19740 __ Viewfrom LATHR ~~~~~A No. 2:1 EDUCATION AND THE LABOR MARKET IN URUGUAY by GeorgePsacharopoulos and Eduardo Velez HumanResourcesDivision TechnicalDepartment Latin America and the CaribbeanRegion 7he World Bank June1992 'A Viewfrom LAIHR' is a series of occasional flyers produced by the Human ResourcesDivision of Latin America and the CaribbeanTechnicalDepartment of the WorldBankfor the purpose of stimulatingdiscussionamongstaff on key issuesfacing the sector. The views expressedhere are those of the authors and should not be attributedto the WorldBank. Abstract This paper uses data from the 1989 Uruguayan Household Survey to investigate the relationshipbetweenearningsand educationin that country. Mincerianearnings functionsfitted to nearly 10,000workers reveal that each extra year of schoolingyields a private rate of return of 9.2 percent, whichis comparableto the returns observedin the more industrialized countries. Among the other findingsof the study: Females realize a full percentagepoint advantageover maleson the return to their educationalinvestmentwhileprivate sector employeesenjoy a nearly five percentagepoint advantageover public sector employees -- a finding that highlightsthe recognitionof the productivevalue of educationby industry. When the full cost of education (both pubic and private) is considered and education is broken down by level of schooling, primary educationexhibitsthe highestrate of return -- nearly double that of secondaryeducation -- whereas graduates of technical/vocational schools and teacher training courses enjoy only minor returns on this type of investment. I I. Introduction Uruguayexperiencedrapid economicgrowth duringmost of the first half of the twentieth environmentduring the past coupleof decadeshas not century, but the overall macroeconomic been favorablefor educationaldevelopment. The average annualgrowth rate in the 1980swas negativein the industrysector and less than one percent in agriculture and services. Stagnation has also meantlittle employmentgrowth in the private sector -- in fact public employmenthas expanded to absorb labor force increases.Y' Central Governmentjobs increased from about 100,000in 1970to 163,000in 1975,accountingfor 26 percent of total employment.Sincethen, however, the share of public employmenthas decreasedto around 20 percent in the 1990s. The structure of the economyalso has important linkages to educationaldevelopment. In Uruguay the agriculturalsectorhas shrunk,losing 30 percent of the workersthat had in 1961; that year there were 210,700 workers while in 1986 agriculturalworkers numbered 153,000. The industrial sector is relativelysmallwith a share of 28 percent of the GDP in 1989. Uruguay is in fact less industrializedthan Latin American countries of similar income levels such as Venezuelaand Brazil, and even relative to countrieswith a lower income level, such as Chile, Mexico or Colombia. The industrial sector has decreased since the 1960s; the manufacture industry share of Montevideo's labor force, for example, dropped from 33 percent in 1970 to 23 percent in 1990. The sector service is the one that shows an increase in the last decades, 1/ Unemployment rates in urbanUruguaypeakedin 1983(14.7percentof the laborforce), sincethen theyhave been dropping to near 8.5 percentin 1990(DGEC,1991). -2- with a substantialpart due to the expansionof the informal sector. The trend toward a service economy is likely to fuel a growing demand for better educated workers. The educationalprofile of the workforceis in fact improving (Table 1). While those with no education, incomplete or completedprimary dropped significantly, the population having secondary and university education has risen sharply. In fact, Uruguay has one of the better educated workforcesin Latin America. Table 1. Labor Force by EducationalLevel, Montevideo, 1969-90 (percentage) EducationalLevel 1969 1975 1985 1987 1990 No education 2.3 1.5 1.1 0.7 0.6 Incompleteprimary 20.1 17.7 13.4 10.5 9.7 Completedprimary 37.2 35.2 25.6 25.7 23.4 Secondary 1st cycle 20.4 21.1 23.2 25.0 24.8 Secondary2nd cycle 5.1 6.6 10.9 10.6 12.7 Vocational(UTU) 6.0 7.5 11.9 11.6 11.7 University 5.7 8.5 10.6 11.7 12.9 Other 3.2 1.9 3.3 4.2 4.1 Source: DGEC, Encuesta de Hogares of respective years. Another aspect that has importantlinkagesto educationaldevelopmentis the demographic profile of the country which resemblesone of a developedcountry. The population growth rate has been extremelylow (0.6 percent per year) in the last decade and as low as 0.2 percent a year during 1970-75, while the average level for Latin America has been 2.6 percent. Of particular relevance