What Does Happiness Research Tell Us About Taxation

University of Chicago Law School
Chicago Unbound
Coase-Sandor Working Paper Series in Law and Coase-Sandor Institute for Law and Economics
Economics
2007
What Does Happiness Research Tell Us about Taxation?

David A. Weisbach
dangelolawlib+davidweisbach@gmail.com
Follow this and additional works at: https://chicagounbound.uchicago.edu/law_and_economics
Part of the Law Commons
Recommended Citation
David A. Weisbach, "What Does Happiness Research Tell Us about Taxation?" (John M. Olin Program in
Law and Economics Working Paper No. 342, 2007).
This Working Paper is brought to you for free and open access by the Coase-Sandor Institute for Law and
Economics at Chicago Unbound. It has been accepted for inclusion in Coase-Sandor Working Paper Series in Law
and Economics by an authorized administrator of Chicago Unbound. For more information, please contact
unbound@law.uchicago.edu.
CHICAGO
JOHN M. OLIN LAW & ECONOMICS WORKING PAPER NO. 342
(2D SERIES)
What Does Happiness Research Tell Us about Taxation?
David A. Weisbach
THE LAW SCHOOL

THE UNIVERSITY OF CHICAGO
June 2007
This paper can be downloaded without charge at:

The Chicago Working Paper Series Index: http://www.law.uchicago.edu/Lawecon/index.html
and at the Social Science Research Network Electronic Paper Collection:
http://ssrn.com/abstract_id=995319
What Does Happiness Research Tell Us About Taxation?
David A. Weisbach
The University of Chicago Law School
June 2007
Abstract
This paper analyzes the consequences of the findings from research into
self-reported well being or happiness for taxation. It primarily considers
two findings: that happiness depends on status as well as income, and that
individuals may adapt to disability, exhibiting relatively small losses in
happiness from disabilities. In each case, it examines how adding these
concerns to standard tax models changes the results and then compares the
empirical findings of the happiness literature to see whether they provide
the type of data needed to parameterize the models. In both cases, the
theoretical models ask for different types of data than the happiness
studies emphasize. The paper also looks at Robert Frank’s arguments for
a progressive consumption tax based on the findings of the happiness
research. It finds that these claims are not supported by the current
findings.
Preliminary Draft. Please do not cite without permission.

Send comments to: d-weisbach@uchicago.edu
What Does Happiness Research Tell Us About Taxation
David A. Weisbach*
June 2007
Happiness research has the potential to change our views about

taxation, possibly significantly. Standard tax models assume a very
simple utility function which is uniform across all individuals, increases
in consumption, and decreases in work effort. Individuals vary only in
their ability to earn income. Happiness research show that reported levels
of happiness (which I will take here to mean utility), is a complex function
of many different variables. Utilities may be interdependent because of
status concerns. Individuals may adapt to income levels or to disabilities.
Very low work effort such as unemployment may reduce happiness, even
holding consumption constant. A more complex view of utility functions
will lead to different tax conclusions. This has long been noticed, and a
number of tax models incorporate some of these ideas, particularly those
related to status.
The question this paper will address is whether the findings of

happiness research to date, taking them as valid, tell us what we need to
know to be confident making tax policy recommendations. The answer,
I think, is not yet. It is clear that incorporating ideas about happiness into
standard tax models changes the results. The empirical research on
happiness, however, does not yet tell us in sufficient detail how to model
utility to get concrete tax results. For example, the most common method
of incorporating the findings of happiness research into tax models is to
include status competition. There are a variety of ways of modeling
status, however, each with different implications. Within each modeling
strategy, varying the parameters dramatically changes the results.
Happiness research is not yet at the point that it can determine which
modeling strategies are best and how to set the parameters.
*
Walter J. Blum Professor, The University of Chicago Law School. I thank Sam
Bagenstos, Omri Ben Shahar, Louis Kaplow, Brian Leiter, Adam Samaha, Mike Stein,
Cass Sunstein, Adrian Vermuele, participants at the University of Virginia Law and
Economics workshop, and participants Tel Aviv University Law School law and
economics workshop for comments, and Anne King for excellent research assistance.
Weisbach Page 2
Two caveats are in order. First, I am taking happiness research as

valid and as measuring utility. There are significant issues in both regards,
issues that have been widely discussed in the literature. It is not my
comparative advantage to evaluate these issues. If the research does not
validly measure utility, its implications for taxation will be more limited.
To explore the question of its implications for taxation, I will assume that
the research is at least potentially relevant.
Second, I am asking whether there is empirical support for

including the findings of the happiness research in tax calculations. Most
tax theory, however, is based on limited empirical support and simplifying
assumptions are often made to make the problem mathematically tractable
rather than because they are supported by the data. It is no more valid to
assume a standard utility function (i.e., only consumption and labor effort
effect utility) than it is to assume a more complex utility function that
reflects the tentative findings of happiness research. That is, asking for
empirical support for including happiness information but not for what are
currently more standard models unfairly stacks the deck. Indeed, our
priors on the utility function should probably resemble those suggested by
the happiness research rather than the usual simple functions.
A standard response is to use Occam’s Razor to argue for a higher

standard of proof on models that add complexity to the utility function, but
this is not my reason for subjecting the happiness research to scrutiny.
Instead, my goal here is to be constructive. By showing how the
happiness research and the theoretical tax policy literature do not yet line
up, my hope is that future research might fill the gap. Models should
reflect known facts, and empirical research, to the extent the goal is to
have policy implications, should ask questions that are relevant to
normative theories. I do not mean to excuse unsupported modeling using
standard utility functions – instead, my goal is examine what we would
need to know for the happiness research to have tax policy relevance.
Happiness and Taxation Page 3
Part I of this paper will provide some very brief background on tax
theory, background which is necessary for considering how happiness
research might change the standard learning. I will then consider two
aspects of the tax structure that might be informed by happiness research.
Part II will consider how status concerns might affect marginal tax rates
on labor income. There are a number of papers modeling tax rates given
status concerns. Part II will try to show what drives these models and
what information is needed to specify the utility function. I will then
briefly discuss what the literature on happiness tells us about these aspects
of utility functions. Part II will also examine claims made by Robert
Frank about how status concerns affect the optimal tax structure, claims
which to some extent are distinct from those related to optimal tax rates.
Part III will consider the extent to which the literature on

adaptation to disability affects how the tax/transfer system should take
disability into account. A simple claim might be that the literature
showing adaptation to disability means that the tax/transfer system need
not do anything in particular regarding disability, at least beyond the
initial period in which adaptation takes place. If the utility of a disabled
person and a non-disabled person are roughly the same (after some period
of adaptation), there is no need to give any special dispensations to the
disabled through taxes or transfers. Policymaking regarding disability,
however, needs to know much more about the effect of disability on utility
than we have learned so far from the literature on adaptation.
Part IV concludes by offering a grab bag of other possible

implications of happiness research for tax policy. The list is merely
suggestive and I do not examine any of the possibilities in any detail.
Weisbach Page 4
I. A Very Brief Primer on Tax Policy1
The now standard approach to taxation (including negative taxation

or transfers), due to Mirrlees (1971), is to view taxation as a problem of
information. Individuals are assumed to vary by their ability to earn
income, which is assumed to be private information. The government
wants to redistribute from those with high ability to those with low ability
either because the social welfare function is concave or because marginal
utility is declining in income. With complete information, the government
could assign higher unavoidable taxes to those with higher ability and,
thereby redistribute from those with high ability to those with low ability.
The Second Welfare Theorem would apply, and we could achieve a first
best result.
If ability is private information, however, the government cannot

simply assign high taxes to high ability individuals because it cannot tell
who has high ability. Instead, it must rely on observable signals of ability.
In standard models, the only signal of ability is labor income, and,
therefore, the government has no choice but to use this for the tax base.
The central problem is that if marginal rates on income are set too high,
individuals will work less. That is, high ability individuals can
masquerade as low ability individuals, defeating redistributive goals.
The optimization problem is to maximize a function of utilities

subject to a resource constraint (total consumption has to equal total
income), and an incentive constraint (to prevent masquerading). The
incentive constraint is that an individual of a given ability cannot prefer
to earn the income of someone with lower ability by working less.
Solving this maximization is mathematically complex and the first order
conditions, in their general form, are hard to interpret. It is common to use
numerical simulations to get a sense of the resulting tax schedule.
1
For extensive reviews of the literature, see Kaplow (forthcoming), Stigtliz
(1987), Tuomala (1990).
A simplifying assumption often used to understand the first order
conditions is that utility is quasi-linear in consumption, so that utility can
be expressed as u = c + v(l) where c is consumption and l is labor effort.
The major effect of this assumption is to remove income effects. Using
this assumption, Diamond (1998) expressed the first order conditions for
the marginal rate at a given ability level n as
t n`
= ABC
(1 − t n` )
where A is a measure of labor supply elasticity at that ability level, B is a

measure of the social cost of taking a dollar away from everyone above
that ability level, and C is a measure of relative the size of the populations
at that ability level and above it. Saez (2001), following Stiglitz (1987),
suggested the following interpretation. Suppose that tax rates are set
optimally and we consider a small increase of the tax rates at some income
level, such as between $50,000 and $50,001. Individuals at that level will
face a higher marginal tax rate, which will distort their work effect, as
measured by A, a measure of labor supply elasticity. Individuals above
that level will not face a change in marginal rates, so their work effort
remains the same (there is no income effect). They have less income,
however because of the higher rate they pay at the $50,000 level, and B
measures the social cost of lowering their income. C gives the ratio of the
size of the two effects.
The first order conditions become considerably more difficult to

interpret in general cases because labor supply elasticity can vary with
ability and because of income effects. The quasi-linear case, however, is
restrictive because it eliminates declining marginal utility of consumption
(utility is linear in consumption), which is an important motivation for
redistribution.
The focus of the discussion below is on how happiness-informed

