The Centre for Social Justice claims that its proposals for welfare reform would get 600,000 households into work. This is an example of the fallacy of composition.
The gist of its story is simple. It proposes to increase the return to working relative to not doing so by increasing in-work benefits for the low-paid. And it claims that decisions to move from welfare to work are quite income-sensitive. Put these together, and you get a big rise in people moving from benefits to work.
There’s a minor problem here and a major one.
The minor one is that there is uncertainty around estimates of the elasticity of labour supply. The CSJ takes its figures (in appendix D of this pdf) from this paper by Stuart Adam. But he says “the economics profession remains far away from a consensus on their magnitude”, and in fact he doesn’t estimate them at all, but rather draws on sources such as this pdf and this pdf.
The major problem is: even if we assume that 600,000 people will want to move from welfare to work, how will they find jobs? Just because such people want to work does not mean that they will find work.
It’s not the case that the extra supply of workers will bid down wages and so create labour demand; the existence of the minimum wage will stop this happening (as we‘re talking about low-skilled workers here, the minimum wage is a biting constraint).
Instead, there’ll only be work for them if the economy grows. But this is a tall order. For example, in the five years to the peak in 2008Q2, employment rose by just 263,000 a year - and this includes a lot of skilled jobs unsuitable for long-term welfare claimants. And this growth came at a time when both banks and the government were throwing cash around like a sailor in a brothel - which they are unlikely to do in the next few years.
This is why I say there’s a fallacy of composition. If we look at a single welfare claimant, it’s quite possible that they would get a job if it paid them to. But what’s true of an individual isn’t necessarily true for all. A big rise in employment requires not (just) effort from welfare claimants, but economic growth too. As Mr Adam writes, the elasticity of labour supply is "endogenous to the state of the economy."
Nor is it valid to extrapolate evidence of smallish changes in benefits (which are the basis for what estimates of work elasticities we have) to large-scale changes of the sort the CSJ proposes.
So, why does the CSJ sully what is in many ways an impressive report with such an implausible claim? I can think of two possibilities, both of which are depressing.
1. It is so wedded to the belief that getting off welfare into work is a matter of individual responsibility that it cannot see that what (might) be true for one needn‘t be true for all. Macro isn’t merely micro writ large.
2. The report’s recommendations - higher spending on support for the low-paid, at the expense of the better off - are hard to sell to the Tory party and the trash press. Its claim that hundreds of thousands will move off welfare is the sugar it needs to coat this bitter pill.
The gist of its story is simple. It proposes to increase the return to working relative to not doing so by increasing in-work benefits for the low-paid. And it claims that decisions to move from welfare to work are quite income-sensitive. Put these together, and you get a big rise in people moving from benefits to work.
There’s a minor problem here and a major one.
The minor one is that there is uncertainty around estimates of the elasticity of labour supply. The CSJ takes its figures (in appendix D of this pdf) from this paper by Stuart Adam. But he says “the economics profession remains far away from a consensus on their magnitude”, and in fact he doesn’t estimate them at all, but rather draws on sources such as this pdf and this pdf.
The major problem is: even if we assume that 600,000 people will want to move from welfare to work, how will they find jobs? Just because such people want to work does not mean that they will find work.
It’s not the case that the extra supply of workers will bid down wages and so create labour demand; the existence of the minimum wage will stop this happening (as we‘re talking about low-skilled workers here, the minimum wage is a biting constraint).
Instead, there’ll only be work for them if the economy grows. But this is a tall order. For example, in the five years to the peak in 2008Q2, employment rose by just 263,000 a year - and this includes a lot of skilled jobs unsuitable for long-term welfare claimants. And this growth came at a time when both banks and the government were throwing cash around like a sailor in a brothel - which they are unlikely to do in the next few years.
This is why I say there’s a fallacy of composition. If we look at a single welfare claimant, it’s quite possible that they would get a job if it paid them to. But what’s true of an individual isn’t necessarily true for all. A big rise in employment requires not (just) effort from welfare claimants, but economic growth too. As Mr Adam writes, the elasticity of labour supply is "endogenous to the state of the economy."
Nor is it valid to extrapolate evidence of smallish changes in benefits (which are the basis for what estimates of work elasticities we have) to large-scale changes of the sort the CSJ proposes.
So, why does the CSJ sully what is in many ways an impressive report with such an implausible claim? I can think of two possibilities, both of which are depressing.
1. It is so wedded to the belief that getting off welfare into work is a matter of individual responsibility that it cannot see that what (might) be true for one needn‘t be true for all. Macro isn’t merely micro writ large.
