Cash isn't the missing ingredient.

From Study May Undercut Idea That Cash Payments to Poor Families Help Child Development by Jason DeParle.  The subheading is Rigorous new research appears to show that monthly checks intended to help disadvantaged children did little for their well-being, adding a new element to a dispute over expanded government aid.

If the government wants poor children to thrive, it should give their parents money. That simple idea has propelled an avid movement to send low-income families regular payments with no strings attached.

Significant but indirect evidence has suggested that unconditional cash aid would help children flourish. But now a rigorous experiment, in a more direct test, found that years of monthly payments did nothing to boost children’s well-being, a result that defied researchers’ predictions and could weaken the case for income guarantees.

After four years of payments, children whose parents received $333 a month from the experiment fared no better than similar children without that help, the study found. They were no more likely to develop language skills, avoid behavioral problems or developmental delays, demonstrate executive function or exhibit brain activity associated with cognitive development.

“I was very surprised — we were all very surprised,” said Greg J. Duncan, an economist at the University of California, Irvine and one of six researchers who led the study, called Baby’s First Years. “The money did not make a difference.”

Lots of substance in the reporting.  It does appear to be rigorously designed, 1,000 participants, randomly assigned treatments, pre-published measures, multiple locations (four, New York, New Orleans, greater Omaha and Minneapolis-St. Paul), etc.  

The general explanation in the reporting (and in the NYT comments) is that $333 a month was insufficient to make a difference.  DeParle does a good job, though, of conveying how perplexing these results are to both the researchers and to those involved in the issue.  

The researchers specified in advance seven measures on which they thought children in high-cash families would outperform the others. But after four years they found no group differences on any of the yardsticks, which aimed for a comprehensive look at child development.

Children in the families getting the higher cash payments did no better on tests of vocabulary, executive function, pre-literacy skills or spatial perception. Their mothers did not rank them more highly on assessments of social and emotional behavior. And they were no more likely than the children in the low-cash group to avoid chronic health conditions like asthma.

Mothers in the high-cash group did spend about 5 percent more time on learning and enrichment activities, such as reading or playing with their children. They also spent about $68 a month more than the low-cash mothers on child-related goods, like toys, books and clothing.

It is possible that with a longer time lens, after more time has elapsed, that later positive impacts will emerge which are not currently detectable.  "The payments continued for more than six years, and future analyses will examine the longer-range effect."  

What was striking to me is that this follows the pattern of so much social policy.  An idea makes intuitive sense, we start funding experimental projects, then deploy more broadly and only eventually start measuring impacts and results.  And only much later do we start designing rigorous reviews.  

And every time we get around to the rigorous review, the sought for benefits disappear.  They don't exist.  Think of Head Start and its variations.  Thirty or forty years now, widely deployed in many states, costing billions of dollars and with much support.  But after many years, we eventually discovered that in every variation in all states - Head Start made no difference to educational outcomes.  

Unrestricted cash support would seem like an obvious and easy solution.  And it appears not to work. 

Extra and earlier educational support would seem like an obvious and easy solution.  And it appears not to work. 

The more rigorous the review, the less evidence of benefit there is.  

Systemic constraints don't seem to be the root cause they have been made out to be, pushing the discussion more towards values, behaviors, and capabilities.  I wonder if anyone has ever looked at Gregory Clark's work as a source for ideas which might inform social policy today and which might have a demonstrable base of what works.