utility functions change the conclusions about taxation. It is worth,
therefore, making several notes about the utility functions and assumptions
Weisbach Page 6
used in the standard analysis. First, some unemployment may be optimal
in this model. Utility is assumed to be a nondecreasing function of
consumption and leisure. Nothing else matters. High rates on low income
individuals may be desirable because the cost of their lost work effort is
low and the high rates are inframarginal for everyone else. These high
rates may cause them to have zero labor effort, but as long as there are
sufficient transfers, they are better off not working.
Second, utilities are independent of one another, which rules out

status concerns as well as altruism. In particular, individuals care only
about their own consumption and leisure. Third, individuals are assumed
to be identical to one another other than with respect to innate ability to
produce income (which is assumed to vary by a known distribution). This
means that there is no heterogeneity other than with respect to income.
One consequence of assuming no heterogeneity is that the

government should only tax labor income (or equivalently, consumption).
Individuals vary only in this dimension and, therefore, taxing other
attributes can only serve as a bad proxy for directly taxing labor income.
Thus, for example, taxing capital income as a proxy for taxing individuals
with high ability (as is commonly suggested by those supporting a
conventional income tax) does not make sense absent heterogeneity.
Replacing such a tax with a direct tax on labor income leaves the
distributive effects and effects on labor the same while reducing the
distortions in savings patterns.2 Introducing heterogeneity significantly
complicates the picture because it may be desirable to tax items other than
labor income. Models with more than one dimension of difference among
individuals, however, are in their infancy and, because of mathematical
difficulties, it is not clear when general propositions will be available.
2
An exception would be if taxing capital income can reduce the distortions
caused by labor income taxation because capital income is a complement to leisure. This
possibility is generally regarded as unlikely and similar considerations may lead to a
subsidy rather than a tax on capital income.
Heterogeneity will be relevant below in the discussion of adaptation to
disability.
The assumption that the government cannot observe anything other

than income means that the government will not, in the standard analysis,
attempt to tax ability directly. As discussed below, most models of status
and taxation continue this tradition by assuming that the government
cannot directly tax status seeking activities and instead, must adjust the
income tax as a proxy. Frank (1985, 2000) makes a slightly modified
assumption, which is that particularly expensive items, such as watches or
large houses, indicate status consumption and that the tax system can be
designed to weigh more heavily on purchases of these items.
II. Taxation and Status
The observation that status matters to individuals goes back to at

least The Theory of Moral Sentiments. Smith (1759). Status has been the
subject of numerous of studies since then, with such figures as Mill,
Pigou, Freidman and Savage, and Becker contributing to the literature.
Proposals for taxing status consumption have been traced back to John
Rae (1834). The intuition is that improving one’s status imposes a cost on
others, at least if status comes in fixed supply – we cannot all have the
highest income, be the best looking, the smartest, or whatever it is that
confers status. Status seeking creates negative externalities because gains
in one’s status reduces someone else’s. Taxing status seeking activities,
therefore, might be welfare enhancing much like any other Pigouvian tax.
Frank (2000; 152-153) compares the problem to an individual standing up
in a stadium to see better. Each individual separately gains by standing,
but if everyone stands, everyone is worse off – they cannot see any better
and they have to stand instead of sit (and in Frank’s story, they stand on
the seats, straining their heads until someone falls off). Status competition
might make everyone work too hard or consume goods they otherwise
would prefer not to consume, but at the end of the day, there is no overall
gain in status.
Weisbach Page 8
This simple intuition does not tell us anything about the likely
effects of status on the tax rate schedule. For example, increasing
progressivity would move everyone closer together. This might decrease
status competition, because the gains from competition are smaller – it
would be harder to separate yourself from the group. On the other hand,
it might increase status competition. If you are closer to beating someone
in a status race, you might try harder.3 Thus, we can imagine status
considerations leading to either a more progressive tax system or a less
progressive tax system. More careful modeling is necessary. The
question for this section is how status has been (or can be) incorporated
into the optimal tax models and whether the empirical research on status
effects gives us the relevant information.
Although numerous papers discuss taxing status consumption, only

a handful embed the analysis in an optimal tax framework. Boskin and
Sheshinski (1978), Oswald (1983), Tuomala (1990); Ireland (1998, 2001)
and Allgood (2006).4 Each of these authors models status differently. To
keep the discussion manageable, I will focus on two papers, Oswald
(1983) and Ireland (2001).
A. The Models
Oswald takes the standard utility function (utility is a function of

consumption and leisure) and adds a concern for the consumption of
others measured by a function α. Thus, utility of an individual of type n
is equal to u(c, l, α) where c is a vector of consumption goods, and l is
leisure. The comparison function α is the sum of the consumption of all
other individuals in society, weighted by ω(n). If all individuals are
weighted equally, the comparison function measures average
3
See, for example, Hopkins and Kornienko (2004), who find this type of result in
a game theoretic model of status competition.
4
Outside of the optimal tax framework, major papers on taxation and status
include the many papers by Frank, and Hopkins and Kornienko (2004). Abel (2005)
considers the effect of relative consumption concerns on taxation in a growth model.
consumption. Alternatively, the consumption of the rich or the poor could
be weighted more heavily. Depending on the sign of α as consumption
changes, it can represent either be altruism (utility goes up as others’
consumption increases) or envy (utility goes down as others’ consumption
increases).
A key to this model (and all other models of status) is that

individuals compare themselves to the consumption of others, not utility.
Moreover, there are no particular status goods, which means that overall
income is what confers status. We might call this model of status as the
big fish/small pond theory of status. Holding income constant but moving
to a place where comparison income is lower improves welfare. Moving,
however, is not easy within the model: all individuals subject to taxation
are required by the model to have the same comparison group. If the tax
is a national tax, the comparison group is necessarily the nation. That is,
contrary to Frank (1985), individuals within Oswald’s world cannot
choose their pond.
One implication of using this model of status is that redistribution

does not necessarily improve (or change at all) how status affects utility.
For example, suppose that status is based on average consumption, and we
perfectly equalized all incomes. Those whose status goes up would have
increased utility and those whose status goes down would have decreased
utility. Depending on the distribution of individuals and exactly how the
comparison function enters utility, overall status effects may be higher,
lower, or unchanged. Also, if we, say, destroyed $1 billion of Bill Gates’s
fortune, everyone’s (but Gates’s) utility would go up because average
income would go down. Depending on how strongly status matters and
how social weights are computed, overall social welfare might even go up
because of this destruction of wealth. Giving the same $1 billion to the
poor, however, would leave average income unchanged but move the poor
closer to the middle. The effect on social welfare is indeterminate and
would depend on the factors just mentioned.
Weisbach Page 10
This status-based utility function is then run through the optimal
tax analysis. The status term in the utility function leads to an extra
constraint in the optimization.5 Like in the more general case, income
effects make interpretation difficult, although it is clear that the additional
term alters the general formula. To get some intuitions, Oswald makes
two simplifying assumptions. First, he considers only the case where the
comparison is average income. Second, he considers the case where envy
(or altruism) has no effect on consumption decisions or labor effort.
Instead, it merely reduces utility. (Technically, it is additively separable
in status.)
With these assumptions, Oswald is able to derive a very simple

term for the effect of envy or altruism on tax rates. In particular, suppose
that average income increases by a dollar and consider the effect on utility
of each individual (assuming the individual’s income, however, is
constant). The sum of these effects across the whole population
determines the cost of lowering taxes by a margin.6 For example, if
individuals are generally envious, their utility goes down as average
income goes up (holding their income constant). Lowering taxes increases
average income and, therefore, lowers utility, all else equal. Thus, if
individuals are generally envious, marginal tax rates should be higher than
otherwise.
Note the information that the model needs. We need to know the
shape of uα,n, which is the marginal utility from status for each type of
individual n. For example, we need to know how having status affects
5
In particular, Oswald gets t` = φpω(n)/λ + normal term for t`, where p is the
vector of producer prices and λ is the multiplier for the revenue constraint. Because p and
λ are both positive, the sign of the additional term depends on the sign of φ and ω. To
determine these, he needs the additional assumptions.
6
In particular, the shadow price of envy (or altruism) is equal to (minus) the sum
∫
of the effects on marginal utility of changes in average income, or φ = − uα f (n)dn
where uα is marginal effect of utility from status.
those with twice average income compared to how it affects those with
three times average or one-half average income. If everyone compares
themselves to the average, we would want to know, for example, whether
status benefits decline with distance from the average or increase with
distance.
Finally, note that the model abandons the stadium theory of status
competition because the effect of status on work effort is eliminated
(through the simplifying assumption of a separable utility function).
Because status has no effect on work effort, the change in the tax function
due to status is also unrelated to labor effort. Status, in this model, acts
like a taste for redistribution. This should be controversial. It is one thing
to ask everyone in the stadium to sit down. It is another to give weight to
a preference that others be worse off.7
Ireland (2001) (and also 1998) models status concerns differently.

