Economic development as opportunity equalization

1
Economic development as opportunity
equalization
12 May 2006
John E. Roemer Yale University
2
The traditional view

• development measured by per capita income
• two possible justifications
• technology proxy
• welfare proxy
• technology view cannot be right imagine
  slaves operating a wonderful technology and
  dumping the product into the ocean

3
The welfare justification

• simplest justification is
• and
• This is utilitarianism plus the assumption that
  welfare is linear in income

4
Even if one accepts the utilitarian premise, the
assumption that welfare is linear in income
denies that some desires or needs are more urgent
than others E.g., suppose, reflecting this
recognition, one chooses Then Clearly this
version puts supreme importance on avoiding zero
incomes, unlike the first approach.
5
Conceptions of development
6
The capability approach ( as captured by HDI)
does not depart from utilitarianism it takes as
a measure of individual utility a sum of income,
education, and health measures and then measures
economic development as the average of these
values across all individuals. So the innovation
by UNDP is not anti-utilitarian, but rather
inclusiveness w.r.t. other traits than
income (Perhaps surprising given A. Sens
involvement in HDI)
7
In contrast, the opportunity measure of
development I will propose here is
anti-utilitarian, but is not inclusive of other
personal characteristics than income. It will,
finally, take income as the measure of a persons
welfare
8
Key Idea of EOp Leveling the playing field
How well an individual does should reflect
his/her effort, not his circumstances,
where Effort comprises those actions a person
can control and Circumstances comprise those
influences he/she cannot control (social
biological environment)
9
The language

• objective the kind of advantage being sought
• circumstances those things that influence
  objective and are beyond a persons control
• effort .... influences within control
• type set of people with the same circumstances
• policy intervention to affect objective outcomes

10
The EOp policy is...

• that policy that renders as equal as possible,
  the distributions of the objective across all
  types
• in other words ones chances for acquiring
  degrees of the objective are independent of ones
  type (i.e, circumstances)
• level the playing field the troughs are
  disadvantage due to poor circumstances

11
Demarcation between circumstances and effort
Consider the problem of equalizing opportunities
for wage-earning capacity, using school finance
as the policy Circumstances could include
parental social class, parental education, race,
gender, IQ, region of country. The more
influences we choose to call circumstances the
more egalitarian will the EOp policy be. The most
conservative view everything is effort, nothing
is circumstance. The most radical view
everything is circumstance, nothing is effort
12
If everything is effort, then the EOp policy is
laissez-faire plus, perhaps, protection against
discrimination If everything is circumstance,
and all people are different, then the EOp policy
is equality of outcomes (each person would be a
type)
13
EOp vs. Meritocracy
Meritocracy generally means rewards according to
talents, even if talents are in-born or socially
derived This contrasts with EOp, in which rewards
are only justifiable in proportion to effort
unlike meritocracy, the source of effort matters
14
What the EOP for income policy is

• Goal to make the distributions of the incomes,
  across types, as close together as possible
• No attempt to decrease the variance within
  distributions this is due to effort, by
  definition
• In many problems, this amounts to maximizing the
  average income of the type that is most
  disadvantaged

15
Example CDFs of income, 3 types of US men
US, 1992 type is level of mothers education.
the value of the EOp program is the area to the
left of the red CDF
16
Contrast this with Denmark

• same three worker types as in the US case

17
and with Spain

• same three worker types, early 1990s

18
According to the measure of econ development I
will propose, Spain is the least developed of
these three countries, and Denmark is the most
developed.
19
An approximate approach
One often does not possess (for DCs) data to
construct these CDFs. A substitute is to
choose that policy that maximizes the minimum
average income across types. Let ? be a space of
feasible policies let be the average
income of type t at policy ?. Then the problem
is I.L.M., the rate of econ development is the
rate of growth of this index
20

• This approach is only interesting if it gives a
  different assessment from traditional growth rate
  of gdp per capita.

21

• On the other hand, there are examples where the
  rate of econ development would be greater using
  the EOp approach, although the level would be
  lower

