FLORIDA TAX REVIEW
VOLUME 1 OCTOBER 1992 NUIBER I

VALUING PERSONAL CONSUMPTION:
Cost Versus Value and the Impact of Insurance

Daniel Halperin"

"An ideal income tax would perhaps differentiate
among individuals according to their talents for using funds
in consumption; but ... income taxes must [be based upon]
measurable quantities! ... [clonsumption, as an element of
income, must be measured ... by outlays for consumption
purposes."'

I. INTRODUCTION

Andy is excited. He has finally been able to get the money together
for an expensive new car. After years of making do with a used car or
cheaper models he is about to pick up a new $50,000 Mercedes sedan. But
alas! Andy's dream of years of driving comfort are soon shattered. The car
is a lemon. It is constantly in the shop for repairs wasting Andy's time and
eventually costing him several thousand dollars. In the end, it never drives as
well as expected.

Would an ideal tax system, as Simons suggests, take some account
of Andy's misfortune? Should we, if we could, somehow determine what the
car is really worth and allow Andy a deduction for the difference between
that amount and his outlay of $50,000? Alternatively, should we ideally
allow a deduction for the extraordinary cost of repairs, the amount spent

* Professor of Law, Georgetown University Law Center. J.D. Harvard. 1961. Like
many law professors, I have been thinking about the issues raised in this article for a long
time. I am grateful to Boston University Law School for inviting me to a workshop which led
to the initial draft of this article. I have benefitted from comments at that workshop, a
workshop at Georgetown, and from comments received at a 1989 Seminar on Current Research
in Taxation sponsored by the Harvard Law School Fund for Tax Research. I also want to
thank my colleague Stephen Cohen as well as Professor Lawrence Lokken with whom I
exchanged early drafts of our respective musings on this subject. Professor Lokken's draft
stimulated my thinking on a number of points. I am particularly indebted to my colleague
Richard Diamond for his suggestion which led to one of the central ideas in this article. My
research assistants, particularly Daniel Luchsinger and Ajay Mehrotra. contributed enormously
to this effort.

1. Henry C. Simons, Personal Income Taxation: The Definition of Income as a
Problem of Fiscal Policy, 119-20 (1938).



Florida Tax Review

beyond what would normally be expected? Is our failure to take account of
Andy's misfortune a concession to administerability rather than a matter of
principle? This article explores the answer to that question as well as whether
recoveries, from insurance or otherwise, should be taxable, if such losses are
disallowed.

There are a number of reasons why there might be a great deal of
difference between the amount of enjoyment derived from a consumer
purchase and the cost to the individual on the market. Consumer surplus,
circumstance of use, variation in quality, events subsequent to purchase, and
changes in market price may explain the discrepancy between actual
enjoyment and cost.

Consumer surplus is the difference between the amount the buyer is
willing to pay and the market price.' The difference occurs because sellers
cannot efficiently determine each buyer's willingness to pay, and hence
cannot extract that amount in price. Thus, it is not possible for a seller, say
of wine, to charge each consumer the amount she is willing to pay. For some,
an excellent bottle of wine might be worth $50, but if there are not enough
such individuals to buy the entire supply, the price will be lower. To
simplify, the market price will be the highest price that will move the
merchandise on hand. If a person would be willing to pay $50, but she can
buy the bottle of wine for its market price of $30, she will be said to have
obtained $20 of consumer surplus.3

Individuals will also differ in the circumstances under which they will
use the product. A bottle of wine enjoyed with a loved one, perhaps on an
anniversary or birthday or in the company of superior food, may take on
additional value. Some purchases will be rarely used, such as an exercise bike
purchased at a time when dreams of self-improvement momentarily took
control.

Items which are supposedly alike may differ in actual quality. Some
automobiles are lemons. There can be differences in taste among wines of the
same vintage. If information is available as to the range of quality but not as
to the quality of any particular item, one should expect the price to reflect
average quality.4 In other circumstances, perhaps due to fraud, all items of

2. Paul A. Samuelson & William D. Nordhaus, Economics 733 (14th ed. 1992).
3. Inclusion of consumer surplus or other non-monetary satisfactions (watching

sunsets) in the tax base, even if feasible, could lead to difficulty in payment of the tax for
those for whom such satisfactions represent a large proportion of their income. Somehow the
impact of the potential tax would have to be taken into account in determining the true amount
of the benefit. See Daniel I. Halperin, Business Deduction for Personal Living Expenses: A
Uniform Approach to an Unsolved Problem, 122 U. Pa. L. Rev. 859, 882-85 (1974). This
matter is not pursued in this article.

4. While the price will reflect average quality in the case of new products, that may
not be the case for used goods. The literature suggests that because everyone believes used

[Vol 1:1



Valuing Personal Consunption

a particular product will turn out to be worth far less than the price charged.
This variation in quality may be reflected by differences in the cost

of upkeep or repairs. If the value of consumption is based simply on
expenditures, an automobile with a terrible repair record would be found,
contrary to common understanding, to produce more enjoyment than one
which is trouble free.5

A difference in value may also arise because of events, possibly
fortuitous, which take place after the initial purchase. An automobile may be
involved in an accident. A bottle of wine can break. A vacation may be
ruined by bad weather, an elegant dinner because of the chef's disposition.
Many arbitrary events can alter the identity between cost and amount of
enjoyment.

Finally, there may be changes in relative market prices. A home or
a work of art may increase or decline in value.6 Arguably these gains and
losses, whether from market fluctuation or otherwise, should be taken into
account in an ideal tax system whether or not the asset is actually sold.

Currently, however, such gains and losses are generally ignored.
Gains on consumption expenditures are recognized only upon sale or other
disposition of the items. While certain medical expenses may be deducted,7

other losses are taken into account only in the event of a "casualty"-a
sudden unexpected and unusual event, such as a collision or fire, as opposed

cars are "lemons" their price reflects the bottom of expected quality. See George A. Akerlof.
The Market for "Lemons": Quality Uncertainty and the Market Mechanism. 84 Q. J. Econ.
488, 489-90 (1970).

5. See Zarin v. Commissioner, 92 T.C. 1084, 1101 (1989) (Tannenwald. J.,
dissenting), rev'd, 916 F.2d 110 (3d Cir. 1990). In an analogous situation. Judge Tannenwald
argued against the proposition "that the more a gambler loses, the greater his pleasure and the
larger the increase in his wealth." Id.

6. Sometimes a product may be sold for more than the purchase price even if the
underlying market does not change. Thus. if a wine improves with age, its price can be
expected to rise as it approaches the ideal time for consumption. For example. a bottle of wine
purchased for $30 may be expected to sell 10 years later for S60 even if there is no change
in the underlying market value for the wine. This increase in price reflects the fact that an
individual who purchases the wine when it is first bottled gives up the opportunity to invest
the sum devoted to the wine.

If he did not expect a return on this investment, he would wait and purchase the
wine when it was ready for drinking. Although the difference between S30 and S60 does not
reflect current use value, it seems related to imputed income on a consumer durable which is
ordinarily not subject to tax.

Suppose, however, the price of the wine actually increases to S100. A sale of the
wine for $100 would produce taxable income of S70, consisting of the imputed return of $30
and an increase in market price of $40.

7. IRC § 213(a).

19921



Florida Tax Review

to an unusually high cost of repair.8 Further, while at one time casualty
losses were deductible in full, such losses now can be deducted only to the
extent that they are truly extraordinary in that losses for the year exceed 10%
of adjusted gross income.9 This article analyzes whether these results are
correct as a matter of principle or merely-as Simons suggests-a concession
to reality?'0

A related question that this article also considers is whether insurance
and tort recoveries that compensate for a loss related to a consumption item
should be taxable. To illustrate, Andy may not be totally out-of-luck. The
repairs are likely to be covered by a dealer warranty. In fact, under recently
enacted legislation, in some states, he may be entitled to a new car to replace
a "lemon."" If Andy's car were destroyed by collision or fire, he would
probably recover the amount of his loss from either his insurance carrier or
the tortfeasor who caused the damage.

Intuitively, I believe most of us would conclude that such recoveries
or benefits from the dealer, a tortfeasor or an insurance carrier should not be
taxable to Andy. After all he is no better off than those who did not suffer

8. Personal losses are deductible only if they "arise from fire, storm, shipwreck, or
other casualty, or from theft." IRC § 165(c)(3). The Internal Revenue Service has defined the
terms "sudden," "unexpected," and "unusual" in the following manner:

To be "sudden" the event must be one that is swift and
precipitous and not gradual or progressive.

To be "unexpected" the event must be one that is ordinarily
unanticipated that occurs without the intent of the one who suffers the loss.

To be "unusual" the event must be one that is extraordinary and
nonrecurring, one that does not commonly occur during the activity in
which the taxpayer was engaged when the destruction or damages
occurred, and one that does not commonly occur in the ordinary course of
day-to-day living of the taxpayer.

Rev. Rul. 72-592, 1972-2 C.B. 101, 101-02.
9. IRC § 165(h)(2).
10. "To abandon amounts paid and market prices as measures is to leave one's self

stranded in the intellectual desert of subjective values and psychic numeraires." Simons, supra
note 1, at 119.

if one purchases a vacuum cleaner and finds that it will not sweep, this
fact must be recognized in the computation of "final income." Such a
proposition may seem too obvious to question; yet one may apply the
same line of argument to purchases of patent cures that cure nothing,
informational literature that misinforms, and almost everything sold by
false representations.

Simons, supra note 1, at 120 (citing Economics and Accountancy 166 (1929)).
11. See e.g., Cal. Civ. Code § 1793.2 (West Supp. 1992); Fla. Stat. ch. 681.10-.i 18

(1991); Pa. Stat. Ann. tit. 73, §§ 1951-1963 (Supp. 1992); P.R. Laws Ann. tit. 10, §§ 2051-
2065 (Supp. 1989); Va. Code Ann. §§ 59.1-207.9 to .16 (Michie 1992).

[Vol 1:1



Valuing Personal Consumption

a loss. There is statutory support for this result?
But insurance recoveries are not always nontaxable, as Hal Millsap

discovered to his chagrin. 3 Millsap was not a lucky fellow. His residence
was severely damaged by fire and became temporarily unusable. Fortunately

for Millsap, his insurance policy provided him with $2,500 for temporary
living expenses, sufficient, he thought, to bear the expense of a motel. But

this turned out not to be true, as the Internal Revenue Service succeeded in

including the proceeds in Millsap's income.' 4

My intuition that most of us would think this unfair turned out to be

correct, as Congress quickly reacted to overturn this result.' 5 In the future,

people whose homes are damaged or destroyed by fire, storm or other

casualty will not be subjected to tax on the portion of their insurance

recovery for living expenses that exceeds their normal living expenses.' 6 The

legislation, however, is limited to the Millsap situation-extraordinary living

expenses resulting from loss of use of occupancy of a principal residence-

and does not cover such cases as car rental to replace a stolen car or even

compensation for loss of use of a second residence.
Can any principle, beyond intuition, explain when and why a

recovery from an insurance company or tortfeasor should be nontaxable? Can

such a principle reconcile non-taxation of such recoveries with the lack of

deduction to Andy if he is uninsured and personally bears the loss? Alter-
natively, would consistency require that a deduction be allowed to the

unreimbursed Andys of the world to reflect the difference between their

situation and the situation of those who were made whole without being
subject to tax? In other words, should the tax system explicitly take account

of any differences between expenditures and actual enjoyment of goods and

services?

12. A specific exemption applies to amounts received on account of death. IRC
§ 101, personal injury or sickness, IRC § 104(a)(3), for living expenses resulting from loss of

use or occupancy of a principal residence due to a casualty such as fire or flood. IRC § 123,
and for an insurance recovery, with respect to property destroyed by casualty, which is

reinvested in similar property, IRC § 1033. The list of exceptions suggests insurance recoveries
could be taxable in the absence of a specific statutory exemption. Most people probably
believe there should be no taxable income to a tortfeasor when the insurance company pays
a victim's claim against her but it is not clear how to reach this result under the present Code.
Boris I. Bittker, A Comprehensive Income Tax Base? A Last Word 126 (1968): see Charles
0. Galvin & Boris I. Bittker, The Income Tax: How Progressive Should it Be? 64 (1974).

13. Millsap v. Commissioner, 387 F.2d 420, 423 (8th Cir. 1968).
14. Id.
15. IRC § 123, titled "Amounts Received Under Insurance Contracts For Certain

Living Expenses."
16. The exclusion is limited to the actual living expenses incurred during the period,

in excess of the normal living expenses that would have been incurred during the period. IRC
§ 123(b).

19921



Florida Tax Review

Initially, it would seem that the answer to these questions requires an
understanding of the proper tax base under an income tax system. An
examination of the literature, as more fully described in Part II, does not,
however, indicate any unanimity of opinion as to the proper role of the actual
value of goods and services enjoyed by a taxpayer in determining the
applicable tax base. Some suggest that consumption is irrelevant-we should
care only about income, namely the amount available for consumption. 7

Others argue that ideally we should look to actual enjoyment of goods and
services-taking into account differences that are both large and measur-
able.8

Since I believe that most significant losses are a result of voluntary
behavior or could be avoided through insurance, I tend to prefer the ex ante
approach. Further, if the ex post perspective were adopted, both equity and
efficiency, in terms of incentive to insure, would require the recognition of
gains (more than expected enjoyment) in addition to losses. Since I believe
this is unlikely to occur, I ultimately conclude that regardless of whether in
principle one favors an ex ante or ex post approach to measurement, con-
sumption should normally be measured ex ante by the amount expended. As
suggested, I reach this result because it can be shown that for administrative
reasons an ex ante approach may in many instances be a better second best
approximation of actual enjoyment than an attempt to measure such enjoy-
ment is likely to be. Finally, I believe that exemption for insurance and other
recoveries can generally be reconciled with nondeductibility of losses for
those who are not reimbursed.

Because the degree of comfort one has with these suggestions may
well depend upon one's view of the ideal tax base, I begin in Part II with a
discussion of the "correct" tax base under an income tax. Although I deal
with the issue on the assumption that we continue the present income tax
base, the question must also be faced under a consumption tax which would
generally be measured as income plus or minus the change in savings during

17. Simons, supra note 1, at 139; see Mark G. Kelman, Personal Deductions
Revisited: Why They Fit Poorly in an "Ideal" Income Tax and Why They Fit Worse in a Far
from Ideal World, 31 Stan. L. Rev. 831, 835 (1979).

18. William D. Andrews, Personal Deductions in an Ideal Income Tax, 86 Harv. L.
Rev. 309, 313 (1972). According to Robert M. Haig:

Modem economic analysis recognizes that fundamentally income is a flow
of satisfactions, of intangible psychological experiences.... If one spends
his dollar [of income] for something more durable than a dinner-say a
book or a pipe-is his true income the book or the pipe, or the series of
satisfactions or "usances" arising from reading the book or smoking the
pipe? There is no doubt as to the answer.... A man strives for the satis-
faction of his wants and desires and not for objects for their own sake.