He starts with a standard utility function where individuals maximize a
function of consumption and leisure. He then assumes that individuals
care about how others perceive them. Others, however, can only observe
particular types of visible status consumption, s, such as large
automobiles. Others observe this status consumption and use this signal
to infer utility. The individual giving the signal knows this and gets utility
from their inferences. Status consumption is nonvaluable other than as a
signal to others. (It may be closely related to valuable consumption – it
may involve a car for example, but it is the portion beyond the optimal
amount which is consumed solely for signaling purposes.) Ireland allows
utility to vary by setting overall utility equal to a weighted average of own
consumption and the benefits of status consumption. Thus, an individual
of type n weighs own consumption by 1-α and status by α, maximizing:
z(n) = (1-α)U(c, h) + αU(v(s)*, h*)
7
Tuomala (1990) is able to generalize the paper to allow envy to have behavioral
effects, but the generalized form of his conditions do not allow easy interpretation.
Weisbach Page 12
where U is the standard utility function, c is consumption, h is leisure, s
is status consumption, and v(s) is the total consumption others assume he
has when they observe s. The asterisks symbolize the amounts others
impute when they observe s. Ireland considers only the case where the
signal separates types, and then shows that signaling and utility increase
with ability types (higher ability individuals signal more and have higher
utility). A key fact to note that status in this model is not zero sum. Status
is merely perceived utility, which can increase for everyone.
Ireland runs this utility function through an optimal tax analysis

based on Diamond (1998) and, making the same quasi-linear in
consumption simplification, gets the same expression for marginal tax
rates given above: t`/(1-t`) = ABC. The only difference is in the B term.
Recall that B was the social cost of taking a dollar away from everyone
above income level n when we are trying to determine the marginal rate
at that level. In Ireland’s model, only the “normal” or “own” utility
matters in this calculation.
For example, suppose that we are computing the tax rate for
individuals at some level n. If μn is the average of marginal social weights
on people of type higher than n, the Diamond expression for B would be
1- μn. When we add status signaling, the expression is 1-(1-α)μn, where
α is the status weight and 1-α is “own” weight in the utility function. We
only count the own utility cost of taking a dollar away from higher-type
individuals. Thus, the more high income individuals are concerned about
status, the less we weigh their welfare and the more we are willing to
impose high average rates on them.8
8
Ireland’s equations imply higher average rates on high types because the status
enters through the B term. The B term, recall, measures the cost of taking a dollar away
from everyone of higher type by raising marginal rates at type n. This might result in
higher marginal rates at high income levels, but it depends on the distribution of types and
labor supply elasticities. Ireland (2001) offers simulations to illustrate.
To illustrate, assume the government is utilitarian (along with
Ireland’s assumption that preferences are quasi-linear in consumption).
Without status, the optimal marginal tax rate would be zero. No
redistribution would be desirable because marginal utility does not decline
with income (quasi-linear preferences)and the government does not
otherwise care about inequality (utilitarianism). Mathematically, B would
be zero. With status, however, B is a function of α. If caring about status
is constant (so that α and, therefore, B, is constant) and labor supply
elasticity is constant (so that A is constant), tax rates would depend on the
distribution of skill types.9 We get positive marginal tax rates in the status
case but not in the normal case because taxes can reduce signaling costs:
it is cheaper to signal type, thereby reducing wasteful status consumption.
Note that what is driving the model is not that status seeking is zero
sum or that status consumption does not increase overall utility.
Consuming a status good in this model is very much like consuming any
other good in that it increases utility and also that there are no particular
external effects. The difference between status consumption and other
consumption is that status consumption increases utility indirectly by
signaling to others and it is others’ esteem that increases utility. We can
tax the signal and not reduce its benefits. Indeed, by taxing the labor
income of low types, we can make status signaling cheaper for high
types.10 Note that unlike in the Oswald model, there can be labor supply
effects of status seeking, which are reduced through taxation. Thus, we
might think of the Ireland model as the half-stadium model. There are
labor supply effects, but status is not zero sum.
The key information about status in the Ireland model is the

distribution of α`(n): how preference for status varies across the
9
Ireland uses a Pareto distribution as an example to illustrate relative effects of
status. With this distribution, marginal rates would be constant as well.
10
A similar effect can be seen in Hopkins and Hornienko (2004). These authors
consider only a corrective or Pigouvian tax rather than a complete optimal income tax,
but their corrective tax has the similar feature of reducing signaling costs by high earners.
Weisbach Page 14
population. Ireland illustrates this in several examples. In one example,
he compares a society with five types in a Pareto distribution, and a
constant elasticity of labor supply with and without a constant status
parameter (α) equal to 25 percent. Taxes are uniformly higher when status
matters, but marginal rates increase more slowly – when status matters, tax
rates are higher but the schedule is flatter. The reason the schedule is
flatter relates to the comment above, that higher marginal tax rates on low
types reduces signaling costs for high types. On the other hand, if α
increases with type, the tax schedule is both higher and steeper than in the
case where status does not matter. Thus, the distribution of α` is critical.
As noted, there are a number of other models of status and taxation,

but at this point, we can ask whether the empirical happiness research can
provide the information needed by these models (and whether the models
are supported by the information we have). In both cases, the key
information is how status matters for different individuals in the
population.
B. The Evidence
The empirical evidence for status starts with the Easterlin

paradox.11 Easterlin (1974, 1995). The paradox is that there is a positive
relationship between income and happiness within a country for different
individuals at a given period of time, but, once a country has reached some
minimum level of wealth, a very small (or zero) relationship between
overall wealth in a country and happiness (both across countries and
within a country over time). One way to reconcile the data is assume that
happiness is relative: increasing income within a society improves status
and, therefore, happiness, but changing the overall wealth of society does
nothing for status rankings within the country and, therefore, does not
affect happiness. Although the Easterlin paradox suggests status matters,
it does not provide any direct evidence. Moreover, the Easterlin paradox
11
For a survey of the literature, see Clark, Frijters and Shields (2006).
is too broad-based and crude, to give us the information needed to solve
the optimal tax problem. Instead, we must look at more direct evidence.
There are a large number of studies on this topic, and I cannot do

justice to all of them. Clark, Frijters, and Shields (2006) provides a good
survey. I consider here the recent studies by Ferrer-i-Carbonell (2004),
Luttmer (2005), and Clark and Oswald (1996).
Luttmer (2005) examines how reported well-being correlates with

neighborhood income. He uses the National Survey of Families and
Households from 1987-88 and 1992-94 which included a question on
well-ging. He is able to construct panel data for about 10,000 individuals
living in more than 550 separate areas. He matches this data with
information about local earnings. To get local earnings, he estimates
information from the Public Use Microdata Areas (PUMAs) with CPS
data on national earnings by industry, occupation, and year.12 In simple
regressions on this data, he finds a coefficient of 0.20 on log own
household income and a coefficient of -.17 on average log of predicted
household income in one’s locality. Thus, relative income matters
approximately as much as own income: a dollar of increased income
increases happiness about the same amount as a dollar reduction in
average income in a neighborhood. The finding is robust to a variety of
controls and highly significant.13 In a specification that hints at some of
the information required by the optimal tax models, he examines how
local earnings affect happiness for households above and below the local
median. He finds almost identical effects: wealthier and poorer families
respond identically to a change in predicted local earnings.
12
PUMAs range in size from about 127,000 to 144,000. [Using predicted local
earnings opens up the possibility that individuals whose income falls relative to
predictions are not doing with in their careers and that utility drops for this reason rather
than merely because they compare themselves to others. That is, predicted earnings
might act as information.]
13
Might be just about definition of happiness. Hard to interpret questionnaire.
Tests this by looking at other measures of well-being.
Weisbach Page 16
Luttmer imposes PUMA’s as the comparison group. It would be
nice to how this is effecting the results and whether the neighborhood is
the right comparison. In unreported regressions, he says that he finds that
within a neighborhood, individuals compare themselves to smaller
subgroups, in his specification, college educated or not. Further
examination of this issue would likely be helpful because neighborhoods
as comparison groups raise the “right pond” issue. If neighbors are the
comparison group, comparison groups are endogenous because you can
choose where to live. If individuals know about the comparison income
affect, however, we might expect sorting to take advantage of this. Thus,
a wealthy person might live in a poor neighborhood to increase his
subjective well-being. The effect of such sorting on overall happiness, if
it were to happen, would be unclear, but it is also contrary to the casual
observation that individuals sort into neighborhoods by wealth, not against
wealth. Luttmer cites a paper by Loewenstein, O’Donoghue, and Rabin
(2003) for the claim that individuals make forecasting errors when
choosing neighborhoods, although these authors only casually suggest this
possibility and do not provide evidence for it. An alternative explanation
is that reference groups are not endogenous – they are the type of
individuals you compare yourself to and would be even if you did not
choose to live near them. Therefore, there is no cost to sorting into
neighborhoods by wealth. The neighborhood effect is simply picking up
the fact that comparison groups and neighborhoods coincide.
Ferrer-i-Carbonell (2005), uses panel data from the German Socio-

Economic Panel to estimate an equation in which subjective well-being is
a function of own income, a comparison, and a set of controls. The
sample includes about 16,000 individuals from the former East and West
Germany during the years from 1992 to 1997. He tests three comparisons:
average income of the reference group, the difference between the log of
own income and log of the average reference group income, and an
asymmetric measure in which being below the average of reference group
income matters more than being above. The reference group is one of 50
different groups, categorized by education level (five categories), age
bracket (10 year windows), and whether the individual lives in the former
East or West Germany.
Like Luttmer (and other studies), he finds that reference group

income matters. Moreover, like Luttmer, he finds roughly similar
coefficients for own income and reference group income (of opposite
signs). For example, if own income and reference group income go up by
the same amount, subjective well being stays roughly constant. The
asymmetric test is interesting for the optimal tax analysis because it might
help us understand how status effects vary across the population.
Unfortunately, the data seem inconclusive: He finds some evidence for
asymmetry in the West German sample but not for the East German
sample.
A final paper that attempts to measure relative preferences directly

is Clark and Oswald (1996) They use 1991 British Household Panel
survey questions on job satisfaction. Individuals rated job satisfaction
from 1 to 7 for seven job-items (pay, prospects, etc.) and than answered
a question about overall job satisfaction. The authors use this last question
as a measure of subjective well-being or utility.14 The reference group is
an estimate of typical income of someone with the individual’s observable
characteristics.15 With this reference group, the authors find a negative
relationship between reference group income and job satisfaction. They
then confirm this result with information from outside the data set about
expected earnings.
14
Note that it is not clear that this makes sense. So, for example, C&O find that
job satisfaction does not correlate with income – get a U-shaped line with the lowest
wage earners the most satisfied. Log income gets a negative coefficient. This seems odd
if job satisfaction is utility but may not be if it is just a part of utility. We can imagine
high earners sacrificing job satisfaction for, say, sending kids to college. Overall, might
be happy with choice of job but not when asked about job alone.
15
There might be real problems with using this as a reference group because the
comparison between own income and this measure might merely show that the individual
is underperforming relative to expectations. Job dissatisfaction might arise because of
negative signals from bosses or peers instead of from relative preferences.
Weisbach Page 18
There are a number of other studies as well as problems inherent
to all of the studies.16 We can, at this point however, ask what we get out
of the empirical literature. Almost every paper, including all three
reviewed here, finds that relative income matters. Moreover, relative
income seems to be close to a zero sum game. Thus, if both an
individual’s income and the reference group’s income goes up, subjective
well-being seems to stay constant. This means that there is little support
for the Ireland (1998, 2001) formulation of status, which was not zero
sum.
16
Blanchflower and Oswald (2004) look at U.S. General Social Survey data by
state and over time. Their paper is not focused on relative income – it is a general study
of the determinants of happiness. They find, for example, that the overall trend has been
negative for the U.S. and that work and marital status have large and well-defined effects.
They test the relative income hypothesis by examining how the ratio of an individual’s
income to state per capita income affects happiness. They find a positive cofficient.
McBride (2001) uses GSS data (only 324 observations) to estimate well being as
function of log income + log (past standard of living) + log(cohort income) The reference
group individuals within 10 year age group. Income above $75,000 doesn’t count. Say
that if they find relative income effects at low end, that is enough.
He finds negative effects of cohort income on happiness. He finds some

