Labor Markets Ethnic in Peninsular Malaysia, 1970

20
The Journal of Developing Areas 18 (October 1983) 1-20 Labor Markets and Ethnic Inequality in Peninsular Malaysia, 1970 CHARLES HIRSCHMAN The classical sociological theory of race and ethnic relations suggests that assimilation-the withering of divisions based upon race and ethnicity-is the long-term expectation in modem societies. This trend is supposed to reflect modem industrial organization, where social mobility is based upon achieved rather than ascribed status. This theory (often labeled as assimilation theory or the industrialization hypothesis) has found only modest support in the decades following its formulation. There is considerable evidence of persistent ethnic inequality and conflict in many countries at all stages of development throughout the modem world.’ The search for an alternative paradigm of race and ethnic relations has not led in a common direction. There have been numerous the- oretical contributions to the study of race and ethnic relations over the past two decades, but none has emerged as dominant.2 Some of these writings have at- tempted to posit alternative paths of evolutionary change in multiethnic soci- eties, while other theorists suggest a broader array of causal variables that in- fluence the relative degree of ethnic inequality in societal rewards (earnings, status) or power. The hypothesis of socioeconomic (or structural) assimilation remains the keystone of much research, but it is expected to be a function of political forces, labor-market structures, and other institutional factors as well as the functional needs of a modern industrial society. Drawing upon both the classical and alternative theoretical frameworks, the present analysis examines the determinants of ethnic inequality in Peninsular Malaysia. The analysis is based upon 1970 census data and incorporates both individual human-capital variables and the structural characteristics of local labor markets. The results suggest that structures of opportunity, measured by labor-market characteristics, explain part of the ethnic differentials in socio- economic roles in the plural society of Peninsular Malaysia. Professor, Department of Sociology and Asian Studies, Comell University. This article is a revised version ofa presentation to the annual meeting ofthe Population Association of America, 26-28 March 1981, Washington, DC. The author gratefully acknowledges a research grant from the National Institute of Mental Health (Social Change in Malaysia, MH-30663), the excellent computer programming of Robert Jackson, and the invaluable comments of John Gartrell and Richard Campbell on an earlier draft of this article. c 1984 by Western Illinois University.

Transcript of Labor Markets Ethnic in Peninsular Malaysia, 1970

Page 1: Labor Markets Ethnic in Peninsular Malaysia, 1970

The Journal of Developing Areas 18 (October 1983) 1-20

Labor Markets and Ethnic Inequalityin Peninsular Malaysia, 1970

CHARLES HIRSCHMAN

The classical sociological theory of race and ethnic relations suggests thatassimilation-the withering of divisions based upon race and ethnicity-is thelong-term expectation in modem societies. This trend is supposed to reflectmodem industrial organization, where social mobility is based upon achievedrather than ascribed status. This theory (often labeled as assimilation theory orthe industrialization hypothesis) has found only modest support in the decadesfollowing its formulation. There is considerable evidence of persistent ethnicinequality and conflict in many countries at all stages ofdevelopment throughoutthe modem world.’ The search for an alternative paradigm of race and ethnicrelations has not led in a common direction. There have been numerous the-oretical contributions to the study ofrace and ethnic relations over the past twodecades, but none has emerged as dominant.2 Some of these writings have at-tempted to posit alternative paths of evolutionary change in multiethnic soci-eties, while other theorists suggest a broader array of causal variables that in-fluence the relative degree of ethnic inequality in societal rewards (earnings,status) or power. The hypothesis of socioeconomic (or structural) assimilationremains the keystone of much research, but it is expected to be a function ofpolitical forces, labor-market structures, and other institutional factors as wellas the functional needs ofa modern industrial society.Drawing upon both the classical and alternative theoretical frameworks, the

present analysis examines the determinants of ethnic inequality in PeninsularMalaysia. The analysis is based upon 1970 census data and incorporates bothindividual human-capital variables and the structural characteristics of locallabor markets. The results suggest that structures of opportunity, measured bylabor-market characteristics, explain part of the ethnic differentials in socio-economic roles in the plural society of Peninsular Malaysia.

Professor, Department of Sociology and Asian Studies, Comell University.This article is a revised version ofa presentation to the annual meeting ofthe Population Associationof America, 26-28 March 1981, Washington, DC. The author gratefully acknowledges a researchgrant from the National Institute of Mental Health (Social Change in Malaysia, MH-30663), theexcellent computer programming of Robert Jackson, and the invaluable comments ofJohn Gartrelland Richard Campbell on an earlier draft of this article.

c 1984 by Western Illinois University.

Page 2: Labor Markets Ethnic in Peninsular Malaysia, 1970

2 Charles Hirschman

Plural Society of Peninsular Malaysia3

Located at the maritime crossroads of Asia, the Malaysian peninsula hashistorically been settled by peoples from distant lands, perhaps most influentiallyfrom the Indian subcontinent.4 The countryside has been populated by the in-digenous Malay population, who were primarily engaged in wet-rice cultivationand fishing along the major river valleys of the peninsula. Further inland, inthe uplands and mountains are small aboriginal populations who were pushedto the interior when the Malay population settled the lowlands about 1,000-2,000 years ago.5Not until the nineteenth century were the roots of the present-day plural

society firmly implanted. British colonialism, which began in the eighteenthcentury on the islands of Singapore and Penang and the historic coastal city ofMalacca (these three initial colonies were labeled the Straits Settlements), ex-panded long-distance trade and supported early capitalist ventures in agricul-ture.6 Perhaps more consequential during this era was Chinese entrepreneurialpenetration, which was primarily directed toward tin mining and agriculturaldevelopment. Early Chinese capitalists, often working in close alliance withMalay rulers, began to import large numbers of Chinese laborers to develop thegrowing opportunities in mining and plantation agriculture.7 By the middle ofthe nineteenth century, Chinese immigration had reached substantial levels,often numbering in the tens of thousands annually. While the numbers of em-igrants were as high in many years, a growing number of Chinese remained tosettle permanently. In the late nineteenth century, the British officially extendedthe colonial mantle to most of the Malaysian peninsula. This period coincidedwith the penetration of British economic interests to the interior (plantationsofcoffee, pepper, sugarcane, for example, and mining were the most important).8Since few of the Malay population were willing to serve as cheap labor in theearly plantations and mines, the pace of immigration accelerated in the lastdecades of the nineteenth century. In particular, the British promoted the im-portation of labor from India, which was also a British colony.9 Immigrationof Chinese laborers continued, as well as a sizable migration stream from Su-matra, Java, and other islands of the then Dutch East Indies-present-day In-donesia.10With the development ofthe rubber industry in the last years ofthe nineteenth

century, Malaya’s economy was transformed." The growth of the automobileindustry in the West created the rubber boom of the first two decades of thetwentieth century, making Malaya the world’s leading producer of rubber anda very profitable colony. Migration from India and China grew apace.12 By 1931,the population of British Malaya, excluding Singapore, consisted of approxi-mately 50 percent Malay, 34 percent Chinese, 15 percent Indian, and 2 percentothers (European, Eurasians, and other small Asian populations).13 As immi-gration was slowed by the Depression of the 1930s and virtually ceased afterWorld War II, a settled Chinese and Indian population gradually began to takeroot in Malaya. In 1970, 53 percent of the population were identified as Malay,35 percent as Chinese, 11 percent as Indian, and percent in the residual categoryofother. More than 8 out ofevery 10 Chinese and Indian residents ofPeninsularMalaysia were bom in Malaysia or Singapore.14As the plural society of Malaya developed in the twentieth century, sharp

ethnic divisions were created in residential, economic, and social patterns. Whilethe bulk ofthe Malay population remained in rural villages, Chinese and Indian