to the education sector is that populationgrowth has been low not only because of low -3- natural growthbut also due to net emigration);the numberof school-agepopulation(6-17 years) increasedonly from 609,000 to 642,000 between 1980 and 1990. Current life expectancyis 72.5 years, significantlyhigher than the 68.6 years observed 25 years ago. The low net reproduction and the high life expectancyproducedan aging population,comparableto the one in developedcountries. Uruguay is a highly urbanizedcountry. Accordingto the 1985populationcensus, 86.2 percent lived in urban areas, maintainingthe pattern that has existed at least since the 1960s when over 80 percent of the populationwas living in urban areas. This trend is expected to continue as the projectionsfor 2000 indicate 91.1 percent will be living in urban areas. I. The Education System Uruguay has mandatoryprimary (six years)and lower secondaryeducation(three years). Preschooleducationhas developedrecently, and around 40 percentof childrenaged three to five were covered in 1989. Primary educationis universal. In fact, gross enrollmentin primary is more than 100 percent and net enrollment reached 88 percent in 1988 (see Unesco, 1991). Repetition, however, is notorious, mainly in first grade where it reaches near 20 percent. Secondary education, which lasts six years, reaches around 60 percent of the school-age population. The first three years constitute the lower secondary education (ciclo bdsico), and upper secondary education is diversified offering two options: a general, and a technical- vocationalcurriculum. University education,that was solely provided by the governmentuntil -4- 1984, has a coverage of about 2 percent of the (19-24) school-agepopulation (up from the 0.4 percent in the 1960s). Enrollment.During 1980-90primaryenrollmentincreasedfrom 331,247 to 351,452, with public schools expandingfrom 277,018 to 294,910, respectively (see Table 2). Enrollment in secondary academic increased from 125,448 in 1980 to 204,198 in 1990, a more significant increase, partially explained because secondary education attendance became mandatory, and because of a change in educationalemphasis from technical/vocational education that has kept a constant enrollment since 1983 (55,259 in 1983 and 56,084 in 1989). University enrollment doubled between 1973 and 1989 from 31,255 to 62,886. Table 2. Enrollment by Level of Education, 1972-90 EducationalLevel 1972 1980 1989 1990 Primary 345 331 350 351 - Public (285) (277) (296) (295) Secondary 145 125 197 204 - Public (116) (97) (161) (167) Technical/vocational 36 43 56 n.a. University 29 34 63 n.a. Sources: For primary and secondary, Departamento de Estadistica de la Divisi6n Planeamiento Educativo, CODICEN. For Technical/vocational and University, DGEC, Anuario EstadIstico. n.a. = Not available -5- Expenditure. Public educational spending amounts to only 2.6 percent of GDP (an increase from the 2.0 percent in 1983), placing Uruguay among the countries with the lowest public educationalspendingin Latin America. The pattern of intra-sectoral allocation--outof the US$211millionsof the 1989public expenditureson education, 41 percent went to primary education, 23 percent to secondaryacademic, 12 percent to vocational and technical, and 21 percent to higher education (19 percent for universityand 2 percent for teacher training). Table 3. Public Expenditureon Educationby Level, 1989 Level of Education US$ millions Percent of total Percent of GDP Primary 86 41 1.1 Secondary 74 35 0.9 - Academic (48) (23) (0.6) - Technical/Vocational (26) (12) (0.3) Higher Education 44 21 0.5 - University (40) (19) (0.5) - Teacher Training (4) (2) (0.0) Other 7 3 0.1 Total 211 100 2.6 Source: Ministerio de Economfa y Finanza (ContadurfaGeneral de la Naci6n), MEF-CGN,Budget Execution Statements. -6- Significantcuts in total educationspendingtook place in the early 1980s with a recovery by the end of the decade; overall, total education has grown in real terms. A comparison of expendituresin 1989 and 1980 shows that growth was relatively high in the public universities, less in secondary education and in technical-vocationaleducation, and slightly negative in primary education (see Table 4). Table 4. Public Expenditure on Education Index, 1980-89 (1984= 100, in constant prices) EducationalLevel 1980 1984 1989 Primary 141.6 100 136.8 Secondary 127.1 100 153.7 TechnicallVocational 131.7 100 136.7 University 137.8 100 202.0 Total 136.1 100 159.8 Source: Ministero de Economfa y Finanzas (ContaduriaGeneral de la Naci6n), MEF-CGN, Budget Execution Statements. Unit cost estimates appear in Table 5. In 1989, the annual expenditure