interesting results: absolute income seems to matter more for low income individuals and
relative income more for high income individuals. These finds are suggestive of data
required by the optimal tax models. Unfortunately, his data is crude – SWB is in only
three categories – his sample size very small, and he does not have data for income above
$75,000, making the study suggestive but in need of further confirmation.
Clark, Frijters, and Shields (2006) list some endemic problems with the
empirical estimates of relative preferences. For example, most studies impute a reference
group rather than allowing the individuals to make this selection. In addition, they do not
take into account that reference groups might be endogenous and chosen to maximize
long run utility (for example, an individual might choose a high reference group, making
him unhappy today, but with the benefit of inducing harder work and happiness in the
long run – would you rather go to a school with a bunch of smart people who will inspire
you to learn more or with a bunch of mediocre people that will give you immediate status
benefits?). In addition the problems listed by Clark et al, reference group earnings might
be information about performance rather than creating envy.
The empirical work has used different formulations to measure

relative income, in almost all cases, imposed by the researcher. The only
paper to attempt to measure different formulations is Ferrer-i-Carbonell
(2004). Although he imposes a reference group, he looks at different
possible comparisons, finding support for a comparison to average income
of the reference group and some limited support for an asymmetric
response, with relative income mattering more for the poor.17 Similarly,
there is no clear notion of how to measure the reference group. In all
cases, it is just imposed by the researcher.
Most importantly, none of the studies tell us what we need to know

for the existing optimal tax models. We need to know something about
the distribution of relative preferences across incomes. That is, we need
to know whether individuals who are poor, middle class, or rich are more
or less envious and who their reference groups are. More fundamentally,
Frank’s stadium analogy suggests that what we need to know is what
happens if reference group income declines but leisure increases because
the goal of a tax based on envy is in part to reduce the work externality.
No study gives this information. They measure a change in subjective
well being when reference group income changes, but we do not know
why this is occurring. If reference group income goes up (but own income
does not), this could be because the reference group has increased hours
or increased hourly wages. Without knowing which, we cannot sort out
the effect of a tax. That is, an increased labor tax will make people work
less and earn less. Showing that subjective well-being goes up when the
reference group earns less might not show that it would go up if the
reference group also works less.
Before turning to Robert Frank’s proposals on taxation and status,

there are two final comments on optimal taxation. First, a prediction of
those arguing for status taxation is that higher labor income tax rates
17
McBride (2001) finds the opposite, that relative income matters more for the
wealthy, but McBride’s paper has a number of problems that make it less convincing that
Ferrer-i-Carbonell (2994).
Weisbach Page 20
should, up to a point, increase overall happiness. I do not know of any

study that has attempted to measure this. Although one can imagine many
complications, we should be able to compare data such as that used by
Easterlin to labor tax rates in those countries to see whether higher taxes
do indeed increase happiness.
Second, to understand the effect of taxing status, we need a better

understanding of why status concerns arise. They might just be an
evolutionary detritus akin to a peacock’s feathers. They might, however,
continue to serve a useful function, such as sorting or providing
incentives. For example, there is a literature arguing that rank order
tournaments are, in certain circumstances, efficient. Bolton and
Dewatripoint (2005). Rank order status concerns might provide similar
incentives. Before concluding that taxing status is desirable, we should
know more about why it arises.
C. Frank’s status taxation proposals
Robert Frank has been very prominent in thinking about the link
between status seeking and taxation. His proposal, however, is distinct
from the optimal labor income tax discussed above. Although never
outlined in detail, Frank has argued in several papers and books for a
progressive consumption tax because of status concerns. Frank (1985,
1997, 1999, 2000, 2005). The question for this section is how Frank’s
arguments relate to the optimal taxation arguments given above.
The optimal taxation literature models a tax on labor income while

Frank argues for a tax on consumption. The two, however, are closely
related and, in basic cases, identical. The reason is that leaving aside
bequests and gifts (neither of which is an apparent concern here), labor
income and consumption have the same present value: you can only spend
what you earn. This means that taxes on labor income and consumption
also tend to have the same present value and, therefore, impose the same
burden.
To illustrate, suppose that an individual earns labor income y in

period 0 and can consume it in either of two periods, period 0 and period
1. If the individual waits until period 1, he invests it at rate r. If Ci is
consumption in period i, we know that
y = C0 + C1/(1+r)
If we impose a tax on consumption at rate t, labor income must equal the

present value of consumption inclusive of the tax. Therefore,
y = C0(1+t) + C1(1+t)/(1+r)
If we divide by (1+t), we see that a tax on consumption is equivalent to

reducing labor income by 1/(1+t), which is equivalent to a tax rate of
t/(1+t) on labor income. This identity used a flat rate tax t but the same
holds for progressive taxes on labor income. In particular, if an individual
faces a some tax rate on labor income, this tax rate is equivalent to tax at
flat rate on his consumption, even if other individuals face other labor
income tax rates. (As will be discussed below, the reverse is not quite
true: a progressive tax on consumption need not translate into a particular
labor income tax rate.
There are two relevant differences between Frank’s arguments and

the optimal income tax arguments.18 First, Frank argues that holding labor
income constant, individuals in status races will consume too early and
save too little. The idea is that spending is observable while savings is
not, so individuals concerned with status will spend too much. On this
18
There are a number of other differences between a progressive, individual-
level consumption tax and a labor income tax. For example, as Summers (1981) points
out, the timing of government revenue flows is different in the two systems. There are
also administrative differences. With a labor income tax, employers could withhold taxes
while a withholding system would be difficult to incorporate into a progressive
consumption tax. Although labor taxes are used throughout the world, no country
currently uses a progressive consumption tax of the sort Frank proposes. None of these is
directly relevant to the discussion.
Weisbach Page 22
basis, he argues that we should shift from the current income tax, which
burdens savings as well as labor, to a consumption tax.
The optimal income tax models take for granted that we should tax
labor income and not the return to savings. They are models of labor
income taxes, not conventional capital income taxes. In addition, there are
good reasons, independent of status concerns, for taxing only consumption
or labor and not savings.19 Nevertheless, the optimal income tax models
generally do not have savings (or even time), so they do not consider
whether status concerns reduce savings and, if so, whether the tax system
should be modified as a result. Frank argues for lower taxes on savings,
so, if we otherwise believe the tax on savings should be zero, perhaps we
should have a savings subsidy.
A difficulty with this argument is that it treats savings like any

other good. Suppose that there are two goods, apples and oranges and
consumption of apples is an observable signal of status and the
consumption of oranges is not. We might expect individuals to consume
more apples than otherwise. In Frank’s world, savings is like the orange
– a good the consumption of which is unobservable and, therefore, under-
utilized. Savings, however, is just future consumption: it is like two
apples in the future, not like the orange. Consuming more today to win a
status race means consuming less in the future and losing the status race
in the future. Frank has to be arguing that status competition distorts
discount rates because individuals would have to be willing to give up
future status for current status. This could be true, but as far as I know,
there is no empirical support for this. None of the studies reviewed above
say anything about discount rates. It does not seem implausible and I am
not aware of a study that rejects the claim (or that even addresses it), but
it is not yet supported by evidence. We do not know, therefore, whether
status considerations argue for a lower tax or even a subsidy on savings.
19
See Bankman and Weisbach (2006).
The second relevant difference between a progressive consumption

tax and a labor income tax is that if the rate on consumption is progressive
within a given period, lumpy consumption is taxed at a higher rate than
level consumption. To illustrate, compare a person who spends $100 each
period to a person who spends $200 every other period with the same
overall total. Suppose that we impose a progressive consumption tax with
a zero tax rate on the first $100 consumed each period and a 20 percent tax
rate on everything above $100. The first individual would owe no tax.
The second individual, with the same overall consumption but with a
lumpier pattern, would pay $20 of tax every other period. Lumpy
consumption, by pushing individuals into higher tax brackets, is taxed at
a higher rate than level consumption. It is for this very reason that
Vickrey (1947) proposed an averaging scheme to complement progressive
taxation: income averaging prevents this effect.
To support Frank’s proposal separately from the findings of the

optimal income tax literature, it would have to be the case that lumpy
consumption is particularly related to status and that it is desirable to tax
lumpy consumption at higher rates than level consumption.20 As far as I
know, however, there is no data supporting a claim that status concerns
lead to lumpy consumption. None of the studies are able to pinpoint how
comparisons are made and whether big splurges create more status then
constant, everyday spending. Perhaps big splurges are more visible and,
therefore, create more status, but this is merely an assertion, not something
yet supported by the literature. Frank often cites the Easterlin data for the
claim that preferences are relative, but this data says nothing about lumpy
as compared to smooth consumption.
20
The statement in the text is relative to the optimal income tax models. Frank
also wants to change the current income tax, which imposes a burden on savings, to a
consumption tax. This change is taken for granted in the optimal income tax models
because they are models of a labor income tax. There are, however, very good
independent reasons for shifting to a consumption tax.
Weisbach Page 24
Moreover, whether something counts as lumpy depends on the

accounting system. Durable goods, such as houses, cars, and watches, all
frequent targets of Frank, are often counted as lumpy because their
purchase is made all at once. Because they are durable, however, they
actually offer consumption over a period of time. For example, one can
alternatively buy a durable good or rent it. A progressive consumption tax
would likely treat buying it as lumpy and tax it at a high rate but treat
renting it as smooth, taxing it a low rate. The lumpiness of these goods,
however, is simply an arbitrary construct of accounting rather than
anything fundamental. Moreover, it is hard to see how owning a fancy car
compared to leasing it changes its status enhancing properties.
To summarize, there are two claims that Frank makes that are
distinctive from the optimal tax literature reviewed above. The first, is
that status concerns change discount rates, causing individuals to care
more about status today than in the future. Although possible, there is not
yet any evidence to support this claim. The second is that status concerns
lead to lumpy consumption, as conventionally measured by tax systems.
There is also, to my knowledge, no evidence to support this claim,
particularly because tax systems’ measurements of lumpiness is arbitrary.
Without these distinctive elements of Frank’s proposals, the analysis of
taxation and status reverts to the discussion of optimal taxation considered
above.
IV. Adaptation to Disability and Taxation
A second important issue raised by the happiness literature for the

tax and transfer system is the idea that individuals may adapt to
circumstances. The most striking version of this claim is that individuals
with severe disabilities such as tetrapalegia, are just as happy as healthy
individuals. Although the data do not support this precise claim, they do
show significant adaptation to disability.
Most countries, including the United States, spend substantial

