School of Economics and Centre for Competition Policy

1
School of Economics and Centre for Competition
Policy University of East Anglia Norwich NR4
7TJ, United Kingdom
Why incoherent preferences do not justify
paternalism Robert Sugden Paper prepared for
seminar at Financial Services Authority, 5
October 2011
2
There is a long tradition of liberal thought in
which the market is seen as an institution in
which privately-motivated individual actions
produce consequences that are socially
beneficial... ... and often more beneficial
than can be produced by deliberate planning.
The most famous statement of this idea is Adam
Smiths Wealth of Nations.
2
3
The invisible hand by directing that industry
in such a manner as may be of the greatest value,
the merchant intends only his own gain, and he
is in this, as in many other cases, led by an
invisible hand to promote an end which was no
part of his intention. Nor is it always the
worse for society that is was no part of it.
Adam Smith, Wealth of Nations, 1776, p. 456.
3
4
The liberal tradition in economics isnt
dogmatically laissez-faire. But it favours
market and market-like mechanisms, even when
dealing with externalities, public goods, or
distributional issues ...
-- cash rather than in-kind transfers
-- Pigou taxes/ tradable permit solutions to
externality problems
-- cost-benefit analysis (as market
simulation) to determine provision of public
goods.
Background idea is still that spontaneous-order
market mechanisms tend to work well, i.e. the
invisible hand.
4
5
In this talk, I defend the idea of the invisible
hand against a new challenge, from behavioural
welfare economics, and sketch an alternative way
of reconciling behavioural and normative
economics. My perspective
-- Im a behavioural economist who is sceptical
of neoclassical rationality assumptions
-- and a normative economist with
classical-liberal, anti-paternalist sympathies
-- Im not arguing for a neo-con or libertarian
reconstruction of normative economics, just
trying to conserve/ refurbish a strand of thought
that has been a major part of economics for over
200 years.
5
6
Why the findings of behavioural economics cause
problems for normative economics
From the early 20th century, the dominant form of
economics has been neoclassical. A fundamental
assumption of neoclassical theory is that
individuals have coherent preferences over all
relevant economic outcomes, and act according to
those preferences ( maximise utility).
Coherent preferences are -- stable (i.e. not
subject to random or arbitrary variation)
-- context-independent (i.e. not affected by
arbitrary changes of framing of decision
problems)
-- internally consistent (i.e. satisfying
rationality principles such as transitivity).
6
7
In neoclassical welfare economics, the normative
criterion is the satisfaction of preferences
-- Fundamental theorems of welfare economics
show that competitive equilibrium is
Pareto-efficient, and that any Pareto-efficient
resource allocation can be achieved as a
competitive equilibrium if combined with
appropriate income transfers. Hence presumption
in favour of markets, with regulation to prevent
fraud and ensure competition.
-- If there are public goods or externalities,
interventions can simulate the efficiency
properties of markets if they are based on
cost-benefit analysis (i.e. use individuals
willingness to pay as the measure of value).
A recurring theme in neoclassical (and classical)
economics anti-paternalism, consumer
sovereignty. In neoclassical economics, this is
seen as equivalent to using preference-satisfactio
n as the normative criterion.
7
8
But the findings of behavioural economics (i.e.
use of research methods and theoretical ideas
adapted from psychology) cast doubt on whether
coherent preferences really exist. In many cases,
individuals economic behaviour reveals
incoherent preferences the incoherencies
(anomalies) are systematic and can be explained
psychologically. For example
-- preferences between two options depend on
which is perceived as the status quo (the
endowment effect, loss aversion)
-- preference between A and B varies according to
whether C is in the opportunity set (decoy
effect if C is clearly inferior to B, but not
to A, adding C to the set makes B more
attractive).
If preferences are incoherent, how can
preference-satisfaction be used as a normative
standard? (And so what is left of the claim
that competitive markets satisfy consumer
preferences?)