per primary student was, on average, US$256, for secondary education US$306, for technical-vocational US$450, for teacher training USS$643, and for higher education it was US$614. At this last level there is ample variation by field of study (i.e., while expenditureper student in agronomy was close to US$2,000, in law, social sciences and economics it was between US$200 and US$300). For a detailed estimationof higher education costs, in particular costs per graduate, see Labadie, (1989). -7- Table 5. Unit Costs of Public Educationby Level, 1984-90 (Current US$/student) EducationalLevel 1984 1985 1986 1987 1988 1989 1989* 1990 Primary 127 130 181 226 239 256 (154.9) 265 Secondary 168 172 225 261 279 306 (185.1) 285 Technical/Vocational 235 263 353 425 436 450 (272.2) 419 Teacher Training 231 310 542 639 743 643 (389.0) 795 University** 355 433 643 655 618 614 (371.5) n.a. Source: CODICEN. For universitycosts MEF-CGN. * In 1989 Uruguayanpesos (in thousands). ** Excludes expenditures on the Hospital de Clfnicas (however 50 percent of the salaries of teaching personnel were included in the expenditure). Average expenditureper student includingHospital de Clinicasis US$896. n.a. = Not available. III. Labor Market and Education One concern of educationplannersis the linkage betweenthe educationalsystemand the labor market, specificallythe basic issue being the matching between the education system's output and the demand for educated labor. Two approachesare generally used to assess what would be externally efficient for the education system to produce: (a) manpower forecasting, (which, after decades of practice, has received repeated and sustained critique (see Youdi and Hinchliffe, 1985, and Psacharopoulos,1991,and World Bank, 1991)and has subsided), and (b) labor market analysis. This later technique,that presents a more reliable guide for educational investment, is used in what follows. -8- Sampleand data description.- The data used in this analysis are drawn from the 1989 Encuesta Nacional de Hogares conducted by the General Administrationof Statistics and the Census (DGEC). The surveyis based on a nationallyrepresentativehousehold sampleof 31,766 individualsconducted in urban areas. We selected those aged 14 to 65 with positive earnings from dependent employment (Y) (aggregate of all payments received from their wage employmentand earnings). Table 6 presents summary statisticsof the main variables used in the analysis. Fifty-nine percent of the sample are male workers, and the average age is a little over 37 years. Years of schooling (S) was constructed in two ways. As a continuous variable by combining the individual's highest level of formal education attended and the last grade completed at that particular level, and as a string of dummyvariables, indicatingthe fact that a person belongs to a respectivelevel. If an individualrepeated a grade, it is not reflected in our measures. With a mean of 8.61 years of education,the sampleis a relativelywell educatedone, with an average worker almost having completedlower secondaryeducation. One percent of the sample has no education, 38 percent has primary education, 35 percent has secondary education, 13 percent has some form of technical/vocational education, and 10 percent has higher education. Experience, constructed in the traditional Mincerian way (Age - S - 6) is 22.3 years. The number of hours workedper week is 45.5. The mean earnings per month is 150,680 pesos for each worker against 433,800 pesos average household income. With almost one in three workers in the public sector, Uruguay presents a case of a large public sector. -9- Table 6 Mean Sample Characteristics Variable Mean Std. Dev. Urban resident .98 .15 Male .59 .49 Age 37.01 12.88 Part-time Students .06 .24 Years of Schooling 8.61 3.58 Illiterate .01 .09 Primary education .38 .49 Lower education .24 .43 Upper secondary .11 .31 Vocational/Technical .13 .33 Teacher training .03 .18 University .10 .30 Private sector employee .66 .47 Public sector employee .29 .45 Years of experience 22.27 13.92 Hours worked per week 45.48 15.67 Earnings (000 pesos/month) 150.68 126.22 Total household income (000 pesos/month) 433.80 509.84 Source: Uruguay Household Survey, 1989. Persons with positive earnings from dependentincome. N - 9,417 Table 7 presents mean earnings of selected variables by gender and sector of employment. The sharpest earnings differentials are due to education, closely followed by gender, and then by sector of employment. Portes, Blitzer, and Curtis (1986) also found that gender, but mainly education, have significant additive effects on income. In their study, - 