On 28 Feb, FT’s columnist Chris Giles attacked the United Kingdom Office of National Statistics in his article titled UK’s official statistics cannot be trusted.

As his title suggests, Giles argues that the ONS is helping the UK government hide some public finances statistics and is losing its independence. His article is misleading to say the least.

Also see Jil Matheson’s reply to FT.

To see this one needs to understand some national accounts concepts. Usually economists and journalists err on simple concepts and Chris Giles does the same – leading him to his strange conclusions.

The government’s deficit is the difference between its expenditures and income (mainly taxes). The difference is financed by borrowing from the markets. This used to be called the PSBR – Public Sector Borrowing Requirement. The government also earns from the central bank’s profits. During any period of measurement however, the government also rolls over its maturing debt and the nomenclature changed to PSNB – public sector net borrowing.

So we had

G – (T + F_cb) = PSNB

where G is the government expenditure including interest payments (and including interest payments to the central bank), T is tax receipts and is the F_cb profits of the central bank (assumed to be remitted in full to the Treasury for simplicity). However, during the financial crisis, the UK government needed to bailout banks and to finance its purchases of newly issued equities, it needed to borrow more. Hence the above relation was no longer valid.

So

G – (T + F_cb) ≠ PSNB

However, the deficit is fundamentally the left hand side. The purchase of equities cannot be thought of as adding to the government’s deficit because the government acquires financial assets in return. Of course this adds to PSNB and hence also to the public debt. So to measure the real effects, the UK ONS proposed a measure PSNBex.

G – (T + F_cb) = PSNBex

This is as it should be – the bailout of banks by purchases of its newly issued equities cannot be said to be equivalent to an expenditure by the government which has a direct effect on aggregate demand. The bailout of banks although has its benefits – in the sense that they will be less constrained in lending – has an indirect effect and it will be captured when banks lend for a house purchases and this will reflect in investment expenditures. Of course the bank bailout has to appear somewhere in the accounts and it does and the UK ONS has been careful in doing it the right way.

In the Euro Area however, this is captured differently in national accounts and this is not the best way in the spirit of the accounts. So when the Irish government bailed out banks, it was measured as a huge rise in the deficit.

So here is from the Eurostat for the year 2010 and you can see Ireland visibly:

(Source: Eurostat. Click to enlarge)

Back to UK national accounts.

Assuming bailouts are in the past, i.e., not in the period of recording and also assuming no sale of equities by the government,

G – (T + F_cb) = PSNBex = PSNB

Now, the Bank of England has been has been purchasing government bonds in open markets as part of its Asset Purchase Facility program (strange word “facility”). It established a fund (or a financial vehicle) called Bank of England Asset Purchase Facility Fund Ltd which funded its purchases of UK government bonds by borrowing from the Bank of England. Over time, the BEAPFF has made profits on its operations and this is reflected in its cash balances at the Bank of England. The UK Treasury recently decided to transfer the cash to its own account (“raid” in the words of Chris Giles).

The cash transfer is an income flow from the Bank of England to the Treasury and hence reduces the PSNBex.

The ONS came with a technical note on how to measure this and Giles decided saw it as a proof that the ONS is not independent!

Although these effects seem intra-government, national accounts (SNA) treats the central bank as part of the financial sector and hence the transfer of the cash from BEAPFF reduces the PSNBex. Hence according to the ONS:

Following recommendations from the National Accounts Classification Committee (NACC), The Public Sector Finances Technical Advisory Group (PSFTAG) and Eurostat, it has been decided that:

• The Asset Purchase Facility will continue to be treated as part of the Bank of England in National Accounts and Public Sector Finance statistics.
• The flows of cash from the BEAPFF to HM Treasury, up to the level of the combined Bank of England’s ‘Entrepreneurial Income’ from the previous year, are treated as final dividends and therefore reduce the level of General Government Net Borrowing by that amount. However, anything above this level will be treated as a special transaction in equity, known as a super-dividend, which will not impact on the level of General Government Net Borrowing. (This calculation is known as the super-dividend test.) Whatever the impact on General Government Net Borrowing, the full value of the payments will impact on the General Government Net Cash Requirement.
• Any flows of cash from HM Treasury to the Bank of England in the future to cover losses made by the BEAPFF will be treated as Capital Transfers and so will impact (in the opposite direction) on measures of General Government Net Borrowing and General Government Net Cash Requirement.
• The BEAPFF as a whole should remain classified as a temporary effect of financial interventions as the end state of the BEAPFF remains unknown.
• The payments between the Bank of England and Treasury should be treated as permanent effects and therefore impact on the headline measures of Public Sector Net Borrowing (PSNB ex) and Public Sector Net Debt (PSND ex), which exclude temporary effects of financial interventions.

All this is consistent with the principles of national accounts – how the central bank is treated and how various flows are measured and recorded etc. The government interest payment to the BEAPFF increases PSNBex and the cash transfer reduces it.

The ONS says clearly that in case the BEAPFF suffers a loss, there will be a capitalization by the government and this will be treated as a capital transfer and will impact government’s accounts – but in the future.

Yet in spite of very clear thinking and action by the ONS, Chris Giles erroneously concludes that “official statistics cannot be trusted”. His critique is vacuous.