The UK government’s Office for National Statistics has got its data on the national debt wrong, and it is deliberate

Posted on

I was aware I was being controversial yesterday when I suggested that the Office for National Statistics was lying about the scale of UK national debt in its bulletin on public sector finances, issued yesterday. I was also aware that I was being provocative. But in the sense that I was suggesting that the ONS were stating something other than the truth - as evidenced by the UK Whole of Government Accounts - I was entirely happy with the claim. After all, if the audited accounts for our government show debt to be other than the ONS claim it to be, because in the audited accounts the bonds owned by the Bank of England under the quantitative easing programme are offset against the total debt to deliver a much lower figure than that which the Office for National Statistics claims exists, then the ONS simply cannot be right.

Someone called Phil Stokoe, who implies he used to work at the ONS, did, however, get very angry in response. His suggestions were (and I quote him to make sure I am being fair):

[The] ONS are publishing a number of different measures for public sector debt, and trying to do so in accordance with internationally accepted statistical methodologies, but these are tempered by other considerations.

I agree they are, but let me be clear: none appear to be right. He continues:

The method for the headline number of debt (Public Sector Net Debt ex) has been basically unchanged for a decade and this was introduced during the global financial crisis.

I agree: the current method was introduced to exclude the debt of nationalised banks and disguise the impact of QE. As he notes:

[I]n 2008 the UK defacto nationalized RBS, Lloyds and Northern Rock, and so these entities were classified into the public sector, that meant their liabilities and their liquid assets entered into net debt — but back then, it was envisaged that these interventions, and the BoE quantitative easing, would be short term affairs — that the banks would be sold back to the private sector (as has happened now with Lloyds) and QE unwound.

But those assumptions did not prove to be good. Many of the nationalised banks failed. 62% of RBS is still state-owned.  And QE has not been unwoud. So the conventions adopted were wrong, as it turned out. And so too, then, was the accounting, which needs to be revised, which is the whole point of my claim. Stokoe claims otherwise, saying:

In addition to this, ONS uses a definition of gross debt that includes three types of liabilities — debt securities (bonds and T-Bills), loans, and currency and deposits. These are all clearly liabilities, and widely agreed that these are debt (EU Maastricht debt includes these three instruments, the IMF Government Finance Statistics Manual definition of gross debt encompasses these and other laibilities too).

Except that's not what they call national debt, as I make clear below. But before doing so, I noted he provided a link to an ONS spreadsheet which suggests that  the UK public sector balance sheet for 2018/19 (which is the last year for which we also have Whole of Government Accounts) looked like this:

I offer this to contrast with the audited Whole of Government Accounts, that looks like this for the same date:

Should we worry that the ONS is, overall, about £900 billion out? I suggest we should be. Only one of these numbers is right, and I'm going to suggest it's the audited ones. And in them the debt, in the form of borrowing, is stated as follows:

Whilst other financial liabilities are stated to be:

Having noted this, let me go back to the claims from Phil Stokoe. He says:

QE has the practical effect of replacing a government debt security liability with a Bank of England currency and deposit liability. There is still a liability there. Debt is only lower by 735 billion as you claim, if you use a narrow definition of debt, and ignore the Bank of England’s own liabilities. As for the 200billion at NS&I, these are clearly liabilities as well, I’m not sure why you’d argue to exclude these.

And he adds:

All of this is complex, all of this depends on the accounting conventions used (ONS is not following IFRS or the public sector equivalent, IPSAS), but statistical standards, in line with the Eurostat manual ESA 10 and the IMF GFSM.

You are objectively wrong to state that ONS are lying, and its not correct to call this false accounting either. You, as much as anyone, now that accountancy is riddled with conventions and differences in treatment which is why the accountancy communities professional in joke is “What is 2+2? What do you want it to be”.

ONS are following a consistent approach, that has been in place for a decade, and its disgraceful for you to question their motives in this way, and you should strongly reconsider this post.

In summary, he demanded that I apologise.

I will not be doing so. There are very good reasons for not doing so.

First of all, debt subject to QE is no longer debt, and accounting standards make that clear. In other words, I have literally nothing to apologise for. His claim, and that of the ONS is wrong.

Second, because a statistical convention agreed in 2008 says otherwise does not make that alternative claim true. I will stick to accounting standards on this occasion as the indicator of what is true and fair.

Third, I can say with absolute confidence that the statistical convention that he refers to is in denial of the truth, as is the ONS as a result. As the March 2019 version of the UK government debt and deficit report for the Office for National Statistics notes:

To claim that the ONS is right when to make it so requires that it deny that a Bank of England subsidiary company is within the government sector is pushing the boundaries of credibility beyond any reasonable limit.