Page 3: Labor Markets Ethnic in Peninsular Malaysia, 1970

Labor Markets and Ethnic Inequality in Peninsular Malaysia, 1970 3

immigrants settled on plantations and in tin-mining towns. As urban centersgrew and developed in response to the economic changes underway, they werelargely populated by Chinese traders and workers.15 Because of a variety ofbarriers, some resulting from cultural and linguistic differences and some fromofficial colonial policy, very few Malays were able to enter the modem urbaneconomy." The British encouraged the sons ofthe traditional Malay aristocracyto becomejunior officers in the colonial bureaucracy but advocated a traditionalway of life for the rural Malay masses. Opportunities for higher education,especially English education-the pathway to social mobility-were extremelylimited for the Malay population.17At the time of formal independence from England in 1957, Malaya was a

deeply segmented plural society. The majority ofMalays were small-scale farm-ers in rural villages, with rice and rubber as their primary crops.18 While a sizableminority ofChinese were employed as wage laborers on plantations and in otheragricultural pursuits, the majority were in urban areas in the secondary andtertiary sectors. Most Chinese had very low incomes, but their concentration inurban areas and an almost ubiquitous presence in commercial activities createdthe impression of their being the "businessmen" ofthe country. The small (andvery heterogeneous) Indian community occupied an intermediate position. Aboutone-half of all Indian men were laborers on rubber (and oil palm) plantations,and the other half were engaged in a variety of manual and nonmanual occu-pations. A higher proportion of Indians than Malays were engaged in "middle-class" occupations in business and the professions.

In addition to the substantial socioeconomic barriers between ethnic groups,there were also major cultural divisions-for instance, each group retained itsmother tongue. English became the medium of communication among thosewith higher education, and "market place" Malay became the lingua franca ofeveryday exchange across ethnic lines." In addition to the language barrier,religion and food preferences (and prohibitions) inhibited the development ofclose interpersonal links across ethnic lines. Most Chinese and Indians, evenafter several generations ofMalaysian residence, retained a strong sense ofethnicidentity, though this was often independent ofany sense of attachment to theirancestral homeland.The reins of political power passed from the colonial regime to the present

multiethnic coalition of political parties in 1957, but authority has been largelyconcentrated in the dominant Malay party.20 The formation ofMalaysia in 1963and a wider group ofparties in the ruling coalition have broadened the politicallandscape, but the basic political framework has not changed. From 1957 to1969, the government tried to foster the relative status ofthe Malay communitythrough expansion of educational institutions and development programs inrural areas. After the 1969 outbreak of ethnic disturbances in Kuala Lumpur(following a particularly emotional election), government policy shifted towardmore direct intervention in the economy to assist the Malay population.21 Onegoal of the "New Economic Policy" is to eliminate the identification of race(ethnicity) with economic activity. In the present study, I only consider thestructure of ethnic inequality in 1970, prior to this major change in policy.

Models of Ethnic Stratification

In recent years, the study of ethnic stratification has drawn upon the logicand methods of the emergent paradigm of social stratification.22 The basic ob-

Page 4: Labor Markets Ethnic in Peninsular Malaysia, 1970

4 Charles Hirschman

jective in such studies is to decompose the difference (gap) in mean levels ofsocioeconomic attainment between two populations in terms of resources, op-portunities, and discrimination." Discrimination is generally measured indi-rectly as the residual gap, after all other social background and human-capitalvariables are held constant in a multivariate analysis. This research strategy isused to measure net inequality as a social indicator24 as well as in comprehensiveanalyses that interpret the impact of various independent variables on racial orethnic differentials in socioeconomic achievement.25 While the usual model forsuch an analysis is the socioeconomic life cycle26 and human-capital variables,there is a wide range of theoretical propositions that can be tested within thisframework.27

Prior research on Peninsular Malaysia has shown that the largest share of theethnic gap in educational attainment (years of schooling) and income (monthlyincome) can be accounted for by differential social and economic origins.28 Spe-cifically, most of the Malay disadvantage results from their predominant ruraland agrarian background. The children of farmers and those who were bom inrural areas, in general, are less likely than their urban-reared cousins to achievehigher socioeconomic positions. Controlling for the additive effects of thesebackground variables reduces the Malay-Chinese gap in educational attainmentby 80 percent and the income gap by 65 percent.29 The occupational attainmentprocess is more complicated. For occupations that are dependent upon formaleducational qualifications, differential social origin was the primary explanationof ethnic inequalities. For other occupations that can be inherited (throughproperty), such as sales and some blue-collar occupations, the "net" ethnic gap(after controlling for background variables) was almost as strong as the originalgross difference. It seems that employers, especially in the small-shop sector thatdominates retail trade, are more likely to hire on the basis of ascriptive criteriaof ethnic identity.30The present study extends the earlier work (with more comprehensive data-

the 1970 Population Census) to include a neglected set ofexplanatory variables.In addition to the individual background variables that might be conceptualizedas resources (or human capital) that individuals carry with them, I include theexternal attributes of labor markets (conceptualized as opportunity structures)as possible explanations of ethnic inequality. In the standard model of socialstratification, one makes the simplifying assumption that there is only one op-portunity structure with an open competition in the exchange oflabor and wagesin the labor market. While this may be true in certain occupations for whichthere is a national system ofvacancy advertising and recruitment, most workersare probably restricted to the opportunities available in the proximate geograph-ical area. To the extent that different ethnic minorities are located in differentplaces (which are unequal in relative opportunities), unequal socioeconomicattainment may be the outcome. This hypothesis is explored with a focus onthe employment sectors with the widest ethnic differentials in Peninsular Ma-laysia.There is a small but growing literature on the effects of labor markets upon

relative levels ofethnic inequality.31 Perhaps most well known is the "city effectshypothesis," which posits that the size and structure of cities influence the pro-cesses of stratification.32 Another variant is the "percent minority hypothesis,"which suggests that the relative level of discrimination is partially dependentupon the minority proportion of the local population.33 Empirical analyses of

Page 5: Labor Markets Ethnic in Peninsular Malaysia, 1970

Labor Markets and Ethnic Inequality in Peninsular Malaysia, 1970 5

the effects of labor markets upon racial stratification in the United States haveyielded different results. Stolzenberg and D’Amico claim that city differences inracial stratification are minimal, while Parcel finds important labor-market ef-fects on racial inequality."

Data and Methods

The present analysis is based upon a 2-percent sample ofthe 1970 PopulationCensus of Peninsular Malaysia.35 Internal evaluation and comparisons with in-dependent data sources suggest that the 1970 Population Census data are ofgood quality. A postenumeration survey reported a census undercount ofabout4 percent of the total population.36For this analysis, ecological or labor-market characteristics were appended to

each individual person record in the data file. Population counts (by ethniccommunity) of the town of residence were added to the individual records ofall persons living in urban areas, as well as the socioeconomic characteristicsof each town (the industrial structure, an indicator of educational attainment,and the level ofmigration).37 These socioeconomic characteristics were estimatedfor the 82 towns with a 1970 population above 5,000 (these data are not availablein published reports). The labor-market variables ofthe industrial structure werebased upon a census question that asked for "usual industry" (as distinct fromthe typical labor-force measure based upon activity during the preceding week).The same set of characteristics (industrial structure, etc.) were also estimatedfor the 70 administrative districts of Peninsular Malaysia (districts are similarto counties in the United States). An additional measure was the size of thelargest city in the district. These labor-market variables were appended to eachindividual’s record as contextual characteristics. Thus each person’s record con-tains information on individual attributes as well as the basic structural featuresof the local area (labor market) in which he/she resides. For the present paper,only adult males between age 25 and 64 who were in the experienced labor forceand reported an occupation, industry, and employment status (based upon thelast week or previous job) were selected as respondents (N 24,443).

