Een demografisch mozaiek, Indonesie 1880 - 1942

A.J. Gooszen

    Research output: Thesisexternal PhD, WU


    This study concerns regional diversity in demographic developments regarding the indigenous population of the Indonesian archipelago during the heyday of colonial rule, i.e. from 1880 to 1942. The central question of the study is: Have demographic developments been more or less uniform in the various parts of the archipelago?

    This concentration on demographic uniformity or diversity is inspired by recent findings concerning demographic developments in Europe in the nineteenth and twentieth century. Research into the decline of fertility in Europe, carried out by Princeton University, yielded the conclusion that regional variation in fertility decline could not be explained by socio-economic variables only. Cultural diversity and peculiarities of the various regions played their part as well. A similar conclusion was found in various studies regarding the determinants of spatial variation in mortality. This means that research into regional demographic diversity has an interest in its own right. By situating this kind of research in a colonial setting an extra dimension is added. It is not only interesting to know to what extent socio-economic or cultural phenomena have determined demographic developments, but also whether colonial or indigenous socio-economic and/or cultural determinants were decisive for demographic behaviour in Indonesia.

    One can speculate on the impact of colonial rule on demographic diversity in Indonesia. On the one hand, it seems plausible that around 1900, when colonial rule had penetrated into the very outskirts of the archipelago, a certain uniformity in living conditions of the indigenous population emerged which enhanced uniformity in demographic developments. On the other hand, colonial exploitation in agriculture and mining could have brought about an entirely new spatial pattern of economic activities and spatial inequalities in development, resulting in new patterns of demographic diversity. Along similar lines, one might conjecture whether colonial rule has diminished or enlarged cultural diversity. The introduction of western education and the rise of Indonesian nationalism as a reaction to colonial rule may have enhanced cultural integration and thereby lessened diversity. An equally valid speculation is, however, that colonial rule brought about the migration of certain ethnic groups, creating new patterns of cultural diversity. Furthermore, it seems plausible that -quite apart from colonial rule- developments in the indigenous society were decisive for demographic behaviour. Hoping to bring speculations on firmer ground, the first chapter of the study concludes with an inventory of recent literature on the demographic history of Indonesia.

    The literature on population growth in nineteenth and early twentieth century Java is reviewed. With respect to the nineteenth century, two views emerge. According to Breman, Wander, and Boomgaard, on the one hand, the growth during the first half of the nineteenth century must have been somewhat more than 1.0% per year. Breman's and - following him- Wander's guess amounts to 1.4%, while Boomgaard, who uses more reliable benchmark years, conjectures 1.25%. The same authors agree that population growth in the second half of the nineteenth century accelerated to more than 1.5 % per year. This speeding up was caused by mortality decline and, according to Wander, by a rise in fertility. The mortality decline was attributed to the success of the vaccination programme against smallpox and to improvements in the transport network. Peper, Widjojo Nitisastro, and McDonald, on the other hand, are of a different opinion. Peper has put forward that a growth of more than 1.0% per year during the first half of the century is very unlikely because the smallpox vaccination had not yet been organized efficiently, since the government had no time or money to spare, involved as it was in the Java war. Widjojo thinks that any acceleration of growth in the nineteenth century is improbable, since living conditions were very poor and health care was hardly provided for. McDonald agrees with them in general, but is slightly more optimistic, putting forward that there must have been a slow but rather steady growth. This contrasts favourably with the varying periods of decline and growth in former centuries. According to his guess, population growth amounted to 1.2 % per year on average during the entire nineteenth century.

    Turning the attention to the early twentieth century, a stronger consensus about demographic developments is seen, but the picture is somewhat vague. All authors (Breman, Wander, Widjojo, and McDonald) agree that the growth rate slowed down in the years 1900-1920, and accelerated again in the twenties and thirties. Breman conjectures that the slowing down was caused by a deterioration in living conditions. Wander is less sure about this, putting forward that there were also economic crises during the nineteenth century, without perceptibly slowing down population growth. The acceleration in growth after 1920 is attributed to improvements in health care. The growth rates mentioned by the various authors amount to 1.0 % per year for the years 1900-1920, 1.5 to 1.6 % in the twenties, and 1.5% in the thirties. A closer examination of these rates show them not to be very well founded. For one thing, the results of the so-called 'Bevolkingsopnamen' (Population Counts) of 1900 and 1905 are not very reliable. Censuses were held in 1920 and 1930. Evidently, the 1930 census was the more reliable one. It seems anybody's guess, however, how much better the response was in 1930. To complicate matters further, population figures with regard to the thirties are very scarce, and the census which had been planned for 1940 was cancelled due to the outbreak of the Second World War in Europe.

    Reliable birth and death rates are very scanty as well. The authors agree that the registration of births and deaths -as far as it has been performed and the results have been published- was very unreliable. Various estimates of fertility and mortality have been made which do not inspire much confidence either. Widjojo's well-founded estimate of the infant mortality rate is an exception. Only this estimate is based on data concerning western Java in the thirties, and can therefore neither be generalized to other parts of Java nor to other time periods.

    Much less is known about population growth, fertility, and mortality in the Outer Islands (i.e. all islands outside Java). Wander and McDonald both suppose that, in the Outer Islands, population growth was faster than in Java in the early twentieth century. Moreover, Wander conjectures that regional diversity in mortality must have been considerable. She expected the lowest rates to be found in accessible areas where the colonial administration was firmly settled, while the remoter areas had the higher rates. These opinions are very speculative however, since the authors do not mention any reliable demographic data. Only in a few areas does demographic history seem to have been better documented. Widjojo mentions Sumatra's east coast in this respect, where registrations were kept at the agricultural estates, while Jones analyses demographic data from Minahasa (north Celebes), where most of the population was christianized in the nineteenth century.

