How does UK expenditure on benefits and services for disabled people compare with spending in other countries? To answer that question, we need a framework in which spending in countries with very different policies and institutions can be assessed on a fair basis. Fortunately, such a framework exists - in fact, there are two frameworks, one from the OECD and one from Eurostat. Although they measure different things - the OECD provides data on expenditure on 'Incapacity-related benefits' and Eurostat on 'social protection benefits in the disability function' - they adopt similar approaches to fundamental issues of classification and definition. If we want to compare the UK with other countries on a like-for-like basis, these are the best sources to use.

Both Eurostat and the OECD are explicit about the conventions they use in order to ensure comprehensive coverage of all the relevant types of expenditure. There is therefore no excuse for garbled comparisons such as this, from welfare reform minister Lord Freud in June: 'I will set a little bit of context by saying that even in these hard economic times this Government continues to spend around £50 billion a year on disabled people and services to enable those who face the greatest barrier to participate fully in society. That figure compares well internationally. We spend almost double the OECD average as a percentage of GDP—2.4% against the OECD average of 1.3%.Only two out of the 34 OECD countries spend more.' (H/T Disability News Service

A cursory glance at the OECD social expenditure database shows that the OECD average expenditure on 'incapacity-related benefits' in 2009 (the latest data available) was 2.6% of GDP, not 1.3%. UK expenditure was 3% of GDP, not 2.4%. 2.4% of GDP in 2009 was not £50bn but £34bn, and even if we apply the 2.4% figure to UK GDP in 2013/14, this only gets us to £38bn. 3% of GDP is not 'double' the OECD average, it is 1.16 times the average. Ten, not two, OECD countries spend more than the UK.

That's five serious factual inaccuracies in one [correction: two!] sentences. The only figure here which is plausible is 'about £50bn' for overall spending, which is reasonably close what you get if you apply the correct 2009 percentage (3%) to forecast GDP for 2013-14. But this figure is patently inconsistent with the other figures Lord Freud cites.

There is no great mystery about how Lord Freud got his figures so badly wrong. He has mixed together figures on total expenditure on incapacity related benefits (about £50bn) with figures on one component of those benefits, referred to by the OECD as 'disability pensions'. The latter does indeed come to 2.4% of GDP in the UK, and it also comes to precisely twice the OECD average for this type of benefit. But this is hardly a fair comparison: on this component Germany spends a mere 0.1% of GDP on disability. Does anyone think the UK is spending 24 times as much as Germany? In fact, Germany spent 3.5% of GDP on incapacity related benefits in 2009.

Why would someone compare disability expenditure across the board using just one component of spending? Well, it's hard to ignore the fact that it is only on this component that the UK could be held to be an exceptionally high spender in international comparison. Looking at total expenditure, the UK is neither especially generous nor especially stingy: it spends less than the Nordic countries, the Netherlands and Germany, and more than France, Spain, Belgium, Italy and Ireland (see chart).

The reasons the UK appears as a high spending country on the OECD's 'disability pensions' component only show how misleading the minister's comparison is. There are a host of ways in which welfare states differ in how they support disabled people. Some systems rely more heavily on benefits in kind - directly providing or subsidising goods and services - than others. The balance between tax funding and mandatory contributions from employers and employees varies. In some systems, most employment support for disabled people is delivered through dedicated programmes while in others most of the support is integrated with wider provision. And so on.

In dealing with this variation the frameworks used by the OECD and Eurostat share a number of features. Both cash benefits (what people in the UK tend to refer to as 'welfare') and benefits in kind are included. Healthcare services are excluded in both frameworks, as are (identifiable) benefits for the retired. Both public expenditure and spending on mandated benefits such as social insurance and sick pay are included. But there are some important differences between the frameworks: Eurostat distinguishes between 'disability pensions' (corresponding to IB/ESA in the UK) and 'care allowances' (in the UK, Disability Living Allowance), while the OECD includes care allowances in the 'disability pension' category; and the OECD includes sickness benefits while Eurostat excludes them.

Both of these differences are relevant to the UK appearing to be a high spender on 'disability pensions' in the OECD data. Conflating 'care allowances' with 'disability pensions' has a big impact on UK spending, because the UK makes more use of this type of benefit than most other countries. This is much clearer in Eurostat's framework, which treats 'care allowances' separately and gives a much lower figure for 'disability pension' spending in the UK. But while the UK spends more on this type of benefit than other countries, it also spends a lot less on benefits in kind, some of which serve a similar function to DLA, than some countries. In 2006 (the last year for which a detailed breakdown from Eurostat is available) the UK spent 0.7% of GDP on care allowances, while Sweden spent 0.1%. But Sweden spent 2% of GDP on benefits in kind compared to a mere 0.4% in the UK. It seems reasonable to infer that DLA in the UK to some extent substitutes for spending on benefits in kind, as of course was intended by successive UK governments.

At the same time, sickness benefits have an impact on the UK's relative position, because the UK spends very little on these benefits - reflecting a shorter entitlement period than many other countries, the fact that the benefit is flat-rate rather than earnings-related and low coverage due to labour market flexibility. In fact, UK spending is a mere 0.1% of GDP compared to an OECD average of 0.7%. There are two effects: on the one hand the UK's ranking in terms of total incapacity related expenditure is lower, but ESA acts as a substitute for sick pay, so the UK's ranking in terms of 'disability pensions' is higher. Comparison needs to take account of how a shorter sick pay entitlement and lower coverage will tend to push up receipt of 'disability pensions', as is clearly the case in the UK, where many people make short-term claims for ESA because their sick pay has run out or because they aren't entitled to sick pay at all. So using the OECD 'disability pensions' category as the basis for comparison is even more misleading, as this component is substituting for both benefits in kind and sick pay.

All of this should make it clear why expenditure on disability needs to be looked at in the round, and why cherry-picking individual components falls into the category of applied data torture (q.v. It also brings out what is really distinctive about the UK system. Compared to other European countries, UK spending on helping disabled people with the additional costs they face (counting both cash 'care allowances' and benefits in kind) looks respectable, lower than Sweden but similar to Denmark and Norway and higher than France or the Netherlands. But the UK spends very little on the benefits which cover basic living costs for sick and disabled people, Incapacity Benefit and ESA, and this is even more marked when we take into account the fact that these benefits are also to some extent substituting for sick pay. Of course under the coalition's plans the UK will be spending less on both types of benefit in the future, which may explain why Lord Freud feels the need to grossly exaggerate the UK's spending relative to other OECD countries.