Iain Duncan Smith wants to shift the definition of child poverty from one based upon low incomes to one based on educational attainment, worklessness and addiction.

What's at stake here is, as Amelia Gentleman points out, an ideological issue: is poverty due to individual failings or to the structure of society?

Of course, some parents of poor children are feckless workshy druggies. But in a population of 64.6 million people, pretty much anything is true of someone.

Two big facts, however, suggest that the link between child poverty and parental failure is weak. One comes from the DWP's own report:

Children in families where at least one adult was in work made up around 64 per cent of all children in low income [before housing costs] in 2013/14 (p46 of this pdf).

Think what it means to be in work. It means you've impressed an employer sufficiently to get hired, and you are managing to turn up roughly on time most days. You have, in short, got your shit together. And yet you're still unable to get your family out of relative poverty.

Secondly, Andrew Dickerson and Gurleen Popli point out that there is zero correlation between material child poverty and a measure of parental involvement based upon facts such as whether parents read to their children or given them regular meal times and bed times. There is, therefore, no link between bad parenting (on this measure) and material poverty.

These two facts suggest another, bigger reason for child poverty. Quite simply, it has become harder for less skilled people to provide for their families. For someone at the 25th percentile of weekly wages, real wages are 6% lower now than they were as far back as 1997*.

In this sense, obstinately high child poverty has its roots in developments in the labour market. Whether it be because of mass unemployment**, deindustialization, the offshoring of low-skilled work, technical change or whatever, the fact is that things have gotten tougher for what used to be called the respectable poor in the last 40-odd years.

It is this fact that Duncan Smith seems to want to gloss over. From the point of view of the ruling class, it is better to question the character of the poor than the health of capitalism. And, sadly, I fear he might succeed in this aim: the egocentric bias means many voters like to think well of themselves and hence less well of others. There will therefore be an audience for slanders against the poor.

However, facts are facts however you try to change the language. You can define your cow as a horse but it still won't win the Derby.

* According to ASHE, gross weekly earnings for the 25th percentile have risen 53.6% since 1997 (when their data begins) whilst the RPI has risen 63.6%: I suspect RPI is a better measure than CPI as it includes housing costs.

** There are 1.81 million officially unemployed, plus 1.3m part-timer workers who want full-time work, plus 2.34m economically "inactive" who want a job. That's a total of 5.45 million.

Another thing: insofar as some poor parents are lazy, there's a question of endogeneity: is the laziness a cause of poverty or is it instead an adaptive preference - a response to their belief that they can't find work?