Implications of behavioural economics for regulatory reform in New Zealand

Abstract

Sapere Research Group and the New Zealand Law Foundation.

Neoclassical economics, which assumes rationality and self-interest, has helped analyse many regulations. But a growing body of evidence about judgements, decisions, and preferences casts doubt on the applicability of these assumptions. Drawing on this evidence, behavioural economics can now supplement neoclassical economics in regulatory analysis.

Full text

Implicationsofbehavioural
economicsforregulatoryreformin
NewZealand

PublishedwiththeassistanceoftheNewZealandLawFoundation


TimothyIrwin
December2010

2
Implications of behavioural economics for
regulatory reform in New Zealand
Published with the assistance of the New Zealand Law Foundation
Timothy Irwin, Sapere Research Group Limited1
December 2010

1 This paper has benefited from comments from Glenn Boyle, Simon Kemp, Graham
Scott, Richard Tooth, and Sally Wyatt, none of whom should be presumed to agree
with its conclusions.

3
Summary
Neoclassical economics, which assumes rationality and self-interest, has
helped analyse many regulations. But a growing body of evidence about
judgements, decisions, and preferences casts doubt on the applicability of
these assumptions. Drawing on this evidence, behavioural economics can
now supplement neoclassical economics in regulatory analysis.
Many regulations in New Zealand are more easily reconciled with
behavioural economics than with neoclassical economics. Delays preventing
impulsive decisions are part marriage and divorce laws, and the right to
change one’s mind is part of the laws governing borrowing and door-to-door
sales. Problem gamblers can ban themselves from casinos, and employees
are automatically enrolled in KiwiSaver. Simple and sometimes emotionally
powerful information disclosure is used to discourage smoking and guide
investment decisions. Behavioural economics also suggests other areas
where carefully chosen default options, simpler and more powerful
information disclosure, or devices that facilitate self-control may be useful.
At the same time, it strengthens concerns that regulations may fail to
achieve their intended effects and doesn’t change the principle that
regulatory decisions should be made only after careful weighing of costs
and benefits.
Yet behavioural economics also raises problems for the weighing of costs
and benefits. Traditional cost–benefit analysis measures costs and benefits
by reference to people’s choices, but to the extent that those choices are
internally inconsistent or premised on mistaken beliefs, choice is a
questionable basis for policy. In response, some researchers propose trying
to separate mistaken and inconsistent choices from those that reflect
preferences, while others propose basing cost–benefit analysis on measures
of experienced well-being, and still others propose using traditional cost–
benefit analysis despite the flaws in its foundations.
Because behavioural economics underscores problems with instinctive
judgements, it helps justify rules that encourage deliberation before
regulations are made. It also suggests ways of improving regulatory
decisions, such as the use of de-biasing training, regulatory premortems,
early and informal reviews of regulatory proposals, and checklists of
questions for assessing proposed paternalistic regulations. Lastly, by
undermining neoclassical economic theory without offering an alternative of
similar scope, it adds weight to calls for more empirical testing of proposed
and existing regulations.

4
Caveat emptor, we declare, let the buyer beware. This is a
policy that presupposes that the buyer is rational enough to
see through the blandishments of the seller, but since we
know better than to believe this myth taken neat, we go on
to endorse a policy of informed consent, prescribing the
explicit representation in clear language of all the relevant
conditions for one agreement or another. Then we also
recognize that such policies are subject to extensive
evasion—the fine-print ploy, the impressive-sounding
gobbledygook—so we may go on to prescribe still further
exercises in spoon-feeding the information to the hapless
consumer. At what point have we abandoned the myth of
‘consenting adults’ in our ‘infantilizing’ of the citizenry?—
Daniel Dennett (2003, 270).
Introduction
1. The traditional approach to economics, which assumes that
people are rational and generally self-interested, has proved fruitful
for the analysis of regulation2. For example, the theory of
externalities and the theory of transaction costs help analysts
examine the problem of pollution and assess whether regulation is
desirable. Advances in the economics of information help assess
whether it is better to tackle pollution by means of taxes or quotas
and, when quotas are preferred, to design markets in those quotas.
Likewise, the theory of games and the theory of industrial
organization make it possible to compare monopoly, oligopoly,
and competition, and to estimate the welfare effects of price
control and restrictions on mergers and acquisitions.
2 ‘Regulation’ is used here in the broad sense favoured by economists to
refer not only to the instruments that lawyers call ‘regulations’ but also to
Acts of Parliament and other rules made by government. See Scott et al.
(2009, 32) and The Regulations (Disallowance) Act 1989, section 2.

5
2. Yet there are problems with the assumptions of rationality
and self-interest that limit the value of neoclassical economics3 in
the analysis of regulation. There is now a large body of evidence
that shows that we are sometimes predictably irrational. It’s not
just that we are imperfectly informed (which is after all rational,
given the costs of information), but that we are biased and
inconsistent. Psychologists have demonstrated, for example, that
we make predictably poor judgments about risks: that we
systematically overestimate the probabilities of some kinds of
events and underestimate those of others. They have also shown
that our decisions are influenced by things that shouldn’t matter.
For example, we can make one choice if we are encouraged to
evaluate our options as gains relative to a bad state of affairs, and
another choice if we are encouraged to evaluate the same options
as losses relative to a better state. Psychologists and economists
have also found that our preferences change in a way that makes us
inconsistent over time. We have a bias toward the present, which
means that even if we want to save and diet we prefer to start
tomorrow.
3. Nor are we always self-interested. We are sometimes more
cooperative and sometimes more vengeful than we would be if we
were purely self-interested. We also care much more about keeping
up with the Jones. In general, our preferences are often ‘other-
regarding’.
3 The best term for this kind of economics is unclear. ‘Rational-choice
economics’ (Posner 1998, 1551) has the virtue of being clear, but it is not
common. ‘Standard’ or ‘mainstream’ are other possibilities, but what is
standard or mainstream changes over time. ‘Neoclassical economics’ is
sometimes used to refer to economic analysis that assumes perfect
information as well as rationality and self-interest. It is used here more
inclusively to refer to economics that allows for imperfect information but
assumes rationality and self-interest.

6
4. In response to this evidence, a new field of research called
‘behavioural economics’4 has emerged, which studies the ways in
which people deviate from rationality and simple self-interest and
investigates the implications of these deviations for markets and
public policy. The field, which draws mainly on psychology but
also on sociology and neurology, is controversial, and how
successful it will be remains unclear. Although experiments have
demonstrated that behaviour sometimes deviates systematically
from rationality and self-interest, it is unclear how large or
widespread or enduring the deviations are in ordinary economic
settings. It is also unclear how well behavioural economists will be
able to explain people’s behaviour with models that retain some of
the simplicity and scope of neoclassical economics. Nevertheless,
behavioural economics is by now an established part of
economics.5
5. In neoclassical economics, regulation serves to fix market
failures. For example, it can reduce air pollution and other negative
externalities when transaction costs mean that people cannot easily
negotiate among themselves to solve the problem. But in
neoclassical economics regulation is not needed to protect people
from their own decision-making flaws or from firms that exploit
those flaws. Consumers may make mistakes, but the mistakes are
4 The name ‘behavioural economics’ is problematic, because neoclassical
economics also seeks to explain behaviour. An alternative that is sometime
used is ‘psychology and economics’. Economic psychology is a
longstanding closely related field of psychology (e.g., Earl and Kemp 2002).
5 Camerer et al. (2003) note that behavioural economics may prove to be a
natural step in the development of economics. The simplest economic
assumptions are that markets are perfectly competitive and that people are
perfectly informed and perfectly rational. Much progress has already been
made exploring imperfect competition and imperfect information. The next
frontier may be imperfect rationality.

7
random, not systematic, and tend to be corrected over time. There
are market failures and government failures, that is, but no
‘consumer failures’. By contrast, behavioural economics allows for
the possibility that paternalistic regulations may be valuable, as
well as identifying new sources of externality.
6. On the one hand, behavioural economic provides support
for existing regulations whose rationale is apparently paternalism,
such as mandatory delays before marriages and divorces, and
cooling-off periods during which consumers can change their
minds about loans or purchases from door-to-door salesmen.6 7 It
also strengthens the case for regulations whose rationale is
probably a mix of paternalism and concern for negative
externalities, such as requirements that motorcyclists wear helmets
6 See KiwiSaver Act 2006 (section 9), Credit Contracts and Consumer
Finance Act (2003), (section 27), Door to Door Sales Act 1967 (section 7),
Marriage Act 1995 (section 24), and Family Proceedings Act 1980 (section
39).
7 It is difficult to know exactly what the rationale is for many consumer-
protection regulations in New Zealand. A recent textbook on consumer law
in New Zealand (Bevan, Dugan, and Grainer 2009), for example, asserts that
the need for such regulations is obvious, but doesn’t argue for them (e.g., pp.
21, 44, 330, 481). An earlier textbook on the same subject (Tokeley 2000,
ch. 1) does provide an argument for consumer protection, but the argument
relies on neoclassical economics and is unsatisfactory. Consumer protection
is said to be necessary because consumers and large firms have unequal
bargaining power, which in turns stems from imperfect competition and
imperfect information. Yet imperfect competition is neither necessary nor
sufficient justification for the consumer-protection laws that the book
discusses. It is true that imperfect competition can justify price control and
restrictions on mergers and acquisitions, but these regulations are beyond the
scope of the book. Moreover, consumer-protection problems may arise even
in markets with numerous suppliers: a study in the United States concluded
that credit-card holders were getting a poor deal despite the presence of
some 4,000 suppliers (Ausubel 1991). Imperfect information may justify
some kinds of information-disclosure regulation, though not necessarily the
kinds that are designed to protect consumers, but it does not justify
mandatory cooling-off periods.

8
and restrictions on the sale and marketing of cigarettes and alcohol.
To take a different kind of example, behavioural-economic
analysis of preferences for relative position helps justify
regulations that slow the ‘arms race’ among consumers, such as
those that prescribe minimum periods of annual leave.
7. On the other hand, behavioural economics also offers ideas
for modifying regulations. If people are poor at understanding
complex information, information-disclosure regulations should
probably prescribe very simple disclosures and possibly
emotionally powerful ones. If people left to themselves save or
insure too little, perhaps regulation should make some savings and
insurance compulsory. If present bias causes people to behave
impulsively in ways they come to regret, perhaps regulation should
encourage more cooling-off periods between impulse and
execution, such as delays between an initial decision to gamble at a
casino and the ability to do so.
8. Behavioural economics may also help make non-
paternalistic regulations more effective. Neoclassical analyses of
regulation emphasize the role of taxes, quotas, and minimum or
maximum prices. The surprising effectiveness of ‘nudges’—small
changes that affect behaviour without altering prices or imposing
quotas—suggests that more use might be made of advertising,
changes in defaults, and other soft approaches.
9. Although behavioural economics helps justify some
regulations, its implications are far from settled. It doesn’t
necessarily imply more regulation. First, officials, politicians,
regulators, judges, and juries are undoubtedly affected by biases,
so behavioural economics provides new reasons to be concerned
about regulatory failure—that is, about the risk that regulation will

9
fail in practice to make things better even though the right
regulation properly enforced would. Second, to the extent that
nudges and other soft approaches are effective, there is less need
for regulatory coercion. Third, other-regarding preferences can
sustain social norms that generate good behaviour even in the
absence of legal rules requiring that behaviour. Indeed, behavioural
economics has brought to light a new kind of regulatory failure:
when people are inclined to cooperate to solve problems, the
imposition of imperfect regulations can sometimes make matters
worse. Finally, because behavioural economics suggests that
consumers may demand too little of some services, such as savings
products, it implies that governments should be wary about
reducing the supply of these products by imposing regulations that
are costly to comply with.
10. Because decisions made after deliberation are likely to be
less biased than gut reactions, behavioural economics helps justify
procedural rules that prevent hasty regulation, such as
requirements for regulatory impact analysis. It also suggests ways
in which those procedures might be improved, including the use of
checklists, and regulatory premortems, and early informal
regulatory reviews.
11. Lastly, behavioural economics calls into question
traditional cost–benefit analysis, the standard economic tool for
judging whether the outcomes produced by a regulation are
desirable. Traditional cost–benefit analysis assumes that
regulations are good to the extent that they generate outcomes that
people would choose. For example, antipollution regulation is
justified if, in a hypothetical world without transaction costs, those
who suffer from the pollution would be willing to pay polluters
enough to persuade them to stop polluting. But if choices are

10
sometimes poorly informed and sometimes simply incoherent,
standard cost–benefit analysis loses its rationale. Judgements about
policies may need to be based on something else.
12. This paper investigates the implications of behavioural
economics for regulation in New Zealand. It argues that
behavioural economics should be taken seriously by those involved
in the design of regulation (section 1) and gives examples of
existing regulations and possible regulatory changes that find some
support in behavioural economics (section 2). It doesn’t, however,
attempt the comprehensive analysis that would be needed to make
recommendations. It also explores how regulations should be
judged if people cannot be assumed to be rational and considers
whether paternalism is sometimes justified (section 3). Lastly, it
argues that behavioural economics supports rules that require
regulatory impact analysis and it makes suggestions for improving
the process of analysing regulation (section 4).
The case for taking behavioural economics seriously
13. To begin with, it’s important understand how behavioural
and neoclassical economics differ. In particular, what exactly are
the assumptions of rationality and self-interest?
14. Rationality implies first that our preferences are internally
consistent. For example, if we prefer A over B and B over C then
we prefer A over C.8 To give the assumption of internal
8 Preferences are also assumed to be complete, which means that for any two
options a decision maker is either indifferent between the options or prefers
one to the other. The link between rationality and completeness and
transitivity is strong enough that the term ‘rational’ has been used to refer to
(continued)

11
consistency some bite, it is also assumed that preferences are stable
over time. In the extreme, people are assumed to have constant
lifetime preferences over their lifetime consumption paths
(Bernheim and Rangel 2007).
15. Rationality doesn’t imply having perfect information, but it
does imply making good use of available information. When faced
with uncertainty, for example, people are assumed to correctly
update their estimates of probabilities in the light of new evidence.9
In strategic situations, they are assumed to be able to put
themselves in others’ shoes and to carry out chains of reasoning of
the form: he knows that I know that he knows that I know ….10
16. People acting according to internally consistent preferences
can be described as maximizing their utility, which might suggest
that rationality implies the pursuit of self-interest. But this is a
any mathematical relation that is complete and transitive (see Corbae,
Stinchcombe, and Zeman 2009, 39).
9 In modeling decisions under uncertainty, probability estimates are assumed
to be updated according to Bayes rule, which says that if the prior
probability of an event A is denoted by Pr(A), the probability conditional on
some new evidence E is given by
Pr( )Pr( )
Pr( ) Pr( )
EA A
AE E
,
where Pr( )
A
Eis the probability of A conditional on E, Pr( )EAis the
probability of E conditional on A, andPr( )Eis the unconditional probability
of the evidence E.
10 Precisely defining rationality is difficult, which makes it hard to draw a
precise boundary between neoclassical and behavioural economics. Indeed,
some work by economists in the neoclassical tradition is motivated by a
view that some people are ‘less rational and calculating’ than others (Salop
and Stiglitz 1977, 493).

