at welfare reform but it's not just for Christmas, that's a medium-term policy and I think you want to be thinking about whether that's a sensible thing to do above and beyond whether there's a particular need to meet the rules in March, which the story may be different in October. What would you do if you were the Chancellor now? What would you do if you were Rachel Reeves in her office today, you know, to calm things down? Well, given my other current roles at the moment, my days of giving clear policy advice, unfortunately, are temporarily suspended at the moment.

9:37A little snippet, just there's something in the boundaries of which you can tell us? Well, I think one thing which she's already indicated that she wants to do soon is obviously to give more detail and more substance to what the government's plans are for economic growth. So there was quite a bit of criticism, obviously, around the initial budget that that wasn't doing much in that department. But essentially, that was a set of decisions around ensuring you had or deciding how you were going to come up with the money to pay for public services and the inheritance on that score. The broader issues around economic growth.

10:09So John Kingman, former Treasury official sort of did a very nice list of sort of the sorts of things that you might be thinking about. So were you in favour of another runway at Heathrow, for example, which is one of the things John is in favour of? Well, that decision has been going around the airport waiting to land for quite some decades. So I think his point is that you might get on with it. But fortunately, airport design is not my speciality. But progress on that sort of thing, decisions on, you know, some areas of tax reform, decisions on what you want to do about education. Do you want to be thinking seriously about sort of the development prospects of the what's called the Oxford to Cambridge corridor?

10:50Another thing that he mentioned there. We've talked about that on this programme a number of times. I mean, I sort of take the view that although, you know, there are issues of a sort of fairness sort since, you know, one of the things that Steph and I are obsessed with is how do you raise living standards in poorer parts of the country? And let's be absolutely clear, Oxford and Cambridge is not one of the more deprived parts of the country. It is also when it comes to particularly technology, both are centres of excellence. They've got large numbers of growing businesses around them. If you join them up, it might have massive economic benefit in terms of generating bigger businesses faster.

11:30It sounds like a no brainer. It sounds like the sort of thing that any sensible government that wants to generate growth ought to sign up for making it easier to get housing and transport developments between Oxford and Cambridge. But rather like putting additional runways in Heathrow, there are some people who are more affected by the costs and the downside of that than others. So taking the decisions to press ahead with that when you have winners and losers is not straightforward. And the losers are the NIMBYs who don't want new housing developments near their lovely cottages. That might be slight oversimplification, but that would be one of the factors.

12:03In terms of the distribution of welfare and economic well-being around the country, that, of course, was one of John's other proposals, is that you should make further progress in decentralising some tax and spending powers to local authorities, to the big authorities, and that requires the Treasury to loosen its grip a little bit over some of those decisions. That was one of the points he made, which is not just about overall productivity and national well-being, but distributing equitably across the country. At the moment, you're chair of the UK Statistics Authority. Perhaps the most important bit of data when it comes to whether it's Bank of England making decisions on interest rates, whether it's actually the Chancellor deciding what kind of stimulus is needed in a budget and where is data on the level of unemployment.

13:04And one of the things that has been really shocking is the extent to which the Office of National Statistics has said it's putting out data that it doesn't have any confidence in. I suppose what I want to know from you is what went wrong and how do we fix this? Because, you know, you get an organisation like the Resolution Foundation saying that we've all been operating on the basis that there are, I think it said there were almost a million people who are currently classified as not available for work, who, in its view, are in work. So that's a massive difference. Well, I think the labour market statistics are, as you say, an absolutely key input into the Bank of England's decision.

13:49There are other ones that matter as well, obviously, the national accounts, GDP measures of that sort. The challenge that's confronted the ONS in constructing the labour market statistics is in part part of a more general issue, which is that it's become increasingly difficult and expensive to get the quality of outputs that you want from surveying households. So this is not unique to the UK. You've seen different countries find it harder to get the response rates that they want from questions. It's not just about the labour market. It could be about wealth and assets and other things like that.

14:24What's going on? Why was it that 10 years ago people would answer these questions and now they won't? This process had been in decline for some time and then accelerated over the coming out of the pandemic. So you have issues like, are people basically as happy to share personal information as they used to be? Are they happy to let interviewers physically into their homes for, you know, 45 minutes and a cup of tea to answer a series of questions? Is it harder to get access to secure gated accommodation? So actually reaching the people that you're supposed to ask. Is there a more general sense of lack of trust in public institutions and government that affects people's willingness to do that?