Robert M. Haig, The Concept of Income-Economic and Legal Aspects, in Readings in the
Economics of Taxation 54, 55 (R. Musgrave & C. Shoup eds. 1959).

NVot. 1:1



Valuing Personal Consmnaption

the year.' 9 I then turn in Part III to the implementation of the ex ante and
expost alternatives without regard to the possibility of recoveries of loss from
insurance, tortfeasors or otherwise. In Part IV, I explore the treatment of such
recoveries. Part V, which concludes with a brief discussion of insurance to
protect against loss of income, explores whether insurance raises total utility
by providing security and peace of mind.

II. CHOOSING THE "PROPER" TAX BASE

I begin by considering whether the reason income is chosen as a tax
base can help determine how income should be defined. Some view income
as the best indicator of ability to pay for the cost of govemmentZ) They
sometimes assert that the tax rates should be set to extract an equal level of
personal sacrifice. 2' By itself "equality of sacrifice" is an ambiguous term.
It can either mean that the loss of units of utility demanded of each individual
be equal ("equal sacrifice") or that each taxpayer should be required to give
up an equal percentage of total utility derived from income ("proportionate
sacrifice"). If utility per dollar of expenditure declines as income rises, the
latter interpretation would require progressive rates.

Others have more explicitly focused on the role of the income tax in
redistributing society's resources. According to Professor Alvin C. Warren,
"an income tax serves to deflect to the government, for the purpose of
providing for public goods and for distribution of the remainder, a progres-
sive portion of each citizen's share of the product otherwise allocated to him
by transfers and the marketplace."'  Professor Thomas Griffith, expressing
a similar view, explicitly grounds his proposals upon principles of distributive
justice.24 He offers two notions of welfare maximization for consideration:

19. See William D. Andrews, A Consumption-Type or Cash Flow Personal Income
Tax, 87 Harv. L. Rev. 1113, 1149 (1974). Under the Andrews proposal. the tax base would
be more or less as it is now decreased by any addition to savings or increased by withdrawal
from savings.

20. See generally, Joseph Bankman & Thomas Griffith. Social Welfare and the Rate
Structure: A New Look at Progressive Taxation, 75 Cal. L. Rev. 1905 (1987).

21. For those who view the payment of taxes as a "coerced contribution to the
government ... a confiscation of property," the equality of sacrifice argument assures that the
pain associated with taxes is apportioned equitably. Walter J. Blum & Harry Kalven. Jr., The
Uneasy Case for Progressive Taxation, 19 U. Chi. L. Rev. 417, 455 (1952). Generally, this
argument ignores the government benefits originating from taxation and concentrates on the
sacrifices taxes entail.

22. Id. at 457.
23. Alvin Warren, Would A Consumption Tax Be Fairer Than an Income Tax?. 89

Yale L.J. 1081, 1090 (1980).
24. Thomas D. Griffith. Theories of Personal Deductions in the Income Tax. 40

Hastings LJ. 343, 345 (1989). Griffith criticizes Professors Kelman and Andrews for not

1992]



Florida Tax Review

maximizing total societal welfare, utilitarianism, or improving the lot of the
least well-off member of society, the Rawlsian maximin. 5

Griffith's model assumes that the marginal utility from any given
expenditure declines as income rises and is the same for all individuals at any
income level.26 If the tax is to be assessed so as to cause the minimum
reduction of total individual utility, it follows that, except for incentive
effects, the tax rate would be 100% at the highest income levels until the
requisite revenue was collected.27 In measuring the base, any expenditures
that do not provide utility, which in this model would include medical
expenses, should be excluded.28

In any event, whether the goal is redistribution, or equal division of
sacrifice, for the goal to be accomplished the tax base must correctly measure
resources. Redistribution requires that we take more from those who enjoy
greater resources in order to provide for those who initially have smaller
claims. Equality of proportional sacrifice also requires that those with more
resources should pay more tax. In either case, two individuals with the same
resources should pay the same tax.

It is not clear, however, what we mean by resources in this regard.
As Henry Simons described it, income is the algebraic sum of consumption29

and accumulation.3' One way of thinking about Simons's maxim is to focus
on the term income, the right side of the equation being merely descriptive
of the only two uses of income. If one takes this view, one might well
conclude that how one spends income and the results of such spending are
irrelevant.3' Income is important not use.

Professor William D. Andrews, on the other hand, focused on the
right side of Simons's formulation32 and concluded that the ideal tax base

explicitly grounding their proposals upon any "coherent normative principle," such as a
principle of distributive justice. Id.

25. Id. Utilitarianism seeks to provide the greatest good to the greatest number; thus,
it is willing to sacrifice individual desires in the name of maximizing society's total benefit.
The Rawlsian maxim, with its bottomup perspective, proposes to maximize the utility of the
marginal members of society, despite the decrease in society's total benefits.

26. Id. at 392-93.
27. Id. at 394.
28. Id.
29. More precisely, Simons defines consumption as "the market value of rights

exercised in consumption," which, if taken literally, means that Simons would not have taken
account of what was actually consumed. Simons, supra note 1, at 50.

30. Simons defines accumulation as "the change in the value of the store of property
rights between the beginning and the end of the period in question." Simons, supra note 1, at
50.

31. See Kelman, supra note 17, at 835.
32. See Andrews, supra note 18, at 313. Of course, such an interpretation of

Simons's maxim makes a "shift toward a more explicitly consumption-oriented tax ... appear

[VOL 1:1



Vahing Personal Constumption

took account of the actual value of goods and services or consumption
enjoyed by the taxpayer.33 According to Professor Andrews, consumption
outcomes ought to be taken into account when it is reasonable to do so, such
as when there are large differences that are administratively feasible to
measure.

34

Professor Andrews's provocative article has led, in recent years, to
a lively debate about the proper role of personal deductions under an income
tax.3

1 The arguments raised in this debate, particularly with respect to
medical expenses36 and casualty losses, 37 are relevant to the more general
question considered herein: Should the value of consumption expenditures
ideally be based on the market price or on the amount of actual enjoyment
to the taxpayer? We now turn to this debate concerning the proper tax base.
Possible bases include earning capacity, power to consume after taking
account of the extent to which earning capacity has been exercised, and
ultimate utility or satisfaction perhaps only taking account of involuntary
losses of utility.

A. Alternative Standards

1. Earning Capaciy.-It might be said that ideally an income tax
should be based upon earning capacity. This method would best reflect the
individual's potential to acquire resources and thus may "best assess the
sacrifice a particular taxpayer is making when he is asked to give a certain
sum of money to the state., 38

Implementation of this concept raises a number of questions. Would

much less radical than it is commonly assumed to be." Andrews, supra note 18, at 317.
Recognition of the importance of the consumption element lends support to my view

that entertainment and other forms of enjoyment derived while engaged in business should
ideally be included in the tax base even if the expenditure is wholly justified for business
reasons. See Halperin, supra note 3, at 862.

33. Andrews, supra note 18, at 313.
34. Andrews, supra note 18, at 329-330.
35. A variety of personal expenditures are deductible under the present income tax,

including medical expenses, IRC § 213 and casualty losses, IRC § 165(c)(3).
36. See Griffith, supra note 24: Louis Kaplow. The Income Tax as Insurance: The

Casualty Loss and Medical Expense Deductions and the Exclusion of Medical Insurance
Premiums, 79 Cal. L. Rev. 1485, 1493-99 (1991); Kelman, supra note 17. at 858-79.

37. See Kaplow, supra note 36, at 1489-93 for a discussion concerning casualty
losses.

38. Kelman, supra note 17, at 841. Although Professor Kelman raises the earning
capacity argument, he ultimately rejects this standard in favor of a net receipts tax. He believes
"that once people deliberately exercise their earning power in the market, the tax system
should measure their relative positions as a prelude to redistribution." Keiman. supra note 17,
at 880.

19921



Florida Tax Review

actual earnings in excess of the amount expected from an individual of a
particular ability also be taken into account? While individuals who spend
their time beachcombing or who leave their money in a mattress would have
some income imputed to them, it is more difficult to determine the correct
treatment of a taxpayer who tries but is unsuccessful, either in investment or
in exploiting the value of her personal services. The assertion that you should
have done better and, therefore, we will include more than you actually
earned in income is troubling, I think, even under a capacity approach.

These questions have not been explored because no one argues that
the tax base should actually be based upon capacity. There are two reasons.
First, such a tax would violate the simple libertarian principle that the state
should not require people to engage in particular activities. For example, if
the tax were based on how much a taxpayer could earn, a law school
professor might be forced to practice law in order to pay the tax. Second,
measurement of earning capacity would be too intrusive. Presumably, in
addition to education level, one would have to measure such attributes as
intelligence, judgment and charm.39

2. Income or Power to Consume.-These concerns could lead one
to reject earning capacity as a tax base in favor of income or realized power
to consume. This approach is responsive to the belief, based on administrative
grounds, that the tax base must be measured mostly by money earned in
market transactions. It also would be simpler to implement and far less
intrusive than efforts to measure either earning capacity or the actual value
of consumption.

Beyond administerability, there is logic to measuring taxable capacity
by power to consume. The assertion that equality of such power is all we can
hope to achieve is appealing. As Kelman puts it, "[o]nce the taxpayer volun-
tarily takes control of resources, her particular subsequent uses of those
resources are irrelevant to tax law., 40 Behavior will be most efficient if
individuals bear the full burden of their errors. At least in terms of expectan-
cy, the individual will get what he pays for and even outcomes will often be
subject to the individual's control. For example, one who does not research
thoroughly before purchasing can expect to get less bang for the buck. In
addition, variations in outcome could in some circumstances be prevented by
adopting less risky behavior or by the purchase of insurance. Although I think
the issue is a close one, on balance, I find the arguments for the income
approach more persuasive than those in favor of the actual enjoyment
approach.

3. Actual Enjoyment.-There is, nevertheless, force to the idea that
the disparities we are concerned with are disparities in actual enjoyment of

39. See generally Kelman, supra note 17, at 855.
40. Kelman, supra note 17, at 835.

[VOL 1:1



Valuing Personal Consumption

real goods and services.4' This may be particularly true if failure to utilize
consumption power to the fullest is not the fault of the individual. In many
cases, consumers constrained by limited resources cannot adequately prevent
losses.4 2

According to Professor Andrews, while the concept of ability to pay
may suggest that taxes should be levied by the receipt of income (which
shows such ability, whether or not it is devoted to personal consumption), an
ideal personal income tax would apportion tax burdens to a taxpayer's
"aggregate personal consumption plus accumulation of real goods and
services. ,4 3 "[I]ncome once earned and received is [not] income whatever
is done with it."44 Andrews's focus on ultimate enjoyment of real goods and
services makes sense "if we think part of the purpose of a graduated income
tax is relative redistribution ... [since] [wihat we mean to redistribute ... must
be shares of real goods and services which persons otherwise would be
consuming .... ,4

According to Andrews, "[w]e rely on money expenditures to provide
a practical measure of the real consumption ... which such spending buys"
only because we cannot measure consumption directly.46 But, "if consump-
tion ... of real goods and services is less than ... money income ... in any
substantial and ascertainable way, ... that discrepancy should be adjusted for
by a deduction from money income." '47

Griffith's focus on minimizing sacrifice would also apparently lead
to a tax base based on actual enjoyment. Griffith appears to agree with
Andrews that an "unvarnished net income tax base would not ... redistribute
to those most in need ... because net income is likely to be a less accurate
measure of need than a tax base that takes into account such items as medical
expenses and casualty losses...."

Professor Stanley A. Koppelman also advocates a more direct linkage

41. Andrews, supra note 18, at 326.
42. The poor are more often victimized by limited outlets, insufficient consumer

information and exploitative rental policies. Cf. Kelman. supra note 17. at 860 n.87 (noting
that the poor are more likely to make "bad buys").

43. Andrews, supra note 18, at 327. Andrews says the tax should be based on con-
sumption of real goods and services because that is what will be ultimately sacrificed to pay
the tax. Andrews, supra note 18, at 327. Many have noted, however, that there would not
necessarily have to be an identity between the tax base and the medium of payment. See e.g..
Halperin, supra note 3, at 883.

44. Andrews, supra note 18, at 325.
45. Andrews, supra note 18, at 326.
46. Andrews, supra note 18, at 327.
47. Andrews, supra note 18. at 325.
48. Griffith, supra note 24, at 384.

19921



Florida Tax Review

between the tax base and theories of welfare maximization,49 such as
Rawlsian or utilitarian philosophy.5" Under his view, "the consumption
component of income involves the exercise of economic consumption power
in a manner intended to produce a current personal benefit."'" While
voluntary expenditures unrelated to profit-seeking activity should be
considered taxable consumption, Koppelman agrees that deductions for
involuntary expenditures may be justified because the involuntary nature of
the transactions may suggest that the taxpayer does not receive personal value
equal to the amount of the expenditure.5" There may be no personal benefit
because "the expenditure compensates for a psychic, noneconomic loss which
necessitated the expenditure." 3

B. Choosing the Base

This discussion suggests that a reasoned argument could be made
either for the view that ultimate enjoyment should be taken into account or
that we should focus solely on expenditures without regard to how things
work out. In all likelihood neither reason nor logic will help us choose
between an ex ante or ex post approach to measuring the benefit of con-
sumption expenditures. This dilemma could be avoided, however, if we could
conclude either that there is not likely to be a material difference in the
results of the two approaches, or that, because of likely imperfections in
implementing the ex post perspective, it is unlikely to do a better job of
measuring actual consumption than the ex ante method. These questions are
discussed in the next section.

III. IMPLEMENTATION OF Ex ANTE OR Ex POST APPROACH

In this Part, I explore a number of points which relate to the question
of whether a tax system which ignores consumption gains and losses is,
nevertheless, likely to give a fairer picture of ultimate satisfaction than one
which uses an ex post, direct, measure of such enjoyment.

I believe that the ex ante approach may be more accurate for a
number of reasons. First, there is no reason to suspect that there would often
be large differences between expenditures and enjoyment. In part this is true
because what at first glance appears to be a loss may on a closer examination

49. See Stanley A. Koppelman, Personal Deductions Under an Ideal Income Tax, 43
Tax L. Rev. 679, 697-704 (1988).

50. Id. at 698-700.
51. Id. at 705.
52. Id. at 709.
53. Id. at 710.

[VOL 1:1



Valuing Personal Consuniption

be merely a reflection of a pattern of consumption. In addition, we might
expect losses and gains to even out. But it seems likely, as is developed
below, that any attempt to measure consumption ex post would not only
exclude some losses while allowing others, but more importantly would
systematically exclude gains while allowing losses. If this is the case an ex
post approach would understate aggregate income and it may well tend to

exclude gains achieved by the very persons whose losses would be taken into
account. Finally measuring consumption ex ante does not preclude taking into
account an increase in the market value of housing and other consumer
durables.

A. Ex Ante and Ex Post Measures May Not Differ Significantly

If there is no reason to suspect that any one individual or class of
individuals is likely to experience unusual differences between cost and actual
enjoyment, it is at least possible that the purchase price will give a fair
indication of overall enjoyment. To pursue this further, let us look at the
reasons previously identified as causes for the difference between purchase
price and outcome.