12
confusion caused by the word ‘utility’.11 In this technical sense of
the word, the most altruistic person can be described as
maximizing her utility. Neoclassical economic theory is consistent
with any kind of preferences, so long as they are internally
consistent. 12
17. Yet many critics believe that neoclassical economists
assume that people are selfish. The Ministry of Economic
Development (2006, 4–5) has written that ‘The traditional (core)
neoclassical assumptions include’ an assumption that ‘[a]ll
individuals act in complete self-interest to maximise their own
welfare and their decisions are not influenced by the welfare of
others.’13 This isn’t true, but it isn’t far from the mark for many
applications of the theory, in which simple assumptions must be
made about people’s preferences in order to derive predictions.14
11 A person whose preferences are complete and transitive can be
represented as maximizing a mathematical function. That function is
conventionally called a ‘utility function’, but the name of the function
implies nothing about the content of the preferences.
12 As David Hume said, ‘’Tis not contrary to reason for me to chuse my total
ruin, to prevent the least uneasiness of... [a] person unknown to me (1739–
1740). And as two game theorists recently wrote, ‘Our methodology remains
unchanged whether our players are Attila the Hun or St Francis of Assisi.
We simply recognize that they have different tastes by writing different
numbers in their payoff matrices’ (Binmore and Shaked 2010, 98).
13 Henrich et al. (2004, 8) refer to the existence of a ‘selfishness axiom’ in
neoclassical economics.
14 In the words of the distinguished neoclassical economist, George Stigler
(1981, 190), ‘the hypothesis’ is that ‘we live in a world of reasonably well-
informed people acting intelligently in pursuit of their self-interests’—
though Stigler assumed that self-interests could include the welfare of a
person’s family and ‘a narrow circle of associates’ (189) . In the nineteenth
century, Edgeworth (1881, 16) said, ‘[t]he first principle of Economics is
that every agent is actuated only by self-interest’. In a recent blogpost,
Caplan (2010) wrote, ‘textbooks love to claim that economics assumes
“optimizing behavior,” not self-interest. But whenever economists do
(continued)

13
18. It is not that any economist believes that people are
perfectly rational or perfectly selfish.15 But analysis requires
simplifying assumptions, and many have judged that any loss of
realism implied by these assumptions is more than offset by the
gain in the scope and simplicity of the theory. And the assumptions
are typically applied with some care; no one thinks it useful to
model chess matches as played by people with perfect information-
processing powers, since such players would know how each game
ended before it began. Moreover, as Friedman (1953) observed, a
theory is tested not by armchair consideration of the realism of its
assumptions, but by the empirical success of its predictions
compared with those of alternative theories.
19. By now, however, there is much empirical evidence that
theories built on the assumptions of rationality and self-interest
often fail to predict behaviour well, and that behavioural-economic
models are more successful than neoclassical models in some
domains. It would be wrong to say that neoclassical economics
was no longer useful. There are domains in which it works well.
(Hence, the joke that Vernon Smith got the 2002 Nobel Prize for
showing that economics worked in the lab, while Kahneman got it
for showing that it didn’t.) Moreover, although behavioural
economists have developed models that attempt to explain
particular aspects of behaviour, such as simple choices under
applied work, they quickly slide to self-interest ... [b]ecause although people
aren’t perfectly selfish, they’re shockingly close.’
15 Indeed, several prominent economists outside the tradition of behavioural
economics have criticized the assumption of rationality. Coase (1984, 231)
writes, ‘Most economists make the assumption that man is a rational utility
maximizer. This seems to me both unnecessary and misleading. I have said
that in modern institutional economics we should start with real institutions.
Let us also start with man as he is.’

14
uncertainty or over time, the models do not have anything like the
simplicity and scope of neoclassical models. Yet the evidence does
suggest that behavioural economics should be taken seriously by
those interested in regulation. The evidence for irrationality and
other-regarding preferences in certain domains is strong, and a
theory that predicts behaviour reasonably well in a particular
domain may be useful to policy makers, even it doesn’t have the
scope of neoclassical economics.
20. The evidence considered here relates to16
 Poor use of information,
 Overconfidence,
 Susceptibility to framing,
 Lack of self-control, and
 Other-regarding preferences.
Poor use of information
21. Psychologists have found that we form many judgements
by using simple rules of thumb, or heuristics, which often serve us
well but sometimes lead us astray. As an example, consider an
early experiment by Tversky and Kahneman (1974). They asked
16 There are many much more comprehensive compendiums of evidence.
Kahneman, Slovic, and Tversky (1982) is a collection of research on
heuristics and biases in judgment. Kahneman and Tversky (2000) is a
collection on biases in decision making. Thaler (1993) is collection of
articles on ‘anomalies’—that is phenomena inconsistent with neoclassical
economic theory—which first appeared in the Journal of Economic
Perspectives. Camerer, Loewenstein, and Rabin (2003) is a collection of
articles on behavioural economics. DellaVigna (2008) is a review of
behavioural economics that highlights evidence from the field, not the lab.
Dawes and Thaler (1998), Gintis et al. (2005), Henrich et al. (2004), and
Croson (2008) discuss other-regarding preferences.

15
subjects to estimate the percentage of states in the United Nations
that were in Africa. Before asking, however, they spun a rigged
wheel of fortune that came to a halt on either the number 10 or the
number 65. Then they asked whether the percentage was higher or
lower than the number revealed by the wheel of fortune. On
average, subjects who saw the number 10 estimated the percentage
to be 25, while subjects who saw the number 65 estimated it to be
45. It seems that the number they saw served as a kind of anchor
from which their estimate could drift only so far. Subsequent
studies confirmed the existence of this phenomenon of anchoring
and insufficient adjustment.17
22. A more recent study by Ariely, Loewenstein, and Prelec
(2003) suggests that anchoring affects markets. They asked
subjects to bid for items whose precise market value was unlikely
to be known, such as bottles of French wine. But first they asked
the subjects to write down the last two digits of their social-
security numbers and to consider whether the value of the item was
greater or less than those two digits considered as dollars. Those
whose last two-digits made up a relatively big number ended up
owning a disproportionate amount of the merchandise.
23. As well as using irrelevant information, we sometimes
make poor use of relevant information. For example, we aren’t
good at using diagnostic tests to estimate the probability that a
patient has a disease. In particular, if we know a test is highly
accurate, we tend to think that a positive result very likely means
that the disease is present, forgetting that if the disease is rare there
17 Tversky and Kahneman (1974) also discusses the availability and
representativeness heuristics. See also Kahneman, Slovic, and Tversky
(1982).

16
will be many false positives.18 In a more dramatic example of the
same tendency to underweight underlying probabilities, we will
sometimes guess that a person is more likely to be a feminist bank
teller than a bank teller (Tversky and Kahneman 1984). In still
other cases, however, we fail to update our estimates of prior
probabilities in the light of new evidence as much as we should
(Edwards 1982).
Overconfidence
24. Many other studies find that we have too much confidence
in our abilities and in our judgements about difficult questions. On
average, we consider ourselves better-than-average drivers
(Svenson 1981). If we’re CEOs, we launch takeover bids in search
of illusory synergies (Roll 1986). If we invest in the stock market,
we think we can do better than the average investor. It may be
worse if we’re male. Using data for nearly 38,000 US households,
Barber and Odean (2001) found that equity investors traded too
much for their own good. They would have earned higher returns,
that is, if they had simply bought and held their shares. Women’s
trading reduced their net returns by 1.72 percentage points a year.
18 Suppose, for example, that a diagnostic test is such that 90 per cent of
those who have the disease test positive and 96 per cent of those who do not
have the disease test negative. Suppose further that 1 per cent of the
population has the disease. The probability that a randomly selected person
who tests positive actually has the disease is only 19 per cent. This can be
seen using Bayes’ rule. The probability of having the disease (D), as
opposed to not having it (N), given the evidence of the test (E), is given by
Pr( ) Pr( )Pr( ) / (Pr( )Pr( ) Pr( )Pr( ))
(.90)(.01) / ((.90)(.01) (.04)(.99)) .185
D
EEDDEDDENN


17
Men’s more frequent trading reduced their net returns by 2.65
percentage points a year.
25. Psychologists have assessed overconfidence in belief by
asking people to state their confidence intervals for the answers to
each of a set of questions. For example, they might ask, ‘What is
the length of the Nile? Specify an interval that you’re 98 percent
sure the true length lies within.’ A possible answer is 3,000
kilometres plus or minus 1,000. Asked 100 such questions, we
ought to get about 98 right, however much or however little we
know about the subject. But we typically get only 60 to 70 right.
Events we judge impossible happen about 20 per cent of the time.
Curiously, however, we actually have too little confidence in our
answers to very easy questions.19
Susceptibility to framing
26. Other studies have found that we change our choices in
predictable ways according to the way the choice is framed, even
though the substance of the choice remains the same. For example,
consider a choice between two gambles, one that offers a high
probability of winning a low prize, another that offers a low
probability of winning a high prize. When asked to choose
between the two gambles, we tend to pick the high-probability
low-prize one. But when asked how much we are willing to pay to
play the gambles, we tend to offer more for the low-probability
high-prize one. When considering our willingness to pay, we seem
to emphasize the prize, which is measured in the same units as
19 See Lichtenstein, Fischhoff, and Phillips (1982), Griffin and Tversky
(2002), and Camerer (1995, 591).

18
willingness to pay, whereas no such effect influences our choice
between the gambles. As a result, our preferences are internally
inconsistent.20
27. Other preference reversals are associated with change in the
reference point against which options are assessed. Kahneman and
Tversky (1981) asked subjects to ‘Imagine that the U.S. is
preparing for the outbreak of an unusual Asian disease, which is
expected to kill 600 people.’ Some subjects were then asked to
choose between two programs, A and B, whose consequences were
described as follows:
If Program A is adopted, 200 people will be saved.
If Program B is adopted, there is 1/3 probability that 600
people will be saved, and 2/3 probability that no people will
be saved.
28. The majority chose program A. Both programs are
expected to save 200 lives, but most subjects were risk averse and
preferred the certain outcome. A second set of subjects was given a
choice between programs C and D, whose consequences were
described as follows:
If Program C is adopted 400 people will die.
If Program D is adopted there is 1/3 probability that nobody
will die, and 2/3 probability that 600 people will die.
20 Psychologists Lichtenstein and Slovic (1971, 1973) discovered preference
reversal in experiments in the lab and in Las Vegas. Economists Grether and
Plott (1979) confirmed their results. For reviews, see Tversky and Thaler
(1990) and Starmer (2008).

19
29. There is no substantive difference between Programs A and
C, on the one hand, and between Programs B and D, on the other.
Yet in the second problem, the majority of subjects chose Program
D (and were thus risk seeking). The only difference between the
two choices is the framing. In the first, the problem is framed in
terms of lives saved, and the reference point is the outcome in
which all die. In the second, the problem is framed in terms of
lives lost, and the reference point is the outcome in which all live.
30. The experiment, like many others, suggests that we are
typically risk averse when faced with possible gains and typically
risk seeking when faced with possible losses. Choices can often be
framed either in terms of possible gains or in terms of possible
losses. Thus our decisions can often be reversed by a change of
framing.21
31. Part of the reason for the inconsistency is that we are loss
averse, giving more weight in decisions to a loss of, say, $100 than
to a gain of the same amount. Loss aversion is associated with an
endowment effect: we tend to value many things more highly
simply by virtue of owning them. Thus college students randomly
21 The standard theory of choice under uncertainty is expected-utility theory.
Consider a lottery (gamble, prospect) that offers chances of winning prizes
1
x
and 2
x
with probabilities 1
p
and 2
p
respectively. Its expected value is
11 2 2
p
xpx. Its expected utility, according to expected utility theory, is
11 22
() ()
p
ux pux, where u is a function that allows for risk aversion or risk
seeking. Early violations of expected utility theory were found by Allais
(1953) and Ellsberg (1961). Kahneman and Tversky (1979) provided
evidence of further problems, and presented prospect theory as an
alternative. Prospect theory say that the utility of the lottery considered
above is 11 2 2
()() ()()
p
vx p vx

where

is a function that transforms
probabilities (for example, increases the weight of low probabilities) and v is
a function transform the prizes in a way that differs from the transformation
u of expected-utility theory in that v depends on the value of x relative to the
reference point and is steeper in the domain of losses than of gains.

20
given coffee mugs tend to value the mugs more highly than do
students not given them. Closely related is status-quo bias: a
tendency for people to prefer the status quo to changes even when,
by objective measures, the change is an improvement.22
Inconsistency over time
32. Many important choices involve tradeoffs between current
and future consumption. Typically, future consumption is
discounted relative to present consumption. There is nothing
internally inconsistent about discounting, and so nothing irrational
about it. But many of us have a bias toward the present that
exceeds the effect of internally consistent discounting.
33. Early evidence of present bias came from Thaler (1981),
who asked subjects to say how much money they would require at
various future dates to compensate them for giving up $15 now.
The responses implied discount rates of more than 300 percent
over a one-month horizon and of only 19 percent over a ten-year
horizon. Other studies have found different rates, but have
replicated the finding of rates that decline as the time horizon is
extended.23 If our discount rates decline in this way, we are time-
22 For an overview, see Kahneman, Knetsch, and Thaler (1991). Among
other things, these results cast doubt of the applicability of the Coase
theorem, which states that in the absence of transactions costs the initial
allocation of rights affects the distribution of wealth but not the ultimate
allocation of rights (Kahneman, Knetsch, and Thaler 1990). For a discussion
of reference-dependence, see also Tversky and Kahneman (1986).
23 For example, Laibson, Repetto, and Tobacman (2007) estimate discount
rates of 15–40 per cent a year for short-run choices and 4 per cent a year for
long-run choices. See Frederick, Loewenstein, and O’Donoghue (2002) for a
review. Strotz (1956) identified the link between discounting and
consistency over time and discussed the idea that intertemporal decisions
(continued)

21
inconsistent.24 On Monday, we prefer to delay consumption from
Tuesday to Wednesday, but when Tuesday arrives we no longer
want to wait till Wednesday. So if we committed ourselves to
waiting on Monday, we regret it. In practice, our decisions involve
a battle between temptation and self-control, which can be thought
of as arising because of a divergence between the interests of the
‘present self’ and ‘future selves’. Impatient choices are said to
impose ‘internalities’ on future selves, by analogy with decisions
that impose externalities on others.
34. Choices over time are complicated by three other factors.
First, we underestimate the power of exponential growth, and thus
underestimate the speed at which savings and debt can accumulate
(Stango and Zinman 2009). Second, cues such as seeing a piece of
cake and visceral influences such as being hungry can make
present consumption particularly compelling (Loewenstein 2000).
Third, we have trouble predicting our future tastes. For example, if
we are asked to choose a series of weekly snacks in advance, we
choose more variety than we subsequently want.25 If we win the
involve conflict between different selves. Other important contributions are
Ainslie (1975) and Laibson (1997).
24 Decisions are time-consistent if future consumption is discounted at a
constant rate. If

denotes that rate, the discount factor is 1(1 )



, and
the utility at time 0t of a profile of future consumption is 0()
Ttt
tuc


,
where c is consumption and ()u

is per-period utility. A simple present-
biased utility function, is 01
() ()
Ttt
t
uc uc


, in which

determines the
extent of present bias. This utility function is called quasi-hyperbolic and is
said to represent beta-delta preferences. An example of a hyperbolic utility
function is 1(1 ) ( )
Tt
ttuc

.
25 For evidence, see Gilbert et al. (2002) as well as Loewenstein,
O’Donoghue, and Rabin (2003), Kahneman, Wakker, and Sarin (1997), and
Kahneman (1994).

