Thursday, December 22, 2011

Nudge - By Definition

By Pelle Guldborg Hansen & Andreas Maaløe Jespersen

Confusing confusion

"The definition of nudge is vague and more work should be done on clarifying this before we can consider..." 

"Nudging is basically about controlling incentives - penalties and rewards..."

These are just some of the remarks we are often confronted with - even by Academics, including people who sit on boards and committees who's function is to hand out money and thus invest and direct future research.

What is most disturbing about this, isn't that these remarks are plain wrong. Rather, it's that they seem to result from people confusing their own confusion with regard to some simple facts and concepts that may quite easily be checked.

Nudge - by definition
On page 6 in both the US and UK version of Thaler & Sunstein's Nudge: Improving Decisions about Health, Wealth & Happiness (2008) it is written that:  
A nudge, as we will use the term, is any aspect of the choice architecture that alters people’s behavior in a predictable way without forbidding any options or significantly changing their economic incentives. To count as a mere nudge, the intervention must be easy and cheap to avoid.
Now superficially, this may seem as a straight-forward definition liable for vagueness. However, what one has to remember, that this definition is coined against the background of behavioral economics. 

If you don't have the time or patience to sit down and work through the literature in this field, this is the rule of thumb: any intervention that would influence an unbounded, unrestricted rational agent, is not a nudge. Hence, a nudge does not invoke incentives - positive or negative.

At a more abstract level, a direct consequence of this is that 'a nudge' is defined as effecting a deviation from mathematically well-defined baseline models. Hence, saying that the definition of nudge is vague is straight-forwardly wrong.

A problem with the original definition   
However, the attentive reader will by now have discovered a flaw in Thaler & Sunstein's original definition (we think it is an unfortunate simplification). 

The payoff-functions of rational agents are affected by other things than mere economic variables. For instance, the expectation of cake, electric shock or social ostracism. Hence, restricting the definition to economic incentives seems wrong. For this reason we usually adopt the definition provided by Hausman & Welch (2010):
Nudges are ways of influencing choice without limiting the choice set or making alternatives appreciably more costly in terms of time, trouble, social sanctions, and so forth. They are called for because of flaws in individual decision-making, and they work by making use of those flaws. 
(Hausman & Welch 2010:126)
God and evil nudges
However, while Thaler & Sunstein as well as Hausman & Welch extends the notion of 'a nudge' to cover any attempt of influencing behavior - well- or ill-intended (in fact, Thaler & Sunstein's notion is even broader) - we suggest limiting the notion to only well-intended for several important conceptual reasons of conistency.*

Nudge - US paperback version 

Nudge US hardcover version
Nudge UK paperback version

This post draws on points from a forthcoming journal article of ours. Thus, if you intend to cite or use points from the above, please contact us.  

No comments:

Post a Comment