Consumer surplus, which measures the difference between willingness
to pay and price, is one reason for the difference between cost and actual
enjoyment. I see no reason why consumer surplus would not be somewhat
equally distributed, at least among people in the same income class. Some
suggest that consumer surplus is not "distributed unevenly across classes;
thus, it poses no vertical equity problems."' On the other hand, it seems to
me that those with large amounts of income would willingly pay much more

than market price for such items as ordinary entertainment and travel." This
income effect would create larger consumer surplus for each dollar spent by
the well-to-do.

Although changes in market prices and variation in the quality of the
particular item among models should be random, choosing wisely among
models and avoiding fraud and misrepresentation might not be. One suspects
that the less well off or the less well educated are more likely to be
victimized by lower quality products as they have less resources with which
to research potential purchases and access to fewer sales outletsY

Still, if the amount of difference between cost and outcome is not

54. Mark G. Kelman, Time Preference and Tax Equity. 35 Stan. L Rev. 649, 657
n.23 (1983).

55. But see Herbert Hovenkamp, Positivism in Law & Economics. 78 Cal. L Rev.
815, 839-40 (1990) (arguing that while the rich may have a greater willingness to pay, that
does not necessarily mean they achieve greater utility from a purchase).

56. See Kelman, supra note 17, at 860 n.87.

19921



Florida Tax Review

likely to be significant among people at a particular income level, it may not
be important that there might be large differences between one income level
and another. These vertical differences can be handled by varying the rates
more than the amount required by the desired degree of progression."
Suppose a rate of 15% was imposed on lower levels of income and a rate of
28% on higher levels of income on the assumption that the tax base correctly
reflected all resources or at least an equal proportion thereof. If it were
believed likely that income at the higher level was understated to a greater
degree, lowering the 15.0% rate to 14.5% and/or raising the 28.0% rate to
28.5% could ameliorate the vertical inequity. I recognize that it may be naive
to believe that rates are chosen solely to achieve a certain distribution. There
may be political or efficiency objections to higher rates even if the distribu-
tion of resources is less than "ideal." For example, high marginal rates may
be said to inhibit investment or work effort. Nevertheless, at least in theory,
vertical inequities can be eliminated through a rate adjustment without
correcting the tax base.

B. Apparent Losses Are Not Always Real

While we all may commonly experience "the minor frustrations of
life," the burdens of serious casualties and illnesses may not be equally
distributed even at a particular income level. Andrews suggests there would
be "large differences in ... [the] magnitude ... [of medical expenses] between
people in otherwise similar circumstances."58 This also seems to be the view
of Professor Boris I. Bittker who defends the deduction for casualty losses as
follows.

In a statistical sense, of course, destruction by fire is
one of the hazards of home ownership "voluntarily" assumed
when the taxpayer chooses to buy a personal residence. But
if a dog can distinguish between being kicked and being
stumbled over, as Holmes asserted, we can properly distin-
guish between the minor frustrations of life-a cigarette burn
in a rug, a dented fender, a quarter lost when fumbling for
change to put in a parking meter-and major casualties
("sudden, unexpected, and unusual" events that do not
"commonly occur in the ordinary course of day-to-day
living").59

57. Griffith, supra note 24, at 360-363.
58. Andrews, supra note 18, at 336.
59. Boris I. Bittker, Income Tax Deductions, Credits and Subsidies for Personal

Expenditures, 16 J.L. & Econ. 193, 197 (1973). Bittker draws some support for the casualty

[ Vol 1: 1



Valuing Personal Consumption

On the other hand, it may be that what appears to be bad luck would
often be a direct result of the consumer's behavior. For example, according
to Kelman, it is plausible that "most medical needs really arise from
voluntary decisions to pursue potentially unhealthful activity '" such as
smoking, skiing, or not closeting oneself away from potentially tortious
automobile drivers. Thus, Kelman asserts that many departures from good
health result from past decisions that gave the taxpayer untaxed benefits."

A deduction for loss of health without inclusion of the gains
from such risky activities will mismeasure how relatively
well-off two taxpayers are.... Recognizing that medical
spending may involve discretion before illness manifests
itself, it is no longer plausible to say that such spending
"restores" the taxpayer to a baseline position enjoyed by
those without medical care "needs."'

Professor Koppelman finds, for different reasons, that most currently
deducted medical expenditures have a strong voluntary component."l Since
those claiming the medical deduction are typically insured against involuntary
costs, he believes generally their uninsured expenditures are at least in some
sense voluntary.64 Thus, it cannot be said that medical treatment merely
restored the pre-illness condition.

Clocking more miles, driving faster than average or improper care of
an automobile are additional examples of voluntary activities that provide

loss deduction from an example which purports to show that you need to reach a peculiar
conclusion to justify determining Simons's income without regard to a casualty loss. In
Bittker's example, a taxpayer who earns S50,000, spends S20,000 on necessities, and S30,000
to replace a $20,000 residence destroyed by fire. According to Bittker. since increase in net
worth is $10,000 (new residence S30,000 compared to old residence S20,000), consumption
must be $40,000 which can only come from $20,000 of necessities and $20,000 of old
residence literally consumed by fire. Bittker, 16 J.L. & Econ. 193, at 196. However, it would
seem more accurate to measure lifetime consumption from a residence by the purchase price
in the year of acquisition. Andrews, supra note 19, at 1157-58. Under this view, the old
residence does not enter into the calculation. There is $50,000 of consumption-$20,000 of
necessities and $30,000 from the new residence.

60. Kelman, supra note 17, at 869.
61. Kelman, supra note 17, at 869. Of course, no one is suggesting this is true of all

illness.
62. Kelman, supra note 17 at 869.
63. Koppelman, supra note 49. at 712-13.
64. Koppelman, supra note 49, at 712-13. Of course this is not true to the extent that

a loss of job leads to loss of health insurance, and an individual is unable to secure a policy
which will cover pre-existing conditions. It is not clear how many people in this position
substantially benefit from the medical expense deduction. In any event the true solution to the
problem is continuing insurance coverage, not tax relief.

19921



Florida Tax Review

unaccountable benefits. The car may depreciate more than expected, but the
owner has also derived more consumption, or increased her leisure time.

Furthermore, differences in the quality of purchases which result
among taxpayers, perhaps even at the same income level, could arise in part
because some individuals spend more time on shopping and research, reading
Consumer Reports and the like. In a sense the better outcome merely reflects
the fact that the expected value among the more cautious group can be said
to be higher than that for those who are more cavalier in their purchases.
However, members of the cavalier group, all other things being equal, are
able to enjoy more leisure time, which, unlike the cautious group, they
presumably value more highly than time spent on research. For example,
members of the cautious group, after expending five hours on research, each
spend $1,000 on an item and the item averages $1,050 in overall quality.
Members of the more cavalier group also each spend $1,000 on a similar
product, but forgo the research and therefore purchase items that average
$1,000 in quality. The cavalier consumers also have five more hours of
leisure which they value at more than $50. Thus, the cautious consumers'
research raised their expectancy for the product that invariably led to a
greater outcome. The cavalier consumers knew, in the absence of research,
they might not get the best product (for example, they had an expected value
of only $1,000) but believed the extra leisure time was worth the difference.
In both cases average ex post results did not differ from what was expected
ex ante. In short, differences in outcome are merely a reflection of a lifestyle
providing additional noneconomic pleasures.

Persons who choose not to insure when insurance is available will get
a greater difference between expectancy and outcome than those who choose
to insure. The same can be said for those who make more risky consumption
choices. While in these cases there may be a real difference between
expectancy and outcome, it seems to me that the concerns which lead one to
prefer a tax base which takes account of actual enjoyment, as opposed to
mere opportunity to consume, may not extend to those whose losses are in
this sense voluntary.65

C. Attempts to Measure Outcomes Will Be Incomplete

In this section, I explore the possibility that even if actual losses can

65. This latter point seems even stronger if what is lost is a luxury item rather than
a necessity which must be replaced. See Koppelman, supra note 49, at 709. We will also see
later that allowing the uninsured to deduct losses is likely to be unfair because they will tend
to have gains which will be ignored. Thus, if uninsured losses could be deducted there would
be a disincentive to insure which would lead to inefficient behavior. Kaplow, supra note 36,
at 1491.

[VOL 1:1



Valuing Personal Consuimption

be identified, an ex ante approach might still come closer to measuring actual
enjoyment than an attempt to measure outcomes. First, past experience would
indicate we might only take account of selected losses. Perhaps more impor-
tantly, although it has not been generally understood, many of the transactions
that can cause losses can also provide gains to others. Since these gains are
unlikely to be recognized, taking account of losses could unfairly skew the
base. This would be particularly troublesome if it is likely that the people
who would tend to have losses are also more likely to have gains.

1. Consumption Windfalls.-It may be that sometimes people make
expenditures that do not increase their consumption. Individuals incurring
such costs may have a claim to a deduction. If so, however, there are others
who enjoy a benefit because they spend less than they expected, to achieve
a certain level of enjoyment.' These people should have additional income.
To illustrate, I present first the purchase and operation of an automobile, and
then turn to medical expenses and casualty losses, which have been the
subject of extensive discussion. 67

Assume Bob purchases a car for $10,000. To determine whether the
purchase is worthwhile, Bob would take account of the costs of operation,
including maintenance and insurance, as well as his initial outlay, and
compare the total costs with the benefits of driving. Assume that Bob expects
his costs to average $5,000 a year, but the car needs extensive repairs over
its life, so the average annual cost is $6,000.6S Annual costs will also be
more than expected if the car depreciates more quickly than assumed. It may
be said that Bob has suffered a loss, not necessarily because he derives less
value from the car, but because he will be able to buy fewer other goods and
services than expected.

On the other hand, assume the cost of operation for Carol, who owns
a similar car, averages only $4,000 annually. Since the benefits from driving
have not decreased (if anything, driving a trouble free car is more valuable),
Carol has an unexpected gain of $1,000. Put another way, she has the ability,
without any diminution of her enjoyment from driving, to buy an additional
$1,000 worth of goods and services as compared to what she could buy if her
car had an average repair record.

If the tax base is to reflect actual consumption, it should include the
$1,000 gain to Carol and allow the $1,000 loss to Bob. Since individuals may
take flawless performance for granted and certainly have no incentive to bring
it to the attention of the Internal Revenue Service, such gains would be even

66. See Daniel Shaviro, The Man Who Lost Too Much: Zarin v. Commissioner and
the Measurement of Taxable Consumption, 45 Tax L. Rev. 215, 232 (1990).

67. See supra notes 17-18 and 36.
68. The comparison is actual to expected costs, not higher cost of repairs on one

make of car compared to another.

19921



Florida Tax Review

harder to measure and take into account than losses.
Medical expenses raise similar concerns. Some argue that a deduction

for medical expenses, in excess of some percentage of gross income, is
justified because such costs do not provide greater than expected consump-
tion.69 Like Bob, who does not get more than the normal benefit from his
car despite the extra expenditure, such individuals are no healthier than
average.

According to Andrews, "[tihe right basis for making interpersonal
welfare comparisons ... is that ultimate object, good health, rather than the
intermediate good..." medical services, which is not a proxy for good
health.7 ° If anything, the relation is apt to go the other way. "As between
two people with otherwise similar patterns of personal consumption ..., a
greater utilization of medical services by one is likely not to reflect any
greater material well-being or taxable capacity, but rather only greater
medical need.'

While it would be impractical to include good health directly,
differences can be partially reflected by allowing a deduction for medical
services that will be used more by those in poorer health. This approach
"reflect[s] differences in health only as they manifest themselves in financial
terms. 72 In other words, the best tax base would be money income plus
some amount to take account of varying degrees of health, including dis-
abilities encountered from birth. If the base is purely money income, those
with poor health would be overtaxed. Since high medical expenses tend to
reflect poor health, a deduction for such expenditures would be a proxy for
taking account of such poor health directly.

Professor Kelman raises a number of objections to Andrews's
argument.73 He disagrees with Andrews's conclusion that expenditures on
medical care are a good proxy for differences in health.74 Essentially,
Kelman believes that people with similar conditions spend different amounts
or none at all on health care, reflecting to some extent general consumption
desires.75

Therefore, a medical expense deduction based on the nature of the
injury or illness would provide a better picture of differences in well-being.
In other words, if B and C suffer from an illness while A is in perfect health,
B and C can both be said to be less well off than A by an identical amount.

69. Andrews, supra note 18, at 335-37.
70. Andrews, supra note 18, at 335-36.
71. Andrews, supra note 18, at 314.
72. Andrews, supra note 18, at 336.
73. Kelman, supra note 17.
74. Andrews, supra note 18, at 336.
75. Kelman, supra note 17, at 869-70.

[VOL 1:1



Valuing Personal Consunption

B may choose to treat his illness with heavy medical expenditures while C
perhaps does not need to because he has adequate insurance or would just
rather console himself by eating and drinking well.

If one, nevertheless, concludes that excessive medical expenditures
fairly reflect differences in well being which should be taken into account in
determining the tax base, it would follow that individuals who are still able
to enjoy good health despite expending significantly less than the average are,
like Carol, able to enjoy unexpected additional amounts of other goods and
services. Ideally, this would be taken into account if excessive medical costs
are to be deducted. Again, however, this is unlikely to occur.

A similar problem arises with respect to what might be called
"casualty gains" 76 which correspond to casualty losses for uninsured persons.
Returning to Bob's $10,000 car, implicitly, the car must have an expected
value of at least $10,000 even though there is say a 1% chance it will be
totally destroyed in the very near future. A car which is not destroyed must
then yield $10,101 of enjoyment.' In other words, $10,000 reflects the
average expected value of all cars, not just cars which are neither victims of
casualties nor lemons.78

Taxpayers who do not have an accident obtain $ 10,101 of automobile
value for $10,000. The one driver in one hundred who suffers a casualty has
a $10,000 loss. If this loss were allowed, the total consumption of the group
would be undervalued unless each person who did not have an accident re-
ported a $101 gain.79 If neither gains nor losses are taken into account the
group as a whole would report the correct taxable income.

Since it would be difficult to explain to taxpayers why a good repair
record, excellent health, or avoiding an accident should result in taxable
income, the type of gains just discussed will not, and perhaps cannot, be
taxed. To the extent that losses and gains could be expected to be incurred
by the same persons, as they would if some people systematically under-
insure or take greater risks in their purchases, we might expect to more
accurately measure individual consumption by outlays than if we tried to take
outcomes into account. Indeed, the notion of allowing losses for such risk-
neutral individuals while ignoring the noneconomic gains from such risky
activity suggests that consumption is best measured by expenditures rather
than outcomes.

Louis Kaplow reaches the same result by a different approach. He

76. Kaplow, supra note 36, at 1502.
77. This is assuming there is no consumer surplus. A 99% chance of enjoying

$10,101 equals $10,000.
78. It would seem that $10,000 is the expected value to the marginal risk-neutral

purchaser. If an individual were risk adverse she might demand an expected value in excess
of $10,000 in order to make the purchase.