22
lottery, we are, after the initial euphoria wears off, only a little
happier than others, while if we become paraplegic, we are,
eventually, only a little less happy. Difficulty in predicting future
states is related to the influence of visceral states: when hungry, for
example, we overestimate how much we will want to eat later—
hence the advice not to shop for groceries on an empty stomach.
35. Further suggestive evidence for imperfect consumer
rationality comes from firms’ interest in psychology (e.g., Earl and
Kemp 2002) and behavioural economics (Welch 2010) and from
the marketing strategies that firms follow. If consumers were better
information processors, for example, firms would have no reason
to favour prices like $99 and $199. And more advertising would
provide information on price and quality and less would depict
cues (e.g., pictures of appealing food).
Other-regarding preferences and social norms
36. Finally, there is evidence of preferences that differ from
simple self-interest. There is nothing irrational about such
preferences (de gustibus non est disputandum—there’s no
disputing taste). So they are fully consistent with the core
assumption of neoclassical economics, but they differ from the
preferences assumed in much applied neoclassical work.
37. One of the best-known pieces of evidence for nonstandard
preferences comes from the ultimatum game (Güth, Schmittberger,
and Schwarze 1982). One player, the proposer, offers part of a
given sum of money to the other player, the responder. If the
responder accepts the offer, the proposer keeps the rest. If the
responder rejects the offer, no one gets anything. The game is not
repeated. If the responder is rational and concerned only to

23
maximize her wealth, she will accept any positive offer. If the
proposer is rational and self-interested, and believes that the
responder is rational and self-interested, he will therefore offer the
smallest possible amount.26 In fact, proposers typically offer 40–50
percent of the available sum, and many responders reject offers
below 30 percent (Camerer 2003, ch. 1). Thus responders, at least,
don’t maximize their monetary payoff. They are prepared to incur
a cost to punish ungenerous proposers.
38. Other evidence about other-regarding preferences comes
from public-good experiments, in which players can contribute to a
project that has net benefits for the group but in which the free-
rider problem means that each player maximizes his payoff by
contributing nothing. In a typical one-shot experiment, four players
are initially given $20. Every dollar they contribute to the project
generates a social return of $2, but this return is divided among
four players, so the private return to contributing a dollar is only 50
cents. The players can keep any money they don’t contribute. The
payoff-maximizing strategy is to contribute nothing (and to hope
that others don’t follow this strategy). But if no one contributes
anything, everyone walks away with only their initial $20, whereas
if they all invest $20, they all get to keep $32.
39. In the first round of a finitely repeated game, the average
player contributes about 50 percent of his per-round endowment
(Croson 2008). There are, however, big differences among players.
Perhaps 30–50 percent of the players contribute nothing, as would
be predicted if people had simple self-interested preferences, while
the rest cooperate to varying degrees (Ledyard 1995). Moreover, in
26 This outcome is not the game’s only Nash equilibrium, but it is its only
subgame-perfect equilibrium.

24
finitely repeated games, cooperation tends to decline over the
rounds of the game, so that by the last round the average
contribution may be only a quarter or so of the endowment
(Croson 2008). The decline may occur partly because the players
are learning that contributing has a negative payoff, but this
doesn’t seem to be the whole explanation. Some players initially
cooperate even though they understand the game and apparently
stop cooperating because they are disappointed by the level of
other players’ contributions (Andreoni 1995, Dawes and Thaler
1998).
40. Indeed, contributions in public-good games may decline
partly because cooperators want to punish defectors and can do so
only by not contributing. To test this explanation, Fehr and Gächter
(2000) allowed subjects to pay to punish another player. When
punishment was possible, cooperative players often punished those
that had defected in the previous round, and the level of
cooperation rose over the rounds of the game instead of falling. By
round six, the average contribution was close to the maximum.
Results such as these suggest that many people are conditional
cooperators and altruistic punishers: they contribute if others
contribute and they punish defectors at personal cost. Put
differently, they comply with a norm that prescribes cooperation
with cooperators and the punishment of defectors.
41. Other research shows that material incentives designed to
encourage cooperation sometimes undermine the effectiveness of
cooperative social norms (Bowles 2008, Fehr and Rockenbach
2008, Frey 1997, Gintis et al. 2005). A striking possible illustration
comes from a study of a group of day-care centres in Israel that
imposed a fine on parents who were late collecting their children
and found that parents became even less punctual (Gneezy and

25
Rustichini 2000). Parents may have thought of the fine as a price
and may have come to view being late as a service that they could
purchase, not as a violation of a norm against inconveniencing
others. A similar result was found in an experimental study of
environmental regulation in rural Colombia: an imperfectly
enforced regulation restricting the use of a limited environmental
resource led to outcomes worse than those under no regulation.27
One explanation is that externally imposed rules reduce intrinsic
motivation (Frey 1997).
42. Other work suggests that people worry about their position
relative to others as well as their absolute position. Traditionally,
economists have assumed that people’s welfare depends on their
consumption over their lifetimes. Generally, more is better, and a
higher income is desirable because it allows more consumption.
But there is evidence that people are very concerned about their
relative position or status (see Heiffetz and Frank 2008). Higher
27 The study is Cardenas, Stranlund, and Willis (2000). Groups of eight
villagers were asked to choose how much to take from a hypothetical forest,
in a set-up in which the players could see that the pursuit of self-interest was
socially suboptimal. The level of cooperation in the games was less than the
social optimum, but more than would be predicted of rational and self-
interested players. The main purpose of the experiment, however, was to test
whether an imperfectly enforced regulation improved cooperation. So in a
second stage half the groups were told that they should take only their share
of the socially optimal amount of the resource and that they would be fined
if they were audited and found to have taken more. But the subjects were
given information that implied that the probability of being audited was low
and that the expected-value-maximizing choice was still to take too much. In
the first rounds of the second stage, regulation worked well, but after a few
rounds people were taking at least as much as before. Cardenas and his
colleagues hypothesized that regulation ‘crowded out group-regarding
behaviour in favour of greater self-interest’ (p. 1731). The other half of the
groups continued to play the game without regulation but now had the
ability to communicate with each other before making their choices. They
cooperated somewhat more than before, and in the contrast to the case of
regulation, the improvement endured.

26
income is desirable partly because it allows more conspicuous
consumption and higher status. For example, we may prefer to live
in a relatively large house in a neigbourhood of small houses than
to live in a somewhat larger house in a neighbourhood of very
large houses (see Frank 2005). We may prefer to earn a relatively
high wage in a low-wage firm than to earn a higher wage in a firm
where the average wage is higher still (Shafir, Diamond, and
Tversky 1997).
What does the evidence imply?
43. Much evidence suggests, then, that behaviour sometimes
deviates systematically from the predictions of neoclassical
economics. It is less clear how widespread, enduring, or important
the deviations are. For example, effects found in some experiments
are not found in others.28 More fundamentally, the relevance for
policy of experimental evidence, which provides the strongest case
against rationality and self-interest, is not always clear.
Experiments allow careful controls, but they differ in many ways
from ordinary markets. The stakes may be small and the situations
unrealistic. In ordinary markets, people may behave differently and
learn from their mistakes. Over time, evolutionary pressures may
weed out less-rational firms. Arbitrage by smart traders may limit
the influence of irrational traders. Overall, markets may be less
affected by irrationality or non-selfish preferences than behaviour
in experiments (List 2003, Levitt and List 2007).29
28 The endowment effect, for example, is not found in a careful recent study
by Plott and Zeiler (2005).
29 There are many critical reviews of behavioural economics. Gul and
Pesendorfer (2008) criticize behavioural economics and neuroeconomics,
(continued)

27
44. Yet it would be wrong to conclude that violations of
irrationality are fleeting or unimportant. Many anomalies have
survived repeated testing; preference reversals, for example,
survived what the sceptical economists doing the research
described as ‘a series of experiments designed to discredit the
psychologists’ work as applied to economics’ (Grether and Plott
1979, 623). Some violations persist even when the stakes are high
relative to people’s incomes (Cameron 1999, Kachelmeier and
Shehata 1992). Moreover, there are now many studies that
document irrationality or other-regarding preferences in real-world
settings (DellaVigna 2008). It is true that when a task is repeated
many times and feedback is prompt, people learn from their
mistakes. But learning is often slow, and for some decisions,
feedback is too infrequent (buying a house), too late (saving for
retirement), or too noisy (investing in shares) to be very useful.
Arbitrage and evolutionary pressures no doubt weed out some
firms, but there are limits to their effectiveness (e.g., de Long et al.
1990). Some markets may be relatively unaffected by individual
arguing that seemingly irrational choices can sometimes be analyzed by
assuming rational choice with non-standard preferences. Fudenberg (2006),
reviewing a collection of articles on behavioural economics edited by
Camerer, Loewenstein, and Rabin (2003), argues that ‘[t]here are too many
behavioral theories, most of which have too few applications’ (p. 697).
David Levine (2009) argues that the relevance of psychological evidence to
economics is exaggerated because psychology focuses on individual
behaviour, especially the abnormal, while economics focuses on normal
market outcomes. Binmore and Shaked (2010) dispute the interpretation of
some experimental evidence on social preferences, focusing on Fehr and
Schmidt (1999). There are also criticisms by psychologists of the
psychological research that underpins behavioural economics. Gigerenzer
(1996), for example, criticizes work on ‘narrow norms and vague heuristics’
by Kahneman and Tversky and argues that ‘fast and frugal’ heuristics are
surprisingly reliable. Finally, there are criticisms by legal researchers, such
as Posner (1998), who, responding to an article on behavioural law and
economics by Jolls, Sunstein, and Thaler (1998), criticizes it as, among other
things, exaggerating human irrationality.

28
irrationality, but it would be unwise to assume that markets are
generally unaffected. Evidence suggests, for example, than
financial and betting markets are reasonably efficient in the sense
of excluding opportunities for easy profit, but that they are not
immune from the effects of individual irrationality (Thaler 2005,
Snowberg and Wolfers 2010). The recent rise and fall of house
prices in the United States may be a case in which individual
irrationality led to global macroeconomic problems (Akerlof and
Shiller 2009).
45. Drawing conclusions about the balance of the evidence is
extremely difficult, in part because it’s impossible for any
individual to comprehend more than a small subset of the relevant
evidence, in part because, as Friedman (1953, 40) noted, ‘[t]he
importance of [economics] to everyday life and to major issues of
public policy impedes objectivity’, and in part because we suffer
from confirmation bias—the tendency to seek out evidence that
corroborates our beliefs and to interpret ambiguous evidence in a
light favourable to those beliefs (Nickerson 1998). However, the
evidence presented by psychologists and behavioural economists
seems to create serious problems for neoclassical economics,
suggesting that policy analysts should be sceptical about the
applicability of theoretical results that assume sophisticated
reasoning on the part of individuals. In addition, the behavioural-
economic results summarized above seem robust enough to use in
designing regulation. At the same time, it would be premature to
dismiss neoclassical analysis such as that mentioned in the
beginning of this paper. Behavioural economics may offer insights
into environmental regulation and competition policy—for
example about ways of getting consumers to reduce polluting
activities—but neoclassical economics remains the only
comprehensive source of guidance in these and others areas.

29
Unlike neoclassical economics, behavioural economics is not a
comprehensive theory that can be used to analyze almost any
policy problem.
Regulatory approaches suggested by behavioural economics
46. This section of the paper looks at just four areas where
behavioural economics has a relatively strong claim to relevance,
namely the regulation of
 Retail savings and investment products,
 Consumer credit,
 Individual insurance, and
 Drugs, gambling, and impulsive behaviour.
Savings and investment
47. The KiwiSaver scheme has clearly been influenced by
behavioural economics. Consistent with research on present-bias, it
presumes that some people have a tendency to save too little.
Consistent with research on status-quo bias, it makes enrolment in
the scheme the default option. By preventing impulsive
withdrawals, it provides a commitment device to aid self-control.30
Consistent with doubts that savings decisions are fully rational, it
30 Laibson (1997) notes that illiquid savings products are a useful
commitment device for present-biased individuals and argues that financial
liberalization, which increases liquidity, may reduce savings. See also
Ashraf, Karlan, and Yin (2006), which finds that a group of hyperbolic
discounters in the Philippines save more when offered an illiquid savings
product.

30
presumes that the savings it generates won’t be offset one for one
by declines in other forms of household saving.
48. KiwiSaver came into effect in July 2007, and by April
2010, some 1.40 million people were enrolled, exceeding forecasts,
while .24 million had opted out.31 Household savings have
increased significantly since the introduction of KiwiSaver
(Reserve Bank 2010), but whether KiwiSaver has contributed to
the increase is unclear;32 the financial crisis may be more
important. Moreover, contributors to KiwiSaver get government
subsidies, so it is hard to distinguish enrolment for behavioural-
economic reasons from enrolment for neoclassical reasons. And
the subsidies for KiwiSaver reduce public savings, so the scheme’s
effect on national savings is also hard to determine. Yet, if savings
decisions are not fully rational, it would not be surprising if
KiwiSaver had increased household savings.
49. Behavioural economics suggests other ideas for savings-
related regulation. One is that subsidies are not the only, and are
probably not the most cost-effective, way of encouraging savings.
As a field experiment in South Africa shows, advertising may be as
powerful as quite large financial incentives. Men who received
junk mail offering a short-term loan were more likely to take up
31 For enrolments, see http://www.kiwisaver.govt.nz/statistics/ks-stats-10-
04-30.html, viewed on 7 June 2010. Inland Revenue Department (2009, 11)
forecast that it would take till 2013/14 for enrolment to reach 1.38 million.
32 Gibson and Le (2008, 1) found that by January 2008 KiwiSaver had raised
total savings only slightly and concluded that it may be a ‘costly and
ineffective solution to a relatively small problem’. It may be too early to tell,
however. Discussing 401k retirement savings, a tax-advantaged employer-
provided scheme in the United States, Benjamin and Laibson (2003) argue
that the long-run increase in savings caused by the scheme has been much
larger than its short-run effect.

