This paper is prepared for staff
use and is not for publication.
The views expressed are those of
the author and not necessarily
those of the Bank.
INTERNATIONAL BANK MOR RE0ONSTRUCTION AND DEVELOPMENT
INTERNATIONAL DEVELOPIENT ASSOCIATION
Economics Department Working Paper No. 47
Shadow Prices for Project Evaluation in Less Develop2ed Countries
June 20, 1969
This essay on Shadow Prices for Project Evaluation
in Less Developed Countries was written, at our invitation,
by Professor Richard S. Weckstein, of Brandeis University,
as part of a larger study of resource utilization problems
in developing countries on which he is currently engaged.
It has been inspired in part by the recent Manual of Indus-
trial Project Analysis in Developing Countries (Volume II)
prepared by Ian M. D. Little and James A. Mirrlees for the
Development Centre of the OECD. The author wishes to ack-
nowledge the helpful and stimulating discussion of many
difficult and controversial issues arising in the use of
shadaw prices in project analysis during a seminar on this
subject held at the World Bank in April 1969 and during a
subsequent stay in Oxford. He is particularly grateful to
Ian Little for the patience and generosity extended to him
as a critic.  He alone, however, is responsible for the views
expressed here.
Sector ard Projects Studies Division
Prepared by:  Richard S. Weckstein (Consultant)



TABLE OF CONTENTS
Page No.
I.   INTRODUCTION .     .  .  .  .  .   . . . . . . . . . . . . . . .  1
II.   PRICE DISTORTIONS IN LESS DEVELOPED COUNTRIES.........            2
III.   A CRITIQUE OF SOME SHA.     PRIE: MEASURES. ........ . .          6
1. Measures of Shadow Prices   .......           . .   . . .  .  . .  7
Shadow Prices from Linear Programning    ........ .         7
Shadow Prices from International Trade Prices    .   .      8
2. Shadow Prices and Persistent Economic Distortiorm   .    .     9
3. The Effects of Inappropriate Shadow Prices . .... . . .       10
Shadow Prices and International Trade  . ....... .         10
Shadow Wiages, Employment and Investment  . . . . e        11
IV.   THE USE OF RELEVANT SHADOW PRICES            ...............     13
1. Shadow Prices for Fragmented Markets . . . ........           14
2. Exaxxples of Choosing Shadow Prices . ............            16
Excise Taxes .............       . .  .  .........         16
Import Dutit s  .........      .  . .   .  .  .  .  .  .  .   . .  17
Private Mono   y.    .    . ..........  ..... . . . . . . . . .  18
Land Values .........   .   . .    .............           19
Public Monopoly   . ...........             . . . . .      20
Wage Rates . . . ...................... 20
Fragnented Grain Markets........        .   . .. . . . .   22
Interest Rates................ ......... ..........        23
3.  Concluding Remarks...................      . . . . . . . .   25



SHADOW PRICES FOR PROJECT EVALUATION IN LESS DEVELOPED OOUNTRIES
I. INTRODUCTION
1.        Shadow prices are synthetic substitutes for the imperfect natural
price products of market economies. They are manufactured to be used in mak-
ing project choices by governments interested in improving the performance of
their economies. Technically they work by giving decision makers more valid
information than natural prices give. But these synthetic substitutes can be
manufactured in a number of different ways and the quality and usefulness of
the products is very sensitive to the process used. In this paper some
fundamentally different processes are examined from the point of view of the
usefulness of the products to consumer governments. The first step in this
evaluation is to diagnose in Part II the weakness of the prices that are
produced naturally by less developed market economies. The next step is to
rate some of the available synthetic product lines, the shadow prices; this
is done in Part III. Finally in Paxt IV there is a brief and general guide
to the use of synthetic prices which focuses on how to avoid the worst pitfalls
in using synthetics.
2.        All economies use prices to convey information about the social
valuation of goods and resources. In some economies prices are cast up by
efficiently organized markets and as a result these prices are an unbiased
measure of' social values. But this may occur only in nythical economies in
which a perfect price system operates. The price systems of developed Western
societies are more or less satisfactory approximations to this ideal, and
highly decentralized decisions concatenate into a socially acceptable whole.
In the nature of underdevelopment, however, the quality of market prices for
the purpose of guiding otherwise uncoordinated decisions is marred by both
systematic biases and random influences with large dispersions. To achieve
tolerable results in the over-all total of decentralized decisions in less
developed countries seems to require either that the price system generate more
valid resource and goods prices, or failing that, a reliable procedure be
designed to provide high quality substitute prices usable at least to guide
governments' decisions and perhaps also those private decisions that have
far-reaching and important effects.
3.        The necessity to use shadow prices in less developed countries is
compound; it arises from the presence of the imperfections well known in
developed market economies plus some additional imperfections that are
important in less developed countries, even if they are not unknown in more
developed market economies. A list of imperfections of developed economies
which are ordinarily responsible for project evaluation errors and which
could in principle be rendered less destructive by the use of shadow prices
includes the following: the presence of unemployed labor anticipated to
persist through much of the life of new investment projects where the market
wage of labor exaggerates the real scarcity of labor; the influence of mono-
poly power over market price, driving prices above or below the competitive
equilibrium values; the existence of a non-optimal income distribution and
hence a non-optimal distribution of purchasing power which in turn affects
the structure of market prices. This particular market imperfection reminds



-2-
us of the fact that the critical evaluation of the prices of a market system
inplies a value premise. In this case it implies the specification of a
"good" distribution of income, and there are similar income distribution
premises implied in judgments applied to other ki.nds of imperfections. Fur-
theriore, the replacement of biased prices with shadow prices implies still
other value premises.
4.        To continue with the list of cormmon price imperfections, there is
a broad categozy of prices that are imperfect because they arise from markets
in which contracts are two-party agreements between buyer and seller and omit
the interests of third parties either as beneficiaries or as sharers in the
burden of the cost. We also include the pricing of "public goods"l, for which
the technical nature of the goods precludes their being priced efficiently
without at the same time altering the distribution of income. In addition,
the influence of government taxation (and su-bsidization) biases prices, even
though it may be the objective in employing such financial means simply to
raise general revenue or to alter the distribution of income.
5.        Despite the fact that these influences which render imperfect the
information conveyed by prices are common in developed economies, no national
administration of industrial investment decisions in advanced countries
attempts systematically to compensate for them by the use of shadow prices.
Where it is felt that such imperfections produce undesirable results, govern-
ments rely on other instruments of control ranging from incentive taxes and
subsidies to direct licensing or prohibitions and to undertaking the desired
activities by government agencies themselves. The economically more advanced
countries may well have abstained from the use of shadow prices to correct
the distortions arising from the above market imperfeutions for reasons having
to do with the specific types of imperfections and the general inconclusiveness
of arguments favoring the use of shadow prices. But the basis for urging the
use of shadow prices on less developed countries is that there are more far-
reaching distortions in at least some critically important prices which are
especially likely to plague the price systems of such countries. The view one
takes of how correctly to establish shadow prices depe-nds on one's understand-
ing of these special weaknesses of the pxice systems of less developed countries.
For this reason it is important to specify in some detail how less developed
market systems generate prices that carry spurious information.
II. PRICE DISTORTIONS IN LESS DEVELOPED COUNTRIES
6.        In contrast with the pr-ice distortions of advanced countries which
are doubtful candidates for general correction by shadow prices, the distortions
of less advanced countries 1/ are likely to be gross and enduring. Rapid
inflation, which is a chronic problem of no advanced country is widespread
among the less advanced. Its influence on prices is most importantly con-
centrated on the international value of the country's own currency because of
1,/ The folowing cryptic listing is borrowed from a more adequate discussion
in I. M. D. Little and J. A. Mirrlees, Manual of Industrial Project Anaaly-
sis in Developing Oountries, Vol. II, Social Cost Benefit Analysis, Organi-
zation for Economic Cooperation and Development, Development Centre, Paris,
1969.