79. Shaviro, supra note 66. at 241.

19921



Florida Tax Review

finds that if uninsured losses are deductible, individuals would be driven by
the tax law to absorb more risk than desired.80 Since the loss deduction
provides free partial insurance, a taxpayer who chooses to insure and forego
the free partial insurance pays for more than the additional coverage acquired.
Therefore insurance may be rejected where it would otherwise be chosen.
Because, all other things being equal, they would prefer to avoid the
additional risk, if there is an opportunity to insure, individuals would ex ante
uniformly prefer lower tax rates and nondeductibility of losses as long as the
rate cut was such as to be neutral, on distributional grounds, compared to
expected losses. In these circumstances, loss deductibility accompanied by
higher rates would leave everyone worse off. Thus, Kaplow is able to find
deductions for medical expenses and casualty losses undesirable without, he
says, going through the type of exercise pursued in this article.8" But it is
important to recognize that the loss deduction would cause an undesirable
change in behavior precisely because of the disparity between the treatment
of uninsured gains, which would not be taxable, and uninsured losses, which
would be deductible.82

2. Only Selected Losses Will Be deducted.-Federal income tax law
has only taken account of medical deductions and losses due to casualty.83

It is not clear what distinguishes casualty losses as defined in the tax law
from other disasters. It may be that the requirement that the loss be "sudden,
unexpected and unusual"' 4 is essential to distinguish extraordinary occur-

80. Kaplow, supra note 36, at 1500 n.68 & 1505.
81. Kaplow, supra note 36, at 1487. While he does not, at least initially, identify the

failure to take account of gains to the uninsured as the culprit, at a later point in the article
Kaplow, citing a draft of this article, notes the idea that a casualty loss deduction is
asymmetric because it does not include in income what he refers to as casualty gain. Kaplow,
supra note 36, at 1502.

82. If gains were somehow accounted for, deductibility of losses would not cause
individuals to under insure. Thus, exploration of issues of income measurement helps explain,
and thereby reinforces, Kaplow's conclusion.

83. Losses from bad debt on bona fide loans, not connected with a trade or business,
are treated as capital losses. IRC § 166(d). It has been suggested that bad debts even from
loans to relatives or friends reduce net worth and if repayment was genuinely expected, the
loss should be deductible. Bittker, supra note 59, at 199.

It seems to me, however, that in many circumstances, a potential gift was contem-
plated. For example, a taxpayer may loan $10,000 to a relative, and while repayment is
expected, he believes there is a 20% possibility that default will occur. While it is possible that
the interest rate is high enough to compensate for the potential default, and the loan would be
made by the taxpayer to a stranger under similar circumstances, it seems more likely that there
are personal motivations and the taxpayer may well be willing, perhaps without explicitly
thinking about it, to make a gift with an expected value of $2,000. When the "gift" turns out
to be $10,000 instead, would there be an argument for an $8,000 loss? If so, is there $2,000
of income if the loan is repaid and the gift becomes costless?

84. See Rev. Rul. 72-592, supra note 8.

[VOL 1:1



Valuing Personal Consimption

rences from normal wear and tear. According to Kelman, however, casualty
losses resemble other disappointments with purchased commodities.s"
Automobiles that are "lemons" may involve as big a loss from expected
return as those which suffer a fire.86

Professor Kelman also finds it hard to distinguish health from other
psychic pleasures. He believes that it is not plausible that taxpayers differ in
health far more than they do in other forms of well being, such as untreated
depression, bad marriage, or a boring job.8 In fact, these other conditions,
unlike poor health, are more likely to be inflicted on the poor. Therefore, it
is not appropriate to take account of decline in health without also comparing
other noneconomic attributes.

In this connection, this article would be remiss and out of step with
much of the recent tax scholarship if it did not refer to Zarin v. Commission-
er.88 As probably every reader of this article well knows, Zarin lost over
three million dollars gambling with chips furnished him on credit by a casino.
When the debt was settled for one-half million dollars, the Internal Revenue
Service charged Zarin with discharge of indebtedness income for the dif-
ference. Zarin lost in Tax Court but eventually won on appeal.

Of concern to this article is whether Zarin has a loss equal to the
difference between what he spent and the presumed enjoyment derived from
gambling.89 As Daniel Shaviro has pointed out, given the amount of money
and time Zarin devoted to gambling (huge sums, twelve to sixteen hours a
day, over a period of several months), it is likely that his actual losses
coincided with the amount he should have expected to lose." Therefore, the
only argument for a loss deduction would be that because he was a compul-
sive gambler, Zarin's actual enjoyment was much less than even the expected

85. Kelman, supra note 17, at 860 n.87; see Boris I. Bittker & Lawrence Lokken.
Federal Taxation of Income, Estates and Gifts 1 34.2 (2d. ed. 1989) (listing of significant
losses which have been denied as not attributable to casualties).

86. Professor Kelman also believes that casualties are more likely to be evenly dis-
tributed than "bad buys," which the poor are more likely to make. Kelman. supra note 17. at
860 n.87. As discussed above, vertical equity may not be a concern, however. See supra note
57 and accompanying text.

87. Kelman supra note 17, at 869.
88. Zarin v. Commissioner, 92 T.C. 1084 (1989). rev'd. 916 F.2d 110 (3d Cir. 1990);

see Babette B. Barton, Legal and Tax Incidents of Compulsive Behavior. Lessons From Zarin,
45 Tax Law. 749, 749-50 nn.4 & 7 (1992) (Zarin is "widely celebrated" in law school texts
and legal commentaries).

89. Perhaps more important is whether Zarin has a non-taxable recovery, regardless
of whether the loss is recognizable. Again, however, this presupposes a loss. But if Zarin
achieved full value from his gambling there is no loss to be recovered. Therefore, the issue
remains the same--should Zarin's mental state be taken into account in determining the value
derived from gambling?

90. Shaviro, supra note 66, at 233.

19921



Florida Tax Review

cost.9' Even if an ex post approach to consumption were to be adopted, it
is unlikely that such subjective feelings, as opposed to differences in outcome
which could be objectively measured, would be taken into account. 92

This reinforces the conclusion that a tax base which ignored losses
could be fairer than one which took account of only some losses, such as
casualties and medical expenses. This would be particularly so if losses of all
types combined tended to be fairly evenly distributed.

D. Lost Paycheck

The previous sections in this part suggest that ex post measurement
should be approached with extreme caution, if not actually rejected. Still, in
limited circumstances, we need not measure consumption by the amount of
income earned. Some who generally prefer an ex ante measurement would
favor a deduction for a taxpayer who suffers the loss of his paycheck on the
way to the bank. This income may be viewed as never effectively received
or appropriated and, therefore, equivalent to never having been earned in the
first place.93

Of course, the reasons that cause us not to base income on earning
capacity do not apply once income has been earned. The amount earned is
measurable, and no one has been forced to work in order to meet his tax
liability. Still, one who is sympathetic to an ex post result but is concerned
about the measurement problem might allow losses to reflect a stolen pay-
check, or perhaps misplaced cash in general, even if they ordinarily disregard-
ed losses once consumption took place. In this view, a taxpayer whose pay-
check is stolen gets no benefit from the transaction and should be entitled to
a deduction.94 Putting aside problems of proof as to whether a loss actually

91. Cf. Shaviro, supra note 66, at 236-39.
92. Shaviro, supra note 66, at 225. If one is sympathetic to Zarin's cause, as most

seem to be, it seems a better course to argue he never purchased more than $500,000 worth
of chips because the casino's course of conduct made it impossible to enforce the debt.
Shaviro, supra note 66, at 244. While such a bargain purchase, if it occurred, would not be
treated as income, this description of events may be hard to accord with the facts. Barton,
supra note 88, at 762-65. However, one way to view the settlement is as a purchase price
adjustment, analogous to IRC § 108(e)(5). Shaviro, supra note 66, at 249.

93. See Kelman, supra note 17, at 859 n.87; William Andrews, Basic Federal Income
Taxation 519 (3d ed. 1985).

94. An analogous situation may arise in the case of a taxpayer who erroneously over-
pays her income taxes. In this case, like the loss of cash, the amount the individual can
actually spend on consumption is less than would be expected based on the income earned.
In the next section, at note 126, I consider whether taxpayers who recover from an accountant
or other person who caused an excessive income tax to be paid can exclude that recovery from
income. If the exclusion is allowed, should a taxpayer who is unable to recover be entitled to
a deduction to recognize, as in the lost paycheck case, that his income does not fairly reflect

[Vol 1:1



Vahing Personal Consumption

occurred, the amount of the loss, since it is money, can be measured and it
is unlikely that the person suffering the loss would have unreported gains of
a similar type.

Nevertheless, it could be that the loss of a paycheck or theft of cash
is a result of a life-style choice, just in the way that some see many medical
expenses. For example, the taxpayer may lose his paycheck when he stops in
a bar on the way home from work. Further, allowing this type of loss and not
others could be viewed as favoritism toward some taxpayers.

E. Sale of Assets Held for Personal Use

Despite the general adoption of an ex ante approach to measuring
consumption, there is no question that if an automobile or a boat is volun-
tarily sold in a market transaction, gain, if any, will be taken into account.
We need to consider whether it is the sale, which makes such gains measur-
able, or whether there is some other attribute that distinguishes this case from
an absolutely wonderful dinner, perfect weather during vacation, or an extra-
ordinarily exciting baseball game, all of which might be worth more than the
price paid.

It must be recognized that regardless of how consumption is
measured, income would be determined according to how things turned out,
rather than by expectancies.95 A lottery ticket purchased for one dollar
probably has an expected value of less than fifty cents, but the winner of a
forty million dollar jackpot will be taxable on the forty million dollars. In
measuring income, ex post results count not ex ante expectations.

But if we are to take account of actual outcome with respect to
income but only expectancy as to consumption, it is necessary to clearly draw
a line between the two activities. This may not be easy.9  Many consump-
tion purchases have an investment aspect. The purchase of a residence, an
automobile, fine wine, art, or jewelry is in many ways an investment similar

his spending power?
95. See Michael J. Graetz, Implementing a Progressive Consumption Tax. 92 Harv.

L. Rev. 1575, 1601 (1979).
96. As Simons puts it:
Consumption, presumably, should be measured in terms of values at the
time of consumption; but to ignore changes in value between time of
purchase and time of use will ordinarily make very little difference. The
difference might be large, of course, where a man purchased choice
beverages and allowed them to acquire the quality and distinction of ripe
age.... At all events, there is suggested here the difficult question of where
to draw the line between acquisition of means and their employment. At
what point shall the idea of gain or loss be dropped. One finds no ready
answer....

Simons, supra note 1, at 119.

19921



Florida Tax Review

to the purchase of a stock or bond. With respect to items such as artworks
and jewelry, and perhaps a residence as well, a purchaser would often count
on appreciation to provide a return on his "investment" in addition to the
imputed income from use. Whether or not this is the case, with respect to
consumer durables it should be explicitly recognized that the true element of
consumption is the value of the use of the item and that the purchase is in
effect an investment.97 Unlike a ball game, dinner or vacation the consump-
tion value of more durable items still exists in the future, and it is possible
by sale to switch one's consumption to other items. This is hard or impossi-
ble to do for most items of consumption. When appreciation occurs, the
ability to shift consumption to another form seems to reflect increased market
power to consume and not merely the level of enjoyment from purchases
after the fact.

Therefore, it would be appropriate to take gains on consumer durables
into account even if unrealized. As Richard A. Epstein has noted, it appears
to be the realization requirement and not the ex ante approach that keeps this
from happening.98 This impact of the realization requirement is particularly
troublesome because the result is exemption rather than, as is the usual case,
deferral.

Exemption could extend to even realized gains under present law.
Because we do not adjust basis for depreciation99 there is gain on sale only
if the selling price exceeds the original cost. Apart from residences, where
gains are most often deferred and eventually forgiven, and art and other
collectibles, which generally do not depreciate, appreciation over the original
cost with respect to consumer items would be extremely rare. But, if we
adjusted basis for depreciation, which we should do whether or not deprecia-
tion is deductible, gains upon sale would be more common.

Apparent gain would reflect not only changes in market price but
such variables as better than average quality or lower than average wear."°

In the latter case, however, if depreciation were properly measured, there
would be no gain to report. For example, assume Agnes's automobile costs
$10,000, should normally last five years and decline in value at the rate of
$2,000 a year. After three years of use Agnes sells her car for $5,000 or
$1,000 above the depreciated basis of $4,000. The gain could reflect a rise
in market value (overall good quality) for automobiles of that vintage,
recognition, based on repair record or otherwise, that Agnes owns a better

97. Joseph A. Pechman, The Brookings Institution, What Should Be Taxed: Income
or Expenditure? 94 (ed., 1980).

98. Richard A. Epstein, The Consumption and Loss of Personal Property Under the
Internal Revenue Code, 23 Stan. L. Rev. 454, 457 (1971).

99. IRC § 167.
100. See Graetz, supra note 95, at 1617.

[Vol 1:1



Valting Personal Consumption

than average car,'0 ' or merely the fact that Agnes has driven fewer than
average miles.

If, however, depreciation in the value of an automobile reflects not
only the period of ownership but also the level of use and care, in the last
case the adjusted basis of Agnes's automobile should not be $4,000 but rather
should be $5,000 to indicate that depreciation over the three years of use was
less than the $6,000 expected. Put another way, Agnes did not get 60% of the
expected use of the automobile over the three years that she owned it; more
than 40% was saved for the future. She had only $5,000 of use and $5,000
of cash from her original $10,000 expenditure, and consequently she had no
gain.102

Professor Epstein recommended allowance for losses as well as
taxation of gains on transfers of consumer durables and residences (after
properly adjusting basis for depreciation).' Whether allowing a loss is
appropriate would again seem to depend on the reason why the loss was
incurred.,'t In addition to the possibility of a decline in market value for
automobiles of that vintage, a loss could reflect the fact that the car had been
damaged in an accident or is known to have suffered above average repairs
or just is not running as well as it should. All these events could reflect
additional consumption on the part of Agnes from more than average mileage
or from a style of driving or of care for the car (leaving more time for other

101. It is unlikely that the used car market would recognize Agnes's better than
average car. Because fraud often goes undetected in the used car market, consumers cannot
differentiate between good and bad cars. Thus, the rational consumer assumes that all cars are
of average quality, and the seller, without a method of proving the superior quality of his
automobile, has little choice but to ask the average price for the car. Eventually, better than
average cars will not be available in the market, the average quality will decline, and the price
which consumers are willing to pay for a car of unknown quality will be reduced. The cycle
will repeat itself driving owners of average quality cars out of the market. David M. Green.
Comment, Due Diligence Under Rule 415: Is the Insurance Worth the Premium?, 38 Emory
LJ. 793, 818 (1989). See generally Akerlof, supra note 4.