8
9
In behavioural economics, a consensus seems to be
developing around a particular response to this
problem libertarian paternalism/ asymmetric
paternalism/ soft paternalism/ behavioural
welfare economics.
This has become influential both within academia
and outside -- Richard Thaler and Cass
Sunsteins popular book Nudge -- Thaler recently
visited 10 Downing Street to advise the
behavioural insight team or Nudge unit on how
to apply his approach to public policy.
This presentation -- presents a critique of
this approach -- gives a very quick sketch of
an alternative approach that I am developing.
Based on a series of papers, including American
Economic Review 2004 Social Choice and Welfare
2007 Constitutional Political Economy 2008
Economics and Philosophy 2008 (with Bruni)
Journal of Environmental Economics and Policy
2009 review of Nudge in International Journal of
the Economics of Business 2009 Economics and
Philosophy 2010.
9
10
NB I am criticising the theoretical approach of
behavioural welfare economics, not any specific
policy proposals that have been made under the
libertarian paternalist banner. The issue I
address is What normative criterion should be
used in assessing proposals for the regulation of
markets?
10
11
Behavioural welfare economics
Two (remarkably similar) manifestos appeared in
2003 Cass Sunstein and Richard Thaler.
Libertarian paternalism is not an oxymoron.
University of Chicago Law Review 2003 The
academic paper that was expanded and popularised
as Nudge. Colin Camerer, Samuel Issacharoff,
George Loewenstein, Ted ODonaghue and Matthew
Rabin (2003). Regulation for conservatives
behavioral economics and the case for asymmetric
paternalism. University of Pennsylvania Law
Review 2003.
Each paper has a legal scholar as a co-author.
Otherwise, a roll-call of the great and the good
of American behavioural economics.
Titles (libertarian paternalism, regulation
for conservatives) signal that the authors will
propose interventions in the economy that have
traditionally been opposed by pro-market thinkers
but the authors arguments will be immune to
their opponents usual criticisms.
11
12
Ill focus on the version of behavioural welfare
economics presented by Sunstein and Thaler.
Richard Thaler
Cass Sunstein
And Ill focus on one of Sunstein and Thalers
central claims -- The findings of behavioural
economics show that paternalism is inevitable
the idea that there are viable alternatives to
paternalism is a misconception the
anti-paternalist position is incoherent, a
nonstarter. So ....
12
13
... planners/choice architects ought to be
consciously paternalistic we argue for
self-conscious efforts, by private and public
institutions, to steer peoples choices in
directions that will improve their own welfare
(Sunstein and Thaler, 2003).
But (especially in more recent work) their
recommendations are designed to make choosers
better off, as judged by themselves. STs
italics (What as judged by themselves means is
part of my topic.)
13
14
Why (according to Sunstein and Thaler) is
paternalism inevitable?
Sunstein and Thaler conceive of themselves as
advising a planner (later choice architect)
who is responsible for designing the presentation
of options to individuals.
Their favourite example the cafeteria. The
cafeteria director chooses the display of the
food items, knowing that this affects customers
purchases.
They say an anti-paternalists recommendation to
the director would be give customers what she
thinks they would choose on their own.
ST say this recommendation is meaningful only
if what the customer would choose can be
defined independently of the directors choice.
But it cant consumers lack well-formed
preferences, in the sense of preferences that are
firmly held and preexist the directors own
choices about how to order the relevant items in
the display. If the arrangement of the
alternatives has a significant effect on the
selections the customers make, then their true
preferences do not formally exist.
14
15
Since the cafeteria director cant avoid
influencing consumers choices, she should
recognise this and make choices that she thinks
would make the customers best off, all things
considered.
This is paternalism, but of an inoffensive kind,
because its only a nudge customers freedom of
choice is not constrained.
-- This respects autonomy (i.e. acceptable to
principled libertarians)
-- and ensures that coherent preferences are
satisfied.
15
16
What is the normative criterion?