-3-
the modern convention of maintaining '?fixed, exchange rates. The combination
of rapid inflation in the less advanced countries and stable prices in advanced
countries generally requires spasmodic adjustmnent of exchange rates that
bridge the two sets of prices. During the period preceding depreciation the
official currency value is increasingly an exaggeration of the international
purchasing power of the currencies of the less developed countries. After de-
preciation, depending on how far the price change goes, the official currency
value may exaggerate, correctly state or understate the international purchas-
ing power of the currency. Because the importance of imports is ordinarily
so great to the small specialized economies of the less developed countries,
this price distortion permeates most decisions and the consequences are exceed-
ingly destructive to such countries' well-being. The opportunity for using
shadow prices to rectify this kind of distortion is apparently attractive.
7.        Similarly, there may be a need to apply shadow prices to the wages
of labor, in couniries in which there is an excess supply of labor in compari-
son with the means for its productive employment. For, in less developed
countries, perhaps quite generally, the wages actually paid employees in in-
dustrial establishments may be greater than the value of the marginal product
of labor at full employment. Under these circumstances the temptation is great
to use a shadow wage, free of the constraints on market wages.
8.        The rate of interest it is appropriate to charge for the use of
capital or to discount future net returms to compute the present value of
those earnings, is generally not the market interest rate which particular
projects must actually bear, for the capital markets are so poorly developed
and little integrated with one another that these rates can hardly be accepted
as measures of the social value of the alternative employments of capital.
More specifically to the general circumstances of less developed countries, it
is found that the rates of interest on capital flowing to industrial projects
from official sources is far below the rates paid on non-industrial projects
which are institutionally less favored. The favored rates are furthermore
likely to be established without adequate allowance for the common anticipa-
tions of inflat;ion that do exert an influence on other rates. These biases
convey by price distortions to capital users a harmful inducement to the over-
use of capital 1/ which leaves to agencies such as development banks the task
of rationing capital. But these agencies and also finance ministries and plan-
ning offices must also rely on a discount rate to make interproject comparisons.
9.        Finally the problem of price distortions associated with monopoly
power in the generally smaller economies of less developed countries may be
more serious than in advanced countries and hence warrant corrected prices
for investment decisions. And some economies that export specialized colmmdi-
ties are in a position to influence the price of their exports in their favor
1/ Low rates of interest on capital flowing throigh official channels can be
thought of as a shadow interest rate, whicih being below the market rate,
provides a basis on which government agencies can justify more capital-
intensive projects. See footnote on p. 14.



-4-
which would frequently justify a national profit maximizing output policy that
can be implemented by the application of shadow prices for these exports to
control unprofitable expansion of national capacity.
10.       It is probably true that this list of market failures would be justi-
fication enough for most economists of an effort to implement a shadow price
system. Yet the impression this and similar lists gives of the problem of
achieving investment choice efficiency in less developed economies is somewhat
over-simplified. It suggests that the market defects can be easily categorized,
from which it is assumed to follow that the corrections required fall similarly
in categories. The facts are, I believe, rather less convenient. The market
systems of less developed countries fail in quite a general way as a system to
integrate the dispersion of activities.  Rather than being unified market econo-
mies, these economies are broken into fragments by numerous barriers and by the
failure, during the early state of their development, of markets to have spread
across the barriers. They are economies made up of many non-competing parts
which are m.isrepresented by general. equilibrium models.
11.       Within such fractured economies the significance of actual prices
will in many cases be highly ambi.guous. In functioning market systems we ex-
pect to be able to intexpret the meaning, say, of the price of an acre of land
as a measure of the social sacrifice imposed, its social cost, when that land
is used to grow some particular crop. Where the price is either not the result
of a market process, or where the market process is arbitrarily circumscribed
by legal restrictions or requirements, the substitution relations implicitly
required for such social cost judgments probably do not exist.
12.       In thf* arket recruitment of inputs for a new project, the ordinary
assumption that permits us to accept market prices as the measure of social
cost is that the order in which resources are diverted from other activities
is at least roughly efficient. That is, the activity from which the resource
is withdrawn receives the least-valued contribution of all users of the resource.
In general, resource diversion requires highly decentralized decisions even in
countries where new projects are chosen by centralized agencies. For this
reason where the assumption of efficient resource diversion is unjustified,
market prices are unacceptable measures of social cost. In a general way all
that can be said about the biases carried by market prices is that if resource
diversion is inefficient, market prices will understate sccial cost. For, if
prices measure the value of resources at the margin of their productive employ-
ment and they are withdrawn fmm the more-valued uses, and not from the least-
valued uses, the losses will be both greater than they might have been and
greater than market prices indicate.l/
1/ To argue that activities do not come within the scope of the price
system, or that they are not integrated with one another through con-
tinuous markets, is not to say that the individuals of less developed
economies are any less "economically" motivated than are individuals
in advanced countries. There are reasons to expect that people generally
are rational in the sense that they maximize some utility function by so
choosirB among the alternatives available to them. NJevertheless, in
less developed countries the marginal rates of substitution among indi-
viduals, and the marginal rates of transformation among producers are
far less perfectly brought into equality than in advanced market economies.



12.       The less developed a country is, the less likely it is that market
price distortions can be removed within any reasonably limited period of
time. For advanced countries most of the oommon price distortions have in
fact been brought to tolerable levels. Unenployment is a small fraction of
the labor force of almost all advanced countries. Monopoly is effectively
countered so that its price distortions are rarely gross, and especially is
this the case when the dynamic influence of innovation on market competition
is taken into account; monopoly destro;4tig innovation is the process par
excellence of economic development. The distribution of income may not con-
form to the ideal any one of us holds, but the distribution is subject to the
effective social control of the polity and it must be unclear how we should
alter the distribution by some new means when in fact we have not altered it
by the already available means.
13.       Even in the most difficult class of value distortions, where im-
portant externalities are present, the operation of a flexible market system
gives continuous opportunity for the internalization of externalities by
changes in the size of business finrs to incorporate more or fewer activities
within a single administration. Those externalities that cannot be thus in-
ternalized are in fact a major basis for the establishment of government
intervention. The scope of government is freely changeable in response to
these shifting needs, and externalities which appear to be unattended may be
a minor subset of the set of all externalities, The main application of
shadow prices in advanced economies is in the evaluation of projects which are
to be undertaken by government. And the argument for the use of shadow prices
there is that by their use such evaluations may be made more satisfactorily
than by a political bargaining method of project choice that is simply devoid
of systematic evaluation. The use of shadow prices in advanced countries is
ironically a step in the direction of increasing reliance on markets. The
arguments for shadow prices do not apply to the evaluation of industrial pro-
jects undertaken within the private sector.
14.       Similar considerations might suggest the linLitations of shadow
prices in less developed countries as a means to deal with the conventional
list of market failures. While the conclusion would be reached with many
more reservations, still it would be plausible to hold that the best policy
is to eliminate the failures rather than to accommodate to them by a compen-
satory shadow price system. To illustrate the way such a case might be pre-
sented, consider the main points of failure in the OECD list. The distortions
introduced by inflation, especially those that spoil the significance of the
exchange rate, can certainly be removed by the termination of inflation, and
they can be mitigated to some tolerable level by a suitable reduction of the
rate of inflation. Whatever the merits of the structuralist case for per-
mitting continuing inflation, we need not conclude that inflation is inherent
in the state of being less developed. All we can conclude is that inflation
may be less costly than the cost of its control, and the fact is that many
less developed countries choose inflation while advanced cowutries do not.
15.       Unermployment comes much closer to being an inherent aspect of being
less developed, at least in countries with large populations in comparison
with the endowment of resources. Still the bias in the market wage in the
modern sector where wages are paid to recruit labor need not necessarily be



6-
distorting. If the market wage is set at the level of per capita income
in the non-market sector., there is a sense in which that wage does measure
the social cost of employment in the modemn sector. Furthermore, if wages
are held constant at that level there is an open opportunity for the expansion
of the modern sector by transferring labor resources from the nond-market
sector. Wage distortion becomes a particularly serious problem when the
industrial wage rises above this level as a result of perhaps government
policy or union power. Then it does become difficult to develop by the in-
creasing use of the available labor resources, for the high wage is an induce-
ment to excessive economies in the employment of labor and it imposes a high
cost on expanding employment in the fomn of increases in current consumption
of a minority of the population.  Yet there are two things to be said about
a situation of inflated industrial wages of this kind. The first is that
industrial wages are in principle subject to control, even in less advanced
countries. The second is that our judgment of the cost imposed is a matter
that implies a value premise with respect to the inter-temporal and the inter-
personal distribution of consumption. If the structure of wages in a less
developed country were established knowingly through some political process,
that result may then convey a kind of legitimacy to the consequences that we
cannot lightly ignore.
16.       A final distortion of the market system commonly found in less
developed countries, that which is associated with export monopoly power,
does not require less devaloped countries to use a shadow price for exports.
For a 'tax can be le-vied on these exports equal to the difference between the
world price and the marginal revenue to the country, with the desired effect
on production.
17.       Griven the market failures of less developed countries, we can
choose either to concentrate on the elimination of the distortions, or to
offset their influence on the way the econony grows by using a system of
shadow prices in project evaluation, or, of course, some comabination of the
two. Those who choose the second, seem to give little weight to the likeli-
hood of economic reform of a kind that would improve the operation of markets
significantly, and they have conmaratively greater confidence in the useful-
ness of special pricing policies in the evaluation of new projects. In
opposition to this view we ought to recognize that economic development and
the requisite reform go together, perhaps inseparably in the absence of a
dominantx centralized economic administration. But the point is not only that
there is a strong relation between improved markets and development, it is
also tha,t in the absence of improved markets there are serious limitations on
the improvements in rational choice it is possible to achieve by the use of
shadow prices. These limitations will become evident as we discuss the sources
and appLications of shadow prices.
III. A CRITIQUE OF SOME SHADOW PRICE MEASURES
18.       The market weaknesses of less developed countries cannot be quickly
changed and there is certainly every need to accommodate to them in the sense
that major investment decisions be made as rationally as possible under these



