Public Administration and Constitutional Affairs Committee
Oral evidence: Governance of Statistics, HC 1820
Tuesday 26 March 2019
Ordered by the House of Commons to be published on 26 March 2019.
Members present: Sir Bernard Jenkin (Chair); Ronnie Cowan; Kelvin Hopkins; Dr Rupa Huq; Mr David Jones; Eleanor Smith.
Questions 285 - 343
Witnesses
I: Clare Lombardelli, Director General, Chief Economic Adviser, HM Treasury; Dr Ben Broadbent, Deputy Governor, Bank of England; and Robert Chote, Chairman, Office for Budget Responsibility.
Examination of witnesses
Witnesses: Clare Lombardelli, Dr Ben Broadbent and Robert Chote.
Q285 Chair: Can I welcome our second panel of witnesses this morning, which is about the governance of UK statistics? I am very grateful for you all attending. Please could you identify yourselves for the record?
Dr Broadbent: I am Ben Broadbent, Deputy Governor at the Bank of England.
Robert Chote: Robert Chote from the Office for Budget Responsibility.
Clare Lombardelli: Clare Lombardelli from HM Treasury.
Q286 Chair: I am extremely grateful for you all coming. Can I start with a very broad question? How effectively do you think UKSA is performing its role in managing ONS as a producer of statistics?
Dr Broadbent: As a very broad answer to a very broad question, “well” is the answer I would give. As you know, we have this task of delivering on a big review by my predecessor, Charlie Bean, but I think the early signs from that, which is a very important project, have been very encouraging. We are, at the Bank, very heavy users of official statistics and we are pleased with the progress that has been made since that review was delivered.
Robert Chote: I agree with that. There is a strong leadership team across both UKSA and ONS at the moment. As Ben said, the Bean Review had a list of outcomes that they thought were a scope for improvement, and I think you get that sense across the institution. There have been other achievements that do not leap above the parapet immediately, but I think they have had a reasonably successful upgrading of their IT system and there is many an institution that cannot say that, which has been quite helpful. I think you get the sense of the overall trust in the statistics system. How much you put that down to UKSA specifically versus the ONS versus the interaction between the two is obviously up for debate, but in general it is working reasonably well at the moment, albeit as you have discussed with other witnesses beforehand, in an odd structure that you probably would not design this way if you were starting with a blank sheet of paper.
Chair: We will come to that. Clare Lombardelli.
Clare Lombardelli: Yes, I agree. It is a tough job. Demands on them are high and it is hugely important, but they do a good job and I agree with what Robert said about a lot of the progress we have seen recently, particularly around some of the implementation of the Bean Review is hugely valuable, even if it is not always stuff that makes the news all the time.
Q287 Kelvin Hopkins: How effectively is UKSA performing its role as an independent regulator of official statistics?
Robert Chote: I think if you were starting with a blank sheet of paper, for exactly the same reason that the odd structure of the BBC Trust was a weird way of trying to run that organisation, combining the two roles together is not the way you would design this if you were starting. That said, I think there was some time ago a recognition of that sort of tension and there was an attempt to sensibly separate out as far as was possible within the legislative structure that they had been given those functions. That is something that has moved further more recently in the wake of the Bean Review with the Office for Statistics Regulation and so on.
There is an issue that Charlie Bean picked up in his review about the degree to which the focus of the regulatory activity has on the quality of the statistics, versus what one might think of the broader quality of the process and perhaps there was more understandably on process earlier on and less on the focus of is this delivering the numbers that you need to understand as best you can what is going on in the real world? If you look at, for example, the length of time it has taken to get the construction data re-designated as a national statistic, getting that over the line, which only happened two or three weeks ago, required you to do things like reviewing HR structures and accessibility issues, which are broader ONS issues, but not necessarily the way you would focus it if your primary interest was dealing with a problem with a set of deflators for a particular measure of economic activity. There is scope for improvement, but within the structure that they have to operate in it is working reasonably well.
Dr Broadbent: Yes, I agree with that. Perhaps the most high-profile difficulty the ONS has had over the last few years has concerned the RPI. That is not a simple matter, how to change it, when to change it, because it is connected with so many other contracts within the economy. Perhaps the thing that people might use most often as a criticism of UKSA in that respect, it is not the easiest thing to have dealt with. As Robert says, if you designed this thing you would not start from here, but some of the criticism of the governance has been misplaced, I think, certainly to the extent that it has focused on the RPI issue, which is fairly complicated.
Chair: We will come to the RPI issue a bit later. Do you have anything to add to the original question?
Clare Lombardelli: The only thing I would add is that movements have been heading in the direction of strengthening the governance recently, certainly post the Bean Review. It is quite early to take a judgment on how that is going, but that has definitely been the direction of travel that we have seen.
Q288 Kelvin Hopkins: The criticism levelled at us is that it is too often slow to tackle issues with statistics. Do you agree? If so, why do you think this is the case?
Dr Broadbent: It depends which issues you mean. Again, I will go back to Charlie’s review for an example, which is does the structure of the statistics and the focus of them keep pace with the change in the shape of the economy? One thing that he identified very clearly was that you probably need it to develop new ways of measuring, for example, digital activity, develop new sources to measure existing activity more accurately. There is a balance to be struck always for any statistical agency between how many resources it devotes to improving the quality of an existing series and how much it devotes to developing new measures of understanding of the economy, which is always changing shape.
I guess it could be said that they were slow to do that. My view would be it is often easier if someone on the outside points that out. As Clare says, this division with the OSR, it has not existed for very long, so as it happens it was the Bean Review that moved that forward. That has probably been the most important shift of resources. It is a moving target, measuring the economy and understanding how it evolves and it is not going to be finished, but I would identify that particular thing as the area where maybe they were slower than they might have been. It is now improving.
Clare Lombardelli: One thing I would add, it can maybe seem time-consuming and frustrating from the outside, but it does take time to develop new uses of statistics, new methodologies for the way in which we are going to measure things. It is right that you take time to think through how that is going to work, make sure you are completely comfortable with it. It does take a bit of thought and time, but it is right to do that. If we are developing new statistical methodologies we cannot really rush it. It is slow and methodical work.
Robert Chote: One thing to add on that would be there is the line of argument that says, “Look, if UKSA had been proactive and ahead of the curve then you would not have needed the Bean Review in the first place”. Now, obviously there are some things that you would have liked UKSA to have picked up that the Bean Review then has to, but I think you should not underestimate the value, as Ben was saying, of having periodically an external view of an organisation on top of whatever internal quasi-non-executive governance structure you have. For the independent fiscal institutions like the OBR, we have a well set up process of international peer and expert review. Every few years each institution will expect to have an outside review and that I think is a useful part of the process.
Clearly the ONS has had over the years quite a lot of external reviews in lots of different areas and that suggests that it is too much of a good thing and that more of that could be captured on the way, but I think there is a value to having these external things. Not least in the case of the Bean Review, it is that it is much easier for an outside body or reviewer to say, “Look, the resources that this institution has are not commensurate with the job it is either supposed to be doing now or that you want it to do” and therefore making the case for greater resources, which came in the wake of the Bean Review and that is obviously helping implement that.