102. If the reduction in the rate of depreciation is due to diligent care and use, it may
be reflected by maintenance expenditures. This is not to suggest that depreciation charged
should be directly related to maintenance costs. Large repair records could reflect accelerated
wear and tear, not prevention of depreciation. In the absence of a method to distinguish be-
tween expenditures made to prevent depreciation and those made because of depreciation, there
may be no objective measurement other than mileage (which would not be entirely accurate)
to adjust depreciation to reflect actual wear.

103. See generally Epstein, supra note 98.
104. It has been suggested to me that loss allowance is inappropriate as long as im-

puted income from use is not taxed. It is not clear that the two are necessarily related. I note
that gains and losses on the sale of state and local bonds are taken into account even though
interest income is exempt. It seems to be true, however, that if imputed income were taxed,
it could reflect an increase in the value of the car or a better than average repair record
depending upon how the amount of imputed income was computed.

19921



Florida Tax Review

pursuits) which led to excessive deterioration or even a major casualty.'0 5

If the loss occurred because Agnes drove greater than the expected
number of miles or failed to properly care for the car, allowance of a loss
would be clearly unjustified, as would recognition of gain in the converse
situation. If the loss was not a result of Agnes's behavior, however, recog-
nition might be appropriate.

One problem with recognition of such losses would be the adverse
selection that would occur from taxpayers selling assets which have declined
in value and holding on to those which have increased. In fact, more frequent
sales of items of low quality might occur even in the absence of such a tax
incentive. Even if realization were not required, comfort with loss recognition
might depend upon whether we would be able to identify comparable gains
because a particular car was better built than average or gratuitously had a
better than average repair record.' °6

In sum, even under an ex ante approach, gains on sales of consumer
durables are taken into account. It would be appropriate, moreover, to recog-
nize gains even in the absence of realization. In theory, an exception should
be provided for gains that are a result of less than average use but it seems
difficult to separate such gains from market changes. Furthermore, perhaps
such gains do not occur often enough to be worth the trouble, if, as is
suggested above, the market does not readily distinguish between a particu-
larly good used car and an average one. 1°7

While symmetry might require deduction of losses, particularly if
attributable to a market decline, a number of concerns arise. If realization is
required, it seems losses would be much more likely to be recognized than
gains. Again, even though the existence of a declining market could be objec-
tively documented, it seems impossible to separate and disallow "losses" that
reflect higher than average consumption in the particular case, and it may be
that it is these items which are most likely to be sold. Moreover, even where
extra consumption is not present, losses, because a particular item has per-
formed below average, are comparable to gains for those whose "costs" are
below average. Since discovering unrealized gains of this type is difficult,
loss recognition would seem likely to systematically understate total income

105. As Professor Kelman has noted, above average medical expenses can also be
the product of lifestyle-smoking, drinking, or engaging in other activities that are inherently
dangerous, such as skiing. Kelman, supra note 17, at 869.

106. There may be, for example, a number of 3-year old cars worth more than
$4,000 (or more than the average market value of that model at that age). Even if realization
were not required, how would such gains be discovered?

On the other hand, cars which suffer a casualty or are otherwise lemons are more
likely to be sold. See generally, Julie A. Vergeront, Note, A Sour Note: A Look at the
Minnesota Lemon Law, 68 Minn. L. Rev. 846 (1984).

107. See supra note 101 and accompanying text.

[VOL. 1:1



Valuing Personal Consunption

for society as a whole.'
If allowance of losses is unwise, as I believe it is, the question is

whether recognition of gains alone, perhaps even if unrealized, produces a
better tax base than if both gains and losses are ignored. The above dis-
cussion suggests that ignoring gains because we cannot allow losses is better
only if those who have gains are also more likely to have losses. There
should be some correlation to the extent gains and losses arise from the
ownership of particular assets or a penchant for more risky purchases.
Nevertheless, since the losses involved here do not necessarily arise from
failure to insure, matching of gains and losses is somewhat less likely.

Moreover, as a matter of equity, or at least perceived equity, there
seems to be little difference between an investment gain in the stock market
and an increased net worth due to appreciation in the housing market.
Certainly with respect to art, we could not exempt gains merely because a
taxpayer designates the item as purchased solely for personal use. Since I
believe that gains must be included in income and that allowance of losses
could be abused, I am forced to grasp for a way to distinguish losses due to
a decline in market value of consumer durables and residences from other
investment losses." 9

Consider that it may not be possible, even in the case of a market
decline, to identify the true amount of consumption derived from housing.
For example, assume an individual purchases a new residence for $100,000.
Assume, for ease of exposition, that the property is not expected to depreciate
during the first year."' The expected cost, and therefore consumption value
of the house, at a 10% interest rate, would be $10,000 a year which is the
imputed income from home ownership or the loss of income from withdraw-
ing $100,000 from other investments.'

If at the end of the year the residence were sold for $95,000, is the
$5,000 decline an investment loss, or can it be considered an additional cost
of consumption? Assume the owner would have been willing to pay $16,000
per year to live in the house. While the $6,000 premium over expected cost
would normally be considered consumer surplus, is it possible to consider at
least $5,000 of this amount as an additional cost of consumption when the

108. More precisely, loss recognition would underestimate total value of goods and
services.

109. The House has recently passed a provision allowing losses on sale of a principal
residence to be carried forward and allowed against future gains. In effect, the purchase price
of the new residence would be increased by the disallowed loss on the sale. See, Barbara
Kirchheimer, Archer Real Estate Measure Called Harmless Enough To Win Passage, 56 Tax
Notes 254 (July 20, 1992).

110. I am ignoring inflation and possible incentive depreciation in order to assume
that depreciation is intended to reflect actual decline in value.

11l. Under current law this imputed income is not subject to tax.

19921



Florida Tax Review

surplus in fact does not materialize to that extent?" 2

The converse of this argument could not be used to justify exclusion
of gain. For example, suppose the residence sells after one year for $105,000.
The owner cannot as easily argue that his true consumption was only $5,000
because if this is the expected value he attached to the residence, he would
not normally have purchased it. For him to have done so, he would have had
to have believed that the market was wrong. The owner who valued the use
of the home above expected cost would have purchased the home in any
event. Thus, while this point does not have overwhelming force, it offers
some justification for asymmetrical treatment of gains and losses which
appeals to me on other grounds.

F. Conclusion

In sum, the case for ex post as opposed to ex ante measurement does
not seem strong enough to warrant taking on the additional administrative
difficulty. There should be little difference in most cases especially if gains
on housing and consumer durables are properly accounted for. Moreover,
much of what appears to be losses actually reflects increased consumption or
preference for leisure. Further, the theoretical advantage of taking account of
actual differences in enjoyment seems to be considerably weakened to the
extent that unfavorable outcomes are the result of failure to insure or other
voluntary action. Finally, particularly because gains are hard to identify, any
attempt to measure outcomes may fall shorter of the goal than if the effort
were not made.

IV. TREATMENT OF RECOVERIES

The previous part of this paper has considered whether a taxpayer
who experiences less than expected enjoyment from an expenditure should
be entitled to a tax deduction for this "loss." We turn now to the treatment
of those luckier individuals who would have suffered such a decline in
enjoyment but for their ability to "recover" from someone else such as the
seller of the product, an insurance company or a tortfeasor.

At first glance, it may appear to be obvious that such recoveries
would be tax-exempt regardless of the approach favored for the treatment of
those who cannot recover for their losses. If the tax base were income or
power to consume, consumption would be measured ex ante by the amount
expended in the purchase of goods or services. Outcomes would be irrelevant.

112. Some may consider it relevant that the $10,000 of imputed income was not in-
cluded in income. See supra note 104. A similar argument could be made for not treating the
excess of the actual over the expected cost of any consumer purchase as a loss.

[Vol 1:1



Valuing Personal Consumption

Under this approach the value of insurance or a dealer warranty could be said
to be the price or premium charged, whether or not the taxpayer recovered
on the policy. The purchaser who insures has merely acquired a more
"reliable," and hence a more expensive product, no different than if she had
purchased a better quality model. In short, there is no "recovery."

Those who prefer a tax base equal to actual consumption would care
how matters turned out. But the purchase of insurance or other protection is,
ostensibly at least, to protect the consumer's expectancy. Unlike the unin-
sured, the insured experiences no difference between such expectancy and
actual enjoyment. For example, if property is damaged or destroyed, insur-
ance will replace it. Therefore, there is no reason to tax any recovery.

I believe, however, that while this conclusion might be at least
generally correct, the explanation is not quite that simple. As we have seen,
even under an ex ante approach, the consumer who transfers what might be
considered an item of consumption, such as an automobile, will be taxable
to the extent that the amount recovered exceeds her basis. Proceeds in excess
of basis may be taxable whether or not the seller's position is improved. For
example, an amount received in return for permission to violate an individ-
ual's right to privacy would be taxable even though such an individual could
be said to be no better off."3 In the case of an insurance or other recovery,
the recovery measured ex post will exceed the cost of the protection which
cost might appear to be the taxpayer's basis.

Ultimately of course, it might seem unfair to include the recovery in
income without allowing a deduction for the loss which the recovery reim-
burses. But this justification for the exclusion of recoveries appears
inconsistent with no deduction for uninsured losses. In short, how can we
justify similar tax treatment for two individuals who seem to be in different
positions in that one recovers for her loss and the other does not?

Therefore, while tax exemption for insurance or warranty recoveries
does seem appropriate under either an ex ante or ex post approach to
measuring consumption, such exemption is not easily explained. After con-
sidering a number of possible justifications, one can better understand the
possible limits of such an exemption.

A. Basis Recovery

1. In General.-Even voluntary transfers would not be subjected to
tax if the amount received did not exceed the taxpayer's basis. For example,
if an automobile which cost $50,000 were sold for $30,000, there would be
no tax on the transaction. As we have seen, this treatment is incorrect to the
extent that the owner's basis is not reduced to reflect depreciation over the

113. Kelman, supra note 17, at 842-43.

19921



Florida Tax Review

period of use. If, however, proper depreciation did not exceed $20,000, tax
exemption of the proceeds would be appropriate.

Some have noted a potential discontinuity if basis is taken into
account for the purpose of excluding the proceeds of sale, but not for the
purpose of allowing a loss. 14 In that connection, we have discussed pre-
viously whether a deduction for a loss would be appropriate if an automobile
were sold for less than its basis after properly adjusting for depreciation.,"5

However that question is determined, it is inappropriate to tax proceeds which
are not in excess of basis. Thus, for example, suppose the automobile pur-
chaser immediately changed her mind and was able to sell the new car for the
original $50,000 purchase price and buy a boat instead. It would seem to be
clearly wrong to include the $50,000 from the sale of the car in income. Total
consumption remains at $50,000, even though it is shifted from an automo-
bile to another form. As long as the taxpayer receives no more than her basis,
consumption does not increase and neither should taxable income.

In the case of insurance or seller warranty, however, the very purpose
of the transaction is to pay a small, fixed amount up front to protect oneself
against the possibility of a very large expenditure later on. While the taxpayer
may seem to be no better off when the insurer or seller makes good after a
"disaster" occurs than she would have been if all went well, an exclusion may
not be explainable on the grounds of basis recovery. The essential nature of
insurance is to pay a little for a large amount of protection.

2. Is Basis Essential?-Under one view, the fact that the individual
is no better off is sufficient without regard to whether there is "technically"
enough basis. Andrews notes that the exclusion of medical services provided
by a charity or by a government welfare program or by a tortfeasor has not
been considered to be a tax preference." 6 He suggests this is because "[tihe
taxpayer is no better off after the whole transaction than before he incurred
his injury and it would be unnatural to view the provision of medical services
in isolation from the injury as producing a taxable gain."....

There are difficulties with this approach, however. It is not consistent
with the treatment of taxpayers who are taxable on a sale of rights, such as
privacy, even though they would not appear to be better off."8 In addition,
as Andrews notes, it may follow from his premise that a taxpayer who pays
for his own treatment should be allowed a medical expense deduction to

114. See, Lawrence Zelenak, The Taxation of Indemnity Payments: Recovery of
Capital and the Contours of Gross Income, 46 Tax L. Rev. 381, 389 n.43 (1991).

115. See supra text accompanying notes 98-106.
116. Andrews, supra note 18, at 334.
117. Andrews, supra note 18, at 334.
118. Andrews's view would also suggest that wages would be exempt to the extent

leisure is lost.

[VOL 1:1



Valuing Personal Consumption

reflect the fact that he is worse off than either a taxpayer who does not get
sick or one who is reimbursed for his costs." 9 While a deduction is allowed
for medical expenses in excess of 7.5% of adjusted gross income, as we have
seen, losses are generally not allowed.

Some would avoid these difficulties by asserting that it is only when
the transaction is involuntary that it is unfair to tax an individual who ends
up no better off. Noting that we do tax a person who voluntarily markets his
privacy rights, even while we exclude tort recoveries for violation of
privacy,12 Kelman sees the distinction between voluntary and involuntary
action as "a political recognition of a basic human resistance to commoditi-
zation."12' According to Kelman, the tax system must confirm a person's
refusal to treat his privacy as a saleable object)'" It is the involuntariness
of the conversion that bars taxability.

Others are not sure. Griffith suggests that once a taxpayer's rights
have been violated, imposing a tax on the proceeds does not necessarily add
to the injury."2 If the injured party is worse off because he is forced to
substitute taxable dollars for an untaxed benefit, perhaps the recovery should
have been increased so as to make him whole.

Further, when the tax law specifically waives current tax on insurance
proceeds under section 1033, it conditions the exemption for the amount
received in excess of basis on reinvestment in similar property.'24 The pre-
sumed need for section 1033 may indicate that in the absence of basis, invol-
untariness does not necessarily make a recovery nontaxable, or at least that
exemption should be conditioned on reinvestment in similar property.

3. Does Basis Exist?-Assuming sufficient basis is a prerequisite to
exclusion of proceeds, to what extent can it be said that there is adequate
basis in the transactions we are considering? If instead of comparing the
insurance proceeds to the premium, we could compare the proceeds to the full
original cost of the insured purchase, as if the item were sold to the insurer,
the recovery, in some instances at least, may be said not to exceed basis.
Thus, an individual whose car was stolen or totally destroyed would be likely
to have a basis in the car at least equal to the amount recovered from
insurance or the tortfeasor who caused the damage.

Something analogous to basis recovery occurs when a taxpayer
merely gets her money refunded. Consider, for example, a resort which has

119. Andrews, supra note 18. at 334.
120. See IRC § 104(a)(2).
121. Kelman, supra note 17, at 842. Kelman rejects the notion that the non-taxability

of tort judgments supports the medical expense deduction.
122. Kelman, supra note 17, at 843.
123. See Griffith, supra note 24. at 381.
124. IRC § 1033(a)(2).

19921



Florida Tax Review

a policy that reimburses 80% of the cost of the hotel if it rains more than
50% of the time during a week's stay. A refund from the hotel would clearly
seem to be nontaxable, even if the refund technically might not amount to a
recovery of basis.