10 - employmentstatus (definedas formal worker, informal worker, and informalemployer), has the strongesteffect on income. Workers with higher educationexperience(includingdropouts) earn about 2.7 times more than illiterates, 1.9 times more than those who have primary education, and 1.5 times more than those who have secondaryeducation. Males, on average, earn about 60 percent more than females. Gender earnings differentialsare particularlylarge among illiterates,with male illiterates earning about 2.2 times more than female illiterates,but also among workers with higher educationexperience, in which category males earn 1.8 times more than females. This last finding is partially explained by the fact that in the last decades Uruguay is the only Latin American country where female enrollment has decreased in careers conducive to high salaries Oike engineering) at the same time that female enrollment has increased in careers conducive to low salaries like social sciences (see Schiefelbeinand Peruzzi, 1991). Earnings differentialsbetween public and private workers follows the pattern observed in Latin American countries. With the exception of upper secondaryand university education, the public sector offers higher pay relative to the private sector. Figure 1 presents the age-earningsprofiles by level of education. In spite of the saw- tooth pattern because of the low number of observations within each education-age cell (especiallyregarding older people with higher education), the level and growth of earnings in Uruguay is very similar to that observed elsewhere in the world. - 11 - Table 7 Mean Earnings by Educational Level (in thousand Pesos per month) Educational Entire Gender Economic Sector Mean N Level Sample Male Female Private Public Mlliterates 85 120 55 86 118 0.0 83 Primary 122 144 83 117 144 5.2 3,589 Secondary 154 181 112 151 165 9.4 4,525 - Lower 149 178 109 144 164 9.1 2,270 - Upper 169 222 125 174 162 10.7 1,027 Tech/Voc (UTU) 148 162 93 144 160 8.9 1,194 Higher 228 312 169 277 190 15.6 1,220 University (URU) 252 319 181 294 210 15.6 910 - Teaching 158 226 149 162 158 15.8 310 Overall 150 177 112 147 164 8.6 9,417 Note: Educational categories include dropouts of the respective level. - 12 - Figure 1. Age - Earnings Profiles by Educational Level (3- year MovingAverage) 400 400. 14~~~ i E 300 a X r 200 04 10 20 30 40 50 60 70 Age - 13 - Exploring earnings variation.- Table 8 reports Mincerian earnings functions (Mincer, 1974)2'fitted to the sampleas a whole and by gender and by economicsector. Column (1) is the classical specification which includes the continuous years of schooling and experience variables. We have also includedthe logarithm of hours worked per week as a compensatory of the first specificationconformwith human capital theory factor. The signof the coefficients and the explanatory power of the model (40.4 percent of the variance-') is consistent with previous research in the Latin American context. The negative sign of the squared term for experience reflects the concave age-earningsprofiles. Mincerian Rates of Return The rationale for restrictingthe last regressionto private sector males is in order to gain some insight to approximatethe returns to education in the competitivesector of the economy (assumingthat in the private sectorearningswould be closer to the productivityof the employee relative to the public sector), and in order to eliminateeffects of possiblediscriminationagainst females in the labor market. In addition, the Mincerian experiencevariable for females is not 2/ The standard is function Mincerian 2 + eLn(H)+ U, Ln (Y) = a + bS + cEX + dEX where: Y = monthly laborearnings, S = yearsof formalschooling, EX = yearsof workingexperience, H = hours workedper week, U = error term. varianceof the modelsrangefrom 29.8percentfor the publicsectorto 43.5 percent 3/ The explained for private sector male workers. - 14 - as good a measure for their actual labor market experience because of work interruptions for family reasons. For the Uruguayancase see Arends (1991). The coefficientsof the years of schoolingvariablein columns (1) to (5) in Table 8, and the differencesbetween successiveeducation dummy coefficientsin column (6) give us a first glimpse on the returns to education in Uruguay, either referring to a typical extra year of education, or to specific educationallevels. These are summarizedin Table 9. The overall Mincerianrate of return (whichis private by construction)is 9.2 percent. This value is typical of that in advanced countries over the last 20 years. Females enjoy a one percentage point advantageover males-- another typicalresult