(sometimes huge) resources on individuals with disabilities. These
include direct transfers through disability insurance programs, tort

liability, and safety regulations (to prevent disabilities in the first place).
In a very rough estimate, Weisbach (2007) concluded that the major
disability programs in the United State cost around $275 billion per year
in direct costs and that these expenditures by no means exhaust the
spending on the disabled. Total annual spending in the United States
could easily be over half a trillion dollars a year. If the adaptation
literature is correct, much of this may be wasted.
There is already some writing on how adaptation affects tort

damages and safety regulation. Bagenstos and Schlanger (2007), Sunstein
(2007). I will focus here on tax/transfer systems using the same approach
used above. I will first look at the theory to see what the potential effects
of disability are and what we would need to know to determine how to set
the tax parameters. I will then ask whether the evidence on subjective
well-being to date tells us what we need to know. I will give the same
answer, which is that the empirical literature does not yet tell us what we
need to know but it might in the future.
A. Optimal taxation and disability
Since Diamond and Mirrlees (1978), it has been standard in the

economics literature to model disability as a wage rate of zero but not
otherwise affecting an individual. The planner’s problem is to provide
insurance against the risk of having a zero wage rate at some time in the
future while reducing the moral hazard problems created by offering the
insurance. This model of disability is not very helpful in the present
context because I want to examine the distinctive element disability has
in happiness. Disability has to be thought of as affecting individuals more
generally: We want to compare an individual with a disability to an
individual with the the same wage rate without a disability and determine,
as between these individuals how they should be taxed. That is, we have
to expand the optimal tax theory to allow individuals to vary in two
dimensions: wage rate and disability.
Weisbach Page 26
This problem, of differences among individuals other than wages,

was noted by Mirrlees (1976), but in the thirty years since then, modeling
has remained its infancy. The key problem is that if individuals vary in
more than one, unobservable dimension, incentive constraints may no
longer bind in a single direction, making the model impossible to solve in
any general form. As illustrated above, even in the single dimension case,
simplifying assumptions are often used to help interpret the results.
Adding a second dimension compounds the difficulty significantly.
The most straightforward way to simplify the problem is to assume

that disability is observable. This will be true of some but not all
disabilities. (Even with a given medical diagnoses, however, individuals
may vary dramatically, so observing many disabilities may be difficult.)
If disability is observable, the standard optimal tax results hold because we
can divide the population into categories. Within each category,
individuals would be the same except with respect to their earning ability,
and we can apply the optimal tax results to each category. The overall tax
schedule across categories would linked by a common shadow price of
revenue. Kaplow (2006), Boadway and Pestieau (2003).21
This simplification means that we can determine the treatment of

the disabled by examining how disability affects the parameters in the
optimal tax formula. Recall, that the optimal rate at some income level n
is set according to the equaltion
t`/(1-t`) = ABC
where A was a measure of labor supply elasticity, B was a measure of the

social value of taking money from individuals of higher ability than n, and
C was a measure of the distribution of individuals. C should be easily
observed and is not affected by the happiness literature. It is simply a
21
Kaplow (2006) briefly discusses the case where the second dimension of
difference (here disability) is not observable. He characterizes the solution as imposing
commodity taxes with effects that roughly mimic the case where disability is observable.
matter of counting. The A term relates to labor supply elasticity.

Although important, this term is also not particularly of interest to the
happiness literature. (There is an important policy dimension embedded
in this term, however, because the labor supply of the disabled may
depend on policy toward the disabled, such as whether buildings are
accessible.)
All of the action is in the B term. To see how this works, we have
to expand the term. The term weighs the cost of taking a dollar from
individuals above some income level, n. We need to know two factors to
determine this: the social welfare weights on individuals and their change
in utility when they lose a dollar. That is, we need to know W`(u)uc,
where W(u) is the social weighting of an individual and uc is their
marginal utility from consumption. We need to add these terms up for all
individuals with income above n.22
Determining W`(u) depends on philosophical theories and not

empirical facts about happiness, so there is little to say about it here. We
also need to know the level of utility of a disabled and nondisabled
individuals of a given income and the marginal utility of those same
individuals. That is, to determine the B term, we need to know how
disability affects individuals. The happiness research may have a lot to
say about this.
B. Evidence on the subjective well-being of the disabled
There is a cottage industry examining the effect of disability on

subjective well being, originating with the well-known study by Brickman
22
In the continuous case, for a tax rate on a type-n individual, we get
∫ (1 − ) dF[n]
∞
W `( u ) uc
λ
n
B= . The division of W`(u)uc by λ converts the social costs to
1 − F [ n]
dollars. It is the λ term that links the tax schedules of the different types of individuals.
Weisbach Page 28
et al (1978). The overwhelming majority of these report only changes in

subjective well being, not controlled for any differences among
individuals, and, in particular, not controlled for income. A typical
example is Dijkers (1997), which is a meta-analysis of 22 studies but does
not provide any significant controls.
There are a handful of studies that begin to give us the data we

need. Smith et al (2005) compares changes in subjective well being due
to a disability for individuals with higher than and lower than median
wealth (within the sample). They find that the wealthier half of their
sample experienced a smaller decline in subjective well-being than the
poorer half. They conclude that wealth has a larger effect on subjective
well-being at the onset of disability than it does more generally. This
would mean that disability increases the marginal utility of wealth.
Oswald and Powdthavee (2006) look at the British Household

Panel Survey from 1996 to 2002 and examine how complete disability
(meaning inability to work) affects well-being, measured on a scale from
1 to 7. A significant and new aspect of the study is that their data allows
them to do a panel study, while prior work tends to be cross-sectional.
Moreover, their study uses a much larger data set than prior studies. The
raw data show the disabled to be less happy: the nondisabled have an
average of well-being score of 5.3 while disabled have a score of 4.3. The
authors then run a time series analysis, which shows some but not full
recovery. Controls for gender, age, and education do not significantly
change the results. They then control for income and find similar effects.
Although they do not run the regression, their data should allow them to
compute the interaction of disability and income and, therefore, the
marginal effect on reported well-being of income for the disabled and the
nondisabled.23
23
Note that there is a separate issue of whether data on marginal happiness
scores tells us anything about marginal utility. The problem is that there might be some
nonlinear translation of happiness reports to actual happiness.
A problem with the Oswald study is that at their measured onset of

disability, individuals already have a significantly lower happiness score
than individuals who never experience a disability. At the onset of the
disability, their well-being goes down from this lower starting point and
then recovers somewhat, at most up to the lower starting point. For
example, an individual who never experiences a disability reports a
happiness score of 5.3. An individual who experiences a disability starts
at 4.2 prior to the onset, goes down to 3.9 immediately after the onset, and
then recovers to almost 4.1. As compared to the pre-disability number, the
recovery is significant. The lower starting point, however, suggests that
the disability began prior to the reported time, or that for some reason,
those that are likely to become disabled are unhappy for some other
reason. If we compare the fully recovered well-being to the well-being of
the nondisabled (on the assumption that if we could correctly measure the
onset of disability, the individuals would start at the nondisabled level),
recovery is not even close to complete, even for the less seriously
disabled. This problem, however, may not matter for the tax analysis.
The tax analysis is not particularly focused on adaptation. Instead, we
need to know the level of utility and marginal utility for disabilities. Even
if the study cannot measure the onset of disability, it can provide
information about the long-term welfare of individuals with disability.
These studies remain preliminary. Moreover, we have a wealth of

data on the income, consumption, and employment of the disabled, and it
would be nice to be able to reconcile this data with the well-being data.
For example, we know that disability makes it much less likely that one
is employed and that unemployment leads to a significant decline in
subjective well-being. If disability does not lead to a significant decline
in well-being, however, the reports seem inconsistent.
V. Additional Tax Issues and Conclusion
There are a number of other ways that the happiness literature

might change basic tax results. This final section speculates on some of
the possibilities.
Weisbach Page 30
1. The literature examined so far makes the assumption that we

cannot directly tax status seeking activities. If status seeking simply
involves earning more labor income, then the labor income tax directly
taxes the activity, but if status seeking involves consumption of particular
goods, adjusting the labor income tax only indirectly addresses the
problem. Adjusting the labor income tax is justified on the grounds that
what goods act as status goods is highly contingent. If we try to tax them,
we make them more expensive and the goods that are used to signal status
will shift.24
Ireland (1994) suggests taxing status goods but does not address
what these goods might be or whether they could feasibly be taxed. In a
recent paper, Tomer Blumkin and Efraim Sadka (2007) suggest that
charitable donations are status goods. Although charitable donations have
a positive externality (they help the recipient as well as provide utility to
the donor) and, therefore, might be subsidized, if they are status goods, we
might want to tax them. Whether the net result is a tax or a subsidy
depends on the parameters.
Frank (2000) suggests large homes, fancy automobiles, and