This approach could possibly extend to "refunds" from a third party.
For example, an insurance company might contract to "refund" the cost of a
vacation in the event of excessive bad weather. The Internal Revenue Service
has accepted the tax free status of tax indemnities from tax counsel who erred
in return preparation." After filing a joint return on the advice of their
accountant, Mr. and Mrs. Clark later learned that their tax liability would
have been almost $20,000 less had they filed separate returns. Their accoun-
tant admitted his error and reimbursed the Clarks for the additional liability.
The Internal Revenue Service acquiesced in the non-taxability of such a
recovery. 126

125. Clark v. Commissioner, 40 B.T.A. 333 (1939), acq. 57-1 C.B. 4; Rev. Rul. 57-
47, 1957-1 C.B. 23.

126. Id. Before it recently reversed its position, the Internal Revenue Service also
agreed to exemption for tax indemnity payments received from a transferor of assets who
guaranteed that the purchaser would be entitled to certain tax treatment. G.C.M. 39697 (Jan.
27, 1988) withdrawn, G.C.M. 39857 (Aug. 15, 1991). In one case, a contracting party failed
to live up to his promise that interest from the package of mortgages transferred would earn
a tax credit under IRC § 936. Priv. Let. Rul. 8748072 (Sept. 3, 1987). In another case, the
transferor under a safe-harbor lease had represented that all the property was "qualified lease
property" when in fact some of it was not. This led to a denial of an expected investment
credit and depreciation deduction. Priv. Let. Rul. 8923052 (Mar. 16, 1989). General Counsel
Memorandum 39697 appears to argue that if the taxpayer had paid the lower tax it legitimately
expected to pay, it would have been able to retain the excess cash without additional tax.
Therefore, the same result should follow when the cash is recovered from the contracting
party.

After first withdrawing the earlier ruling, Priv. Let. Rul. 9014404 (January 5, 1990),
the Internal Revenue Service has now reversed its position, Priv. Let. Rul. 9226032 (March
26, 1992) and Priv. Let. Rul. 9226033 (March 26, 1992). The Service reasoned that unlike
Clark, the taxpayers involved in the private rulings had paid the proper tax under the facts as
they existed. Amounts received from a third party to offset such tax payments are taxable
under the principle of Old Colony Trust Co. v. Commissioner, 279 U.S. 716 (1929).

The Internal Revenue Service's action has been criticized. William L. Raby, IRS
Change of Position May Up the Cost of Tax Malpractice, 56 Tax Notes 195 (July 13, 1992);
Lewis M. Horowitz, Excludability of Tax Indemnification Payments Threatened by Recent
Change in IRS Position: PLR 9014046, 49 Tax Notes 799 (Nov. 12, 1990). Professor Zelenak,
however, agrees with the Service. He points out that if the earlier position had been
maintained, a corporation might be able to represent that the interest on its bonds is tax-
exempt. It could then claim that the additional payments it must make when the representation
turns out to be false are non-taxable tax indemnity payments as opposed to taxable interest.
Zelenak, supra note 114, at 398-99.

It may be, however, that the issue is more complex. Suppose, if proper advice had
been given, adverse tax consequences could have been avoided without changing the

[Vol. 1:1



Valuing Personal Consniption

No matter how far the concept can be expanded, however, basis
recovery will not suffice as an explanation whenever an individual pays a
small amount for protection against a potentially large future cost. Examples
would include the value of repairs under a dealer warranty,'- a regular
airline ticket provided to a traveler unable to make the return flight by
charter,1 S and payment by an insurance company of a judgment entered

economics of the transaction, or in Priv. Let. Rul. 8748072 (September 3. 1987), other
mortgages were available that had not been purchased by taxpayers seeking benefits under IRC
§ 936.

127. The basis in the original part may not be sufficient particularly if depreciation
were required.

128. Assume a charter flight is available for S 1,000. If the traveler cancels, the ticket
price is non-refundable. If she is stranded overseas, and is unable to make the return
connection, a regular ticket will cost $800. Assume the traveler will value the trip at Sl,100.
Her calculation of whether it makes sense to purchase the charter ticket will be as follows:

Percent Chance Expected Cost Expected Value
Charter Both Ways 90% $ 900 S 990
Charter One Way 5% 90 55""
Trip Canceled 5% 50 0

$ 1,040" S 1,045""

If the expected cost is less than the expected value, the charter flight is a sensible
purchase as shown. But the risk-adverse traveler may be unwilling to make the purchase since
there is a 10% chance of a substantial loss. Even a risk neutral person would find the trip
inadvisable if the risk of a separate return increases to 6% which would increase the expected
cost to $1,048. She may, however, be able to purchase insurance for, say $90, which will
reimburse the cost of the charter flight if she cannot make the trip, or supply a return ticket
if she is unable to make the return flight. If the insurance is purchased, the calculus is as
follows:

Expected Cost Expected Value
95% x $1,090 = $1,035.50 95% x $1,100.00 = SI,045
5% x $90 = $ 4.50""

$1,040.00

Viewed from an ex ante perspective, the insurance is worth no more than $90 and
the value of the ticket and the insurance is worth no more than $1,090. What should be the
result if the insurance provides reimbursement of $1,000 or if it pays the cost of a return flight.
$800?

"The expected cost is $1,000 (the cost of the charter) + S40 (the expected
cost of the return flight for the "'stranded" traveler) (5% of $800) = $1,040.
In other words, the cost is $1,000 for charter whether canceled or not and
an additional $800 when there is a separate return, which has a 5% chance
of occurring.
.. The expected value is 95% of $1,100. assuming a separate return does
not indicate a reduction in value of the trip. This may not be true if, for
example, the traveler returns early because of illness.

19921



Florida Tax Review

against a driver held liable in an accident.
The recovery reimburses the insured against a cost that she would

otherwise have to incur. But to claim this is merely a refund of that outlay,
which in many cases will not actually occur, seems to stretch basis recovery
beyond recognizable limits. In fact, this suggestion seems more like an argu-
ment for recovery of a loss.

In sum, basis recovery cannot explain exclusion of all recoveries. In
fact, since the value of the recovery measured ex post clearly exceeds the
amount paid for the protection, basis recovery may not in general be a good
explanation for exemption for insurance, warranty or tort recoveries. Another
justification for non-taxation of insurance recoveries would be needed to deal
with situations where basis is insufficient.

B. There Is No Recovery

The question of whether a recovery is taxable could be avoided if we
can determine that there is in fact no recovery by the taxpayer. An individual
who insures or obtains a warranty purchases a different product, since she
now has a one hundred percent chance of full enjoyment for the price paid.
The difference in value of the two items is merely the cost of insurance. The
cost to the insurance company or the seller should be irrelevant.'29

For example, an automobile has an expected cost of operation which
includes the initial purchase price, the cost of gasoline and repairs, and
potential liability to injured parties. One presumably measures this expected
cost against the expected value in determining whether to make the purchase.
But there is a risk that the actual cost might be greater than expected.
Insurance (or a seller's warranty) merely guarantees that actual cost of
operation will be closer to expected cost. Thus, once the premium is paid for
insurance or a warranty, the purchaser is out of the picture. Her cost of
operation is fixed.

In a sense, this is just another way of explaining the ex ante
approach. Value is measured by the cost in a market transaction, here the sum
of purchase price and insurance. As described above, however, the ex ante
approach does not preclude taking account of gains on consumption items
whenever there is further market transaction, for example a sale of a
residence which would enable the taxpayer to substitute a different form of
consumption. In the situation we are now concerned with, there may be a

*** If ticket price is reimbursed when the traveler must cancel, expected
cost is $90 (insurance) + $950 (95% of $1,000 cost of ticket).
129. Kaplow, supra note 36, at 1500-01 n.68 (citing an earlier draft of this article).

Professor Kaplow compares an insurance company to a lessor that promises to make an asset
available regardless of whether a casualty occurs. Kaplow, supra note 36, at 1501 n.68.

[VOL 1:1



Valuing Personal Consumption

gain in the sense that the recovery exceeds the cost of the insurance or
warranty. Nevertheless, while the recovery does provide, in some sense at
least, a market transaction, it does not necessarily permit such consumption
to vary.

In the case of a seller warranty, there is no opportunity to alter the
form of consumption. A vacationer who can stay an extra, hopefully less
rainy, week at a resort which provides this option when there is excessive
rain during the initial week, or a car or television set owner exercising his
rights under a product warranty can in no way change the mode of consump-
tion. They are merely enjoying the original purchase, a working car or
television set or a week of good weather, at the agreed upon price, which
price, because of the seller's warranty, was higher than it otherwise would
have been. It is immaterial that the vendor's costs with respect to a particular
buyer may exceed the price paid.

In the case of payment of liability by an insurance company, one of
the costs of operating an automobile, the potential liability to injured parties,
has been "purchased" at a fixed price. That price has not changed nor can the
money be used for any other purpose. The insurance company pays the
injured party directly.

The analysis is more strained in the case of a charter flyer, who is
stranded in Europe because she is unable to make her return flight. Her
insurance may provide her with cash in lieu of a return ticket. Although she
receives cash, she is likely to have little discretion in how to use it.

In any event, this approach would not explain all situations. If, when
it rains, the vacationer gets her money back from the resort rather than an
additional week, she clearly has money which could be used for other
pursuits, as does the driver who recovers when an automobile is destroyed by
a collision. While in these cases there may be said to be basis equal to the
amount of the recovery, exemption cannot be explained on the grounds that
the consumption choice remains unchangeable.

In other situations the original form of consumption cannot be
maintained. An individual whose privacy is violated is reimbursed in dollars
rather than in restored privacy. In the event of a fire that makes a residence
unusable, the insurer provides cash to meet additional expenses incurred as
a result of being denied access to the residence. The insured becomes a
renter, not a home owner. Thus, once again we have an incomplete theory for
exemption of recoveries.

C. Deduction for Insurance Premiums

There may be another way to explain exclusion for certain recoveries.
It is arguable that if insurance or warranty proceeds are properly taxable, the
premium or cost should be deductible. Taking account of neither is simpler,

19921



Florida Tax Review

in part, because it alleviates the need for additional insurance to cover the tax
liability. 30 It is also revenue neutral if we can assume similar marginal
rates would apply to deduction of premiums and taxation of proceeds.

The premise is that if the value of insurance must be measured ex
post to produce gain to those that recover, the value would have to be
similarly measured for those that do not suffer a loss. If the value in the latter
case is zero, 131 there would be a loss. If this loss is deductible, the insured,
except for the impact of progressive rates, could, at the same net cost, provide
additional insurance to cover the tax liability on the recovery. For example,
assume an individual could purchase $100 of insurance for $1 to protect
against a potential $100 loss. If the proceeds were taxable, at a 28% marginal
rate, he would need $138.80 in insurance to have enough left after tax to
cover the loss, but this would still only cost $1 if the premium ($1.39) were
deductible. The net cost to a person who does not suffer the loss would be
$1 whether one adopted the premium ($1.39) deductible and proceeds taxable
or the premium ($1.00) nondeductible and the proceeds nontaxable approach.

The identity disappears if we take account of the insurance com-
pany's expenses, investment income' 32 and profit or loss on the transaction.
If, taking these factors into account, the payout exceeds the premiums
collected, clearly the revenue suffers if neither the premiums nor the proceeds
are taken into account in determining the insureds' taxable income. On the
other hand, it is much more likely that, in the long run, premiums will exceed
the proceeds in order to cover the insurance company's expenses and provide
a profit. When both gains and losses are excluded, this "loading charge" is
effectively not deducted by the policy holders.

This fact, as well as the effect of a change in tax rates, indicates that
there would not necessarily be an identity between the two approaches con-
sidered in this section. This makes it necessary to consider further which
approach is more appropriate.

To the extent progressive rates are intended to reflect greater ability
to pay, they would seem to be misapplied to loss recoveries which replace
other expenditures. Moreover, it would be burdensome if taxpayers had to
predict the applicable tax rate in order to determine the adequacy of insur-
ance. Professor Kaplow has also noted a potential for serious moral hazard

130. Kaplow, supra note 36, at 1501.
131. This assumption is questionable. See infra p. 37.
132. The discussion generally assumes insurance against a contingency which will

or will not occur instantaneously. If there is a delay, the insurer will invest the premiums and
earn investment income. Ideally this investment income should be taxed to the insured. See
generally Daniel I. Halperin, Interest in Disguise: Taxing the Time Value of Money, 95 Yale
L.J. 506 (1986). This matter will not be pursued further in this article.

[Vol 1:1



Valuing Personal Conslinption

if a drop in marginal rates caused a taxpayer to be over-insured. 33 This
may suggest that even if it were theoretically more correct to allow a
deduction for premiums and impose a tax on recoveries, it could be better to
do neither.

Moreover, the latter approach may in fact be more appropriate. As
noted above, exclusion of the proceeds is not equivalent to a deduction for
the premiums to the extent that the premiums exceed the proceeds. However,
since to the extent of the loading charge the insured pays more than the
expected value of the purchase in order to avoid the possibility of a large
loss, perhaps the value of insurance to those who do not suffer a loss can be
said to be their share of the insurance company's loading charge. This would
not be unreasonable even if ex post the risk related premium can be said to
be worthless." Achieving this result when a deduction for the premium is
allowed would entail the additional difficulty of separating out the nondeduct-
ible loading charge. For these reasons, exclusion of proceeds and nondeduct-
ibility of premiums would seem to be a better choice than inclusion of
proceeds and deductibility of premiums, assuming one or the other were
theoretically correct.

There are, however, reasons why deductibility of premiums might not
be generally accepted. As noted below in section D, the insurance premium
would reflect the difference between the expected value of the product and
the value if the event insured against does not occur. Because of the insur-
ance protection, in all circumstances, the insured achieves the value deter-
mined as if the insured event did not occur. Thus, the total amount spent, the
cost of the product plus insurance, equals the value received. Treating the
insurance premium as a loss, in a sense, ignores the excess of the total value
over the price of the product, just as treating insurance proceeds as income
ignores the decline between price and actual enjoyment when a loss occurs.

In any event, justifying non-taxation of recoveries on the grounds that
it compensates for nondeductibility of premiums would not apply to recov-
eries from a tortfeasor. It could not be argued that if premiums were deduct-
ible and proceeds taxable, the parties could adjust by increasing the size of
the policy. Here any adjustment, such as increasing the recovery to the victim
to compensate for taxes, could have real consequences. For these reasons, I
do not rely on this argument to justify exclusion for recoveries.

133. Kaplow, supra note 36, at 1493-1504 (insurance would exceed the sum of the

potential loss and the lower than expected taxes). Kaplow defines "moral hazard" as the
"tendency of individuals to take less effort to control costs when some or all of their costs are
borne by others, such as the insurer." Kaplow, supra note 36, at 1494.

134. William Vickrey, Insurance Under the Federal Income Tax. 52 Yale Li. 554.
557 (1943).

19921



Florida Tax Review

D. Recovery Offset by Loss

The difficulty of explaining exemption for recoveries could be
avoided if we acknowledge that the recovery itself might produce a gain.
Exemption still could be justified on the grounds that the recovery reimburses
the taxpayer for a corresponding nondeductible loss. If the loss were allowed,
the two transactions would be offsetting.