circumstances. To further that objective we ought to seek prices that convey
more infonaation and less lnoise" than some of those that are generated by
undeveloped and imperfect markets. There are a nwuber of distinct alternative
methods of generating shadow prices fmm which to choose. The line that
divides the main alternative methods is drawn on the basis of how much it is
expected can be achieved in the making of efficient project decisionis by the use
of shadow prices. The temptation is to find unbiased measures of social cost
and gain in prices that approximate the prices that would obtain under condi-
tions of general equilibrium in perfect markets.
1. Measures of Shadow Prices
a       s ro =Linear Pror
19.       In fact the original use of the term "shadow pricest" comes from
an analytic abstraction of a natural market process, linear programming, that
generates resource-scarcity weights at the same time as it finds an efficient
allocation of those resources to satisfy a given objective function. Linear-
programming planning models may be constructed for actual economies, and from
these shadow prices can .be generated. This possibility gives substance to the
temptation to find once and for all the very prices that tell us the correct
information about social scarcity, no matter what actual prices may be, and no
matter what the actual activities of the economy are.
20.       But the temptation to use such shadow prices is a delusion. If in
fact the results of such a programming model were to become the directives for
the actual activities of the economy then, strictly speaking, prices would not
be necessary, for the sector allocations would all have been made by the pro-
gramming simulation and the accompanying shadow prices would merely record
the consequent comparative scarcities. If the results of the planning model
were interpreted to be merely guidelines to the activities within sector
aggregates, and specific projects still had to be selected, the shadow prices
generated out of the model would be inapplicable to project evaluation for
they would be scarcity measures of resource and activity aggregates where
projects must be evaluated in terms of individual resources and goods values.
Furthermore, there are severe limitations in suchi mocls as actual planning
guides relating to the assumption of linearity, the crudity and datedness of
the statistical information, the excessive simplicity of the objective function
and the arbitrary time horizon during which the objectives are optimized.
Even if shadow prices generated from such models could be used for project
evaluation they would be far from unbiased.
21.       More inportant than all these objections, however, is the fact that
no less developed country (to say nothing of advanced countries) is in fact
organized and directed in accordance with such a planning model even in
countries in which such models have been constructed. The critical question
is, what would the relevance be of shadow prices derived from an optimal
planning model to an economy which is not in practice governed in accordance
with the model? Is it efficient to employ resources in a project, or to
choose new projects on the basis of social costs measured under conditions
of "optimally efficient" allocation which is nonetheless fictional? We
require second-best pricing here ratder than optimal pricing. For in general



-8-
in a non-optimal world we may fail to achieve an improvement in the total
system by adopting locally a policy that would be optimal if applied
universally. Furthermore, the information "optimal' prices convey is likely
to be entirely erroneous in a gt;nerally non-optimal situation. Actual re-
course scarcities, those that result from the way in wiich resources are
being used, are sinply different from scarcities under conditions of efficient
allocation, and they will not be revealed in the solution to an optimizing
planning model.
Shadow Prices from International Trade Prices
22.       General equilibrium programming methods are neither correct nor a
practical source of shadow prices. An alternative source of optimal prices
that is however also feasible is the method proposed in the OECD Manual ...
In this proposal morld commodity prices are used as prox prices for project
evaluation.  For goods that are actually traded internationally, either im-
ported or exported, there is little difficulty in identifying the relevant
world prices, for they are either received or paid by the country (tariffs
and most export taxes are disregarded). For non-traded goods and resources,
the goods are separated into traded and non-traded components, the traded
component is valued at world prices, and the non-traded component is again
separated and proceed again as with the first separation, eventually summing
the world value of traded parts.   The value of any residual non-traded part
is ascertained from its contribution to some traded good. An exception is
made for the value of labor, which is discussed below (page 11 ff.). The
rationale supporting this method of valuing resources and goods is that the
opportunity cost of a resource in any use whatever is, at the margin, the con-
tribution it makes, or might have made via foreign trade to the balance of
payments.  In this approach to shadow pricing of resources, the foreign trade
sector is not limited to its influence on the exchange rate. Foreign trade
is the window through which all internal values are compared with those in the
rest of the world.
23.       It is an ingenious method, both for its rationale and its feasibi-
lity of administration.  Yet it is a system of valuation, like the programm-
ing model source of shadow prices, that is based on the questionable assump-
tion that individual decisions should always be guided by optimal prices.
It assumes that the meaning of cost is foregone opportunity under the ideal
circumstances in which any resource could indeed be used to produce for
export or substitute for imports. It further assumes that the actual trans-
formations facing all parts of the economy have been brought into marginal
equilibrium with world rarkets. And it assumes that each individual in the
country has brought his substitution schedule into equality at the margin
with the structure of world prices. Wherever the Manual ... acknowledges
that these counterfactual assumptions are unrealistic, it justifies the pro-
cedure with the argument that better policies would bring about these
circumstances.
24.       Even in economies with considerable resource mobility and the
capacity to accomplish transformations rapidly, disequ.ilibria persist for
long periods, and indeed create the incentives which induce the reallocation
of nesources to new activities that eventually eliminate the disequilibria.



-9-
A pricing policy that anticipates this adjustment process 'by the establish-
ment of the utlimate equilibrium prices by fiat goes far toward unhinging
the entire adjustment mechanism. Ideally marginal cost-price equality should
be approached as efficient activity levels are reached, the remaining gaps
serving as continuing guides in the feedback control mechanism. This market
related mechanism may work poorly or not at all in less developed economies,
but it may hinder the establishment of such mechanisms to make project choice
irmnune to continuing local inflluences by sanctioning the use of prices that
remain insensitive to local conditions.
25.       The technique proposed in the OECD Manual ... is a little like the
method reportedly used to establish trading values among the planned economies
of Eastern Europe. For very different reasons;, there prices are also unreliable
measures of resource and goods value, and whatever use may be made of arbitrary
official prices within each country, they are unacceptable as a basis for ex-
change outside each country. Confronted with ignorance of true internal cost
and value, trade negotiators adopt (capitalist) world prices for the purpose
of es-6tablishing exchange values. Perhaps East European trade is facilitated
thereby. But an important distinction needs to be made. The inefficiencies
of those economies and the ignorance in which decentralized decisions must be
made cannot be rectified by the fiat of using world prices. The prices of the
outside world do not measure the marginal rates of production transforation
or consumption substitution within the Eastern European coiuntries. In a
partial equilibrium sense the export industries and import conpetitive indus-
tries may, in part, adjust to outside prices. But in the nature of those
economies, and in the nature of the economies of less developed countries,
there is not the further indirect adjustment to frontier activities. The OECD
method would bring world prices inside the economies of countries adopting it,
to the extent project investments are, based on them, which is a step toward
a kind of integration with the world economy not actually taken in Eastern
Europe where trade, but not production or investment, are apparently affected
by trading prices. It is regrettable that the price problemjs of the Eastern
European countries cannot be so easily solved. They could have knowledge of
their own scarcities anad values without the experience of markets to reveal
their values. But it cannot be done.
2. Shadow Prices and Persistent, Economic Distortions
26.       Misunderstanding about this may be traceable to a difference of
view about non-optimal economic activity in less developed countries. There
does seem to be some evidence in the Manual ... of a kind of optimism about
market failures, that it is limited, accidental and temporary. In an example
in which a tire factory inivestment decision would be viewed favorably if rubber
inputs are valued at world prices and unfavorably if domestic prices are used
(p.17), it is argued that world prices are the correct basis for decision. The
argument against using the higher domestic price is that one error should not
be permitted to impose additional errors on the econonry. Now if it is to be
acknowledged that the high-cost rubber indutry is an error, does that mean
the supply of domestic rubber to the tire industry will not be expanded when
required by further increases in the capacity of the tire industry? If it does
not imply this, if rubber production is not to be limited because it is an error,