In the following analysis, I first describe the major axes of ethnic socioeco-nomic cleavage, with cross-tabulations of ethnicity by occupation and industryin 1970. Then, I estimate multivariate models with a focus on the ethnic gapin three key areas: the proportions in white-collar occupations, in the commercialsector, and in the manufacturing sector.

Ethnic Differences in Occupational and Industrial Structures

The major differences in economic roles between Malay, Chinese, and Indianmen in Peninsular Malaysia are shown in tables and 2, which contain theindustrial and occupational composition ofthe experienced labor force in 1970.38The most fundamental ethnic difference is in the agricultural sector. More

than 60 percent of Malay men are in the agricultural sector (about equallydivided between rice and rubber), while one-quarter of Chinese men and littleover one-third ofIndian men are in agriculture (rubberand oil palm plantations).Mining and the secondary sector of manufacturing and construction are indus-tries that are disproportionately Chinese. Malay men are substantially under-represented in commerce. Only 5 percent of Malay men are engaged in com-merce, but over 20 percent of Chinese and 15 percent of Indians are in thissector. Industrial divisions within the tertiary sector show quite varied ethnic

Page 6: Labor Markets Ethnic in Peninsular Malaysia, 1970

Charles Hirschman

TABLEINDUSTRIAL COMPOSITION OF MEN, AGE 25-64, IN THE EXPERIENCED LABOR FORCE, BY ETHNIC

COMMUNITY, PENINSULAR MALAYSIA, 1970

AgricultureRice or mainly riceRubberFishingOther agriculture

MiningManufacturingFood, beverage, tobaccoTextiles, wearing apparelWood, wood productsPaper, rubber, minerals, metalsMachinery, transport equipmentMiscellaneous

ConstructionCommerce

Retail tradeWholesale trade

Transportation, communications,utilities

ServicesGovernment administrationPolice and armed servicesEducationMedical and communityPersonal and recreationOther services

Total (percentage)’1(N)

TotalPopulation"

46.314.021.43.07.93.48.41.50.72.12.21.40.52.912.39.92.4

7.219.44.33.34.21.64.02.0

100.0(24,443)

Malay

62.125.324.83.88.21.74.10.90.41.21.20.30.11.55.34.21.0

6.219.14.94.84.61.52.01.2

100.0(12,761)

Chinese

26.01.3

13.92.88.06.2

16.02.41.34.23.83.21.05.7

22.018.23.7

7.516.61.41.13.91.25.93.2

100.0(8,560)

Indian

37.30.3

30.10.16.82.55.11.60.60.41.70.50.21.1

15.310.84.5

11.227.58.42.73.73.17.71.9

100.0(2,880)

SOURCE: 1970 Population Census of Peninsular Malaysia, 0.02 Sample.The total population includes the small fraction ofpersons who are not Malays, Chinese, or Indians.The percentage distribution excludes those who did not report an occupation or industry.

distributions. Indians are somewhat overrepresented in transportation, com-munications, and utilities, a legacy of earlier times when the colonial admin-istration recruited laborers (and clerks) for the railroads and public works fromIndia. A higher proportion of Malays are employed in the armed forces andpolice, but Chinese and particularly Indians are well positioned in other serviceactivities.The occupational data in table 2 are not completely independent of the in-

dustry divisions. In fact, several occupations such as sales and agriculture closelyparallel comparable sectors in the industry classification. Occupations are meantto denote the actual work or skill distinctions of workers, as opposed to thenature ofthe enterprise or the product produced, which is tapped in the industrymeasure. Malays lag behind Chinese and Indians in white-collar roles (profes-sional, administrative, and clerical workers), but the gap is not as great as inother occupations. A close look at the detailed occupational categories showsthat most Malay professionals are teachers. While teachers are a large categoryof the Indian and Chinese professional workers, non-Malays have higher pro-portions in the higher-status occupations of doctors, lawyers, engineers, and so

Page 7: Labor Markets Ethnic in Peninsular Malaysia, 1970

Labor Markets and Ethnic Inequality in Peninsular Malaysia, 1970

TABLE 2

OCCUPATIONAL COMPOSITION OF MEN, AGE 25-64, IN THE EXPERIENCED LABOR FORCE, BY

ETHNIC COMMUNITY, PENINSULAR MALAYSIA, 1970

Professional and technicalDoctors, engineers, etc.TeachersOther professionals

Administrative and managerialClericalSalesManagers and proprietorsSalesmen and shop assistantsStreet vendors

ServiceProtective serviceOther service

Craft and productionMinersTransport operativesPaid employeesSelf-employed and family workers

LaborerAgricultural

Forestry and fishingManagers and supervisorsPaid employeesSelf-employedFamily workers

Total (percentage)’(N)

TotalPopulation*

5.70.91.13.71.74.510.65.63.11.98.24.24.0

20.02.05.19.63.34.5

44.83.80.9

12.824.62.7

100.0(24,443)

Malay

5.00.53.70.81.12.94.32.90.80.68.56.12.4

12.40.94.45.31.94.5

61.34.60.4

13.039.63.7

100.0(12,761)

Chinese

6.61.22.43.82.45.9

19.29.46.33.56.31.25.1

31.94.06.015.96.13.6

24.23.80.98.39.13.0

100.0(8,560)

Indian

5.30.92.22.31.37.4

13.46.53.83.1

11.63.68.0

19.81.65.9

10.71.67.6

33.60.13.2

26.13.60.5

100.0(2,880)

SOURCE and NOTES: Same as table 1.

on. Only in protective-service occupations (police, armed forces, guards) andagriculture do the Malay proportions exceed the Chinese and Indian figures.Indian and Chinese men have more diversified occupational structures than dothe Malay men. Over the two decades prior to 1970, there have been only modestreductions of the ethnic differences in economic roles. The Malay occupationaland industrial structures have become more "modem" in the sense ofreductionsin agriculture and increases in other sectors, including white-collar occupations.These same processes, however, have occurred at a faster pace among the Chineseand Indian populations.39

Rather than explain all the ethnic differences in occupational and industrialstructures, I have chosen to focus on three selected positions that are most visibleand central to current policies aiming at a restructuring of society: the sectorsof manufacturing and commerce, and white-collar occupations. Manufacturingis seen as the modem sector that expands during the process of economic de-velopment. Commerce represents the most obvious sign of ethnic inequalityand the source of considerable interethnic antagonism. Malay participation inwhite-collar occupations (professional, administrative, and clerical occupa-tions)-the most prestigious and well rewarded in society-lags behind Chineseand Indian levels. Elimination of these disparities will be necessary for the goalofan integrated society.