    Some literature exists about migration regarding Java, as well as the Outer Islands. This pertains to the so-called transmigration, resettlement schemes of the colonial government, and the labour migration between Java and the agricultural and mining estates in the Outer Islands. Furthermore, various authors (Wander, Widjojo, and Hugo) have analyzed the lifetime migration data that were collected in the 1930 census. Widjojo gives a brief description of the most significant migration streams, and Wander tries to characterize the various streams according to distance and destination (rural/urban). Wander and especially Hugo stress that most migration in the period 1880-1930 occurred in consequence of colonial rule.

    The review of literature on the demographic history of Indonesia has revealed that there are many lacunae in the knowledge concerning the period 1880-1942. Most authors suggest or state explicitly that colonial rule had an impact on demographic developments. Since they do not pay much attention to regional diversity, it is far from clear whether colonial rule did enhance or diminish diversity. In order to be able to fill some of the gaps in the knowledge about Indonesia's demographic history, it seems necessary to initially scrutinize the sources of demographic data thoroughly. The results of this endeavour are described in Chapter II.

    During colonial times, most demographic data were collected by civil servants. Consequently, these data were arranged according to the administrative division in force. Part of the Netherlands Indies was governed under a system of indirect colonial rule. Here, the imprint of colonial rule was still of very little consequence during the last decades of the nineteenth century. Attempts of the colonial government to strengthen its hold caused some bloody wars, the one in Aceh lasting thirty years being the most well known. The endeavours to limit the extent of the indirectly ruled areas by force or by peaceful means continuously caused administrative revision of the territory. This is a source of discontinuity in demographic statistics. After about 1910, indirect rule was not confined much further. Ibis did not mean, however, that administrative redivision came to an end. On the one hand, the government needed various revisions of the administrative division in the process of gradually claiming more power in all areas (directly and indirectly ruled). On the other, it had to administer the colony with as few civil servants as possible. Changes in the distribution of the population or in the infrastructure could therefore prompt a redivision. With regard to Java, most administrative redivisions are comparatively well documented. Concerning the Outer Islands however --and especially the indirectly ruled areasinformation is often lacking about the number of inhabitants or the size of the area involved in the redivision. This is a serious impediment to the reconstruction of demographic trends.

    The administrative entities which are used most frequently in this study are the residence (residentie) and the subdivision (afdeling). The residences were the entities on the highest level in the administrative territorial hierarchy. Each residence was divided into several subdivisions, while these in turn were divided into still smaller units.

    The various sources of demographic data and their shortcomings are discussed only briefly here, since a detailed account in English is given in volume XI in the series "Changing Economy in Indonesia" (P. Boomgaard & A.J. Gooszen, 'Population Trends 1795-1942').

    Figures about population size can be derived from the afore-mentioned Population Counts, which have been taken every five years from 1880 to 1905 inclusive. Most parts of the Outer Islands were not included in these Counts. The first census was organized in 1920. It was of a rather simple design and could not be taken in many areas outside Java. The 1930 census was much more elaborate, and had been carefully prepared and well organized. From the point of view of the demographer, it is a pity, however, that no data were gathered on the age structure of the indigenous population. A division was only made between adults, older children, and young children who could not walk yet, the latter category presumably between 0 and 1½ years old. This census was the first one to be executed nearly everywhere in the archipelago. For many areas it yielded the first and also the last reliable population figures the colonial government ever obtained. This does not mean that no insight whatsoever existed on the population size of these areas. Local tribal chiefs or rulers of more complex societies certainly kept a fairly good account of the number of their subjects who owned them taxes and services. However, only fragments of this knowledge actually reached the Dutch functionaries. The native rulers had their reasons not to reveal too much to them. On the other hand, Dutch civil servants were possibly not too keen on knowing more than strictly necessary to carry out their limited administrative tasks.

    With regard to Java, another source of data on population size is available, which has not been used by other authors. It concerns the population figures published annually by the Public Health Service from 1912 to 1933. These figures are not reliable enough to reflect the actual population sizes at the time, but -except for the first few years- they are of a fairly constant quality and are therefore useful to compute growth rates.

    As has been stated by other authors, the registration of births and deaths was very deficient. In the thirties, a system of birth and death certificates was introduced in many Javanese residences, especially in Central Java. At that time, it was felt that this system would improve registration considerably. However, later research revealed that this optimism was misplaced. In the 1970s, when the same system was used, births and deaths were still registered very incompletely. At the same time, registration quality showed considerable regional variation. Birth and death rates can be traced in another way, i.e. in various reports from civil servants, mostly about medical or epidemiological research. Part of this material has not yet been used in the literature on the demographic history of Indonesia. The reports concern small groups and areas in various parts of the archipelago.

    With regard to migration, new material can be found in the extensive reports about the research conducted in the first years of the twentieth century into the deteriorating economic situation of the indigenous population of Java (the so-called 'Mindere Welvaarts Onderzoek').

    The search for demographic data has brought to light that a reconstruction of regional diversity in demographic developments during the last sixty years of colonial rule can only be fragmentary, especially for the years 1880-1900 and 1930-1942. Considering this limitation, it was decided to selectively answer the question about the impact of colonial rule on demographic developments by investigating a few strictly-defined themes. With regard to fertility, two case studies have been carried out, one about the fertility of Javanese estate labourers on Sumatra's east coast, the other on marriage and divorce among the christianized population of the Minahasa (northern Celebes). As far as mortality is concerned, the discussion is turned to western health care and its possible contribution to mortality decline. With respect to migration, Hugo's study is taken as a starting point, and the question is raised whether migration took place without being influenced by colonial rule.