31
the loan if the letter included a photo of a woman’s face, and the
photo increased take-up as much as a 25 per cent reduction in the
interest rate (Bertrand et al. 2010).
50. Information-disclosure regulation is another area for which
behavioural economics has implications. Information-disclosure
regulations are of course common and uncontroversial. New
Zealand law requires that those seeking funds from investors
provide a clear disclosure of the nature of the investment and its
costs and risks.33 Further disclosure requirements have been
introduced in the wake of the financial crisis and the failure of
many finance companies, including a requirement that non-bank
deposit takers obtain and disclose a credit rating.
51. But, while information-disclosure regulations are common
and uncontroversial, they are not easy to reconcile with
neoclassical economics, even though neoclassical economics
routinely deals with the implications of imperfect information. At
least in simple neoclassical models, uninformed buyers are still
rational and are therefore sceptical. They assume that a firm that
fails to disclose information about its product has something to
hide. As a result firms with good-quality products have an
incentive to disclose information. Ultimately the sellers of all but
the worst products should choose to disclose information: there
should be no need for regulation to require it (Milgrom 2008).
52. By contrast, the assumption that people are not just
uninformed but also less than fully rational makes it more likely
that information-disclosure regulations, which influence not only
33 The Securities Act 1978 and the Securities Regulations 1983 require that
investors receive an investment statement and, if they request it, a more-
detailed prospectus.

32
what people know but also what they pay attention to, may be
useful. Among other things, information-processing problems
suggest that disclosure for consumers should be very simple.
53. Drawing on arguments from behavioural economics, the
‘Squam Lake’ group of finance professors recently proposed that
investment funds provide a simple disclosure modeled on
mandatory food labeling (Kenneth French et al. 2010, ch. 6). Food
labels, designed to fit on small surfaces, are shorter and more
focused than typical financial disclosures, and thus better attuned
to the limits of human information processing. Table 1 sets out part
of the investment disclosure label proposed by the Squam Lake
group. (The rest of the label explains the information in the table).
The proposal is notable not only for what it includes but for what it
excludes, namely information on past returns. The reason for the
exclusion is that many retail investors mistakenly assume that high
past returns indicate high future returns, which generally isn’t true
(Carhart 1997). The group also recommends that ‘Whenever an
advertisement or other disclosure about an investment product ...
reports an average prior return, it ... also include a standardized
measure of the uncertainty associated with the average’ (p. 61). In
contrast to past returns, fees and expenses do help predict future
returns (higher fees being associated with lower returns) and are
thus included in the Squam Lake disclosure. Similar rules could be
applied to KiwiSaver and other investment funds.

33
Table 1 Squam Lake Group’s proposed disclosure label
Source: French et al. (2010, 64).
Note: The possible 10-year payoffs are the mean and selected percentiles of the estimated
distribution. The mean is assumed to be higher than the 50th percentile because the
distribution of typical investment returns is skewed to the right. Turnover is the percentage
of the firm’s holdings bought and sold in a year (the higher the percentage, the higher are
realized capital gains, and thus in the US taxes, and the higher are unmeasured costs
associated, for example, with bid–ask spreads). Annual volatility is the standard deviation
of annual returns.
54. Other simple disclosures might also be contemplated.
Finance companies and other issuers of risky debt might be
required to disclose not only the promised interest rate on their
debt but also an estimate of the expected interest rate. The expected
rate takes account of the probability of default and can be
estimated from credit ratings.34
55. Behavioural economics could also inform prudential
regulation of savings in banks. Since 1996, prudential regulation
has emphasized monitoring by depositors over monitoring by the
34 If rating changes follow a Markov process, the probability of default over
a given number of years depends only on the initial credit rating and the
annual ratings-transition matrix, the elements of which are the probabilities
of the possible changes in ratings. Ratings-transition matrices are published
by ratings agencies.

34
Reserve Bank.35 Evidence from the United States suggests that
private monitoring may be effective (Flannery 1998). And in New
Zealand regulated information disclosure facilitates private
monitoring while the absence until recently of deposit insurance
has encouraged it. Neoclassical economics suggest that reliance on
private monitoring may suffer from a free-rider problem, because
for small depositors the cost of analysing information to determine
whether a bank is sound probably outweighs the benefit, even
though the benefit of the analysis to depositors as a group
outweighs the cost. Conventional political economy also suggests
that the absence of formal deposit insurance isn’t entirely credible:
when it comes to deposit insurance, the joke goes, there are two
kinds of governments—those that provide it and those that think
they don’t. Evidence of poor information-processing powers casts
further doubt on the efficacy of private monitoring—though of
course doesn’t imply that monitoring by a government agency will
be better. In addition, evidence of risk-seeking in the domain losses
raises the possibility that banks facing losses may take excessive
risks. If so, regulators have another reason to intervene quickly to
control the risk-taking of troubled financial institutions benefitting
from implicit or explicit government guarantees.
56. Although regulation may help investors make good choices
among investments, it also raises the cost or reduces the
availability of investment products—which may be particularly
unhelpful if present bias causes people to invest too little. One
possible cost of regulation is created by hindsight bias, the
phenomenon that an event that has occurred seems in hindsight to
have been more predictable than it really was (Fischhoff 1982).
35 See Reserve Bank Act 1989, Part 5, and Grimes (1996).

35
The success of an investment may be entirely unpredictable. If it
fails, however, its failure will tend to seem predictable. Those who
make mistakes can thus seem negligent when they are merely
unlucky. Regulation needs to ensure that unsuccessful investment
promoters are not judged on the basis of hindsight (Rachlinksi
1998).
Consumer credit
57. Research that suggests that people save too little also
suggests that they borrow too much. Indeed, excess borrowing may
be a bigger problem than insufficient savings, because financial
firms have an incentive to encourage both savings and borrowing,
which is helpful in the case of savings but not in the case of
borrowing. That is, banks’ advertising and sales promotion might
help correct a consumer failure that leads to insufficient savings,
bringing the level of savings closer to what would be observed in a
market of fully rational consumers, whereas their advertising and
sales promotion might exacerbate a problem of excess borrowing.
58. Some people appear to borrow for longer than they should
on credit cards. Overconfidence in their ability to pay off their
balance in full may cause them to accept cards with low annual
fees and high late fees and interest rates (Ausubel 1991). Low
required minimum payments may then serve as an anchor that
causes them to pay less than they otherwise would (Stewart 2009).
Underestimation of the exponential growth of an unpaid interest-
accruing balance may further reduce their monthly payments
(Stango and Zinman 2009). On the other hand, cardholders get
monthly feedback on the effects of their credit-card decisions and

36
frequent opportunities to change course, and appear to learn from
their mistakes (Agarwal et al. 2008).
59. Table 2 illustrates a disclosure required in the United States
by the Credit CARD Act of 2009, which may protect cardholders
that are vulnerable to these biases, without imposing large costs on
credit-card issuers or limiting the freedom of sophisticated
cardholders. (It may, however, raise fees for the sophisticated,
because, if the vulnerable are encouraged to reduce their interest-
bearing balances, banks can no longer subsidize fees for the
sophisticated with the profits made from the vulnerable—as
discussed below.)
Table 2 Illustration of disclosure required by US Credit CARD Act
2009
Note: The disclosure relates to a card with a balance of $5,122.45, a minimum required
payment of $103, and an annual interest rate of 8.9 percent.
60. In addition to prescribing information disclosure by
lenders, the government could increase its provision or
subsidization of information for individual borrowers or for
consumers generally. It already provides a little such advice
through the Ministry of Consumer Affairs and it subsidizes the
Citizens Advice Bureau. It might increase those subsidies or also
subsidize other organizations with similar goals, such as Consumer
NZ. The fact that information is a public good (non-rival in

37
consumption and non-excludable) creates a neoclassical argument
for subsidizing or providing information, but governments usually
choose to subsidize or provide only a very limited range of
information, because markets often find ways of overcoming the
non-excludability problem (there are plenty of commercially
supplied books, newspapers, magazines, and informative
websites), and because the imperfections of public subsidies or
public provision must also be taken into account. But if public-
good problems are compounded by flaws in consumer decision
making, so that consumers don’t always know what kind of
information they should ask for, there is another argument for
public subsidies or provision.
61. The accompanying case study on the regulation of
consumer credit (Tooth 2010) considers these issues in more detail.
Insurance
62. Insurance is another regulatory area where the application
of behavioural economics may be fruitful. First, insurance law in
New Zealand has been said to be ‘singularly bereft of legislative
protection for consumers’ and ‘characterised by judge-made law
which makes utterly unrealistic expectations of consumers’
(Grainer, Bevan, and Dugan 2010, 266). Second, psychologists and
behavioural economists have extensively investigated the way
people make decisions in the presence of uncertainty.
63. According to prospect theory, consumers are loss averse
and tend to place too much weight on events with small
probabilities. According to research on heuristics and biases, they
overestimate salient risks, such as those that tend to make the
news. They are also extremely, arguably irrationally, averse to

38
small risks.36 For all these reasons, it would not be surprising if
consumers paid too much for insurance against small risks.
Experimental subjects have been found to pay as much for air-
travel insurance against terrorist acts as for comprehensive air-
travel insurance (Johnson et al. 2003).37 And there is a widespread
view that many consumers overvalue extended warranties.38
64. But underinsurance against large risks may be a bigger
problem. Insurance against some important risks, such as the loss
of one’s car or house is available yet not always fully used, leaving
people vulnerable to large avoidable losses. Underinsurance
against floods has been observed in the US, even when insurance is
subsidized (Johnson et al. 1993). Although the overweighing of
small probabilities encourages people to insure against these risks,
overconfidence may have a more-than-offsetting effect. On
average, people tend to underestimate risks that are under their
own control, such as the risk of causing an accident and may
therefore underinsure these risks (Weinstein 1989, Sandroni and
Squintani 2007).
36 Rabin (2000) shows that, in the framework of expected-utility theory, the
degree of risk aversion necessary to explain a risk-averse choice in a small-
stakes gamble (e.g., refusing even odds of wining $11 and losing $10)
implies crazily risk-averse choices in larger-stakes gambles (e.g. refusing
even odds of winning $1 million and losing $100). See Rabin and Thaler
(2001) for an overview. Loss aversion helps explain the ‘risk averse’ choice
in the small gamble.
37 This result is actually most consistent with a specific psychological theory
of probability judgment called support theory (Tversky and Koehler 2002).
38 Consider the episode of The Simpsons in which Homer has a crayon
hammered into his nose to lower his I.Q. ‘The surgeon knew the operation
was complete when Homer finally exclaimed: “Extended warranty! How can
I lose?”’ (Camerer et al. 2003)

39
65. A possible response to underinsurance is to make some
kinds of insurance compulsory, either by regulating for the
purchase of private insurance or by establishing tax-funded public
insurance. Compulsion is supported to some extent by traditional
analysis of externalities (the likelihood of public assistance being
provided to the uninsured) and adverse selection (the problem that
the high risk are more likely than the low risk to seek insurance).39
Another option is to make full insurance a default option, perhaps
for employees or for people seeking insurance. Involving
employers in insurance as well as retirement savings would require
a fundamental change in the way insurance is sold, however, and
would demand stronger evidence of the extent and significance of
underinsurance. It is not further considered here, although it could
be examined in a case study. Making full insurance the default for
those seeking insurance may be useful, but it could be done by
insurers themselves without regulation, perhaps with some
encouragement by governments seeking, at least, to reduce
externalities.
66. Given concerns about consumers’ insurance decisions and
uncertainty about whether biases for over-insurance or under-
insurance predominate, the best option may be for the government
to provide or require the disclosure of more information on risks.
Estimates of the probabilities of events might also be disclosed in
ways that are relatively easy to understand. In particular, estimates
39 There is a theoretical argument that suggests that overconfidence may
weaken the case for making insurance compulsory. Rothschild and Stiglitz
developed a neoclassical model in which imperfect information in an
insurance market creates adverse selection and in which regulation requiring
insurance is a Pareto improvement. Sandroni and Squintani (2007) add
overconfidence to this model and show that such regulation is no longer a
Pareto improvement over no regulation.

40
appear to be easier to understand if it they are expressed in terms
of natural frequencies. Thus ‘7 in every 1000 people like you will
die next year’ appears to be easier to understand than ‘Your chance
of dying next year is 0.007.’ Evidence from medical studies
suggests that graphical representations of risk may aid
understanding (Lipkus and Hollands 1999). A ‘Paling palette’
indicates a probability of .007 by showing 1,000 stylized
individuals in diagram, 7 of them highlighted (Paling 2003).
Estimates of losses in the absence of insurance might also be
disclosed.
67. Incidentally, information presented in natural frequencies
also reduces errors in updating probability estimates in the light of
new information—for example, estimating the probability that a
person has a disease given a positive test result (Gigerenzer 2003).
Figure 1 illustrates, using a diagram inspired by Fountain and
Gunby (2010). The example is the same as that discussed in
footnote 18, but the diagram is more intuitive than the mathematics
of the footnote.
Figure 1 Reasoning with natural frequencies is easier than with
probabilities
1000
people
10 990
disease nodisease
9 1 40 950
postivenegativepostivenegative
Chanceofhavingthe 9
diseaseiftestedpositive 9 + 40
≈19/100=

41
Drugs, gambling, and impulsive behaviour
68. Another area where the case for applying behavioural
economics is strong is regulation of addictive substances and
possibly addictive activities such as gambling. Here, present-bias,
cues, and visceral influences on decision-making are associated
with ‘hot-state’ choices by addicts that are widely recognized to
diverge from their interests, even while other people, in ‘cold
states’, are able to make prudent decisions. Standard neoclassical
theory has something to say about the supply of drugs—for
example about the effects on price of cartels and government
interventions to restrict supply. The neoclassical theory of rational
addiction (Becker and Murphy 1988) can also explain some
features of addiction, such as binges and cold-turkey withdrawals,
and the effect of drug prices on the use of drugs, but it cannot
explain others, such as repeated, unsuccessful attempts to quit and
addicts’ beliefs that their choices are mistakes.40
69. In some behavioural-economic models of addiction, the
optimal tax on drugs takes account not only of externalities but
also of internalities—the harm that a problematic drug user
imposes on his future self (Gruber and Kozsegi 2001). Once
internalities are accounted for, such taxes may even be progressive,
since they may disproportionately help low-income consumers to
reduce consumption. But in other behavioural-economic models
the optimal rate of the tax is actually less than that implied by
externalities (see Bernheim and Rangel 2007), in part because cue-
driven consumption in hot-states is not very responsive to price.
40 For introductions to addiction by behavioural economics, see Loewenstein
and Rick (2008) and Bernheim and Rangel (2007, section 2).

42
Moreover, taxes, like prohibitions, do not have the libertarian-
paternalist benefit of allowing normal consumption by non-addicts.
70. A common rationale for existing bans on the advertising of
alcohol and cigarettes is that advertising encourages young people
to start consumption. The significance of cues suggests that it may
be useful because it helps addicts avoid impulsive consumption.
Similar reasoning might suggest that the government should reduce
advertising for its own gambling scheme, Lotto. The significance
of cues also suggests that it is useful to create counter-cues, such as
viscerally charged warnings on cigarette packs. Similar warnings
might accompany alcohol and gambling. Such warnings may,
however, may reduce the pleasure that non-addicts get from
consumption or make addicts feel guilty without much changing
their behaviour (Loewenstein and O’Donoghue 2006).
71. Regulation could also facilitate self-control, without
preventing use. The Gambling Act 2003 allows people who
identify themselves as problem gamblers to ban themselves from a
casino. Indeed, a man who recently thought he had won $60,000
was denied the prize because he had previously added his name to
a list of self-banned gamblers (a more-effective policy would have
stopped him gambling in the first place).41 In addition, gamblers
could be required to buy a gambling debit card. Money put on the
card could be used after a week’s delay to buy gambling chips. The
gambler would choose how much to put on the card, but once his
chips had been used, he would have to stop gambling until new
money was put on the card and another week had passed
(Benjamin and Laibson 2003).
41 ‘Gambler’s ban costs him $60,000 prize’, Michael Fox, Dominion Post, 6
August 2010.

