You are on page 1of 18

American Economic Association

Migration, Unemployment and Development: A Two-Sector Analysis


Author(s): John R. Harris and Michael P. Todaro
Reviewed work(s):
Source: The American Economic Review, Vol. 60, No. 1 (1970), pp. 126-142
Published by: American Economic Association
Stable URL: http://www.jstor.org/stable/1807860 .
Accessed: 19/01/2013 02:02
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp
.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.
.
American Economic Association is collaborating with JSTOR to digitize, preserve and extend access to The
American Economic Review.
http://www.jstor.org
This content downloaded on Sat, 19 Jan 2013 02:03:00 AM
All use subject to JSTOR Terms and Conditions
Migration, Unemployment and
Developmnent: A Two-Sector
A nalysis
ByJOHN R. HA RRIS A ND MICHA EL P. TODA RO*
Throughout mnanyless d eveloped econ-
omies of the world , especially those of
tropical A frica, a curious economic phe-
nomenon is presently taking place. Despite
theexistence of positive marginal prod ucts
in agriculture and significant levels of ur-
ban unemployment, rural-urban labor
migration not only continues to exist, but
ind eed , appears to be accelerating. Con-
ventional economic mod els with their
singular d epeind enceon theachievement of
a full employment equilibrium through
appropriate wage and price ad justments
are hard put to provid e rational behav-
ioral explanations for these sizable and
growing levels of urban unemployment in
the absence of absolute labor red und ancy
in the economy as a whole. Moreover, this
lack of an ad equate analytical mod el to
account for the unemployment phenome-
non often lead s to rather amorphous ex-
planiations such as the "bright lights" of
the city acting as a magnet to lure peas-
ants into urban areas.
In this paper we shall d iverge fromthe
usual full employment, flexible wage-price
mod els of economic analysis by formulat-
ing a two-sector mod el of rural-urban
migration which, among other things,
recognizes the existence of a politically
d etermined minimumurban wage at levels
substantially higher than agricultural
earnings.' Weshall then consid er theeffect
of this parametric urban wage on therural
ind ivid ual's economic behavior when the
assumption of no agricultural labor sur-
plus is mad e, i.e., that the agricultural
marginal prod uct is always positive and
inversely related to the size of the rural
labor force.2 The d istinguishing feature of
this mod el is that migration proceed s in
response to urban-rural d ifferences in
expected earnings (d efined below) with the
urban employment rateacting as an equil-
ibrating forceon such migration.3 Weshall
then usetheoverall mod el for thefollowing
purposes:
1) to d emonstrate that given this po-
*
The authors areassistant professor of econonmics,
Massachusetts Institute of Technology and research
fellow, Institute for Development Stud ies, University
College, Nairobi, respectively. They would like to
thank the Rockefeller Found ation for making possible
their research on economic problems of East A frica.
Peter Diamond , Richard Eckaus, Joseph Stiglitz, two
anonymous referees, and the managing ed itor mad e
valuablecomments on a previous d raft. Theauthors, of
course, areresponsiblefor remaining errors.
1
For some empirical evid ence on the magnitud e of
these real earnings d ifferentials in less d eveloped
economies, seeReynold s, Berg, Hend erson, and Ghai.
2 Wed o not makethespecial assumption of an agri-
cultural labor surplus for the following reasons: Most
available empirical evid ence to d ate tend s to cast
d oubt on the labor surplus argument in thecontext of
those economies of Southeast A sia and Latin A merica
wheresuch a surplus would bemost likelyto exist (see
Kao, A nschel, and Eicher). Moreover, few if any
economists would seriously argue that general labor
surplus exists in tropical A frica, thearea to which this
paper is most d irectlyrelated .
3
For a d ynamic mod el of labor migration in which
urban unemployment rates and expected incomes play
a pivotal rolein themigration process, seeTod aro. How-
ever, unlikethepresent mod el which attempts to view
themigration process in context of aggregateand inter-
sectoral welfare consid erations, Tod aro's mod el was
strictly concerned with the formulation of a positive
theoryof urban unemployment in d eveloping nations.
A s such, it d id not specificallyconsid er thewelfareof the
rural sector, nor was it concerned with the broad er
issues of economic policy consid ered in the present
paper.
126
This content downloaded on Sat, 19 Jan 2013 02:03:00 AM
All use subject to JSTOR Terms and Conditions
I-IA RRIS A ND TODA RO: MIGRA TION A ND DEVELOPMENT 127
litically d etermined high minimumwage,
the continued existence of rural-urban
migration in spite of substantial overt ur-
ban unemployment represents an econom-
ically rational choice on the part of the
ind ivid ual migrant;
2) to show that economists' stand ard
policy prescription of generating urban
employment opportunities through theuse
of "shad ow prices" implemented bymeans
of wage subsid ies or d irect government
hiring will not necessarily lead to a welfare
improvement and may, in fact, exacerbate
the problem of urban unemployment;
3) to evaluate the welfare implications
of alternative policies associated with
various back-to-the-land programs when
it is recognized that the stand ard remed y
suggested by economic theory-namely,
full wage flexibility-is for all practical
purposes politically infeasible. Special
attention will begiven here to the impact
of migration cumunemployment on the
welfareof therural sector as a wholewhich
gives rise to intersectoral compensation
requirements; and , finally,
4) to arguethat in the absence of wage
flexibility, an optimal policy is, in fact, a
"policy package" includ ing botk partial
wage subsid ies (or d irect government em-
ployment) and measures to restrict free
migration.
I. TheBasic Mod el
The basic mod el which weshall employ
can be d escribed as a two-sector internal
trad emod el with unemployment. The two
sectors are the permanent urban and the
rural. For analytical purposes we shall
d istinguish between sectors fromthepoint
of view of prod uction and income. The
urban sector specializes in the prod uction
of a manufactured good , part of which is
exported to therural sector in exchange for
agricultural good s. The rural sector has a
choice of either using all available labor
to prod ucea singleagricultural good , some
of which is exported to theurban sector, or
using onlypart of its labor to prod uce this
g,ood while exporting the remaining labor
to theurban sector in return for wages paid
in theformof themanufactured good . We
arethus assuming that thetypical migrant
retains his ties to the rural sector and ,
therefore, the income that he earns as an
urban worker will be consid ered , fromthe
stand point of sectoral welfare, as accruing
to therural sector.4 However, this assump-
tion is not at all necessary for our d emon-
stration of the rationality of migration in
the face of significant urban unemploy-
ment.
The crucial assumption to be mad e in
our mod el is that rural-urban migration
will continue so long as the expected urban
real income at the margin exceed s real
agricultural prod uct-i.e., prospective rural
migrants behave as maximizers of expected
utility. For analytical purposes, we shall
assume that the total urban labor force
consists of a permanent urban proletariat
without ties to the rural sector plus the
available supply of rural migrants. From
this combined pool or urban labor, we
assume that a period ic rand omjob selection
process exists whenever the number of
available jobs is exceed ed bythenumber of
job seekers.5 Consequently, the expected
4
In tropical A frica especially, this notion that mi-
grants retain their ties to theruralsector is quitecom-
mon and manifested by the phenomenon of the ex-
tend ed family systemand the flow of remittances to
rural relatives of largeproportions of urban earnings.