Page 8: Labor Markets Ethnic in Peninsular Malaysia, 1970

8 Charies Hirschman

Models of Ethnic Stratification

In the multivariate analysis ofethnic inequality, I regress the dependent vari-ables (three dichotomous dependent variables: proportion of workers in white-collar occupations, proportion of workers in commerce, proportion of workersin manufacturing) on ethnicity and other individual and labor-market variables.The statistical technique used is analysis of covariance, where most of the in-dependent variables are categorical, but several labor-market characteristics (citysize and proportions in various industries) are coded as linear variables. Theuse of categorical or dichotomous dependent variables in multivariate analysisis a subject about which "we now understand more about the difficulties thanwe do about their solution.’140 While some authorities claim that ordinary-least-squares techniques are not appropriate for dichotomous dependent variables,others conclude that the usual regression approach leads to essentially the sameresults.41 For the present analysis, I have experimented with alternative statisticaltechniques, and the results do not differ from those using ordinary least squares.For ease of interpretation and because ethnic gaps in the form of percentage-point differences (based upon using probabilities as the dependent variables)seem to be closest to the conceptual framework that guides the research, I presentonly the results from the least-squares analysis.The individual-level variables used in this analysis are: age (25-34, 35-44,

45-64), birthplace (foreign, east-coast or northern states, west-coast states), ed-ucation (none, primary schooling, lower secondary, Lower Certificate of Edu-cation or above), and duration ofresidence (nonmigrant, migrant with less thanyear of local residence, migrant with l-to-4 years of local residence, migrant

with 5-to-lO years oflocal residence, migrant with 11 or more years ofresidence).The distribution of the population (men, age 25-64 in the experienced laborforce) by these variables is shown in the last (right-hand side) column of table3. These variables generally tap human-capital and other individual resourcesthat are important for socioeconomic attainment. It is important to note, how-ever, the specific substantive meaning of these variables in the Malaysian con-text.The birthplace variable divides the native-born by region oforigin. The west-

coast states (Penang, Perak, Selangor, Negri Sembilan, Malacca, and Johore) aregenerally considered to be the most developed with better educational facilities,while the east-coast states (Kelantan, Trengganu, and Pahang) and the northeast(Kedah and Perils) are considered the least developed. The educational variablesorts on the basic cutting points of formal schooling. The primary schoolingcategory includes all those with l-to-6 years of schooling. Lower secondaryschooling is equivalent to grades 7, 8, and 9. Upon completion oflower-secondaryschooling (after 9 years of schooling) all Malaysian students take a nationalexam that determines continuation to the next educational level. Those whopass the exam receive the Lower Certificate of Education (LCE)-a very im-portant credential for employment in the formal sector. Nonmigrants are personswho still live in the locality where they were bom. Migrants are subdivided bylength of residence in their current locality (a city, town, or village).Labor-market characteristics are represented by three groups ofvariables. The

first is an urban-rural classification based upon a definition of an urban placeas a gazetted area with a population of 5,000 or more. The next variables arecity size (coded to the nearest 1,000) for those living in urban areas (defined bya population of 5,000 or more), and the size of the largest town in the district

Page 9: Labor Markets Ethnic in Peninsular Malaysia, 1970

Labor Markets and Ethnic Inequality in Peninsular Malaysia, 1970 9

for individuals living in rural areas. Conceptually, the population size ofa townmay tap the volume and range of economic opportunities. For those living inan urban area, the town of residence probably represents the appropriate labormarket. For those residing in rural areas, the appropriate labor market is cer-tainly greater than the local village, but the boundaries may depend on availabletransportation as well as distance. With the idea that rural residents may beexposed to more opportunities ifthey live in close proximity to a large city, thepopulation size of the largest town in the district (coded to the nearest 1,000)is assigned as a labor-market characteristic for all rural residents. For the city-size variable, all rural residents have a zero score, and all urban residents havea zero score on the largest-town-in-district variable (essentially these are inter-action terms).The final set of labor-market characteristics are indices of the industrial com-

position of the usual work force of the town of residence (for urban residentsonly) and ofthe district (for rural residents only). Seven indicators are includedin each set and coded to the nearest percentage of the usual work force in padi(rice cultivation), rubber, fishing, manufacturing, commerce, transport and com-munication, and services. These variables are based on the industry of thosewho worked "most of the time last year."The independent variables are entered in a sequence of five equations for

each dependent variable. The first equation only includes ethnicity as an in-dependent variable. Malay is the omitted category, and the Chinese and Indianvalues represent the difference between the Malay proportion (in the dependent-variable category) and the other ethnic communities. The second equation addsin the other four individual-level independent variables (the first category ineach set is omitted). The labor-market characteristics are then added sequentiallyin the next three equations. Changes in the ethnic coefficients show the con-sequences of holding constant the effects of the other variables in the equation.Statistical significance is measured by a metric regression coefficient that is twicethe size of its standard error (approximately the 0.01 level).

White-Collar Occupations. The five equations predicting attainment in white-collar occupations (proportion of work force in professional, administrative,and clerical occupations) are presented in the first five columns of table 3. Thefirst reduced-form equation in column shows the "gross" (or observed) ethnicdifferential in the proportion in white-collar occupations. Chinese men are about6 percentage points and Indian men are 5 percentage points more likely thanMalay men to be in a white-collar position (the small "others" ethnic categorywas also included in the regression equation, but the coefficient is not presentedhere). This ethnic differential is the same as that available from a comparisonof percentage distributions in table 1. While the ethnic gap is a fairly modestone, it is very visible and of great political importance. In the second columnor equation, when all individual-level background variables are included, theethnic gap is reduced to statistical insignificance. It is evident that the differentialethnic achievement in white-collar occupations is largely due to unequal edu-cational resources. The introduction ofthe 3 sets oflabor-market variables showsthat the net (direct) ethnic effect is a slight Malay advantage (about 2 percentagepoints) in white-collar attainment. In other words, distribution ofthe population(especially rural-urban distribution) presents a modest obstacle for Malay par-ticipation in white-collar occupations, net of all other variables, but this is bal-anced out by a net Malay advantage.

Page 10: Labor Markets Ethnic in Peninsular Malaysia, 1970

10 Charles Hirschman

TABLE 3

INDIVIDUAL AND STRUCTURAL EFFECTS ON THE PROPORTIONS W WHITE-COLLAR OCCUPATIONSIN THE COMMERCIAL SECTOR AND IN THE MANUFACTURING SECTOR

AMONG MEN, AGE 25-64, IN THE EXPERIENCED LABOR FORCE, PENINSULAR MALAYSIA, 1970

PROPORTION WHITE COLLAR

2 3 4 5

INDIVIDUAL CHARACTERISTICSEthnicityMalayChineseIndian

Age25-3435-4445-64

BirthplaceForeignEast/NorthWest

EducationNonePrimaryLower secondaryLCE or above

Duration of residenceNonmigrantMigrant, < yr.Migrant, 1-4 yrs.Migrant, 5-10 yrs.Migrant, 11+ yrs.

STRUCTURAL CHARACTERISTICS

TownCity size X town (000)District town X rural (000)Town industry X town (percentages)PadiRubberFishingManufacturingCommerceTransportation & commerceServices

District Industry X Rural (percentages)PadiRubberFishingManufacturingCommerceTransportation & commerceServices

Constant

R1 (adjusted)

5.8* -0.1 -1.9* -2.2* -2.0*5.0* -1.1 -2.2* -2.3* -2.3*

1.8* 1.6* 1.8* 1.5*2.9* 3.0* 3.0* 2.9*

0.5 1.2 1.5* 1.2*1.4* 1.8* 1.8* 1.9*

4.4* 3.9* 3.8* 3.8*21.4* 19.6* 19.4* 19.2*66.5* 64.1* 64.0* 63.6*

-0.3 0.1 -0.1 -0.21.3* 1.6* 1.3* 1.2*1.3* 1.6* 1.4* 1.4*0.6 0.5 0.5 0.5

6.1* 6.9* 6.7*0.00 -0.01*0.01* 0.01*

0.0-0.1*-0.1-0.1*0.00.00.1*

0.00.00.0

-0.20.10.20.0

9.0 -1.8 -3.0 -3.4 -4.2

1.0 37.7 38.4 38.4 38.5

SOURCE 1970 Population Census of Peninsular Malaysia, 0.02 Sample Tape.Symbol indicates omitted category of each set ofdummy variables.Symbol indicates statistically significant.