    The reconstruction of regional diversity in demographic developments starts with the two demographic processes which are relatively well documented, i.e. population growth and migration. These processes are described from a bird's-eye view in Chapter III.

    In precolonial and early colonial times, there must have been a considerable amount of migration between the large islands, Java, Sumatra, Borneo, and Celebes, and the rest of the archipelago (henceforward denoted as 'the eastern archipelago'). Some migration streams were curtailed or diverted to other destinations when the Netherlands' East India Company (VOC), and later the colonial government, disturbed and destroyed power and trade relations. Besides this indirect impact on migration, the colonial government tried to control migration and mobility directly by requiring travel passes. This measure turned out to be quite ineffective and was abolished gradually, till most travel and migration was free by the beginning of the twentieth century. Attempts were also made to control outmigration to neighbouring countries, most of them colonized as well at that time. Labour recruitment for agricultural and mining estates in these colonies was effectively restricted. However, the government of the Netherlands Indies did not succeed in prohibiting its subjects to migrate on their own account to Malaya, where they could enlist at the British labour recruitment offices.

    In the early twentieth century, the number of spontaneous and recruited labour migrants to other colonies (including Surinam, one of the Dutch colonies in the West Indies) was greatly surpassed by the number of outmigrants from Java to the agricultural and ruining estates on the Outer Islands, especially Sumatra. Estate agriculture and mining, financed with private western capital, had developed in Java as well as in the Outer Islands since 1870. After 1900, however, the estates really flourished when the demand for rubber rose sharply and petroleum mining became profitable. Local labour not being sufficient in the Outer Islands, the estate owners turned first to China (mostly via Malaya) and, after the turn of the century, more and more to Java to recruit labour. The resulting labour migration got the aspect of forced migration since the labourers, who often had been lured to the estates under false pretences, were severely punished if they did not comply with the very strict rules of their contracts. They were housed on the estates and hardly received any opportunity to mingle with the local population.

    Migration to the estates was by no means the only migration between the large islands during the first decades of the twentieth century. Smallholders' commercial agriculture induced much labour migration as well. These migration streams are not as well documented as the migration to the estates. It is clear, however, that a lot of seasonal migration took place from Banten (western Java) to south Sumatra due to the pepper cultivation in the latter area. Besides, in central Sumatra, smallholder rubber cultivation attracted migrants from Borneo, Java, and Celebes. Just like the estate labourers, these migrants were not integrated in the local population.

    Migrationto Sumatra contributed to sustaining and presumably accelerating population growth on this island. The exact growth rate, however, is difficult to ascertain. Before 1920, growth rates can only be estimated for the south Sumatran residences Lampung, Bangka, and Bengkulu, and for the residence Sumatra's west coast. Elsewhere, population figures are too unreliable and/or computations are hampered by administrative redivision without the required documentation. During the last decades of the nineteenth century, population growth varied from 0.2% per year in Bengkulu to 1.5% per year in Bangka. It seems plausible, therefore, that the growth rate for this area as a whole was not much different from the one in Java at the same time, where it is estimated to have been 1.1 % at the least and not more than 1.5 % per year. After the turn of the century, the growth rate in Java declined to 1.0% in the period 1900-1920, and remained the same during the twenties. In the south Sumatran residences Lampung and Bengkulu, the rate of population growth accelerated during those decades, while the population of Bangka followed the Javanese trend.

    With regard to the other islands, Borneo, Celebes, and the eastern archipelago, up to 1920, population growth can only be computed for two areas, i.e. Minahasa (northern Celebes) and the small island of Amboina, together with the still smaller neighbouring islands of Saparua and Haruku in the eastern archipelago. For the years 1880-1900, both areas showed growth rates which again resembled those on Java, amounting to 1.4% for Minahasa and 1. 5 % per year for Amboina. In the first two decades of the twentieth century, this growth slowed down to 1.3 and 0.6% per year, respectively. During the twenties, an acceleration followed to 1.6% in Minahasa and 1.2% in Amboina.

    Although it is impossible to compute population growth for the population of Sumatra, Borneo, and Celebes as a whole during the twenties (estimates varying from 0.7 to 2.7% per year for Sumatra, 1.4 to 2.8% per year for Borneo, and 1.8 to 3.3% per year for Celebes), one can safely conclude that at least on Borneo and Celebes, population growth must have been faster than on Java at that time.

    Concerning the thirties, population figures are very fragmentary. With regard to Java, a growth rate of more than 1.0 %, but 1.5 % per year at most, can be estimated. This means an acceleration of growth compared to the twenties. On the contrary, available data on the Outer Islands (i.e. West Borneo, Palembang on Sumatra, Minahasa and the Sangir and Talaud isles north of Celebes, and the island of Bali in the eastern archipelago) seem to indicate a decline in growth in the thirties compared to the twenties.

    At the end of Chapter III, attention is paid to one aspect of population distribution, i.e. urbanization. In the various counts and censuses, the criteria for considering a settlement as urban are often vague. Moreover, they are obviously not always the same. Therefore, one criterion is used here, being the number of indigenous inhabitants. Only settlements with 20,000 or more indigenous inhabitants are considered as towns.