43
72. A similar policy might be applied to drugs to which access
was legal but controlled, with irrevocable self-prohibitions or with
self-prohibitions that could be revoked after a delay. Smokers, for
example, might put themselves on such a list, at least if the
purchase of cigarettes required the presentation of identification.
The policy would not be intended to prevent smokers from
smoking, but to help quitters avoid impulsive recidivism. Perhaps
banks could be required to offer a credit card that would require
the cardholder to approve transactions above a threshold value a
day or two in advance. Current law provides a cooling-off period
for borrowing, but giving people the right to reverse a decision
within three days is probably less powerful, because of status-quo
bias, than requiring them to plan for the decision in advance.
Substantive criteria for assessing regulations: Behavioural
welfare economics
73. This discussion of regulations suggested by behavioural
economics glosses over a problem. To evaluate a proposed
regulation, we need both a descriptive (‘positive’) theory of
behaviour, such as that provided by behavioural or neoclassical
economics, and a normative (that is, ethical) theory. The
descriptive theory tells us how a regulation will affect outcomes of
interest, and the normative theory tells us whether or not the
change is good. For example, the descriptive theory might predict
that a night-time curfew on flights at an airport would lead to a
certain reduction in noise along with a certain change in the
number and timing of flights. The normative theory then says
whether the predicted outcome—reduced noise and the new
schedule of flights—is, on balance, better or worse than the status
quo, or equally good. Behavioural economics undermines the

44
normative theory that traditionally underlies economic analysis of
regulation, namely welfare economics and its offshoot, cost–
benefit analysis. How then can proposed regulations be evaluated?
74. Sometimes the problem may be merely theoretical. If there
is enough agreement on the goal of a regulation, there is no need
for a theory that allows for tradeoffs between different goals (e.g.,
less noise and more-conveniently scheduled flights). But an
advantage of neoclassical welfare economics and cost–benefit
analysis is that they provide a principled way of trading off
different goals when there is disagreement. Formal cost–benefit
analysis is quantitative, and can therefore provide precise answers.
If its assumptions are accepted, and its application is feasible, it is
very powerful. But even when no quantitative analysis is
undertaken, standard welfare economics is influential because
economically trained analysts are likely to think of themselves as
making an informal estimate of whether the benefits of a regulation
outweigh the costs, benefits and costs being understood in terms of
willingness to pay and willingness to accept, and thus in terms of
preferences. Behavioural economists have therefore begun to
consider how neoclassical welfare economics and cost–benefit
analysis might be reconstructed in the absence of the assumption of
rationality.
75. In the normative theory that is part of neoclassical
economics (welfare economics) outcomes are assessed by
reference to the same preferences that are used to predict
behaviour. What is good is equated with what people choose or
42 In a discussion of behavioural economics and neuroeconomics, Gul and
Pesendorfer (2008) argue that welfare economics is not in fact normative in
this sense. In their view, economists, as economists, should seek only to
(continued)

45
would choose if given the opportunity. Thus, an airport curfew is
clearly desirable (a Pareto improvement) if some people prefer the
outcomes under the curfew and no one prefers the status quo. Few
regulatory changes are Pareto improvements, and the Hicks–
Kaldor criterion is commonly applied instead. According to this
criterion, which is fundamental to cost–benefit analysis, a
regulatory change is desirable if those who gain from it would be
willing to compensate those who lose, even if no compensation is
paid. (The idea is that when the government makes a large number
of regulatory choices each of which creates a net gain, there
shouldn’t be many people who experience a net loss, and that any
remaining concerns about the distribution of well-being should be
addressed by the system of taxation and benefits.)
76. Assessing what people would choose can be difficult, but
analysts have developed techniques to provide at least rough
answers. How much people would pay to reduce noise pollution,
for example, might be estimated by giving them a carefully
formulated questionnaire43 or by comparing the prices of houses
near airports and houses in quieter neighbourhoods.
77. But if choices do not reflect reasonable beliefs and
consistent preferences, then they are an imperfect guide to the
desirability of outcomes. Some other criterion might be better.
Research in behavioural welfare economics, as it is sometimes
understand the world and not to improve it. Their view is probably not
standard.
43 The reliability of answers to questions about willingness to pay has,
however, been criticized both by neoclassical economists who prefer to rely
on preferences revealed by behaviour and by psychologists who believe that
the answers merely reflect attitudes (e.g., Kahneman et al 1993 and Kemp
2002).

46
called, seeks to establish such a criterion (Bernheim and Rangel
2008, Loewenstein and Haisley 2008). Like neoclassical welfare
economics, behavioural welfare economics deals with ethics, and
thus overlaps with the concerns of philosophers.
Choice free from error and inconsistency
78. One approach to normative analysis in the presence of
irrationality is to continue to defer to choice, but to try to weed out
choices that don’t reflect preferences. For example, choices based
on mistaken beliefs can be ignored.44 Perhaps workers’ willingness
to accept certain dangerous work is suspect because workers
underestimate the risks (Akerlof and Dickens 1982). Choices that
depend on framing or other aspects of the problem that should be
irrelevant can also be ignored (Bernheim and Rangel 2009). For
example, in the problem of the Asian disease set out above, if we
switch between saving 200 lives for certain and taking a chance on
saving all 600 according to the framing of the problem, both
44 This approach is proposed by Köszegi and Rabin (2008). As an
illustration, consider a choice between two bets. In one, you get an apple if a
coin is tossed and lands heads up, otherwise nothing. In the other, you get an
orange if the coin lands tails up, otherwise nothing. Suppose you normally
choose the bet for an apple, which suggests that you prefer apples to
oranges, but that you choose the bet for an orange after a long series of
heads. If your beliefs are assumed to be correct, either you have inconsistent
preferences or a complex and unusual preference for oranges preceded by
certain coin tosses. A simpler explanation is that you suffer from the
gambler’s fallacy. If so, your choice of the bet for an orange after a series of
heads should not be taken as indication of your valuation of apples relative
to oranges. By contrast, Gul and Pesendorfer (2008) argue that the best way
to deal with certain behavioural anomalies is indeed to assume more-
complex preferences, in particular over consumption and the choice sets
from which consumption is chosen.

47
choices could be discarded as suspect. The choices that count
would be those robust enough to survive changes in framing.45
79. This approach has the advantage of respecting consumer
sovereignty insofar as it isn’t undermined by ‘consumer failure’,
but it is not clear that is practical. If choice is very susceptible to
framing and other irrelevant matters, for example, the criterion of
unambiguous choice might not be very discerning: many
regulatory options might rank equally. Moreover, it is one thing to
describe a few cases in which choices clearly do not reflect
preferences; it is another to imagine that cost–benefit analysts
might generally be able to discern people’s preferences reliably
other than by observing their choices. Lastly, even the relatively
simple neoclassical approach is hard to apply to many regulations,
and this behavioural-economic alternative is more complex.
Subjective well-being
80. An alternative is to assess outcomes according to a measure
of subjective well-being, such as happiness or experienced utility.
This idea goes back to Bentham (1789), but it fell out of favour
with economists because of apparently insurmountable difficulties
in measuring subjective well-being and the doctrine of
behaviourism (not to be confused with behavioural economics) that
mental states are not a proper subject of scientific inquiry. As a
result, economists came to focus on choices. Problems with
choice-based criteria, however, have prompted renewed interest in
45 There are related philosophical approaches. One is to count preferences
made under ideal conditions. Another is to assume that people have second-
order preferences about their first-order preferences, such as a preference for
not wanting to smoke, and to count only preferred preferences.

48
measuring subjective well-being and in using the measurements in
the design of policy.46
81. Kahneman and others have experimented with measuring
experienced utility (pleasure and pain) as it is reported over the
course of a day and thus building up a measure of the hedonic
value of the day. Others have focused on creating better surveys of
reported happiness and related measures, and giving those surveys
more prominence in the monitoring of economic performance.
Stiglitz, Sen, and Fitoussi (2010, ch. 2) recommend that national
statistical offices routinely ask survey questions on subjective well-
being, and do so in a way that allows analysis of the relationship
between subjective well-being and objective circumstances.
82. Despite progress in the measurement and explanation of
subjective well-being, assessing regulations only by reference to
estimates of their effects on happiness would be very difficult.
Even if it were feasible, it might have some counterintuitive
implications. For example, it might justify reduced concern for the
disabled on the grounds that they adapt to their circumstances and
are not much less happy than the able-bodied.
Using traditional cost–benefit analysis without believing all its
assumptions
83. A third alternative, espoused by Sunstein (2000), is to
recognize that traditional cost–benefit analysis has flawed
conceptual foundations but to use it anyway, at least for some
decisions, because it is better in practice than the alternatives.
46 See, for example, Layard (2005) and Kahneman, Wakker, and Sarin
(1997) and, for a review, Frey and Stutzer (2002).

49
Indeed, Sunstein argues that behavioural economics provides a
better justification than neoclassical economics for traditional
cost–benefit analysis:
Cost–benefit analysis is often justified on conventional
economic grounds, as a way of preventing inefficiency. But
it is most plausibly justified on cognitive grounds—as a way
of counteracting predictable problems in individual and
social cognition. Poor judgments, by individuals and
societies, can result from certain heuristics, from
informational and reputational cascades, from thinking
processes in which benefits are ‘on screen’, but costs are
not, from ignoring systemic effects of one-shot
interventions, from seeing cases in isolation, and from
intense emotional reactions. Cost–benefit analysis serves as
a corrective to these cognitive problems (p 1059).
Libertarian and asymmetric paternalism
Behavioural welfare economics raises the spectre of paternalism,
which is anathema to many.47 It is not that behavioural economics
47 Kant, for example, described a paternalist government as the ‘greatest
conceivable despotism’ (1793/1991, 74). And when Alexis de Tocqueville
considered what kind of ‘despotism’ might arise in a democracy, he
imagined a state that is ‘responsible for securing … [the] enjoyment [of its
citizens] and watching over their fate’ and whose ‘power is absolute,
thoughtful of detail, orderly, provident, and gentle’ but ‘daily makes the
exercise of free choice less useful and rarer, restricts the activity of free will
within a narrower compass, and little by little robs each citizen of the proper
use of his own faculties’ (1835–1840, book 4, ch. 6). John Stuart Mill (1859)
contended that governments could legitimately restrict people’s freedom
only to prevent harm to other people: ‘The sole end for which mankind are
warranted, individually or collectively in interfering with the liberty of
action of any of their number, is self-protection. His own good, either
physical or moral, is not a sufficient warrant.’ Robert Nozick’s (1974)
rights-based political philosophy naturally has no place for paternalism. But
(continued)

50
necessarily supports paternalist regulations. When competition is
strong, the irrational may still get a good deal from an unregulated
competitive market.48 To take a trivial example: if you irrationally
believe that a particular orange has miraculous benefits, and are
willing to spend all your money to get one, you’re likely to be
exploited, but if you equally irrationally believe that all oranges
have miraculous health benefits you’ll be fine because oranges are
supplied competitively (Benjamin and Laibson 2003). More
generally, problems caused by an imperfect market must be
compared with problems caused by an imperfect government, and
behavioural economics strengthens concerns about governments as
well as markets. But, at the very least, behavioural economics
undermines one reason for opposing paternalism, and is therefore
‘anti-anti-paternalist’ (Jolls, Sunstein, and Thaler 1998). In
practice, moreover, the policy proposals of behavioural economists
are often (mildly) paternalist.
84. Accordingly, behavioural economists have attempted to
distinguish acceptable from unacceptable paternalism.49 The best-
John Rawls’s quite different principles of justice also give priority to liberty
(1971/1999, 220). His first principle, which takes priority over the better-
known difference (or maximin) principle, is that ‘Each person is to have an
equal right to the most extensive total system of equal basic liberties
compatible with a similar system of liberty for all.’
48 Sugden (2004) develops a model in which consumers do not have
coherent preferences and markets are ‘collections of money pumps operated
with the intention of extracting value from consumers’ but in which the
‘overall effect of these money pumps is benign, not because consumers are
induced to form coherent preferences, but because of the effects of
competition among arbitrageurs’ (p. 1015). Other theoretical and empirical
evidence suggest that the competition doesn’t necessarily protect irrational
consumers (Ausubel 1991, Gabaix and Laibson 2006, Grubb 2009, and
Brown, Hossain, and Morgan 2010).
49 In doing so, they have joined an older tradition in philosophy. Mill (1859,
ch. 5), for example, allowed a person’s liberty be infringed to protect that
(continued)

51
known variety of soft paternalism is the libertarian paternalism of
Thaler and Sunstein (2003). Libertarian paternalism allows
‘nudges’ that encourage people to choose one alternative but leave
them free to choose another. Default enrolment in KiwiSaver is
libertarian, because employees can opt out, and paternalist,
because enrolment is chosen as the default on the grounds that it is
good for many people who nevertheless won’t chose it. Similarly,
mandatory information disclosure preserves the liberty of
consumers to choose, but encourages them to make informed
decisions. (The policies do limit the liberty of employers or
sellers.)
85. An attraction of libertarian-paternalist regulations is that
they may help the vulnerable (less rational), while imposing little
or no cost on the sophisticated (more rational). Sophisticates who
work out that enrolment in KiwiSaver makes no sense for them can
easily opt out, so their choice isn’t affected by the default.
Asymmetric paternalism is a generalization of libertarian
paternalism that allows paternalistic regulations as long as the
costs they impose on the sophisticated are small relative to the
benefits they create for the vulnerable (Camerer et al. 2003).
86. Asymmetry is potentially important because research by
psychologists suggests that there is variation among individuals in
person’s welfare if the person is ‘delirious, or in some state of excitement or
absorption incompatible with the full use of the reflecting faculty’. Another
form of paternalism says that it is permissible to stop people from acting on
mistaken beliefs (see Dworkin 1972, 2010), which is consistent with the
program of developing a measure of choice-based welfare cleansed of the
effect of errors. This kind of paternalism would clearly allow information-
disclosure regulation. Perhaps it would also be consistent with regulations
that allowed people to do certain things, such as investing in a complicated
financial product, only if they could demonstrate understanding of the
product.