The same thing happened when we were last externally reviewed. That was an opportunity for a broader case to be made, not just self-interestedly by ourselves and our non-executives, that there was an argument that said that we needed more resources to do the job we were supposed to be doing as well as we should be and some additional things as well. I do not think the presence of the Bean Review per se is an illustration or an indication that the governance structures failed.
Q289 Kelvin Hopkins: You anticipated my final question, which is about what changes, if any, you would propose to improve UKSA’s effectiveness as a regulator? My view anyway is there has been a general resistance to having external regulators and governance. Indeed, they would like external regulation themselves, so is there not a case for a separate body that is not effectively part of the statistical institution in itself?
Dr Broadbent: Arguably it is separate. As Robert said it is not ideal. There are two different bits of it and there is a bit that is meant to regulate and a bit that executes the measurement and delivery of statistics. I would point out this is a slightly different aspect of the regulation of statistics directed externally rather than internally, but I would point out that I am quite impressed by the way that UKSA has defended the correct use of statistics externally, which is also part of its job. It is not simply a matter of making sure that the production is free of interference, which is one of our primary tasks as regulator and that the quality is good and so forth, but that when they are used externally they are used correctly and that is done without fear and quite effectively.
Q290 Kelvin Hopkins: It is interesting to hear the point of view about UKSA from yourselves about regulation. There are others who have been slightly more critical, but I look forward with interest to what changes might happen in the future.
Robert Chote: I think it is a question that from where you are starting now, do you get enough of a gain from the process of ripping up the legislation and separating these functions more formally? Obviously the difficulty is that it is only when something goes wrong in the system that that crystallises, and it is harder to make that case when things are going reasonably smoothly. There is a spectrum of the degree of internal review and regulation you could have. I think Charlie talked about it. You have the IMF, with independent evaluation offices nested within the institution. That is pretty much the model the Bank has as well, so you could move further in that sort of direction if you wanted to within, I presume, the same legislative structure you have at the moment.
If you wanted to change the legislation obviously there is a question for Parliament about whether you wanted to vote. When the legislation creating us was there, the Government took that as the opportunity to bolt on a whole lot of things they wanted to do to the NAO to the same Bill and they took that as an opportunity to make time for it, effectively speaking. The system is not so obviously broken at the moment that you can see that impetus, but it is not how you would start if you were designing it with a blank sheet of paper.
Q291 Kelvin Hopkins: Perhaps the mild criticisms in reasonably recent times are enough to wake them up and perhaps put a bit more energy into regulation.
Robert Chote: Yes, and obviously the issue with any regulatory system is you do not see the low-scale successes. The difficulty comes when there is a difficult situation that comes along and then is that a failure of the situation or is it one of those instances where you can expect to pluck off most of the low-hanging fruit, but then there is an occasion where that gets more difficult? Whether that would have been very different if you had had a more separated organisation in a whole variety of public policy areas, mistakes can be made, and co-ordination failures can happen when things are nested in the same organisation or when they are in separate ones.
Q292 Chair: We picked up quite a lot of low confidence in UKSA as a regulator. What do you think, Clare Lombardelli, it could do to address that lack of confidence?
Clare Lombardelli: I am quite interested to hear you think it has picked up that low level of confidence. It is understandable that people who are users of statistics would have views on that. It is a pretty tough job to produce statistics to a point at which no one is ever going to criticise you for them.
Q293 Chair: So basically it is okay?
Clare Lombardelli: I would not say it is okay, but they are accountable directly to Parliament. They are accountable to you and Committees such as yours and others have put a lot of scrutiny on them and I think the question is how they will respond to that.
Robert Chote: One distinction that is worth making is that there is this regulatory function as distinct with ONS and its role over the broader statistical system. There is clearly more of a challenge establishing confidence in what is coming out of Government Departments as distinct from what is coming out from ONS, and UKSA has less of a role directly in responsibility for that. It is harder for them to help the statistical profession nested in individual Departments’ set strategies and work programmes than it is for ONS as well. There is a bit of a distinction there. Whether the sense of lack of confidence that you hear from people is reflecting their lack of confidence in its scrutiny of ONS or the fact that they just have less of a grip on the remainder of the GSS might be part of the issue.
Q294 Chair: What could it do better within the framework?
Robert Chote: On the GSS side, Charlie was making in his review the point that having the heads of professions within Government Departments, ensuring that they are of sufficient seniority, that they carry weight in the decisions and the discussions with Ministers and other officials—
Q295 Chair: So that would be the job of the regulator to oversee that?
Robert Chote: I do not know formally whether it has the right to do that, but I think there is a very important role if you are the head of the statistical profession and as the National Statistician you are obviously embodying that role in a broad sense as well as in a formal legal sense. What network of support, what assistance are you providing to heads of profession in Departments, who obviously have a conflict of interest between their responsibility to produce data that is appropriate for the public good and also as civil servants to serve their Ministers and their Department?
You are never going to be able to wish away that potential conflict, but I think it would be a very important role if you were running UKSA—and I know that David sees the importance of this—to be providing support, practical, as far as you can, moral where you need to, being responsive. If heads of profession or other statistical professions in Departments feel that they are being asked to do inappropriate things that you go in there, sometimes publicly, maybe more often privately, and make it clear that that will not stand.
Q296 Chair: That is an interesting suggestion.
Dr Broadbent: Do you mind if I second that? It will naturally become an area where UKSA will have to become more involved, not least because the ONS, as again recommended by Bean, is beginning to use statistics produced by other bits of Government in ONS releases and it announced something measuring turnover or activity from say tax data. Obviously it is crucial that those are as high quality as they can be, and because once the ONS produces them and releases them they will probably move financial markets, that there is a very clear process for not pre-releasing, not allowing anyone in the Department to see them before they are released and so on and so forth. This will naturally be an area of Government statistics where I think UKSA is drawn in and should have strong oversight of those processes.
Q297 Mr David Jones: Could we discuss further the unusual structure of UKSA? Is it possible for UKSA to achieve maximum effectiveness as both the regulator and the producer of statistics or to what extent would you say it is hampered by this unusual dual function?
Robert Chote: As we were saying earlier, I would not have designed it this way if we were starting with it and it is not the only institution that has had difficulty. The BBC Trust is the other example of something where the blurring of the responsibility for regulation and for championing gets stuck together. That is quite a good example of where it had proved quite difficult. Where you have problems of different sorts within the BBC, the lines over which the sorts of areas that the trust took responsibility for, they were having to look more and more at the things that you would imagine that a board was responsible for. That is not an ideal world to get into.
As we have said, already prior to the Bean Review and more subsequently to that, there has been a determined effort within that overarching legal structure to get this as separated and functioning. You have the Office for Statistics Regulation, you have people with clearly defined responsibilities there. You could go further, as I say, with an Independent Evaluation Office internal model. You could go for something that is more explicitly separated, the model of having a more beefed-up operation inside the organisation on the IMF. I presume the Bank model is that it is only by having them in there that they understand how it all works. You have to be inside the fence to be able to see what is right and what is wrong. The OBR model is more that you have a completely external organisation, but you make sure that you employ quite a few people who have, to mix the metaphors, either made the sausages or buried the bodies on the other side and therefore you do not get the wool pulled over your eyes in that way.