Page 11: Labor Markets Ethnic in Peninsular Malaysia, 1970

Labor Markets and Ethnic Inequality in Peninsular Malaysia, 1970 11

TABLE 3 (Continued)

PROPORTION IN COMMERCE PROPORTION IN MANUFACTURING POPULATION

2 3 4 5 ^ 3 4 5COMPOSITION

16.7* 13.8* 10.8* 10.6* 10.6* 11.9* 11.3* 9.2* 8.7* 8.7*10.0* 5.4* 3.7* 3.4* 3.4* 0.9 0.7 -0.5 -1.0 -0.9

0.5 0.2 0.2 0.1 -3.2* -3.5* -3.4* -3.4*0.7 0.7 0.7 0.5 -3.2* -3.1* -3.1* -3.0*

-9.0* -7.8* -7.3* -8.5* -1.4* -0.6 0.3 -03-11.2* -10.5* -10.5* -10.1* -0.6 -0.1 -0.1 -0.3

5.6* 4.8* 4.7* 4.7* 1.4* 0.8 06 0712.6* 9.6* 9.2* 9.0* 3.6* 1.4 0.7 0.78.2* 4.3* 4.0* 3.7* -1.2* -4.0* -4.6* -4.8*

-5.2* -4.5* -4.6* -4.2* -1.0 -0.5 -1.2 -1.1-4.2* -3.6* -3.8* -3.4* 2.8* 3.1* 2.4* 2.3*-2.5* -2.1* -2.2* -1.8* 0.5 0.9 0.4 0.5-0.5 -0.8 -0.8 -0.5 -0.7 -0.9 -1.0 -1.0

523512

392833

172954

23625II

456121126

10.0* 8.9 12.4* 7.2* 8.3* 6.50.01* 0.00 0.01* 0.000.01* 0.02* 0.03* 0.03*

35

-0.1*-0.2*-0.2*-0.2*0.4*

-0.1-0.1

0.0-0.1-0.10.4*0.1*0.2*0.0

0.1* 0.1*0.0 0.1*O.I* 0.3*

-0.3* 0.4*0.2*-0.2*0.5* 0.8

-0.1 -0.1

5.3 11.4 9.5 9.1 4.9 1.2 6.0 4.6 3.3 -2.5

5.5 7.5 9.1 9.3 10.0 4.0 4.6 5.8 6.5 7.4

Page 12: Labor Markets Ethnic in Peninsular Malaysia, 1970

12 Charles Hirschman

There are a number of other important predictors of participation in a white-collar occupation. Older workers are somewhat more likely to be in a white-collar job as are the native-born, especially those bom in the west-coast states(historically more developed). The single most important variable, however, iseducational attainment. Primary schooling has been a modest influence, andlower-secondary schooling has a substantial effect, but it is those with the LCEcredential (11 percent of the male work force) who are most likely to work ina white-collar job-more than 40 percentage points above those with lower-secondary schooling and more than 60 percentage points above those with noschooling. The influence of education is not mediated or even substantiallyaltered with all other variables included.

Migration status has a very modest effect. Those who have migrated but havelived from l-to-10 years in a place are slightly more likely to be in a white-collarjob. Net ofall individual variables, living in an urban area has a significanteffect, raising the probability ofbeing in a white-collar position by 6 percentagepoints. For urban residents, size of the city has no effect, but for rural residents,the size of the largest town in the district is significant, though fairly modest.For every addition of 1,000 population in the largest district town, the proba-bility of being in a white-collar job was increased by 0.01 percentage point. Inother words, living near a large town of 120,000 rather than one of20,000 meansa 1-percentage-point increase in the probability ofa rural resident being a white-collar worker.Towns with higher proportions of workers in the rubber or manufacturing

sectors have a modest depressing effect on the probability of white-collar em-ployment, while the proportion in services has a positive effect for urban resi-dents. Since the industry-composition variables are scored in percentage terms,a percentage-point change in the industry composition means a 0.1-percentage-point effect on the probability of employment of a specific type. Interestingly,controlling for the industry composition, city size has a modest negative effecton the probability of white-collar employment.The Commerce Sector. There is a much wider ethnic gap in the proportion

of workers in commerce. Indian men are 10 percentage points more likely thanMalay men to be in commerce (includes both retail and wholesale trade), andChinese men have a 17-percentage-point edge. This differential is about thesame in 1970 as it was in 1947.42 The latent tension between the ubiquitousChinese shopkeeper and the Malay peasant is probably the greatest source ofethnic antagonism in the country.

Controlling for the individual-level variables reduces the Indian and Chinesecoefficients by almost one-halfand one-fifth, respectively, oftheir original levels.Even with these variables held constant, ethnic differentials remain wide: Chinesemen are 14 percentage points more likely than Malay men to be in commerce.The inclusion of the labor-market variables reduces the net Chinese-Malay dif-ference by another 3 percentage points (more than a one-fifth reduction). Mostof this is due to the basic urban-rural distribution. The city-size and industry-composition characteristics do not appreciably reduce the ethnic gap further.The effects of several individual and labor-market variables are significant.

Foreign birth is strongly associated with an entrepreneurial vocation. The easeof entry into petty retailing may be the reason that foreign-bom Malaysians aredrawn to the commercial sector. Having more education facilitates entry intocommerce, but only up to a point. The strongest net effect is for those with

Page 13: Labor Markets Ethnic in Peninsular Malaysia, 1970

Labor Markets and Ethnic Inequality in Peninsular Malaysia, 1970 13

lower-secondary schooling. Those with an LCE educational certificate may findother sectors of the economy more attractive. Nonmigrants who have lived alltheir lives in the same place are most likely to be in commerce, but migrantswith 10 years of local residence have almost the same level of employment incommerce. On the surface, this finding seems to conflict with the positive as-sociation for the foreign bom (who are migrants, by definition). I think, however,that these characteristics reveal different aspects of the causal process. Thosebom abroad may be more likely to find entrepreneurial jobs attractive (andperhaps have the necessary motivation for very long hours with low remuner-ation), but shopkeeping probably requires a considerable period of local resi-dence to develop personal ties and access to opportunities.Labor-market variables also have significant net effects. Residence in an urban

area adds 10 percentage points to the probability of commercial-sector em-ployment, net of individual variables. Of course, a higher population densityexpands the local market for trading. City size and the size of the largest townin the district have positive effects for urban and rural residents, respectively.Attributes of the economic structure of towns and districts, as indexed by in-dustrial composition, also have significant effects on the probability of employ-ment in commerce. Most obvious is that towns and districts with larger distri-bution in the commerce sector raise the probability ofindividual workers beingin this sector. This effect is partially an artifact of the variables, although thetwo census items are independent.43 A higher proportion in a primary activity(padi, robber, or fishing) or manufacturing lowers the probability ofemploymentin commerce. For rural residents, however, the effect of the aggregate share inpadi or fishing increases the probability ofbeing in commerce, as does the shareof employment in the transport and communications sector. I will offer somespeculation on the patterns later in the paper.The Manufacturing Sector. The ethnic gap in the manufacturing sector is

somewhat different. The levels of participation are equivalent for Malay andIndian men, but Chinese have an almost 12-percentage-point advantage. Theintroduction of individual-level variables has almost no effect upon the mag-nitude of the Chinese coefficient (reduction from 11.9 to 11.3). The first labor-market variable, urban-rural residence, reduces the Chinese-Malay gap by 2percentage points-by about 20 percent. Introduction of the next two variables,city size and size of largest district town, further narrows the gap by half apercentage point, but the industrial composition has no measurable impact onthe net ethnic differential.