10-
then the true social cost of rubber is measured by the higher domestic price.l/
If, on the other hand, no such expansion need be expected, and new supplies
of rubber would be imported at world prices, then there is no need to continue
to protect the domestic price of rubber and the correct policy would permit
the actual domestic price to fall to meet import competition. A shadow price
would be unnecessary. But failing such an actual price policy, the case for
using world prices depends upon the correctness of the second assumption., and
it cannot be assumed correct automatically. Furthermore, as a general surmise,
it seems rather much to assume that errors are to be acknowledged and rectified
in future decisions. There is much experience to suggest that this is not
always the way it goes. The circumstances that lead to such "errorst" persist.
If there were a reliable tendency for developing countries to return to optimal
patterns of production, there would be very much less need for shadow prices in
the first place. The role of shadow prices is precisely to guide the adjustment
to persistent non-optimal situations. This is the second-best proposition again.
27.       Inefficient forms of pn:duction are not entirely the product of past
error, as in the case of the rubber industry, above. IMuch of the inefficiency
of less developed countries is the consequence of the state of underdevelopment
itself.  It is a prior condition with which we start and it is more or less
ubiquitous. It affects the way all resources are used and the way transforma-
ticns from one kind of production activity to another takes place. It is
gratuitous to assume that better policies will in some miraculous way presently
rectify the terms on which actual decisions are made.
28.       In addition some of the ttinefficient"t productive activities of less
developed countries as well as of advanced countries, are matters of social
policy and, being the result of choice and not error, are indeed likely to be
highly durable. It is not even clear that it is right to recommend that such
policies be corrected. For, many investment choices of the past were made
in response to a social distribution preference to benefit a region, a jgroup
or an individual, even at the expense of national income which is a summation
of unweighted individual incomes. Correctly conceived shadow prices should
incorporate a measure of actual domestic preferences. Clearly the prices of
the outside world cannot be expected to measure these inside values where
there has not been an opportunity for free adjustment of activities to the
markets of the world.
3. The Effects of Inaropriate Shadow Prices
Shadow Prices and International Trade
29.       The choice of which goods are to be imported and which exported, a
critical choice if a country is to benefit from international trade, is sensi-
tive to the way costs are presented for investment decision, that is, the system
1/ How costly an error is this? The value of the loss will depend on the
extra consumption of rubber (or tires) induced by using the lower world
price of rubber. Given the elasticity of the demand for tires, e, the loss
would be L     qi q, p[p/p (e + 1) - 1], where q and p are the given
doimstic values of consumption and price, and A p is the difference between
the domestic price and the world price of rubber. If the tires are also
exported, the loss is f pX + the domestic loss.



of resource pricing. The fundamental basis for efficient international
specialization is the difference in the primary resource scarcities of one
country in comparison with the rest of the world. The presence of a com-
paratively plentiful supply of one type of resource, which is indicated by
a relatively low price for it, induces the embodiment of that resource in
goods which can be exported profitably into the world market. Frequently
the primary resource is only indirectly embodied in an exportable good, by
being first embodied in a non-exportable intermediate good the price of which
is in turn low because a) its cost of production is low and b) it cannot be
exported to the world market. But when this is the route by which a compara-
tively plentiful factor influences the choice of exports, the low domestic
cost of these intermediate goods is the critical price required to generate
a profitable supply of exportable goods.
30.       Now what is an exportable and what is not is determined by apparent
comparative cost, by transportability, and by specific market opportunities.
For example, the local price of sugar may remain below the "world price" be-
cause sugar is marketed under restrictive marketing conditions. That which is
exported is sold at a different price from the local price and the two prices
do not necessarily converge. If, under these circumstances, sugar is valued
at world prices when it is employed to produce, say, rum, there may be a
costly error. For the use of the high world price of sugar in rum production
may well indicate that it is unprofitable to export rum. In effect, to use a
world price for an ambiguously exportable good imposes, by fiat, an equaliza-
tion of the price of a comparatively plentiful local factor with the price of
the factor in theworldmarket. Through this kind of error a country may be
discouraged from entering into world trade on the basis of its fundamental
resource strengths.
31.       Symmetric errors are possible that would encourage a country to
export goods for which its resource endowment is inappropriate by obscuring
the locally high cost of production of intermediate goods by the substitution
of lower world prices for local prices. Similarly, imports can be perversely
selected by shadow prices that make domestic productions appear more expensive
than imports. Where there are complex and widespread market imperfections
that open up these possibilities the potentiality for social loss is quite
real and great care must be taken, as the Manual ... recognizes, to avoid the
mechanical use of such general rules as using world prices to value "tradable
goods."
32.       These observations about world trade pitfalls where shadow prices
are used should not be interpreted to mean that ava-ilable market prices will
reliably avoid similar trade perversities. They will not. There are no doubt
numerous examples of welfare losses occurring under the guidance of market
prices, frequently assisted, no doubt by policies misguided in still other ways.
However, shadow prices in less developed countries should be sensitive to the
market and price  anomalies that commonly occur.
Shadow Wages, Employment and Investment
33.       In the proposal for the valuation of labor the OECD Manual ...
attempts to incorporate the valu,as of the economic authorities. The scheme



12 -
for the valuation of labor is based on the assumption that the wage paid in
industrial employment is higher than the value of the marginal product of
labor in traditional agriculture from which expanding industrial employment
is supplied. But the lower marginal product is not the measure of the social
cost of expanding industrial employment because of the assumption about the
way national authorities regard the distribution of present consumption and
present investment (future consumption). It is assumed in the Manual
that the authorities may in some cases prefer to increase future consumption
at the expense of present consumption., that they may prefer more rapid econo-
mic growth than that which is chosen by individual decisions to save and invest
through the market, and that they may lack the instruments to bring the actual
ratio of investment to incoma into equality with their desired ratio. But if
a government is unable to prevent a net rise in consumption when a worker is
transferred from traditional agriculture to industry, increasing industrial
employment raises the ratio of consunption to income even while it increases
income. Hence the Manual ... includes in the shadow wage a part of the added
consumption as the social cost of industrial emnployment.
34.       Reservations may be held about the assumption of governments'
inability to tax to limit consumption increases, but given the assumption,
the social cost of industrial employment is higher than the direct loss of
output in traditional agriculture and the7ormulae provide a means for estimat-
ing the correct social cost accordingly. This is a proposal for finding a
shadow price of labor that accommodates to local circumstances, to the
authorities' values, to the structLral discrepancy between two sectors and to
the fact of limited tax power. It is a means to establish a second-best
shadow price rather than one built on the counter-factual assumption of general
equilibrium.
35.       There is, however, an inconsistency in the argument that arises in
connection with the undertaking of selected projects whose social profitability
exceeds private profitability on account of a shadow wage below the market
wage. Such projects will incur, for at least a part of their labor force, an
actual expense in excess of the market value of the labor product, regardless
of how profitable the project may appear when valued by shadow prices. This
deficit, insofar as it is not offset by contributions to total profit from
internal project sources, must be made up by subsidies for which there must
be tax counterparts. But the shadow wage is based on the assumption that the
government's capacity to tax is practically restricted and hence a desired
rate of saving and investment is not achieved. Pmjects are to be valued, in
such circumstances, by the use of an inducement wage rate, but this neglects
the necessity to match the investment this evaluation method would justify with
added required taxation. It is clear, once the finance side of the argument
is recognized, that the extent to which investment may be encouraged by means
of shadow wages must be related to the availability of tax revenue and to the
value of that revenue in other uses, including the financing of still other
investments rather than financing the extra consumption of workers employed at
inducement wages. The state of the government's budget is given no place in
the calculation of the shadow wage rate, although the general social value of
investment and future consumption are taken into account in the OECD M4anucal ...
(Chapters XIII, XIV), a problem ignored in many other pleas for using shadow
wages in place of market wages.