    In 1890, only 2.0% of the indigenous population of Java lived in towns of which there were a total of nine at that time. Thirty years later the number of towns had risen to 29. The degree of urbanization (i.e. the percentage of the population living in urban settlements) then amounted to 4.2%. During the twenties, urbanization increased further until, in 1930, 33 settlements could be called urban, housing 5.2% of the indigenous population. On the less densely populated Outer Islands, the number of towns was much smaller and the degree of urbanization lower. The only exception was Borneo, where the degree of urbanization seemed to have been 5.4 % in 1930. It is possible, however, that an inaccurate count of the rural population could account for this figure.

    All three themes discussed in Chapter III, i.e. migration, population growth, and urbanization, are investigated in more detail in Chapter IV. With regard to migration on Java during the years 1880-1905, the qualitative data collected during the 'Investigation into Declining Prosperity' (Mindere Welvaarts Onderzoek) are analyzed. This brings to light that there had been a lot of outmigration from the residence Banten in western Java. Besides the seasonal migration to southern Sumatra, which has already been mentioned, many people left Banten for Batavia. In turn, many people from the residence Batavia were attracted to the residence Priangan in the south of West Java. Presumably they migrated in response to the demand for labour on agricultural estates in Priangan, or in order to reclaim new farm land on their own account. In south Central Java, there was a lot of migration between the neighbouring residence Banyumas and Kedu. Many people from Kedu also migrated to Sumatra's east coast, where agricultural estates developed rapidly. East of Kedu, in the residence Madiun, many people left for the residences Kediri and Pasuruan, situated still farther east. Pasuruan and the adjacent residence Besuki attracted many migrants not only from Madiun, but also from Kediri and from the north coast residences Semarang, Rembang, and Madura.

    With regard to inmigration in Besuki, some very detailed quantitative data are available. These data, which refer to 1906, reveal a strong spatial concentration of migration streams: most migrants came from a few neighbouring districts and settled in one or two villages. This is seen as an indication of chain migration.

    The analysis of (quantitative) lifetime migration data, collected at the census of 1930, reveals that inter-residence migration in Java in the years 1905-1930 showed many similarities with the migration movements in the preceding twenty-five years. Furthermore, these data are used to define so-called migration regions, clusters of continuous residences or subdivisions which have been connected by inter-residence (c.q. inter-subdivision)migration, inmigration from the same area of origin, or outmigration to the same destination. Clusters of residences or subdivisions which were similar with respect to a lack of in- and outmigration are considered as migration regions as well. On each of the large islands, Java, Sumatra, Borneo, and Celebes, several migration regions are defined. Next, urbanization and rural-urban migration in each region are investigated, together with the ethnic composition of its population. From this analysis, the following picture emerges.

    Banten in western Java was an outmigration region consisting of one residence only. The most important destination of the migrants was south Sumatra (see maps 4.6 and 4.7). There was no urbanization in this region, but many outmigrants left for the nearby town of Batavia. The West Javanese residences Batavia, Buitenzorg, Priangan, Cirebon, and the subdivision Brebes; (belonging to the residence Pekalongan) exchanged a lot of migrants. Together they formed a region with much internal migration, but not much in- or outmigration. There were several towns in the region, of which Batavia and Bandung were the most important. They expanded rapidly during the years 1900-1930. Rural-urban migration contributed considerably to this expansion. Part of the migrants to the towns came from outside the region. In eastern Java, a similar region with much internal migration but only few in- and outmigrants can be discerned. It consisted of the residences Surabaya, Bojonegoro, Madura, Madiun, Kediri, Malang, and Besuki. Urbanization started here somewhat later than in the West Java region, and most rural-urban migrants came from inside the region. Surabaya and Malang were the big towns in this region. Along the north coast of Central Java, the residences Jepara-Rembang, Semarang, and Pekalongan (minus the subdivision Brebes) formed a region characterized by a slight outmigration. In 1930, the population of this region was more urbanized than that of West Java. Except for Semarang, the towns in this region only grew slowly during the years 1900-1930, their heyday of expansion having been earlier. In the southern part of Central Java, the residences of Banyumas and Kedu, together with the so-called Principalities (Vorstenlanden), Surakarta, and Yogyakarta, can be considered as one region with a lot of outmigration, especially to the agricultural estates on Sumatra's east coast. The degree of urbanization was lower here than in the north coast region or in the regions West and East Java. Nevertheless, the towns were expanding, particularly Yogyakarta and Surakarta, and rural-urban migration was considerable during the first three decades of the twentieth century.

    Three inmigration regions can be defined on Sumatra of which the so-called Estate area (Cultuurgebied) on Sumatra's east coast attracted the largest number of migrants, mainly from Java (see map 4.9 and 4.11). This region consisted of the residence Sumatra's East coast minus its subdivision Bengkalis plus the subdivision Eastern Aceh (belonging to the residence Aceh). There was only one town in the region, Medan, which had developed since the turn of the century due to the booming expansion in estate agriculture. The subdivision Bengkalis and the adjoining residences Riau and Jambi formed another inmigration region. Here most inmigrants came from southeast Borneo, Sumatra's West coast, and south Celebes. The demand for labour on rubber and copra smallholdings and the opportunity to open up new land were the most important attractions here. There were no towns in this region (i.e. with more than 20,000 indigenous inhabitants) in the first decades of the twentieth century. The third inmigration region on Sumatra consisted of the southern residences Lampung, Bengkulu, Bangka, and Palembang. Here smallholdings as well as estate agriculture and mining and resettlement schemes (transmigratieprojecten) attracted migrants, mostly from Java. There were two towns, Palembang and Teluk Betting. Along the west coast of Sumatra, three residences, Aceh (minus subdivision Eastern Aceh), Tapanuli, and Sumatra's West coast, were characterized by more out- than inmigration. Together they are considered as an outmigration region. The most important destinations were Sumatra's east coast and Malaya. The only town in this region was Padang, which housed a lot of rural inmigrants in 1930. Presumably many of them only stayed in town a few years and returned home afterwards.