in most countries. What is of extremeimportance is the fact that the returns in the private sector of the economyexceed those in the public sector by almost 5 percentage points. Such finding, as elsewhere, gives confidenceon the productive role of education in the sense that no private employer would maintain in the payroll more educatedpeople if their wages did not somehowcorrespondto increased productivity. Amongthe differenteducationallevels, lower secondary(whichin this case is the last cycle of compulsory basic education exhibits the highest rate of retum, 13.1 percent. The returns thereafter drop by the level of further education, the lowest being those for teacher training (negative). Secondarytechnical vocationaleducation exhibits a very low rate of return of 1.4 percent, which is significantly lower than what has been found elsewhere (for the Venezuelan case see Fiszbein and Psacharopoulos, 1992). - 15 - Regarding teachers, they receive a very low premium, as in most Latin American countries, and this may be attributed to the part-time nature of their profession.' In a recent census (Censo Nacional de Docentes de Educaci6n Media) for example, only one in three teachers mentionedthat his/her salary as a teacher was the main source of family income (see CODICEN, 1990). Perhaps teachers enjoy non-monetaryrewards we were not able to capture in the incomevariable (e.g., free housing). One shouldalso note the high unit cost of teachers' education due to the decliningenrollmentsin normal schools (enrollmentsdeclinedfrom 5,287 in 1985 to 2,361 in 1990). This has resulted in underutilizedhuman and physical capacity in the normal school; some normal schoolscurrently have enrollmentsas low as 20 studentswith a capacity to attend 200 studentsor more. 4/ Accordingto the survey, teachers work on average 33.5 hours per week, whereas other professions work between 43 and 49 hours per week. - 16 - Table 8 Mincerian Earnings Functions Variable Entire Gender Economic Sector Private Sample Males Females Private Public Mes Males Constant .352 1.348 .460 -.103 1.934 1.934 Years of Schooling(S) .092 .091 .102 .109 .060 Experience (EX) .045 .056 .041 .050 .030 .069 EX-squared -.0006 -.0007 -.0006 -.0007 -.0003 -.0001 Log Hours .803 .551 .715 .873 .530 .642 Primary (Base)* .000 Lower Secondary* .394 Upper Secondary* .688 Technical/Vocational (UTU)* .352 Teacher Training* .471 University (URU)* 1.158 R2 .404 .388 .418 .434 .298 .435 N 8,623 5,064 3,560 5,865 2,562 3,313 Mean 5 8.7 8.3 9.2 8.1 9.6 8.1 Note: All coefficientsare statisticallysignificantat the 1 % level or better. * Dummy Variables Regarding technical/vocational education,the low returns must reflect current scarcities in the labor market. Sapelli (1988) shows that the stagnation of the manufacturingsector had adverse implicationsfor job opportunitiesfor UTU graduates. He also notes that enrollment in private sector training programs was increasing,probably as the result of a decline in quality of UTUs programs. - 17 - The figures shownin columns(2) and (3) indicate that private rate of returns are higher for females (10.2 percent versus 9.1 percent) even though females earn less. This finding has been consistently found in the literature (Psacharopoulos,1985) and it is due to the lower foregone earnings of females. It has also been reconfirmed in the case of Uruguay (Arends, 1991). Experiencehowever is more rewarded for males. In column (4) and (5) the results by economic sector indicate that the returns to education are significantlyhigher in the private sector (10.9 percent) than in the public sector (6.0 percent). This result supports the productivity-enhancing role of education since the more competitivesector is rewarding more the more educatedworkers. Also consistentwith previousanalysis is that returns to experience (i.e., growth of earnings) are lower in the public sector. Finally, column (6) is an expandedearnings function for public sector males- where schoolingis disaggregatedinto a series of dummyvariablesto estimatereturns to investmentin the various levels of education. The results of the modifiedspecificationare consistentwith the previous pattern. Although returns continue to rise with educational level the trend is not smooth (i.e., returns to male teachers in the private sector are lower than for workers with secondary education). Full method. The number of observations available in this sample permit us to also estimate the returns to educationusing the discountingformula, i.e. finding the interest rate that brings the discountedactual (and non-regression-smoothed) net age-earningsprofiles equal to zero. In such