- 13 -
36.       Setting shadow wages below the market wage would be even less rele-
vant in an economy in which the addition of an industrial job in one industry
is accompanied by some contraction of industrial employment somewhere else in
the market sector. This would be unfortunate where a large pool of low-
productivity workers continues to exist in the traditional sector, but the
actual cost of using industrial labor in these circumstances should be re-
flected in their wage.
37.       Such local knowledge cannot be had in advance of direct experience
and specific studies made within less developed countries. These studies
should focus on resource recruitment and product use associated with projects
of the past to discover how the particular economy has actually performed.
Resource diversion patterns are not generally planned and are subject only to
indirect control. How the controls work must be observed. There may be stable
patterns for each country so that once they have been studied for recent his-
torical cases the patterns may be made the basis for establishing resource
values in shadow prices. We may, of course, be unlucky and regular patterns
may not be revealed so that it wi-l be difficult to anticipate how and from
whence resources will be released in future expansions. We may also discover
that some countries are much like others, all others or only others that fall
into some more or less homogeneous category. With so little knowledge of this
kind it is impossible to know what to expect.
38.       However much work this may requixe, the usefulness of shadow prices
lies in the information of just this kind encapsulated in them. If we are to
employ shadow prices in project evaluation we must acquire the knowledge on
which to establish them. There is no need to set prohibitively rigorous
standards of accuracy. Much infonaation of this kind would not be worth the
cost of obtaining it. Many specific resources are too uninportant for any in-
vestment project to warrant much study. Many inputs are actually imported and
it is only necessary to agree on a shadow price of foreign exchange to value
them. It may even be appropriate to conduct benefit-cost studies to determine
what information is worth learning. But the first task is to acquire experience
in conducting resource diversion studies and to learn about the general patterns
of resource diversion.
IV. THE USE OF REEVANT SHADOW PRICES
39.       Although relevant shadow prices require specific knowledge of the
economies in which they are to be applied, they share some of the problems of
identifying them, which can usefully be explored in a general way. The most
common circumstance  that suggests the need to substitute a shadow price for a
market price is, as discussed before, the fracturing of markets so that there
are two (or more) prices from which to choose. Markets may be fractured by
institutional divisions, by the use of taxes, subsidies, import duties, quotas,
rationing, the influence of monopoly or monopsony, and price discrimination.
To a private finn, for example, the price of an inported good includes the duty.
But a government user of such a good may ask whether it should regard tbhe cost
of using a unit of the imported good as the c.i.f. price including the duty or,
in consideration of the fact that the duty contributes to the government's own



-14 -
revenue, as simply the c.i.f. price, net of duty?  The problem is only slightly
more complex if there is also a domestic source of the imported good. A simi-
lar problem is raised in choosing a social discount rate for the evaluation of
a project by the government. Should it be guided by the cost of capital re-
sources which may be the price of borrowing either at home or abroad? Or
should it impose upon its own projects a higher rate closer to the rate applied
by private firms which pay a business profit tax? Where direct controls limit
imports (or other activities) the market is fractured into a market clearing
price that is above the supply price. Should project appraisal use the common
market price or is social cost more truly measured by the supply price? These
problems are ubiquitous, and we should recall that each of the policies or
conditions that has the effect of fracturing markets inevitably distorts an
allocation mechanism.
40.       As a general caveat, when taxes., subsidies, direct controls, or other
protective and revenue intruments are considered, the cost of distortion of
resource allocation ought to be weighed with the advantages and other disadvan-
tages. Many of the objectives of such policies can be achieved with minimal
distortion to the price system. Much price system distortion and destruction
is the outcome of policy choices made in ignorance of the cost of destructive-
ness to the price system, and once markets have been fractured there is no way
to regain all the benefits of a unified price system, even by the judicious
application of shadow prices in project choice.
1. Shadow Prices for Fragmented Markets
41.       The principle of efficient allocation that bears on these problems,
is the Principle of Equal Valtation (PEV).l/ An econony can achieve net gains
by transferring resources from low-valued uses to higher-valued uses. Where
there are two (or more) prices of a single resource it will, in consequence,
be employed in production with sinLlarly different marginal contributions,
thus preventing the efficient use of the xesource in the econony as a whole.
The PEV requires therefore that all classes of users of each resource be charged
the same price.
42.       This is an important principle and. its application will prevent a
good deal of mischief, 2/ especially the kind, of costly mischief that is
possible once an economic administration is released from the inhibitions of
1/ In the absence of some criterion which would have general applicability
based possibly on a desirable intersectoral allocation ratio, or the
utility function of govemments making project choices, the Paretian
standard of efficiency is adopted here.
2/  The separation of interests between the public and an official decision
maker is analyzed in Louis de Alessi, "Some Implications of Property
Rights Structures for Investment Choices Within the Govemment," American
Economic Review LIX, March 1969, pp. 13-24. Insofar as shadow prices are
witin the diseretionary choice set of an official rather than objective
data given to him, the choice may be made for the maximization of his
welfare function rather than thi social welfare function.



market price by the seductiveness of being able to substitute shadow prices.
For example, a country with a high-cost, protected rubber industry may find
it unprofitable to establish a rubber tire industry unless the project is
evaluated with a shadow price for rubber at the c.i.f. price of imported
rubber net of duty. That xould violate the PEV, and in fact it would be a
costly policy if a rubber tire industry were to involve an expansion of the
high-cost donestic rubber source. If it were not to be supplied by more home
production, but by imports, then it should be asked if rubber tim production
is the best use to be made of greater rubber inports? If the PEV is followed,
the question will be asked.
143.      Now on the other hand prices can be biased in their representation
of the actual terms on which resources are available., whether or not such
terms themselves are the result of optimal conditions. Moreover, generally
such biased prices are coupled with detailed constraints on the adjastments
that ate made in the form of rationing or the use of other direct controls
that partially sterilize the influence of price. In these circumstances we
can imppmve project evaluation by the use of a shadow price that is based on
an estimate of social cost and benefit. In this kind of case there is nothing
lost in applying a shadow price to the evaluation of a particular project
because the rest of the econonxy has not adjusted to the nominal market price,
and in fact equalization of cost and use values is more likely to be achieved
by shadow prices applied to specific projects.
44.       Frequently the problem which presents itself here is the existence
of more than one value of the resource within the econom;y because there are
barriers to equalization.  The influence of a tax, monopoly and rationing are
all similar in preserving differences in the value of a resource to its user
and the cost at the margin to its supplier. Wherever such multiple values arise
the correct value with which to calculate social cost is the one (or more than
one) that applies to the part of the market from which the sacrifice is made,
or to which the benefit goes. Where it is feasible to do so, the shadow
price should be a weighted average of the prices actually paid and rec,sived
where the weights reflect actual quantities. For example, an expansion of
electrical generating capacity may draw on fuel oil from new drillimg within
the country, by contracting the flow of oil to rail transport where coal is
competitive and by increasing imports. The cost of oil from each of these
sources may be different because of the use of taxes, tariffs and administra-
tive a1locations. The differing values sacrificed in the three sources are
all relevant, although for practical reasons it may be undesirable to estimate
each. But no general rule can tell us which simplifying assumption is best
in all cases.
45.       The variety of imperfect and fractured markets is large and it will
clarify the applicationL of principle to shadow pricing to choose illustrative
cases from the main types. In these examples we follow a procedure that can
usefully be adopted by the project analyst, consisting of four broad steps to
be initiated once it is determined that the narket price is not useful for
project evaluation and must be replaced by a shadow price. The following
examples, however, will include some market prices that can be used for project
evaluation, and the point of the discussion will be to distinguish those prices
that are usable from those that are not. The first step, a), is to ascertain



- 16
whether or not the market imperfection responsible for the unacceptable price
is subject to refom, on the general ground that it is better to change the
market price for all users than to use a substitute price for a single
decision. The evraluation of projects may be expected to provoke useful dis-
cussions of market policy if project analysts are alert to the opportunity.
The question, moreover, is strictlJy relevant to project evaluation, for the
expectation that market reform may occur in the future ought to be taken into
account in the present evaluation.
46.       The second step, b), is applicable in cases in which the unusable
price is not the result of a market imperfection but of some government policy
that affects the price. Before adopting a shadow price it should be ascertained
whether or not the policy can be, or is likely in the future to be, changed.
If both inquiries yield negative results, then we nrust find a way of acconmnodat-
ing to not having a single optimum price applicable over the entire economy.
We must choose a sector of the econoimy., step c), within mich to achieve an
efficient allocation. As a rule the object in making such a choice should be
to choose a sector within which there is actual resource mobility, to accept
the natural divisions that already exist in an economy that is imperfectly
integrated. Such divisions may be based on geographic regions among which
labor is relatively immobile, or governmental ministries among which capital
budget allotments are determined according to political preference, etc.
That decision made, the final step, d), is to estimate and adopt a shadow price
that satisfies the PEV within that subsector of the economy and encourages sub-
optimizing allocation within that sector.
2. E2amples of Choosin Shadow Prices
Excise Taxes
47.       A simple kind of market imperfection is that which is introduced by
the imposition of an excise tax on an intermediate good. The effect is to
destroy the marginal equivalences between producer and user on whom the tax
falls. The project analyst skips step a) of the procedure and surveys the
possibilities in step b) by reference to the canons of good taxation by which
the influence of the tax may be rendered less harmful to efficient resource
use. However, assuming opportunities for market improvement ar'e not present,
step c) indicates that a choice must be made to honor the marainal equivalences
either vertically or horizontally. Suppose we are dealing with the price of
machinery used as inputs in the project being evaluated, and produced in domes-
tic firms. Preservation of the vertical equivalence means to equate the cost
of producing the machinery and its productivity at the margin, while horizontal
equivalence means the equation of the marginal productivity of machinery in all
uses. Inequality vertically means an error in the total quantity of machinery
produced and used, while permitting efficiency in the allocation of the
available supply among alternative uses. The reverse error results from
permitting horizontal inequality. The guidelines suggested above help to
choose the more important equivalence. Where there are many users and where
the choice followed in the present case is not sure to become universal, it
is best to optimize over the uses of machinery rather than within the produc-
tion chain. Then, in step d), the tax is included in the shadow price. But
with this choice there is the unavoidable penalty of violating the marginal-
cost/marginal-produc tivity equivalence.