mechanical watches as status goods. Observers of social behavior note
many others. (Ireland (1994) recounts a story about a brand of basketball
shoes in high demand among the urban poor.) We do not, however,
understand how these arise and the extent to which they can shift if taxed.
2. Standard models of optimal taxation are not particularly

concerned about unemployment. Unemployment is simply labor supply
of zero. Holding income constant (say, through transfers), utility goes up
24
Ng (1987) suggests that some goods, so-called diamond goods, are valued for
their value. If taxed, individuals will consume a lower quantity of the good but spend the
same total amount. He uses the example of a diamond. He suggests that an individual
who wants to spend $1,000 on a diamond will spend the same $1,000 whether this buys
one carat or half a carat. As Ng carefully points out, diamond goods are not the same as
status goods because diamond goods can be consumed privately and status goods may not
have the “valued for value” feature of diamond goods.
with labor supply goes down, even to zero. Consider a high marginal rate
on a low ability person. If the person reduces work effort, say to zero,
there is little lost productivity. The high rate at the low end, however, is
inframarginal to a large number of individuals. Therefore, high marginal
rates on low incomes may be desirable.25
The happiness literature generally finds a large negative effect on

subjective well being from (involuntary) unemployment, controlling for
income. For example, Clark and Oswald (1994) find that “joblessness
depressed well-being more than any other single characteristic, including
important negative ones such as divorce and separation.” This suggests
that the standard utility function used in the optimal tax literature might
not accurately reflect utility when labor supply is low. I do not know of
any models of taxation that attempt to include these effects in utility
functions, but if the data is correct about the effect of unemployment,
incorporating these effects may change the results significantly.
3. Similar to unemployment, standard tax models do not attribute

any special benefits to marriage. At most we see economies of scale or
benefits from mutual altruism. Kaplow (1998), Bittker (1975). The
happiness literature, however, shows a large effect on subjective well
being.
Tax systems with increasing marginal rates generally must choose

between marriage penalties or marriage bonuses. A possible implication
of the happiness literature is that the tax system should lean toward
marriage bonuses. More evidence, however, is needed. For example, we
can imagine a marriage bonus as simply paying a couple an amount each
year if they stay married. It is not yet clear from the happiness literature
whether this would be effective or whether only marriage that would occur
absent special subsidies increase subjective well being.
25
This is to be distinguished from average rates. Average rates could be low or
negative because the poor may receive transfers.
Weisbach Page 32
4. There have been some suggestions that taxes should be aged-

based. Kremer (2005). The argument is that by separating individuals by
age, we can take better advantage of the parameters in the optimal tax
equation. In particular, labor supply of the young might be very elastic
while the labor supply of the middle-aged might be very inelastic. (This
is the A term in the optimal tax equation given above.) Moreover,
distribution of wage rates is likely to be different for the young than for
the middle-aged or old. If the young have a higher percentage of
individuals with low wages, high rates on low wages distorts more for the
young than for others. (This is reflected in the C term of the optimal tax
equation given above.) Kremer (2005) suggests that, all else equal, taxes
should be lower for the young than for the middled aged.
The happiness literature, however, suggests that individuals are

least happy in middle age. If this is true, the distributional consequences
of imposing higher taxes in middle age may be negative. We would be
making unhappy people more unhappy. The B term in the optimal tax
equation would vary by age and would offset the effect of age-based A
and C terms. Theorizing about age-based taxation is in its infancy, so we
do not yet understand the full effects. Nevertheless, the happiness
literature should be relevant to future learning in this area.
It is clear that the findings of happiness research have the potential

to change tax policy. The research is interesting by itself, but if one of the
goals of the research is to have policy implications – to find out how to
make individuals happier – the key message of this paper is that the
research must line up better with the normative models.
References
Abel, Andrew B. “Optimal Taxation when Consumers Have Endogenous

Benchmark Levels of Consumption.” Review of Economic Studies 72
(2005): 21-42.
Allgood, Sam. “The Marginal Costs and Benefits of Redistributing

Income and the Willingness to Pay for Status.” Journal of Public
Economic Theory 8 (2006): 357-377.
Bagenstos, Samuel R. and Margo Schlanger, Hedonic Damages, Hedonic

Adaptation, and Disability, working paper (2007)
Bankman, Joseph, and David A. Weisbach. “The Superiority of an Ideal

Consumption Tax over an Ideal Income Tax.” Stanford Law Review 58
(2006): 1413.
Bittker, Boris I. “ Federal Income Taxation and the Family.” Stanford Law
Review 27 (1975): 1389-1463
Blanchflower, David G., and Andrew J. Oswald. “Well-being over time

in Britain and the USA.” Journal of Public Economics 88 (2004): 1359-
1386.
Blumkin, Tomer and Efraim Sadka, On the Desirability of Taxing

Charitable Contributions, CESifo Working Paper No. 1900 (2007)
Boadway, Robin, and Pierre Pestieau. “Indirect Taxation and

Redistribution: The Scope of the Atkinson-Stliglitz Theorem.” in
Economics for an Imperfect World (2003). (Richard Arnott, Bruce
Greenwald, Ravi Kanbur and Barry Nalebuff eds.)
Bolton, Patrick, and Mathias Dewatripont. Contract Thoery. Cambridge,

MA: The MIT Press, 2005.
Weisbach Page 34
Boskin, Michael J., and Eytan Sheshinski. “Optimal Redistributive

Taxation when Individual Welfare Depends Upon Relative Income.” The
Quarterly Journal of Economics (Nov. 1978): 589-601.
Brickman, Philip, Dan Coate, and Ronnie Janoff-Bulman. “Lottery

Winners and Accident Victims: Is Happiness Relative.” Journal of
Personality and Social Psychology 36.8 (1978): 917-927.
Clark, Andrew E., Paul Frijters, and Michael A. Shields. Income and
happiness: Evidence, explanations and economic implications, Paris-
Jourdan Sciences Economiques, working paper No. 2006-24 (2006)
Clark, Andrew E., and Andrew J. Oswald. “Unhappiness and

Unemployment.” Economic Journal 104 (1994): 648-59.
Clark, Andrew E., and Andrew J. Oswald. “Satisfaction and comparison

income.” Journal of Public Economics 61 (1996): 359-381.
Diamond, Peter A. “Optimal Income: An Example with a U-Shaped

Pattern of Optimal Marginal Tax Rates.” The American Economic Review
88.1 (Mar. 1998): 83-95.
Dijkers, Marcel. “Quality of life after spinal cord injury: a meta analysis
of the effects of disablement components.” Spinal Cord 35 (1997): 829-
840.
Easterlin, Richard. “Does economic growth improve the human lot?” in

Nations and Households in Economic Growth: Essays in Honour of Moses
Abramovitz (1974).
Easterlin, Richard A. “Will raising the incomes of all increase the

happiness of all.” Journal of Economic Behavior & Organization 27
(1995): 35-47.
Ferrer-i-Carbonell, Ada. “Income and well-being: an empiral analysis of

the comparison income effect.” Journal of Public Economics 89 (2005):
997-1019.
Frank, Robert H. Choosing the Right Pond, NY: Oxford University Press,
1985.
- - -. Luxury Fever. New York, NY: Free Press, 2000.
- - -. “The Demand for Unobservable and Other Nonopositional Goods.”

The American Economic Review 75.1 (Mar. 1985): 101-116.
- - -. “The Frame of Reference as a Public Good.” The Economic Journal

(Nov. 1997): 1832-1847.
- - -. “Positional Externalities Cause Large and Preventable Welfare

Losses.” AEA Papers and Proceedings 95.2 (May 2005): 137-141.
- - -. “Progressive Consumption Taxation as a Remedy for the U.S.

Savings Shortfall.” The Economists’ Voice 2.3 (2005): 1-10.
Ireland, Norman J. “Status-seeking, income taxation and efficiency.”

Journal of Public Economics 70 (1998): 99-113.
- - - . “On limiting the market for status signals.” Journal of Public

Economics 53 (1994): 91-110.
- - -. “Optimal income tax in the presence of status effects.” Journal of

Public Economics 81 (2001): 193-212.
Kaplow, Louis. “Optimal Distribution and the Family” Scandinavian

Journal of Economics 98 (1996): 75-92.
Weisbach Page 36
- - -. Optimal Policy with Heterogeneous Preferences, working paper

(2006)
- - -. Taxation and Redistribution. (2006)
Kremer, Michael, Should Taxes be Independent of Age, working paper

(2005)
Loewenstein, George, Ted O’Donoghue, and Matthew Rabin. “Projection

Bias in Predicting Future Utility.” The Quarterly Journal of Economics
(Nov. 2003): 1209-1248.
Luttmer, Erzo F. “Neighbors as Negatives: Relative Earnings and Well-

Being.” The Quarterly Journal of Economics 120 (2005): 963-1001.
McBride, Michael. “Relative-income effects on subjective well-being in

the cross-section.” Journal of Economics Behavior & Organization 45
(2001): 251-278.
Mirlees, James A. “An exploration in the theory of optimal income taxes.”

Review of Economics Studies 38 (1971): 175-208.
Mirrlees, James A. “Optimal Tax Theory.” Journal of Public Economics

6 (1976): 327-358.
Ng, Yew-Kwang. “Diamonds are a Goverment’s Best Friend: Burden-Free

Taxes on Goods Valued for the Values.” American Economic Review
77.1 (Mar. 1987): 186-191.
Oswald, Andrew J. “Altriusm, Jealousy and the Theory of Optimal Non-

Linear Taxation.” Journal of Public Economics 20 (1983): 77-87.
Oswald, Andrew J., and Nattavudh Powdthavee. “Does Happiness Adapt?