However, the reverseflow, i.e., rural-urban monetary
transfers is also quite common in cases where the
migrant is temporarilyunemployed and , therefore, must
be supported by rural relatives. For an excellent d is-
cussion of this phenomenon froma sociologicalpoint of
view, seeGugler (pp. 475-78).
6 The qualitative conclusions of the mod el d o not
d epend on the precisenature of the selection process.
We have assumed rand omselection not merely for
analytic convenience but also because it d irectly cor-
respond s to an appropriated ynarnic construct d eveloped
in Tod aro's 1969 article. There it is shown that over
time expected and actual earnings will converge to a
positive number even though therateof job creation is
less than therateof migration so that unemployment is
increasing.
This content downloaded on Sat, 19 Jan 2013 02:03:00 AM
All use subject to JSTOR Terms and Conditions
128 'rlE A MERICA N ECONOMIC REVIEW
urban wagewill bed efined as equal to the
fixed minimumwage(expressed in terms
of manufactured good s) times thepropor-
tion of theurban labor forceactuallyem-
ployed (see equation (6)).
Finally,
we
assume perfectly competitive behavior
on thepart of prod ucers in both sectors
with the further simplifying assumption
that the price of the agricultural good
(d efined in terms of manufactured good s)
is d etermined d irectly by the relative
quantities of the two good s prod uced .
Consid er now thefollowing formulation
of themod el.
A gricultural Prod uction Function:
(1) XA
k-(IA A ,
L,
KA ),
4
q > 0
q"t
< 0
where,
XA
is
output
of the
agricultural good ,
NA
is the rural labor used to
prod uce
this output,
7L is thefixed availabilityof land ,
KA is thefixed capital stock,
q' is thed erivativeof q with respect of
NA , its onlyvariablefactor.
Manufacturing
Prod uction Function:
(2)
XM
-
f(A Tm, KM),
f' > 0, f" < 0
where
XM is theoutput of themanufactured
good ,
NM is thetotal labor (urban and rural
migrant) required to prod ucethis
output.
KM is fixed
capital stock,
and
f' is thed erivativeof
f
with respect to
N._ , io nlr%"I rorsl fnofnrr
Price Determination:
(3)
(
XM
0
where
P, thepriceof theagricultural good in
terms of the manufactured good ,
(i.e., theterms of trad e) is a function
of therelativeoutputs of agricultural
and manufactured good when the
latter serves as numeraire.6
A gricultural Real, WageDetermination:
(4)
WVA P*q
where
WA ,
theagricultural real wage, is equal
to the value of labor's marginal
prod uct in agricultureexpressed in
terms of the manufactured good .
Manufacturing Real TWage:
(5)
,- -
f' WM.
The real wage in manufacturing, ex-
pressed in terms of manufactured good s, is
equated with the marginal prod uct of
labor in manufacturing becauseof profit
maximization on the part of perfectly
competitiveprod ucers. However, this wage
is constrained to begreater than or equal
to thefixed minimumurban wage. In our
analysis, we shall be d ealing only with
cases in which
f'
- Wm(i.e.,
thereis never
an excess d emand for labor at the mini-
mumwage).
Urban Expected Wage:
(6)
Wu I, 1
It is interesting to notein this context that sociologist
Gugler who has spent consid erable time stud ying
labor mnigration in A frica has recently conclud ed that
rural-urban mi0ration is essentially an economic
phenomenon that can be portrayed as a "game of
lottery" in which ruralmigrants cometo thecity fully
awarethat their chances of find ing a job arelow. How-
ever, thegreat d isparitybetween urban and ruralwages
makes thesuccessfullocation of an urban salaried job so
attractive that unskilled migrants arewilling to takea
chance(pp'.
472-73). Seealso Hutton.
6
A sufficient, but not necessary, cond ition for this
assumption is that all ind ivid uals in theeconomyhave
the same homothetic preference map. A gain, the as-
sumption is mad efor analytical convenience. The qual-
itative conclusions of our analysis will remain unaf-
fected und er several plausibleassumptions about d istri-
bution of incomeand tastes.
This content downloaded on Sat, 19 Jan 2013 02:03:00 AM
All use subject to JSTOR Terms and Conditions
HA RRIS A ND TODA RO: MIGRA TION A ND DEVELOPMENT 129
wheretheexpected real wagein theurban
sector, Wu, is equal to thereal miniimum
wageWM ad justed for theproportion of
the total urban labor force (permanent
urban plus migrants, d enoted as
N.)
ac-
tuallyemployed , NM/NU.7 Onlyin thecase
of full employment in the urban sector
(NM=N.) is the
expected wageequal
to
theminimumwage(i.e., W,= WM).
Labor End owment:
(7) NA + NU
=
NR + N,
N
There is a labor constraint which states
that thesumof workers actuallyemployed
in the agricultural sector (NA ) plus the
total urban labor force(NU) must equal
the sumof initial end owments of rural
(NR)
and
permanent urban
(N.) labor
which in turn equals the total labor en-
d owment (N).
EquilibriumCond ition:
(8) WA
=
wit
Equation (8), an equilibriumcond ition, is
d erived fromthehypothesis that migra-
tion to theurban area is a positivefunc-
tion of theurban-rural expected waged if-
ferential. This can bewritten formallyas
(9) Nu U pq
?Vu~Nu
{' > O, 41(0)
=
O
where]VUis a timed erivative. Clearlythen,
migration will cease only when the ex-
pected incomed ifferential is zero, thecon-
d ition posited in (8).8 It is important to
note that this assumes that a migrant
gives uponlyhis marginal prod uct.9
Wethus have8 equations in 8 unknowns
XA , XM, NA , NM, WA , WU, N.
and P.
Given the prod uction functions and fixed
minimumwage WM, it is possible to solve
for sectoral employment, the equilibrium
unemployment rateand , consequently, the
equilibriumexpected wage, relative out-
put levels and terms of trad e. Let us
analyze how such an unemployment equi-
libriumcan comeabout.
The essence of our argument is that in
many d eveloping nations the existence of
an institutionally d etermined urban min-
imumwage at levels substantially higher
than that which the free market would
allow can, and usually d oes, lead to an
equilibriumwith consid erable urban un-
employment. In our mod el migration is a
d isequilibriumphenomenon. In equilibrium
WMNM/NU= Pq' and migration ceases.
(See A ppend ix I for proof that this equi-
libriumis stable.) Now we know from
equation (5) that in thecompetitive urban
manufacturing sector, WM=J'. XVealso
know fromequation (7) that
N-NA = N.
and fromequation (3) that P= p(XM/
I
This assumes a veryparticular formof wageexpec-
tation, namely that the expected wage is equal to the
average urban wage. A lthough this is a convenient
expression to work with, wecould bemoregeneral and
make the expected wagesomefunction of theaverage
urban wage. Ind eed , theonlyrestrictions on such a func-
tion that arenecessaryfor our results arethat, ceteris
paributs, the expected wage varies d irectly with the
minimumwageand inverselywith the unemployment
rate.
8
V(0)
= 0
is purelyarbitrary. If, instead , weassume
;p(a)
=0 wherea can takeon anyvalue, migration will
ceasewhen the urban-ruralexpected wage d ifferential
is equal to a. None of the subsequent analysis is af-
fected qualitatively by specifying a=0. Equation (8)
would merelybewritten as WA +a =
V,.
I
Other assumnptions could be mad e. Much of the
literaturehas stressed that in peasant economies pro-
d ucers receive their average prod uct which is higher
than their marginalprod uct. Ind eed , this is at theheart
of thewell-known Lewis and Fei-Ranis mod els. How-
ever, these mod els ignore the migration d ecision and
seemto assumethat migrants continue to receivetheir
shareof peasant prod uction yet migrateonlyif jobs are
actually available. In much of A frica it appears that
migrants continue to receive income fromland after
migration and commonly hirelabor to work on their
farms in their absence. There is also a consid erable
groupof land less ind ivid uals who work on farms for
wages. Thus it would appear that our assumiption is not
unreasonable. The anIalrsis could easilybemod ified to
makeearnings foregoneequal to averageprod uct, bow-
ever.
This content downloaded on Sat, 19 Jan 2013 02:03:00 AM
All use subject to JSTOR Terms and Conditions
130 THE A MERICA N ECONOMIC REVIEW
NA
NA
H
v
zI \ Ha
I
I
|
\NA +NM=&
X I I\
NM
N*
Na
NM
FIG 1.
XA 1). Therefore,
wxecan rewrite our
equi-
libriumcond ition (8) as
f'NM
(8')
1D=
P(XM/XA )q'
- I = 0.
N
-
NA
SiniceXM and XA arefunctions of NM and
NA respectively, b is an implicit function
in NA and NM which, for any stated min-
imumwage, can be solved for the equi-
libriumcombination of agricultural and
manufacturing employment. From this
solution thelevels of urban unemployment
and commod ity outputs can also be d e-
termined . There will be a unique equi-
libriumassociated with each possible value
of the minimumwage, and the locus of
theseequilibria is plotted in Figure 1 as the
line 1=O in NA , NM space.10 The line NA
+ NM = N in Figure1 is thelocus of full-
employment points.
Point Z is theonlyequilibriumfull-em-
ployment point in Figure1 at which
NM
workers would be employed in manu-
facturing and N* in agriculture. Points on
thelocus 1= 0 east of Z areinfeasibleand
will not beconsid ered further, whilepoints
to the west of Z areassociated with min-
10
In Figure1 wehaveassumed that
d .NA
d M
=
[NM /14NA
] >
0
d A Tm M
although this need not necessarily hold true. Dif-
ferentiating (8') partiallywith respect to NA Wefind that
A ,pfqf2 2 pqif
pq
A
q2 +pqNA p
which is unambiguouslynegativesinceq" <0 and p'>0.
Differentiating (8') partiallywith respect to NM wefind
that
1 f'NN
77LW
XM
which is less than, equal to, or greater than zero as
1 f'Nm>
- + P -
= 1,
77LW XM <
where
d NmW.
?7LW = -
d W. NM
is thewageelasticityof d emand for labor and
d P XM/XA
d
(XM) P
is theelasticity of the terms of trad ewith respect to a
changein relativeoutputs. It follows, thereforethat the
slopeof the locus of equilibria, d NA /d NM d epend s on
therespective employment and priceelasticities.
A sufficient cond ition for 4?NM to benegative (making
d NA /d NM positive) is for the wage elasticity of em-
ployment to beless than one, a situation which recent
empiricalstud ies suggest is likelyto exist (seeErickson,
Harris and Tod aro (1969), and Katz). However, even if
77LW exceed s unity, d NA /d NM can still bepositive pro-
vid ing priceelasticity is sufficientlyhigh. The logic of
thesecond itions is clear. If V7LW iS less than one, a d e-
clinein theminimumwagewill lower the urban wage
bill even though employment and output increase.
This causes theexpected urban wageto d eclinethereby
red ucing theexpected rural-urban earnings d ifferential
which gives rise to reverse migration and increased
rural employment and output. If V7LW exceed s unity, a
fall in the minimumwage is accompanied by an in-
creased urban wage bill and , hence, a higher expected
urban wage. However, the expected rural-urban earn-
ings d ifferential can either increaseor d ecreasein this
case d epend ing on the movement in terms of trad e
which raises the value of the marginal prod uct in
agriculture. For example, if
77LW
were1.5 and thewage
share of manufacturing output (f'NA r/XM) were .50,
then an agricultural price elasticity greater than 0.67
would hesufficient to maked NA /d NM positive.
This content downloaded on Sat, 19 Jan 2013 02:03:00 AM
All use subject to JSTOR Terms and Conditions
HA RRIS A ND TODA RO: MIGRA TION A ND DEVELOPMENT 131
imumwages higher than thefull-employ-
ment wage. Thereis a monotonic mapping
such that higher minimumwages areas-
sociated with points on 4 = 0 lying farther
to thewest. Thus wecan d emonstratethat
thesetting of a minimumwageabovethe
market-clearing level causes an economyto
settleat a point such as H in Figure1. A t
H, N'A workers areemployed in agricul-
ture, NM in manufacturing, and NU-NM
workers areunemployed . It is evid ent that
theminimumwagecauses a loss of em-
ployment and henceoutput in both sec-
tors.'1
It is important to notethat even though
an equilibriumat point H represents a
suboptimumsituation for theeconomyas
a whole, it d oes represent a rational, utility
maximizing choicefor ind ivid ual rural mi-
grants given the level of the minimum
wage.
Onefinal point might beraised at this
juncture. So far wehaveassumed that the
urban minimumwageis fixed in terms of
themanufactured good . What if, instead ,
theminimumwagewerefixed in terms of
the agricultural good ? We would then
substitutefor equation (5):
(5') 1. '
> WM
p
Substituting (4), (5'), and (6) into (8) we
get theequilibriumrelationship
(11) (Pi). NrM
P.q
=
INu4
Wecan then imaginean economystarting
initially at the point on the prod uction
possibilities frontier at which XM is that
for which equation (5') is satisfied and
assumethat
(4$).NM
Pq'<
NU
at that point. Theequilibriumpoint will
again bereached through a simultaneous
raising of Pq' and lowering of W,
in re-
sponseto migration. A s relativeagricul-
tural output falls, P will rise. This in turn
will causeoutput of themanufactured good
to fall as well, sinceprod ucers will pro-
d uceupto thepoint
thatJf'=
WM P which
rises in terms of themanufactured good .
Note that
f'
can be raised onlythrough
output restriction
(sincef"
< 0). Theref ore,
in general, wewould find that imposition
of a minimumwagegives riseto an equi-
libriumcharacterized by unemployment
and loss of potential output of both good s.
A new locus V'= 0 will bed efined in Figure
1 such that thepoint on V' correspond ing
to anygiven minimumwagewill bewest of
thecorrespond ing point on 4b.
A lthough our initial assumption is a bit
easier to hand le, theprincipal conclusion
remains unaffected if we make the min-
imumwagefixed in terms of theagricul-
tural good . Equilibriumis onlyachievable
with unemployment. A ctual minimum
wagesetting is usuallyd onewith reference
to somegeneral cost of living ind ex, and
food is thelargest singleitemin thebud get
of most urban workers. (SeeMassell and
Heyer, and theNigeria report.) Hence, the
second casemaybesomewhat morereal-
istic. Notethat in thefirst casethe"true"
real wagewas red uced somewhat by the
rising agricultural price, whilein thelatter
caseit is increased bythefalling relative
priceof themanufactured good .
11
If d NA /d NM < 0, which webelieveto beempirically
unlikely, this statement would have to bemod ified . In
such a case, increasing theminimumn wagewill d ecrease
manufacturiing employment but will increase agricul-
tural employment and output. Unemployment will
result fromtheinmposition of a miniimumwagebut we
can no longer assert that the level of unemployment
will increase concomitantly with the level of the
m-iiinimumwage.
This content downloaded on Sat, 19 Jan 2013 02:03:00 AM
All use subject to JSTOR Terms and Conditions
132 THE A MERICA N ECONOMIC REVIEW
III. Implications for Development Policy
A . Planning in Terms of Shad ow Prices
Thestand ard solution to theproblemof
an institutionally d etermined wage that
is higher than theequilibriumlevel is to
employlabor in thepublic sector accord ing
to a shad ow wageand /or to grant a payroll
subsid yto privateemployers that equates
privatecosts with this shad ow wage.12 Two
main problems arisewith this prescrip-
tion: first, how can oned eterminetheap-
propriateshad ow wage? and , second ly,
what aretheimplications of executing such
a schemewhen theinstitutional wagewill
continueto bepaid to theemployed ? Our
mod el can shed light on both of these
issues.
In a static framework the appropriate
shad ow wageis the opportunitycost of
labor hired bytheind ustrial sector. Hence,
if labor is hired to thepoint that its mar-
ginal prod uct in ind ustryis equated with
theshad ow wagewhich in turn is equated
with themarginal prod uct in
agriculture,
marginal prod uctivity of labor will be
equal in both sectors, a necessarycond ition
for an optimal allocation of resources. Na-
turally, this assumes a positivemarginal
prod uct in agricultureand sufficient factor
mobility to ensure full employment of
labor. Theexistenceof urban unemploy-
ment, however, suggests that theremaybe
a pool of labor that can betapped without
sacrificing output. Consequently,
it might
besuggested that even though agricultural
labor is fullyemployed at peak seasons, the
appropriateshad ow wagefor urban labor
is likelyto beonethat is lower than the
marginal prod uct in agriculture. This
would becorrect if the two labor forces,
urban and rural, wereseparatenoncom-
peting groups. In linear programming
terms, therearetwo labor constraints and
each maywell havea d ifferent associated
shad ow wage.
Now, the essenceof our mod el is that
the two sectors areintimatelyconnected
through labor migration. If onead d itional
job is created in theind ustrial sector at the
minimumwage, the expected wage will
riseand rural-urban migration will bein-
d uced . In A ppend ix II it is shown that
more than one agricultural worker will
likelymigratein responseto thecreation
of onead d itional ind ustrial job. Hence, the
opportunitycost of an ind ustrial worker
will exceed themarginal prod uct of an ag-
ricultural worker. On theother hand , an
increasein agricultural incomewill ind uce
reversemigration with no d iminution of
ind ustrial output. Thus, the opportunity
cost of labor is lower to the agricultural
than to theind ustrial sector!
Theliteraturehas been strangelysilent
for themost part about thefull implica-
tions of using shad ow-wagecriteria. In a
static context, Stolper has pointed out that
financing subsid ies or losses of public enter-
prises gives rise to fiscal problems, but
unfortunatelythis issuehas not yet been
pursued in sufficient d etail.'3 If theproblem
is consid ered at all, theanalyst usuallyas-
sumes that a systemof nond istorting lump-
sumtaxes is available. Little, Lefeber, and
12
Hagen (p. 498) states, "a subsid yper unit of labor
equal to thewaged ifferential [between agricultureand
ind ustry] will increase real income further [than a
tariff] and if combined with freetrad ewill permit at-
taining an optimumoptimorum." Bard han (p. 379)
similarlyad d s. "Thebest remed yfor themisallocation
caused by a wage d ifferential is ... an appropriate
subsid y to the use of labor in the manufacturing in-
d ustry." It is important to recall that this argument is
d epend ent on variableproportions prod uction functions.
If prod uction coefficients arefixed , a wagesubsid ywill
have no effect in the short run. The classic statement
of this caseis byEckaus. Bard han explores its implica-
tions for subsid yin a d ynamic context. Both of these
papers, however, posit surplus labor in agriculture, an
assumption wed o not wish to makein an A frican con-
text.
13
Lefeber assumes that a wage subsid y can be
financed bya profits tax, whileother writers, e.g. Hagen,
Bard han, and Chakravarty never even consid er the
problem. Even Littleand Mirrlees who present an excel-
lent d iscussion of how to calculatea shad ow wagenever
mention thefiscalproblems of implementation.
This content downloaded on Sat, 19 Jan 2013 02:03:00 AM
All use subject to JSTOR Terms and Conditions
HA RRIS A ND TODA RO: MIGRA TION A ND DEVELOPMENI' 133
XM
T
-
0
FIGU$EE 2
Little and Mirrlees have pointed out that
in a d ynamic setting, the extra consump-
tion arising frompayment of the institu-
tional wage d iverts resources frominvest-
ment to consumption; thus some of the
foregone future consumption should be
consid ered in calculating theshad ow wage.
In our mod el, payment of the minimum
wage to ad d itional ind ustrial workers will
ind ucemorerural-urban migration. There-
fore, implementation of a shad ow-wage
employment criterion will have important
effects on the level of agricultural output
and on urban unemployment. The argu-
ment can be clarified with reference to
Figure 2.
The initial equilibrium, given the min-
imumwage, is at point Dwith output of
themanufactured good restricted to OXM*.
If ind ivid uals d id not migratein response
to expected waged ifferentials, theeconomy
could prod uct at point E, but migration re-
d uces agricultural output to thelevel OQ.
Thetheoryof shad ow pricing suggests that
with an appropriatewage subsid y (or
public-sector-hiring rule) the economy
could moveto point L on theprod uction
possibilities frontier which, with the pos-
ited social ind ifferencemap, is the opti-
mumposition. Welfarewould beincreased
froma level U1 to a higher level U4.
In thecontext of our mod el, such a point
is unattainable. Theeffect of implementing
a shad ow wagewill beto increaseprod uc-
tion of themanufactured good . But crea-
tion of an ad d itional job at theminimum
This content downloaded on Sat, 19 Jan 2013 02:03:00 AM
All use subject to JSTOR Terms and Conditions
1 4 THE A MERICA N ECONOMIC REVIEW
wage will ind uce some ad d itioinal migra-
tion (see A ppend ix II) fromthe rural
sector and therefore agricultural output
will fall.
HIence,
movement fromD can
only bein a northwest d irection. The line
DK in Figure 2 is the locus of all such at-
tainable points and it is evid ent that there
is onlyonepoint, K, at which therecan be
full employment of the economy's labor
resources. A t that point theexpected wage
will be equal to the minimumwage since
there is no urban unemployment. There-
fore, the marginal prod uct in agriculture
will have to be equal to the minimum
wage. But, with the subsid y, the marginal
prod uct of labor in manufacturing will be
lower than in agriculture, hence K lies in-
sid e the prod uction possibilities frontier.
(In the extreme case in which marginal
prod uctivity in agriculture can never beas
high as theminimumwage, K will coincid e
with T, the point of complete specializa-
tion in manufactures.) This situation will
certainly not meet the cond itions for a
general optimumwhich can bemet onlyat
L. Thus, implementing a shad ow wage
criterion to the point that urban unem-
ployment is eliminated will not generally
bea d esirablepolicy.'4
However, some level of wage subsid y
will usually lead to an improvement. In
Figure 2 it is clear that point J, with a
welfare level
U2,
will be preferable to D.
The criterion for welfare maximization,
d erived in A ppend ix III, is the following:
(d Vu
(12) f =Pq(
Note what this means. Creating one ad -
d itional job in the ind ustrial sector in-
creases output byf' but, since increased
employment will raisethe expected urban
wage, migration will be ind uced in an
amount d Nu/d NM. The right-hand sid e of
equation (12) states the amount of agri-
cultural output sacrificed because of migra-
tion. Thus the shad ow wage will be equal
to this opportunity cost of an urban job
and the amount of subsid y will be WM
-f'.
So long as f'> Pq' (d NU/d NM), ag-
gregate welfare can be increased by ex-
pand ing ind ustrial employment through
subsid y or public sector hiring. Clearlythe
more responsive is migration to ind ustrial
employment, the higher is the social cost
of ind ustrialization and the smaller is the
optimal amount of subsid y. In many A fri-
can economies it is likely that
d Nl\7/d NA m
exceed s unity. If so, it will be optimal for
the marginal prod uct of labor in ind ustry
to behigher than in agriculture and urban
unemployment will be a persistent phe-
nomenon so long as minimumwages are
set above a market-clearing level.
The d iscussion so far has ignored two
other ad verse effects of using a shad ow
wage. A s mentioned earlier, several writers
have noted that paymnent of a subsid ized
minimumwage to ad d itional workers will
increase total consumption, thereby re-
d ucing the level of resources available for
investment. If foregone future consump-
tion is positively valued , the opportunity
cost of ind ustrial labor will behigher than
ind icated in equation (12) and the shad ow
wage will be raised correspond ingly. Fur-
termore, wage subsid ies or public enter-
priselosses must befinanced and if revenue
cannot be raised through costless lump-
sumtaxes, the opportunity cost of raising
taxes must be consid ered . Both of these
effects will red uce the d esirable amount of
subsid ized job creation in the ind ustrial
sector.
It is interesting to note that this mod el
implies d ifferent opportunity costs of labor
to thetwo sectors. Whilethecreation of an
ad d itional job in the urban area red uces
14
A s shown in A ppend ix III, DK is not uniformly
convex. Therefore, K maybethebest attainable point
in some cases and the first-ord er cond itions may not
ensureoptimality. A s d rawn in Figure2, moving from
D to K represents a worsening of welfare, but this
clearlyis not a necessaryconclusion.
This content downloaded on Sat, 19 Jan 2013 02:03:00 AM
All use subject to JSTOR Terms and Conditions
HA RRIS A ND TODA RO: MIGRA TION A ND DEVELOPMENT 135
agricultural output through ind uced mi-
gration, ad d itional employment can be
generated in the agricultural sector with-
out red ucing manufacturing output. If
this phenomenon is not taken into account,
stand ard application of investment criteria
is likely to be biassed in favor of urban
projects.
B. Migration Restrictiont
A n alternative approach to the problenm
of urban unemployment is to physically
control migration fromthe rural areas.
Such controls have recently been intro-
d uced in Tanzania and have been used for
some time in South A frica.'5 Other coun-
tries, such as Kenya, are giving serious
consid eration to instituting such a policy.
A lthough we personally have grave reser-
vations about the etlhical issues involved
in such a restriction of ind ivid ual choice
and the complexity and arbitrariness of
ad ministration, it seems d esirable to in-
vestigate theeconomic implications of such
a policy.
Looking at Figure 2 it is obvious that
with the minimum wage such that in-
d ustrial output is OXm*, prohibition of mi-
gration in excess of the labor required to
prod uce that output will allow the econ-
omyto prod uceat point E. Themovement
fromD to E arising fromrestriction of
migration lead s to an unambiguous aggre-
gate welfare improvement provid ing ap-
propriate lump-sum red istribution is ef-
fected . Since such compensation is no-
toriously d ifficult to carryout in practice,
it will beuseful to examiinethewelfareim-
plications of such a mnoveon each of the
t\vo sectors in theabsenceof compensation.
Recall that the two sectors wered efined
to bea permanent urban groupand a rural
sector that prod uces both agricultural
good s and exports labor to the urban area
in exchange for wages in the formof
manufactured good s.'6 In Figure3 theline
T'S' represents prod uction possibilities for
theagricultural sector when labor export is
allowed . If its entirelabor end owment is
d evoted to agricultural prod uction, it can
prod ucea quantityOS'. However, byex-
porting its labor, the agricultural sector
can "prod uce" the manufactured good
(wages arepaid in theformof this good ).
Hencethis prod uction possibilities frontier
d epend s on market forces (wagelevels and
unemployment) as well as on purelytech-
nological factors. Theamount of agricul-
tural output foregoneif a unit of labor is to
be"exported " is its marginal prod uct; the
amount of manufactured good s obtained
bytheexported labor unit d epend s on the
wage, theamount of employment obtained
bytheexported unit, and its effect on em-
ployment of previouslyexported units.
In ad d ition to these prod uction pos-
sibilities, therural sector also has theop-
portunityto trad esomeof its agricultural
output with thepermanent urban sector
in exchangefor manufactured good s. Cor-
respond ing to each point on theprod uction
possibilities frontier T'S', thereis a d e-
terminatepriceof the agricultural good .
Themanner in which alternativeconstella-
tions of prod uction and trad eaffect the
16 SeeHarris and Tod aro (1969) for an analysis of the
T'anzanian program.
16
In consid ering the welfareof the rural sector as a
wholewearemaking thetacit assumption that thereis
red istribution of good s between ind ivid uals in this
sector. This is a very strong assumption. Yet there is
consid erable evid ence fromtropical A frica that em-
ployed urban migrants repatriatesubstantial portions of
their earnings to their kinsmen remaining
in the rural
areas and conversely that income both in cash and
kind is received byunemployed migrants fromkinsmen
remaining on thefarm. To theextent that theextend ed
family systemd oes red istribute good s between mem-
bers, this assumption maybetenableas a first approxi-
mation. A s Gugler (p. 480) has pointed out, it is appro-
priateto view theextend ed familyas maximizing its in-
come by allocating its members between agriculture
and urban wage employment. A lthough there is some
evid ence that growing numbers of urban workers are
settling permanently and grad ually eliminating rural
ties, it will be many years before such ties arecom-
pletelysevered .
This content downloaded on Sat, 19 Jan 2013 02:03:00 AM
All use subject to JSTOR Terms and Conditions
136 THE A MERICA N ECONOMIC REVIEW
XM
x
A
A O
OA XA X? XA XA S !
XA
FIGURE 3
sector's welfarecan beillustrated byFig-
ure3.
D' correspond s to theinitial unemploy-
ment equilibriumD (Figure2). A t that
point the rural sector as a whole "pro-
d uces" XA 0 and Xmo of thetwo good s. It
also has the opportunityto trad eat the
pricePO. By trad ing someof its agricul-
tural output to thepermanent urban sector
for ad d itional manufactured good s, it con-
sumes XA O XMO and achieves a welfare
level of UR. Restriction of migration re-
sults in thesector's prod ucing X]I XM '. If
it could still trad eat pricePO, the agri-
cultural sector would clearlybebetter off.
But this is impossible. A t E' (which cor-
respond s to E in Figure2), the priceof
agricultural good will fall to P' and with
trad ethebest consumption bund leattain-
able by the sector is XtA , XM which cor-
respond s to a lower level of welfareU0.
(Note that if p' d id not cut T'S' there
could beno incentiveto migrateat E'.)
It can beshown that Pq' (1- I/ ) (where
77
is thepriceelasticityof d emand for the
agricultural good ) is the amount of the
manufactured good sacrificed bytherural
sector as a result of removing oneworker
fromprod ucing the agricultural good
which could havebeen exchanged for the
manufactured good at the market price
1/P. This quantityis less than thevalue
of labor's marginal prod uct in agriculture
(Pq') sincethered uction in output has a
This content downloaded on Sat, 19 Jan 2013 02:03:00 AM
All use subject to JSTOR Terms and Conditions
HA RRIS A ND TODA RO: MIGRA TION A ND DEVELOPMENT 137
favorableterms-of-trad eeffect. If thed e-
mand for theagriculturegood is inelastic
(n <1)
we reach the
startling conclusion
that the sacrificebecomes negative! This
is, of course, thefamiliar proposition that
aggregatefarmincomemaybeincreased
by red ucing output. The d irect gain in
manufactured good s achieved bytherural
sector through exporting an ad d itional unit
of labor is
WMNM/N.,
the expected urban
wage. But ad d itional migration, by in-
creasing unemployment, red uces theearn-
ings of all migrants alread yin theurban
labor forcebya factor (1 -R), whereR is
thefraction of thetotal urban labor force
supplied bytherural sector.'7
A s long as Pq' (1-n) < WMNM/Nu (1-
R) thewelfareof therural sector will bein-
creased byallowing migration even though
unemployment ensues and theeconomyas
a wholesacrifices output. SincePq' and
WMNM/IN,
are
always positive
and R <
1,
ad d itional migration will always benefit
therural sector when X < 1. In general, the
lower is Pq',
t7,
or R and the higher is
WMNM/NU, the more will the rural sector
benefit fromtheopportunityto migrate.
Fromthe foregoing, one can conclud e
that although migration restriction will
improveaggregatewelfareof theeconomy,
given plausiblevalues of x7 and R, sub-
stantial compensation to therural sector
will be required if it is not to be mad e
worseoff byremoving theopportunityfor
free migration. The permanent urban
labor forceclearlywill bemad ebetter off
bybecoming fullyemployed at thehigh
minimumwagewhilealso being ableto
buyfood at a lower price. Each unit of
labor exported by the rural sector will
similarlyearn morebut this gain will be
offset byred uced total labor exports and
lower agricultural prices. Whether or not
this will betrued epend s, of course, on the
values of the specific parameters of the
economy. If
?7
is sufficientlyhigh, therural
sector could be mad e better off by re-
stricting migration in theabsenceof com-
pensation, but this seems veryunlikely.
C. A Combination of Policies
It has been shown that either a limited
wage-subsid y or a migration-restriction
policywill lead to a welfareimprovement.
Which of thetwo policies will lead to the
better position cannot bed etermined with-
out knowing all the relevant parameters
for a particular economy. It is d ear, how-
ever, that neither policyaloneis capableof
moving theeconomyto theoptimumthat
could beachieved with competitivewage
d etermination (point L in Figure2).
A t first sight it mayseemstrangethat
with a single market failure, the wage
level, a singlepolicyinstrument is unable
to fullycorrect thesituation.18 Thereason
is that thewageperforms two functions in
this mod el. It d etermines both thelevel of
employment in the ind ustrial sector and
theallocation of labor between rural and
urban areas. Whilea subsid ychanges the
effective wage for d etermination of in-
d ustrial employment, so long as thewage
actuallyreceived by workers exceed s ag-
ricultural earnings therewill bemigration
and urban unemployment. Restriction of
migration prevents the minimumwage
having its effect on unemployment but
d oes nothing to increasethe level of in-
d ustrial employment. Therefore, if the
optimumposition is to beachieved , a com-
bination of both instruments will haveto
17
If theurban unemployment wereexperienced only
bymigrants, this termwould equal zero sincethetotal
amount of earnings through labor export would becon-
stant. It can be positive only because the permanent
urban labor force shares in unemployment, thereby
red ucing its shareof theconstant wagebill in themanu-
factured good ind ustry. A n interesting extension of the
mod el would be to incorporate d ifferent employment
probabilities for thepermanent urban and migrant rural
labor forces and then to check thesensitivity of results
with our more simplified assumption of
equal prob-
abilities.
"I
Wewish to thank a refereeof this Review for d raw-
ing this to our attention.
This content downloaded on Sat, 19 Jan 2013 02:03:00 AM
All use subject to JSTOR Terms and Conditions
138 THE A MERICA N ECONOMIC REVIEW
beused . In ord er to reach point L a wage
subsid ymust beinstituted such that in-
d ustrial employment will increaseto the
extent that with full employment themar-
ginal prod uct of labor will be equal in
manufacturing and agriculture. The sub-
sid ywill bepositiveand equal to thed if-
ferencebetween theminimumwageand
marginal prod uctivity. A t that point W.
= WM and WM>Pq'. Therefore, ind ivid -
uals would still find it in their interest to
migrateand thepoint will not beattain-
ableunless migration is restricted .
Theagricultural sector has to bebetter
off at L than at E sinceeach ad d itional
unit of labor exported earns thefull min-
imumwage, marginal prod uctivityin ag-
ricultureis less than theminimumwage,
and thepriceof theagricultural good rises.
Whether the agricultural sector is better
off at L than at D, however, d epend s again
on theparametric values of themod el."9
It can bestated with certaintythat the
amount of compensation need ed to make
therural sector no worse
ofF
than at Dwill
beless at L than at E, and , furthermoreit
should beeasier to financesincetotal in-
comeis greater.
Even so thefiscal requirements of sub-
sid y(or public enterpriselosses) and com-
pensation cannot be taken lightly.20 A
government mayfind it d ifficult to find
nond istorting taxes capableof raising suf-
ficient revenue. Perhaps a head -tax on all
urban resid ents would befeasiblealthough
this too raises the question of how min-
imumwages are set (unions in tropical
A frica have, in some
cases,
successfully
fought to maintain thereal after-tax wage).
A tax on rural land is ruled out if there
must be net compensation to the rural
sector which, in theabsenceof pureprofits
in manufacturing, leaves an urban land
tax as theremaining potential id eal tax.
A ll of theabovesuggests that altering
theminimumwagemayavoid theprob-
lems of taxation, ad ministration, and inter-
ferencewith ind ivid ual mobilityattend ant
to thepolicypackagejust d iscussed . In-
comeand wages policies d esigned to nar-
row therural-urban wagegaphavebeen
suggested by D. P. Ghai, and Tanzania
has formallyad opted such a policyalong
with migration restriction. In the final
analysis, however, thebasic issueat stake
is reallyoneof political feasibilityand it is
not at all clear that an incomes policyis
anymorefeasiblethan thealternatives.
A PPENDIX I
Proof of Stabilityof Unemployment
Equilibrium
In ord er to provethat our urban unem-
ployment equilibriumis stable, we can
d ifferentiate
Vt
(equation (9)) with respect to
Nu remembering that d NA , = - d NA ac-
cord ing to (7). Wethereforeobtain
d Nu ,
WrNWMNM
-N-= ()
- +
Pq
(1.1)
d N
A uJ
aXA ]
Stability requires d NR/d N.<O which is
satisfied if
WMNM
-
Pqt
ap
(Nu))2
d XA
(q)2
l9 A s d rawn in Figure2, L must represent a higher
welfarelevelthan Dfor theruralsector sinceP rises
and thesector prod uces moreof both good s. In fact if L
lies along TS north of theraygoing through Dthere
willbean unambiguous sectoralwelfareimprovement.
However, if L lies south of therayon TS, theruralsec-
tor could beworseoff than at DsinceP falls.
20 This argument coincid es with thestatement by
Stolper (p. 195), "It should benoted , however, that
even at best theapplication of shad ow prices lead s to the
substitution of oneproblem, thebud get, for another one,
an imperfect market."
Wewould not go as far as Stolper in rejecting out of
hand anyuseof shad ow priciIng becauseof thefiscal
implications. Thegeneralpoint is valid that onecannot
d isregard the consequences of implementation of
shad ow-pricecriteria if actualprices or wages continue
to d ivergefromtheshad ow prices or wages.
This content downloaded on Sat, 19 Jan 2013 02:03:00 AM
All use subject to JSTOR Terms and Conditions
HA RRIS A NI) 'TODA RO: MIGRA T1ON A ND DEVELOPMENT 139
The right sid e of this inequality is un-
ambiguouslypositivesinceq" <0. Henceour
assumption that d P/OXA <O Will ensure
stability and , ind eed , is stronger than
necessary. The ad justment mechanismmay
be mad e clear by the following phase
d iagramin, which thefunction i is plotted .
Its positive slope reflects the hvpothesis
that migration flows will increasewith the
magnitud eof theurban-rural expected wage
d ifferential. In Figure4,
4t'
is plotted und er
the assumption that
4t'O)
= O, hence the
horizontal intercept is at the origin (in
general theintercept would bea). Further-
more, wehavearbitrarilyassumed that t,6 is a
linear function. The arrows show the d irec-
tion of ad justment in accord ancewith (1.1).
If WlmNM/Nu-Pq'>O, then Nu>O but we
know that if A U> O, the expected wage
d ifferential will d ecreasesince
d N9/d N,<0.
A d d itional migration by increasing N,,
without affecting Nm1
will red uce the ex-
pected urban real wage through increased
unemployment. Coiicomitantly, the trans-
fer of labor out of agricultureraises q' and
red uced agricultural output also causes P to
rise. Thus migration red uces the expected
waged ifferential to zero and equilibriumis
achieved when thereis no further incentive
for migration. See Tod aro for a more d e-
tailed analysis of this process in a d ynamic
setting.
NL;~~~~N
i0 (Gk 4)
FIGu2E 4.
A PPENDIX II
Differentiating the equilibriumcond ition
(8) with respect to
NM,
recalling that d N,u
--d NTA , weobtain theexpression
d N,,
d JA T,r ,
WA I
f
- qp
_
N.
X
WMNM ,,p"
q,, gXM
Vu
pq
q
pXA 2
Defining the elasticity of d emand for the
agricultural good as
. aXA
P PXA
(112)
A q
= - _
-_
-_
Cl PA XA P'X
(1I.1) can be rewritten as
tVM pq'f'
d Nu NuT 77A XM
(11.3) U _ ___ _ _ _ _ __ _ _ _ _ _
d NMTm
WMNM
- pq
p(q" )
A TU7z2 'qA XA
Differentiating the expression partially
with respect to its various arguments it can
be shown that d N,/d Nm
will vary d irectly
with WM, NM, r1A and inverselywith p,
q',
fI, NU,
and
q".
In
general,
the
greater
is the
urban-rural wage d ifferential, and the less
sensitiveareprices and marginal prod ucts in
agriculture, thegreater will bethemigration
ind uced by creation of an ad d itional job. If
the minimum wage exceed s agricultural
earnings, (11.3) will generally be positive
and , with parameter values relevant for
manyA frican economies, will exceed unity.
When d Nu/d NM> 1, creation of an ad d i-
tional job at the minimumwage will in-
crease the absolute level of unemployment
although the rate of urban unemployment
will have to fall. This can beseen by con-
verting (11.3)
to an elasticity measure.
d NNU Ni
(11. 4) -*-c
d Nm
N,
This content downloaded on Sat, 19 Jan 2013 02:03:00 AM
All use subject to JSTOR Terms and Conditions
140 THE A MERICA N ECONOMIC REVIEW
W'MN,M NMpqf'
\TU2 NA 7'A qXM
<1
WMNM ,q
p(q
+
)
Nu t1A XA
since q" <0.21 To give an example of what
this means, suppose that ani economy ini-
tially has an urban unemployment rate of
25 percent. If in responseto thecreation of
100 ad d itional ind ustrial jobs, 125 ad d itional
ind ivid uals migrate to the urban area, the
absolute number unemployed increases by
2 5 although the unemployment rate will
d rop, sincethemarginal unemployment rate
is onlv 20 percent.
A PPENDIX III
If minimumwages are maintained and
migration takes place in accord ance with
equation (8), aggregate welfare will be
maximized if the following Lagrangean ex-
pression is maximized :
Q
= U
(XA , YXM)
+
Xl[q(N
- Nu) - XA ]
+ X 2[ f(IT)
-
XMI
q
(
-
)A T
*qA T r1T,) - V }___M
where Uis the social welfarefunction and
the succeed in(g terms are the constraints
imposed by equations (1), (2), and (8) (re-
call that NA =N-N-N fromequation (7)).
Maximizing (111.1) wveget the following
first-ord er cond itions:
d Q~ auI
(IIT.2) = -A l = 0
aXA aXA
O?
3iU
(III.3) =
-X2
-
0
aXM aXM
Ou
-
I~
fq'
(III.4) aN = -
Xlq' + L3' q2
- pq,
+ WmNm ]
0
(111.5) NM
-X2f/
+
jif
X ]
WM
=
and thed Q/OXi =0 (i= 1, 2, 3) which ensures
that the constraints hold .