The Nudge approach requires a criterion of
better off, all things considered, as judged by
herself. Compare neoclassical welfare economics,
which has (what was assumed to be) a well-defined
and objective (i.e. observer-independent)
criterion revealed preference.
ST spend much more time discussing how people
can be nudged than on how the planner decides in
which direction to nudge them (just unspecific
appeals to cost-benefit analysis without
saying that standard CBA uses preference-satisfact
ion as the criterion).
But their official position (as far as I can
decipher it) is to use an informed desire or true
preference criterion...
16
17
Starting point for Nudge individuals make
pretty bad decisions decisions they would not
have made if they had paid full attention and
possessed complete information, unlimited
cognitive abilities, and complete self-control.
In 2003 paper individuals are treated as not
acting in their own best interests if their
decisions are ones they would change if they had
complete information, unlimited cognitive
abilities, and no lack of willpower.
Implication a persons best interests, as
judged by himself correspond with the
preferences he would reveal if he paid full
attention and possessed complete information,
unlimited cognitive abilities, and no lack of
willpower.
Reasons for concern about this informed desire
criterion...
17
18
Reasons for concern 1. It uses a normative and
contestable concept of an ideally rational agent.
Full attention, complete information, unlimited
cognitive ability, complete self-control are all
normative concepts.
E.g. the obese customer who buys the Home Wrecker
at Hillbilly Hot Dogs.
On his first day in Huntington, W. Va., Jamie
Oliver spent the afternoon at Hillbilly Hot Dogs.
He learned how to perfect the Home Wrecker, the
eaterys famous 15-inch, one-pound hot dog (boil
first, then grill in butter). For the Home
Wrecker Challenge, the dog gets 11 toppings,
including chili sauce, jalapeños, liquid nacho
cheese and coleslaw. Finish it in 12 minutes or
less and you get a T-shirt. Times magazine 6
Oct 2009
18
19
What is true information about diet/ health
links? (In a model, complete information is an
objective concept but not in the real world!)
What is correct reasoning about health risks?
What is the distinction between changing ones
mind and lacking self-control?
Very easy for the nudger to attribute his own
judgements to the ideally rational version of
the nudgee. Sunstein and Thalers attributions
are very casual
19
20
E.g. arguing for nudges to advocate nudges
against obesity-inducing diets, smoking,
drinking Sunstein and Thaler just report
familiar statistics about the associated health
risks and conclude With respect to diet,
smoking, and drinking, peoples current choices
cannot reasonably be claimed to be the best means
of promoting their well-being.
No suggestion that ST need to verify that
nudgees are making bad decisions as judged by
themselves.
20
21
ST sometimes claim that nudgees have expressed
desires to make the choices they will be nudged
to make and so have acknowledged a self-control
problem. But again, very casual argument
The New Year resolution test How many
people vow to smoke more cigarettes, drink more
martinis, or have more chocolate donuts in the
morning next year?
On saving they cite survey evidence that
two-thirds of employees describe their savings
rate as too low while only one per cent
describe it as too high. ST such statements
are not meaningless or random they show that
people are open to a nudge and might even be
grateful for one.
Subtext Were only nudging, so paternalism
doesnt need rigorous justification.
21
22
Reasons for concern about the informed desire
criterion 2. It assumes that, deep down,
individuals have rational preferences.
ST are assuming that, inside every behavioural
human being, there is a neoclassical rational
agent. The rational agents optimal choices are
frustrated by imperfections that are external
to it, (lack of attention, imperfect information,
imperfect cognition, imperfect self-control).
This idea doesnt fit with the methodology of
behavioural economics the concept of true
preferences comes from rational choice theory,
not pychology.
Information, cognition and willpower are inert
without desires to act on. Superhuman agents who
had perfect information, perfect cognitive powers
and perfect willpower would still have to deal
with their actual desires, which are matters of
psychology, not rationality. Anomalies in
human decision-making (i.e. deviations from
rational-choice theory) may reflect the structure
of desires (e.g. loss aversion), not error.