22
WDR 2006 Equity Development
Argues for the EOp approach to measuring
development, and presents much data types are
defined by educational level (unfortunately not
of parents), sex, urban vs. rural outcomes are
income, infant mortality, nutrition, and life
expectancy. Most important document thus far
arguing for the EOp approach to econ development
23
Equity vs. growth
Equity, here, means equality of opportunity
equity is achieved to the degree that
is maximized. What does growth mean?
Growth of gdp/capita. Why be interested in
this? Only if one is a utilitarian of the first
kind. Thus, given the SWF
growth should be construed as growth of
where
24
Therefore the formulation equity vs. growth
is an incoherent one. It presumes that, on the
one hand, one has a concept of justice (equity),
and on the other , one is not interested in the
degree to which a society approaches justice over
time, but rather with the growth rate of some
other measure of welfare/justice. (For we have
already ruled out an interest in technological
measures per se.)
25
Equity vs. efficiency
Efficiency means one of three things for an
economist technological efficiency Pareto
efficiency social efficiency The maximization
of any SWF will entail Pareto efficiency and
hence also technological efficiency. Hence, if
equity is to be traded off against efficiency,
then the effic in question must be social
efficiency
26
But equity is social efficiency both mean the
maximization of ones chosen SWF So what do
people mean when they say theres a trade-off
between equity and efficiency? They mean between
equity and utilitarian efficiency -- i.e.,
maximization of average income. Again, this is a
misformulation. If one has a concept of equity
which is construed as maximization of a SWF then
there can be no such trade-off.
27
An exception.
There is a possible exception, but it requires
that our concept of equity NOT be construed as
maximization of a SWF. E.g. Suppose by equity
we mean equality of incomes. This conflicts with
efficiency, in the sense that equality may be
Pareto dominated by another allocation (i.e.,
maximin incomes). Some philosophers do take
equality as the desideratum (L.
Temkin) Guaranteeing rights might be another
exception
28
WDR 2006 makes this error
Greater equity is thus doubly good for poverty
reduction through potential beneficial effects
on aggregate long-run development and through
greater opportunities for poorer groups within
any society (p.2) If the opportunities faced
by children like N. are so much more limited than
those faced by children like P. or S., and if
this hurts development progress in the aggregate,
then public action has a legitimate role in
seeking to broaden opportunities.(p.3) Third,
the dichotomy between policies for growth and
policies specifically aimed at equity is false
(p.10)
29
These quotations from WDR 2006 show the
pervasiveness of the utilitarian assumption in
economic thinking. Despite its advocacy of EOP
as a development measure, the WDR repeatedly
falls back upon an implicit utilitarian
conception of development. Either that, or
opportunismon which more below.
30
The coincidence of EOp and gdp/capita optimal
policies??
The last quote maintains that the best way to
maximize gdp/capita is to maximize equity.
This is almost surely false. The simultaneous
optimization of two SWFs occurs only as a
singularity. The statement is opportunistic. It
is perhaps true that equalizing opportunities
will increase gdp/capita, but not that doing so
will maximize the rate of growth of gdp/capita
31
Neoliberal opportunism
Thus the equity-with-growth school is guilty of
opportunism. But so is the neoliberal school,
which maintains that unregulated markets are the
optimal policy for improving the welfare of the
poor. Exactly the same critique applies That
would be a singularity. Neoliberal policies are
perhaps the policies which maximize the welfare
of the entrepreneurial class it would be
singular were they to maximize the welfare of the
poor or the workers.
32
How important is inequality of opportunity?

• What fraction of inequality in a country is due
  to inequality of opportunity?
• there is a method by which one can decompose the
  CV of a distribution into that part which is due
  to inequality between the types and that part
  which is due to intra-type inequality.
• Doing this for the OECD countries, and
  partitioning each countrys workers into three
  types according to parental education gives the
  following

33
Surprisingly, perhaps, very little inequality is
due to inequality of Opp, so defined. Of course,
one would like to do this with a larger set of
circumstances than just parental education
(method from Roemer (2006), Decomposing
inequality)
34
China

• However, the situation in poor countries is
  worse. In China, there is intense discussion of
  regional inequality. The China Human Development
  Report 2005 presents income data for 31 regions
  of China.
• I calculate that 35.6 of income inequality in
  China is due to the circumstance of regional
  residence.
• This allows us to calibrate what really bad
  inequality of opportunity is When it accounts
  for one-third or more of total inequality

35
Brazil
WDR 2006 calculates that 30 of income inequality
in Brazil is due to inequality of opportunity,
where type is defined by the persons education.
Of course, this is not a could way of defining
type. One should use parents education.
Evidently, this is the best the data permitted.
36
Summary
1. Equality of opportunity is a pervasive if not
ubiquitous view of justice today. If you ask a
person whether inequality between A and B is
justified, she is likely to answer, It depends
on how hard they each tried. This, in a
nutshell, is the EOP view inequalities are
justified if they arise from differential effort
and only in such as case. Policy should level
the playing field to compensate those with poor
circumstances.
37

• The measure gdp/capita flows from two
  assumptions
• utilitarianism linearity of
  utility in income
• If one replaces utilitarianism with maximization
  of the EOp objective but takes income as the
  objective then the measure of development should
  be .
• A more radical proposal would be to equalize
  opps for welfare where we take
  . This would entail the social objective
• where

38

• equity vs. growth and equity vs. efficiency
  are phrases which pre-suppose two different
  social ethics. These presuppositions are almost
  always implicit, never clarified by the author.
• A coherent way of compromising between the ethics
  of equal-opportunity and utilitarianism is to
  choose as the SWF
• Here, we do define types. If ?1, we have
  utilitarianism. If ?-?, we have EOp.

39
Political economy
Even if one advocates EOP, one might say growth
of gdp/cap is necessary for political
feasibility of policy. This is an imprecise
statement. To be simplistic, what is necessary
is that the policy improve the welfare of some
given fraction, p, of the popn.
(Simplistically, p0.5 in democracies.) But
then the right approach is to maximize ones SWF
subj. to the constraint that at least fraction p
benefit. I.e., political feasibility does not
ipso facto require that one be a utilitarian.
(end)