There are models there. At the moment, as I say, the system does not look broken enough to suggest there is an urgent need to come to this with a legislative solution that splits it apart, but as I say, I would not have designed it this way.
Q298 Mr David Jones: No, but do you think it would have been more effective if this dual function had not been imposed upon you?
Robert Chote: It would have been clearer for people to understand. In this room are the majority of people who could tell you that the ONS is the executive office of UKSA and this is not widely understood outside. That has not stopped there being a generalised improvement in confidence in the notion that if something is badged as a national statistic that is probably a good thing and that there is probably a process behind that. I am sure that would have been more obvious to outsiders and cleaner in that sense. Whether it would have been more effective depends in part on this point about whether the separation loses you something in terms of the ability of the regulatory body to know exactly what is going on and to be engaged in that on a timely basis. The problem is we cannot rerun history and see how it would work that way.
Dr Broadbent: It is pretty early days of the post being structured and I would suspect—I do not know, because I do not know what precise criticisms you have had—that some of them are criticisms of particular statistics of whether they are accurate or how often they are revised or issues concerning the RPI, which are not necessarily anything to do with governance but are to do with the ability to produce good statistics, the quality of them. Do you have the right surveys, do you have the right techniques, and do you have the good people to understand them? That is not necessarily a governance issue.
Chair: Really?
Dr Broadbent: Not necessarily, no.
Q299 Chair: If you have the wrong people or they are doing the wrong job, or they do not understand their responsibilities?
Dr Broadbent: That maybe, that there are certain processes that you might be able to improve, but that is an executive decision of how to produce better quality statistics. That is the job of the management of the ONS. It is not necessarily a failure of governance. Quite often I suspect, and maybe this is unfair, but sometimes the criticism of statistics presupposes some perfect world where we could measure these things completely accurately.
Chair: Yes, I understand.
Dr Broadbent: That is not true, so I am not saying there is not a valid criticism. I am only saying that this separation within UKSA that Robert referred to is pretty recent and I suspect that the criticisms relate to things that certainly predated and may or may not be improvable and may or may not have anything to do with governance. It may be to do with construction data. It may be the governance of that was wrong and maybe it was not. Maybe it should have taken place sooner, that they realised that the early construction data was flawed and there were things you could to improve things.
Q300 Chair: That is the sort of criticisms that people make.
Dr Broadbent: Yes, but it is not clear to me that the answer to that is to tear up the structure.
Q301 Mr David Jones: Mr Chote mentioned that arguably there are some benefits of having this joint structure. Dr Broadbent, do you see any other benefits?
Dr Broadbent: As you said, one of them is that we happen to be starting from there. I think Robert is right that presumably good regulation involves someone who is independent and disinterested in whether the management is feeling good or the wellbeing of the thing they are overseeing and interested in the output, but equally the regulator has to be informed enough, as Robert says, to understand what the issues are.
As he says, we at the Bank now have an Independent Evaluation Office modelled on the structure of the IMF. That office reports directly to the Chair of Court, not to any executive in the Bank, and writes reviews, decides what it is going to look at, looks at them with the help of external experts. It is pretty important that the IEO in our case has the ability to spend time with and talk directly to people in the Bank to understand what they are doing and to understand whether they may be able to do it better. It would be harder to do that if you were wholly separate. There is a balance to be struck. As I said, this new structure seeks to strike some sort of balance. There is this separate OSR and I think it is probably too early to say whether it is could work or not, frankly. It has not been around for long enough.
Clare Lombardelli: I agree with that. As I said, it is a fairly new structure. It followed from one of the recommendations of the Bean Review. It is being implemented, but I think given the time it takes for these systems to bed in and show their impact, it is a bit soon to make a judgment on the Office for Statistics Regulation one way or the other. There is clearly, as Robert and Ben have said, pros and cons about the degree of separation you have between the producer and the regulator function. These are in the detail quite technical issues and so you do need to have a level of understanding and appreciation of how you go about doing this.
Q302 Dr Rupa Huq: Returning to the RPI fiasco, the House of Lords Economic Affairs Committee report delineates a litany of failures. In the written evidence it was estimated that it was costing £1 billion a year to not fix this flawed—that was Chris Giles who said that in his evidence. Again, it comes back to the failures of the regulator. The Committee concluded that UKSA could be accused of failing its statutory duties to promote and safeguard the quality of official statistics. Would you agree with that assessment?
Dr Broadbent: No, I do not think I would agree with it. It is a highly complicated issue because the RPI is in statute, used for all sorts of indexation purposes and has been for many years and you cannot just get rid of it very easily. I do not believe it is as much as £1 billion a year. What happened essentially was that there was a particular change in the method of collection, the bit of the price index, closed prices, which amplified very much an existing flaw in the RPI. It is well-known and has been known, frankly, for decades. That produced a jump in the rate of inflation of the RPI relative to better price indices. It is not an ongoing cost. It pushed up the value of index-linked gilts at that point. There was at that time essentially a transfer from the taxpayer to the holders of these instruments.
As you know, the complication is difficulties with the measurement of the RPI and the flaws in the RPI. People looking at them looked at them for many years. UKSA did a consultation in 2013, after which it concluded it was not going to make wholesale changes. That decision was endorsed by the Johnson Review. They were in a very difficult position. The nub of the issue was if it had gone and arbitrarily and without any further consultation said it was going to correct all the flaws that it sees in the RPI, that would have had very significant implications, not just for holders of gilts, but many other contracts linked to the RPI. The issue is much bigger therefore than a simple statistical failing of a particular statistic. It was bound to and is bound to come back to the Government.
As you say, the House of Lords Economic Affairs Committee has now looked into this. I think it has drawn broadly the right conclusions and I hope somehow the nettle is grasped and a way to co-ordinate a change is found. I would not say it is simply a failing of the statistical regulation. The issue is much bigger than that because it has implications for public finances, for all sorts of other contracts, for student loans and so forth and it is not an easy decision to take.
Q303 Chair: You are making it sound as though the UKSA has much wider responsibilities than it has to worry about the effect on gilt holders.
Dr Broadbent: Yes. You might say it should not have had any concern with that.
Q304 Chair: There is nothing in the legislation to suggest it has any other responsibilities except overseeing the proper production of statistics. If it is a wrong statistic, does that not beg the question to whom is UKSA answerable to?
Dr Broadbent: I do not know. It pre-empted or anticipated certain difficulties, but as I say, it reached that conclusion, not to make wholesale changes to this in 2013, but rather to develop and champion better price indices like CPI and CPIH, which it has been doing. Of course, for our part, our inflation target is expressed not in terms of the RPI or any bit of it but the CPI.
Clare Lombardelli: It is a judgment for Parliament to take about its statutory duty. The Lords report came up with a number of recommendations to the Government, to UKSA. As Ben said, this is very complicated, with quite a lot of impact, so you would want to think them through. It is important to recognise that there are obviously public sector implications, but the vast majority of the uses in value terms of RPI are in the private sector. There are literally trillions of pounds of contracts linked to it. All of that needs to—
Q305 Chair: Yes, my mobile phone bill has just gone up by RPI, thank you very much.
Clare Lombardelli: I can believe it. If you think about pension recipients, that will probably be the biggest use. You have people on defined benefit, people who receive pensions that will be linked to RPI. I think Ben is right to say the use is extension. I would say rather than think about it in terms of the cost, there is a distributional consequence of this, that is what you would need to think through and why the Lords report is welcome progress in this.
