The great news revealed in the recent NHS 2010-11 accounts is that commissioners spent significantly less on administration that year than was previously assumed.

The bad news for the health secretary is that reduces the overall administration cost baseline from which his reforms were supposed to extract huge net savings, from £5.1bn to just £3.95bn.

The DH reckons it can reduce that to £3bn by 2014-15, but such a reduction would equate to just £2.1bn accumulated “savings” over the period.

Not only is that much less than the £5.2bn saving originally flashed around the Cabinet table, it’s also somewhat shy of the £2.5bn the Treasury is quietly expecting to be the true up-front impact of the reforms. Oops.

Now watch carefully what the revised September impact assessment does next.

Trick one: having the cake and eating it. Quickly dismissing the evidence in the accounts that primary care trusts spent just £2.4bn on admin in 2010-11, the impact assessment counters that, surely, PCTs actually meant to spend £2.76bn. Thus it takes £2.7bn as the PCT baseline, but also chalks up £240m of the difference as a year-one “saving”.

Trick two: transubstantiation. The remaining £100m is left unclaimed as a “provision”. But keep your eyes on that £100m, ladies and gentlemen, and you will see it transform into £400m by 2014-15, simply by being left in the inflated “baseline” against which each year’s progressively reduced spending is measured. The £240m “saving” similarly multiplies, magically becoming £1.2bn by 2014-15. That’s £1.6bn saved already, and the bill hasn’t even been passed yet.

Trick three is really just a tidying up job. Although there was none in the January impact assessment, suddenly a bit of contingency is deemed necessary. So, the figure £206m gets plucked from mid-air and also shoved in the “baseline” whereupon it too multiplies as a saving each year.

Once other costs are added, these magical adjustments to the facts take the full 2010-11 admin spend from the £3.95bn in the accounts to a fictional overall admin “baseline” of £4.5bn.

But more importantly, they boost the “savings” claimable by 2014-15 from a likely £2.1bn to a stupendous £4.5bn. That’s £2.4bn conjured out of nowhere, and before your very eyes.

Abracadabra: How the September Impact Assessment conjures £2.4bn savings from nowhere

 2010-11 (£m)2011-12 (£m)2012-13 (£m)2013-14 (£m)2014-15 (£m)Total over period (£m)
Actual total admin spend 2010-11 (the real baseline)3,9543,9543,9543,9543,95419,770
Sum added to 2010-11 “baseline” by Sept IA*5465465465465462,730
DH claimed “baseline” total admin spend4,5004,5004,5004,5004,50022,500
Planned admin spending4,2603,8573,6133,2813,00018,011
Actual savings (actual 2010-11 spend V planned admin spend)-973416739542,065
Savings planned against inflated “baseline”, claimed in Sept IA2406438871,2191,5004,489
Difference in savings2405465465465462,424

*Additions are: £240m both claimed as 2010-11 PCT admin “saving” and added into baseline; £206m general contingency; £100m PCT contingency.

Source: Coordinating Document for the Impact Assessments, DH, September 2011, p54-6

Sally Gainsbury is a news reporter for the Financial Times.