2. The report’s recommendations - higher spending on support for the low-paid, at the expense of the better off - are hard to sell to the Tory party and the trash press. Its claim that hundreds of thousands will move off welfare is the sugar it needs to coat this bitter pill.
I don't agree with your view on the minimum wage, but clearly that would give the CSJ a simple get-out that would appeal to many Tories - scrap the minimum wage. So I wonder if you are reading too much into their 600,000 figure.
Posted by: Matthew | September 17, 2009 at 03:19 PM
Hmm... I can see politicians of any, all and no party being slightly unenthusiastic this year and the next three about a measure which promises to increase unemploymnet counts by 500,000 plus.
Posted by: Diversity | September 17, 2009 at 06:33 PM
I have been reading back and forth in that report since yesterday and it is surprisingly wimpy. Asking for the NMW to be scrapped would be a shoe-in, if he also had the guts to call for a radical reduction in total withdrawal rates.
While his idea of sort-of universal benefits (with little or no distinction out-of-work and in-work benefits) is a good step in the right direction, his vaunted 55% benefit withdrawal rate is in addition to normal PAYE (after a disregard to really confuse matters, of course) so it still comes to 69%.
If, instead of keeping 30% or less of £5.74, lower earners were keeping 50% of £4, that would increase the number of hours that employers are prepared to pay for and/or reduce the cost of employing people, everybody's happy.
To actually try and model this is making yourself a hostage to fortune, but that's no reason to assume that he is actually wrong.
Plus he doesn't realise how easy it would be to have universal benefits with a total 50% withdrawal rate, the PAYE system is already geared up for this. Workings here:
http://markwadsworth.blogspot.com/2009/09/centre-for-social-justice-part-2-k.html
Posted by: Mark Wadsworth | September 17, 2009 at 09:54 PM
The fallacy with all these straightline projections is that the available jobs are not where the unemployed (and the long term sick on benefits)are. Take the inner cities - maybe excepting London - and the former industrial and mining areas; just where is the work? How does someone who is long term unemployed get to what jobs exist, which are normally at a distance from where folk stay? If you take Iain Duncan-Smith's road to Damascus in Easterhouse, where's somewhere there going to find a job? Maybe central Glasgow 10 miles away but at basic wages.
The theory may be grand but the simple practicality falls over.
Posted by: Richard T | September 18, 2009 at 08:45 AM
Although the report is a hefty 369 pages long and contains a lot of detail, its underlying assumptions indeed look very shaky.
In the absence of the full model formulae it really does seem that CSJ assume that all that is required to create 600,000 new jobs is to tweak the benefits system to reduce financial 'disincentives to work' - AFAICS no account whatsoever is taken of labour market conditions and particularly of the regional factors Richard points out here.
Costs are always easier to calculate than benefits and I am also deeply sceptical of the benefits they have calculated to offset those very hefty (but probably still radically underpriced) costs.
Given how many of the reports authors are members of free market think tanks they also show a touchingly naive faith in the ability of bureaucracies to radically reform themselves out of existence in just three years.
For a start the IT implications are probably as if not more complex as those of the NHS IT system - and look where that is after nearly a decade of development.
However one very interesting aspect is the suggestion that rather than withdrawing benefits on starting work these would be recouped by employers through PAYE - and would thus be genuinely universal benefits (at least to anyone whose ever signed on).
This and the shift from 51 to 2 (why not just one?) benefits would constitute a big step towards a negative income tax/MGI - although the authors are rather coy about this.
Posted by: Roger | September 18, 2009 at 09:41 AM
If you are having an in-work benefit system then you need some form of minimum wage as a long-stop.
This is known as the Speenhamland problem, after an earlier experiment with wage supplementation for the poor (1790s to 1844). The net result was to drive down wages at the expense of raising taxes, to, according to the Poor Law report of 1844, a level at which no worker could survive without accepting Poor Law subsidy.
The converse problem is responsible for much of the complexity of Housing Benefit - Landlords have an incentive to raise rents to the maximum Government will subsidise.
The other main part of the 1844 report recapitulated regularly by both main parties is that the level of out of work benefits should be lower than the lowest income from a worker in the marketplace (otherwise known as Making Work Pay). Minimum wages work together with benefits to provide a socially acceptable minimum level of income.
Posted by: Paul | September 21, 2009 at 01:39 PM
" Macro isn’t merely micro writ large."
What?! When I stand up at a play or a sporting event I can see better. But you're telling me that if everyone stands up together we can't all see better?
Posted by: peter | September 24, 2009 at 09:58 PM