Q306 Chair: Robert Chote, how happy are you that our public finances are managed on the basis of what the Lords report called index shopping? Things that it wants to go up more it links to RPI, like train fares and other costs that people have to pay, and then it links benefits like pensions and unemployment benefit to CPI, so they go up rather less. Do you think this is an example of effective regulation of statistics?
Robert Chote: That is unlikely to be a coincidence, that policy has responded to those differences in that sort of way. Obviously in a long-term situation it is evidently unsatisfactory that the ONS is publishing something that it believes to be a bad measure of what it is supposed to be doing.
Q307 Chair: But we can conclude that that is what it has allowed to happen?
Robert Chote: The strategy—my colleagues are closer to this than I am—has basically been to say, “We will provide other better measures of inflation. We will point out very publicly that this one is not a very good one and we will encourage public and private sector agents to use the better alternatives that we are producing”. I think one lesson of the experience—to take the example way back of RPIJ, which was trying to deal with some of the formula issues in RPI, and with CPIH—is that notwithstanding those statistics being available, not as many people maybe as it would have anticipated have followed the advice to use those instead of the other ones.
CPIH, which I think is the ONS’s preferred measure now, is not widely used. It is not something we even forecast because it is not material to the public finances. But as Ben says, you want to grasp this nettle and the problem with grasping nettles is that they sting. Can you move to a more satisfactory position, both for the management of the public finances and for the integrity of ONS producing numbers that it believes are good measures of things, but in a way that deals with all of the sorts of contractual difficulties and makes that as smooth an adjustment as possible? The fact that the House of Lords report in itself was market-moving is a concrete illustration of the complexities and the distributional consequences of even the suggestion that you are moving there.
Dr Broadbent: I have no doubt that if there had been no implications of this beyond simply people are interested in prices, and even if there had been particular taxes that tended to be linked to that—after all, the Government does not oblige themselves to index these things; they take a fresh look at the budget every year. It is not as if they have locked this in and prices that are worked out as RPI plus X, the X can change and routinely does.
Say there had only been that and there had not been this huge stock of index-linked gilts, there had not been pensions sitting there for years that are indexed to this particular index and there had not also been the requirement to produce the RPI in statute that UKSA would have either ditched it or just renamed it or said, “The RPI will now be what we call the CPIH”, that would have happened a long time ago. The process should have started, this very difficult co-ordination problem about how to move the whole thing to a better measure of prices, that complicated process, given all these dependencies. Maybe you would have reached that sooner if UKSA had done this, but I rather doubt it. The complicated problem would have remained. Ultimately it can only be the Government and Parliament that think through these very difficult questions because, as Clare says, there are distributional implications and potentially legal implications as well.
Q308 Dr Rupa Huq: How well and through what mechanisms does UKSA engage with you, as key stakeholders, to understand what you need?
Dr Broadbent: I talk regularly, more often to the ONS. The Bank has, as any heavy user would, contacts at all sorts of levels, usually technical contacts. We want to understand exactly how it constructs the CPI and which bits it measures and how and so forth. We want to understand, when it changes methodologies and the national accounts, exactly how it is being done. We have an important change coming this year in the Blue Book, something called double deflation, which is a welcome advance, but we need to understand exactly how it is doing it and where it is getting the information from. The contacts are extensive with the actual producers of those statistics, I would say, more than UKSA.
Robert Chote: Most of our interaction is with the whole variety of user groups that are focused on particular sets, if it is the national accounts user group or inflation ones or public finances and so on. There is in addition, following on from one of Charlie’s recommendations, a high-level stakeholder engagement group that meets, I think, twice a year. It is not tested yet because there has not been a problem that that has needed to feed back on. It is a conduit right at the very top to get stakeholder concerns to the top of UKSA and the top of ONS and for that to come back.
Charlie represents OBR himself on that group. That is useful, but obviously its true usefulness will be demonstrated as and when some serious issue comes along. Most of the interaction is, as I say, at a working level with people who are experts in those particular sets of numbers and the corresponding maybe officials at the Bank, maybe fewer officials at the OBR who deal on inflation or relativity or labour markets or whatever it may be.
Clare Lombardelli: It is exactly the same for the Treasury. We have multiple, working-level discussions going on about the particular statistics that Robert talked about. Those are all technical discussions about how those statistics are compiled and therefore how can they be interpreted and those sorts of things.
Q309 Eleanor Smith: We have received evidence to this inquiry describing the information of Tetrapartite Group as a, “Secretive and arguably biased towards excessive Treasury and Bank influence”. What is your response to that description?
Dr Broadbent: We have just described how we have multiple interactions and conversations with the ONS about technical things. This one happens to have a name, it is called Tetrapartite, but it is purely that. It is a technical group where we go along—I have never been; it is a relatively junior level. We ask the ONS, “Exactly how is this bit of index compiled, what are you going to do there, how do you measure air fares?” It has nothing to do with policy and I see absolutely no reason—it does not decide anything. It is information gathering, information exchange basically from the ONS to us about how these indices are compiled. I think there are, if not full minutes, then at least agendas and some description of what has gone on published by the ONS. But it is purely technical information, it does not have a policy function.
Robert Chote: The last one, for example, discussed quality adjustment for strawberry prices, whether you needed to take account of fact of were strawberries of different geographical origins comparable in terms of using them as generating an input into inflation. That gives you some sense of the high policy that goes on in this group.
Clare Lombardelli: The membership and the terms of reference are all published on ONS’s website if you want to see them.
Q310 Eleanor Smith: That is what we were going to ask. You obviously do not go to the meetings, so you cannot tell us what happens in that group. How effective is that group as a mechanism for you to engage with ONS?
Dr Broadbent: As I say, there are some things I go to or I will go and talk to my counterparts at the ONS. I see Jonathan Athow not often, but regularly. This is a technical group and such groups exist throughout Government, where people are talking to each other to understand things. That is what this is, and it means that when a Bank of England economist stands up in front of the Monetary Policy Committee and says, “Here is our near-term inflation forecast and we have done this and this” that person knows what he or she is talking about because they have the information from the ONS. That is the value of it. I do not need to be there, and I would not have the time to go to all such meetings that the Bank is engaged in. It is not that sort of meeting, it is not a policy meeting.
Q311 Eleanor Smith: Can I ask then, since it is not a secretive meeting, why are the agendas and the papers not published of that group?
Dr Broadbent: I think the papers are published.
Robert Chote: For those sorts of relatively junior level—our inflation economist goes to this and I think his deputy directors, at most, from Treasury, and whatever the equivalent would be at the Bank. If it was something that was generating policy and conclusions—most of the significant methodological discussions—again, you will both be closer to this than I am.
There are other groups that consider those more formal issues about the sorts of big questions about the difference between the way you calculate RPI and CPI. It is not going to be dealt with by this group. There are other more formal channels through which that takes place with a wider range of participants. This is, as I say, all experts but it is relatively junior group of people discussing those sorts of technical issues and keeping an eye on this. It is obviously very important for us as institutions. We both have to produce short to medium-term inflation forecasts, and knowing whether there is something strange going on in fruit prices is quite important at that level, but it is not Bilderberg.