- 17 -
48.       This treatment of an excise tax is based on the assumption that
revenue is the government's tax objective. If in fact the tax rate on
machinery is close to the rate on other goods, the argument to include the
tax is even stronger, for then its inclusion permits rescource use optimiza-
tion across all the inputs so taxed. That is, L' avoids a bias against the
use of particular types of inputs that are subject to the evaJlation procedure.
In the case of a special tax on a particular good intended to discourage ibs
use, such as on a soot-producing fuel, then including the tax is obviously
appropriate, for it is intended that the vertical equivalence be violated.
Notice that there is a reverse conclusion where the excise is levied on the
output of the project being evaluated. lius optimizing across goods, the
price to apply to the output for evaluation should be net of tax, such as
would be received from the sale of the good by any other firm whose product
is similarly taxed. And again, if the tax is not intended to raise revenue
but to control the consumption of the good, it is even more clearly appropriate
to adopt the price net of tax.
Imort Duties
49.       Should import duties be included in the price of an input when a
project is evaluated? Some tariffs are adopted for revenue and others for
protection and we may judge the possibility, under step b), of price reform
accordingly. Unusually high rates that protect no domestic industry may be
destructive of industry that uses such imports and the best remedy would be
to lower the tariff.  It is certainly more cumbersome, although in some
situations perhaps also more feasible, to evaluate a project without the duty
charge and later obtain either a specific duty exemption or an offsetting
subsidy. But if the tariff is not set at an unusual level and other industries
pay comparable rates on other imports, optimization across industry users of
this and other inputs would lead to th6 conclusion that the correct price of
the input should include the duty. The alternative of vertical optimization
for a single project and input, the equalization of marginal import cost and
input productivity, by excluding the duty, would then require arbitrary and
special tax-subsidy transfers and probably would also entail horizontal mis-
allocation losses.
50.       If the tariff is set to protect a domestic industry that would not
be able to survive against open import competition, there is a temptation to
omit the duty (or to subtract it from the price of the domestic product) in
calculating the prospective yield of a new project so as not to impose the
inefficiency of one industry as a penalty on another potentially profitable
industry. But the argument is unsound. So long as the protected industry does
manage to sell its output to customers who pay the high, protected, price,
making an exception of some new prospective industry user would justify an
investment with a less productive use for the input than is standard. The
situa-bion would be even more costly if it were possible to argue that a number
of new industries will be established on the basis of a net-of-duty input
price and that therefore there will exist a class of users within which mar-
ginal-productivity equivalences will be preserved. For in this case the true
domestic marginal cost of the input throughout the new class of industry will
exceed its marginal productivity, and the net social loss of the protected
input-supply;ing industry could be greater than the net social benefit in the



- 18
new industry, for a possible net national loss. Catastrophic, perhaps, but
not extraordinary.
51.       If there is excess capacity in the protected industry and it is
desired to find new uses for the good, an inducement might be arranged in the
form of deducting the tariff in the price used for project evaluation.
However, a more flexible price policy would be a more direct means of achiev-
ing the objective. But the choice made will determine which industry will
require a subsidy. If the protected industry's price is cut, it will require
a subsidy., whereas if the input using industry must pay the full price after
it has been evaluated with a net-of-tariff cut price, it may have to receive a
subsidy. In either case care should be taken to avoid the expansion of social
net-loss industries by means of subsidies to the exten-t and in a way that they
can "compete" in thve export market and provide benefits to the rest of the
world at national eapense.
Private Monopoly
52.       The presence of monopoly provides both false temptation and genuine
opportunity for the application of shadow prices in less developed countries,
but not always in equal proportions. If the project being evaluated is ex-
pected to exercise some monopoly control over the price it charges for its
output, we cannot very well correct for the expected monopoly price by sub-
stituting one that would obtairn under competitive conditions, for the output
will in fact be restricted and the marginal social value of the output will
accordingly be greater than the competitive price. If the authorities are
aware of the prospect that the investment in question is likely to be operated
as a monopoly, they might more appropriately consider policies by which compe-
titive price and output might be achieved. For, if there is monopoly restric-
tion on output there is no effective way to have improved the results by having
used hypothetical prices in the investment decision. If a monopoly price is
expected to rule, then the consumption of the good will adjust to that price
and the monopoly price measures the marginal social benefit of the project even
though the marginal social cost is below the benefit.
53.       Similarly, in the case of a project the price of whose output will
be established by the necessity of competing with a substitute service over
which a firmn with monopoly power has control, the appropriate service price to
use in estimating social benefits must be based on that monopoly price. For
example, the benefit from a taxi service to supplenent basic urban transport
of a city bus system is related to the price and use of the bus system; the
higher the bus price, the larger the benefit of an added -unit of taxi service.
The fact that the bus-service price is set monopolistically is qui.te incidental
to the determination of tai values. If possible, an analyst might evaluate
an expansion of the bus service based on a "competitive"' price schedule, as an
alternative to an investment in taxis. But if that is rejected and the bus
system is to be taken as a datum, then so must the monopoly price. In the
evaluation of a taxi service the benefits depend on the alternatives available
and it does not matter whether the restriction on altemnatives is the conse-
quence of the niggardliness of nature or the restrictions of a monopolist.



- 19 -
54.       Where an investment is to contribute to the exports of a country
and the country has some degree of monopoly power over the price of the good,
then there is an opportunity for applying a shadow product price that differs
from the price that is expected to obtain in the market. Under these circum-
stances the price to an individual or small pmducer is the world price (plus
or minus taxes), whereas the net value of a unit of the product from the point
of view of the country is less than that by the amount the price must be
lowered to induce the world to accept the new unit. To encourage a correct
volume of investment in new capacity to produce the export good, a shadow
price equal to the marginal national revenue should be applied in investment
valuations. In such uses shadow prices, perhaps inplemented by imposing an
export tax, are more effective an(d more economical than is a directly adminis-
tered restriction on national exports of the good, with or without restrictions
on production.
Land Values
55.       There is a temptation commonly found irrestible to regard the value
of land, especially urban land, at; the manifestation of monopoly, and there-
fore, taking advantage of the possibility of using shadow prices to evaluate
projects, to declare the value of land to be nil. This is an ideological
error, a class of error in fwhich one familiar element is the Marxian-socialist
elimination of land value. Although this particular element is recognized as
an error by many economists, there is a surprising amount of confusion about
the class. The confusion relates to both the source of high land prices and
the consequence of ignoring them. Land prices can, of course, be raised by
land-market monopolies, and no doubt that is the best explanation of' some
interesting examples. Yet it is one of the most reliable predictions that
land prices (even discounted for general price level changes) will rise as
economic development takes place. The reliability of the forecast stems from
the natural limit on good locaLtion and the increasing cost of finding suitable
.3ubstitutes. There is nothing more classically scarce and hence more appro-
priately tagged with a risirng scarcity measlure, its price, than is land in
choice locations. Given its scarcity, its efficient use is of increasing im-
portance and therefore some mLethod of land valuation is needed. But if the
valuation of land is to have an influence on land usage, the prospective land-
using projects must be evaluated to the realistic assunption that such projects
entail the deprivation of still other project users of valued sites. Generally
the market price of land should be honored.
56.       A less subtle motive for wishing to be rid of the reatraining effect
of high land values is the desire to deny the influence of those users respons-
ible for the rise in land prices on the course of economic development. In
some specific circumstances it may be the luxury demand for residential sites
and in others the demand for land by the private sector that a government may
wish to deny. The specific motives vary, and shunting out the influence of
market prices can be an effective means of periitting the objectives of the
government to dominate over the objectives of others. While no outside expert
ought to presume to block an attempt of this kind, it would be a misuse of ex-
pertise to provide a spurious intellectual justification, based, say, on the
g        availability of land and posibility of monopoly restriction, for
ignoring market land prices and establishing instead zero, or arbitrarily low,
land shadow prices.