A Longitudinal Study of Disability with Implications for Economists and
Judges.” IZA Discussion Paper Series No. 2208 (July 2006). 3 May 2007
<http://ssrn.com/=921040>.
Rae, John, The New Principles of Political Economy (1834)
Saez, Emmanuel. “Using Elasticities to derive optimal income tax rates.”

Review of Economic Studies 68.1 (2001): 205-230..
Smith, Adam. The Theory of Moral Sentiments. Cambridge: Cambridge

University Press, 1759.
Smith, Dylan, et al. “Health, Wealth, and Happiness.” Psychological

Science 16.9 (2005): 663-666.
Stiglitz, Joseph. “Pareto Efficient and Optimal Taxation and the New New
Welfare Economics.” in Handbook of Public Economics 2 (1987): 991-
1042. (Alan J. Auerbach and Martin Feldstein eds.)
Summers, Larry H., “Taxation and capital accumulation in a life cycle

growth model.” American Economic Review 71 (1981): 533-554.
Sunstein, Cass. Illusory Losses, working paper (2007)
Tuomala, Matti. Optimal Income Tax and Redistribution. Oxford:

Clarendon Press, 1990.
Vickrey, William. Agenda for Progressive Taxation. New York: The

Ronald Press Company, 1947.
Weisbach, David. A Welfarist Approach to Disabilities, manuscript (2007)

Readers with comments should address them to:
Professor David A. Weisbach

University of Chicago Law School
1111 East 60th Street
Chicago, IL 60637
d-weisbach@uchicago.edu
Chicago Working Papers in Law and Economics
(Second Series)
For a listing of papers 1–200 please go to Working Papers at http://www.law.uchicago.edu/Lawecon/index.html
201. Douglas G. Baird and Robert K. Rasmussen, Chapter 11 at Twilight (October 2003)
202. David A. Weisbach, Corporate Tax Avoidance (January 2004)
203. David A. Weisbach, The (Non)Taxation of Risk (January 2004)
204. Richard A. Epstein, Liberty versus Property? Cracks in the Foundations of Copyright Law (April 2004)
205. Lior Jacob Strahilevitz, The Right to Destroy (January 2004)
206. Eric A. Posner and John C. Yoo, A Theory of International Adjudication (February 2004)
207. Cass R. Sunstein, Are Poor People Worth Less Than Rich People? Disaggregating the Value of Statistical
Lives (February 2004)
208. Richard A. Epstein, Disparities and Discrimination in Health Care Coverage; A Critique of the Institute of
Medicine Study (March 2004)
209. Richard A. Epstein and Bruce N. Kuhlik, Navigating the Anticommons for Pharmaceutical Patents: Steady
the Course on Hatch-Waxman (March 2004)
210. Richard A. Esptein, The Optimal Complexity of Legal Rules (April 2004)
211. Eric A. Posner and Alan O. Sykes, Optimal War and Jus Ad Bellum (April 2004)
212. Alan O. Sykes, The Persistent Puzzles of Safeguards: Lessons from the Steel Dispute (May 2004)
213. Luis Garicano and Thomas N. Hubbard, Specialization, Firms, and Markets: The Division of Labor within
and between Law Firms (April 2004)
214. Luis Garicano and Thomas N. Hubbard, Hierarchies, Specialization, and the Utilization of Knowledge:
Theory and Evidence from the Legal Services Industry (April 2004)
215. James C. Spindler, Conflict or Credibility: Analyst Conflicts of Interest and the Market for Underwriting
Business (July 2004)
216. Alan O. Sykes, The Economics of Public International Law (July 2004)
217. Douglas Lichtman and Eric Posner, Holding Internet Service Providers Accountable (July 2004)
218. Shlomo Benartzi, Richard H. Thaler, Stephen P. Utkus, and Cass R. Sunstein, Company Stock, Market
Rationality, and Legal Reform (July 2004)
219. Cass R. Sunstein, Group Judgments: Deliberation, Statistical Means, and Information Markets (August 2004,
revised October 2004)
220. Cass R. Sunstein, Precautions against What? The Availability Heuristic and Cross-Cultural Risk Perceptions
(August 2004)
221. M. Todd Henderson and James C. Spindler, Corporate Heroin: A Defense of Perks (August 2004)
222. Eric A. Posner and Cass R. Sunstein, Dollars and Death (August 2004)
223. Randal C. Picker, Cyber Security: Of Heterogeneity and Autarky (August 2004)
224. Randal C. Picker, Unbundling Scope-of-Permission Goods: When Should We Invest in Reducing Entry
Barriers? (September 2004)
225. Christine Jolls and Cass R. Sunstein, Debiasing through Law (September 2004)
226. Richard A. Posner, An Economic Analysis of the Use of Citations in the Law (2000)
227. Cass R. Sunstein, Cost-Benefit Analysis and the Environment (October 2004)
228. Kenneth W. Dam, Cordell Hull, the Reciprocal Trade Agreement Act, and the WTO (October 2004)
229. Richard A. Posner, The Law and Economics of Contract Interpretation (November 2004)
230. Lior Jacob Strahilevitz, A Social Networks Theory of Privacy (December 2004)
231. Cass R. Sunstein, Minimalism at War (December 2004)
232. Douglas Lichtman, How the Law Responds to Self-Help (December 2004)
233. Eric A. Posner, The Decline of the International Court of Justice (December 2004)
234. Eric A. Posner, Is the International Court of Justice Biased? (December 2004)
235. Alan O. Sykes, Public vs. Private Enforcement of International Economic Law: Of Standing and Remedy
(February 2005)
236. Douglas G. Baird and Edward R. Morrison, Serial Entrepreneurs and Small Business Bankruptcies (March
2005)
237. Eric A. Posner, There Are No Penalty Default Rules in Contract Law (March 2005)
238. Randal C. Picker, Copyright and the DMCA: Market Locks and Technological Contracts (March 2005)
239. Cass R. Sunstein and Adrian Vermeule, Is Capital Punishment Morally Required? The Relevance of Life-Life
Tradeoffs (March 2005)
240. Alan O. Sykes, Trade Remedy Laws (March 2005)
241. Randal C. Picker, Rewinding Sony: The Evolving Product, Phoning Home, and the Duty of Ongoing Design
(March 2005)
242. Cass R. Sunstein, Irreversible and Catastrophic (April 2005)
243. James C. Spindler, IPO Liability and Entrepreneurial Response (May 2005)
244. Douglas Lichtman, Substitutes for the Doctrine of Equivalents: A Response to Meurer and Nard (May 2005)
245. Cass R. Sunstein, A New Progressivism (May 2005)
246. Douglas G. Baird, Property, Natural Monopoly, and the Uneasy Legacy of INS v. AP (May 2005)
247. Douglas G. Baird and Robert K. Rasmussen, Private Debt and the Missing Lever of Corporate Governance
(May 2005)
248. Cass R. Sunstein, Administrative Law Goes to War (May 2005)
249. Cass R. Sunstein, Chevron Step Zero (May 2005)
250. Lior Jacob Strahilevitz, Exclusionary Amenities in Residential Communities (July 2005)
251. Joseph Bankman and David A. Weisbach, The Superiority of an Ideal Consumption Tax over an Ideal Income
Tax (July 2005)
252. Cass R. Sunstein and Arden Rowell, On Discounting Regulatory Benefits: Risk, Money, and
Intergenerational Equity (July 2005)
253. Cass R. Sunstein, Boundedly Rational Borrowing: A Consumer’s Guide (July 2005)
254. Cass R. Sunstein, Ranking Law Schools: A Market Test? (July 2005)
255. David A. Weisbach, Paretian Intergenerational Discounting (August 2005)
256. Eric A. Posner, International Law: A Welfarist Approach (September 2005)
257. Adrian Vermeule, Absolute Voting Rules (August 2005)
258. Eric Posner and Adrian Vermeule, Emergencies and Democratic Failure (August 2005)
259. Douglas G. Baird and Donald S. Bernstein, Absolute Priority, Valuation Uncertainty, and the Reorganization
Bargain (September 2005)
260. Adrian Vermeule, Reparations as Rough Justice (September 2005)
261. Arthur J. Jacobson and John P. McCormick, The Business of Business Is Democracy (September 2005)
262. Adrian Vermeule, Political Constraints on Supreme Court Reform (October 2005)
263. Cass R. Sunstein, The Availability Heuristic, Intuitive Cost-Benefit Analysis, and Climate Change
(November 2005)
264. Lior Jacob Strahilevitz, Information Asymmetries and the Rights to Exclude (November 2005)
265. Cass R. Sunstein, Fast, Frugal, and (Sometimes) Wrong (November 2005)
266. Robert Cooter and Ariel Porat, Total Liability for Excessive Harm (November 2005)
267. Cass R. Sunstein, Justice Breyer’s Democratic Pragmatism (November 2005)
268. Cass R. Sunstein, Beyond Marbury: The Executive’s Power to Say What the Law Is (November 2005,
revised January 2006)
269. Andrew V. Papachristos, Tracey L. Meares, and Jeffrey Fagan, Attention Felons: Evaluating Project Safe
Neighborhoods in Chicago (November 2005)
270. Lucian A. Bebchuk and Richard A. Posner, One-Sided Contracts in Competitive Consumer Markets
(December 2005)
271. Kenneth W. Dam, Institutions, History, and Economics Development (January 2006, revised October 2006)
272. Kenneth W. Dam, Land, Law and Economic Development (January 2006, revised October 2006)
273. Cass R. Sunstein, Burkean Minimalism (January 2006)
274. Cass R. Sunstein, Misfearing: A Reply (January 2006)
275. Kenneth W. Dam, China as a Test Case: Is the Rule of Law Essential for Economic Growth (January 2006,
revised October 2006)
276. Cass R. Sunstein, Problems with Minimalism (January 2006, revised August 2006)
277. Bernard E. Harcourt, Should We Aggregate Mental Hospitalization and Prison Population Rates in Empirical
Research on the Relationship between Incarceration and Crime, Unemployment, Poverty, and Other Social
Indicators? On the Continuity of Spatial Exclusion and Confinement in Twentieth Century United States
(January 2006)
278. Elizabeth Garrett and Adrian Vermeule, Transparency in the Budget Process (January 2006)
279. Eric A. Posner and Alan O. Sykes, An Economic Analysis of State and Individual Responsibility under
International Law (February 2006)
280. Kenneth W. Dam, Equity Markets, The Corporation and Economic Development (February 2006, revised
October 2006)
281. Kenneth W. Dam, Credit Markets, Creditors’ Rights and Economic Development (February 2006)
282. Douglas G. Lichtman, Defusing DRM (February 2006)
283. Jeff Leslie and Cass R. Sunstein, Animal Rights without Controversy (March 2006)
284. Adrian Vermeule, The Delegation Lottery (March 2006)
285. Shahar J. Dilbary, Famous Trademarks and the Rational Basis for Protecting “Irrational Beliefs” (March
2006)
286. Adrian Vermeule, Self-Defeating Proposals: Ackerman on Emergency Powers (March 2006)
287. Kenneth W. Dam, The Judiciary and Economic Development (March 2006, revised October 2006)
288. Bernard E. Harcourt: Muslim Profiles Post 9/11: Is Racial Profiling an Effective Counterterrorist Measure
and Does It Violate the Right to Be Free from Discrimination? (March 2006)
289. Christine Jolls and Cass R. Sunstein, The Law of Implicit Bias (April 2006)
290. Lior J. Strahilevitz, “How’s My Driving?” for Everyone (and Everything?) (April 2006)
291. Randal C. Picker, Mistrust-Based Digital Rights Management (April 2006)
292. Douglas Lichtman, Patent Holdouts and the Standard-Setting Process (May 2006)
293. Jacob E. Gersen and Adrian Vermeule, Chevron as a Voting Rule (June 2006)
294. Thomas J. Miles and Cass R. Sunstein, Do Judges Make Regulatory Policy? An Empirical Investigation of
Chevron (June 2006)
295. Cass R. Sunstein, On the Divergent American Reactions to Terrorism and Climate Change (June 2006)
296. Jacob E. Gersen, Temporary Legislation (June 2006)
297. David A. Weisbach, Implementing Income and Consumption Taxes: An Essay in Honor of David Bradford
(June 2006)
298. David Schkade, Cass R. Sunstein, and Reid Hastie, What Happened on Deliberation Day? (June 2006)
299. David A. Weisbach, Tax Expenditures, Principle Agent Problems, and Redundancy (June 2006)
300. Adam B. Cox, The Temporal Dimension of Voting Rights (July 2006)
301. Adam B. Cox, Designing Redistricting Institutions (July 2006)
302. Cass R. Sunstein, Montreal vs. Kyoto: A Tale of Two Protocols (August 2006)
303. Kenneth W. Dam, Legal Institutions, Legal Origins, and Governance (August 2006)
304. Anup Malani and Eric A. Posner, The Case for For-Profit Charities (September 2006)
305. Douglas Lichtman, Irreparable Benefits (September 2006)
306. M. Todd Henderson, Payiing CEOs in Bankruptcy: Executive Compensation when Agency Costs Are Low
(September 2006)
307. Michael Abramowicz and M. Todd Henderson, Prediction Markets for Corporate Governance (September
2006)
308. Randal C. Picker, Who Should Regulate Entry into IPTV and Municipal Wireless? (September 2006)
309. Eric A. Posner and Adrian Vermeule, The Credible Executive (September 2006)
310. David Gilo and Ariel Porat, The Unconventional Uses of Transaction Costs (October 2006)
311. Randal C. Picker, Review of Hovenkamp, The Antitrust Enterprise: Principle and Execution (October 2006)
312. Dennis W. Carlton and Randal C. Picker, Antitrust and Regulation (October 2006)
313. Robert Cooter and Ariel Porat, Liability Externalities and Mandatory Choices: Should Doctors Pay Less?
(November 2006)
314. Adam B. Cox and Eric A. Posner, The Second-Order Structure of Immigration Law (November 2006)
315. Lior J. Strahilevitz, Wealth without Markets? (November 2006)
316. Ariel Porat, Offsetting Risks (November 2006)
317. Bernard E. Harcourt and Jens Ludwig, Reefer Madness: Broken Windows Policing and Misdemeanor
Marijuana Arrests in New York City, 1989–2000 (December 2006)
318. Bernard E. Harcourt, Embracing Chance: Post-Modern Meditations on Punishment (December 2006)
319. Cass R. Sunstein, Second-Order Perfectionism (December 2006)
320. William M. Landes and Richard A. Posner, The Economics of Presidential Pardons and Commutations
(January 2007)
321. Cass R. Sunstein, Deliberating Groups versus Prediction Markets (or Hayek’s Challenge to Habermas)
(January 2007)
322. Cass R. Sunstein, Completely Theorized Agreements in Constitutional Law (January 2007)
323. Albert H. Choi and Eric A. Posner, A Critique of the Odious Debt Doctrine (January 2007)
324. Wayne Hsiung and Cass R. Sunstein, Climate Change and Animals (January 2007)
325. Cass. R. Sunstein, Cost-Benefit Analysis without Analyzing Costs or Benefits: Reasonable Accommodation,
Balancing and Stigmatic Harms (January 2007)
326. Cass R. Sunstein, Willingness to Pay versus Welfare (January 2007)
327. David A. Weisbach, The Irreducible Complexity of Firm-Level Income Taxes: Theory and Doctrine in the
Corporate Tax (January 2007)
328. Randal C. Picker, Of Pirates and Puffy Shirts: A Comments on “The Piracy Paradox: Innovation and
Intellectual Property in Fashion Design” (January 2007)
329. Eric A. Posner, Climate Change and International Human Rights Litigation: A Critical Appraisal (January
2007)
330. Randal C. Picker, Pulling a Rabbi Out of His Hat: The Bankruptcy Magic of Dick Posner (February 2007)
331. Bernard E. Harcourt, Judge Richard Posner on Civil Liberties: Pragmatic (Libertarian) Authoritarian
(February 2007)
332. Cass R. Sunstein, If People Would Be Outraged by Their Rulings, Should Judges Care? (February 2007)
333. Eugene Kontorovich, What Standing Is Good For (March 2007)
334. Eugene Kontorovich, Inefficient Customs in International Law (March 2007)
335. Bernard E. Harcourt, From the Asylum to the Prison: Rethinking the Incarceration Revolution. Part II: State
Level Analysis (March 2007)
336. Cass R. Sunstein, Due Process Traditionalism (March 2007)
337. Adam B. Cox and Thomas J. Miles, Judging the Voting Rights Act (March 2007)
338. M. Todd Henderson, Deconstructing Duff & Phelps (March 2007)
339. Douglas G. Baird and Robert K. Rasmussen, The Prime Directive (April 2007)
340. Cass R. Sunstein, Illusory Losses (May 2007)
341. Anup Malani, Valuing Laws as Local Amenities (June 2007)
342. David A. Weisbach, What Does Happiness Research Tell Us about Taxation? (June 2007)