Dr Broadbent: No, it is not. When I say these are published, forgive me, here are some of the things that I think are in the public domain: agenda items, thanks to FOI requests. I will give you a list of the last few: HPI progress update; airfares; pilot price collection; trade prices; methods; double-chain link update; business prices update; processing private rental data. These are not policy things, they are not deciding, “We are now going to switch wholly from this to that”. It is not that level or kind of meeting. It is purely technical, and it is purely information exchange.
Q312 Eleanor Smith: Is the group needed?
Dr Broadbent: Yes.
Clare Lombardelli: To be clear, it is absolutely needed so that the people working on the statistics can understand the way in which the ONS is using the statistics. It is definitely needed. It is very much a working-level meeting. There are multiple meetings that happen across different statistics, in the same sort of way—as I think Ben said, those meetings just will not have a name. But the key thing is it is not a decision-making body. When it comes to proposals around information, you have the advisory panel for consumer prices, there is a stakeholder one, there is a technical one. They put proposals to the National Statistician, but this is very much just a working level. It is the sensible business of Government you would expect to be going on to make sure everyone understands that is actually happening here.
Q313 Eleanor Smith: What relationship does this group have with the public-facing stakeholder groups and does it undermine them?
Robert Chote: Sorry, does it undermine them? No, it is dealing with a different set of issues. This is much more granular, technical, as the statisticians are turning the handle on this month by month. Are there strange things going on that we as users may say, “We are trying to come up with a forecast for this, but we cannot understand why petrol prices appear to have moved as they have done in the index when you can see what they have done on the high street”? It is those sorts of issues.
It is not something—and we would not do this anyway—where we would come and say, “We think this is a fundamentally flawed index and you ought to target something else”. It is much more down level. It is complementary to rather than undermining. It is part of the efficient plumbing of the system and ensuring that key users—for us, particularly, we have to make forecasts of this—can understand how the technology is working, the technology or the numbers is working.
Q314 Chair: I think the concern arises here that UKSA is not just part of Government, it is meant to be an independent body. If there is discussion going on between other parts of Government and UKSA, it should be transparent.
Dr Broadbent: This will be ONS, not UKSA. It is down the chain for them as well as down the chain for us.
Q315 Chair: Yes, but that comes back to the dual role of UKSA, does it not? You referred to other user groups where more policy discussions take place. The RPI/CPI User Group expressed its view that, “It is not clear why there is a need for the separate Tetrapartite Group, whose minutes are not published, especially when it is possible for these organisations to be represented on the stakeholder panel, as is currently the case for the Treasury and the Bank of England”.
Dr Broadbent: I think that confuses again what this thing is for. It is much lower level. It is not the same as the user group. The people who come from the Bank—as I say, I do not go, I do not have time—are not conveying to the producer of the statistics a view about how they should be produced. That is not the purpose of this.
Q316 Chair: Where do these policy discussions that Robert refers to take place?
Dr Broadbent: They would not take place at the Tetrapartite Group.
Chair: Where do they take place?
Dr Broadbent: I have never, I can tell you, had meetings with Athow.
Chair: I am not making personal accusations.
Dr Broadbent: No, I understand. I am saying, if you can generalise from my experience, I have never suggested to the ONS anything about the way it should collect price statistics. I was in the past on—I cannot remember what it was called now—a similar advisory group a long time ago, before I came to the Bank. It was national accounts and a variety of users who came, and that was a consultation. The ONS would ask them—and maybe the RPI group and the users’ group are the same, again, I do not go to those—“What do you think about how we should define this? How would it intersect with that?” They are technical, again, but they are more conceptual than what this Tetrapartite Group is.
Q317 Chair: What you have told us is how difficult it is to adjust the measure of RPI because it has such huge consequences.
Dr Broadbent: Yes.
Q318 Chair: We agree that is not an explicit responsibility expressed in the legislation, but somehow UKSA as a regulator has understood this. Who has conveyed this information to UKSA, so it has finished up being accused by a House of Lords Committee of failing in its statutory duties? Where did it gain this understanding from?
Dr Broadbent: On the particular question of the RPI?
Chair: Yes.
Dr Broadbent: That is a very distinct question. It is not really connected to what we were talking about a moment ago.
Chair: No, but it is a very interesting question. Somehow it has been persuaded not to do anything about RPI.
Dr Broadbent: I do not think it is true that it has not done anything.
Q319 Chair: That is a different question. We are six years after this problem really emerged big time and you say it has been very sensible not to do anything.
Dr Broadbent: No, I did not say that. I said I in some ways understand the approach it has taken.
Chair: Where have they gained this understanding from?
Dr Broadbent: I have also said the fact that we have had the RPI and it is flawed, and it is costly, it would be simplistic to say this is just all and entirely and only the fault of the regulator. It is a very difficult position to get out of. I do not know exactly how it arrived at the decisions it has made. It made those decisions some time ago not to make enormous reforms of the RPI, but instead to develop alternatives, maybe in the hope that all the things that are linked to the RPI would somehow migrate to a much better index. As Robert said, that has not happened, and I can understand that too. It is very difficult. As long as everyone else is attached to this index, it is very difficult for one entity—
Q320 Chair: I quote again from the RPI/CPI User Group, “The Treasury indirectly got its way by replacing the RPI with the CPI except of course where it was to their advantage to retain the RPI. Arthur Barnett gave his perception that the Treasury and the Bank of England are using their privileged position to influence statistical methodology to ‘show them in a good light’”.
Dr Broadbent: In what way am I using my—
Chair: I am asking you to respond to that.
Dr Broadbent: I disagree. I do not think I am using my privilege. I have no privileged position to determine how those statistics are developed.
Robert Chote: Governments can choose. As you have pointed out yourself, given the range of possibility that is available to them, it is perhaps not hugely surprising what you attach things that raise money to and what you attach things that cost money to. In that sense you have not changed, there is a range of numbers and the Government respond in that sort of way.
Correct me if I am wrong, but I think the Johnson Review of the inflation numbers was commissioned by either UKSA or ONS. That was not imposed upon them. In the wake of that, this idea of essentially allowing RPI to wither on the vine, having a range of other numbers and simultaneously saying, “I would not use that one if I were you” and having alternatives there to encourage was a view that was generated within the statistical system with the help of a commissioned study and it reached that judgment on that basis. The Johnson Review was not imposed on it by—
Chair: Do you want to respond to that point?
Clare Lombardelli: There are two different distinct questions here. There is how statistics are produced and the best way to do that. Rightly, the ONS is independent and UKSA regulates it on that and that is something that is a decision for it, not something that the Treasury would get involved in. I do not think the Bank of England would either. We see it as hugely valuable that that is independent and that it makes its decisions about the best way to measure statistics in that way. How statistics are then used is of course a decision for the Government and that is a separate question. I do not agree that on the production there is any sort of issue.
Robert Chote: On our home turf, the ONS decisions were taken on the classification of housing associations and on the move to the new treatment of student loans, neither of which are obviously advantageous to the Government and the presentation of the public finances.