52
their susceptibility to some biases. Perhaps not surprisingly,
students at the Massachusetts Institute of Technology are better
than students at the University of Toledo at avoiding the trap in
questions such as, ‘If it takes 5 machines 5 minutes to make 5
widgets, how long would it take 100 machines to make 100
widgets?’ (Frederick 2005). Moreover, people who fall in to the
trap seem to be more impulsive and more likely to make time-
inconsistent decisions. This is not to say that biases afflict only the
less educated: they have been found among doctors, statisticians,
and mathematical psychologists. It is dangerous simply to assume
that a bias affects some groups but not others.
Cost–benefit analysis for paternalist regulations
87. Asymmetric paternalism lends itself to a kind of cost–
benefit analysis. Good paternalist policies are those that create
benefits to the vulnerable (less rational) that exceed the costs
imposed on others, including sophisticated (more rational)
consumers, firms, and the government (Camerer et al. 2003).50
50 In symbols, good paternalistic regulations are those for which
(1 )pB p C I , where p is the proportion of people who are
vulnerable, B is the net benefit created by regulation for the vulnerable, C is
the net cost imposed on the sophisticated, I is per-person implementation
costs, and  is per-person lost profits. How B and C are to be measured, so
that they are expressed in the same units and can thus be added to each other
and to I and  , is of course a crucial question. They might measure
willingness to pay (or accept) in dollars, where willingness to pay (or accept)
is measured in ideal conditions, namely after the adjustments for mistakes
and ambiguous preferences, as discussed by Köszegi and Rabin (2008) and
Bernheim and Rangel (2009), respectively. They might be measured in utils,
where weights are attached to each person’s utility scale to ensure
commensurability (see, e.g., Yaari 1981) or according to some objective
measure of well-being (from neuroscience?). If none of these options is
(continued)

53
88. According to this criterion, if everyone is sophisticated, no
paternalist policy has a net benefit. More generally, paternalism is
more likely to have a net benefit the larger the fraction of the
regulated population that is vulnerable to the problem. Thus, the
question of the prevalence of deviations from rationality in the
regulated population is important, and restricting the scope of the
regulation to certain groups may be helpful. This is consistent with
the application of some investor protections to ordinary
individuals, but not to institutional or ‘habitual’ investors, and the
application of some consumer-protection law to consumers but not
to firms buying their raw materials.51 Assessment of the net
benefits to the vulnerable must take account of costs imposed on
the vulnerable. For example, even if many unsophisticated people
save too little, there may be some for whom KiwiSaver is
inappropriate and who remain involved out of inertia.
89. The costs for sophisticated individuals come in two kinds.
First, non-libertarian paternalist policies may prevent them from
buying some goods and services. For example, they might lose the
ability to use drugs prudently (e.g., to take them without getting
addicted) or to make risky investments that are nevertheless
prudent for them (e.g., because they constitute a small expected-
return-increasing addition to a diversified portfolio). Second,
sophisticated consumers may be the beneficiaries of firms’
schemes to exploit the vulnerable. If an industry is competitive,
and new firms enter when there are supernormal profits to be
made, no firm can expect to sustain supernormal profits. If so, the
practical, the criterion is a heuristic that organizes impressionistic
assessment.
51 See Securities Act 1978 (section 3) and Credit Contracts and Consumer
Finance Act 2003 (section 11).

54
ultimate beneficiaries of the exploitation of vulnerable consumers
must be other consumers, not the firms. If the hotel industry is
sufficiently competitive, for example, sophisticated travellers may
get accommodation below cost by avoiding the overpriced
telephone and minibar. If the credit-card market is sufficiently
competitive, sophisticated cardholders may get transaction
services and short-term credit for free by paying off their balance
each month.
90. It is natural also to take account of the costs incurred by the
government in implementing the regulation and the costs incurred
by firms in complying with it. The costs incurred by firms include
administrative costs such as preparing, printing, and distributing
information disclosures and any profits lost from having to change
business strategy.
Objections to libertarian and asymmetric paternalism
91. Libertarian and asymmetric paternalism have elicited a
number of objections. First, there are traditional concerns about
government failure (Glaeser 2006). Officials, ministers, and
regulators may in principle be able to select beneficial paternalistic
regulations, but they may have little incentive to do so. By
contrast, even irrational individuals have an incentive to improve
their own welfare. Moreover, the information required to design
beneficial paternalistic regulation might be difficult or impossible
to obtain. These problems, which are well known, are seen to be
more serious once the cognitive limitations of officials, ministers,
regulators, and judges are considered. According to Glaeser (2006,
13), ‘the flaws in human cognition should make us more, not less,
wary about trusting government decisionmaking’.

55
92. Second, there are concerns that paternalism may reduce the
learning that comes with choice and trial and error (Mill, 1859)
especially when default options are (reasonably) taken to convey
information about what is optimal. Along similar lines, McCaffery
and Baron (2004, p. 434) argue that ‘In private markets, arbitrage
mechanisms, which allow some to profit from the biases of others
... can be expected to reduce the effects of bias… In the public
sector, however, the absence of any simple, general arbitrage
mechanism ... gives reason to believe that the adverse effects of
cognitive biases can persist for long periods of time.’
93. Third, there are doubts that behavioural economists
correctly identify people’s true interests when decisions are
inconsistent over time. Many paternalist policies aim to reduce
behaviour that is enjoyable in the short run but harmful in the long
run. Implicitly, these proposals give more weight to a person’s
long-term interests than to his short-term interests. But does this
preference for the long-term have any justification (Whitman
2006)?
94. Fourth, there are concerns that soft paternalism is the first
step on a slippery slope that will lead to the banning of gambling,
alcohol, homosexuality, and more (Glaeser 2006, Rizzo and
Whitman 2009).
95. These objections all express concerns that the benefits of
paternalism are likely to be smaller, or its costs larger, than its
proponents think. None rules out paternalism in principle. Whether
government failure is likely to be greater or smaller than consumer
failure is something that needs to be assessed case by case.
Although hard paternalism reduces learning, libertarian
paternalism, at least, allows room for it (Loewenstein and Haisley

56
2008). Although short-term interests should not be ignored, there is
reason for giving more weight to long-term interests (see Bernheim
and Rangel 2007, 14). That there is a slippery slope from soft to
hard paternalism is an argument for vigilance, not for refusing all
soft paternalism. Yet the objections do call for caution.
96. Other objections, if accepted, are decisive. Sugden (2005)
argues against paternalism on the grounds that the role of
government should be to maximize people’s opportunities, not
their welfare. Mill (1859, ch. 3) likewise argues that ‘a man’s
mode of laying out his own existence is best not because it is the
best in itself, but because it is his own mode.’ If autonomy or
opportunity is given enough weight, at least some forms of
paternalism are ruled out.
97. If autonomy is not considered sacrosanct, we are left with
no single criterion for evaluating regulation, but instead a variety
of criteria, including consistency with unambiguous, informed
choice, effect on subjective well-being, and consistency with
traditional cost–benefit analysis. This is conceptually untidy,
compared with the world of neoclassical welfare economics. And it
leaves governments with considerable freedom of action: in the
absence of a unique criterion for judging whether a regulation is
desirable, they have more latitude to pick and choose among the
criteria to suit their prior preferences.
98. In practice, however, the lack of conceptual clarity is not
always be a problem—or that different from the current situation.
When there is widespread agreement about the goal of regulation,
and relatively little concern about its side effects and costs, the
insights of psychologists and behavioural economists can be
applied without philosophical problem. Working out ways of

57
ensuring that tourists drive on the left might be an example. When
people assign different values to different effects of a regulation,
behavioural economics may still inform decisions even if it doesn’t
lead to an unambiguous recommendation. And in New Zealand at
least, few regulations are currently subject to quantitative cost–
benefit analysis along neoclassical lines, so it is not as though
regulations are currently judged against a clear and unique
criterion.
Behavioural economics and the process of regulatory decision-
making
99. As important as substantive criteria for judging proposed
regulations are procedural rules and informal practices that govern
their development. There are several sources of procedural rules in
New Zealand. One is the Cabinet Office’s Cabinet Manual, which
requires that proposed regulations be accompanied by regulatory
impact analyses (Cabinet Office 2008, section 5.71). The Cabinet
Office’s Guide to Cabinet and Cabinet Committee Processes
(CabGuide), which provides more detail, states, ‘The government
wants to ensure that proposals involving regulatory options are
subject to careful and robust regulatory impact analysis’ and it
requires that statements setting out the analysis describe the status
quo, the problem that is being addressed and the objectives of
regulation, a range of options for addressing the problem, and the
‘costs, benefits, and risks’ of those options. 52 Proposals identified
by a ‘preliminary impact and risk assessment’ as likely to have a
52 See http://cabguide.cabinetoffice.govt.nz/procedures/regulatory-impact-
analysis.

58
‘significant impact or risk’ are reviewed by a Regulatory Impact
Analysis Team in the Treasury.
100. The Government has recently tightened some of these
requirements (Cabinet Office 2009), and the Treasury has
published a handbook offering more detailed guidance on the rules
set out in the CabGuide (Treasury 2009). Among other things, the
Treasury’s guidance encourages the quantification of costs and
benefits and refers to the Treasury’s (2005) guidance on cost–
benefit analysis, although in practice most analyses do not include
much quantification.
101. Another source of guidance is the Legislation Advisory
Committee (2001–2007), which has produced a checklist of 83
questions for the developers of regulations to consider (pp. 14–18),
such as ‘Has the policy objective been clearly defined?’, ‘Have
those outside the Government who are likely to be affected by the
legislation been consulted?’, ‘Has sufficient time and consideration
been given to the preparation of the legislation?’ ‘Have vested
rights been altered? If so, is that essential? If so, have
compensation mechanisms been included?’ Is the legislation
consistent with the New Zealand Bill of Rights Act 1990?’ And
‘Should the legislation provide a right of appeal?’
102. In recent years proposals have been made for a regulatory-
responsibility law that would establish regulatory decision-making
rules in law, in the image of the Fiscal Responsibility Act (now
part of the Public Finance Act). The most recent version of the
regulatory-responsibility bill, prepared by Scott et al. (2009),
includes six principles that regulations should satisfy. The
principles are set out under the headings ‘rule of law’, ‘liberties’,
‘taking of property’, ‘taxes and charges’, ‘role of courts’, and

59
‘good law-making’. The liberties principle is particularly
interesting in relation to paternalistic regulation. In context it reads
as follows:
(1) The principles of responsible regulation are that, except
as provided in subsection (2), legislation should ...
(b) not diminish a person’s liberty, personal security,
freedom of choice or action, or rights to own, use, and
dispose of property, except as is necessary to provide for,
or protect, any such liberty, freedom, or right of another
person…
(2) Any incompatibility with the principles is justified to the
extent that it is reasonable and can be demonstrably justified
in a free and democratic society.
103. Regulatory-responsibility legislation is often supported on
the ground that too much regulation is passed, imposing excessive
costs on businesses and, directly or indirectly, citizens. A related
but distinct concern is that regulation sometimes responds too
hastily to scandals and crises—that it follows the logic of the PM
of Yes, Prime Minister, who said during one crisis: ‘Something
must be done. This is something, therefore we must do it’.53
Hoping to avoid this trap, the New Zealand government has said
that it will ‘[r]esist the temptation or pressure to take a regulatory
decision until [it has] considered the evidence, advice and
consultation feedback.’54
53 http://www.yes-minister.com/polterms.htm.
54 ‘Government statement on regulation: better regulation, less regulation’,
issued in 2009 by the Minister of Finance and the Minister for Regulatory
Reform.

60
104. Concern to avoid hasty regulation predates behavioural
economics, but is reinforced by it. Procedural constraints, such as
requirements for regulatory impact or cost–benefit analysis, slow
decision-making and encourage deliberation, potentially reducing
hindsight bias, the overweighting of vivid, recently realized risks,
and other problems with intuitive decision-making. As Sunstein
argues, behavioural economics provides a justification for the use
of cost–benefit analysis even as it undermines its conceptual
foundation. It may also suggests ways in which the requirements
for deliberation might be improved.
105. One area where it is relevant is the analysis of consumer-
protection law and other regulations that have a rationale that is at
least partly paternalistic. For example, it may help decision makers
judge whether an exception to the liberties principle of the
regulatory-responsibility bill ‘is reasonable and can be
demonstrably justified in a free and democratic society’.
Libertarian-paternalist regulations are consistent with the liberties
principle, because they do not infringe individuals’ liberties. But
what about regulations that do limit individual liberties, and
probably cannot be justified on externality grounds, such as those
that require a cooling-off period before a decision is made? Here, it
seems reasonable to require that the asymmetric-paternalist
approach to cost–benefit analysis set out above be applied. Some
questions relevant to such an analysis are set out in table 3.

61
Table 3 Checklist of questions for paternalistic regulations
Extent of the problem and possible benefits
a) Is there evidence that people systematically make decisions
in the domain covered by the proposed regulation that do
not further their own interests?
b) Do the people making the apparent mistakes consider, on
reflection and with good information, that their decisions
are mistakes?
c) Is the domain of the proposed regulation one in which
people have good opportunities to learn from their
mistakes, or one in which feedback is rare or noisy or too
late to be useful? Is it one in which competition between
firms is likely to protect people?
d) Is there evidence that some people are vulnerable to the
mistakes while others are not?
e) Has a similar regulation been tested empirically, in the lab
or in randomized field trials or by econometric methods?
Possible costs and problems
f) Is the regulation likely to create indirect costs for the
intended beneficiaries?
g) Does the regulation reduce the ability of sophisticated
consumers (or investors) to pursue their interests?
h) Is the proposed regulation likely to strengthen or undermine
any social norms that help solve the problem that regulation
is intended to solve?
i) Do officials, ministers, and regulators know enough, and
have sufficiently strong motivation, to implement a
regulation that improves the decisions overall?
Quantifying the costs and benefits
j) What is the expected net benefit of the regulation to the
vulnerable?
k) What is the expected costs to the sophisticated?
l) What is the expected administrative cost of the regulation?
m) What is the expected cost imposed on firms by the
regulation, including the cost of lost profits?
n) What is the best estimate of net benefit of the regulation?

62
106. Another area where behavioural economics may help shape
regulatory analysis is environmental protection. As discussed
above, research on social norms shows that we are sometimes
better at cooperating in tragedies of the common than would be
expected if we were simply self-interested and that regulations that
create material incentives to behave well can sometimes undermine
social norms, and thus have a muted or even negative impact. If an
imperfectly designed or implemented regulation is worse than no
regulation, one option is of course to do nothing. Another is to
strengthen the regulation or its enforcement so that the material
incentives it creates are strong enough to compensate for any
reduction in people’s propensity to cooperative voluntarily. A third
is to design and introduce the regulation in a way that promotes
social norms. An example of a regulation that is apparently
successful in this respect is an Irish levy on plastic bags provided
by shops. Designed mainly to reduced visual pollution, the levy
was introduced after extensive consultation and publicity, and
seems to have been effective not only because it increased the
price of the bags to consumers but also because it helped create a
norm against unnecessary use of bags.55
107. More generally, Gintis et al (2005, 4) argue that the design
of regulation must take account of possible interactions between
material incentives and social norms:
55 See Convery, McDonnell, and Ferreira (2007). A recently imposed tax of
5 US cents a bag in the District of Columbia has been said to be effective
despite its small value for another behavioural-economic reason, discussed
by Ariely (2010, ch. 3), namely the salience of the difference between a zero
price and any positive price.