case we can also add the social cost of providing education at a given level to the beginningof the age-eamingsprofiles and thus estimate the rates of return from the social view point. - 18 - Table 9 Mincerian Returns to Education (percent) Reference Private Rate of Return (percent) Gender Males 9.1 Females 10.2 EconomicSector Private 10.9 Public 6.0 Private Sector Males by EducationalLevel Lower secondary(vs. primary) 13.1 Upper secondary (general) (vs. lower secondary) 9.8 (vs. primary) 11.5 (UTU) (vs. lower secondary) Technical/Vocational 1.4 (vs. primary) 5.9 Teacher Training (vs. upper secondary) negative University (vs. upper secondary) 9.4 The results appear in Table 10. As elsewhere,primary educationexhibitsthe highest rate of return, secondary educationa lower rate, and university educationthe lowest rate among the three main levels. Within secondaryeducation,the technicalfield exhibits a lower rate of return relative to general education. As already found with the Mincerian method, teacher training shows a negative return. - 19 - Table 10 The Returns to Education -- Full Method (percent) EducationalLevel Private Social Primary (vs. illiterate) 19.1 15.2 Secondary(vs. primary) 9.8 8.0 - General 11.4 8.5 - Technical/Vocational (UTU) 8.2 6.2 Teacher Training (vs. secondary) negative negative University (URU) (vs. secondary) 8.1 6.5 Unemployment One issue is how educationrelates to the chancepeoplehave to be unemployed,and once unemployedhow long they have to wait for finding a job. In order to answer these questions we worked with a larger samplefrom the householdsurveythat includesnot only those who are employed for labor earnings (used in the above analysis), but also those who are unemployed and are looking for a job. This samplein fact corresponds to the definition of the 14-65 years economicallyactive populationor labor force. rate at the time of the surveywas 5.7 percent. But as shown The overall unemployment in Table 11, the level of education a person has relates to his/her chance to be unemployed. rate (7.9 percent), followedby lower secondaryschool Illiterateshave the highest unemployment graduates. Upper secondary school graduates have about the same unemploymentrate than secondarytechnicalgraduates, 6.6 and 6.2 percent respectively. Confirminga pattern observed - 20 - at least since the mid 1980steachers have the lowest unemploymentrate (1.2 percent). Due to the above mentioneddeclinein normalschoolenrollmentonly around 500 new teachers graduate from normal schoolseveryyear. This numberis relatively small to replace the teachingworking force that on average has been in the education system for more than 16 years (the retirement rate is relatively high). The second column in Table 11 refers to the sub-populationof those in the labor force who are either unemployedor are lookingfor work. The mean waiting time among this group is 31 weeks. But as shown in Table 11, there is strong differentiationby level of education. University graduates seem to search longer than any other group. This could be partially explained by the ability of higher income families, who are over-represented among the university-educated,to support their children until they find a suitablejob. It also reflects the low cost of attending university. The last two columns of the table show that the average number of weeldy hours by the urban labor force was 42.6 (45.5 including total jobs), with teachers and illiterate people working less and secondarytechnicalgraduates working more. It is very difficult with the data in hand to determine the extent to which the observed incidenceand length of unemploymentis involuntary (genuine joblessness) rather than reflecting job search, i.e. people voluntarily remaining unemployed in order to improve the wage offer they will get. Although some voluntary unemploymentcould be attributed to the university graduates, this theory would be very hard to support regarding those with lower levels of education. - 21 - Table 11 Unemployment Characteristics of the Active Population by Educational Level Educational Level Unemployment Looking for Average Weekly Hours in Rate Work (percent) (weeks) Primary Job All Jobs Illiterate 7.9 26.6 38.8 40.1 Primary 5.0 26.9 44.3 46.0 Lower Secondary 7.2 30.1 43.6 46.0 Upper Secondary 6.6 32.5 41.1 43.5 Secondary Technical (UTTU) 6.2 29.4 45.8 49.1 Teacher Training 1.2 25.0 28.5 33.2 University (URU) 3.8 44.9 36.6 44.1 Overall 5.7 30.6 42.6 45.5 (N) (13,600)!y (1,001DV 8,650 8,667 Population 14-65 years old, employed, unemployed, or looking for work for the first time. P' b/ Unemployed or looking for work. =Equity The above analysis