The one on the housing association is an interesting example of reflecting the fact that the Government were exerting control over these bodies. ONS concluded, because it is control rather than ownership that matters, that these things should be counted as part of the public sector. The Government then took decisions on the regulation of these bodies and was publicly explicit about saying, “We have done just enough here to ensure that when the ONS revisits this it will knock it back on to the other side of the fence again”.
You might say that does not look very edifying, but it is all quite transparent as to what went on there. On the student loan side, again we highlighted back in 2017 that the existing treatment of student loans was effectively creating a fiscal illusion and flattering the underlying position of the public finances. That was followed up by the Treasury Committee and the Lords Committee as well. Clearly UKSA and the ONS did not lead off in that process, but having grasped the nettle they are now pursuing that very well and very effectively. They have taken it through the Eurostat process efficiently, they are engaging well with the stakeholders who will be affected by this and we talk to them a lot about how we think this is eventually going to affect the numbers. I am sure in an ideal world the Government would rather they were not about to add £12 billion or so to the core measure of the budget deficit.
Chair: By the way, you grasp a nettle in order to avoid being stung.
Robert Chote: It depends a bit what side of the nettle you are grasping.
Q321 Mr David Jones: Ms Lombardelli, I would like to ask you a few questions about the relationship between the Treasury and UKSA, which is set up as an independent arm’s-length body. How do you ensure that the Treasury does not have an undue influence over UKSA through the funding arrangements that develop?
Clare Lombardelli: In the same way that we do that for multiple independent bodies, in that sense of complete separation. The Treasury has an interest in making sure that funding is affordable and value for money, but beyond that we do not in any way overlap on the policy issues. We would fund it in exactly the same way that we would fund other independent bodies. The OBR is another one we fund. Obviously all independent bodies have to go through the public finance system, so in that sense we agree their funding, but beyond that—
Q322 Mr David Jones: How often do Treasury officials meet with UKSA or the ONS?
Clare Lombardelli: We meet with the ONS very regularly. As the previous testimony shows, there are multiple working-level conversations that are happening virtually continuously around the particular economic statistics that we use. It is a completely separate set of people who would deal with any of the funding questions around the ONS. I do not know, but my guess would be there would be quite infrequent discussions they would have, whereas on the use of statistics and on the economic statistics that is happening fairly regularly.
Q323 Mr David Jones: What sort of things are discussed?
Clare Lombardelli: All things. It might be how are you measuring investment data, how has that changed, why has it changed. In the national accounts there are lots of conversations going on all the time. Ben mentioned double deflation. There are huge amounts of conversation about what will that mean in practice, “How are you going about measuring it so that we can understand it?” It is really to help us, as users of statistics, understand how we should be interpreting the statistics that the ONS are producing, in the same way that I imagine other economic bodies are using them.
Q324 Mr David Jones: Do you think that there may be a risk that the Treasury is having an excessive influence over UKSA and how do you manage that risk?
Clare Lombardelli: As I say, the majority of our conversations are with the ONS rather than UKSA itself. We are a heavy user of economic statistics, as are other people. The ONS is responsive and has those conversations with us and is open and will listen to those in the way that it does with lots of other users. Of course we do encourage it to focus on the quality of its economic statistics because that is of the most value to us.
Q325 Mr David Jones: Do you think there is a risk that it might be perceived that the Treasury might have an undue influence over UKSA?
Clare Lombardelli: I do not think so. If you think about it, I think the risk goes the other way. If we were not talking to it about its economic statistics, I think that would be a problem. As users, it is really important that we talk to it so we can understand. We are advising Ministers to make economic policy and we need to understand the statistics on which we can base those, and I think that is a good thing. I would be quite worried if you thought the Treasury should not be talking to the producers of the statistics and therefore had less understanding of what those statistics told you because I think that would risk sub-optimal policy formulation.
Q326 Mr David Jones: Reverting to RPI, in response to the Lords report the Chancellor has said, “The Government are discussing the relevant issues with the UKSA”. When are you expecting to hear UKSA’s recommendations on how to address the issue of RPI?
Clare Lombardelli: The Chancellor has said that he is considering the Lords report and will have recommendations for both the Government and for UKSA and we are going to respond in full to that. I cannot talk in any more detail about that.
Q327 Mr David Jones: He also said there were discussions continuing over these issues. That is the case?
Clare Lombardelli: Yes. As you would expect, as Ben has said, these are complicated issues so we will want to discuss them.
Mr David Jones: Do you think April?
Clare Lombardelli: He has committed to responding in April.
Q328 Kelvin Hopkins: It is UKSA’s statutory duty to produce and safeguard, “official statistics that serve the public good”. Why was it necessary for the Chancellor of the Exchequer to commission the Bean Review of economic statistics?
Clare Lombardelli: Measuring the economy as it is changing is always quite difficult, capturing demand economy. There is a question, and it came out of the productivity plan at the time, about is productivity being correctly measured, is there an issue around how the economy is changing and are the statistics keeping up with that? On the back of that, the review was commissioned to take a look at this issue. As Robert said, it is useful to have independent expertise take a look at something that is a puzzle, in a sense.
Q329 Kelvin Hopkins: I will come on to productivity in a second, but what effect has the Bean review had?
Clare Lombardelli: I think it has had multiple effects. We have discussed at length here the Office of Statistical Regulation. That is one of the things that came out of it. I think that is an improvement. There have also been a couple of other things that came out that were really good. The Data Science Campus you have heard about. This is about 50 or so data specialists, if you like, looking at statistics and working out how we can improve our statistics by doing things like looking at big data. Some of the things that Ben has talked about, about administration data and public administration data and can they help us to better understand the economy and that sort of thing, work is underway. Some of it is already being used. VAT is now being included in the national accounts. That means you have a much better sample. That means that your statistics are likely to be more accurate and more reliable. Those sorts of things are happening.
There is also the Economic Statistics Centre of Excellence that it set up. That is a group that gets together to talk about future things that you might want to think about to futureproof the system as the economy continues to change. It is looking at some of the interesting things about beyond GDP and those sorts of questions. There is quite a lot that has come out of the Bean Review that is being worked on and there is quite a lot of new things happening and progress is there. It takes a bit of time, but it is all quite valuable and interesting stuff that should improve statistics.
Q330 Kelvin Hopkins: You mentioned productivity. The ONS produced an economic statistics and analysis strategy in response to the Bean Review. How effective has this been, for example, in improving productivity statistics to address what is described as the “productivity puzzle”?
Dr Broadbent: As you know, the rates of growth and productivity across the advanced world, but particularly in this country, have been low to minimal since the financial crisis. Some of that—and this is part of what Bean talked about—may be because you are failing to capture economic output and the growth in economic output in some less surveyed parts of the economy and newer parts of the economy, the digital economy and so forth. As an economist, I very much doubt that that under-measurement would account for anything more than a minority of the puzzle.
The puzzle will remain, but obviously we need to always try to keep up with structural changes that go on in the economy, what counts as output, are we able to observe it properly, keep track of how it is changing and so forth. I think it is extremely important that we do that, but I was just cautioning that this is not going to answer in any substantial way that productivity puzzle. It may be a part of it, but I doubt it is going to be a big part of it, in my view.
Clare Lombardelli: Not least because the economy is modernising across the world and other countries will also have the same issue. But some of the things that the ONS has done in terms of providing additional supplementary analysis around some of their productivity releases is valuable and interesting.
