Nudge: Improving Decisions About Health, Wealth, and Happiness (Hardcover)
by Richard H. Thaler and Cass R. Sunstein (2008)

# Hardcover: 293 pages
# Publisher: Yale University Press (April 8, 2008)
# Language: English
# ISBN-10: 0300122233
# ISBN-13: 978-0300122237

############################

BEHAVIORAL behavioral economics
By  Paul Brandon
Emeritus Professor of Psychology
Minnesota State University,  Mankato

A note:  this is not a general review of the book, but rather an analysis of the ways in which the system presented in the book is congruent with that of another system: behavior analysis.  Were it a general review I would take issue with many of the suggestions made by the authors.

`Nudge' appears to be mainstream behavioral economics; economists attempting to make their analysis more realistic by replacing the completely rational `Economic Man' of classical economic analysis with a more realistic version derived from the decision theoretical work of Tversky and Kahneman.  Nothing theoretically radical here.  In fact, its roots can be traced back to John B. Watson (yes, that Watson), the founder of advertising psychology.

The authors' point is that classical economics is in fact a poor guide to action if we want to be effect in changing what people do.  Instead they propose some empirically supported strategies for behavior change which they derive from some recent work in the area of decision theory, although (as I will point out) many of the strategies they propose can be more parsimoniously derived from other conceptual systems.

In fact, this is a good example of parallel evolution.
Thaler and Sunstein are responding to the same environmental situations as behaviorists do, and it is not surprising that the common set of contingencies (observed patterns of human behavior) occasion similar verbal behavior.  What has surprised me as I read the book is that my typical response is not `how dumb', but `of course.'  They suggest generally the same courses of action that I would as a behaviorist; I (and more so behavioral behavioral economists) might suggest some detail differences, but by and large what they suggest makes good behavioral sense.
Of course, I differ with some of their recommendations, and would argue that they have cherry-picked the data that these recommendations in areas like education are based on.

As I've indicated, there is another field of behavioral economics.
This is a group of behavior analysts (Skinnerian or Radical Behaviorism) who have incorporated some of the concepts of economics such as value delay discounting and demand elasticity into the terms of behavior analysis, and can be looked at as an application of proven behavioral principles.  Specifically, they avoid coercion by emphasizing existing stimulus control rather than programmed reinforcement or punishment.  Thaler and Sunstein's distinction between Automatic and Reflective Systems maps easily onto the behavioral distinction between contingency and rule governed behavior: behavior governed by direct contact with outcomes versus behavior controlled by learned verbal rules.

A 'nudge' itself sounds very much like what a behaviorist would call a prompt -- the minimal stimulus necessary to produce a behavior that is already likely to occur.  Much of their discussion seems to concern how to prompt behavior rather than compelling it with obvious rewards and punishments -- good behavioral practice.

One might say that it is a plea to identify and use existing stimulus control (the way we learn to behavior in a given way in a given situation) as well as some descriptions of common patterns of this sort of influence on our behavior.  For example, their RECAP procedure (a simple and readable breakdown of costs) makes the costs and benefits of things like mortgages and credit cards more discriminable.

Much of their analysis can be put into a structure of concurrent chains.
They identify two (occasionally more) alternative behavioral sequences involved in some choice situations (such as whether or not to pollute), and then discuss ways in which the contingencies of the competing chains might be altered.  This could be as simple as changing the controlling stimuli to make the desired behavior more likely, or if necessary to decrease its response cost, or as a last resort, to more directly modify the consequences in terms of reinforcing or punishing outcomes.
The fact that a chain is always involved; people always have a choice in the sense of available alternative behaviors, allows them to maintain their `paternalistic libertarian' stance.  Thus, they support the use of cap-and-trade systems as an approach to pollution control since it would embed some positive reinforcement contingencies which would result in a decrease in the choice of polluting industrial practices, rather than simply mandating a target (they use the phrase "command-and-control") for pollution reduction.
This sort of distinction is sometimes more apparent than real; a choice between two possible behaviors is present in both cap-and-trade and direct regulation.
In the one case, the choice is between paying a tax for polluting on the one hand, and not polluting on the other.
In the case of direct regulation, the choice is between polluting and suffering some penalty for violating the regulation, versus not polluting.
Thus, the difference between the two situations is the nature of the contingencies, not the presence or absence of choice.

As a related point, they talk about the problems involved with too many choices (Barry Schwartz has written on this, although they don't cite him).  Medicare D plans are a good example: it's almost impossible to make a good choice given the number and complexity of alternatives.


In general, the book's analysis is very consistent with the predictions and recommendations that a behavior analyst would make, to the point where I wonder if the ghost of Israel Goldiamond is still roaming the halls of the University of Chicago where the authors work.
For example, they've discovered the old behavioral tactic of having a person surrender control of a sum of money to someone else.  If some goal is met the money is returned; if not it goes to someplace aversive to the client.

One of the strengths of the book's approach to human behavior is the acknowledgment that we make choices for a reason -- and that these reasons can often be identified.  There is no such thing as purely autonomous choice independent of outside influences.  Therefore, our choice is not between influence and lack of influence, but whether we will study and use these influences to better our lives.  Ignorance is not bliss, nor is it healthy.
Thaler and Sunstein sugar coat this by using the label `libertarian paternalism', but underlying their argument is the assumption that behavior is not ultimately autonomous, and that the job of `choice architects' (those who manage behavior in any sense) is to identify the aspects of the environment that control actions of interest, and to change behavior by changing the relevant aspects of the environment.