referred exclusively to efficiency issues in the labor market, i.e. how education relates to employment and the earnings of those who have been educated. With the data set in hand it is possible to expand the analysis to equity issues by concentrating on a different sub-sample: those who are 10 years old and above and report themselves as being students. The household survey contains 5,129 such persons the characteristics of which appear in Table 12. - 22 - The mean age of the different student groups documents the extent of grade repetition in the Uruguayansystem of education,e.g., whereas the mean official age of those who are in upper secondaryeducation shouldbe 17 years old, the mean actual age of the group is 20. The same applies to university students, their actual mean age being 25 against an official age of about 21. Column (2) in Table 12 shows the mean household income of the respective student groups. There is a distinct pecking order with those who attend university having double the family income of those who are in primary education. It is also of extreme interest that the mean household income of those who study technical/vocationalsubjects is the lowest among education is for the poor. all groups. In Uruguay, as elsewhere, technical/vocational The respondents were asked whether they contribute payments in the school they currently attend. The figures shown in the last column of Table 12 show an amazing stratificationof the incidence of paying by level of education. Those attending primary and lower secondaryschoolsare more likely to contributefor their education,whereas only one tenth of one percent contribute to their university education. This is prima facie evidence on the inequity of the present education financing arrangementsin Uruguay. - 23 - Table 12 Characteristicsof the StudentPopulation EducationalLevel Age Household Fee Paying N Income/month Student (000 Pesos) (percent) Primary 11.9 380 12.6 2,209 Lower Secondary 15.3 472 17.0 1,661 Upper Secondary 20.0 552 9.9 518 SecondaryTechnical(UTU) 18.6 339 0.0 262 Teacher Training 23.6 485 2.9 34 University(URU) 25.1 665 0.9 445 Overall 15.4 450 12.1 5,129 V. EducationalInvestmentPriorities The above results are signals as to where resources for education could be used most profitably at the margin. The expansionof primary education should be priority number one. It is reminded that in spite of the gross enrollment ratio of more than 100 percent, the net enrollmentratio was only 88 percent in 1987. Althoughboth, private and social rate of return are somewhatlower than the average found for Latin American countries(see Psacharopoulos, 1985), they are high enough to indicate the high priority that should be given to investmentin this sub-sector. The expansionof primary education is also good for equity purposes. General secondaryeducationalso exhibits a sizeable social rate of return indicatingthat there is room for expansion. The relatively low social rate of return to technical/vocational secondary education is a red flag regarding further expansion of this level, especially in its - 24 - present form. Perhaps the quality or "relevance"is bad. We were not able to differentiate by specialty. This needs further investigation. It is recommendedthat systematicanalysisof labor market informationtogether with tracer studies be used to improve the imbalances betweenthe technical schools and the labor market. The overall social rate of return of 6.5 percent regarding university education, similar to that in countries such as the United States, indicates that the country might have reached equilibriumregardingthis level of education. The precedingmacro-analysis indicatesthat higher education may not be an investmentpriority. The negative rate of return to teacher training requires further investigation. - 25 - REFERENCES Arends, M., "Women'sLabor Force Participationand Earnings: the Case of Uruguay," in G. Psacharopoulos and Z. Tzannatos (eds.), Women's Employment and Pay in Latin America, The World Bank, 1992 (Forth.) CODICEN,Diagn6sticoy Polfticasde Formaci6ny PerfeccionamientoDocente,Administraci6n Nacionalde Educacion Paiblica,ConsejoDirectivo Central, Montevideo, 1990. DGEC, Encuesta Continua de Hogares 1990, Direcci6n General de Estadfstica y Censos, Montevideo, 1991 Fiszbein, A., and Psacharopoulos,G., "A Cost Benefit Analysis of EducationalInvestmentin Venezuela, 1989," Economicsof EducationReview, 1992 (forth.) Labadie, G.J., El Gasto Universitarioy el Costo de los Egresados,Serie Descripci6ne Indices, CERES, Montevideo, 1989. Mincer, J., Schooling,Experienceand Earnings,NationalBureau of EconomicResearch, 