Q331 Kelvin Hopkins: I do not want to start a debate on economics, but we have low levels of investment and access to relatively cheap labour and that might affect the fact.
Anyway, what has been the impact of recent ONS initiatives, including the Economic Statistics Centre of Excellence and, as you mentioned, the Data Science Campus? Have they been beneficial and what has been the effect?
Clare Lombardelli: Yes, very. As I have just said, the Data Science Campus has been really useful in looking at how can we use some of these larger datasets, particularly some of the datasets we have around, for example, tax data, and how can you use those to improve the statistics that you are using all the time in the national accounts. A tangible benefit means our statistics will be better.
The Economic Statistics Centre of Excellence again I would say is making real progress. It played a big role in, for example, how you deal with double-deflation issues and it advised on some of that. To explain, that is a mix of people. You have a lot of academics on there. NIESR and Cambridge University are very involved. It is about making sure you have the best statistical techniques and methods you are using. Double deflation would be a good example of that.
It is doing some interesting work around beyond GDP. This is the question about are you accurately measuring the other capitals correctly and those sorts of things. There are good things coming out of these things. They take time, so you would expect to see more results over time, but they are both very good initiatives that have improved statistics.
Q332 Kelvin Hopkins: How much is there a sense that the Treasury is looking over the shoulder of UKSA and ONS because it wants them to perform in different ways and be more help to the Government, in a sense?
Clare Lombardelli: I do not think we have different objectives in what we want from the ONS and UKSA than everyone else. We want and need accurate, timely, reliable statistics. That is what it is in the Treasury’s interests for UKSA and the ONS to produce. It is not in our interests for them to do anything other than that because we are trying to make policy in the most informed way that we can. That is why we are supportive of them and want to see them succeed and flourish. We are dependent on their statistics to inform our policymaking.
Q333 Kelvin Hopkins: The fact is that some economic statistics are uncomfortable to Government because they want to put themselves in a good light and sometimes the statistics do not put them in a good light.
Robert Chote: There is no evidence, in the way that these things are being produced, that they are being produced in a way that is favourable to Government. Obviously an important check on that is that you have both the Bank and us as independent institutions using these numbers all the time. I can tell you that if we had a sense that we were not getting kosher numbers because the Government were interfering with ONS and getting it to produce them in strange ways, you would certainly be hearing about it from us. All of these exercises are complicated. Measuring what is going on in the economy and in society more broadly is a very complicated thing to do. As Ben said earlier, you can identify these are the exact features of what a high-quality statistic is.
All of these things involve trade-offs, of which the current student loan is a good example, where you have a system at the moment where the virtue is that you can be wrong, but quite precise about it. But we are going to move to a system that is going to capture the real world better, but that is necessarily going to be more approximate and the outturn data will have more forecast judgments embedded in it. Some people may say that the former is higher quality and others that the latter is. The idea that there are not perfectly reasonable, difficult judgments to make that will be guided by things other than, “Does this make things look good or bad for the Government?” these are complicated things to do.
I think the student loan one is a good example. We are going to get a better picture of the underlying health of the public finances, but possibly at the cost of having greater volatility in outturn estimates of the budget deficits. People will turn around and say, “These figures did not used to move around like this before, something has gone wrong”. No, it is the trade-off you are making between something that, as I say, may be more volatile, more frequently used, but most of the time is giving you a better picture of fundamentally what it is you are trying to measure than something where you can tick a lot more boxes in a technical sense, be precise. It is all very robust, there is little ambiguity around the numbers, but it is giving you a wrong answer.
Dr Broadbent: A perfectly measurable wrong answer. I can second that. The Bean Review specifically says there is no evidence at all that in the production of the statistics there is any undue interference from any arm of Government. What it did say is occasionally it is in the use of the published statistics that everything does not work perfectly, that a Minister might stand up and give either the wrong number or a distorted number. UKSA has come down on that on several occasions in a way that I think is very good. As for the idea that the numbers themselves are distorted by these sorts of things, there is no evidence of that at all. I have been a user of UK statistics for—
Kelvin Hopkins: I am not going to pursue this, but we are all in favour of encouraging numbers, absolutely, and independence, but sometimes Governments want numbers with which they are comfortable, even if they are not quite kosher. I will leave it there.
Q334 Chair: The UN publishes a generic law of official statistics, which is based on the UN’s fundamental principles of statistics. While it was intended for developing countries, it has been adopted as a basis for good practice in many countries. One of the things it recommends is that there should be an annual and multiannual framework of statistical work reviewing statistics. The annual programme would be a list of activities that would be undertaken.
UKSA produces a five-year strategy, which is very much its mission and operations, and a three-year business plan that is very ONS-specific. What UKSA currently does not do is a forward programme of work across the whole statistical system. How much do you think this would be of value to statistics users and to the statistics system as a whole?
Robert Chote: Obviously setting out the strategy has a value and a transparency of itself. There is an interesting issue about what UKSA can be doing. If you were to ask many heads of professional statisticians working in Departments as to what their sense was of UKSA’s input into the work programmes within a particular Department of the statistical profession, the people working that area will have all sorts of other pressures coming in on them. Having an UKSA presence and input into that that the statisticians can refer to and discuss when the Departments are meeting more broader priorities would probably be a good idea. It comes back to this idea of, where possible, UKSA being able to strengthen the hand of the statistical professions in Departments.
Clare will know more about this. In addition to the statistical professional, there is a greater emphasis on the analytical professions more generally, which would include economists and social scientists as well, because obviously there is a shared set of values that you want people to be adhering to on that basis.
I think in addition to what you hear from many heads of profession about the value of the code of conduct, refreshed in 2018, as something that you can point to when you feel your emphasis is not being put in the right areas, that you are being made to do uncomfortable things, but maybe there is less of an UKSA input into what is the work programme of statistical improvement within education, health, transport, whatever. That is probably consistent with that overall approach you describe.
Chair: Dr Broadbent is nodding. Clare Lombardelli.
Clare Lombardelli: I completely agree that anything that strengthens the professional function of statisticians across Government is hugely welcome, as it is for other analytical professions.
Q335 Chair: How could UKSA make the decision-making about its forward programme more engaging across Departments as a whole and more transparent?
Robert Chote: I am not close to the how the GSS organises itself internally and the relationships there. We deal with statisticians and officials working in some particular Departments and they do not get that sense.
Q336 Chair: How worthwhile is it exploring that?
Robert Chote: I think it would be worthwhile doing so, yes.
Clare Lombardelli: Yes. You have the National Statistician coming in, I think, as a witness.
Chair: We certainly do.
Clare Lombardelli: It would be one to raise with him. I obviously only see what happens in the Treasury in terms of statistics, so I am less sighted than Robert and Ben.
Dr Broadbent: The Bank, probably like the Treasury, as users we are almost exclusively focused on ONS and macroeconomic data and less on the areas that Robert was talking about. It sounds as though UKSA already has a forward plan for the ONS but maybe less for the GSS.
Robert Chote: Obviously in all of these Departments there are resource choices to be taken. For example, on the public finances side—and I think this would be shared with the Treasury—there is a view that it would be great to have more timely and more robust data on local authority finances as a feed-in.