    On both Borneo and Celebes, in- and outmigration was concentrated in the coastal areas, leaving the interior undisturbed as non-migration regions (see maps 4.13 and 4.15). The subdivisions Singkawang and Pontianak formed an inmigration region on the northwest coast of Borneo, while in southeast Borneo the subdivisions Banjarmasin, Southeast coast, and Samarinda were characterized by Inmigration as well. The origin of the Inmigrants in the first region is not specified in the census, but the people attracted to the southeastern region are known to have come mostly from Java and Celebes. Many of them found employment on agricultural or mining estates or on smallholdings. Adjacent to this region, the subdivision Hulusungai was characterized by heavy outmigration to Sumatra and the neighbouring southeast coast.

    On Celebes, the southwestern and southeastern peninsula were both outmigration regions as well. From southwest Celebes, Bugis and Makasarese left in all directions, but especially to Borneo and north and central Celebes. From the Sultanate of Buton in the southeast, many people migrated eastward to the subdivision Amboina. On the northern peninsula, the Minahasa area attracted inmigrants from neighbouring Gorontalo and from the Sangir and Talaud isles in the north. These three areas together formed a region where outmigration outnumbered inmigration, Java and south and central Celebes being the main destinations.

    In the eastern archipelago, in- and outmigration to and from the various subdivisions, each consisting of many small islands, was numerically insignificant. Presumably there was a lot of internal migration between those islands, but data about these movements are very scarce. It is clear, however, that the islands of Bali and Lombok together can be considered as a non-migration region.

    In 1930, towns with more than 20,000 indigenous inhabitants were still lacking in the eastern archipelago, the largest settlement being Den Pasar in Bali which did not have more than about 15,000 indigenous inhabitants at that time. Celebes had two towns then: Makasar in the southwest and Manado in the Minahasa area in the north. In 1930, a relatively large part of the small population of Borneo lived in the three towns on this island, Banjarmasin, Pontianak, and Balikpapan. Apart from an underestimate of the rural population, which to some extent might have been responsible for the relatively high degree of urbanization, this has to be attributed to local concentrations of employment outside agriculture, particularly in mining.

    With regard to the ethnic composition of the population, the Outer Islands showed much more variety than Java, where besides the Javanese, only the Sundanese in Banten and West Java and the Madurese in East Java were numerically significant groups (see also figures 4.3 and 4.5). On all islands, but especially on Sumatra, ethnic diversity was enhanced by inmigration. Little integration was to be seen between the various ethnic groups living in the same area. Probably a deliberate colonial policy to keep its subjects divided contributed to this lack of integration.

    Throughout Chapter IV, attention is paid to the impact of colonial rule on migration patterns. Some migration can be directly attributed to the creation of employment on agricultural or mining estates, or in colonial enterprises in various towns. Besides, people were recruited for the colonial army, or they migrated to become domestic servants in European households. However, a lot of migration cannot be considered as a reaction to colonial activities. This holds true with regard to migration towards smallholder agricultural areas, like central Sumatra. Besides, rural-urban migrants often found employment in the indigenous trade or service sector of the economy. Furthermore, migration cannot solely be characterized as a reaction to economic opportunities perceived or expected elsewhere. In some societies, a tradition for outmigration seems to have existed, connected with specific characteristics of the social structure. This seems to have been the case among the Minang-kabau of Sumatra's West coast and also among the Bugis, Makasarese, and some Butonese groups in south Celebes.

    Turning the attention to population growth in the various migration regions, it is only to be expected that regional differences in growth are determined to a large extent by migration. We can expect to find the highest growth rates in inmigration regions and the lowest rates in outmigration regions, while non-migration regions keep an intermediate position. If we happen to find differences in growth that do not coincide with this picture, we detected regional differences in natural increase. In this way, it came to light that in south Celebes (the southwestern and southeastern peninsula combined), the population must have had a larger natural increase during the twenties din in the regions north coast of Java, northwest Sumatra, and north Celebes. All these areas were characterized as 'moderate outmigration' regions. The fact that population growth was much faster in south Celebes than in the other regions lead us to the conclusion that natural increase must have been greater there too. Along the same lines, it is concluded that south Central Java had a larger natural increase than Banten, Hulusungai, and the north coast of Java. Besides, central Borneo and central Celebes must have had a larger natural increase than East and West Java, and Bali & Lombok, while in the latter region natural increase was certainly lower than in East and West Java. Of course, not all differences in natural increase between the various migration regions could be brought forward in this way. It had not been detected yet whether migration and natural increase contributed to a relatively slow rate of growth in outmigration regions or to a relatively fast growth rate in Inmigration areas. This can only be decided after more insight has been gained into the components of natural increase, i.e. the crude birth and death rates, discussed in Chapter V and VI, respectively.

    Since reliable data on the annual number of birth are extremely scarce, Chapter V starts with a search for alternatives. For christianized groups, the annual number of baptisms per 1000 is supposed to give a fairly accurate estimate of the crude birth rate. This is due to the fact that the underregistration of the total number of Christians tended to correspond with the underestimate of the number of births resulting from infant mortality before baptism. Another alternative, i.e. the annual number of first vaccinations against smallpox per 1000 inhabitants, is discarded. The percentages of babies that remained unvaccinated is very difficult to guess. It is only clear that on Java this percentage was constantly changing from 1912 onwards, when a gradual reorganization of the vaccination program started. A third way to gain insight into crude birth rates, namely to estimate them from stable population models, turned out to be useful for a few areas where enough demographic data were available to be used as parameters for estimation. Furthermore, crude birth rates can be estimated by using data from the census of 1930 and computing the number of young children who could not walk yet per 1000 inhabitants at that time. This young children ratio reflects the number of births in the years preceding the census, but is also influenced by infant and child mortality during those years. Besides, for various small groups and areas, birth rates can be derived from reports by civil servants and missionaries.

