Skip to content

ONS adopt the Hybrid model

December 19, 2018

Updated at 4pm, 19/12/2018

On Monday, the Office for National Statistics announced its chosen approach to classifying students loans in the national accounts.

It has chosen to adopt the option known as “Hybrid” or “Option 3”. It will probably go on to be known as the “Partition” model as its main innovation is to split student loans into a part estimated to be repaid (“lending”) and a part expected to be written-off (“spending”). An estimate for the latter will be classed as capital expenditure and be recorded at inception.

In effect, expenditure on loans has been brought forward from when the accounts are closed (and losses known) to when the loans are issued (and losses can only be estimated). There will be an as yet undetermined method for reconciling actuals, which may see the partition (specific to each cohort year) move both ways over the life of loans. (Some of the documentation refers to loan “cancellations” but this is misleading as the part determined to be “spending” will continue to exist from the perspective of borrowers.

I had a first go at writing about the new treatment for wonkhe. You can read it here.

What that article tries to make clear is that ONS have only announced a general approach and the detail of the new treatment is yet to be determined. We can expect a fully worked up methodology in June 2019 and some official estimates of the impact on the deficit then.

What we got on Monday was ONS gesturing towards potential implications for the headline public sector finance statistics though reference to a £12billion hit to the deficit as calculated by the separate Office for Budgetary Responsibility in its October Economic and Fiscal Outlook.

It turns out though that OBR’s interpretation of “hybrid” differs from ONS.

I have spent the last two days trying to work out the differences.  But in simple terms (or at least as simple as I can make it), it comes down to the following:

In the new accounting treatment, there has to be an identity between repayments and the combined total of  interest treated as income and the portion of loans considered to be “lending”.

That is,  Repayments = Interest as Income + Lending

Repayments are fixed in relation to the models run by the Department for Education (both OBR and ONS use these for their independent projections).

It is therefore a decision as to how you determine the relative balance of lending and interest as income. Whichever you determine first leaves the other fixed as the residual outcome of the equation.

OBR chooses to determine Lending first. Its method:

Calculate the ratio of Repayments to Loan Outlay plus Interest Charged. In figures used by both bodies

Repayments = £18bn
Outlay = £16bn
Interest Charged = £30bn

Repayments then represent roughly 40% of total. So 40% of initial outlay is lending (£6.4bn) and 60% is spending.

Having determined lending then you have a deflator (the same ratio above) which turns interest accruing each year against outstanding balances into interest recorded as income (£11.6bn over the lifetime of loans).

ONS take a different tack. They determine the Interest as Income figure first, with a much more complicated method, still only indicative. It basically involves assuming the “lending” balance is cleared by the mid-2050s and working back from that using projections of repayments and interest charged to borrowers to calculate a starting lending balance. They reach the conclusion that lending is roughly 50% of outlay. A 10 percentage point difference from OBR.

Repayments = £18bn
Interest as Income = £9.7bn
therefore

Lending = £8.3bn

which makes Lending roughly half of the £16bn outlay

If you are issuing £15bn of loans each year, that difference between methods equates to a variance in capital expenditure of £1.5bn.  With some of that offset by differences in Interest as Income. Perhaps counter-intuitively the method that generates lower Lending records a higher level of Interest here (but remember the key accounting identity).

In their October EFO, OBR provided a split for the £17.1bn they estimated would be added to the deficit in 2023/24 by the new accounting treatment (on current projections).  A £12.5bn reduction is due to the upfront capital expenditure (against £22.6bn of loan outlays), a further £4.6bn is due to the reduction in interest as income (from £8.3bn to £3.7bn).  This is the capital expenditure for the loans issued that year; the interest is derived from that accruing on all loan balances.

OBR has confirmed they will adopt whatever ONS decides on. ONS are very clear – what they outlined on Monday is provisional and still needs work and refinement. In particular, they will have to fix a method for revising the figures when actual repayments are known and diverge from initial projections.

So while the general principle is outlined – expected losses will be declared upfront as expenditure – the £12bn figure should only be read as a ballpark impact. What is also still to be determined is how that figure breaks down into a reduction in income (lower interest) and an addition of capital expenditure (write-offs brought forward). This may matter in terms of the scope any government has for choosing different measures of the deficit (in an effort to avoid a big hit to their headline fiscal targets).

The headline for my piece for wonkhe declared that fiscal illusions have gone, but the complexity of the new approach may leave us still dealing with the nebulous.

ps ONS’s accounting changes are UK-wide.

pps. for the sake of clarity I have tidied up some minor discrepancies between the OBR and ONS figures.

 

 

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: