ECONOMIC DEVELOPlMENT AND THE LABOR MARKETIN MEXICO

By

Peter Gregory

Series: Studies in Employment and Rural Development No. 70

Division: Employment and Rural Development

Department: Development EconomicsDevelopment Policy StafLInternational Bank for Reconstruction and Development

This paper was prepared by the author as part of a larger

study of RPO 671-84 and is to appear in a volume of contributed papers

in honor of Professor L. G. Reynolds, Y=ale University.

The views expressed are those of the author and the report may

not be published or quoted as representing the views of the World Bank

or its affiliated organization.

Washington, D. C. August 1981

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

aL

in X'ICxic,

I. Introduction

The period of' acceleratcd growch and modernization in !eIxico dates

fromi about 1940. The groun6-:ork for change had alread'y been laid by the

sweeping changes flowingr from thc. preceding thirty years of revolutionary

activ ity that produced far reaching social and institutional chaniges. The

onset of o'Drld '?ar II brought withi it a disruption of .raditional sources

of supply for industrial goods and stimiulated the e::pansion of manufacturing

production for intetnal consumnDtion. The large-scale land distribulion

prograz= of the Cardenas regime during the preceding decade resulted in

a substantial incrci-Se inr the land under cultivaLion andn ila agricult-ural

output. Thus, t:he decade saw all of the principal s-ct,ors of the economy

growing in tandem fashion. As is evidenced in Table 1, the rates of grou't h

of thie three broad sectors showed only small differences.

Followzing the end of t1he war and a return to more competitive condi-

tions in producL v=:ar'.kets, the government opted for a policy of fostering

the furtlher growth of domescic industry behind protective tariff walls

and import controls. It also embarked on a l.arge-s3c1L investmcnt pro-

gram to provide infrastructure for an eoerging :Mc "errn com-irrcial a-r-Icul-

tural sectcr. Large-scale irri-ation and rond-building projects in the

sparsely populated nor:li anCd northwestern provinces prcided a stimulus

for a continuc., e:epansiocln of a'-ricultur-nl output for both riontestic con-

(Millions of pesos of 1950)

1940 1950 1940-50 1960 !970

Sectors 1I cNr GNP Annual Pate N P CNp

of Clhange

PPr ra r:

Bankr. of N'2xico 11,433 2.2 16,473 3. 7

UnIkel 5,221.0 9,242.0 5.9 13,916.61 4.2 17,606.77 2.4

S--Con da r

Br.nL of eico 20;956 5.3 48,727

Uni' e1 6,788.58 12,466.0 6.3 24,603.00 7.0 52,203.2 7.8

TercIa!]y

Bank: of Mexico 39,404 7.4 76,278 6.8

un 1k;el 10,930.80 19,352.13 5.9 35,695.57 6.3 S2,',30.62 e-

'rc.t ,.1

}ianrk of Mexico 71,793 5.7 141,478 7.0

Unikel 22,940.38 41,060.13 6.0 74,215.18 6.1 152,250.59 7.5

1/ The sectors are composed of the following sub-sectors: Primary: Agriculture, cattle, forestry, fishing.

Secondary: Manufacturing, construction, mlining, petroleum, electricity. 'rertiary: Transportation, Co1i-

!:.U:dcations, comirerce, govcrnmeflt, othler services.

Source: L. Unikel, El Pesarrollo Jrbano de Miexico, 2nd Edition (Mlexico: El Colegio de Mtexico, 1978) ,.T bles

VI-k5 to A8. The revised GNP estimrates originatte witi the Banco de tixico as reported by C. Reynolds,

"A Shift-Share Analysis of Regional and Sectoral Productivity Growth in Contemporarry Mexico,"

January 1979 ((mimeo), p. 27.

L .

sumption and Lor export. I,:ith the e::hau23tion of "casy" import substitutiou,

incentives tw!ere adopted for domestic producti"on of consum.er durable goods and

thcenforintcrm.lliate and capital goods. Throughout .,ost of the past four decades,

Mfexico succeceled in maintaining a high and stable rate of grow.th, one that

faltered significantly only during thle middlO of the past decade.

While the acheivernents of the MNexican economy are surely impressive,

nevertheless, there remain areas of concern. A rapidly growing population

of labor force age poses a formidable challenge to the economny's ability to

absorb it in productive employment. The pattern of development, particularly

in agriculture, has been dual. While a highly productive cornmercial agricultural

sector has emerged, the bulk of the rural population is to be found in areas 'with

scarce land and water resources and in which increases in productivity have been

modest. Rural poverty remains a serious problem; some students of Mexico hold

that the incomes of a sizable rural population have not shared in the general

increase that characterizes the country as a whole, or have shared only to an

insignificant extent. The relative distribution of incomes continues to rank

among the most unequal in Latin America and appears to be emerging as an acute

social and potitical issue. The dualistic character of agriculture is reflected

in the large differences bet.een the productivity of land and labor among the

various classes of agritultural holdings. In 1970, the output per hectare on

farm units cmploying "modern technology1' was four times that in subsistence

farm units (those which destined only a small part of output for markets).

The ratio of output per worker was even greater, or the order of 20 to one. 1/

1/ E. Hernandez Laos and J. Cordova Chalvez, "Estructura de 11 Distribuciondel In-reso oi Mexico," Comorcio _:L:terior 29:5 (M1a) 1979, p. 506.

-4-

The dual clia ruc tr of thle sector, -,oreCov'r, has stron, re-ional asscciations.

In 1970, for exampie, the output per hectare ,;is 20 times greater in Sonora

thani in Quinti<ana Roo. ln 12 other states, the productivitLy of land was no

greater thain one fourtth that of Sonora. 2/ SLic'i large differences in the

productivity of diffcrenit classes of far= units and regions may be *expected

to give rise to similar differences in agricultural inco.es.

Intrasectoral differences in productivity ard earnings among productive

units exist in other sectors, as might be expected, though they are not as

pronounced as those indicated in agriculture. In 1975, value added and total

remunerations per worker in manufacturing establishments employing over 750

workers were 3.8 and 3.1 times greater respectively than in small shops with

5 or fewer employees. 3/ In the service sector, the differential in the rate

of earnings between small (5 or fewer employees) and large (over 50 employees)

establishments was more modest, on tthe order of 2.1 to 1 in 1975. 4/ In

1970, the differential in earnings between small and large establishments in

cb.merce was slightly narrow.er at 1.9 to 1. 5/

2/ L. Uniel>e, El Desarrolo 'Urbane dIe e ico (Me-xico Citv: El Coleoio deMexico, 1976), .pp)endix. Table VI--A1S.

3/ Secrataria ie- Pro'gr-!nacion y vPrcsupuest6 Censo Industrial. 1976(Datos de 1975) (Mlexico: 1979) Vol. 1, TablFe 8, P. 167. %'hen the 1975census provides data on remuneracions and total e-ployM ent by establishmentsize, it does not separate out the unre: unerated component of totalemploynent oni a siriilar basis. Trie lacter component w.as estimated on thebasis of miore cot:iplete data in earlier censuses by applying the pastratio of unro7%unc.±rated personnel per establish,;cnt in each size class tothe 1975 est3blishments of equivalent sizes.

4/ SecretaLr a de Industria y Comercio, Direccion General de Estadistica,Censo dI Sctvicios, 1976 (Daros die 1975) (Mc>ico: 1977), Table 17, p. 305.

5/ Sccretar•al do IndLstria y Cocmorcio, Direccion General de Estadistica,VI CLIso Corcial 1 1971 (Datos de 1970) ('M:cx.ico: 1975), Table 22, p. 403.

More Impressive than the differences in productivity and carniings within

cach of t!;e tion-ar'ricultural soctors enuimrated here arc the differences

betwcen thoem and a-riculture. In 1970, value added por worker in the former

was 5.2 times gre.itcr than in the latter; by 1976 it is estimated to have increased

to 6.6. 6/ It is no wonder, then, that most of the poverty in Mexico is to be

found in the rural areas. On the basis of the 1975 incomie and expenditures

survey, it has been estimated that approximately 52 percent of the poor families

were in the agricultural sector. Stated inaother terms, 76 percent of all

families in the sector lived in poverty. 7/ The prevalence and durability of

large urban-rural income differentials are likely to have been important factors

underlying a large rural-urban migration as well as a tendency of rural residents

to seek seasonal employment in urban labor markets. IJnikel. estimates that

4.5 millionpersons moved from rural to urban areas between 1950 and 1970. 8/

There is a widespread recognition of the fact that many Mfexicans are pre-

cariously employed in jobs of uncertain duration and at low levels of productivity

6/ L. Gomez Oliver, "Crisis agrlcola, crisis de los com.pesinos, " ComercioExterior 2S:6 (June, 1973), p. 721. The estimated increase in the ratiomay be exaggerated, for the labor fOrce size assumed for the agriculturalsector would seem to be larger than is lik-ely.

7/ Centro Nacional de Informacio'n y Estadisticas del Trabajo, Encuesta deIngresos v Castos nimiliarcs, 1975 (Mexico, D.F..: 1977) The poverty linein Mexico is dravni at a househ.old income equal to haLf the estimated nationalmean, which, in 1975, would have been \rcx. S1,621 per month, equivalent inpurchasin, pDover to approxi:nately US $1,315 per e.ar per family. Thesefigurcs shiould be taken as being only roughly i.ndicative. The 1975 surveyappears to have underestimated severely the incomes of some groups, largelyin the form of incomee.in kN-ind. The results of thie 1977 surveY, which.arethouogit to be of bettor quality, are not yet available by sector of employmentof tlhe hlead of household.

8/ Unikel, on., cit.,p. 213. In 1950, thlC ppulation mensured 26.5 nmillions.B) 1970 it hiad g,,ruvwi to 50.5 millions.

-6-

anl wae,s. ',That is disturb ing, howecver, i; thle belief prcxn.illng a.mong

g(wOvrru.mnt officials, scholars, and oLher o' scrvcrs sthat L.tzo ri, cnt cor-

ditions are not o;ll,, "poor" but that chay hz%e deterioraited over time.

'A I;vans and James conclude a.fter surveayng a number of cm.plov; cnt studies,

t!( re is cbund;ant evidence that Mex.:ico ... is plag-e'd with a serious and

Wort;ening labor surplus problem." 9/ Unempl.oym-.nt anda undere=rloyment are

SQ'.rri as having increased, with the lattcr variously estimated as characte-

rizIng from 25 to 52 percent of the labor force during the early 1970s.

Slhouild this perception of deterioration prove, in fact, to be correct, then

tihe prospects for improving the material well-being of the bulk of the

world's population in the less developed countries of the workd would seem

grilm indeed. Afte:r all, if under the conditions of sustained grc.wth

achieved in 'Mexico, gencral irprovec-.ent in eziploymenet conditions cac&io; be

attnined, then whal: hope can be held out for the workers of most LDCs that

have achieved much less imoressive rates of growth?

While one finds a widespread acceptance of the preposition that de-

terioration has occurred, it is harder to find convincing empirical support

*92/ J.S. Evans zend. D.D. James, "Conditions of Employment a-nd Income Distri-bution in Mexico as Incent ives for Mexican 1.igration to the United States:Prospects to the End of the Centurv." T-ternicior,21 M'igration Rev%.iew 13:1(Spring 1979), pp. 4-24. For aooraisais of em-ployment conr-itiors, seealso Secretaria de Trabajo y Pravisidn Social, 1 ra r-.a_Nacional deEmpleo, 1:930.-'), Vol. 1, ?rcscn.acion -v Da`Ln''nosctico (--xico, D.F. N Nov-ember, 1979); Snali rzejo R. ,:'L .: yo subocu-.acijon en "M;xico," ComcercicExterior. 22:5 (M,'ay, 1972), pp. 411-16; Saul Trejo R., "II desempleo an1I1cxico: charactoristicas g.;rales," Co7..rcio :.!:t:crior 24:7 (Jul,, 1974)pp. 730-3S; Pro-r.,:7a 1Regional dcl mplolco 1''ara A.merica Lati-na y el Caribc,The Frnlovmcnt Prni',lcm in Latin .-*.:ericn: FPcts., Otutloo:s and Policies(Santic ,o C l lt>); upo Ge -t;uto dl Probl,.:-a del Emplco, ElProb] cma Cc ,;:'cl('i.'on:il c-n :' cc : i ttud' v VLc :nr:vJ;c jones, Versio'n Prc-15.riL;,ar (1':cioo, . c.: COuO do CL:ira i.cci.u:_;ica, 19734)

for it. Most of the appraisals of the ;Mexican labor market focus on its

profile at a mom:ent of tinie and conclude fro;m the "unsatisfactory" state of

employr.cnt that detcrdorat4cn lhas occurred. Deterioration, howev.er, implies

change over timc, change wlhich cannot be inferred from conditions at a moment

in time or even from a cursory review of labor force aggregates in isolation.lO/

I sense a lamentable nbsence of analysis of the evolution of the labor market

over time inits several dirmcnsions, and of how it has responded to and faci-

litated the notable growth and change observable in the level and structure

of output. Almost completely neglected is the course of wages over time,

sometlhing that would appear critical to any evaluation of the changing

quality of employiLent. This essay represents an attemopt to redress these

shortcomings of most current appraisals. of employment cond *itions in Mexico.

It repcrts on some preliminary findings of a larger effort to analyze the

evolution of the Mexican labor mar'ket and focuses on several aspects that

may be indicative of the nature and degree of qualitative change in em-

ployment conditions over time. In the sections that follow I will disc'uss

the changing structure of the labor force, the course of sectoral preducti-

vities and wages, and measures of underutilization of the labor force.

10/ An example of the latter is the conclusion dra;..'n from reviews of thechanging sectoral distribucion of the labor force that there has beenancver e::pansion of tertiary sector emplcovmlent. I'he criteria employedin reachin?; the coniclusion are not made exnlicit nor are the data dis-aggre-ated in a way that might indicate support for it. This questionis addressed in the section that follows.

II. Structure of the Labor Force, Productivity, and WT!its

Ilistorical analyses of the 2M1exican lTbor force rely on tle decennial

censuse; of population as the principal data source, for, until recently,

these have provided thc only comprehensive measures of the labor force.

Ideally, one would hope that the censuses would provide a reliable data

base for tracing the cvolut..on of Lhe labor force in the process of economic

development, charting faithfully the course of labor force pacticipation

rates, the shifts that occur in employment among sectors and occupational

groups and in the emplovment status of the labor force. Unfortunately,

the Nlexican censuses do not provide an unambiguous basis for assessing the

changes that have occurred. In part, this is the restilt of changing classi-

ficatioiis of the labor force over time that obscure the true e-xtent of

shifts or of changing practices in the classification of certain population

groups within the labor force. A substantial problem is posed by the coding

and processing errors that resulted 'in the reporting of an outsized aari-

cultural and total labor force in the censuis of 1960. Further difficulties

are posed in the interpretation of the 1970 census by the very large number

of individuals in the labor force who could not be assi;gned to an economiic

sector of employrmnn. The widesprcad awareness of these slhortcomings has led to

several attempts to adjust the data wuith rather different results. i1/

11/ Colc,io de Dc:xico, Din.icn do l2 Polhlacion oii Mc':ico (MXexico, D.F.,1970); 0. Altiniir, "La '':cicion dLe la Poblacion Eco'Lomic2mente Activa deNtexico," DCo- (:rafia v I Econcen•, VIII :1 (1974), pp. 50-83; Earl L. McFarland,Jr., "Service E:pIcN\.nent in Mc'xico," (roimeo).

-9-

For various reasons, I have concluded that the adjustments proposed by

Oscar Altimir are more persuasiv. thlan those of others and, therefore, rely

heavily on his data. 12/

Table 2 presents the adjustcd distribution of the labor force over the

principal econcmic sectors based on the population censulses dating from

1940 to 1969. 13/ The changing structure of the labor force is a reflection

of the pattern of ecornomic development since 1940. The 1040s were a decade

of rapid growith in output, in the labor force as a whole, and in employment

in each of the principal sectors. Over the follow.ing two decades the in-

dustrial and construction sectors retained tlheir role as "leading" sectors

in the growth of employment with rates well in excess of the growth of labor

force. The commercial sector show7s declining rates of growth in each succes-

sive decade, while the service sector shows a decline and then a recovery in

its rate of growth. According to Altimir's adjusted census figujres, the

*growth of the agricultural labor force over the 1950-69 period amounted to

less than one-half of one percent per year. In my opinion, this is likely

to understate somewhat the increase in the agricultural labor force that

12/ A full discussion of the merits and demerits of the various approaclheswill be included in a forChcoring researchl report on the M1exican labormarket in the process of prcparation .or the Rural Development and Em-ployincnt Division of The I!orld Bank.

13/ The 1970 census distributed the labor force on the basis of the prin-cipal employment lheld during 1969.

Table 2: SECTORA, DISTRIBUTION OF THE LABOR FORCE

1940-1969 AND RATES OF CHlANGE

Labor Force (in thousands) Rates of Chanae (Percent)Sector 194(1 1950. 19c( 1969 1940-:$i) 1950-()0 19(-60 19!^0-69

Agriculture, forestryand fishing 3,832.4 4,864.9 5,048.3 5,292.7 2.4 0.4 0.5 1.1

Manmufacturing, electricityand extractive 709.7 1,237.5 1,760.3 2,829.1 5.7 3.6 5.4 4.9

Co;structicn 115.9 263.8 414.2 609.8 8.6 4.6 4.4 5.9 c

Commerce and fJnance 413.7 732.6 1,083.4 1,397.0 5.9 4.0 2.9 4.3

Other services 786.7 1,246.4 1,906.7 2,826.5 4.7 4.3 4.5 4.5

Total 5,858.5 8,345.2 10,212.9 12,955.1 3.6 2.0 2.7 2.8

Sources:, 1950-1969, Altimir, on. cit, Table i6. Altiniir's distribution of the laboi, force includes his redistributionof the unclassified workers over the several sectors.

1940, L. Unikel, El Desarrollo Urbano de Mexico, 2nd Edition (.Mexico:. El Coleglo de Mexico, 1978)Table VI- A9.

occurred buteccn 1960 and 1969 relat.Wc tLo 1950. ?':l Hlowcvor, I do not be-

lieve that a more accurate measure of the agricul.tural labor force wouldl

modif) si,-nificantly the relative importance off the sector as a source of

employmenit.

The growth of cLte tertiary sector has been a source of concern to some

who view it as representing an "undue" extansion of very low-productivity

employmentsnn 15/ The conventional "explanation" of this growth is posed

in terns of a burgeoning labor force incapable of finding satisfactory wage

employment being "forced" into a marginal employment and productive status

in the subsectors of easy entry in conumerce and services, often in the

status of a self-employed or unremunerated fa-mily worker. Essentially,

these observers would seem to hold to the view that ecployment growth in

certain sectcrs has been entirely supply determined rather than a response

to shifts in demand.

14/ W4hereas in the 1950 cer.sus, all ejidatarios were automatically classi-fied as part of the agricultural labor force, the 1960 census attempted,with only palrtial success, to verify thE principal employment of e;ida-tarios. Since many of thiese held princiDal c;ployments outside agri-culture, the reclassif ication of these in 1960 would reai:ce the 1960total relative to the classification'tised in 1950. Furthermore, Al-timirintroduces addition.al adjustments to compensate for the inmperfect classi-fication of the ejidtLtarios as well as for the codincg and processingerrors that afflicted that census. The 1970 census appears to haveundercnmuncrated the agricultural labor fcrce. This is su',-es ted by theobservation that unusually low laibor force participation rates werereported for men in thc prinie lzbor force age -roup_s in states whichare hcavily agricultural. In additioni, the 1970 figure would have beenbiased downv:ard with respect to 1950 by the greater success in properly.4lassifying tle ej jidatari os.

15/ For example, see MacFarlatnd, on. ci.t.. Trejo (1974), on. cit., andPr o g rnma .Nci n na1 d o_LEpkco, on. cit.

-12-

Unfortunatcly, the criteria e:mploned for dcciding thiat tertiary sector

growth has been "ex:ccssivc" are niot explic.tely set Lorth and verified em-

piriCally. Even a cursory look at the sectoral raLtes of growth, how;ever,

would not seem to indicate an "cxplosivc" growth of maorginal employments.

Both the commercc and servi.ce sectors, for example, expanded at rates below

those of the "productive" sectors of industry and construction. Further-

more, much of the expansion of the service sector has been in technical

and professional occupational categories. Government employment has expanded

at approximately the same rate as total service employment and accounts for

just under a fourth of the total. Finally, the growth of both services and

cormmerce are likely to be closely related to the growth of urban population,

.so that the growth in the latter would be exTpected to shift the demand for

labor in theafrmcr 16/In the case of Mtxico, the urban population, defined

in the census as those residing in corAnmunities, with a population in excess

of 2,500, grew at a rate.of 4.3 percent a year between 1940 and 1970, a rate

that exceeds the growth of tertiary employment.

However, I would maintain that one canniot judge whether the growth of

the tertiary sector has been "excessive" mr.erely by looking at its absolute

size. More instructive would be the course of pioductivity and incom:nes

generated in the sector over time. Examination of these variables would also

16/ Thiose who subOscribe to the view that cmployme.xinont in the tertiary sector

is supply dctcrmined naturally ecmphasize the impact on labor supply of

the rapid growth of the urban populatio'i.

-13-

providle a t:-st of thc widcsprc|;l notion that cmoloymcnt expansicn in this

sector hlas been "supply" deter-,ined and, thcrefore, "c:*:cessive." Co-nsider

the implications otf t'he "labur surplus" and "overcrcwdi n" characteriz ations

.of the gro. th of tertiary employiient. If the wages of the unskilled and

unorganized labor force had been at the subsistence level at th;e beginning

of the period under observation, then we should observe the expansiom of

employment in the sector proceeding at a constant level of real w'ages, or,

in the short run, at wages falling below subs-istence. If wages began at a

level above the subsistence level, then one shlould observe a decline in

wages toward the subsistence level. If one could obser%ve declines in real

wages, this would provide support for the deterioration hypothesis, at.

least within the sector. 17/ One would expect the changes in wages to re-

flect the course of both the marw inal and the average productivity of labor.

Sinceitis possible to subject ithe hypothesis to empirical testing, let us

do so, beginnling first at a macroeconomic level..

In Table 3,I trace the growth of output, labor force, and total factor

productivity over the 1940-69 period for three broad sectors of the economy.

One of thc striking fcatures of the developr:nc-t process in Mexico has been

the broadly based increases in output and productivity across sectors. All

the major sectors have participated in the developmental process, though to

varying dcgreos, in each decade since 1940. The 1940s and 1950s saw strong

17/ A decline in wages within the tertiary sector would not provide suf-ficient grounds for concludin- that avera2,e cmloyment conditions haddeteriorated if thc growth of the tertiary' sector occurred at the ex-pense of an even lower-wage agricultural sector.

TAF, LE 3Output, Eivployncnt and Total FLoctwr Productivity

in Nexico, 1n.0-_1969 (iDGvO pesos)Rates of Chl-ln,2e

(percent per year)1 1940 1950 1960 1969 1940-50 1950-60 1960-69

Priwmary Sector:

Ou 1 pu t ptr i j) $ 3 5 ,5413 $15A442 $23,970 $32,912 6.1 4.5 3.6Labor Force

(thousands) 3,832 4,867 5,048 5,293 2.4 .4 .5Ouputpr- ;rvsrer$ 2,229 $ 3,173 $ 4,748 $ 6,218_ 3.6 4.1 3.0

Secondary Sector:2

Output(mniilj.) $12,447 $23,467 S43,933 $94,362 6.5 6.5 8.9Labor Force 826 1,490 2',175 3,439 6.1 3.9 5.2Outputperworker 15,069 $15,750 $20,199 $27,439 .4 2.5 3.5

Tertiary Sector: 3..

Output(mnills.) $27,663 $48,061 $34,127 $153, 469 5.7 5.8 6.9Labor Force 1,200 1,983 2,990 4,223 5.2 4.2 3.9Outputperw,jorker.22a053 $24,176 $28,136 $36,341 .5 1.5 2.

Total Gross Domestic r'r odtuct:

Output .tmills`.)$48,653 S86,973 $150,511 $277,400 6.0 5.6 7.0Labor Force 5,858 8,345 10,213 22. 95 3.3 2.0 2.7Output per workers8,305 $10,422 $14,737 $21,413 2.3 3.5 4.2

r;otes:

1. The primary sector includes agriculture, livestock,r, forestry, and fishina2. Includes mininc, petroleum, minufacturing, constructiotn, and electric power generation.3. Includes corrirerce, finance, transportatiorn, communications, government, and other

services4. Since the toLals have been adjusted for the value of inter-industry financial

transactionS, they iuay be less than the suum of the sectorai products.

Sources: Output data for 1940 and 1950: L. Unikel, et. al., El Desarrolilo Urbano deMe'xico (>!ex.ico, D.F: Siglo XXI, 1976); for, 1960,,anco de fdxico.Prn.iictCInterlo 3-_utoy G... sto 1970-73 (Iexico, n.F.: 1979), p. 61; for l17U,Banco de Nexico Informe Anual 1977 (Mlexico, D.F.: 1978) Statistical Appendix

,-able 3

Employment data refer to the adjusted d,ta of 0. Altimir, "Le Mledicion de laPoblacion Eornc;iicamente Activa," Demoqrafia y Econo-ia iVII:1 (1974),PP. 50-83.

15 -

increases in total output and output pcr worker in the arifcultural sector

as the large Investments made in rural infrastructure bc;f;an to yielcd returns.

During the 1 9 60s this sector demonstrated less dynamism as tthC limits of

cultivatable laind were reached and the more obvious or easy investr.ment op-

portunities in rural infrastructure (e.g., large scale irrigation projects,

trunk roads, etc.) were exhausted. Gains in bo th output and output per

worker were only modest. One should not overlook the fact, however, that

the gainis may not have been evenly distributed over the sector; recall the

interregional differences that exist in the productivity of agriculture to

which I referred above.

The secondary sector's expansion of output by 89 percent during the

decade of the 1940s was accompanied by an almost. equally large increase in

employment. Given the limited access to capital goods from abroad due to

the Second Wqorld W-Tar, the growthl of output was apparently achieved wit7h

little or no deepenin. oL capital. Stable or declining real mitimum. wages

fron their highly inflated levels of 1940 and declining real earnings during

that decade may also hiave been a factor encouragincg the expansion o F sector

employment. 18/ The following decade produced an equally large increase ifn

18/ S. Perrak-is, "The Labor Surplus tNodel and Wage Behavior in >Ie::ico," In-dustrial R5elations 11:1 (Februarv, 1972); C.N. Revnolds, Thle Vc::ican

E,cono00V (New, Hi;aven: Yale University Press , 1970), pp. 85-86; a much moreconservative esti;mate of. tho probable extent of the decline in rcalwag"Os is offered by T. Kin-, Me:xico: annustriali:ticn nd Tradlc Polic esSince 19l!0 (L.ondcn: OCford LT n.ivc.rsity Pr-s, 1970) ? 26-27. S S:-.arvof th(e literature on wages during tLle 1940s and 1950s is contained in P.StcuisY, "'Tlhe IMpact of Institutional F;actors on Urban Labor .. arkets,"S'tudi(.s in F.-m.ploymlent and Rural Dcveloprnnt 7No. 27 (Washington: TheWork Blank, luly, 1975) mimeo, pp. 65-68.

- 16 -

output as M'I::ico actively pursued policics of im:port i,ub.stiiLution. In con-

trast with the decade of the 1940s, lhowever, ei.;plo:ment grok.th lag-cd that

of output. I,ithli the restoration of normall access to capi.tal goeds fraon

abroad, capitol deepening must liave been a contrIbuting f.ctor to the ¢sharp

rise in output per worker of 28 percent over t:he decade. The rate of growth

of the sector labor force declined relative to the previous; decade but still

managed to increase by 46 percent. The 1960s saw an acceleration in the

rates of growzth of sectoral output, employment, and productivity.

The tertiary sector has been a full participant in and contributor to

growth. Output increases in each decade closely approximatted the gro;uth of

GDP. This sector persistently reported the highest absolute levels of pro-

ductivity in the economy and the rate of increaje in productivity increased

in each decade. These increases might be e:xpected to follow from an in-

creasing weight of high skill employments within L'a sector, as well as

from increases in physical productivity that have occurred in some compDonent

industry groups such as transportation and co;:..munications. Obviously any

expansion in marginal activities that may have occurred within this sector

was too small to offset the impact of these more favorable developments.

In short, the broad sectoral data indicate a shift in employment from

the sectcr with the lowest productivity, agriculture, to those of higher

*productivities. 'Moreover, this shift has not depressed the rate of increase

in producltivity to inconscuential levels in those sectors undergoing the

most rapid growth in employr.,ent. In fact, the rate of increase irn output

per worker is seen as accelerating over the 29-year period. Such a pattern

of developmcnt, on its face, would not seem to bh consistent, with a general

- 17 -

dctcrioration of cmployiint conditions.

On tic otler hand, it could be argued jdstifiably Lhat. large- in-

crenses in productivity in the agregate are not inconsistcJIt with a pirallel

decline in the absolutc productivities of so5nLc component grouns wtithin any

of the sectors. It may well be that thc econo.-:ly is becoming increasing1y

dualistic in nature, witlh productivity and wage declines in part of each

sector being more than offset by increases in other parts. Incone distri-

bution data for 1963, 1968, and 1977, however, would not seem to be con-

sistent with such a development. The share of income accruing to the bottom

40 percent of households in these years varies within a very narrcw range

of 10.5 and 10.9 percent of the total. 19/ To the extent that the share of

income flowing to the bottom deciles has reinained virtually constant over

time, this implies that their incomes grew at approximately the sarae rate

as the average per capita income. Nevertheless, the income distribution

data for Medxico are not free of major deficiencies and may, therefore, mask

the true course of wages of the marginal labor force. 20/ Furthermore,

19/ E. Hernandez L. and J. Cordova C., "Estructura de la distribucion delinsreso en Mle'xico," Comerciao Extrior 29:5 ('May 1979), p. 507.

20/ All of the estimates of income distribution for Mexico are derivedon the basis Of assui!DtionS tlhat, by necessity, involve a high dogreeof arbitrariness. This arises frcm the lrge discrepanc'., wnich hadincreased over time, between thc estimates of national income basedon the hou-,ohold incom..e , and c:rmnditures studies ind those derivedfrom tihe national incomCe accounts. Since the former sere as thebasis of estimating thc distribution of incomo, it is necessary todistribute this discrepancy over houscholds in sor.e manncr. DiffCer-ent sclolars have adpLted thieir o;, n m, thod of distributing, the dis-crepancy twith rather sig,nificantly different results for tlhe shiaresof particular deciles.

- 18 -

since income distribution dala are co..,pi]Cd on a hc. schold rathler t:h.n on

individual basis, they cannot be ex;pected to reflect c:icjrately tlhe course

of wages.

We can c:.plore the possibility th-at the terms of erlpo-mcnt have de-

teriorated for the "m.arginal" labor force by c.amining thoi course of earnings

in that part of the non-agricultural labor --arket that is generally -ackno;-

ledged to be governed by supply and demand conditions rathler thLan by insti-

tutional interventions such as legal minimuzi wages or collective bar-aining.

While legal minimua wages are supposed to be obligatory in all employments,

evasion is acknowledged to be widespread ar-mong small employers. To the ex-

tent, therefore, that the wages paid in small estab'ishments are mar'.ket

determined, they should also provide a rougt guide to the earnings oJ the

self-employed since the latter represent the opportunity cost of labor to

the small-enterprise sector. In Table 4, I present data tracing the course

of carnings of workers emaployed in enterprises of different sizes as well as an

estimateot the returns to labor of the self-employed and unremunerated family

labor in establislim.ents employing no paid worklers. The data are dra=n from

three sectors, industry, services, anid commerce and are drawn from th-.e

sectoral censuses that are taken every five- years. Problcms of comparability

preclude going back beyond 1960 for industry and services, while the un-

availability of thc censuses for co=r-erce for othier years. limit my observa-

tions only to 1965 and 1970.

The data reveal substantial increases in the real average total re-

munerations of employees in all size strata and in all sectors. W ithin

Table 4. Average Real Annual Total Rem-unerations in Services and

*o.Conrerce by Establishrlent Size 1960-75

Sector and Size Average Total Remunerations Percent

of Establishment (Pesos of 1960) Chiange

1960 1965 1970 1975 1960-75

Establishirr.er:ts withno paid employees:

Service 6,315 5,625 6,918 8,245 30.6Indtustry 2,710 5,1-3 5,102 7,058 160.44Cc- i: erc* 4,089 7,501 - 83.4

Esta1blisl.hments withpaid eiip loyces:

Setrvices, 1-2 workers1 5,397 - 9,799 8,655 60.4

Industry, 1-5 workers 4,973 6,005 6,855 9,610 93.24

Co.:.erce 1-2 workers 7,197 7,973 - 0.8

Services, 3-8 workers 7,869 - 10,437 11,096 41.0

Industry, 6-25 worlers 8,119 8,940 10,534 12,977 59.8

Industry, 1-25 w-orlkers5 6,644 7,945 9,308 12,003 80.74Come: er c, 3-8 workers - 9,941 10,957 - 10.2-

Services 9 or more wqorkers 16,470 - 18,637 18,04b 9.6

Inclustry 26-100 workers 10,272 11,969 14,419 17,185 67.3Indtustry, 101-500 worlkers 11,854 15,058 18,148 22,1i6 86.6Industry, over 500 workers 12,877 18,241 23,225 28,236 119.34Ccnrnerce, 9 or more workers - 16,611 16,595 - 1

Notes:1. 1-3 workers in 19602. 4-10 workers in 1960-3. 11 or more workers in 19604. Percent chlange 1965-70

Sources: Censo de Servicios, 1961, 1966, 1971, 1976

cLnbo Tn.111strii., 1961, 1966, 1971, 1976( if:;i ;i. r I ;1 i, 1966,1971.

-20 -

the service sector, wages in the small cstal-)iLi;h:i ents gained relative to

the large, oiies. 21/ The same is true in ind:Ustry; workers emlployed in es-.

tablishr.ecits with five or fewer -.orizers gained relativc to those in all

size categories up to 500 w:or,errs per establish.:ent. A similar narrowing

of differentials is observable in conmerce over the five-year period avail-

able. The censuses do not providc a measure of the returns to labor in es-

tablishments with no paid employees. Thus, the earnings of the self-employed

zare not reported as such. Nevertheless, T made an attempt to estimate the

returns to all labor -employed in establishmTnents with no paid employees by

deducting fronm gross income all recorded outlays for inputs and other ex-

penses as well as an inputed return to the capital employed. The residual

would represent t'le returns to labor which are reducible to a per-worker

basis by dividing the totals by the number of persons employed in such es-

tablishments. The results are presented in the first category appearing in

Table 4. While the margin of error in estimating the returns in any one

year may be substantial, the upward trend in returns is so strong that the

direction of change is not likely, in fact, to have beeni negative. Thus,

21/ It should be observed that, between 1970 and 1975, the real earnings ofworkers in the smallest size service establishments recorded a signi-ficant decline in contrast to ali other catcorics and sectors. Thiscould h2rald a bcginni.g- of a sccu'lar a decline in the quality of em-ployment concit.ions w.ithin that sector or it could be a tenmporary de-cline associated with a 1ig of twages b^!ind an accelerating rate ofinflation. The discussion of 6or.,2stic servant wac-es thiat follows be-low; sug,gests that declines in real w:a^,c Llhat occurred about 1974 and)4975 may have been only a temporary phencniunon.

- 21 -

on thie basis of thelse disan,<gre,a-tvd data, su-pzort for the prusu:.ntion of

deteriorationl is lackin-. On the contrar-, the real wg.ae position oi wor'et rs

in the rio5t scnsitive part of t'he 1cbor market would <:cc:n to indicate a

considerable improvement in emp1loyrmeat conditions. 22/

A closer look. at one category of service employment, domestic service,

provides an additionial basis for evaluating changes in employment conditions.

Domestic service represents an occupation of easy entry, frecjuently the

source of first employment in urban areas of' migrants from rural areas. As

such, it is reasonable to expect that it should mirror changes in labor

market conditions for unskilled female workers generally. Domestics com-

prised the largest single group withini the service sector in 1960, account-

ing for about 540,000 workers. From 1950 to 1969, the number so ermployed

increased at an annual rate of just under 3 percent. Since total employ-

ment in the service sector was expanding atarate of 4.4 percent, do=estic

employment declined as a proportion of total service sector employm.ent.

Again, whAiether this grot;th was simply the produce of a labor surplus pressing

into any available employment can be tested by reference to the course of

wages over time. W1hile data are not available over as prolonged a period as

we would like, those that are at hand do 'not yield an impression of a secular

22/ It might be argued that the establisL-n.ents likel to be included in acensus are thlose v:which are "rcgisterod1 " and, therefore, not representa-tiv'e of the wage conditicns prevailin,g in less for.-.a lly organized enter-prises. However, to thle extent to which tie for-er are not bound byinstitutionial intcrvention in ttc wage detecinationprocess, tlhen they

,may be expected to be competitive in the labor market with the latter.

2 2.

decline in earnint,s. Onc study of the domestic labor r.lrhet in alcn:ico City

compiled the wages offered to dor.c,:tics in tlie classificd advert-semeent

section of tvio Icadin- newspapers, Fx.:el-ior andc Univers.-il. 23/ Detw.eun

1963 and tLhe early 1970s, the real caslh wage of domestic servants in

Mexico City e.cre estimated to havc risen on the order of 25-35 percent.

While this raLe of increase laggcd belhind thc increase in the real legal

minimum wage for IMexico City of about 50 percent by 1970, it did not com-

pare unfavorably with the rate of increase in wages in consumer goods manu-

facturing industries of about 36 percent. The course of real cash wages,

however, took an abrupt turn followiing 1972 as inflation began to accelerate.

By 1975, real cash wages had fallen by about 19-27 percent. 24/ MIost of

the decline occurred in 1974 and coincided with the more than quadrupling

of the rate of inflation in that year as compared with the rates chiaracteris-

tic of the first three years of the decade. TIhe decline in real cash wages

may simply reflect the lack of experience with iniflation in a society which

had enjoyed at least two decades of relative price stabilicy. Lacking so-

phistication or any organizational support, domestic servants may not have

23/ "An•lisis del 'Mercado de los Servicios Domesticos en iMexico," Cuadernosde Empleo, 1 (1976), pp. 51-109.

24/ The range of incrcases and decreases in remunerations reflect the useby the aLltlhors of the study of tn'o diffcrent metlhods to estimate therate of channge. The low.er values given are derived from the data asrecorded in each) vear of the survey. Iowcvcr, since thie annual valuesshow;ed rather sharp and irregular year-to-ycar variations, a trend linewas fitted to the nomninal wagc data by lcast squares method and a seriesof cstirated aninual rem-unerations were derived. Ti,- larger values inthe ran.ges appearing in the text are attributable to the wage seriesso deriv ed. i ppLI., p. 94-95.

Table 5. Average Monthly Cash Wages Offered Domestic Servants-, Mexico City, 1963-80.

Method I 1Method IIMonthy Wage in Index of Real Estimated 1lonthlly Wage in Inde:e of

a Current Pe;os Wages Current Pe5;os R eal WI. ea(1963=100 (19 63=100

19',3 277 100.0 258 . 100.01964 288 . 101.8 283 10(. 3

1 5 320 109.0 308 11 .6

19c,G 349 114.1 333 117.0

1967 360 114.5 358 122.3

19§3 363 113.9 383 128 .9

1969 356 107.4 408 132.2

1'J70 /4 47 128.2 433 1' .5

1971 427 116.1 458 133.7

1972 478 123.6 4 /,83 1 54.1I

1973 511 118.7 . 508 12. .81974 532 101.1 533 1(o.7

1975 615 99.8 558 97.3

1V.^,, July 3,260 198.8

_'>0, July Mini-r-:1 1'aoe Offered 2,500. 152.4

1. Netliod I is a simple avetage of the sample offers in each year.

2. TMetIhod II estimated an -annual wage from a regression line fitted by least squares to the sample averages. Zothmiethod_- Lrereported as publishfed.

Source: "Analisis del Mercado de los Servicios Domesticos en Mexico," Cuadernos de Empleo, 1(Mexico, D.F.:

Secretaria de Trabajo y Previsio'n Social, 1976) pp. 55-114.

The 1980 data were culled from July, 1980 issues of Excelsior and Universal.

- 24h -

adjusted their r ;c:rvation price: s£fficic:utl r. pidI to prevent a declire.

Thc fact that tlhey receive a very substixntial propcrt;ion of their rem.unera-

tion in kind, i.e. room and bo.r.d, may ae h.ave teo-n`d to bluzht their

awarcne;s of the full extent of the decline in. the - .rchasing po'wer' of their

cash wage.

The course of don.mestic scrvant re-uneratiLons since 1.975 i-.. not re-

corded in any statistical source in ''Texico. A's a fi-rst approxim-ation to the

measure of the change whiclh has occurred sir.ce 1975, I culled a sample of

cash wages offered for live-in household help from E::cclsior and Uiniversal,

the same twqo newspapers employed in the survey cited aibove. I obtained

78 observations from issues published in July, 1980. These sug-ested that

the loss suffered in the middle 6f the decade had not only been recou?ed

but that substantial further gains hCad been made. The average real :-oney

wage offered in July, 1980 was tn.rice as great as that of 1975 and 55 per-

cent greater than that OL 1970 (See Table 5). Lest my sample not be re-0

presentative of the offered wa-2 but biased upward, I also ccmpared the

lowest frequently quoted wage with the past recorded. ave-:a-e offers. .s the

minimuni, I adopted a cash wage of 2,500 pesos whiich was the bottom o. the

2,500-3,000 peso incerv.alwfthin which just 50 percent of the.observations

fell. (Only siix recorded t'7age offers were for less than 2,500 pesos per

month.) Even this wage, in real terms, proved to be 19 percent greater than

the previous highest average cash wage, that for 1970. Thus, the course of

real-wages for domestic servants does not appear to be consistent w ith

labor market conditions characterized bv "overcro-wtding" in marginal

-25

emnployrents..

III. One'Mplo;/tnt and Unlderemployment

A crit'calissue for observcz. of Mexi:an development in reccnt years has

been the deoree to whiich the available labor resources have been utilized.

As indicated earlier, it is commonly asserted that unemnloynent and under-

employment have been increasing over time. The einpirical foundations 'upcn

which this conslusion rests include the following elemenits: 1) a decline

in labor force participation rates that is int2rpreted as reflecting a res-

* ponse to increasingly scarce employment opportunities; 2) the secular rise

in rates of open unemployment reported by the population censues; 3) the

census measures of the number of months in the year or days in the refe-

rence week worked by labor force members; and 4) the concentration of large

numbers of workcrs employed at low wages. We consider each of these in turn.

As is clear from the data presented in Table 6 the overall participa-

tion rate of the population 12 years and older has declined over the past

three censuses. A substantial decline in male participation rates was only

partially offset by an increase in female rates so that the total partici-

pation rate declined from 49.5 in 1950 to 43.5 percent in 1970. However,

almost the whole of.that decline is associated with an almost equal in-

crease in the proportion of the population over 12 years old enrolled in

school from.6.5 in 1950 to 12 percent in 1970. This increase of 5.5 per-

centage points accounts for all but .5 percentage point of the decline in

the participation rate. The other significant reduction in participation

- 26 -

rates occturs aimon- mon zmtedi Si:Ly an;l over. IT.n view of tlhei extenEion of

reLircr.ent benufits to a wider segmcent of the ag,ing popul:itiun, withdraw..;al

from the labor force miglht be interpreted as a voluntary measure rather than

a response to shortage:,; of employm'nt opportunitic;.

The reported decline in labor force participa'tion rates may also reflect

shortccming,s in the enumeration process. 25/ The difficulties c f

accurately assess,ing the labor forcc st:atus of populations in developing

countries are too familiar to be reviewed here, a:.Id these are not entirely

absent in the case of MIexico. In addition, the interpretation of reported

changes over time are complicated by differences in the concepts employed,

the nature of questions asked, and the time of the year in which censuses

are taken. In particular, the declines reported for 1970 seem suspect. We

have already referred to the unusually low participation rates reported for

men in the prime labor forca age groups in states which are predominantly

agricultural. An anomaly that raises doubts about the female participation

rates is found inma comparison of the rate based on the whole of 1969 and that

based on activity in the week prior to the census. One would normally expect

that a labor force defined on the basis of activity at any time during the

preceding year should. be larger than one based on labor force status during

a single week. Yet, the 1970 census reports a larger female labor force

during the week preceding the census.

25/ An excellent ann.lvsis of the measurement of the ZM1exican. labor force andan. interpretation of the changes is offered by Donald B. K;eesing, "Em-ployment and Lack of E:nployiment in 'Me:ico, 1900-1970," in QuantitativeLatin Americ.an Stl½s: Mlethods and F?indin(-s, .J.W. 'Wilkie.:and K. .Rucdle,eds. (Los Angeles: UCLA Latin Ar;c'rican Center, 1977), pp. 3-21.

( Table 6: LABOR -RCE AND PARTICIPATION RATES, 1950-69, BY SEX

Nur-ier Participation 1960(Thousands) Rate (Percent) No. 1/ LFPR 1/ No. 2/ LFI'R 2/ No. 3/ LITIP;

Total 8,345.2 49.5 11,253.3 51.1 10,631.2 48.3 10,212.9 46.5

Men 7,207.6 88.2 9,235.0 85.1 8,732.0 80.5 8,496.2 78.7

Ilo-.cn 1,137.6 13.1 2,018.3 18.0 1,899.2 16.9 1,7Ž-6.6 15.4

1969 1970 (January)_

No. LFI-R No. LFPR

Total 12,955.1 43.6 12,909.5 43.5

-en 10,488.8 71.7 10,255.2 70.1

Wolrdn 2,466.3 16.4 2,654.3 17.6

Sources: 1950 VII Censo de Poblacion.1960 1/ VIII Censo de Poblacion as reported by Altimir.

2/ Coleglo dc Mc'xico, Dinanmcna de 1;i Poblaclon en Mexico ns reported by Altim!r.3/ 0. Altimair, "L.a I.edicl6n de la I'oblacin Lconoinicamnente Activa de Mcxlco, 1950-70,"

Deminografia y Ecoiton:la, VIII: 1(1974) pp. 50-83.1970 IX Censo de lPoblaclon.

- 28 -

Finally, a subsccluent source of inforli..tion scrve: furtlhcr to cast. doub t

on the extent of decline in participation rates. Bc inning in 1973, a wide

range of labor force data has becn provided by quarnerly surveys of housc-

holds for the thrce major urban concentrations of the country, Mfoxico City,

Guadalajara, and .Iontorrey. Additional rcgions of the co.untry have been

added over tine until, at the end of the decade, the whole country was be-

ing suveyed. Unfortunately, the only comparison that can be made of 1973

survey participation rates with those of the census is that for the Federal

District; census data are not available for the other twto metropolitan

areas in isolation from the states in which they are located. For the Fede-

ral District, however, the survey data report participation rates for both

sexes well above the census rates for the whole year of 1969( See Table 7.)

Whereas tlhe census participation rate for mcn was 71.5 percent, the average

of four quarterly surveys yielded a rate of 75.1 percent; ornly anong men

over 55 are the survey rates lower than tllose of the census. The absolute

and relative differences in the two rates are even greater for women, 29.7

to 35.0 percent. 26/ While it is not unusual for labor force surveys to

report a larger labor force than population censuses these differences

would seem to be unusually large, partictilarly if participation rates are

26/ Secretarla de Pro-ramacio'n y Presupuesto, Encuesta Continua de Manode Obra, Vol. 1, 'Number 1., .1973. Zcginning in 1978, the survey pub-lication was renamed Encuesta Continua do Ocunacion.

Table 7.: LABOR FORCE PARTICIPAtION RATES' IN TH1E FEDERAL PTSTRICt BY AGE GROUP AND SEX, 1969 AND 1973

Ag,e GroupsTotal 12-19 20-24 25-34 35-44 45-54 55-64 65 and Olver

tales

1969 71.7 28.0 80.0 94.9 96.4 94.5 85.9 57.9

1973 75.1 37.9 85.9 98.1 98.2 95.8 80.8 .51.0

Fen-ales

1969 29.7 23.8 44.2 33.0 31.2 29.2 '23.6 13.8

1973 35.0 29.6 50.7 39.5 39.0 34.0 24.7 12.0

JI'

Sources: Secretaria de Industria y Comercio, Dlreccion-General de Estadistica, IX Censo Ceneralde Pollacio1u, Fesumnen Ceneral (MlexicQ: 1972); Secretaria de Prograr.acidn y Prcsupui.sto,'Etcuesta Continua de Mno de Ohra, 1973 (Mexico: 1977).

- 30-

are bcleved to be ini secular di-cline.

Cbmparisons of census -and survey partic:J.pation ratcs in other regiors

can bc made for a later point in the decadtc. Table 8 presents a couparison

of participati on raLes reported for the week prior to the census (January,

1970) with those reported by the surv,y for the first quarter of 1978.

Data are presented only for men and are disaggregated into three of the

prime labor force age groups. The margin by which the survey data exceed

the census participation rates is notable in all of the regions reporting

in 1978. In slhort, because of the large increase in school attendance

among the population of labor force age and because of the possibility that

the 1970 cenus seriously underreported the size of the labor force, the

reported decline in participation rates cannot be accepted as having been

established as fact. It follcws t:hat the alarming conclusions draun from

the reported trend wxith respect to the degree of ut:ilization of available

manpower resources is likewise open to serious question.

The second basis for the view that uneinployEmeni: has been increasing

secularly rests on 'tle rates of open unemployment reported by censuses

since 1940. Clearly, these have been increasing as can be seen from Table

9. However, while the 'rate appears to have quadrupled between 1940 and 1970,

the 3.8 percent rate fcor the lat:ter year can hardly be considered to have

reached a troublesorme level. In view of the doubling of the proportion of the

urban populationi in thc total from, 30 to 59 percent, it is reasonable to

expctt an increase in frictional untemploy-ment. Indeed, the 1970 level, if

it could be accepted as an accurate mcasure of unemployment, could be

-31 .

Tib le_3 LAO oc"PT?.:AT2 XISO ACE! Cc7=17S,J-

jq4;1~7 N F: 71,T rL'A 7T .7 l978 i Y 11.GICN.

Region 25-34 35-44 45-54

Federal District-

1970 92.9 92.1 92.31978 95.8 98.0 95.3

North

1970 87.0 89.3 88.11978 89.9 91. 8 92.7

Northeast

1970 90.3 91.7 89.91978 95.0 97.5

Northwzest

1970 90.5 89.2 88.21978 97.0 98.5 93.9

Central-N:'orth

1970 86.9 89.1 88.51978 93.4 94.8 93,.7

Peninsular

1970 90.1 91.8 91.71978 99.0 98.6 98.6

* Centr.a1-P<c 4 f4 '

1970 84.5 86.2 85.31978 94.1 95.8 93.8

Central

1970 87.4 88.7 87.61978 94.9 95.8 . 92.6

Sources: Conso cle Pobl:i( ion, 1970

SccreuirfIn do Pros ramncicm v Prcestuo.uesto, F.ncuectta ContintivnSohu. OcIlii, Vol. 7, Tvimustrc 1, 1978.

- 32 -

Table 9: LNE='L0Y'I';T IN' ':XIC0, 1940-70

_________t'nE.:ov'.",t !vnteYear Total .Lale Ferale

1940 1.0 1.0 0.17

1940-a 3.1 - -

1950 1.3 1.3 1.2

1960 1.8 1.9 1.3

1970 3.8 2.8 7.5

Notes:

1940 - Census includes only those unemployed for atleast a frcnth at the time of the census.

1940-a - The derivation of this estimate is not clear.It appeared in an unemploymer.t series pub-lished in the e-nuar-o Estadistico for 1938and 1940, as repor-ed by Wilkile.

1950 - Of the 105,177 une=ployed workers, the censusreported 32,030 unzrploycd for 12 weeks orless and 73,147 who had been seeking employ-ment for 13 weeks or core.

1960 - The.measure of the labor force used here in-cludes the Altimir adjustm6c.nt. Of the182,088 unecployed only.

Sourcc:s:* Censo de Poblaci6n, VI - IX;

1940-a, J.r.> ''kie, "tes Hypotheses for Sta-tistic;al ]c. st- h .n Recent Mexican Hlistory,"Latin Anmeicicin Reiszrch Reviv.c 6:2 (Sum;m ir, 1971),Table 4.

- 33 -

considered to be close to the frictional level.

Whether the trcnd can be accepted as an accurate reflection of the

past is, however, open to qucstion. Cha.i,;cs in the ccnccpts employed and

in the questions asked have varied from one census to anothcr.rendering

.comparisons over time difficult to interpret. 27/ Certainly, the census

measures for 1970 have proved to be conservativc when co:mpared to thc

findings of the labor force surveys of the 1970s. For example, the census

reported the rate of open unemployment in the Federal District to be on

the order of 5 percent in the week preceding the census. In contrast,.

the first of the houselhold surveys of the labor force, that for the: first

quarter of 1973, reported that 7.2 percent of the labor force was unem-

ployed. However, important differences in concepts eiployed render,an

interpretation of this increase dif"ficult to rake. For example, in the

census, a respondent's status was defined by reference to a single week,

whereas the surveys extended the reference period for the definition of

unemploymient to two months. This difference alone could account for a

substantial part of the difference betwqeen the two measures. In any case,

the data for the subsequent years of the survey do'not reveal any consistent

tendency for unemployment to increase in any of the regions during the 1970s.

Unemployment increases did occur as the economy slid into recession in 1976,

but the recovery which followed quickly reduced the rates to pre-recession

levels..

27/ K Keesinc, discusses at length some of the difficulties of interpretingthe various cenisus measures, op. cit., pp'. 5-7.

- 34 -

Of inLerest is tlle relatively shlort duration of reported iuiomployncnt.

In all of thc ceiisus years, only about onc percent of the labor force was

unemployed for as long as onc month. Long-term uncemployment of over thir-

teen weeks was particularly Ciriustial; in 1970, only 13.3 percent of the

unemployed were in this category. As in most countries, the incidence of

unemployment was greatest among the yoUng. In contrast to the national

average unemployment rate of 3.8 percent reported by the 1970 census,

that for teenagers was almost 6 percent while that for labor force mem-

bers aPged 20-24 was 4.4. 28/ The first surrvey reports even greater dis-

parities in the age-specific unemploy-ment rates in the large metropolitan

areas. For exanple, in the Federal District, in the face of an average

rate of 7.2 percent during the first quarter of 1973, the teen-age rate was

18.4 and that for young adults aged 20-24 was 10.1 percent. 29/ In con-

trast, that for the prime age groups was on the order of 2 to 3 percent.

The 1970 census data for the District yield a teen-age rate of 11.2 percent

and a young adult rate of 5.9 percent as compared to the average rate of 5

percent.

The vies; that a secular increase in unemployment may have occurred

may be deduced from the surveys. Earlier, we suggested that such an in-

crease, if it had indeed occurred, could be a reflectioni of the increasing

28/ Ccnso de Poblacion, 1970,.Table 32..

29/ Enctiesta Continua cle Mano de Obra, op. cit., Table 2A.

35

uarb.,tiization of the country anld a subsc-quent incrcase in frictional tin-

cmphl;t ,,. A42 crossctionS analysis of r he regcional uneiployment ratus

would tend to Fupport tlis view. The highest rates of unemploymcent arC to

.bc founI in the metropolitan urban areas of the country, intcrmediatc rates

in the s-maller urban centers, and the lowest in the least urbani-ed re-

tions. Since the urban areas have incrcased tlicir relative weight in the

Coutitry as a whole, some increases in the national average rate of unemr-

PlOyment could be expected to occur. The significance of an increase due

to f;uch a structural change in the distribution of the labor force, how-

ever, is surely less ominous than secular increases in rates within regions.

To (late, the labor force surveys have not recorded such increases.

NIeasured open unemployment rates in Mexico would thus appear to have

been below those frequently observed in other. developing countries. Con-

sequtently, concern with the utilization of labor resources in Mexico has

centered not so much on open unemployrment as on the presumed high level of

underemployment. Unfortunately, most of the discussions of underemploy-

ment of labor in MIexico reflect the same conceptual and measurement pro-

blenis that prevail generally. As a-result, one finds widely disparate

.estimates of undercinplcvmcnt depending on w:hether these are based on mea-

sures of the prcporti.on of a year or week worked by the employed, on mea-

sures of adequacy of incomes earned, or a combination of both. For example,

a Study group formed to study the employmcnt problem in the early 1970s

adopted a com-plc:: definition of underemployment that inclucdcd almost all

unpaiid family workers, those with incomecs below official regional minimum

Wage levels, and workers employed for considerably less than a full year.

- 36 -

The gnr!ityl, eSt4:,anv,,,- e rf br _ln:t.1s.b d on 1970 ccnsus data. ranc-

ed frc:_. 27.6 to 44.8 percent of thc labor force. 30/'

Mm.pirical data of time wur;:cd per ti:ne unit. are available for various

points it, time. The 1970 census recordad Lte n-:-.Ther of monclis w.orked dur-

ing 1969. The 1950 census recorded the number of days w orked during a re-

ference weck. Finally, since 1973, the labor force surveys have captured

information regarding the numbcr of hours worked during the referencc week.

Unfortun;ately, the different measurements employed make it impossible to

calculate a trend in the possible degree of underemployment over time.

Only since 1973 can a consistent series be said to e:.rist for a few major

urban areas. Estimates of underemployment in the agricultural sector have

been derived by a-plying labor input coefficients to regional croppirhg

patterns to estimiate the "required" quantity of labor' inputs; converting

these to full employment equivalents and subtracting them from the measured

labor supply yields the "labor surplus."31/ For those who define underem-

ployment in terms of the level of earnin,gs,sources of inforrmnation include the

census of 1970 and a number of household surveys of incomle and e:xpendituras

which have been undertaken since 1963.

30/ Crupo de Estudio del Problema del Emipleo, oT. cit., Table 2. For otherestimates, see p.6 and note 9above.

31/ L.MI. Goreux and A.S. Mtanne, Mutlti-1vcyl ?1Iannin,: Case Studies inMdxico (Amsterdam: North-Holland i'ubiis;ing- Co.,.1973).

.

- 37 -

Unfortuflatcl-l, all of tlicfc appro.nclics to the rca surc.ent of uremploy-

ment. posc seriou.s problTms of interpretation. Those that definc undercm-

ployment by reference to time worked as a proportion of some refercnce

period considered to be "nor-mal" cannot, in the absence of additional in-

formation, distinguish between those dcparturets fronii full-time employment

that are voluntary from those thiat are involuntary; presumably only the

latter qualify as underemployment. Nor can the entire proportion of 'work;ers

working less than a full year as reported by the 1970 census be taken to

represent a measure of underemployment. Some part of these workers re-

present individuals who entered the labor force for the first time after

the beginning of the year or who retired permanently before the end of the

year; on the basis of the age distribution of the population and labor

force I estimate that approximately five percent of the labor force, as

measured in the 1970 census, was composed of individuals in one of these

two categories. 32/ Unless an adjustment is introduced in the raw data on

months worked to account for this sub-group, estimates of underemployment

-will be exaggerated.

The'estim;3tion of the degree of underemployment which exists in the

.ural sector of Mekico is extrcmely difficult to undertake with confidence.

'ne estimation of an. agricultural "surplus" labor force derived by employ-

-yr,ig input, coef ficients does not necessari~.ly moan that that "surplus"

2/ It will be recalled that one of the -measures of the labor force em-

ploycd inclided all t-hose who liad .:o.rlcd at anyv tirme during 1969. It

-was to thlis grouips that a nuestion regardincg the tiumber of months

worked was addressed.

-38-

reinains idle wliczi not einploYcd in agricultural production. Employment in

otlier forms of home production may occupy some part of the time not spc!nt

in agricultural production of final goods. Furthermore, movcment of labor

frori agricultural to non-agricultural employment is a co-rmon phenomenon in

Mexico so that what may appear to be a surplus to the agricultural sector

may not be a surplus to the economy as a whole. Finally, some part of the

agricultural labor force which is av3ilable for peak season employment may

simply not be available for employment at other times.

The measurement of underemployment by reference to income levels

likewise leaves a good deal tio be desired. MSany of those below the income

level adopted to identify the underemployed may not be physically under-

eaployed at all, in the sense that they are available for additional hours

of work. Indeed, they'may be working ardously long hours at socially use-

ful tasks. Their incomes are low because their marginal productivity is

low, which in tiirn, is a function of the relative availabilities of labor

and of physical and human capital. It is precisely this prevalence of low

productivity and inc;^e that is the principal identifying characteristic

of underdevelopment. The redefinition of poverty as underempl-y-ment

creates a misleading imoression of the degree to which available labor re-

sources are not being utilized in productive employlernt. An additional

difficulty with the income criterion for defining underemployment is the

arbitrariness of the standard employed. For example, the legal minimum

wage'is frequently talken as the dividing line between underemployment and

"full" employment. Yet, this wage h.is no uniclue quality that objectively

dis;tinguil.-,hes air.nn thte tvo classe.' of worKers'. By merely raising th

legal minirmum wage rolative to "going" wages or the opjportunity cost of

labor, the nur-ber of "undcorCni?.?oy(:l" woclkers will be ilicreased evLn though.1

no objective change has; occurred in their eniployment or income situation.

Conversely, if the legal minimum is allowed to drop relative to the oppor-

tunity cost of labor, the degree of unde2rcmployment will appear to have

declined. 33/

Thus, not only is the concept of underenployment an elusive one, but

mechanistic approaclhes to its measure:ment are likely. to prove rather mis-

leading. This becomes evident if one examines some of the more detailed

information derived from the household iabor force sureys. While these

surveys have not been in progress long enough to provide a basis for

observing conclusively the secular course of employment conditions, they

do yield newt and important insights into various observable labor market

phenomena that counsel caution in the use of raw or aggregated data. The

section which follows reviews some of the findings based on these surveys

that have a relevance for some of the questions raised irn this paper to date.

IV. Recent Labor Market Observationis, 1973-78

Earlier we discussed the possible significance of the observed decline

- 33/ Employing this standard in 'exico would yield an astronomical propor-.tion of tthe labor force as underemp)1oyed in 1940 when the legal mini-mun had becn pushod up to a very high level relative to average in-comes. During thc succeeding decade, thiis proportion would have fallen

,as the real value of the minimum wage declined

- 40. -

in labor force participation rates over timc, nioting that it could be in-

terpreted as a product of the incroased availability of educational oppor-

tunities ratler than of an insufficiency in the availability of erriployments.

Tlle survey data now permit us to carry the analysis furLller and to bring it

fonrard to 1978. WNe have :indicated that the smurveys, from the start, have

reported higlher labor force participation rates for both scxes than did

the 1970 census. At the same time, however, the surveys do confirm and

extend the secular declining trend in the participation rates among the

youngest and oldest age groups of men, trough the impact on the aggregate

participation rate has become quite small. As observed in the intercensal

periods, declinLes in participation rates among young males continue to be

associated with increases in school enrollments. Among women, the partici-

pation rates have varied within a narrow range for the areas surveyed since

*1973 with no evidence of a trend. One of the striking aspects of this sta-

bility- is the fact that it has occurred in spite of a much greater in-

crease in school enrollment among young females than among males. Prior to

1970, the rate of school attendance among females of post-primary school

age was well below that of males, by approximately a third. During the

1970s, hotwever, fem.ale enrollments increased more rapidly than male, so

that, by 1978, the gap had narrowed substantially. WThat is significant,

- however, is an indication that there was not an equal tradeoff between

increases in school enrollment and declincs in labor force participation

rates. Amnong females, for example, labor participation rates in 1978

were higher than tllcy were reported to be by the census in spite of even

largcr increnses in tho! rate of schlool cinrol.,,c,nt. This was alsr; true

among males, thougah to a lesser cxtent. Thiis can bc shuwn by co;.;paring tllc

chaniges in labor force participation races by sexc with the charnge in the

sum of those rates and the ratc of sclhool atctndance. I'hat the data in

Table 10 suggest is that employment opportuinities for the young c!.:panded

during the 1970s at a rate greater than the increase in tlle non-school

population. 34/ Only if one intcrprets the increase in school enrollments

as a reluctant choice by young people of a non-preferred activity, can the

labor force data be interpreted as reflecting a deterioration in ,lob

opportunities for young potential entrants to the labor force.

One of the constant preoccupations of st;udents of Mlexican ec&nomic

development is the fear that employment opportutnities have not increased

as rapidly as the labor force thus giving rise either to rising uneinploy-

ment rates or declining labor force participation rates. This fear has been

encouraged by the larger rates of unemployment uncovered by the labor force

surveys than had been reported by the 1970 census. In fact, while the levels

in the first quarter of 1978 were higher in five of the eight reportinig

regions than they were in January, 1970, they were only mnarginally greater

by and large. See Table 11. Hlowever, in the aggregate, in all but one

region tthC absolute increases in the rates of unemployia.ent laggcd well behind

34/ Even if the 1970 census data are biased toward an underenumeration ofthe labor -force, the uncderenur.eraLion is unlik;ly to have been suffi-cientla large so as to eliminate the large growth in.the su-m of therates of labor force particf pation and school attenudance that is re-

*corded in Table lOfoi all surveyed regions of the country.

- 42 -

TABLE 1]: '! 1:,' Y Cc'Ai1T((7; o) l1C!r,:r- c ON

RATES AN\ D SCUI0O!. LL>UiNTS, JA'.UA Y, 1970 1'AND 197,1,

FIRST QUAI'TL2Y, B'Y RCGT:):; O;N ; SK*:

LzO)or Fo.c ) Ln ;or Io r- Prc :::iuit ;::

Region Total laie i .'aoia1 :*cae Male a

D. F.

1/1970 48.4 70.6 24.7 63.9 89.2. 36.5

I/1978I 51.6 71. 8 33.9 71.3 92.9 52. 4

egRion Nortia

1/1970 41.7 67.9 15.5 54.5 82.0 21.0

1/1978 40.6 63.3 19.5 62.5 87.6 39.1

Region Noreste

1/1970 43.3 70.0 17.0 56.4 85.1 28.1

1/1978 44.2 70.1. 18.7 62.7 90.9 35.0

Region Norceste

1/1970 43.4 68.9 17.4 56.9 83.6 29.4

I/1987 44.7 72.0 18.7 64.9 92.9 38.3

Region CUntroNorte

1/1970 42.0 69.9 14.2 52.7 82.1 23.4

1/1978 43.7 72.4 15.3 58.6 88.2 28.3

ReLz. Pinst.sl.ar

1/1970 43.9 74.4 13.3 54.2 86.4 22.2

1/1975 48.0 79.a 17.2 63.1 96.2 31.0

~~.__Centro-PjC.

1/1970 42.4 69.3 16.5 54.1 82.9 26.4

1/1978 46.0 72.5 22.2 62.0 90.5 36.8

Region Centro 1/

1/1970 42.7 70.1 15.6 5 3. 82.4 24.4

I/1978 44.9 71.7 19.8 61.3 90.0 35.0

1/ Includles thoset:al.nipLo5 t%;hich, for- part ot the ve:uico City 'otrcpolitan 3:ca but

which lie in thc Scate oc "! :lco.

Sc)tjr c8;: !camc' i.- Tnbulc S.

-43-

TABJI,' 11: O '. Y OF C!{A\>:(tEs IT'I;: .':.\:iY , l iO VX.,1) JAt!N-

ARY, 1978 1: lAO CN)!CL P:XJ tC r .;it1:: x:;i !:D .::PLOY::::T'RMTES AXP iKS ';D AS A P . :oU\ It j to:: 0 J; POML UL.T ION

12 YEARS OT-D A.;D OVi :.'

(JaiuLary, 1973 l8ss Jarnuziry, 1970)

Total 14: I e Female

Z Point Chianges ; Point iUncs ; Point ChangcsLFPR U LFPR U- LFPR U

D.F. 3.2 1.5 1.2 1,7 9.2 1.4

Norte -1.1 -0.2 -4.1 0.1 4.0 -0.4

Noreste 0.9 0.5 0.1. 0.8 1.7 0.2

Noroeste 1.3 0.2 3.1 0.7 1.3 -0.3

Centro-Norte 1.7 -1.1 2.5 -1.1 1.1 -1.1

Peninsular 4.1 -0.3 5.4 0 3.9 -0.6

Centro-Pac. 3.6 0.4 3.2 0.5 5.7 0.3

Centro 2.2 0.2 1.6 0.7 4.2 -0.2

Sources: Scine as Tabie 8.

- 44-

the increzises in tlie labor force participaLi.on rates. In other words, in-

creascs in labor force participation. rates were associated witlh increased

employment rates rather than sir-:p1ly incicased uirmo1o.:ent rates. If the

data are disasgregatcd by sex, some differences do appcar, however. Among

men, in tlhree regions, the Federal District, the Northcastern and ;Northern

Rcgions, ttic increases in the unemploymont rate excceded those in the parti-

cipation rates, suggesting a relatively less favorable evolution of employ-'

ment conditions for men there. In all the other regions, the increase in

participation rates far outrun the increases in unemployment. On the other

hand, the changes in the female rates suggest a universal and significant

improvement in employment opportunities. In spite of significant increases

in participation rates over those of' the census, the rates of unemployment

showed declines in five of the eight regions, while in the others they in-

creased far less than the increases in participatior. rate.s. Indeed, there

is a very striking direct relationship between the increase in participa-

tion rates and changes in the uneemployment rate. For example, the largest

increase in the unem,ployment rate over 1970's was in the amount of 1.4

percentage points-in the Federal District. But this was associated with

much greater increase in the labor force participation rate of 9.2 percent-

age points. W4hat this seems to suggest is a widespread and significant im-

provement in emlploeTcnt opportunities, especially fc.r women. The appea-

rance of employment opportunitics induces more women to enter the labor

forec. The rapid influx of women might be e::pected to lead to an increase

in frictional unemployment and may account for the rather larger increases

in uncmployr,mnt A.itc in the regions of i-ost ropid graxswth of fcmale eiploy-

ment.

This apparent rcspotisiveness of wom.en to thlc expansion of employme.nt

opportunities may be inLcrpreted as indicating Lle existence of a large

pool of hidden unemploy-ment, or of d scouraged wor'Kers, workers who are de-

sirous of emrployment bul w.-ho do not seck employmcnt out of a conviction

that vacancies do not exist. That this is characteristic of many of labor

force age is a cornon belief in NMexico. Support for this view would seem

to exist in data from the labor force. survey which asks individuals who

appear to be out of the labor force whether they are available to accept

employment even though thcy have taken no steps to seek jobs during the

preceding twvo months. Among the surveyed populations of the three large

metropolitan areas, approximately two percent of the male respondents and

six to seven percent of the female respondents answer in the affirmative. 35/

In absolute terms, wqomen represent about twqo-thirds of the "hidden unem-

vloyed." The incidence appears to be heaviest anmong teenage women. Over

-he first six years of the surveys, there is nc discernable trend in these

*roportions. The rates seem to fluctuate within a narrow band of t 1 per-

* _,ntagc point for men and a slightly wid6r band for women. These "ireasured"

'sponses are widely interpreted as representing accurate indicators of

)or :force status; therefore, it is argued that the measured open

.* ; Those proportions are equtivalent to about 2.5 and 8.5 percent of themale and fcmale measurcd labor force respectively.

4

- 46 -

unc?ploy,m,cnt ravces scriously understatc the "true" c::,Lont of unc-ploymcnt.

On the othcr lihnd, it may be arguted that. rcsponscs to such hypothetical

questions anout zv:'ailability for eniployivnt ar^ likely to bc mislcading,

that more persons are li'kely to answer in the affirmative thani rre actual-

ly available. Until recently, we had no way of testing the significance

of these responses. However, in a recent publication by the authors of

the survey, more detailed data were presentcd based on the survey of the

second quarter of 1978. 36/ In this publication the proportions of the

respondents who had been identified as the "hidden unemployed" were further

classified by their availability for immediate employment. Only a sniall

minority proved to be available, 4.1, 6.7, and.5.8 percent in the Mexico

City, Guadalajaa-, and 1Ionterrey metropolitan areas rcspectively. The

remainder posed conditions on their availability. For example, among

women the most frequently posed condition was of the nature, "if I can

find someone to take care of my house or children." Among males, the

greatest impediment to immediate acceptance of enmloyinent was their status

as students. This same status was of sccondary importance among women,

though ranlcing well behind household responsibilities as an obstacle. In

brief, then, severe doubts are posed about the reliability of the "hidden

unemployment" datuni as a true measure of unemployment.

36/ Secretaria de Prograinacln.n V PrcsUPLu0esto, Informicion Basica sobre laEstrLctLirn , C.r-7ctcrist .cais dol E;:4co v cl El-sc::ico en I-n- -,rcasMctroolit.f;-i ,- las Ci:i!UIs d.u !V,ico, (;undalajara, v >lonterrey(Dle.xico, D.F.: no date). Tables 59-61.

* \ - 47-

WIc notcd e:,]. i'Lr that o:nv apro;LOach to tLi' mcasutrcont of tli u.lder-

utillzi;tiun of . ibo.:iI,r forcce cl;u.si.fics tI; urndurcl;..plco'.'d all those workers

whose work woech: Ln tl:c referencc priod is slhorter than something zappro-xima-

ting a "norz-t;l" -.cc'ck of 40 ho rs. In Table 12 we present data on tlhe propor-

tionl of thlc eMp>-vec1( labor forcc .orking J.cs thIlan 40 hour:; a; rep)orte.d by th!e

labor force sur-:-y or the second quarter of 1973. The information is dis-

aggregated by rr-gion and sex. As can be secn, significant proportions of the

employcd labor force are employed less than 40 lhours per week. Atmong employed

men, the propor:ion varies over regions from 7.4 to 16.8 percent. Among

women, the propr!rtion varies between 18.2 and 47.3 percent. Ilowever, these

proportions cannot be taken to represent a measure ot underemployment since

a considerable proportion of those work.;ing short hours do so vNoluntarily.

Among men, from ha-f to twqo-thirds of those working short hours do so volun-

tarily. The range for w-omen is from almost two-thirds to three fourths. In

the final column I record the proportion of the employed labor force which

is involuntarily et:ployed for fewer than forty hours. As canl be seen, the

presumed level of underemployment is drastically reduced below the raw pro-

portions. Anong: ;ien and women, the ranges are 3.3 - 7,5 and 3.5 - 14.8 per-

cent respectively across regions. If these measures wqere to be reduced to

full-time equivalcncs, their significance would be reduced even further. An

interesting obse-atioll is the absence of a consistent relationship between

the degree of urb.nization of the region and the incidence of involuntary

underemploymcnt. One michlit have hypothesized that underemployment might

prove to be more severe in regions with rcl.atively larga agricultural popu-

L.ations since thc rural labor force is thought to excced by a considerable

margin the technizal requirements of current levels of agricultural output.

- 43s-

On the contrary -;c),o, regions ,.ith relativ.,ly large rural sectoI:s are am.nig?

those slio,.;ing the lo:;;-st rate:. of involinicary part-timio_ cnPlot::.cnt, at least

among men.

Thu-,, since past estimates of the ainount of undere;-;ployment have not been

able to tak'e account of the voluntary cofl-5onent ofL .ess-than-full-time employ-

ment, those estimates are likely to overstate th2 voIum.e of underutilization

at any point in tirme. Furthetmorc, since there is no ccmparability in the

measurcs of underutilization employed by the various censuses and surveys

taken at different poilnts in time, it is impossible to state writh any degree

of certainty either the direction or degree of' change that has occurred over

time.

IV. Conclusion

This essay has examined a range of erpirical info=ation which i-s rele-

vant to an evaluationi of the changes that have occurred in the "Iexican labor

market in the process of post-lWorld Wlar II economic development. Contrary

to the prevailing perceptions of a secular deterioration in the conditions

of employment, the data analyzed here point to a much more optimistic conclu-

sion. The rapid and sustained growth of the economy has been accomapanied by

a very substantial slift of the labor force from the agricultural sector,

where productivity is relatively low, to the secondary and tertiary sectors,

where productivitics are much higher. Furthermore, this large shift has been

aclhieved without depressing the productivity or wrages of labor in the expand-

ing sectors. Our analysis of the course of remunerations and produictivities

in those parts of the nonngricultLural sectors that are lik^ely to reflect thie

uninhibited forces of suppl)y and dermand in the labor market, recorded substan-

tial. improvements in botlh of these measures. Finnlly, I have set forth reasons

for beliievhig thiat tic> degree of Ul icrutiliz2tin' of labor rcsources m.,ay be

considerably sm.aller thall r-ost CstiMr:lLCS 'Lasve held.

To be sure, challeugiti- preble-,; rcrain to be addressed. The la'bor

force is growing at a rate approachilfl thrce perccnt per yc,ar. Its absorb-

tion at ever higher avcrage levels of productivity will require higli anld

sustained levels of invct!tm.cnt. The educational' endowinment of the labor

force is quite low in comparison to other semi-industrialized countries,

and much of the eduction provided is of questionable quality. Pockets of

abject rural poverty remain only marginally affccted by the clhanges occurr-

ing in the society at large. But these problems are less likely to appear

unsurmountable if proper recognition is given to the substantial improve-

ments that have been achieved in the past and a fuller appreciation is gained

of the way in which labor markets function.