In spite ofthe modest attentuation ofthe ethnic gap, the individual and labor-market variables do have significant effects on participation in the manufac-turing sector. Younger workers, below age 35, are most likely to be in the man-ufacturing industry, as are those with average levels of education (which isreduced to insignificance with the labor-market variables in the equation) andmigrants with l-to-4 years of local residence.Most interesting are the labor-market effects. An urban area, particularly a

large one, has a modest but significant positive effect upon the probability ofresidents being employed in manufacturing. The effects of urban location andcity size are partially mediated by the industrial composition, which includesthe shares in manufacturing, commerce, and transportation/communication.For rural residents, the size ofthe largest town in a district is strongly associatedwith a high probability ofemployment in manufacturing,,as are several features

Page 14: Labor Markets Ethnic in Peninsular Malaysia, 1970

14 Charles Hirechman

TABLE 4

INDIVIDUAL AND STRUCTURAL EFFECTS ON THE PROPORTION IN WHITE-COLLAR OCCUPATIONSIN THE COMMERCIAL SECTOR AND IN THE MANUFACTURING SECTOR

AMONG EMPLOYED MEN, AGE 25-64, BY ETHNIC COMMUNITY, PENINSULAR MALAYSIA, 1970

INDIVIDUAL CHARACTERISTICS

Age25-3435-4445-64

BirthplaceForeignEast/NorthWest

EducationNonePrimaryLower secondaryLCE or above

Duration of residenceNonmigrantMigrant, < yr.Migrant, 1-4 yrs.Migrant, 5-10 yrs.Migrant, 11 + yrs.

STRUCTURAL CHARACTERISTICS

TownCity size X town (000)District town X rural (000)Town industry X town (percentages)PadiRubberFishingManufacturingCommerceTransportation & commerceServices

District industry X rural (percentages)PadiRubberFishingManufacturingCommerceTransportation & commerceServices

Constant

R2 (adjusted)

SOURCE and SYMBOLS: See table 3.

PROPORT

Malay

0.81.7*

2.01.7

4.3*20.8*66.2*

-1.0-0.6

0.5-0.2

8.0*-0.010.03*

0.00.0

-0.2*-0.2*

0.00.10.1

0.00.00.0

-0.20.00.20.0

-3.6

37.2

[ON WHITE

Chinese

2.3*4.8*

-1.12.2*

3.5*20.3*61.4*

-0.93.3*3.1*1.5

4.0-0.01*0.00

0.1-0.1

0.10.00.0

-0.20.2*

0.10.00.0

-0.10.20.4

-0.1

-7.9

37.0

COLLAR

Indian

2.43.4*

4.9*4.5*

1.411.2*63.3*

1.01.8

-1.8-1.1

11.5-0.010.02

-0.1-0.3-1.0*-0.5*

0.00.20.1

U.U0.0

-0.3-0.2-0.2-0.1

0.1

-1.3

45.5

Page 15: Labor Markets Ethnic in Peninsular Malaysia, 1970

Labor Markets and Ethnic Inequality in Peninsular Malaysia, 1970 15

TABLE 4 (Continued)

PROPOR

Malay

-0.21.2*

-5.4*-7.7*

3.0*4.5*1.5

-1.4-0.5-0.90.3

7.8*0.000.00

-0.1*-0.1*-0.3*0.00.10.1

-0.1*

0.00.00.0

-0.10.10.10.0

9.5

2.1

TIONIN CC

Chinese

1.31.5

-4.7*-7.2*

6.6*13.6*6.5*

-7.4*-8.6*-3.3*-2.1

19.1*-0.01*0.04*

-0.2-0.2*0.0

-0.2*0.4*0.10.1

0.4*0.2*0.3

-0.40.31.3*

-0.2

0.1

5.2

)MMERCE

Indian

-3.1*-5.4

-23.1*-20.7*

8.7*9.6*6.3*

-2.30.60.41.7

8.70.010.02

1.4*0.01.0*

-0.20.8*0.0

-0.2

0.10.00.5

-0.10.30.1

-0.1

9.0

16.3

PINM

Malay

-2.9*-3.1*

-0.1-0.5

-0.10.6

-3.3*

-0.70.20.1

-0.3

3.80.000.03*

0.0-0.1*-0.10.4*0.00.00.0

0.00.00.1*0.3*

-0.2*0.9*

-0.2*

4.8

3.4

’ROPORTIOANUFACTU

Chinese

-5.1*-4.3*

-0.4-0.3

2.9*1.8

-6.2*

-1.06.0*1.7

-1.2

12.9*0.000.03*

0.20.1

-0.10.5*0.2*0.5*0.2*

0.3*0.2*0.6*0.9*

-0.20.60.1

-0.6

4.7

N

TUNG

Indian

-1.2-0.7

-1.80.3

0.22.2

-2.2

0.74.3*0.5

-0.6

6.70.000.02*

-0.3-0.1-0.10.40.10.10.0

0.10.10.30.4

-0.10.9

-0.2

-3.5

1.3

POPULA

Malay

382934

44749

286237

546131116

18

TION COMT

Chinese

432730

251065

15647

15

415

11834

56

OSITION

Indian

323137

475

47

20587

15

237131343

45

Page 16: Labor Markets Ethnic in Peninsular Malaysia, 1970

16 Charles Hirschman

of the districts’ sectoral distribution: padi, rubber, fishing, manufacturing, andtransportation/communications. Without giving a detailed interpretation ofthesespecific coefficients, it is clear that the right mix of available employment op-portunities increases the level of participation in manufacturing, net of othervariables.

Effects of Individual and Labor-Market Characteristics and OccupationalAttainment within Each Ethnic Community

Table 4 shows the complete (or final) equations, separately for the three majorethnic groups, for each of the dependent variables. Comparing the same model(and coefficients) across ethnic communities, it is possible to address the ques-tion: is the process of attainment the same or different (technically, are thereinteractions between ethnicity, social background, and labor-market outcomes)?

In general, the basic predictors of white-collar occupational attainment arefairly similar across ethnic lines, with higher education being the key determi-nant. Native birth appears to be a bit more important for Indian men andmigration status for Chinese men, but the differences are fairly small. In termsof labor-market variables, urban residence is a positive factor for all, thoughthe coefficient is larger for Malay and Indian men. City size has a modest negativeeffect for all three populations, but the size of the largest district town is im-portant for rural Indian and especially Malay men. Larger shares in the urban-manufacturing and fishing sectors have negative effects for Malay and Indianmen. A larger services sector is positively associated with a higher probabilityof white-collar employment, but the effect is largest for Chinese.

Predicting the probability of employment in commerce shows some differentpatterns of determinants across ethnic communities. While age has a slightpositive effect for Malay and Chinese men, the reverse is true for Indian men.Foreign birth is an important determinant of participation in commerce for allgroups, but much more significant for Indian men. The general U-shaped patternof educational coefficients is shared by all groups, but with a greater peak forChinese men, and the negative effect of short- and medium-term duration ofresidence of migration is really only evident for Chinese men. Most interestingare the patterns of labor-market variables. Urban location is twice as importantfor Chinese men as for Malay and Indian men. The size of the largest town inthe district is much more important for rural Chinese men. A concentration inthe primary sector generally has a discouraging effect on commercial employ-ment for urban residents (but not for Indian men). The expected positive effectofthe share ofthe urban commerce sector is not present for Malay men. Amongrural residents, a large share of district employment in the primary sector gen-erates entrepreneurial activity for Indian and especially Chinese men, but notfor Malays. The presence of a large district proportion in transport/communi-cations has a very strong effect for Chinese men.Why are all these patterns so strong for Chinese men, but not for Malay men?