    Combining these latter data with baptism rates and estimates from stable population models, a fragmentary picture of regional differences in crude birth rates is obtained. This picture is compared to the one of the young children ratios, which could be computed for nearly all migration regions, for several ethnic groups within those regions, and for various towns (see map 5.1). Comparison leads to the conclusion that the regional diversity in young children ratios is a fairly good indicator of regional differences in crude birth rates in the twenties.

    So, without knowing the exact birth rates, each region can be assigned a relative position concerning its birth rate level. Recalling the conclusions of Chapter IV concerning regional diversity in natural increase, it is stated that the relatively large natural increase in south Celebes compared to the north coast of Java, northwest Sumatra, and north Celebes cannot be attributed to a higher crude birth rate. This means that south Celebes must have had a lower death rate than the other three 'moderate outmigration' regions. Similarly, only a relatively low crude death rate can have accounted for the large natural increase that was found in south Central Java in comparison with the north coast of Java, Banten, and Hulusungai- The same holds true for most of the other regional differences in natural increase that could be detected in Chapter IV. In some cases, however, it was the birth rate that caused the difference in natural increase, solely or in combination with the death rate. This has been the case for instance in central Celebes and West Java, compared to Bali & Lombok. In Bali & Lombok, a relatively small natural increase was found that could at least be partly attributed to a lower crude birth rate than in central Celebes and West Java.

    The second part of Chapter V is dedicated to an exploration of the relation between migration and crude birth rates. The aim of this analysis is twofold. Firstly, it is meant to detect whether migration and crude birth rates contributed to relatively high rates of population growth in inmigration regions and relatively low rates in outmigration regions. Secondly, we can only hope to unravel crude birth rates further and lay bare fertility levels if we take the impact of outmigration on the composition of the population into account. Studying young children ratios by migration region, it is apparent that no straightforward relationship existed between birth rates and migration. Relatively high birth rates are to be seen in outmigration regions, as well as in inmigration and non-migration regions, while relatively low ratios can be found in some inmigration regions and in some non-migration regions as well. This is not very surprising, however, since the impact of migration on the composition of the population was not the same everywhere. When a lot of outmigrants left only temporarily, many of whom were married men who went alone leaving their wives and children behind, as was the case in Banten, this resulted in a relatively high crude birth rate compared to, for instance, south Central Java, where many outmigrants were not yet married. On the other hand, inmigration was of course more conducive to high birth rates where the sex ratio of the inmigrants was rather equal and many of them were young couples than if most inmigrants were single men. Unfortunately, not much is known about the age and marital status of migrants.

    It is clear, however, that not all regional differences in young children ratios, and therefore in crude birth rates, can be accounted for by differences in the composition of the population according to age, sex, and marital status. Differences in fertility must have been at stake too, since regional variation is still considerable when the number of young children is related to married women only. To gain more insight into the relation between migration, composition of the population, crude birth rates, and fertility, the attention is focused on four inmigration regions (the Estate area, central and south Sumatra, and southeast Borneo) and the areas of origin of most of their migrants. By decomposing the population of each of the four regions into various ethnic groups, the impact of migration on birth rates is seen more clearly, since 'migrant groups' can be isolated (i.e. groups mainly consisting of recent migrants and their descendants). Then it turns out that the composition of the various groups according to sex, and presumably age and marital status, does account for some of the differences in young children ratios between the groups. Besides, there must have been differences in fertility as well. In the Estate area, south Sumatra, and southeast Borneo, the Javanese and the Bugis had relatively low fertility compared to the indigenous groups in those regions on the one hand, and to the population in their areas of origin (Java and southwest Celebes respectively) on the other. This was partly accounted for by the difficult socio-economic position of the Javanese and the Bugis, which in turn was assumed to be a concomitant of their migrants' status in general and -referring to the Javanese- their position as estate-labourers in particular. Besides, it was supposed that many Javanese migrants, originating from the lower socio-economic classes in Java, continued the habit of these classes to create small families. Furthermore, it seems plausible that living conditions in the Estate area and southeast Borneo were rather unfavourable for all inhabitants (migrants and indigenous population), adversely affecting their fertility. Better economic opportunities were to be found in central Sumatra. The fertility of the migrant groups in this region, Javanese, Bugis, and Banjar, seems to have been at the same level as in their areas of origin. In the case of the Banjar, fertility was probably even higher in central Sumatra than 'at home' in Hulusungai and southeast Borneo. Maybe the rather favourable economic conditions in central Sumatra enabled the migrants to support larger families.

    In general, regional differences in fertility are not directly determined by economic conditions, but only indirectly via variety in marriage patterns and regional diversity in fecundity problems and in the practice of birth control. These direct determinants of fertility are supposed to be influenced by cultural values as well as socio-economic conditions.

    It is difficult to gain insight into the extent to which infecundity determined regional diversity in fertility. Since agents of infecundity like venereal diseases and presumably endemic malaria and malnutrition were not ubiquitous it might be gathered that infecundity varied spatially as well. Still greater uncertainty must be professed with regard to regional variation in the practice of birth control. The variety in ways and means applied to limit the number of offspring certainly was considerable. Much less is known about the frequency with which the various methods were used in various parts of the archipelago, and the impact they had on fertility. Furthermore, we can only guess at the reasons and motives of people to practice some kind of birth control. As mentioned before, economic hardship may have prompted a limitation in family size. Besides, cultural values like the condemnation of pregnancy outside wedlock might have been influential.