The book follows a basic pattern:
The authors first describe how a rational and knowledgeable Economic Person (an Econ) would approach some decision situation, then document how real people (Humans) actually behave, and finally describe some way of restructuring the situation (a Nudge) that would cause Humans to behave in a way more to their long term benefit.

Some examples of the behavioral processes involved:

Contingency traps
One behavioral concept that they appear to be developing is that of the contingency trap.
This is the observation that immediate consequences (reinforcers) are more effective than delayed ones.  Many human problems are due to the fact that a given action usually has more than one consequence.  If the immediate consequence is reinforcing, but the long term outcome is harmful, we have a contingency trap.  Health risks like overeating and smoking fit into this category.  Again, the problem is how to rearrange the environment (particularly the social environment) to provide prompts for behaviors with delayed positive outcomes rather than immediate ones.

Rule governed behavior
On of the ways out of a contingency trap is to bring behavior under the control of verbal rules.
Often this works by making contingencies more obvious.
In the Nudge context, this is described by terms such as consumer education, getting people to forgo immediate reinforcers that have more delayed and diffuse long-term effects.  They apply this to a free market (cap-and-trade) approach to pollution, to overcome the political difficulties involved in getting legislation enacted.

Social Influence
Not surprisingly, the authors devote a chapter to this topic; again it's pretty straightforward.
In behavioral terms (not theirs) social influence can be divided into two categories: modeling (doing what one sees others doing, and getting reinforced for doing) and social reinforcement (direct peer pressure; approval or disapproval of one's actions).  An additional process is referred to as `priming': prompting an initial step in the sequence of actions necessary to achieve a goal.
Despite their `libertarian' stance, they don't seem to find peer pressure unacceptable as a way to change behavior, although their preference is for various forms of prompting.

Response cost
T & S have discovered that the cost of doing something affects the likelihood of taking action.
In particular, people tend to follow the Law of Least Effort (this concept goes back to Thorndike).  Among other things, this results in behavioral `inertia' since doing nothing (no change) requires less effort than doing something (such as changing one's asset allocations)
Thus, in discussing `opt-in' vs `opt-out' methods of having people enroll in programs such as retirement savings, they recommend structuring the programs so that the default action is enrolling in a program, with a positive action required to opt out of it.  The same analysis is applied to the task of increasing organ donations.
Another point:  the default payment options on credit cards are a minimum payment that maximizes the company's interest income.  Their alternative: make it an equal effort forced choice between minimum payment and complete balance payment.


A good example of the way in which they apply these principles is the Save More Tomorrow program developed by Thaler and Schlomo Benartzi.
This program is based on the principle of making a verbal commitment now (a behavior with a low response cost) to do something in the future which has a high response cost.  This is very similar to Hayes’ ACT (Acceptance and Commitment Therapy).
In this case, they start with a standard savings plan where people sign up for an automatic periodic transfer of funds to a savings account.
The added detail is a commitment to a gradual increase in the size of the deduction (some echos of progressive ratio reinforcement schedules here), thus reducing the aversiveness of a large initial payment.
They also talk about ‘inertia’ as a factor both in maintaining behavior and resisting change; reminiscent of Nevin’s Behavioral Momentum Theory, which might provide some additional suggestions for developing this program.

            =======

Some of their predictions seem a bit self contradictory.
While making a big deal of the difference between "Humans" and "Econs", they seem to assume that simply providing information about costs and benefits of behavior will be enough to change people's choices.
A good example of this is their analysis of medical malpractice lawsuits.  They seem to feel that people might trade off their right to sue in exchange for lower medical fees.  This seems to assume that Humans are Econs and make rational choices based on economic considerations.
An alternative might be the well known psychological principle of minimizing the maximum loss (a mini-max strategy).  People will trade off the lowest long-term cost in exchange for minimizing that maximum possible loss.  That is why few people will choose to self insure, even if that results in the lowest predicted long term cost.

Finally, the whole idea of `libertarianism' (paternalistic, maternalistic, or otherwise) seems to contradict the deterministic assumptions of behaviorism.
From the point of view of the behaviorist, freedom is just another word (see Baum's text for a good analysis of the uses of the term).  People do not behave freely in the sense of being autonomous actors, but rather their behavior is determined by the interaction between genetics, experience, and consequences.  Our goal is not to maximize freedom (except in the limited sense of absence of aversive control), but rather to promote the use of contingencies which will maximize human welfare.

            =======

My final question is, therefore, do we as Behavior Analysts have a unique contribution to make in the sense of suggesting solutions that are basically and profoundly different from those suggested by non-behaviorists, are do we (at least in the case of social engineering) tend to converge on the same basic solutions?
In the latter case, our contribution would not be in the uniqueness of our analysis but in the efficiency of its implementation.  As I have tried to demonstrate, a non-behavioral analysis based on some nonsystematic decision principles and some canny observations can be more parsimoniously analyzed in behavioral terms.  In addition, we might avoid the consequences of pursuing some of the false distinctions implicit in the concept of libertarianism.
This, in turn, should lead to a more effective implementation.

#################

Maximizing versus satisficing: Happiness is a matter of choice.
Schwartz, Barry; Ward, Andrew; Monterosso, John; Lyubomirsky, Sonja; White, Katherine; Lehman, Darrin R.
Journal of Personality and Social Psychology. Vol 83(5), Nov 2002, 1178-1197.

http://www.amazon.com/review/R3FC6CUFXEZTWC/ref=cm_cr_old_cmt_rd