The welfare cap: A political gimmick with very real risks
The ‘welfare cap’ announced by the Chancellor in the CSR is widely, and rightly seen as a political gimmick rather than the much-needed strengthening of the armoury of fiscal discipline it was presented as. This doesn’t mean that it won’t have real-world consequences however. The government’s intention is to write the cap into legislation so it acts as a ‘binding policy constraint’. The spending round document states: ‘The independent OBR will judge the Government’s performance against the cap and the Government will take steps to enshrine this new framework in law.’
If implemented in the way the spending round document suggests (and that is a big ‘if’, as Christian Guy suggests) the main implication is not greater fiscal discipline but roughly the opposite- placing expenditure decisions at the mercy of forecasting error. Any substantial downside errors in spending forecasts will lead automatically to benefit cuts. Because downside errors are more likely in tough economic times, and because of the way the cap is designed, cuts will tend to be pro- rather than counter-cyclical when the economy is at its weakest, reinforcing economic downturns rather than offsetting them.
A look at the coalition’s forecasting record since 2010 shows that this is not a risk to be taken lightly.
Details on the cap in the spending round document were sketchy, but we know a few important things. It will be set in nominal terms over a five year period, thus leaving benefits exposed to inflation risk. As expected, it won’t include the state pension. There is a nod to maintaining the automatic stabilisers- that is, benefits which rise in response to economic downturns- but this appears only to involve Jobseeker’s Allowance and associated benefits: ‘In addition, the cap will take account of the automatic stabilisers by excluding the most counter-cyclical elements of welfare, such as JSA and any passported expenditure. All other social security and tax credits expenditure will be included.’ There will be some sort of margin to avoid small fluctuations in forecast leading to policy changes – what is meant by small is not laid out however.
The claimed rationale for the cap is ‘to protect against structural medium term deterioration in welfare’, ‘deterioration’ here clearly meaning ‘spending increase’. This makes little sense. There has been no structural – as opposed to cyclical – change in overall benefit spending or in working age spending since 1997, so a cap is clearly unnecessary to achieve this objective. And the cap will apply to cyclical as well as structural expenditure increases if the only working age benefits excluded from the cap are Jobseeker’s Allowance and associated benefits: tax credits and housing benefit seem to be included.
These benefits do indeed pose a problem for the coalition, as they have grown far more than the government and OBR forecast in 2010, but the pressures behind this growth are quite clearly cyclical. As Chris Goulden has noted ‘The suggestion that spending on these benefits does not reflect economic cycles is hard to take seriously.’
It’s worth remembering just how the coalition expected the economy to develop at the time of the last spending review – and what actually happened. The chart shows the forecasts in December 2010 and March 2013. By 2012, on the 2010 forecasts, we should have been seeing rising real wages and GDP growth of 2.6% : instead, we saw wages continuing to fall and growth of next to zero. On the revised expectations in the 2013 forecasts the drivers of working age benefit expenditure won’t be going away. Earnings are forecast to fall by a further 1.4% in 2013, and unemployment is set to stagnate to 2015.
Source: OBR Economic and fiscal outlook Dec 2010 and Mar 2013
In fact, in the last quarter of 2012, social security spending as a share of GDP hit levels only previously seen under the Major government in 1993, the previous peak of postwar spending. Given this record, it is understandable that the government should wish to distract attention from its economic performance by evoking an imaginary ‘structural’ increase in social security. That the increase is imaginary is clear from the chart: measured on a consistent basis, overall spending showed only seasonal movements from 2000 up to the last quarter of 2008 and the onset of recession
Sources: Public Sector Finances Supplementary Tables series ANLY (social security spending) ; Quarterly National Accounts series YBHA (GDP at current market prices). GDP data from 1997 incorporates revisions published 27 June 2013.
Given the forecasting record to date, it is worth asking what would have happened had the coalition had introduced a cap at the last spending review.
Back then, the OBR forecast that social security and tax credit expenditure in 2014-15 would be £207.6bn in nominal terms. At the time of the 2013 budget, it forecast spending of £212.4 bn, £4.8bn higher. Here a technical point is unfortunately unavoidable: these figures are not strictly comparable, as the 2013 forecast does not include Council Tax Benefit, which has been devolved to local authorities. The 2010 forecast did include CTB. This means the rise in spending above plan is understated.
Excluding CTB in both years to make the figures consistent shows that spending in 2014-15 is set to be over £9bn more than forecast in 2010. These are nominal figures, not taking account of inflation – but as the cap is to be set in nominal terms, it’s these figures that count.
Source: OBR Economic and fiscal outlook Dec 2010 and Mar 2013; DWP Benefit expenditure tables Dec 2010 and Mar 2013
So the record on forecasting social security spending under the coalition is far from impressive: even in the context of total spending of over £200bn, a forecast error of £9bn is getting into real money. JSA accounts for £1.5bn of the forecast error. There were also modelling errors in the pension spending forecast, which accounts for a further £660m. These are the only major spending lines that won’t count towards the cap, so government would be nearly £7bn over the limit in 2014-15- in other words, it would have had to impose a further £7bn of cuts had there been a cap in place.
The big drivers of spending exceeding forecast levels were housing benefit, tax credits and employment support allowance. As noted, the first two of these are quite clearly driven by the worse than expected economic context. With employment support allowance, government forecast bigger falls in caseload than actually occurred because the impact of successful appeals against assessment decisions had not yet been recognised. The government also decided to have the assessment process reviewed by Professor Malcolm Harrington and implemented many of his recommendations – so the extra spending above plan on ESA is at least in part an intended consequence of policy choice. Unexpectedly high inflation also played a role, although by 2014-15 this wasn’t the most important factor.
So had there been a cap in place in 2010, the coalition would have had to reduce spending by an additional £5bn purely because of cyclical spending pressures which were not taken into account- making a joke of the claim to be protecting the ‘automatic stabilisers’- and by a further £2.1bn due to other forecasting errors. Do we have any reason to believe that the government or the OBR’s five year forecasts for 2014-2019 will be any more reliable? No: but we have reason to believe that the consequences of forecast error will be much more serious once the cap is in place.
What the data on spending under Labour suggests is that in stable economic conditions, overall social security spending is quite predictable: and because it is predictable, there is no need for the gimmick of a welfare cap to maintain stability. It is when there is an economic downturn that spending becomes harder to predict, as the coalition has found to its cost. This means that the real effect of a cap will be to limit the ability of social security to respond to economic conditions. This consequence could to some extent be mitigated by including cyclically driven housing benefit and tax credit spending in the cap. But let’s not fool ourselves that that would be anything more than patching a measure which really has no convincing rationale for existing in the first place, other than the Chancellor’s medium term political objectives.