15:07Now, I think that has been an issue across lots of countries. It's been more of a difficulty here than in others. But why? Why is it more of a difficulty here than in other countries? A couple of reasons. One relative to some countries is that answering the questions is not compulsory. So in Australia, you are required to answer questions of this sort. in Singapore you can be fined up to£600 for not essentially filling in the equivalent of the Labour Force survey. What's the position in America? Good question. I don't know. I don't think that's a compulsory... But there doesn't appear to be a crisis in America, is my point.

15:45Well, they also spend quite a lot more on their equivalent of the Labour Force survey, I think. I think the two, in addition to the compulsory, the other issue that I think has been a challenge for the Labour Force Service. It's a relatively long survey. It takes a relatively long time to complete it. Over the years, I think, you know, understandably governments have wanted to collect more information on different things and that's got larger. So the ONS recognised this several years ago and put in train a move to something which they call the Transformed Labour Force Survey, which is essentially online first rather than going to people's homes first.

16:23and the aim then also is to have something... Sorry, that the individuals will simply be sent, what, a digital form? Is that what you're saying? That's right. You can then follow that up physically if you need to. So that is already being tested at the moment. There are challenges with that, again, because, for example, if you have a relatively long online survey, it's perhaps more likely that people will sort of give up on it halfway through than if they've got somebody sitting in front of them with a mug of tea in their hand. so the ONS is at the moment sort of trialling different options of if you have a shorter one of these are there particular questions that make people go gulp I'm going to save this file now and not carry on with it and so those options are being explored at the moment and we're going to be in a period for some time of essentially dual running this new model which looks like being the long-term solution but continuing to operate the labour force survey.

17:19Now there were particular challenges as the response rate, the proportion of people who you ask who answer the thing, declined and declined sharply after the COVID. It didn't recover as quickly after the pandemic. The ONS has put more resources into that. It's now bringing the response rate for the existing system higher up. But one of the challenges in both moving to the new system and in the existing one is that if you're trying to make improvements, the way these surveys work is that you start off by asking people some questions and then you go back in each of the subsequent four quarters to re-ask those questions.

17:58So it's what's called a wave structure in the panel. So if you're testing or designing some new feature into this or a simplification, then you have to let it work its way through that process, which is why the ONS has warned people that for the time being, you end up with a sort of relatively volatile picture coming out of the Labour Force survey and that you ought to, at the same time, look at other sources of labour market information, particularly the real-time information that you get out of HMRC, Revenue and Customs, because of the PAYE system, and also workforce jobs. So the bank and the OBR and others at the moment will be essentially looking at the range of numbers here to try to generate a full picture.

18:39And that's going to be true for a while. I'm getting really exasperated here because what blows my mind in all this is that we're asking people to do a survey. Like surely there's enough information out there captured from all the different kind of public bodies to do with, you know, tax revenues, benefit claiming, council tax, everything else. Why do we need to ring people or go around the house and ask them questions? I mean, it really can't be challenging to design some kind of AI model that essentially processes all of the data that Steph is talking about in real time, every second of the day, such that you, I mean, as she says, the notion that you want somebody going around with a mug of tea or even being sent a form is just absurd.

19:27Well, I think Steph identifies one of the big trends that you've seen across a whole series of statistics is making more use of administrative data where you have it. So, as you say, pulling together, and this is a big debate, for example, about how you might think about doing censuses in the future. Can you get enough information from combining health records, social security records, etc. together? Pulling all of that information together and sharing it is not quite as straightforward as shoving it into chat GPT, I'm afraid. But there's clearly scope to use those sorts of techniques. I mean, one challenge coming back to the distribution around the country is that there are some parts of the population that we do and should care about that are less tied into the administrative data.

20:17And so one of the reasons for needing survey activity is to ensure that there are particular types of people in the country that you don't lose or, you know, basically become invisible to the data that you're trying to collect. So I think for the foreseeable future, you are seeing partly because of the pressure on response rates, a greater desire to make as much use of the sort of administrative data that Steph's described there as you can. But the idea that you can basically abandon surveys altogether, I think is unlikely for the foreseeable future. Surely, though, like the private sector have a wealth of data on us, which is easily collected.