The loss in some cases would be the original purchase price that
failed to produce value. If an automobile suffers a casualty, the recovery on
the policy equals the cost of the destroyed automobile. The loss and gain
balance out. The amount reimbursed in the event of bad weather on vacation
would also equal the original cost which was "lost." In other situations the
loss would be the additional amount expended as a result of the injury as
opposed to the cost of the item which becomes unusable. For example,
consider the charter traveler who purchases a round-trip ticket for $1,000 and
cannot use the return flight.'35 The unused portion of the ticket cost $500,
while the "value" of the recovery would be the cost of a one-way ticket, say
$800. The loss would be $800, the additional "unexpected" expenditure which
added no value. In the case of a fire damaged house, if there is to be a
deductible loss, it would have to be the additional money spent on the hotel
and food over what would have been spent had the fire not occurred.

This justification for exemption of recoveries requires an explanation
as to why a deduction for losses would be denied to those who choose not
to insure. The following is suggested. As noted above, if an uninsured tax-
payer does not suffer any loss, the value of the actual outcome exceeds the
expected value. Unless such gains are accounted for, it would be inappropri-
ate to allow a deduction for uninsured losses. 3 6 Symmetrical treatment of
winners and losers requires that the uninsured be denied a loss deduction
even if it were appropriate to effectively allow a loss to those who are
insured.

To further explain this argument, return to the previous example
where we considered a $10,000 car which has a 1% chance of being a total
loss. In these circumstances, the insured taxpayer pays $101 for casualty
insurance, representing a $100 premium for the insurance to cover the car and
a $1 premium for insurance to cover an insurance policy on the replacement
car. If there is an accident resulting in a total loss he collects $10,101 which
would enable him to replace the car and purchase a new insurance policy.
The $10,000 gain on the insurance recovery offsets the $10,000 casualty loss.
If no accident occurs there is, perhaps, neither a gain nor a loss as the owner
paid $10,101 to guarantee that she will enjoy the full $10,101 value of the car

135. See supra note 128.
136. See supra notes 76-79 and accompanying text.

[Vol 1:1



Valuing Personal Consiunption

as opposed to paying $10,000 for a 99% chance she will do so. Put another
way, the $101 gain over the expected value of the car is offset by the $101
cost of the policy.

The uninsured driver who suffers a casualty has a $10,000 loss with
no offsetting recovery on the policy. The treatment of insureds (exemption
of recoveries) suggests that we should allow a deduction for this uninsured
loss. But allowing this loss could systematically undervalue consumption of
the uninsureds as a group. Recall that the ninety-nine uninsured taxpayers
who do not have an accident obtain $10,101 worth of value for $10,000. For
every uninsured $10,000 loss, there are ninety-nine $101 gains. If neither
gains nor losses were taken into account, uninsureds as a whole would report
the correct taxable income. Further, to the extent people who systematically
do not insure could be expected to have equal amounts of gains and losses,
they too would report the correct amount of taxable income.

Moreover, if the benefit to the lucky uninsureds were to go untaxed
while losses were deductible, the tax system would systematically favor self-
insurance. A taxpayer who is risk neutral would be indifferent between a 1%
chance of a nondeductible $100 loss and a certain outlay of SI for insurance.
On the other hand, if the loss were deductible and insurance premiums were
not, 37 the potential loss would be reduced while the cost of insurance
would remain one dollar. This could cause even the risk averse person, who
would normally purchase insurance, to self-insure. Therefore, it may be
logical, on the grounds of both equity and efficiency, to deny a deduction to
the uninsured even if we effectively allow a loss for an insured individual by
non-taxation of recoveries.

While this approach would not explain loss disallowance to an
individual who had no opportunity to insure, it should be recognized that
insurance-like protection against a difference between expectancy and out-
come could be achieved in a variety of forms." Individuals can obtain a
product warranty, lease from an owner who bears responsibility for losses or
repairs, purchase a less destructible model'39 or, in the extreme, forego
risky activities altogether. Consider, for example, what would have happened
to the Clarks if they negligently prepared their own return." When they
hired an accountant, part of the fee was for the assurance of accuracy or
malpractice liability for error. If they had prepared their own return they
might have made an error that was not reimbursable. If they did achieve the

137. Section 106 of Code exempts premiums on medical insurance paid by an
employer, effectively allowing a deduction for premiums.

138. Kaplow, supra note 36, at 1489-92.
139. With an abundance of consumer information available, it is well known that

automobiles manufactured by certain car manufacturers have better records of durability and
performance.

140. See supra note 125 and accompanying text.

19921



Florida Tax Review

best result, however, they would in some sense have a gain-payment of less
taxes than the average self-preparer. If this gain is not taxed, perhaps an un-
reimbursed loss should not be deductible.

In summary, it is possible to explain why those who forego an oppor-
tunity to insure should be denied the loss deduction which is effectively
allowed by exempting recoveries. Since insurance in some form is generally
available, general disallowance of unreimbursed losses may be justified.

E. Summary and Conclusion

A number of explanations have been offered to reconcile non-taxation
of recoveries by way of insurance or otherwise with nondeductibility of losses
suffered by the uninsured. These include the application of basis, the exis-
tence of an offsetting loss and the argument that there is in fact no "recovery"
for the consumer.

Consistent with my belief that the value of consumption should
generally be measured by the costs incurred in a market transaction, I am
comfortable with the view that the "recovery" can generally be "ignored" by
a taxpayer who is insured, protected by seller warranty, or reimbursed by a
tortfeasor. 14' She has originally purchased a different more reliable product
and her consumption is measured by the amount paid for that product, which
amount includes the cost of insurance. The cost and hence the "value" of
potential liability to victims, for example, is determined by the insurance
premium rather than the uncertainties of future events.

This approach fits most easily with situations in which the recovery
is not directly received by the taxpayer or at least must be used in connection
with the original activity. In such cases it can be said that, while there may
be a market transaction, the taxpayer does not have a new consumption
choice. Thus, a dealer warranty can be used only to fix the car. Payments
pursuant to liability insurance go directly to the victim. Blue Cross will only
reimburse actual costs. The traveler stranded by his charter flight ostensibly
has more choice if recovery is in cash, but nevertheless must return home.

This analysis will not cover all cases, however. The victim of a
collision need not replace his auto. Vacation insurance payable in cash could
be used for other purposes. Cash cannot restore violated privacy. Basis
recovery may explain exemption in some circumstances, but, perhaps because
the recovery would exceed the price or premium paid for protection, basis
would sometimes be insufficient.

Some would argue that, regardless of basis, an individual who is no
better off as a result of the transaction should not be subject to tax. While this

141. If reimbursed by an employer, the value of that protection should be taxable.
Contra Andrews, supra note 18, at 334 (suggesting employer reimbursement be tax-free).

[Vol 1:1



Valuing Personal Conswnmption

seems too broad, a less drastic version-the idea that all involuntary receipts
that leave the taxpayer no better off should be tax-free-has appeal.

In assessing this view, it is useful to consider how an individual
could be said not to be "better off' if her basis is less than the amount
recovered. For example, consider an individual who purchased a residence for
$100,000. The property increases in value to $150,000 before being destroyed
by fire. While insurance proceeds of $150,000 would leave the taxpayer no
better off, it is clear that, in the absence of reinvestment, the amount received
in excess of basis, $50,000, would be taxable.' 42 The fire becomes the occa-
sion to recognize the previously unrealized gain. "No better off" in this
context must mean, therefore, that even though basis is inadequate, there is
no "unrealized gain." How could that occur?

Two possibilities may be suggested. First, the recovery reimburses the
taxpayer against an expense, subsequent to the original purchase, which
provides no additional utility. Examples include a tort judgment from negli-
gent operation of an automobile, a return flight from Europe when the
traveler cannot for some reason use the previously booked charter and, per-
haps, the cost of a hotel room when fire makes the taxpayer's residence
unusable. Second, in the absence of injury, the individual had access to a
form of tax-free consumption; for example, privacy, use of one's body, or
imputed income from use of a residence or one's own services.

The first possibility seems easier to deal with. In most of these cases,
as noted above, the recovery does not enable the consumer to vary the form
of purchase and thus can reasonably be considered irrelevant. Even if this is
not the case, it seems unfair for a person who is no better off than he would
have been had he not been injured or suffered a loss to bear a larger tax
burden. Thus upon comparison to the uninjured, exemption seems required
so as not to tax the injured more heavily. One way to justify exemption is by
assuming that for any income there is an offsetting loss. Thus, any taxable
income from the reimbursement of the cost of a vacation, ruined by bad
weather, would be offset by a "loss" for the difference between the original
cost of the vacation and the actual benefit derived. This might suggest,
however, that the injured party who does not recover should have a deduct-
ible loss.' 43 Nevertheless, while it might be more appealing to allow a
deduction to the unreimbursed individual in order to treat all three situations
consistently, we have seen that there are reasons for denying a loss for those
who choose not to insure. Since insurance in some form is available with
respect to most purchases, the denial of deductions for unreimbursed losses
does not necessarily determine how "insured" losses should be treated. If this
is so, exemption for recoveries can more comfortably be justified by the

142. IRC § 1001(a).
143. Cf. Zelenak, supra note 114, at 388.

19921



Florida Tax Review

assumption that there is a corresponding loss which should be allowed to
those who are insured.

There would be cases, however, where a loss could not be found. For
example, to the extent recovery for physical injury, such as loss of a limb,
exceeds additional out-of-pocket costs, there would be no additional expendi-
ture for the injured person to deduct as a loss.

This analysis suggests that the most difficult cases involve the substi-
tution of money for tax-free benefits such as the use of one's body or the
imputed income from home ownership. It is at least difficult, if not impossi-
ble to say that there is no change in the original form of consumption;
there is clearly no basis to recover and in the case of one's body no offsetting
loss. The home ownership issue was considered in Millsap v. Commission-
er145 and McCabe v. Commissioner.146

Millsap owned his own home and appliances and enjoyed tax-free
imputed income 47 from their use and from his own services in cooking,
cleaning and doing laundry. The home became temporarily unusable because
of fire, and as a consequence Millsap's access to such imputed income was
interrupted. Millsap, however, was covered by insurance that reimbursed the
cost of temporary living quarters, as well as the additional expenses of
laundry and food.

It would appear that the additional costs incurred by Millsap after the
fire would in effect reflect the loss of the imputed return from his residence
and its contents, the absence of his input in cooking and cleaning, which
made food and laundry more expensive, and the depreciation on his residence
for the period of absence.'48

Despite the fact that Millsap seemed no better off, the Internal
Revenue Service concluded that the recovery was taxable to him. 49 It

144. Use of proceeds to rent living quarters could be considered similar enough to
home ownership so that there is no change in the original form of consumption. However,
money in lieu of privacy is clearly different.

145. 387 F.2d 420 (8th Cir. 1968).
146. 54 T.C. 1745 (1970).
147. Michael J. Graetz, Federal Income Taxation Principles and Policies, 152-54 (2d

ed. 1988).
148. The assumption is that the hotel must charge enough to cover its operating costs

for utilities, laundry and meals, depreciation of its assets and a return on its investment.
Millsap's operating costs in the absence of the fire would be similar except for the value of
services. Thus, his excess costs cover imputed income from property and services and
depreciation. If Millsap's residence continues to depreciate despite being unoccupied, he must
bear this cost plus his share of the hotel's depreciation expense, which indicates why this cost
should be reimbursed even though the amount he normally spends on utilities, for example,
need not be.

149. Millsap v. Commissioner, 46 T.C. 751 (1966), aff'd, 387 F.2d 420 (8th Cir.
1968). Accord McCabe v. Commissioner, 54 T.C. 1745 (1970); Arnold v. U.S., 289 F. Supp.

[VOL. 1:1



Valuing Personal Consumption

asserted that while the taxpayer could enjoy tax-free imputed income from the
use of his residence and his own services, the entire cost of rent, food and
laundry must be purchased with after-tax income. Moreover, if Millsap rented
the residence, cash received would be taxable even if used to rent another
home. Therefore, cash received from an insurance company was taxable
because, in a sense, it could be said to be effectively renting the property.'

In sum, what Millsap was seeking was continued tax-free treatment
for an amount equal to the imputed value of the goods and services he
enjoyed prior to the fire, even though the fire deprived him of his access to
this in-kind income and the insurance company replaced it with cash. The
Internal Revenue Service, on the other hand, viewed the tax-free benefits as
lost. In McCabe the court recognized the dilemma, but opted for inclusion of
insurance proceeds in the absence of basis. It specifically rejected an exclu-
sion put forth on the grounds that the proceeds were in fact a substitute for
a nontaxable type income.' I think the Tax Court got the question exactly
right in McCabe, namely, Whether one should be taxable when he has been
forced against his will to shift the form of consumption from a nontaxable
form to a taxable one?

This issue also arose in a private letter ruling5 2 in which the
Internal Revenue Service dealt with a tenant in a rent controlled building who
received $20,000 in exchange for giving up his apartment and all claims
against the owner for fraud and for scheming to end rent control by violating
rules against seeking non-tenant buyers. The Service held the $20,000 to be
taxable.

If the taxpayer were otherwise able to retain his rights to the apart-
ment, at least temporarily, the $20,000 compensates him for higher rents he
will have to pay elsewhere for similar quarters; he is not better off than he
would have been if he stayed in his apartment.' 3 The $20,000 payment
merely reimburses the extra cost of the more expensive apartment which
provides no additional utility. The tenant has, however, exchanged access to
an apartment at a below market rent, a form of nontaxable benefit, for cash

206 (E.D. N.Y. 1968). But see Conner v. U.S., 303 F. Supp. 1187 (S.D. Tex. 1969. aff'd in
part and rev'd in part, 439 F.2d 974 (5th Cir. 1971); Taylor v. U.S.. 28 AFTR 2d 6108 (N.D.
Ala. 1971).

150. If this analogy is appropriate, Millsap should be entitled to a depreciation
deduction, which would mean that his taxable income would equal the lost imputed income
from property and services.

151. McCabe, 54 T.C. at 1748.
152. Priv. Let. Rul. 8952030 (September 29. 1989).
153. Alternatively, the $20,000 might reflect a discount from market being provided

to existing tenants who buy these apartments. Under this view, the S20.000 is akin to the profit
he might have obtained by purchasing the apartment and reselling it as soon as he was free
to do so. This would seem to be taxable.

1992]



Florida Tax Review

which would ordinarily be taxable. If the transaction is voluntary, it would
seem clearly taxable, but the answer is more difficult if the tenant was in
someway coerced, perhaps by the landlord's fraudulent behavior. It is
appealing to permit the substitution to be tax-free when it occurs without the
individual exercising any choice. 1"4

Cash substituted for a limb or for the pleasure of being pain-free
seems to me to be properly nontaxable.155 The Millsap case seems harder
because while we all enjoy tax-free use of our bodies, renters and home-
owners are treated very differently by the tax law.'56 Should Millsap retain
the more favorable homeowner treatment just because his shift to rental status
was against his will? 157

Proper measurement of income would require that imputed income
from home ownership be included in the base. The exclusion must be
defended on administrative grounds or more likely because the public would
not accept the inclusion. Therefore, there is no theoretical way to determine
how far to extend exemption for imputed income. If public perception is the
key, I would guess insurance recoveries in the Millsap circumstances and
similar situations, such as to temporarily replace a stolen car, would be
exempt as indicated by the prompt enactment of section 123 to overrule
Millsap.