63
In 1754, five years before the appearance of [Adam]
Smith’s Theory of Moral Sentiments, David Hume advised
‘that, in contriving any system of government … every man
ought to be supposed to be a knave, and to have no other
end, in all his actions, than his private interest….’ However,
if individuals are sometimes given to the honourable
sentiments about which Smith wrote, prudence recommends
an alternative dictum: Effective policies are those that
support socially valued outcomes not only by harnessing
selfish motives to socially valued ends, but also by evoking,
cultivating, and empowering public-spirited motives.
108. The challenge is to work out how to do that, while keeping
in mind that regulation needs to be able to cope with knaves in the
regulated population and the possibility of knaves among the
regulators. Although clear-cut solutions do not appear to be
available, the research suggests that governments should be wary
of implementing regulations that don’t enjoy reasonably
widespread support and that it should use consultation to build
support for regulations, being willing to compromise on what
seems to be the best scheme if opposition reflects popular opinion,
not just the views of well-organized interests.
109. A risk with requirements for regulatory analysis is that they
create extra work without changing decisions and are therefore
viewed merely as box-ticking exercises. Research on social norms
and intrinsic motivation suggests that this is particularly likely if
the requirements are externally imposed and run counter to
traditional practice. Research on confirmation bias suggests that
requirements for analysis that are applied after officials have
reached a view may be especially ineffective. Together, the two
lines of research suggest that mandatory regulatory impact
analyses undertaken shortly before the submission of a proposal to

64
Cabinet should be supplemented by critical analysis early in the
development of regulation, but that this analysis should not be a
formal requirement, but rather something encouraged by skilful
management.
110. There is also a case for general-purpose debiasing of
regulatory decisions, that is, for training courses designed to help
officials and ministers make better judgments and decisions.
Effective debiasing appears to be difficult, but there is some
evidence that awareness of biases and consideration of their effect
during decision-making can improve decisions. Bazerman and
Moore (2009) is a text written with debiasing in mind. Table 4 sets
out a list of debiasing questions that officials might consider at
some point in the development of a regulation.
Table 4 Debiasing questions for developers of regulation
a) Is the evidence for the problem based on a sample that is
large enough to draw conclusions from?
b) Are we likely to be giving too much weight to a problem
because of its vividness or recent prominence in the media?
c) What problems might we be ignoring because they are not
salient?
d) Does hindsight bias make the problem seem obvious when
it is not?
e) Do we dislike a proposal for change because we are
suffering from status quo bias?
f) Does the preferred option still look the best if we frame the
choice differently?
g) Are we overconfident in our judgments about the problem
and the possible solutions?
h) Suppose the regulation fails: what is the most likely reason
for its failure?

65
111. A debiasing device that has proved popular with some
businesses is the project premortem, in which it is assumed that a
project has failed and reasons for its hypothetical failure are sought
(Klein 2007). The idea is that this makes it easier for dissenters to
point out the project’s weaknesses, especially if it has high-level
support. Policy development in New Zealand usually succeeds in
bringing arguments against new policies into the open, in part
because consultation ensures that those who expect to lose from a
regulation have an opportunity to state their case. But there may
still be value in fostering internal debate by means of a regulatory
pre-mortem. A policy team could be asked to suppose that a
regulation had failed to achieve its stated aim and to speculate on
what had hypothetically gone wrong. Analogously, policy
ministries might periodically hold internal inquiries designed to
elicit views on missing or inadequate regulations. Analysts could
be asked to imagine that a scandal had occurred that had revealed a
gap or shortcoming in existing regulation and then given a chance
to identify that gap or shortcoming.
112. Finally, because judgements about the effects of regulation
are difficult, we are likely to be overconfident about them.
Regulation is often the subject of strong disagreement, partly
reflecting differences in values, but also reflecting differences over
the facts. Both the supporters and the opponents of a regulation
probably overestimate the strength of their case. Knowledge of the
tendency to be overconfident may encourage more empirical
testing of proposed regulations, in laboratory experiments or
randomized field trials. Experiments can’t answer all relevant
questions, including questions about the effect of policies on social
norms. They also conflict both with the view that all citizens
should receive equal treatment and with the desire of policymakers
to appear knowledgeable. But the debate about KiwiSaver, for

66
example, would now be more informed if trials had shed light on
the effectiveness of, among other things, default enrolment without
subsidies.
Conclusion
113. Psychologists and behavioural economics have by now
presented enough evidence to demonstrate that human behaviour
deviates systematically from the rationality assumed by
neoclassical economics and the self-interest assumed in many of its
applications. The nature and extent of the deviations will remain a
matter of debate for many years to come, but the existing findings
can already help inform the design of regulation. Among other
things, they suggest greater use of simpler information disclosure,
carefully chosen default options, and devices that impose a delay
between the impulse to borrow or gamble (say) and the ability to
do so. It doesn’t follow that neoclassical economics can be ignored
in the design of regulation. In many areas it remains the best or
only economic theory available. However, the evidence against
neoclassical economics—and the absence of any behavioural-
economic theory of comparable scope—do imply that more weight
should be given to empirical testing in the design and evaluation of
regulation.

67
References
Agarwal, Sumit, John C. Driscoll, Xavier Gabaix, and David Laibson. 2008.
Learning in the credit card market. NBER Working Paper No. 13822.
Ainslie, George. 1975. Specious reward: A behavioural theory of
impulsiveness and impulse control. Psychological Bulletin 82(4): 463–496.
Akerlof, George A., and Robert J. Shiller. 2009. Animal spirits: How human
psychology drives the economy and why it matters for global capitalism.
Akerlof, George A., and William T. Dickens. 1982. The economic
consequences of cognitive dissonance. American Economic Review 72(3):
307–319
Allais, Maurice. 1953. Le comportement de l’homme rationnel devant le
risque: Critique des postulats et axiomes de l’école Américaine.
Econometrica 21(4): 503–546.
Andreoni, James. 1995. Cooperation in public goods experiments: Kindness or
confusion. American Economic Review 85(4): 891–904.
Ariely, Dan. 2010. Predictably irrational: The hidden forces that shape our
decisions. Revised and expanded edition. Harper Perenial.
Ariely, Dan, George Loewenstein, and Drazen Prelec. 2003. ‘Coherent
arbitrariness’: Stable demand curves without stable preferences. Quarterly
Journal of Economics 73–105.
Ashraf, Nava, Dean Karlan, and Wesley Yin. 2006. Tying Odysseus to the
mast: Evidence from a commitment savings product in the Philippines.
Quarterly Journal of Economics 121: 635–672.
Ausubel, Lawrence M. 1991. The failure of competition in the credit card
market. American Economic Review 81(1): 50–81.
Barber, Brad M., and Terrance Odean. 2001. Boys will be boys: Gender,
overconfidence, and common stock investment. Quarterly Journal of
Economics 116(1): 261–292.
Bazerman, Max H., and Don A. Moore. 2009. Judgement in managerial
decision making. Seventh edition. John Wiley & Sons
Becker, Gary S., and Kevin M. Murphy. 1988. A theory of rational addiction.
The Journal of Political Economy 96(4): 675–700.
Benjamin, Daniel, and David I. Laibson. 2003. Good policies for bad
governments: Behavioural political economy.
Bentham, Jeremy. 1798. Principles of morals and legislation.

68
Bernheim, B. Douglas, and Antonio Rangel. 2007. Behavioural public
economics: Welfare and policy analysis with non-standard decision makers.
Chapter 2 of Diamond and Vartianen (2007).
Bernheim, B. Douglas, and Antonio Rangel. 2008. Behavioural public
economics. The New Palgrave: Behavioural and experimental economics,
edited by Steven N. Durlauf and Lawrence E. Blume.
Bernheim, B. Douglas, and Antonio Rangel. 2009. Beyond revealed
preference: Choice-theoretic foundations for behavioural welfare
economics. Quarterly Journal of Economics 51–54.
Bertrand, Marianne, Dean Karlan, Sendhil Mullainathan, Eldar Shafir, and
Jonathan Zinman. 2010. What’s advertising content worth? Evidence from
a consumer credit marketing field experiment. Quarterly Journal of
Economics 125(1): 263–305.
Bevan, Bill, Bob Dugan, Virginia Grainer. 2009. Consumer law. Wellington:
LexisNexis.
Binmore, Ken, and Avner Shaked. 2010. Experimental economics: Where
next? Journal of Economic Behavior and Organization 73(1): 87–100.
Bowles, Samuel. 2008. Policies designed for self-interested citizens may
undermine the ‘moral sentiments’: Evidence from economic experiments.
Science 320: 1605–1609.
Brown, Jennifer, Tanjim Hossain, and John Morgan. 2010. Shrouded attributes
and information suppression: Evidence from the field. Quarterly Journal of
Economics 125(2): 859–876.
Cabinet Office. 2008. Cabinet Manual. Department of the Prime Minister and
Cabinet. New Zealand Government.
Cabinet Office. 2009. Minute of decision: Delivering better regulation and less
regulation. Annex to CAB Min (09) 27/11
Camerer, Colin F. 1995. Individual decision making. In The handbook of
experimental economics, ed. John H. Kagel and Alvin E. Roth, 587–704.
Princeton, NJ: Princeton University Press.
Camerer, Colin F. 2003. Behavioral game theory: Experiments in strategic
interaction. Princeton, NJ: Princeton University Press.
Camerer, Colin, George Loewenstein, and Matthew Rabin (eds). 2003.
Advances in behavioral economics. Princeton University Press.
Camerer, Colin, Samuel Issacharoff, George Loewenstein, Ted O’Donoghue,
and Matthew Rabin. 2003. Regulation for conservatives: Behavioral

69
economics and the case for ‘asymmetric paternalism’. University of
Pennsylvania Law Review 151: 1211–254.
Cameron, Lisa A. 1999. Raising the stakes in the ultimatum game:
Experimental evidence from Indonesia. Economic Inquiry 37(1): 47–59.
Caplan, Bryan. 2010. The Cartoon Introduction to Economics: Nice, but short
on ridicule. Blog post, 7 January 2010.
Caplin, Andrew, and Andrew Schotter (editors). 2008. The foundations of
positive and normative economics: A handbook. Oxford University Press.
Cardenas, Juan Camilo, John Stranlund, and Cleve Willis. 2000. Local
environmental control and institutional crowding-out. World Development
28: 1719–1733.
Carhart, Mark M. 1997. On persistence in mutual fund performance. Journal
of Finance 52(1): 57–82.
Coase, Ronald H. 1984. The new institutional economics. Journal of
Institutional and Theoretical Economics 140: 229–231.
Convery, Frank, Simon McDonnell, and Susana Ferreira. 2007. The most
popular tax in Europe? Lessons from the Irish plastic bags levy.
Environmental Resource Economics 38:1–11.
Croson, Rachel T. A. 2008. Public goods experiments. In Steven N. Durlauf
and Lawrence E. Blume (eds), The new Palgrave dictionary of economics.
Second edition. Palgrave Macmillan, 2008.
Dawes, Robyn M., and Richard H. Thaler. 1988. Anomalies: Cooperation.
Journal of Economic Perspectives 2(3): 187–197.
De Long, J. Bradford, Andrei Shleifer, Lawrence H. Summers, and Robert J.
Waldman. 1990. Noise trader risk in financial markets. Journal of Political
Economy 98(4): 703–738.
DellaVigna, Stefano. 2008. Psychology and economics: Evidence from the
field. Journal of Economic Literature 47(2): 315–372.
Dennett, Daniel. 2003. Freedom evolves. Penguin
Diamond, Peter, and Hanny Vartiainen (editors). 2007. Behavioral economics
and its applications. Princeton: Princeton University Press.
Dworkin, Gerald. 1972. Paternalism. The Monist 56(1): 64–84.
Dworkin, Gerald. 2010. Paternalism. Stanford Encyclopedia of Philosophy.
Earl, Peter E., and Simon Kemp. 2002. The Elgar companion to consumer
research and economic psychology.
Edgeworth, Francis Y. 1881. Mathematical psychics.

70
Edwards, Ward. 1982. Conservatism in human information processing. Ch. 25
of Kahneman and Tversky (1982), pp 359–369.
Ellsberg, Daniel. 1961. Risk, ambiguity, and the savage axioms. Quarterly
Journal of Economics 75(4): 643–669
Epstein, Richard A. 2006. Behavioral economics: Human errors and market
corrections. University of Chicago Law Review 73: 111–132.
Fehr, Ernst, and Bettina Rockenbach. 2003. Detrimental effects of sanctions on
human altruism. Nature 422: 137–140.
Fehr, Ernst, and Klaus M. Schmidt. 1999. A theory of fairness, competition,
and cooperation. Quarterly Journal of Economics 114: 817–868.
Fehr, Ernst, and Simon Gächter. 2002. Altruistic punishment in humans.
Nature 415: 137–140.
Fischhoff, Baruch. 1982. For those condemned to study the past: Heuristics
and biases in hindsight. Ch. 23 of Kahneman, Slovic, and Tversky (1982).
Flannery, Mark J. 1998. Using market information in prudential bank
supervision: A review of the U.S. empirical evidence. Journal of Money,
Credit, and Banking 30(3): 273–305.
Fountain, John, and Philip Gunby. 2010. Ambiguity, the certainty illusion, and
the natural frequency approach to reasoning with inverse probabilities. New
Zealand Economic Papers.
Frank, Robert H. 2005. Positional externalities cause large and preventable
welfare losses. American Economic Review Papers and Proceedings 95(2):
137–141.
Frederick, Shane, George Loewenstein, Ted O’Donoghue. 2002. Time
discounting and time preference: A critical review. Journal of Economic
Literature 40(2): 351–401.
Frederick, Shane. 2005. Cognitive reflection and decision making. Journal of
Economic Perspectives 19(4): 25–42.
French, Kenneth R., Martin N. Baily, John Y. Campbell, John H. Cochrane,
Douglas W. Diamond, Darrell Duffie, Anil K. Kashyap, Frederic S.
Mishkin, Raghuram G. Rajan, David S. Scharfstein, Robert J. Shiller, Hyun
Song Shin, Matthew J. Slaughter, Jeremy C. Stein, Rene M. Stulz. 2010.
The Squam Lake report: Fixing the financial system. Princeton University
Press.
Frey, Bruno S., and Alois Stutzer. 2002. What can economists learn from
happiness research? Journal of Economic Literature 40: 402–435.

71
Frey, Bruno. 1997. A constitution for knaves crowds out civic virtues.
Economic Journal 107(4): 1043–1053.
Friedman, Milton. 1953. The methodology of positive economics. Essays in
positive economics. Chicago University Press.
Fudenberg, Drew. 2006. Advancing beyond advances in behavioral
economics. Journal of Economic Literature 44: 694–711
Gabaix, Xavier, and David Laibson. 2006. Shrouded attributes, consumer
myopia, and information suppression in competitive markets. Quarterly
Journal of Economics 121(2): 505–540.
Gibson, John, and Trinh Le. 2008. How much new savings will KiwiSaver
produce? University of Waikato, Department of Economics, Working Paper
03/08. March.
Gigerenzer, Gerd. 1996. On narrow norms and vague heuristics: A reply to
Kahneman and Tversky. Psychological Review 103(3): 592–596.
Gigerenzer, Gerd. 2003. Reckoning with risk. Penguin.
Gilbert, Daniel T., Elizabeth C. Pinel, Timothy D. Wilson, Stephen J.
Blumberg, and Thalia P. Wheatley. 2002. Durability bias in affective
forecasting. Chapter 16 in Gilovich, Griffin, and Kahneman (2002).
Gilovich, Thomas, Dale Griffin, and Daniel Kahneman. 2002. Heuristics and
biases: The psychology of intuitive judgment. Cambridge: Cambridge
University Press.
Gintis, Herbert, Samuel Bowles, Robert T. Boyd, and Ernst Fehr (editors).
2005. Moral Sentiments and Material Interests: The Foundations of
Cooperation in Economic Life. MIT Press.
Glaeser, Edward L. 2006. Paternalism and psychology. University of Chicago
Law Review 73: 133–
Gneezy, Uri, and Aldo Rustichini. 2000. A fine is a price. Journal of Legal
Studies 29(1): 1–17.
Grether, David M., and Charles R. Plott. 1979. Economic theory of choice and
the preference reversal phenomenon. American Economic Review 69(4):
623–638.
Griffin, Dale, and Amos Tversky. The weighing of evidence and the
determinants of confidence. Ch. 13 of Gilovich, Griffin, and Tversky
(2002).
Grimes, Arthur. 1996. Public standards and private monitoring: New Zealand’s
new banking supervision regime. Agenda 3(3): 277–286.