- 20 -
Public- Mnopoly
57.       Public monopoly often presents a price policy quite different from
private monopoly. Where governments are able to control the price at which
its services are supplied, it may be motivated to achieve some specific social
objective rather than to maximize its net return. But the diverse objectives
of governments as sellers of services implies that those services may be sold
at below equilibrium prices, requiring the service to be rationed by a non-
price mechanism. Here there is an uniambiguous case for using a shadow price
in place of the actual price charged to and paid by the user. For prices to
which it has been impossible to adjust because of rationing are likely to be
equal neither to the marginal cost of production nor the marginal benefit to
the user. The outputs of public utilities are commnly "socially" priced and
rationed. Water from irrigation works is conmmnly sold to farmers at arbitrary
prices and the amounts they may purchase are also restricted in some arbitrary way.
58.       An analyst may contribute to the rational organization of such ser-
vices by encouraging a reconsideration of the pricing policies of the agencies
responsible, step b) of the recommended procedure, before seeking to set a
shadow price. But if that attempt fails then step c) requires that a choice
be made of the domain over which resource use efficiency is to be achieved. If
it is assumed that the resource to be used in the project has a fixed supply
either because of the restrictions of nature or pi-Jic,y and there is no possibi-
lity for achieving efficiency by assigning more resources to the production of
the service, then it will not be possible to equalize costs of the service at
the margin with the benefits in the project from the use of the service. In
this case, it is better to attempt to equalize marginal benefits among all uses
of the given service.
59.       To find the correct shadow price we must ascertain the use or uses
from which the service is actually to be withdrawn, which ideally will be the
least valued ones. But we must be prepared for the possibility that the re-
source will be diverted from some higher valued use. This is possible, if not
most likely, precisely under these circumstances in which prlices are not used
to allocate the service but rather the resource is assigned by some administra-
tive system. Having discovered which activities are to be deprived of the
service we must estimate the value of the service in that use. This is the
amount which the user would be just willing to pay to retain the service. It
may be possible to approxmte this figure by the cost of the next best substi-
tute service, which is likely in fact to replace the service once the new pro-
ject is constructed. It is apparent that under some operating conditions a
potentially profitable project may be rejected when evaluated according to this
procedure. Nevertheless, it -uld be wrong to adopt a less stringent method
of setting shadow prices on the inputs in order to put a new project in a more
favorable light if in fact it cannot be exected to operate at such low costs.
This method of setting shadow prices calls attention to the real basis of an
unfavorable evaluation and either prevents the wasteful undertaking or encour-
ages a reform in allocation procedures to reduce the real cost of new projects.
Wage Rates
60.       The mispricing of labor in project evaluation can be among the most
pernicious of shadow pricing practices. The presence of surplus labor that in



- 21 -
some countries could not be fully employed by almost any set of policies
terpts pmrject evaluators to accomrplish by shadow pricing the reform in the
use of labor that has not been accomplished in actual employtant practices.
The temptation is made especially plausible, and may in some cases even be
justifiable,, by the arbitrary way wages are set in the labor market. This
may happen in cases where either union power or government legislation suc-
ceeds in raising the wages at which labor can be employed to unrealistic
levels, unrealistic, that is, because they encourage labor-saving methods,
and the production of goods that require little labor in their production
alt-hough mass unemployment continues to exist. If "thight" wages are the pro-
duct of the exercise of arbitrary market power then there is sufficient basis
for using shadow wages to evaluate projects and bo deal with the problem of
subsidizing employment that must be paid for at higher market wages. The
problems associated with this policy have been discussed above in Section III.
It is more likely, however, at least in many countries, that wages are set by
non-arbitrary market processes even in the presence of unemployment, and can-
not be regarded simply as the consequence of the exercise of power that is
insensitive to economic circumstances. If this is so then the use of shadow
wages to offset actual wages will neither accomplish the intended improvement
in employment nor guarantee a more efficient selection l.>f projects.
61.       In specific situations there are a number of well-established and
fairly well-understood reasons why industrial wages may exceed the marginal
productivity of the labor occupied in subsistence agriculture. Skills and
attitudes of considerable importance in organized wage enployment differentiate
the workers of the oountryside fmm workers who have spent years (even) in
Ilunskilledif jobs. There are transfer costs incurred in recruiting potentially
available labor from subsistence fanming. In some places preferences for the
freedom of non-wage employmient must be o-vercome by wage premiums. Attitudes
vary and the willingness to work for wages is far less among some classes,
castes, tribes, generations than among others, yet the wage must rise to over.-
come the resistence of the most reluctant among those recruited.l/ In addition
to these restrictions on the supply of labor that occur normally in an open
competitive labor market, there are non-conpetitive restrictions that are yet
not the gross effects of labor union power or labor legislation. Some skills,
for example, are preempted by custom and expectation by some (tribal) groups,
and the supply of those skills cannot easily be expanded by drawing workers
from other groups. In other cases there are more or less professional tasks
that are restricted to those who qualify by virtue of some specific preparation.
The labor forces of advanced countries are elaborately divided into groups
between which competition is restricted by devices that receive the approbation
of the community, hence it ought not to be surprising that the labor markets of
less advanced countries are similarly structured.
l/ R. S. Weckstein, "Wages and Labor Scarcity," in Deveopment Policy-
Theory and Practice, ed. G. F. Papanek, Harvard University Press,
Cambridge, 196d.; pp. 319.342.



- 22
62.       Where the wages actually paid in industry are higher than the
marginal productivity in subsistence agriculture as a result of any of these
phenomena, and not sinply the result of arbitrary power in setting wages, it
is unclear what, if anything, is the operational meaning of "labor surplus."'
Wages that respond to these phenomena of preemption, preference, institutional
standards, transfer costs, etc., convey information of the actual tenrs on
which labor is made available to industrial employers. In cases of absolute
restriction on the supply of a particular kind of skilled labor, the correct
wage is that which clears the market for the labor supply, even if we regard
the income to such a restricted group as a rent. Efficient allocation of that
restricted skill requires that the rent be charged and paid (although it could
be taxed away). Setting some lower shadow wage while leaving unchanged the
entry restriction9 will not expand employment and it will make it more diffi-
cult to assign efficiently the available workers to activities.   It will also
remove one element of relevant discrimination among new projects, their real
labor cost.
63.       It is tempting, but incorrect, to argue that where a labor surplus
exists, as employees of established firms are diverted into new projects, the
ranks of this group can, and generally will, be filled from the open pool of
underemployed, and that therefore the real cost of using such workers is the
loss of output of the workers who replace the diverted workers, i.e., the mar-
ginal product of the subsistence fa=r workers.   Hence the shadow wage, accord-
ing to the argument, is this marginal product. If in fact the worker group
involved is expandable by recruitment from below,the market supply price is
likely to reflect the terms on which such replenishment takes place. If the
replenishment were costless and the skill differences among early and late
comers were unimportant, the supply price would not tend to rise, as it comon-
ly does, when such employment increases.   If the wage is conveying information
about these transfer costs and skill differences, then the actual industrial
wage may be a correct measure of social cost and it should be used in froject
evaluation.
64.       This is not entirely satisfactory, for it leaves unresolved the
waste of artificial restriction, irrelevant standards, the immobility that
comes from ignorance, the inequity caused by selfish groups protecting their
own petty advantages, the transfer costs that reflect the failure to improve
the organization of the entire labor market. All these in turn impose their
costs by tilting project selection toward relatively unproductive choices and
leaving a legacy of people increasingly incapable of coping with their lives
while they remain in the thrall of their traditional economies. Nevertheless,
we mrust distinguish between the usefulness of shadow wages when they correct
arbitrary labor pricing, and the harm that can be done by using shadow wages
as though they were a panacea for the imperfections that infest the economies
o0 developing countries.
Fragmented Grain Markets
65.       In addition to the barriers that divide the labor markets from one
another in less developed countries there are numerous other divisions of
importance, such as the divisions frequently found in the grain markets ,and in
the capital markets. The first of these is usually caused by the policies of