I find this result acceptable while at the same time I applaud the
Service's conclusion regarding the tenant who lost his rent controlled
apartment. The distinction may be that the tenant had a special advantage,
access to a rent controlled apartment. While we generally do not tax the value

154. This is similar to Kelman's view that the involuntary nature of the transaction
is controlling although I found Kelman's reasoning unpersuasive. See supra text accompanying
note 123.

155. Contra Griffith, supra note 24 at 374; Joseph M. Dodge, Taxes and Torts, 77
Cornell L. Rev. 143, 182 (1992). The utilitarian approach would justify exclusion of personal
injury recoveries only if the impact of personal injury increased the injured recipient's need
for income. "If the victim ... has no greater economic need as a result of the injury ...
payments should be subject to taxation." Griffith, supra note 24, at 374.

See also William A. Klein, Tax Effects of Nonpayment of Child Support, 45 Tax
Law Rev. 259, 271 (1990).

156. Imputed income from home ownership is not taxed while rent is not deductible.
Regs. § 1.262-1(b)(3). In one sense at least, tax exemption is more easily explainable in
Millsap on the grounds that the loss incurred from the expenditure on his hotel room (which
provided no additional utility) could offset the proceeds. Loss of privacy does not necessarily
increase one's costs.

157. In an analogous situation, Simons justifies present IRC § 119, which excludes
the value of lodging from income when an employee is required to reside on the business
premises for the "convenience of the employer," as providing benefits similar to home owner-
ship to those who, because of the nature of their work, are not free to buy their own home.
Simons, supra note 1, at 123-24.

[Vol. 1:1



Valuing Personal Consuiption

generated by below-market purchases, there seems to be no reason to contin-
ue favorable tax treatment when the special status is lost. Millsap, on the
other hand, is a member of a much larger group-homeowners. He merely
seeks to continue to be treated like all the rest. Still, if Millsap loses, he is
merely treated like all renters. Thus, I am less comfortable with tax-
exemption for Millsap than I am with tax exemption for payments for
physical injury or loss of privacy.'58 Ultimately, the conclusion may reflect
intuition rather than logic.

It was also suggested above that tax-free treatment for insurance
recoveries could be defended on the grounds that premiums should be deduct-
ible if recovery is taxable. The validity of this argument, which was
previously questioned, will be developed more fully in the next section.

V. INSURANCE AGAINST Loss OF INCOME

As a final matter, I want to briefly consider whether the discussion
thus far leads to any preliminary conclusions concerning taxation of insurance
that provides protection against loss of income as opposed to protection of an
item of consumption. In the case of an item of consumption, I have argued
that generally proceeds from such insurance should not be included in income
and the premium should be nondeductible. Does this conclusion apply to
insurance against loss of income as well?

Under current law this treatment-nondeductibility of premiums and
excludibility of recoveries-is accorded premiums on and proceeds of insur-
ance to protect against loss of income from "personal" risks of illness or
death.1 59 On the other hand, premiums on insurance to protect against
"business" risks are deductible160 and in that situation the proceeds are
included in income. Thus, a storekeeper who buys business interruption
insurance is taxable on the proceeds, but the premium is deductible. The
Internal Revenue Service would also allow a deductible business expense to
an employee who sought protection against loss of income caused by a busi-
ness risk-from on the job injury or unemployment caused by an industry
decline.16' But according to the Service, the cost of protection against a loss
of income from personal causes such as illness or death caused by a non-job
related injury or disease would not be deductible. 62

158. Replacement of lost earnings, of course, should be taxable.
159. IRC §§ 101, 104(a)(3), 262; Regs. § 1.262-1(b)(1l. The value of a limited

amount of group-term life insurance provided by an employer is excluded from income under
IRC § 79, which is equivalent to a deduction for the premium and an exclusion of proceeds.

160. IRC § 162.
161. G.C.M. 39016 (January 18, 1983). See Rev. Rul. 81-193, 1981-2 C.B. 52.
162. Rev. Rul. 81-193, supra note 161.

1992]



Florida Tax Review

This distinction has little practical importance, since where a
deduction is not allowed, the proceeds have been excludible and, as we have
seen, the two approaches can produce equivalent results. As described
above, 163 as long as brackets are the same, the insured can adjust to the
taxability of the benefit by the purchase of additional insurance. If the
premium is deductible, the net cost does not increase. Disallowance of a
deduction for premium payments and treating the proceeds as exempt is
merely a simpler way to achieve the same result as a deduction for the
premium and taxation of recoveries.

In fact, tax-free treatment of mortality gains on life insurance cannot
be justified on the grounds that the taxpayer is no better off than she would
be if the insured lived (if the insured lived, earnings would be taxable), or
that the gain on insurance would be offset by a deduction for lost earnings
(failure to earn cannot lead to a deduction). The only viable explanation for
the tax-free treatment of mortality gains is that the exemption is balanced by
a denial of a deduction for the premium, which would be allowed as a cost
incurred for the production of income if life insurance proceeds were tax-
able.' 64 Some, however, who argue that life insurance proceeds should be
taxable, view the premiums, not as a deductible cost of producing what
should be taxable income, but rather as a nondeductible personal expense to
achieve comfort and peace of mind. 65

This discussion suggests two questions for consideration. First, is it
possible that the purchase of insurance, which increases utility by reducing
risk, should be viewed as enlarging the tax base? Second, if the purchase of
insurance does not create any additional overall income, 66 is it more appro-
priate to tax the proceeds and deduct the premium or to exempt the proceeds
and treat the premium as nondeductible?

Turning to the second question, there is, I believe, a distinction
between life and disability insurance on the one hand and, for example,
collision coverage on the other. As we have seen the purchase of casualty
insurance increases the value expected from the automobile-it means full
value will be obtained in all circumstances, not just in the absence of a
casualty. 167 It seems clear, therefore, that a deduction for the insurance
premium would not correctly measure the enjoyment of an individual whose
car is undamaged. If the premium is to be deducted, there must be an off-
setting amount of income, the difference between the value produced by the

163. See supra text accompanying note 131.
164. Robert B. Harris, Comment, Compensation for Loss of Income and Its Taxation,

34 Nat'l Tax J. 135 (1981).
165. William D. Popkin, Taxing Personal Insurance: The Case of Tax Audit

Insurance, 4 Va. Tax Rev. 379, 403 (1985).
166. Warren, supra note 23, at 1087.
167. See supra note 129 and accompanying text.

[VOL 1:1



Valuing Personal Consumption

car (outcome) and the purchase price (expected value).
Life insurance presents a different picture. Since in this case the

excess of actual over expected income is subject to tax, there may be said to
be a double counting of income if the premium is not deducted. Assume a
taxpayer, who has a 99% chance of surviving through next year, will earn
$10,101 if she does. Thus, assuming she is risk neutral, the expected value
of her future income is $10,000. If $10,101 of life insurance is purchased at
a cost of $101, there is now a 100% chance of receiving this expected
value-salary or life insurance proceeds of $10,101 less premium of SI01.
However, if she lives what will be taxed is the actual amount she earns,
$10,101, not the expected value. Therefore, unless the premium is deductible,
the purchase of life insurance would increase her expected income. Further-
more, if the insured dies, there is no loss to offset the proceeds. The
insurance provides additional value. Therefore, in the case of life insurance,
deduction of the premium and taxation of proceeds is at least in theory more
appropriate than the opposite approach which was suggested for casualty
insurance.

As noted, however, a deduction for the premium means everyone will
be taxed at the expected value. Is this the proper income for an insured who
lives? Is true income just $10,000 since she must pay $101 to assure she will
receive the expected income in all circumstances or does such a person
actually earn $10,101?

We can examine the question by comparing two people who survive.
A is uninsured and has $10,101 to spend on food and drink. B is insured and,
after paying the premium, has only $10,000 to so spend. Should A pay more
tax than B, or does B have $101 worth of peace of mind or security?

It might appear that B has as much consumption as A. He merely
chooses to spend some of his resources on insurance which makes him feel
more secure. It therefore seems hard to argue that B, if he lives, has less than
A, if he lives. They both have $10,101. If B were to die, however, it seems
that the insurance should be valued at $10,101 not its ex ante value of $101.
After all, unlike A's family when A dies uninsured, B's heirs can consume
at least $10,000 worth of goods and services when B dies insured.

If these suggestions are followed the purchase of insurance would
increase overall income." If there is no insurance, there would be a total

168. Gambling raises a similar issue. Suppose C gambles Sl on a 100 to I shot
(assume no take for the house or the state). In the absence of the bet C has a 100% chance of
$1 equaling $1. With the bet C has a 1% chance of 100, which also equals 1.

Life insurance would be seeking to protect against risk---to change a 99% chance
of $100, into a certainty of $99. When one gambles he is moving from certainty to a risky
situation. In both cases, however, the expected value of resources does not change.

If gambling winnings are taxable while losses are not deductible, taxes nevertheless
increase. The justification for this result may be the loser, despite his loss, has SI worth of

1992J



Florida Tax Review

of $1,000,000 for each one hundred people, $10,101 for each of the ninety-
nine who live and zero for the one who dies. But if everyone is insured, total
income would be $1,010,100 ($10,101 not $10,000 for each).

Insurance, by reducing risk, does increase overall utility but if this
argument for nondeductibility of the premium and taxation of proceeds is
valid why would it not extend to business interruption insurance as well. In
the case of insurance against business risk, by allowing a deduction for
premiums we do not include the value of peace of mind in the tax base.

Perhaps one way to resolve this apparent dilemma is to recognize that
A and B are not identical. A is risk neutral. He views a 99% chance at
$10,101, and a 1% chance that he will earn nothing, as equivalent to a 100%
guarantee of $10,000. Therefore, he need not insure.

B being risk averse considers the former option less valuable than the
latter. To B the expected "value" of the former option is less than $10,000.
Therefore, when B guarantees that he will receive $10,000, the expected
"value" of his income increases, which would not be inconsistent with a
higher tax burden. However, what B has achieved through insurance (an
expected value of $10,000) is merely what A already has because of his
greater tolerance for risk. Therefore it is not necessarily clear that B should
pay a greater "expected" tax than A. This would occur if all Bs were taxable
at $10,101,169 while As would only be so taxed if they lived. Perhaps we
need not worry about discrimination between A and B, if most people are in
fact Bs. But if that were the case, it might not matter what tax base were
chosen for B.

Apart from equity considerations, taxation of insurance proceeds
without a deduction for premiums could be said to create a disincentive to
insure. A risk neutral individual would certainly not want to increase her
expected tax burden without an increase in the expected value of income.
However, in the case of B, the additional security from alleviating the risk
could offset the additional tax burden.

For example, assume the tax rate is 20%. If B insures, his pre-tax
cash income after insurance is $10,000 and his tax liability is $2,020.20 based
on $10,101 of income. In all circumstances B or his heirs will have $7,979.80
in cash after tax. If B did not insure, he would have $8,080.80 if he lives
(80% of $10,101). The expected value, given the 99% chance of survival, is
$8,000, which is greater than the amount available to the insured if the
premium is not deductible. However, if B is risk averse, he might not

consumption from the excitement of gambling. A consumption value to gambling seems
plausible whether or not the same can be said for life insurance.

169. Equivalent treatment in the case of casualty insurance would require not only
that premiums not be deducted, but that, in addition, an amount equal to the premium be
included in income.

[VOL 1:1



Valuing Personal Consumption

consider a 99% chance for $8,080.80 more valuable than $7,979.80. After
taking account of risk, he might prefer a lower expectancy. Thus, he might
still insure.

In the end, I believe I would continue to allow insurance proceeds to
be tax free as long as the premium is not deductible."'u Since the govern-
ment would not be placing a charge on diminution of risk, this approach
would be certain not to hinder a decision to insure. It would also avoid
difficult questions about deductibility of what has clearly always seemed to
be business expenses-such as expenditures for business interruption insur-
ance.

171

VI. SUMMARY

This paper has shown that measuring consumption by outlay is
generally supportable even if one believes that ideally outcomes should be
taken into account, and that non-taxation of recoveries which exceed basis
can be reconciled in most cases with nondeductibility of losses for the
uninsured.

If consumption is measured by outlays, insurance is generally just
another outlay, albeit for a different and more expensive product, and any
recovery under the policy can in most cases be thought of as merely a means
of preventing an increase in the costs of an activity, which activity has not
changed. Cash which is freely available for other purposes, particularly if it
replaces normally tax free benefits presents more difficulty.

If outcomes matter then non-taxation of recoveries can be justified by
asserting that there should be a deduction either for the premium or for the
loss that has been incurred. The latter rationale does not necessarily support

170. It was noted above that exclusion of proceeds combined with non-deductible
premiums is not the same as taxability and deduction if the insurance company increases its
premium to cover expenses and profit. See supra note 133 and accompanying text. For
example, assume in the above case that the premium is S102. If the proceeds are excluded
those who live are taxed on S10,101 whether they are insured or not. Those who die would
not be taxed in either event. Insurance does not change the expected tax liability. However,
if the proceeds are taxable and the premium of S102 is deductible, the tax base for those who
insure, both expected and actual would be reduced to S9,999. It seems better to assume that
income does not in fact decline when insurance is purchased. This would be achieved if the
loading charge were non-deductible, or more directly by excluding the proceeds and treating
the entire premium as non-deductible.

171. In fact, if business insurance is held to be non-deductible, would similar
reasoning cast doubt on the deduction for other items like security guards? Suppose a business
determines that in the absence of a robbery, it will earn S 100,000 per year. There is. however,
a 50% chance of a theft, which will reduce its income to S90,000. Suppose hiring a security
guard for $5,000 will prevent the robbery, so that actual and expected income will be S95,000.
Could it nevertheless be claimed that taxable income is S100.000?

19921



50 Florida Tax Review [Vol. 1:1

deductions for the uninsured. Denial of such deductions follows from the
failure, which has not generally been recognized, to tax gains when actual
costs are less than expected. Since gains and losses can be thought to ordi-
narily balance out, it is best to take neither into account.

Deduction of premiums on insurance to protect against loss on con-
sumer purchases would, however, not correctly measure income. It can only
be justified as a balance to another "error"-taxation of proceeds without
recognition of the offsetting loss. The issue of whether premiums on policies
which protect against loss of income should be deductible may be more
difficult. This article offers only a tentative position in favor of such deducti-
bility, assuming proceeds were to be taxed.

Thus, the tax system, properly, does not take account of Andy's
unfortunate experience with his Mercedes. On the other hand, if he could
recover from the dealer, the amount received should be tax-exempt.