    With respect to marriage patterns, the prospects for a reconstruction of regional diversity are somewhat brighter. Investigations are confined to one aspect of the marriage pattern, i.e. age at female first marriage. Marriage being universal everywhere in the archipelago, an impression of the regional variation in age at first marriage can be gathered from the proportion of women never married, as recorded at the census of 1930. Of course this proportion does not reflect regional diversity very clearly, since we cannot control for regional differences in age composition of the population. The picture it suggests is confirmed, however, by evidence of contemporary missionaries and civil servants, and by information collected in surveys conducted in the nineteen sixties and early seventies. In those surveys, data were gathered on individual marital histories, which gave insight into age at first marriage prevailing during the thirties and forties. In those years and presumably also during the twenties, it seems to have been customary for girls in Java to marry very young, i.e. before the age of seventeen at the latest. In West Java and Madura, marriage often took place even before the menarche. In north and south Celebes and in the eastern archipelago, girls married later: in the Minahasa area presumably between the age of 17 and 23, and on the island of Amboina in the eastern archipelago, seldom before they had reached the age of 21. The effect of these regional differences in age at first marriage on fertility levels is supposed to have been slight, since in the case of very early marriage sexual intercourse tended to be postponed, while in societies where later marriage was customary, premarital sexual relations were fairly common and premarital pregnancy was not condemned.

    The reconstruction of regional diversity in fertility levels and their direct determinants is concluded by an inventory of the scarce data available on the exact levels of completed fertility. This brings to light that regional variation in completed fertility is about four to seven children (see map 5.2).

    The last part of Chapter V is dedicated to the impact of colonial rule on fertility. First, the fertility of Javanese estate labourers in the Estate area is discussed. Since the entrepreneurs had a decisive influence on all aspects of the existence of their labourers, it is hardly surprising that they also influenced their fertility. This impact took various shapes however. The policy regarding recruitment, housing, labour contracts, and health care, especially for pregnant women and babies, varied considerably between the estate companies. Some companies stimulated family formation and provided adequate housing in order to create a more or less permanent labour force, while others thought it more profitable to have only a small group of female labourers. Many of those were destined for prostitution or became concubines of unmarried western employees. This variety in policies created considerable differences in crude birth rates and fertility rates between the Javanese labourers of various companies.

    The second case study concerning the impact of colonial rule on fertility refers to marriage and divorce among the christianized population of Minahasa. In Minahasa, the church as well as the civil administration advocated christian values and rules with regard to marriage and divorce. Their attempts were often somewhat halfhearted, since they tried to combine a strict adherence to christianity with a respect and understanding of local values and customs. Consequently, and perhaps also because the Minahassans showed a strong preference for their own culture, marriage habits remained fairly unchanged. This meant that living together without a marriage certificate remained customary and many children were born outside (christian) wedlock. Therefore, it is concluded that the impact of christian values on fertility must have been negligible in Minahasa.

    In Chapter VI, mortality trends in various parts of the archipelago are discussed. The attention is first turned to Java and, in the second part of the chapter, to the Outer Islands.

    With regard to Java, the annual number of deaths by residence, published by the Health Service based on desa-registration, are an important source. These figures have been published from before 1880 to 1895 and from 1911 to 1940. They are used to analyze annual fluctuations in the number of deaths by migration region, and to detect middle-long-term trends in mortality levels by region. During the first period, 1880-1895, annual fluctuations in the number of registered deaths were severe, particularly in Banten and West Java. These fluctuations were mainly due to outbreaks of famine and 'fever' (presumably malaria) (see Figure 6. 1). Besides, the eruption of the Krakatau volcano in 1883 caused many deaths, illness, and starvation.

    Looking at the average number of registered deaths per 1000 inhabitants in five-year periods from 1880 to 1895, no clear trend can be observed. If the 1870s are taken into account, as can be done for the north coast region, some residences in south Central Java and the West javanese residence Cirebon, a decline of mortality becomes apparent. This decline cannot be attributed to better economic conditions. It is supposed that local improvements in water supply lowered mortality due to cholera, typhus, and dysentery to some extent. Moreover, efforts to control malaria by the supply of quinine and by local improvements in drainage might have had some success. The vaccination program against smallpox had been rather effective, at least since the 1850s, but till the end of the century the disease had not yet been eradicated and the campaigns still met with various difficulties. It does not seem likely, therefore, that mortality due to smallpox declined further during the last decades of the nineteenth century.

    After the turn of the century, colonial policy changed. Henceforward, more attention was paid to improving the living conditions of the indigenous population. The effect of this so- called 'ethical' policy on economic conditions was not very marked. The government could only preclude new outbreaks of famine after the one that had raged in Central Java in the years 1900-1902. With regard to health care, more success can be noted. After 19 10, the fight against smallpox and cholera was intensified and mortality due to these diseases declined. At the same time, pest broke out in East Java, and although much manpower and money were spent to control it, the disease slowly spread westwards over Java (see Figure 6.2). However, the efforts of the Health Service did prevent mortality from rising steeply due to pest. Meanwhile another disease caused a sharp rise in the number of deaths. This was the influenza epidemic which broke out in 1918. It raged particularly in East and south Central Java and in the north coast region. Afterwards, during the twenties and thirties, annual fluctuations in the number of deaths were only slight (see Figure 6.3). This might be seen as an indication of mortality decline, a decline which becomes more apparent if we look at the five-year average number of registered deaths per 1000 inhabitants. Not only the attempts to eradicate smallpox and cholera contributed to this decline, but also large-scale, scientifically based measures to control malaria, which started in the twenties. At the same time, propaganda campaigns for better hygiene were organized, which were intensified in the second half of the thirties.