Are Chinese men more able to take advantage of trading opportunities or dothey already have an entrenched position from earlier times? This question isone that lies at the heart of the theoretical and policy disputes over ethnicinequality in Malaysia.A similar pattern of differential labor-market effects is evident in the com-

parison of the three ethnic equations predicting the proportion in the manu-facturing sector. Residence in an urban area is a more consequential variable

Page 17: Labor Markets Ethnic in Peninsular Malaysia, 1970

Labor Markets and Ethnic Inequality in Peninsular Malaysia, 1970 17

for Chinese men in the attainment ofmanufacturing-sector employment. More-over, the attributes ofa place, as indexed by the industrial composition variables,are more strongly associated with manufacturing employment for Chinese men.For instance, among urban residents the size of the manufacturing sector en-hances the probability of employment in manufacturing for all ethnic groups,but the probability of Chinese manufacturing employment also responds tolarger shares in commerce, transportation/communications, and services. Amongrural residents, there are some positive effects in the Malay coefficients, but theeffects are generally much more significant for Chinese men.

Discussion and Conclusions

This analysis has not completely explained the sources of ethnic inequalityin Peninsular Malaysia. It has, however, revealed that access to opportunitystructures, as measured by the characteristics of labor markets, are importantvariables to consider in both scientific and policy analysis, indeed as importantin some instances as individual human-capital variables that are at the core ofmost interpretations.

In this study, I focused on differential ethnic participation in two key sectorsand an occupational category that represent some of the salient divisions of thesegmented plural society of Peninsular Malaysia. The explanations of the dif-ferentials in these areas do not necessarily exhaust the range of significant di-visions of ethnic inequality, but these three indices do show the varied com-bination of factors that account (or partially account) for important aspects ofethnic inequality.The explanation of unequal proportions in white-collar occupations is the

most straightforward. Essentially the modest ethnic differentials in high-statusoccupations can be explained by variations in individual characteristics, par-ticularly educational attainment. Controlling for the set of individual variables,ethnic differences are essentially reduced to zero. While this may not be true inevery specific white-collar occupation, the gap is likely to be very small becauseof the few numbers involved in these occupational categories. After controllingfor the individual variables, the urban-rural distinction shows that Malays re-main at a slight disadvantage in the attainment of white-collar occupationsbecause of residential patterns. This, however, is cancelled out by a modest net(direct) positive effect of being Malay on the probability of being a white-collarworker. Were it possible to separate private- vs. public-sector employment, thesource of the preferential hiring might be revealed (the hypothesis being thatthere may be Malay preference in government hiring, but Chinese preferencein the private sector).More complex are the wide variations in ethnic proportion in commerce and

manufacturing. In neither sector do individual or labor-market variables explainmore than a fraction of the substantial ethnic differences. This does not meanthat individual or labor-market variables are unimportant, only that the redis-tribution of the additive effects of these variables would not eradicate ethnicinequality in these sectors. Note, however, that differential patterns of rural-urban residence is a very important determinant of employment in manufac-turing and commerce. It is also an important partial explanation of the Malay-Chinese differences in employment in these sectors. While the size ofthe labor-market and the industrial-composition variables are important in their ownright, they do not add to the narrowing of the ethnic differentials.

Page 18: Labor Markets Ethnic in Peninsular Malaysia, 1970

18 Chartes Hirschman

Some insight into the reasons for the net Malay-Chinese gaps in the manu-facturing and commerce sectors is provided in table 4. Chinese (and to a lesserextent Indian) opportunities for employment in manufacturing and commercewere much greater in urban areas than was the case for urban Malay men (netof all other variables). In addition, the structure of employment of places (dis-tricts and towns) was more strongly related to Chinese than Malay outcomes.

Why were Chinese men more able to take advantage of the emergent oppor-tunities? While it is impossible to address this question with the data in hand,it may be that Chinese are better positioned to take advantage of the diversestructure ofemployment in towns. For instance, family ties, including the own-ership and management of many small shops and factories, may present op-portunities for Chinese employment that are not available to Malays, even ifthey live in the immediate vicinity. This is a hypothesis for future inquiry.

NOTES

’The classical assimilation theory is represented in Robert Park, Race and Culture (Glencoe, IL:Free Press, 1950); and Talcott Parsons, "Some Theoretical Considerations on the Nature and Trendsof Change of Ethnicity," in Ethnicity: Theory and Experience, ed. Nathan Glazer and Daniel P.Moynihan (Cambridge: Harvard University Press, 1975), pp. 53-83. For a general critique of theindustrialization perspective, see Herbert Blumer, "Industrialization and Race Relations," in In-dustrialization and Race Relations, ed. Guy Hunter (London: Oxford University Press), pp. 220-53. An overview ofcontemporary ethnic conflict is given in Martin 0. Heisler, ed., "Ethnic Conflictin the World Today," The Annals 443 (September 1977): 1-160.

’Major theoretical works include: Michael Hechter, Internal Colonialism: The Celtic Fringe andBritish National Development (Berkeley: University of California Press, 1975); Edna Bonacich, "ATheory of Ethnic Antagonism: The Split Labor Market," American Sociological Review 37 (October1972): 547-49; Ernest A. T. Barth and Donald L. Noel, "Conceptual Frameworks for the Analysisof Race Relations: An Evaluation," Social Forces 50 (March 1972): 333-48; Stanley Lieberson, "ASocietal Theory ofRace and Ethnic Relations," American Sociological Review 26 (December 1961):902-10; Richard Schermerhorn, Comparative Ethnic Relations: A Framework for Theory and Re-search (Chicago: University of Chicago Press, 1978); and Pierre L. van den Berghe, The EthnicPhenomenon (New York: Elsevier, 1981). For a review of this literature see Charles Hirschman,"Theories and Models of Ethnic Inequality," in Research in Race and Ethnic Relations, vol. 2, ed.Cora B. Marred and Cheryl Leggon (Greenwich, CT: JAI Press, 1980), pp. 1-20.

’The term Peninsular Malaysia refers to the 11 states ofMalaysia on the mainland ofAsia, excludingthe states ofSabah and Sarawak on the island of Borneo. Prior to the formation of Malaysia in 1963,Peninsular Malaysia (or West Malaysia) was known as Malaya, an independent country from 1957.During the colonial era, Malaya was the general term for the area under British colonial rule. Singaporewas often considered part ofMalaya, though it was administered separately in the post-World War IIera (Singapore was briefly part of Malaysia from 1963 to 1965). For the present paper, refer toPeninsular Malaysia or Malaya, depending on the historical period, but with the same geographicalreference area (excluding Singapore).

’Alastair Lamb, "Early History," in Malaysia: A Survey, ed. Wang Gungwu (New York: Praeger,1965), pp. 99-112.

’B. W. Hodder, Man in Malaya (1959; reprint ed., London: University of London Press, 1968).

’Kennedy G. Tregonning, The British in Malaya: The First Forty Years 1786-1826 (Tucson: Uni-versity of Arizona Press, 1965).