- 23 -
local or regional authorities in their effort to achieve security in the
provisioning of food grains whenever there is a threat of famiine. Under
these conditions embargoes on export of grain from "surplus"t regions to
"deficitt" regions invariably create a geographical grain-price structure
that raises problems in the evaluation of agricultural investments. Clearly
it would be best for the countxy to avoid such security-oriented local policies
by depriving regions of the authority to control internal trade. Failing that,
we must consider the possibility that a single anticipated grain price for the
en.tire country may not be optimal, and in aidition the location of grain spe-
cialization that would be optimal for the country with free arbitrage in grain
may be a disaster if there are to be restrictions on grain arbitrage. The
correct price expectations on which to make planting decisions are rather com-
plicated, but the surplus regions can easily be encouraged wastefully to over-
invest by including in their expectations price levels that are only likely
to occur in deficit regions. And deficit regions would undexinvest in grain
plantings if they neglect the influence of import limitations imposed from
without on their expected price. Clearly no single national or, a fortiori,
international price will give acceptable guidace.
Interest Rates
66.       Comrrparable divisions are commnly found in the markets for capital,
even in advanced countries, and in less developed countries where the capital
markets are both less developed and arbitrage among them less effective, the
consequent multiplicity of interest rates is even greater. Nevertheless,
them is a temptation to seek a single rate to apply to the evaluation of a
pmject, and often a single rate is believed to be applicable throughout the
economy. There are, however, two classes of fact that prevent a single rate
from being appropriate to all projects or, depending on the desired precision
with which projects are to be evaluated, a single rate even being correct for
each project taken by itself. The first is that the real capital cost, under
conditions of separated capital markets, is not uniform for the totality of
investable capital. Or to put it another way, the social cost of the delay
of net returns may not be unifomn for all components of the benefits and costs.
Also the social cost of delayed returns may be different for one project than
for another. The second class of fact is that capital in the government's
budget may not be allocated so as to equalize the returns among component
parts either because of an administrative bias or because that is the policy
of the government.
67.       It is often necessary for a project to obtain financing from more
than one source and at more than one set of terms, Yet we cannot take as the
cost of capital the rate that must be paid for the most (Pcpensive part of the
capital, for the parts are usually not independent and the higi*st rate is not
the marginal cost of capital. If the financing is arranged in a package, as
is common, then the average interest rate may be a better measure of capital
cost than the highest rate. But that is too simple a version of the capital
cost. Project financing may actually cover only the direct expenditures
within the firm undertaking the project and the purchased inputs nmunt be
financed in the open market by t1ie supplying firms. If, as is likely, these
firms lack access to the privileged capital market from which the project is
primarily financed, if their credit standing is less excellent, if their market



- 24 -
for capital is local, less well developed and unconnected with favored markets
because of capital rationing there, then part of the project must be financed
at rates of interest above the rate charged on the part financed directly.
Hence to apply the direct financing rate to the valuation of the entire pro-
ject's net returns will underestimate the true social cost of delay, at least
insofar as the social cost of delay is inferred from the market rate of interest.
The higher actual rates paid by the suppliers of inputs reflect the higher
social costs of the capital sacrifice required for the project. The use of
capital paid for at high rates in some local capital market implies the diver-
sion of capital from other local users of the capital who presumably would have
paid the higher rate. We cannot suppose that funds will flow into that market
from the lower rate markets to make up the loss, for there must be a limit on
the supply of such low-priced capital and it must be rationed to the acceptable
users. But even if t-hat were not entirely the case, the fact that some such
capital does flow to the higher rate markets does not imply that the price will
equalize. The local rate, except for the influence of local monopoly, is likely
to reflect correctly the terms on which the funds do move from one market to
another.
68.       These local interest rates are therefore a good measure of social
cost, if we disregard the possible importance that should be given to external-
ities in the evaluation of the social cost of delayed returns. To take account
of externalities would carx-y us in the direction of applying a rate that may
be even lower than the market rate of interest in the most favored markets.
The case for using such low rates is a speculative argument that if individuals
were able to make their decisions3 to save and invest in contract with the rest
of society so they could each know that his decision would be responsible for
a major improvement to his eventuLal heirs, he would prefer to save and invest
more than he reveals he would in his individualistic response to the market
rate of interest. However, if a low interest rate were accordingly adopted
for project evaluation, there wuLd be the necessity either to raise the level
of saving by taxation, or to ration too little capital by non-price means.
The argument is a justification for an unspecified amount of increased taxation
without a test of the correctness of the policy. The relevance of the policy
in the present context is that suich social rates ought to be applied in a
uniform way, not alone to the pinjects being systematically evaluated, but to
projects whose capital source is limited to high interest rate markets as well.
Otherwise there is bound to be a significant biasing of the projects undertaken,
in favor of large projects, industrial projects, projects financed by foreign
funds or government funds, etc. whatever is the basis on which projects are
chosen for systematic evaluation and receive the benefit of the social interest
rate policy, against all others.
69.       Retuning, finally, to tae objections to a single, all-purpose,
shadow rate of discount, the second class of facts referred to above is that
governments do not always or evern usually, allocate capital resources among
the sectors of the economy so as to equalize net returns, and they fail to do
this because, given the systems of accounting for net returns, goveraxuents
adopt policies to favor some sectors over others. Where this is so it is
clearly inconsistent with the government's policy to apply a universal dis-
count rate to the net returns of projects in sectors in which otherwise
different rates are the rule. This would be more or less obvious if govern-
ments' policies with respect to sector allocations of capital were expressed



by specifying different rates of interest to be used in project evaluation
and -then allowing the amounts of capital to be allocated to implement the
policy. But the most common practice is to reverse the sequence, to allocate
sums of capital and to allow consequentL discrepancies in the net returns.
The discrepancies are then not recognized as having been intended. But
logically, and practically, the sequence is immaterial and we are bound to
infer the intent of governments from either form. Governments wish to favor
some sectors over others and this can be done, and is done, by means of dis-
crepant rates of discount. Project evaluation ought to conform to these
intentions.
3. Concluin    Remarks
70.       The use of shadow prices can under some unfortunate circumstances
contribute to a further fracturing of an economy and decreasing of efficient
project choice. This may happen where a number of agencies each follows its
own rules and guidelines in the evaluation of projects, each believing some-
thing different about the correct way to establish shadow prices. It would
probably be better for aLL to be bound by a belief in the sanctity of market
prices, for at least there would be a good chance then that a project would
not be discriminated aga;.nst simply because it was within the domain of the
wrong agency. Fracturin6 of an economy is particularly likely if outside
agencies insist upon following their own price-making rules while inside
agencies follow others. Conformity to local practice may be better than being
right according to some abstract ideas held in lonely isolation.
71.       After the project has been approved and the investment is committed
there is the subsequent need to evaluate the performance of the management
and to provide appropriate incentives for efficient operation. Where shadow
prices have been used in the choice of the project a problem of some interest
then arises for the subsequent evaluations. Since projects that are socially
profitable when evaluated with shadow prices may be unprofitable against the
standard of ordinary market prices, subsidies are required for private firms
and deficits must be tolerated in the case of public firms. If the subsidies
are paid in relation to the deficit or according to some bargain between enter-
prise and gover-nment not directly related to the shadow price market discrepancy,
then the profit-loss incentive to efficient performance is destroyed. Profit
maximizing firms will not correctly economize in th.e use of scarce resources.
There is therefore a good deal of importance in the way subsidies are arranged
if projects are to be made responsive to changing conditions and if errors are
to be discovered and something is to be done about them,.
72.       In whatever way the subsidies are paid, there is another problem
raised by the need for a subsidyj, namely, the equity with which it is financed.
The need to finance a subsidy by some new tax or other source of revenue will
redistribute income, and given the common institutional constraints on the
fiscal powers of less developed govements the redistributions available may
not include any desirable ones. But we would not be justified in recommendiLng
a project on the basis of a benefit-cost evaluation without also weighing th'ze
associated redistributions required to finance a necessary pmject subsidy.
LIEven if it can be shown that the net benefit's of the project are sufficiently
high to warrant the investment, the expectation of a redistribution of income



- 26 -
can conceivably offset such gains and they must therefore be included in the
evaluation. The need to finance a deficit will,, in addition to the equity
problem, make a new contribution to the biases already present in the price
system. This will be difficult to weigh explicitly into the project values
or in the balance by which the use of shadow prices is judged, except that we
know it must be a cost and not a benefit.
73.       We conclude tentatively, that less developed economies have badly
distorting price systems that are an inevitable challenge to economic admninis-
trators to guide the choice of projects by improved methods. The improvemernt
of natural prices is obviously the prime means by which to gain allocation
improvements. But here the best is the enemy of the good and optimal prices
that assume a unifomr economy subject to general equailibrium analysis will not
move us in the right direction.  Instead we need specific local knowledge of
true scarcity values for the shadow prices we use to replace some bogus natural
prices.  Finally, the opportunity to create prices is a shaxp and potentially
dangerous instrument and we must learn the fine art of judicious selection
where there are multiple prices, and we must preserve the incentive role of the
market system at the same time as we improve its information role.