Policy, & the aspirations gap

Helping people into work can do more harm than good. This is the message of a new paper by Andrew Oswald and Richard Dorsett.

They studied the Employment Retention and Advancement programme introduced by the DWP in 2003. This offered extra training and temporary financial incentives to work longer hours to some randomly-chosen single parents but not to others. It was therefore a rare example of an RCT in economic policy.

But here's the thing. Oswald and Dorsett find that although this intervention raised the earnings of those in the ERA programe, it also led to "substantially lower well-being", even five years later. They conclude:

From a well-being perspective, randomly assigning in-work benefits appears to have hurt rather than helped.

The reason for this, they suggest, is that the temporary extra cash paid to participants raised their reference income and so increased aspirations, and the subsquent gap between high aspirations and actual incomes increased their financial worries and decreased happiness.

We should read this in the context of Andrew Clark's recent finding, that people do not adapt to poverty, because their aspirations don't fall sufficiently to match their low incomes. In raising aspirations, the ERA was therefore scratching an open wound.    

One inference from this could be that the best way to help the poor is not through temporary micro incentives but through macro policy - job creation and permanent redistribution.

But there might be a wider message. Given politicians' tendency to over-promise and under-deliver, perhaps almost all policies reduce well-being by creeating an aspirations gap. Personally, I think this would be too big an inference - because nobody seriously believes politicians' promises. But it is perhaps a warning that even quite benign policies might inadvertently prove to be harmful. 

Comments

[Luis Enrique]

Is "aspiration gap" another way of saying loss aversion? The experience of having been better off and becoming worse off is more unpleasant than having been worse off all along?

So this intervention could have raised well being if it had not included temporary financial incentives that produced a temporary income boost?

[chris]

@ Luis - in a way, yes. But it's only a loss relative to the temporary bonus. Relative to pre-ERA incomes, it's a gain. The fact that it's regarded as the former, not the latter, tells us that the reference frame matters a lot.

[Botzarelli]

Interesting. I wonder whether the inverse of this might not also apply. That is, if people are told that they are badly off and things are getting worse, over time they start to see this as normal so any improvements they actually experience seem better than they might otherwise have done. A bit like a slower acting version of the changes to the chocolate ration in 1984 - a painful reduction in an already meagre portion of a small pleasure followed by a small increase that is still below the original ration feels better than merely having a smaller reduction in the ration in the first place.

If true, this would have serious adverse implications for the "cost of living crisis" theme developed by Labour if there is continuing good economic news and individual optimism rises. That people might still be worse off than they were in 2010 or 2007 would be masked by the fact that they'd long ago stopped aspiring to the level of wealth they had in those increasingly distant days.

[Wpaul63]

Well but all our consumer driven economies create aspiration gaps - "Keep up with the Joneses" - in order to nudge those in their target markets to close the gap between themselves and those who are already ahead...those who have the resources (cash or credit) to indulge in the consumption of whatever good(s).
So maybe we should begin to address the behavioral side of this dynamic, no? Teach our children that their old ehr... vintage video game console is just fine , that they look great in their 3rd generation hand-me downs, that a rusty car and more cash in the pocket is more advantageous than spending on a new car that will instantly evaporate 30% as soon as it is first driven, etc, etc.

[Dave Timoney]

Context is all (hence the weirdness of a chocolate ration in 1984 - sic). Teaching your children that hand-me-down clothes and furniture are fine, and a rusty Land Rover is ideal, has certainly worked for the upper classes over the years.

There is a whiff of a class agenda here, specifically the idea that raising the expectations of the poor is inadvertantly cruel. Oswald and Dorsett's citing of the Cambridge-Somerville Youth Study from the 1930s is notable in that they select this one interpretation ("don't get their hopes up") and ignore others. I imagine IDS will not look unkindly on this paper.

An alternative interpretation is that income volatility is a bugger. Occam's Razor would suggest you don't need to bring aspirations into it.

[Andrew]

Wow I'm impressed!

[Keith]

I suspect any reading of literature over the years would make the same point better. The theme of the wealthy family fallen on hard times being worse off than the perpetually poor is a well trodden path.

Wittgenstein probably had this idea in mind when he gave all his wealth to his sister. Rather than a randomly selected poor family.

I think Politicians may very well be reducing happiness as some foolish people probably do believe the spin. The Tory party can hit them on the head repeatedly but they still vote Tory. And you thought masochism was illegal!

[Jim Birch]

If all people didn't believe politicians spin they would stop. AFAICS there's a chunk of the voting population who believe what politicians say (and thus are incidentally screwing the political system up for the relatively sane rest of us.) Politicians are almost by definition people who can create believable narratives out of any crap, the embodiment of Argumentative Theory:

http://www.edge.org/conversation/the-argumentative-theory

[rogerh]

Were members of the control group revisited as part of this survey? - from my quick read it seems not. Given the rather small statistical effects measured I should think a similar result could be extracted from the control group.

So a definite maybe.

[Andrew]

Rogerh - I think they were. Though it is difficult to interpret the methods (what is wrong with a section called "Methods", eh? economists? rather than interspersing it between discussion and introduction) the main analysis is on a comparison between study and control group survey answers at 2 and 5 years post-intervention. They don't mention blinding the survey askers, or a placebo intervention, which would have been doable, but you can't have everything.

[rogerh]

@Andrew
You are right - but the report is a bit confusing re method. I had thought they had re-run a survey over the original actors and then mashed the data. It seems not, what seems to have happened is a re-hash of the original data - a lot less interesting imho.