If individuals dont have true preferences,
STs claim that they are only helping people to
choose what they really prefer is an illusion.
22
23
An alternative approach
Starting point notice the mismatch between the
liberal tradition (Hume, Smith, JS Mill...) and
the idea of defending the market by taking the
viewpoint of a planner, for whom individuals
choices are just a source of data. Why is the
addressee of normative economics the planner?
Who else could be the addressee?
Each of us as private individuals.
So, in relation to the market (and proposals for
regulation), each of us can ask What does the
market do for me? Does it work in my interests?
A contractarian defence of an economic system
shows each person severally that it tends to give
him what he wants there is no need to imagine a
planner.
23
24
Behavioural and neoclassical welfare economics
interpret the invisible hand argument as claiming
that markets are effective in satisfying given
preferences. (First we specify preferences, then
we ask whether the market satisfies them.)
I propose that we reformulate the invisible hand
argument as claiming that markets allow
individuals to satisfy their preferences,
whatever those preferences turn out to be. (We
evaluate the market from the perspective of
individuals who do not yet know what their
preferences will be.)
Preference inconsistencies disable the first
approach, but not the second.
24
25
Reconstructing the first fundamental theorem of
welfare economics
This theorem is generally seen as the core
neoclassical statement of the invisible hand
argument. Roughly, it says that competitive
markets are efficient in satisfying
preferences. Ill show how the theorem can be
reconstructed so that it doesnt refer to given
preferences. (Just the first step in a bigger
project ...)
First, Ill state the theorem more precisely, as
applied to a very simple exchange economy.
25
26
In this economy, there are many
individuals. There are many goods, all of which
are private. There is a fixed stock of each
good, so the only economic problem is to divide
these stocks between individuals. Any division
of goods between individuals is an allocation.
We start with an initial allocation (individuals
endowments). Individuals are then able to
exchange goods by mutual consent.
Competitive equilibrium is a list of prices, one
for each good, such that all markets clear (i.e.
for each good, total amount offered for sale
total amount demanded). Trade at these prices
brings about a new allocation.
Its assumed that each individual has a given
preference ranking over all bundles of goods, and
acts on this. An allocation is Pareto-optimal if
no feasible reallocation of goods between
individuals would make some individual better off
and no one worse off (in terms of their
preferences).
The theorem tells us every competitive
equilibrium is Pareto-optimal.
27
A sketch of a proof of the first fundamental
theorem
Deliberately, the proof is not quite standard.
It proves as much as possible using only the most
minimal assumptions about preferences, and
introduces rationality assumptions at the very
end. This allows us to identify desirable
properties of the market that dont depend on
rationality assumptions.
Preliminary. Competitive equilibrium can be
defined without using the concept of preference
(or utility). Conventionally, demand and
supply (and hence market-clearing) are defined
in terms of utility-maximising choices. But all
we need to assume is that individuals make
decisions about how much to buy and sell at the
prices that are on offer. These decisions need
not reveal consistent preferences (e.g. desired
holdings of goods may depend on endowments,
and/or on arbitrary framing features). All I
assume about preferences is that one good
(money) is always desired.
27
28
A possible doubt Competitive equilibrium is
defined as a list of market-clearing prices. In
neoclassical welfare economics, this is treated
as an idealised representation of the outcome of
real markets. For this idea to be plausible, do
we need to assume that individuals have
consistent preferences?
No. Its sufficient that trades are mediated by
profit-seeking professional traders (who are
rational in their professional activities). No
one needs to have consistent preferences over
bundles of goods. I show this in a paper in
American Economic Review 2004.
But (very relevant for financial regulation)
rational traders are assumed to be
non-colluding and profit-seeking, and to have
correct expectations (in equilibrium). Argument
would not work if traders did not bear the
downside risks of their trades.