20:55And in real time, as you say, I disagree with you in the sense of actually be quite hard to do this. I really don't think it would. It just needs a platform in which people, you know, different, whether it's the food bank. and when they get a referral, could they not just type it into the system? They do it into their own spreadsheet. Can that just not be a centralised thing where everything just goes into the same pot and some AI thing just works it out and gives you a summary of every single person? It doesn't feel like if a retailer can do that, if a retailer can know that next month I'm probably going to need a pair of tights, surely the government are going to be able to work it out from everything else.

21:36that's one question not asked in the labour force survey as far as i'm aware well i think one one set of concerns around that and pulling this together is around privacy and ethics of this are you able to ensure that if you are pulling in data that you're linking it to other sources of data from other people that's you're doing that in a way which will start to identify individuals and the processes and the safeguards that you would need to put in place to ensure that that was done safely. It wasn't compromising people's security, the security of their financial information, the security of their health information is not to be underestimated.

22:14And indeed, it's one of the reasons why it has been so challenging even to get data shared from within government, let alone from the people deciding whether you need to buy new tights, because different departments are anxious that if they share their data with other departments, what happens if there is a breach? What happens if the data are used in ways that they think are inappropriate? So that is really not to be underestimated as a challenge. On AI, there are areas in which it is being explored and used. One of the challenges with moving to the new online TLFS, an existing challenge, is people answering the question, what occupation are you, what industry are you in, which people will have a different sense of unless you give them an exhaustive list of options.

23:00And one area, I think, where AI is being explored is if you look at all the answers that people give to those, can you get that to sort things out into a more rigorous assessment of what people's occupations are, what industries are, which is important to the granularity and the detail you're trying to get out of the information of the labour market, for example. Well, there's plenty more stats to talk about. So sit tight and we'll have a quick break.

23:30I mean, I think given we have dwelled on this quite a long time, it is just important to explain why it matters when data on how many people are employed or not employed or sign themselves off, why this matters. Because the amount of so-called slack in the labour market, the amount of unused capacity in the economy is an absolutely central judgment when it comes to where the Bank of England sets interest rates. And it has been a source of considerable concern to me, particularly since this is also data they feed into their economic models, that we cannot rely on the information that's been coming out from the Office of National Statistics on all of that.

24:19I mean, at a time when there is, understandably, not a lot of confidence that the Bank of England has been setting interest rates at the right level. I mean, one of the things that I'm slightly concerned about, we just today, as we're recording, got the latest inflation data. That actually does show that when it comes to core inflation and services inflation, that now it has come down rather more than people expected. You know, the economy is patently bumping along, stagnating. You know, I'm one of those people who thinks that the Bank of England should have cut interest rates at its last meeting.

24:56We're now going to have to wait till February to see whether it's going to cut interest rates. So against a backdrop where many of us think the judgment of the Bank of England in terms of where to set interest rates is not what it should be. The lack of robustness in the data is another source of anxiety. It's certainly a source of frustration. I spent 10 years, as you know, being in charge of the OBR's forecasts over five years, of which the question of how much spare capacity is there, how much is that, how is that going to evolve, was a key judgment in every one of the 20 or so forecasts that I had to preside over.

25:30and one rule that we always made for ourselves is that we would never write down now is a uniquely difficult time to do a forecast because there are always reasons why it's difficult to do a forecast and obviously the ones that you can tackle or that you can improve it's always worth looking at that and trying to do that as best you can but it is a fact of life that your understanding of where the economy has been as well as where it is and as where it is going to be evolves over time so if If you think back to when we were in short pants economic reporters, I remember Kenneth Clark's first budget, 1993, going to the Treasury Select Committee.

26:11At that stage, it looked as though the UK economy had shrunk by 4 % during the course of the 1990s recession. It ended in a double dip. By the time I went to the OBR, look back at what the numbers were saying, it was down to 2.5 % and there was no double dip. So it's important as for newspaper reporters as well as for other people to recognise that when you're seeing the next month's GDP numbers, less so with inflation, obviously, because it's not revised. These are the early drafts of economic history that can end up changing quite a bit. And one needs to recognise that. We talked about that only recently, the fact that we've set so much store by data that comes out on any particular day.

26:55And it often takes years for us to be told, actually, the economy, you know, we all made this huge thing as journalists that the economy was in recession. And then, you know, a few years later, you discover the economy wasn't in recession. It doesn't inspire confidence, does it? Well, it's a feature and not a bug. So take the national accounts and the GDP story. you get more information on that months, years after the first stab at it. So the most obvious example is that you takes quite some time to get information from self-assessed income tax returns, which tell you a lot more about the level and distribution of income.