The development of the export economy and the role of immigrant labor are presented in R. N.Jackson, Immigrant Labour and the Development of Malaya; 1786-1920 (Kuala Lumpur: Govern-ment Printer, 1962); and Khoo Kay-Kim, The Western Malay Stales 1850-1873 (Kuala Lumpur:Oxford University Press, 1972).

G. C. Alien and Audrey G. Donnithome, Western Enterprise in Indonesia and Malaya (London:Alien and Unwin, 1954).

’Kemial Singh Sandhu, Indians in Malaya: Some Aspects of Their Immigration and Settlement1786-1957 (Cambridge: Cambridge University Press, 1969); Michael Stenson, Class, Race, and Co-

Page 19: Labor Markets Ethnic in Peninsular Malaysia, 1970

Labor Markets and Ethnic Inequality in Peninsular Malaysia, 1970 19

lonialism in West Malaysia: The Indian Case (Vancouver University of British Columbia Press,1980).

’"Jackson, Immigrant Labour and the Development of Malaya; W. L. BIythe, "Historical Sketch ofChinese Labour in Malaya," Journal of Malaya Branch ofthe Royal Asiatic Society 20 (Part I, 1947):64-125; J. Norman Fanner, Colonial Labor Policy and Administration: A History of Labor in theRubber Industry of Malaya, 1910-1941, Monograph ofthe Association for Asian Studies (New York:Association for Asian Studies, 1960).

"John Drabble, Rubber in Malaya, 1876-1922: The Genesis ofthe Industry (Kuala Lumpur OxfordUniversity Press, 1973).

"Swee-Hock Saw, "Trends and Differentials in International Migration in Malaya," Ekonomi(KualaLumpur) 4 (December 1963): 87-113.

"Charles Hirschman, Ethnic and Social Stratification in Peninsular Malaysia, Arnold M. and Car-oline Rose Monograph Series (Washington, DC: American Sociological Association, 1975), p. 9.

’Charles Hirschman, "Demographic Trends in Peninsular Malaysia, 1947-1975," Population andDevelopment Review 6 (March 1980): 111.

"Heng-Kow Lim, The Evolution ofthe Urban System in Malaya (Kuala Lumpur Penerbit UniversitiMalaya, 1978); M. S. Sidhu, "Chinese Dominance ofWest Malaysian Towns, 1921-1970" Geography61 (January 1976): 17-23.

"Charles Hirschman, "Sociology," Malaysia Studies: Present Knowledge and Research Trends, ed.John Lent (DeKalb: Center for Southeast Asian Studies, Northern Illinois University, 1979), pp. 1-58.

"Charles Hirschman, "Political Independence and Educational Opportunity in Peninsular Ma-laysia," Sociology of Education 52 (April 1979): 67-83.

"Charles Hirschman, "Occupational and Industrial Change in Peninsular Malaysia, 1947-1970,"Journal of Southeast Asian Studies 13 (March 1981): 9-32.

"Ungku A. Aziz, "Political and Economic Books on Malaya: A Review Article," Malayan EconomicReview 5 (April I960): 25-30.

"K. J. Ratnam, Communalism and the Political Process in Malaya (Kuala Lumpur University ofMalaya Press, 1965).

"Mid-Term Review ofthe Second Malaysia Plan, 1971-1975 (Kuala Lumpur Malaysia GovernmentPrinter, 1973).

"Peter Blau and Otis Dudley Duncan, The American Occupational Structure (New York: Wiley,1967); David Featherman and Robert Hauser, Opportunity and Change (New York: Academic Press,1978).

"Hirschman, "Theories and Models of Ethnic Inequality."

"U.S. Commission on Civil Rights, Social Indicators of Equality for Minorities and Women (Wash-ington, DC: U.S. Commission on Civil Rights, 1978).

"Otis Dudley Duncan, "Inheritance of Poverty or Inheritance of Race?" in On UnderstandingPoverty: Perspectives from the Social Sciences, ed. Daniel P. Moynihan (New York: Basic Books1969).

"Otis Dudley Duncan, "Discrimination against Negroes," The Annals 371 (May 1967): 85-103.

"See David Featherman, "The Socioeconomic Achievement of White Religio-ethnic Subgroups:Social and Psychological Explanations," American Sociological Review 36 (April 1971): 207-22, fora test of psychological factors in achievement; and Erik Olin Wright, "Race, Class, and IncomeInequality," American Journal of Sociology 83 (May 1978): 1368-937, for a test of class analysis ofblack-white inequality.

"Hirschman, Ethnic and Social Stratification."Ibid., pp. 55, 66.

"Ibid-, pp. 68-71.

"Arne L. KaUeberg and Aage B. Sorensen, "Sociology of Labor Markets," Annual Review of So-ciology 5 (1979): 351-79.

Page 20: Labor Markets Ethnic in Peninsular Malaysia, 1970

20 Charles Hirschman

"Charles W. Mueller, "Socioeconomic Achievements and City Size," The Process of Stratification:Trends and Analyses, ed. Robert M. Hauser and David L. Featherman (New York: Academic Press,1977), pp. 249-69.

"The original hypothesis was suggested by Robin Williams, The Reduction oflntergroup Tensions(New York: Social Science Research Council, 1947). For more recent studies, see Hubert Blalock,"Percent Non-White and Discrimination in the South," American Sociological Review 22 (December1957): 677-82; and Parker Frisbie and Lisa Neidert, "Inequality and the Relative Size of MinorityPopulations," American Journal of Sociology 82 (March 1977): 1007-30.

"Ross Stolzenberg and Ronald D’Amico, "City Differences and Nondifferences in the Effect ofRace and Sex on Occupational Distribution," American Sociological Review 42 (December 1977):937-50; Toby Parcel, "Race, Regional Labor Markets, Earnings," American Sociological Review 44(April 1979): 262-79.

"The data file of the 2-percent sample was produced by the Malaysian Department of Statisticsand is available to research scholars upon request to the Chief Statistician, Malaysia.

*An Interim Report on the Post-Enumeration Survey (Kuala Lumpur Malaysia, Department ofStatistics, 1974).

"The 1970 figures were obtained from R. Chander, 7970 Population and Housing Census of Ma-laysia: Community Groups (Kuala Lumpur Malaysia, Department of Statistics, 1972). The size ofeach town (and ethnic composition) for the year 1957 was available in the 1957 Population Censusof the Federation of Malaya, reports 2-12 (Kuala Lumpur: Department of Statistics, Federation ofMalaya, 1959).

"For a detailed analysis of these patterns from 1947 to 1970, see Hirschman, "Occupational andIndustrial Change."

"Ibid.

-Eric A. Hanushek and John E. Jackson, Statistical Methods for Social Scientists (New York:Academic Press, 1977), p. 180.

"Michael W. Gillespie, "Log-linear Techniques and the Regression Analysis ofDummy DependentVariables," Sociological Methods and Research 3 (August 1977): 103-22; David Knoke, "A Com-parison of Log-linear and Regression Models for Systems of Dichotomous Variables," SociologicalMethods and Research 3 (May 1975): 416-34; James Morgan et al.. Five ThousandAmerican Families:Patterns of Economic Progress, vol. (Ann Arbor Institute for Social Research, University of Mich-igan, 1970), appendix E; Leo A. Goodman, "The Relationship between Modified and Usual Multiple-Regression Approaches to the Analysis of Dichotomous Variables," in Sociological Methodology1976, ed. David R. Heise (San Francisco: Josscy-Bass Publishers, 1975), pp. 83-110.

"Hirschman, "Occupational and Industrial Change."

"The community characteristic is based upon the "usual industry" of employment, while the de-pendent variable is based upon the industry ofemployment last week (two separate census questions).