Substituting (111.2) and (111.3) into
(III.4) and (III.5) we get
(III.6)
-
C9XA
1'VM ,,f'
aq
qf~~~~~
Nu q
WMNTM fq,
-N
2
-q +P
2
We know that in equilibrium (aU/aXM)/
(OUIOXA ) =1/P and it has been shown in
A ppend ix II that the right-hand sid e of
(III.6) is equal to d N,/d Nm. Therefore(III.6)
can berewritten as
(III.7) f = Pq' d N
d NM
which is the cond ition used in the text to
d eterminethe optimal wagesubsid y.
Cond ition (III.7) can also bewritten as
-f'
d XM
_p ==
(III.8) I
d N, d XA
q
d Nm
d A TM
Weknow that - P is equal to themarginal
rate of substitution between the two com-
mod ities and d XM/d XA is themarginal rate
of transformation. Hence (III.8) states the
familiar cond ition for optimality: equate
marginal rates of substitution and trans-
formation. d XM/d XA is theslopeof theline
DK in Figure2 and it clearlywill benega-
21
Wearegrateful to Peter Diamond for d eriving this
expression.
This content downloaded on Sat, 19 Jan 2013 02:03:00 AM
All use subject to JSTOR Terms and Conditions
HA RRIS A ND TODA RO: MIGRA TION A ND DEVELOPMENT 141
tive. However, its d erivativewith respect to
NM,
(1I1.9)
d
(d -X)
d Nm
d Nu d N)
d 2Nu
-q f ~ I q'l+f'q'
d NM d NM d NM2
{,d N u
2
(q d NM)
is of ind eterminatesign sincef", q"<0 and
d 2N./d NM2
will generally be negative as
well. (III.9) must bepositive if the effective
prod uction possibilities frontier (DK) is to
beconvex, a cond ition that is likelyto hold
but the possibility of concavity as full em-
ployment is approached must beconsid ered .
Theslopeof DK in Figure2 seems plausible
on a priori ground s.
REFERENCES
P. K. Bard ham, "Factor Market Disequilib-
riumand the Theoryof Protection," Ox-
ford Econ. Pap. (New Series), Oct. 1964,
16, 375-88.
E. J. Berg, "WageStructurein Less Devel-
oped Countries," in A . D. Smith, ed ., Wage
Policy Issues in Economic Development,
Lond on 1969.
A . Callaway, "FromTrad itional Crafts to
Mod ern Ind ustries," ODU: Universityof
Ife Journal of A frican Stud ies, July 1965,
2.
S. Chakravarty, "TheUse of Shad ow Prices
in ProgrammeEvaluation," in Rosenstein-
Rod an, ed ., Capital Formation atnd Eco-
nomic Development, Lond on 1964.
Y. S. Cho, Disguised Unemployment in
Developing A reas, with Special Reference
to South Korean A griculture, Berkeley
1960.
R. S. Eckaus, "TheFactor-Proportions Prob-
lemin Und erd eveloped A reas," A mer Econ.
Rev., Sept. 1955, 45, 539-65.
J. Erickson, "Wage Employment Relation-
ships in Latin A merican Ind ustry: A Pilot
Stud yof A rgentina, Brazil, and Mexico,"
International Labour
Office, 1969, type-
script.
J. Fei and G. Ranis, Development of the
Labor Surplus Economy, Illinois 1964.
D. P. Ghai, "Incomes Policyin Kenya: Need ,
Criteria and Machinery," East
A fr.
Econ.
Rev., June1968, 4, 19-35.
J. Gugler, "TheImpact of Labour Migration
on Society and Economyin Sub-Saharan
A frica. Empirical Find ings and Theoret-
ical Consid erations," A frican Social Re-
search, Dec. 1968, 6, 463-86.
E. E. Hagen, "A n Economic Justification of
Protectionism," Quart. J. Econ., Nov. 1958,
72, 496-514.
J. R. Harris and M. P. Tod aro, "Urban Un-
employment in East A frica: A n Economic
A nalysis of Policy A lternatives," East
A fr. Econ. Rev., Dec. 1968, 4, 17-36.
and , "Wages, Ind ustrial Em-
ployment, and Labour Prod uctivity: The
Kenyan Experience," East A fr. Econ. Rev.
(New Series), June1969, 1, 29-46.
J. P. Hend erson, "WagePolicy in A frica,"
Paper prepared for d eliveryat theA frican
Conferenceon Economics, TempleUniver-
sity, mimeo, A pril 1968.
C. R. Hutton, "TheCauses of Labour Migra-
tion," in Gugler, ed ., Urbanzizationt in Szub-
Saharan A frica, Kampala 1969.
C. H. C. Kao, K. R. A nschel, and C. K.
Eicher, "Disguised Unemployment in A gri-
culture: A Survey," in C. K. Eicher and
L. W. Witt, ed s., A griculturein Econzomzic
Development, New York 1964, 129-44.
J. M. Katz, "Verd oorn Effects; Returns to
Scale, and the Elasticityof Factor Substi-
tution," Oxford Econ. Pap., Nov. 1968, 20,
342-52.
L. Lefeber, "Planning in a Surplus Labor
Economy," A mer. Econ. Rev., June 1968,
58, 343-73.
W. A . Lewis, "Economic Development with
Unlimited Supplies of Labour," M1anchester
Sch. Econ. Soc. Stud ., May1954, 22, 139-
91.
I. M. D. Little, "The Real Cost of Labour,
and the ChoiceBetween Consumption and
This content downloaded on Sat, 19 Jan 2013 02:03:00 AM
All use subject to JSTOR Terms and Conditions
142 THE A MERICA N ECONOMIC REVIEW
Investment," in P. N. Rosenstein-Rod an,
ed ., Pricinzg and Fiscal Policies: A Stud y
in Md ethod , Cambrid ge1964, 77-91.
- and J. A . Mirrlees, Manual
of
In-
d ustiial Project A nalysis, Vol. II, "Social
Cost Benefit A nalysis," Paris 1969.
B. F. Massell and J. Heyer, "Household
Expend iture in Nairobi: A Statistical
A nalysis of Consumer Behaviour," Econ.
Develop. Cult. Change, Jan. 1969, 17, 212-
34.
L. G. Reynold s, "Wages and Employment in
a Labor-Surplus Economy," A mer. Econ.
Rev., Mar. 1965, 55, 19-39.
W. F. Stolper, Planning Without Facts: Les-
son's in ResourceA llocation fromii Nigeria's
Development. Cambrid ge1966.
M. P. Tod aro, "A Mod el of Labor Migra-
tion and Urban Unemployment in Less
Developed Countries," A mer. Econ. Rev.,
Mar. 1969, 59, 138-48.
Nigeria, Report of the Commission on the
Review of Wages, Salaryand Cond itions of
Service of the Junior Employees of the
Governmnents of theFed eration and in Pri-
vateEstablishments 1963-64.
This content downloaded on Sat, 19 Jan 2013 02:03:00 AM
All use subject to JSTOR Terms and Conditions