1974 Portes, A., Blitzer, S., and Curtis, J., "The Urban Informal Sector in Uruguay: Its Internal Structure, Characteristics,and Effects," World Development, Vol. 14, No. 6,pp. 727- 741, 1986. Labor Psacharopoulos,G., "From ManpowerPlanningto Labor Market Analysis,"International Review, November 1991. Psacharopoulos,G., "Returnsto Education: A Further InternationalUpdate and Implications," Journal of Human Resources, Vol. 20, pp. 583-601, 1985. Sapelli, C., Evaluaci6n de la Conexi6n entre Sistema Educativo y Mercado de Trabajo, con Enfasis en UTU y la Universidadde la Repiublica.Manuscrito, CERES, Montevideo, 1988. Schiefelbein, E., and Peruzzi, S., "Oportunidadesde Educaci6n para la Mujer. El Caso de America Latina y el Caribe," in Bolettn Proyecto Principal de Educaci6n en America Latina y el Canibe,Unesco-Orealc,No. 24, pp.51-78, 1991. UNESCO, Statistical Yearbook, 1991, Unesco, Paris, 1991. Youdi, R., and Hinchliffe, K. (eds.), The Practice of Manpower ForecastingRevisited, Paris: Unesco, InternationalInstitute of EducationalPlanning, 1985. World Bank, Vocationaland TechnicalEducation and Training:A Policy Paper, 1991. I - Continuedfromback Page No. 16 'Wiat do we think aboutHealth CareFinance in Lain America and the Caribbean? by Philip Musgrove, September1991 No. 17 'Population Growth,Externalitiesand Poverty' by Nancy Birdsall and Charles Griffin, September1991 No. 18 'Wage Trends in Latin America' by AlejandraCox Edwards, September 1991 No. 19 'Investment in ScienceResearchand Training:The Caseof Brazil and Implications for Other Countries' by Laurence Wolff, with contributionsfrom George Psacharopoulos,Aron Kuppermann,CharlesBlitzer, GeoffreyShepherd, Carlos Primo Braga and AlcyoneSaliba,September 1991 No. 20 'Prenatal and PerinatalHealth Care:A DiagnosticInstrunent' by Francisco Mardones, September1991 No. 21 'MaternalAnthropometry in PrenatalCare:A New Maternal Weight Gain Chart'by Pedro Rosso, September1991 No. 22 'Povertyand Inequalityin Latin Americaand the CaribbeanDuring the 70s and 80s: An Overview of the Evidence' by Dominique van de Walle, September 1991 No. 23 'Social Indicatorsin Latin Americaand the Caribbean.A Compilationof Statistics from 1970 to the Present' by George Psacharopoulosand Bill Wood, October 1991 No. 24 'ICEEX - A StudentLoan SuccessStory in Colombia' by SamuelCarlson, October 1991 No. 25 'EducationalDevelopmentand Costingin Mexico, 1977-1990: A Cross-StateTune- Series Analysis' by Juan Prawda and George Psacharopoulos,November1991 No. 26 'A Cost-BenefitAnalysisof EducationalInvestmentin Venezuela,1989' by Ariel Fiszbein and GeorgePsacharopoulos,November1991 No. 27 'EducationalDecentralizationin Lain America:LessonsLearned' by Juan Prawda, March 1992 No. 28 'Education and the Labor Marketin Uruguay' by GeorgePsacharopoulosand Eduardo Velez, June 1992 Viewsfrom LATHR No. 0 "The Magnitudeof Poverty in Latin America in the 1980s" September, 1990 No. 1 'An Ounce of Preventionis WorthHow Much Cure? Thinkingabout the Allocationof Health Care Spending' by Philip Musgrove, September1990. No. 2 "Decentralization and Educational Bureaucracies"by Juan Prawda, November, 1990 No. 3 "WhiatShould Social FundsFinance?:PortfolioMix, Targeting,and Efficiency Criteria"by Margaret E. Grosh, December 1990 No. 4 Balance in Chile:7he ISAPRES anstitucionesde Salud Previsional)Health "Financial CareSystem and the Public Sector" by Philip Musgrove,January, 1991 No. 5 'Population,Healthand NutritionIssues in the Latin Americanand CaribbeanRegion and the Agendafor the 90's" by Oscar Echeverri,January, 1991 No. 6 "Populationand Family Planningin the 1990's: ReconcilingMacro and Micro Issues' by Bruce D. Carlson, February, 1991 No. 7 "TheFeasibilityof StudentLoans in Latin America: A Simulation"by Samuel Carlson and GuozhongXie, March, 1991 No. 8 'Transformingthe Vicious Circle- The Costsand Savings of School Inefficiencyin Mexico by SamuelCarlson, April 1991 No. 9 "Colombia's "EscuelaNueva". An EducationInnovation"by Eduardo Velez, May 1991 No. 10 wHealthTechnologyDevelopmentand Assessment:Do LAC Countimes Have a Choice?"by Oscar Echeverri,June 1991 No. 11 "TheRecurrent CostFactor in the PHR Sector" by Jacob van LutsenburgMaas, July 1991 No. 12 "TheBurden of Death at Different Ages: Assumptions,Parametersand Values" by Phillip Musgrove, August 1991 No. 13 "Government Expenditureon Social Sectors in Latin America and the Caribbean: Statistical Trends" by HongyuYang, August 1991 No. 14 "FromManpowerPlanningto Labor MarketAnalysis" by George Psacharopoulos, September1991 No. 15 "An Update on Cholerain the Americas' by FranciscoMardones, August 1991 - Continued on inside Page