Fourth, to claim that the standard set in 2008 is, in that case, an objective basis for the assessment of the truth is, very politely absurd. Implicit in that claim are a number of assumptions that I think are wrong. The first of those is that there was no politics involved in misrepresenting national statistics at the time when the commencement of QE programmes was deeply controversial and required political cover from national statistics, which has been maintained ever since despite the obvious need for reform. The second is the assumption that such statements are beyond political influence, which is very obviously false when all accounting is politically laden as to meaning. The third is that the statement was ever right, when it has now become apparent that QE programme has not been temporary or likely to be reversed, as was incorrectly assumed (or claimed) in 2008. It is  now,m very obviously, permanent and highly unlikely to ever be reversed, changing the whole nature of the issue. And therefore, there is, fourthly, an assumption that the failure to revisit the convention in the light of subsequent behaviour and evidence is not itself a failure on the part of those using this framework, which failure has itself resulted in misrepresentation of the truth. And fifth, there is a failure to recognise that this misrepresentation has been deliberately maintained by vested interests to achieve a political goal even though they must know it is wrong, as the accounts show.

But perhaps as significantly, I simply think the claims about the ONS data that Phil Stokoe makes are wrong. In fact, the ONS agrees with me on that. Phil Stokoe claims I am wrong because debt must include central bank reserves. But that is not in any way what the ONS are saying. They do not say this because it's clear that they do not even understand many of the related issues. For example, they make this quite extraordinary false claim:

But that is not how cash is made available to the government. This year almost all of the government's borrowing requirement will be covered by QE, and as the Bank of England says of QE:

So, the UK government does not, to get cash, sell financial instruments as the ONS claims. Instead, it asks the Bank fo England to create the cash for it. And that is exactly what is happening. If they cannot tell basic narrative truths what else should we believe about their data? My guess is that they have also not read how the Bank of England says money is created. Nor, I suspect, have many who have commented on this issue. But that is an issue I will address in another blog.

To return to the ONS, what it says is this regarding debt:

Then they add this:

No explanation as to precisely how these sums are calculated, or reconciled, is supplied: indeed, despite publishing a figure that the ONS knew would attract headlines (debt of £2,004 billion) it made no attempt to explain the composition of that number or what it might actually mean, barring these vague estimates. And please do not refer me to spreadsheets headed by code to say they are the explanation: an explanation in code that is only known to ONS insiders is not an explanation, at all.

What is more, the House of Commons Library, still insist that:

In fact, the best data says that NS&A balances are about £177 billion at present. But if that's true (and that's audited accounting information, so I am going to rely on it) then what is very clear is that the £2,004 billion of debt the ONS says exists, less the £194.8 billion they say is attributable to the Bank of England, less the £177 billion attributable to NS&I does not reconcile to their figure of  £1,681 billion of gilts outstanding by some £49 billion, the cause for which difference is left unexplained, hanging in the air, giving rise to considerable doubt that any of the numbers they refer to of their own creation actually have much substance to them, and may be subject to either significant degrees of error, or are simply balancing numbers, the composition of which they do not know or are at least not willing to publicise, which in itself is a significant failing on their part.

But whatever the true numbers are what is clear is that Phil Stokoe's claim that national debt includes central bank reserves according to the ONS is wrong. The ONS ignores central bank reserves in publishing this figure, just as it ignores quantitative easing and just as much as it ignores the Bank of England Asset Purchase Facility Fund for purely political reasons, as I have suggested throughout my arguments.

I should add just one further point, though. It is that when the House of Commons Library issued a note on the National Debt, also on Friday, they did recognise that QE had some impact because they noted this:

That is, they recognized that the APF - the Asset purchase Facility - has an impact on debt costs because it cancels debt, and yet the ONS cannot. That I can suggest that the Whole of Government Accounts are right and the ONS is wrong is, then, something which some other parts of government are also capable of recognising.

So, do I have reason to apologise in that case?

Very clearly not. My claim is right: national debt does not include central bank reserves, and the ONS actually agrees.

But what is more, the quantitative easing debt can and should be offset against gross debt owing as stated by the ONS, as happens in the Whole of Government Accounts, but which only does not in the ONS data because of a purely political decision to assume that the Bank of England's purchases of government guilts take place outside government when very clearly they occur right at its heart, with explicit Treasury permission.

And the ONS data does not, in any case, remotely reconcile to audited information and that published on Friday was put into the public domain without any proper explanation that suggested how it was comprised, which itself was negligent on the ONS's part.

Phil Stokoe's demand that I apologise is, then, a demand that I am very happy to ignore.

I won't demand that he apologises. That's not my style.

But I may be raising this issue with the ONS.

And other blogs will follow on the issue of central bank reserves and the national debt.


Addition on 23/8/20:

Phil Stokoe and I are now in discussion.

He is no longer at the ONS.

I am giving him a right of reply to this post and he intends to write one. 

We are also planning to discuss the issueS it addresses, and which I too will be addressing in blogs to come. Whether that will be on or offline is not yet decided. 

I will leave him to decide if he wishes to disclose his current position.