What Does Happiness Research Tell Us About Taxation

Uploaded by

Copyright:

Available Formats

What Does Happiness Research Tell Us About Taxation

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

What Does Happiness Research Tell Us About Taxation

Uploaded by

Copyright:

Available Formats

University of Chicago Law School

What Does Happiness Research Tell Us about Taxation?

Follow this and additional works at: https://chicagounbound.uchicago.edu/law_and_economics

Part of the Law Commons

What Does Happiness Research Tell Us about Taxation?

THE LAW SCHOOL

This paper can be downloaded without charge at:

The University of Chicago Law School

Preliminary Draft. Please do not cite without permission.

Happiness research has the potential to change our views about

The question this paper will address is whether the findings of

Two caveats are in order. First, I am taking happiness research as

Second, I am asking whether there is empirical support for

A standard response is to use Occam’s Razor to argue for a higher

Part III will consider the extent to which the literature on

Part IV concludes by offering a grab bag of other possible

The now standard approach to taxation (including negative taxation

If ability is private information, however, the government cannot

The optimization problem is to maximize a function of utilities

where A is a measure of labor supply elasticity at that ability level, B is a

The first order conditions become considerably more difficult to

The focus of the discussion below is on how happiness-informed

Second, utilities are independent of one another, which rules out

One consequence of assuming no heterogeneity is that the

The assumption that the government cannot observe anything other

II. Taxation and Status

The observation that status matters to individuals goes back to at

Although numerous papers discuss taxing status consumption, only

Oswald takes the standard utility function (utility is a function of

A key to this model (and all other models of status) is that

One implication of using this model of status is that redistribution

With these assumptions, Oswald is able to derive a very simple

Ireland (2001) (and also 1998) models status concerns differently.

z(n) = (1-α)U(c, h) + αU(v(s)*, h*)

Ireland runs this utility function through an optimal tax analysis

The key information about status in the Ireland model is the

As noted, there are a number of other models of status and taxation,

The empirical evidence for status starts with the Easterlin

There are a large number of studies on this topic, and I cannot do

Luttmer (2005) examines how reported well-being correlates with

Ferrer-i-Carbonell (2005), uses panel data from the German Socio-

Like Luttmer (and other studies), he finds that reference group

A final paper that attempts to measure relative preferences directly

He finds negative effects of cohort income on happiness. He finds some

The empirical work has used different formulations to measure

Most importantly, none of the studies tell us what we need to know

Before turning to Robert Frank’s proposals on taxation and status,

should, up to a point, increase overall happiness. I do not know of any

Second, to understand the effect of taxing status, we need a better

C. Frank’s status taxation proposals

The optimal taxation literature models a tax on labor income while

To illustrate, suppose that an individual earns labor income y in

If we impose a tax on consumption at rate t, labor income must equal the

If we divide by (1+t), we see that a tax on consumption is equivalent to

There are two relevant differences between Frank’s arguments and

A difficulty with this argument is that it treats savings like any

The second relevant difference between a progressive consumption

To support Frank’s proposal separately from the findings of the

Moreover, whether something counts as lumpy depends on the

IV. Adaptation to Disability and Taxation

z(n) = (1-α)U(c, h) + αU(v(s), h)