72
Grubb, Michael D. 2009. Selling to overconfident consumers. American
Economic Review 99(5): 1770–1807.
Gruber, Jonathan, and Botond Kozsegi. 2001. Is addiction rational? Theory
and evidence. Quarterly Journal of Economics (November): 1261–1303.
Güth, Werner, Rolf Schmittberger, Bernd Schwarze. 1982. An experimental
analysis of ultimatum bargaining. Journal of Economic Behavior and
Organization 3: 367–388.
Gul, Faruk, and Wolfgang Pesendorfer. 2008. The case for mindless
economics. In Caplin and Schotter (2008)
Heffetz, Ori, and Robert H. Frank. 2008. Preferences for status: Evidence and
economic implications. Johnson School Research Paper Series No. #05-09)
[Prepared for Jess Benhabib, Alberto Bisin, and Matthew Jackson (eds),
Handbook of social economics. Elsevier]
Henrich, Joseph, Robert Boyd, Samuel Bowles, Colin Camerer, Ernst Fehr,
Herbert Gintis. 2004. Foundations of human sociality: Economic
experiments and ethnographic evidence from fifteen small-scale societies.
Oxford University Press.
Hume, David. 1739–1740. Treatise of human nature.
Inland Revenue Department. 2009. KiwiSaver Evaluation, Annual report July
2008–June 2009. Prepared by Evaluation Services. September.
Johnson, Eric J., John Hershey, Jacqueline Mezsaros, and Howard Kunreuther.
1993. Framing, probability distortions, and insurance. Journal of Risk and
Uncertainty 7: 35–51. Reprinted in Kahneman and Tversky (2000).
Jolls, Christine, Cass R. Sunstein, and Richard Thaler. 1998. A behavioral
approach to law and economics. Stanford Law Review 50: 1471–1550.
Kachelmeier, Steven, and Mohamed Shehata. 1992. Examining risk
preferences under high monetary incentives: Experimental evidence from
the People’s Republic of China. American Economic Review 82(5): 1120–
1141.
Kahneman, Daniel, and Amos Tversky. 1979. Prospect theory: An analysis of
decision under risk. Econometrica 47(2): 263–291.
Kahneman, Daniel, and Amos Tversky. 2000. Choices, values, and frames.
Cambridge: Cambridge University Press.
Kahneman, Daniel, Ilana Ritov, Karen E. Jacowitz, and Paul Grant. 1993.
Stated willingness to pay for public goods: A psychological analysis.
Psychological Science 4:310–315.

73
Kahneman, Daniel, Jack L. Knetsch and Richard. H. Thaler. 1990.
Experimental tests of the endowment effect and the Coase theorem. Journal
of Political Economy 98: 1325–1348.
Kahneman, Daniel, Jack L. Knetsch and Richard. H. Thaler. 1991. The
endowment effect, loss aversion, and status quo bias. Journal of Economic
Perspectives 5(1): 193–206.
Kahneman, Daniel, Paul Slovic, and Amos Tversky. 1982. Judgment under
uncertainty: Heuristics and biases. Cambridge: Cambridge University
Press.
Kahneman, Daniel, Peter P. Wakker, and Rakesh Sarin. 1997. ‘Back to
Bentham’? Explorations of experienced utility. Quarterly Journal of
Economics 112: 375–405.
Kahneman, Daniel. 1994. New challenges to the rationality assumption.
Journal of Institutional and Theoretical Economics 150: 18–36. Kahneman,
Daniel. 2003. A psychological perspective on economics. American
Economic Review 93(2): 162–168.
Kant, Immanuel. 1793. Political writings. Cambridge. Edited by Reiss.
Kemp, Simon. 2002. Public goods and private wants: A psychological
approach to government spending. Cheltenham: Edward Elgar.
Klein, Gary. 2007. Performing a project premortem. Harvard Business
Review, September.
Köszegi, Botond, and Matthew Rabin. 2008. Revealed mistakes and revealed
preferences. In Caplin and Schotter (2008).
Laibson, David, Andrea Repeto, and Jeremy Tobacman. 2007. Estimating
discount functions with consumption choices over the lifecycle.
Laibson, David. 1997. Golden eggs and hyperbolic discounting. Quarterly
Journal of Economics 112(2): 443–477.
Layard, Richard. 2006. Happiness: Lessons from a new science. Penguin.
Ledyard, John O. 1995. Public goods: A survey of experimental research.
Chapter 2 of John H. Kagel and Alvin E. Roth (eds), Handbook of
experimental economics, Princeton, NJ: Princeton University Press.
Legislation Advisory Committee. 2001–2007. Guidelines on process and
content of legislation. 2001 edition and amendments.
Levine, David K. 2009. Is behavioral economics doomed? The ordinary versus
the extraordinary. Max Weber Lecture. 8 June.

74
Levitt, Steven D., and John A. List. 2007. What do laboratory experiments
measuring social preferences reveal about the real world? Journal of
Economic Perspectives 21(2): 153–174.
Lichtenstein, Sarah, and Paul Slovic. 1971. Reversals of preference between
bids and choices in gambling decisions. Journal of Experimental
Psychology 89: 46–55.
Lichtenstein, Sarah, and Paul Slovic. 1973. Response-induced reversals of
preference in gambling: An extended replication in Las Vegas. Journal of
Experimental Psychology 101: 16–20.
Lichtenstein, Sarah, Baruch Fischhoff, and Lawrence D. Phillips. 1982.
Calibration of probabilities: The state of the art to 1980. Ch. 22 of
Kahneman, Slovic, and Tversky (1982).
Lipkus, Isaac M., and J. G. Hollands. 1999. The visual communication of risk.
Journal of the National Cancer Institute Monographs 25: 149–163.
List, John A. 2003. Does market experience eliminate market anomalies?
Quarterly Journal of Economics 118: 41–71.
Loewenstein, George, and Emily Haisley. 2008. The economist as therapist:
Methodological ramifications of ‘light’ paternalism. Chapter 9 of Caplin
and Schotter (2008).
Loewenstein, George, and Scott Rick. 2010. Addiction. In The New Palgrave
Dictionary of Economics. Second Edition. Eds. Steven N. Durlauf and
Lawrence E. Blume. Palgrave Macmillan.
Loewenstein, George, and Ted O’Donoghue. 2006. We can do this the easy
way or the hard way: Negative emotions, self-regulation, and the law.
University of Chicago Law Review 77:
Loewenstein, George, Ted O’Donoghue, and Matthew Rabin. 2003. Projection
bias in predicting future utility. Quarterly Journal of Economics 118(4):
1209–1248.
Loewenstein, George. 2000. Emotions in economic theory and economic
behaviour. American Economic Review: Papers and Proceedings 90(2):
426–432.
Madrian, Brigitte, and Dennis Shea. 2001. The power of suggestion: Inertia in
401(k) participation and savings behaviour. Quarterly Journal of
Economics 116(4): 1149–1187
McCaffery, Edward. J., and Jonathon Baron. 2004. Heuristics and biases in
thinking about tax. In Proceedings of the 96th Annual Conference on
Taxation (2003), 434–443. Washington: National Tax Association.

75
Milgrom, Paul. 2008. What the seller won’t tell you: Persuasion and disclosure
in markets. Journal of Economic Perspectives 22(2):115–131.
Mill, John Stuart. 1859. On liberty.
Ministry of Economic Development. 2006. Behavioural analysis for policy:
New lessons from economics, philosophy, psychology, cognitive science,
and sociology. October 2006.
Nickerson, Raymond S. 1998. Confirmation bias: A ubiquitous phenomenon in
many guises. Review of General Psychology 2(2): 175–220.
Nozick, Robert. 1974. Anarchy, state, and utopia. Blackwell
O’Donoghue, Ted, and Matthew Rabin. 2003. Studying optimal paternalism,
illustrated by a model of sin taxes. American Economic Review, Papers and
Proceedings 93: 186–91.
Paling, John. 2003. Strategies to help patients understand risks (Education and
debate) BMJ 327: 745–748.
Pesendorfer, Wolfgang. 2006. Behavioral Economics Comes of Age: A
Review Essay on Advances in Behavioral Economics. Journal of Economic
Literature 44: 712–721.
Plott, Charles R., and Kathryn Zeiler. 2005. The willingness to pay–
willingness to accept gap, the ‘endowment effect’, subject misconceptions,
and experimental procedures for eliciting valuations. American Economic
Review 95(3): 530–545.
Posner, Richard. 1998. Rational choice, behavioural economics, and the law.
Stanford Law Review 50(5): 1551–1575.
Rabin, Matthew, and Richard H. Thaler. 2001. Anomalies: Risk aversion.
Journal of Economic Perspectives 15(1): 219–232.
Rabin, Matthew. 1993. Incorporating fairness into game theory and
economics. American Economic Review 83: 1281–1302.
Rabin, Matthew. 2000. Risk aversion and expected utility theory: A calibration
theorem. Econometrica 68(5): 1281–1292.
Rachlinski, Jeffrey J. 1998. A positive psychological theory of judging in
hindsight. University of Chicago Law Review 571.
Rawls, John. 1971. Theory of justice
Reserve Bank of New Zealand. 2010. Financial stability report.
Rizzo, Mario J., and Douglas Glen Whitman. 2009. Little brother is watching
you: New paternalism on the slippery slopes. Arizona Law Review

76
Roll, Richard. 1986. The hubris hypothesis of corporate takeovers. Journal of
Business 59(2): 197–216.
Salop, Steven, and Joseph Stiglitz. 1977. Bargains and ripoffs: A model of
monopolistically competitive price dispersion. Review of Economic Studies
44(3): 493–510.
Sandroni, Alvaro, and Francesco Squintani. 2007. Overconfidence, insurance,
and paternalism. The American Economic Review 97(5): 1994–2004.
Scott, Graham, Paul Baines, David Caygill, Richard Clarke, Jack Hodder, Don
Turkington, and Bryce Wilkinson. 2009. Report of the regulatory
responsibility taskforce. September.
Sen, Amartya. 1987. Ethics and economics. Blackwell
Shafir, Eldar, Peter Diamond, Amos Tversky. 1997. Money illusion. Quarterly
Journal of Economics 112(2): 341-374.
Snowberg, Erik, and Justin Wolfers. 2010. Explaining the favourite–long shot
bias: Is it risk-love or misperceptions. Journal of Political Economy 118(4):
723–746.
Stango, Victor, and Jonathan Zinman. 2009. Exponential growth bias and
household finance. Journal of Finance 64(6): 2807–2849
Starmer, Chris. 2008. Preference reversals. The New Palgrave: Behavioural
and experimental economics, edited by Steven N. Durlauf and Lawrence E.
Blume.
Stewart, Neil. 2009. The cost of anchoring on credit-card minimum
repayments. Psychological Science 20(1): 39–41.
Stigler, George J. 1981. Economics or ethics. The Tanner lectures on human
values, delivered at Harvard University, April 24, 25, and 28, 1980.
Stiglitz, Joseph, Amartya Sen, and Jean-Paul Fitoussi. 2010. Mis-measuring
our lives: Why GDP doesn’t add up. The report by the commission on the
measurement of economic performance and social progress. New York and
London: New Press.
Strotz, Robert H. 1956. Myopia and inconsistency in dynamic utility
maximization. Review of Economic Studies 23(3): 165–180.
Sugden, Robert. 2004. The opportunity criterion: Consumer sovereignty
without the assumption of coherent preferences. American Economic
Review, 94(September 2004), 1014–1033
Sunstein, Cass R. 2000. Cognition and cost–benefit analysis. Journal of Legal
Studies 29(S2).

77
Sunstein, Cass R., and Richard Thaler. 2003. Libertarian paternalism is not an
oxymoron. University of Chicago Law Review
Svenson, Ola. 1981. Are we all less risky and more skilful than our fellow
drivers? Acta Psychologica 47: 143–148.
Thaler, Richard H (editor). 2005. Advances in behavioural finance, Volume II.
Princeton: Princeton University Press.
Thaler, Richard H. 1981. Some empirical evidence on dynamic inconsistency.
Economic Letters 8: 201–07.
Thaler, Richard H. 1993. The winner’s curse: Paradoxes and anomalies of
economic life. New York: Free Press.
Tocqueville, Alexis. 1835–1840. Democracy in America.
Tokeley, Kate. 2000. Consumer law in New Zealand. Butterworths.
Tooth, Richard. 2010. The regulation of consumer credit.
Treasury. 2005. Cost benefit analysis primer. New Zealand Government.
Treasury. 2009. Regulatory impact analysis: Handbook. New Zealand
Government
Tversky, Amos, and Daniel Kahneman. 1974. Judgement under uncertainty:
Heuristics and biases. Science 185: 1124–1131.
Tversky, Amos, and Daniel Kahneman. 1981. The framing of decisions and
the psychology of choice. Science 211(4481): 453–458.
Tversky, Amos, and Daniel Kahneman. 1984. Extensional versus intuitive
reasoning: The conjunction fallacy in probability judgement. Psychological
Review 91: 293–315. Reproduced as ch. 1 of Gilovich, Griffin, and
Kahneman (2002).
Tversky, Amos, and Daniel Kahneman. 1986. Rational choice and the framing
of decisions. Journal of Business 59(4), Part 2, S251–78.
Tversky, Amos, and Derek J. Koehler. 2002. Support theory: A nonextensional
representation of subjective probability. Ch. 25 of Gilovich, Griffin, and
Kahneman (2002).
Tversky, Amos, and Richard H. Thaler. 1990. Preference reversals. Journal of
Economic Perspectives 4(1): 193–205.
Weinstein, Neil D. 1989. Optimistic biases about personal risks. Science 246
(4935): 1232–33.
Welch, Ned. 2010. A marketer’s guide to behavioural economics. McKinsey
Quarterly. February.

78
Whitman, Glen. 2006. Against the new paternalism: Internalities and the
economics of self-control. February 22, 2006. Policy Analysis no. 563
(Cato)
Yaari, Menahem E. 1981. Rawls, Edgeworth, Shapley, Nash: Theories of
distributive justice re-examined. Journal of Economic Theory 24(1): 1–39.