28
29
Step 1 of proof In competitive equilibrium, all
opportunities for voluntary transactions have
been made available to individuals (severally).
Let Z be the initial allocation. Let X be the
allocation reached in a competitive equilibrium,
by trades from Z made at the price list P. What
would it mean to say that some opportunity for
voluntary transaction had not been made available?
Suppose that were so. Then there would be (a)
a feasible allocation Y (not the same as X),
reachable from X by some composite transaction
and (b) no party to that composite transaction
(i.e. no individual whose bundle in Y is not the
same as his bundle in X) has been offered his
part of the transaction and has rejected it.
But in fact, for any Y satisfying (a) at least
one party to that transaction was offered his
part of it through the market, and chose not to
take it.
29
30
How do we know this?
X (the outcome of the market) and Y (the outcome
of the composite transaction) contain exactly the
same goods. (In this economy, the only possible
transactions are exchanges.) So value of Y at
market prices value of X at market prices.
For each party to the transaction we can ask
whether his Y-bundle is worth more or less (at
market prices) than his X-bundle. Clearly, it
cant be the case that every party gains value
in the transaction.
So there must be at least one individual i, who
is a party to the transaction, and whose Y-bundle
is worth no more than his X-bundle.
But i had the opportunity to buy his Y-bundle (or
one unambiguously better) at the market prices,
and chose not to do so.
Which proves In competitive equilibrium, all
feasible opportunities for voluntary transactions
have been made available to individuals.
30
31
Step 2 of proof If each individual acts in
accordance with a stable preference ordering,
then competitive equilibrium is Pareto-optimal.
For simplicity, I assume that preferences are
strictly convex (i.e. there is a uniquely optimal
choice from every budget constraint) this isnt
essential.
From Step 1, we know that if X is the outcome of
the market, then for every other feasible
allocation Y, there is at least one individual i
who chose not to take his Y-bundle in exchange
for his X-bundle. So, i prefers his X-bundle.
So, every feasible reallocation of X makes at
least one individual worse off. QED.
31
32
The significance of this proof
Behavioural welfare economists claim that the
invisible hand argument is undermined if
individuals lack coherent preferences. But
coherent preferences are needed only for Step 2.
Step 2 converts In competitive equilibrium, all
opportunities for voluntary transactions are made
available into Competitive equilibrium is
Pareto-optimal. This step switches from
individuals perspectives to the planners
perspective. The planner is trying to maximise
each individuals welfare she treats given
preferences as indicators of welfare so she
needs to show that opportunities for voluntary
transactions translate into the satisfaction of
given preferences.
But For each of us as individuals, assessing
what the market does for us, isnt All
opportunities for voluntary transactions are made
available sufficient? Do we need the planners
viewpoint?
32
33
Some generalisations of this result 1.
Generalise to many time periods.
We show that the market makes available all
feasible opportunities for voluntary
transactions, where these include exchange
between an individuals period 1 self and
period 2 self. The opportunity criterion
treats these as opportunities for the individual
as a continuing person (or locus of
responsibility). E.g. The wine economy
arbitrageurs anticipate systematic changes in
individuals preferences.
33
34
2. Generalise to public goods?
An example the argument extends to economies in
which emission of pollutants is regulated by cap
and trade systems. Given any cap on emissions
(treated as an exogenous constraint) a cap and
trade regime makes available all feasible
opportunties for voluntary transactions in
private goods, given that constraint. Again no
assumption that preferences are coherent.
34
35
Conclusion A competitive market provides maximal
opportunities for voluntary transactions, i.e.
transactions that individuals choose to make at
the time they make them.
Or let us say that a person is willing to pay
for a good if she is willing to give up what
would induce others (or herself at another point
in time) to supply it. Then my answer to the
question What does the market do for you? is
The market gives each person, rational or
irrational, what she wants and is willing to pay
for, when she wants it and is willing to pay for
it.
35
36
Thank you for listening.
36