Chair: That is very relevant to the next question coming up.
Q337 Dr Rupa Huq: Do the official statistics that exist that are out there inform you sufficiently for the decisions you need to make?
Dr Broadbent: This relates partly to something I said earlier. One always wants better data. It is a dream that you have a perfect, up-to-date measurement of exactly what is going on in the economy. That will never be entirely possible. Obviously one always wants improvement. The ONS is about to publish indicators that come out very quickly from tax data that are related to economic activity. That will give us, helpfully, an early steer if you are about to go into a recession or an economy is recovering or something, maybe these give you a little advance notice about it, which would be very helpful. But they are not going to be perfect because they are not perfect measures of output and you are going to get more information over time. It is bound to be the case that things are revised. I cannot give a discrete answer to your question is what I am trying to say.
Q338 Dr Rupa Huq: Could do better, but adequate?
Dr Broadbent: Yes, but one should accept that with the best statistical agency in the world with the best resources it is not going to be possible to produce the perfect answer, exactly the number you want to know, and it is never going to be revised. That is not going to be possible. When you say “better”, I am not talking about the performance of the people who produce them, I am talking about the numbers themselves. One wants progressively more information that allows more accurate and more timely measurement to the things we care about, which are economic activity, price inflation, cost inflation and so forth.
It relates partly to something I said earlier. I do not think one should imagine, if you see a number that is subsequently revised or is volatile, that this is necessarily the fault of the person producing the number. Robert gave an example to do with student loans. Sometimes it may be; sometimes it may not be. What we hope for realistically is a gradual continual improvement in the breadth, the accuracy and the timeliness of the numbers about the economy. Are they adequate? In some ways I am a hard taskmaster. I never say they are adequate, but I do not think that is the fault of the people producing them necessarily.
Clare Lombardelli: The most important priority for us would be around the ONS maintaining and improving the quality of its core economic statistics. As Ben said, this is hard to do. For example, Robert talked a bit about what is happening on construction, where that data has had to be improved to regain national statistics status. That is very welcome. Trade is another area of where there is a long history and it is very difficult, but that is an area where further work is needed, and the ONS recognises that and that is happening. Those are the sorts of areas where we would prioritise.
Q339 Dr Rupa Huq: You just said the economy is modernising, so things like the gig economy, zero-hours contracts, labour force studies are a bit clunky and slow to react to those.
Dr Broadbent: Yes, and also that particular set—one can give individual examples—the response rate has been falling, which is of some concern to us. Again, can it respond quickly? It takes time to send out a survey, to get responses back. Would you want to change it so as to ask a particular question about people’s working habits? Possibly.
Q340 Dr Rupa Huq: I think there is a lot more precarious employment beyond getting a gold clock after 60 years.
Dr Broadbent: How do you measure that? What are the questions you ask in that survey that would elicit that information? These things are not that straightforward.
Q341 Dr Rupa Huq: Is it possible to do? You could keep a record over a span of time what it averages at.
Dr Broadbent: Yes, you make a snapshot with that survey. You ask particular households what they have been doing over a particular period of time. You would need a new question and a different question to try to find out what you are getting at.
Robert Chote: We, like our colleagues, are obviously very interested in the core national accounts data and inflation and labour market figures. We are always dealing with the first draft of history each month when you get new sets of numbers out. We are conscious, in the same way as forecasters, you know that your forecasts are going to swing around, so is your understanding of what has happened in the past.
There are areas. A good example within the national accounts would be business investment, which has been a particular focus, obviously, in the wake of the referendum, what is going on to that, how is it responding, how could it potentially respond. This is an area where the numbers swing around a lot and they can be revised to a considerable degree, changing the pattern or the story that you thought you had, managing to explain to yourself satisfactorily about how the economy was moving. Obviously you would like to get rid of that, but you are not going to. At one level, for forecasters to complain about revisions is like sailors complaining about the sea. That is what we are on and that is the world in which you have to live. You would also always like to see improvements.
There are areas where I think you would like to see more. I mentioned local government, which in terms of getting a good picture of the public finances is always something about which you would like to have more robust data earlier. It is often long after the end of the fiscal year before that is firmed up. There are also areas where we are getting new and interesting information in. This is part of the theme around the exploitation of administrative data.
A good example there would be real-time information from HM Revenue and Customs on essentially PAYE. When we produced our most recent set of revisions to the public finance numbers, that data were able to tell us that one of the reasons why income tax receipts were coming in more strongly than expected was because this data shows you that you had very rapid growth in incomes at the top end of the distribution. Not simply the people with lots of dividends to play with who can shift those things around, this is employees. It is the top of the employee earnings distribution. Those sorts of new bits of information are coming along and are more helpful over time, but you would always like more. There are areas of weaknesses and there are areas of new strength.
Q342 Dr Rupa Huq: To what degree do you find yourself drawing on other sources and additional analytical capacity to make up for those gaps?
Robert Chote: Take local authority stuff. In addition to whatever data the Department collects, the quarterly returns and so on, you also want to talk to local government associations, CIPFA, academics with a particular interest in what is going on in local authorities. At that level this is primarily for the judgment we have to make about whether local authorities are adding to or reducing their reserves in deciding what to spend, given the income they have, and therefore how that feeds through into overall estimates of public sector borrowing and so on. That is a good example of something where you do want to try to get a bit of anecdotal—you have to be sensible about what weight you want to place on that, but that can be a useful addition to whatever it is you get out of the concrete return. Of course the Bank has a whole agency.
Dr Broadbent: We have a business survey and we make labour market surveys. We have an idea of how well they have performed in the past to weight them accordingly. We have a network of agents, so-called, across different parts of the country who talk regularly to businesses and who facilitate visits from Committee members to talk to those same businesses. The agents themselves do so. I would not say it is compensating for a deficiency in the statistics, but it is certainly a very useful supplement and may on occasion in particular give you a more timely steer on where the economy is going, and a richer understanding of why certain things are happening.
Dr Rupa Huq: It can be qualitative evidence as well?
Dr Broadbent: Absolutely, absolutely.
Q343 Chair: You have been very helpful to us. Thank you very much for being with us today. This inquiry is about governance and governance is as much about leadership and culture. Are there any closing comments you would like to give us about your view of the leadership and culture of the UKSA?
Dr Broadbent: My personal reflection would be that the leadership, certainly at the top of the ONS and UKSA, has been very good. We have mentioned the Bean Review a lot because I think it has been the perfect vehicle for something that is not overly critical, constructively critical, and it has given an energy. Obviously it has given more financial resources as well. The leadership I think has been very good, to be honest, for four or five years.
Robert Chote: I would agree with that. The challenge is obviously getting that to permeate its way down through the organisation and to get that sort of innovative spirit there. In addition to the innovative spirt, those people need to have the capacity in addition to the day job of turning the handle to have the time. It is partly about the resources and about the priorities that the leadership team puts there. Clearly a near-term very important decision is who the new National Statistician is going to be. I think John has been very effective as a leader. It is a hell of a combination of things to be responsible for in that job, in a governance and in a leadership sense. I wish the best of luck to whoever it is who ends up doing it.
Chair: Clare Lombardelli, a last word?
Clare Lombardelli: I have nothing to add to what the others have said.
Chair: Thank you all very much indeed.