    Mortality decline showed some regional diversity. After a universal decline in the first half of the twenties, to be considered as an aftermath to the influenza epidemic, mortality declined further in East Java, north coast, and -although slightly- in south Central Java during the second part of the twenties. In Banten and West Java, however, mortality levels stagnated at that time. During the thirties, a decline likely occurred in all regions, although locally, particularly in Priangan (West Java), the number of deaths rose markedly due to a severe outbreak of pest in the early thirties.

    Turning the attention to population growth, it is noted that in Banten and West Java growth slowed down in the early twenties to speed up again during the years 1925-1929. This leads to the conclusion that changes in mortality levels did not have a perceptible impact on population growth there. Since we characterized West Java as a non-migration region, it is assumed that changes in fertility were decisive for the growth rate in this area during the twenties. In the outmigration region Banten, the same seems to hold true with regard to the years 1925-1929. In the regions East Java and north coast, mortality decline is seen to go hand in hand with an acceleration in population growth during the twenties, while the same can be said of south Central Java in the years 1920-1924. In the second half of the twenties, growth slowed down in this last region. This can be attributed to heavy outmigration.

    Although trends in mortality could be reconstructed, the exact level of mortality is very difficult to ascertain, because of the unreliability of the death registration. Moreover, the quality of this registration varied considerably between regions. Consequently, it is impossible to assign relative positions to the regions and to discern regions with 'relatively high' or 'relatively low' mortality rates.

    One might ponder over regional diversity in the impact of western health care measures. It can be supposed, for instance, that in the region north coast mortality particularly declined due to malaria control, or that in south Central Java mortality decline was especially marked during the thirties because of very active hygiene propaganda. This kind of reasoning is speculative though, since information about the location and spatial range of health care measures is very scanty indeed. It seems probable, however, that the spatial range of various measures has been quite restricted. Therefore, western health care is supposed to have caused more intra- than interregional diversity in mortality.

    One example of this intraregional diversity is explored further, namely the differences in mortality between the town of Batavia and its surrounding region West Java. Early in the twentieth century, the crude mortality rate had been much higher in the town than in the region as a whole. This holds true even if we account for a difference in the reliability of the registration. Around 1930 the difference had vanished and the crude death rate was even lower in Batavia than in West Java. This was not due to better survival chances in town, but to the fact that the town's population was much younger on average. When comparing infant mortality rates, it seems very probable that mortality risks were higher in Batavia than in West Java as a whole. Considerable spatial variation in infant mortality rates was found within the town, rates being particularly high in the slum-like outskirts.

    Data about mortality on the Outer Islands are very fragmentary. It is possible, though, to gain some insight into the spatial variety in infant mortality from data collected in various surveys (see map 6. 1). Wander's hypothesis that during the twenties and thirties mortality rates were lowest in the areas where the colonial impact on economic and health conditions was greatest, while the highest rates could be found in the remotest areas, cannot be confirmed by these data. Among some isolated tribes in the eastern archipelago, infant mortality was only slightly higher than in the Karo-lands in the residence Sumatra's East coast, where economic conditions were favourable and western health care was relatively well provided for. In some remote places in New Guinea even lower rates were found than in the Karo-lands and Minahasa, while especially in the latter area the colonial impact on health care was rather marked.

    A comparison of the Outer Islands with Java regarding western health care leads to the following observations.

    In the twenties and thirties, most Outer Islands' residences counted more physicians relative to the number of inhabitants than the residences in Java. Since the Outer Islands were far more sparsely populated, this did not mean that all people there had better access to western medical care. As far as the spread of epidemics is concerned, the low population density was an asset in some cases. Control on hygiene and health care being very strict in harbour towns, 'imported' diseases like pest and cholera could be checked and prevented from spreading in the sparsely populated hinterland. With regard to smallpox, on the other hand, low population density was a drawback, since it hindered an efficient organization and supervision of vaccination campaigns. Consequently, various outbreaks of smallpox occurred during the 1910s and '20s. Malaria control was attempted as vigorously in the Outer Islands as in Java. Nevertheless, the disease remained endemic in various areas and caused high mortality in many places, especially in south Sumatra.

    It is stated further that, due to the low population density, the spatial range of western health care measures probably was even more limited in the Outer Islands than in Java.

    One group of the indigenous population received more medical attention than any other in the whole archipelago, namely the estate labourers in the Estate area. The chapter concludes with a discussion of mortality and western health care among that group.

    From about 1910 onwards, most entrepreneurs in the Estate area strived to secure good health for their labourers. Considerable investments were made to establish modem hospitals and to improve housing and hygiene. On long-established estates, these efforts met with some success: mortality declined sharply until, at the end of the twenties, it reached a level that was in all probability somewhat lower than in Java. On newly-opened land, mortality remained very high indeed. Compared to the amount of energy and money spent, the results of all efforts were not very impressive, even if only long-established estates are taken into account. This was partly due to the fact that attention was strictly limited to the labourers, while the members of their families remained unprovided for. Consequently, infectious diseases could spread easily. New labourers arriving from Java formed another source of infection.

    The last chapter of the study provides a synopsis of major findings. It is concluded that, although colonial rule had a considerable impact on demographic developments during its last sixty years, its influence should not be overemphasized: much demographic behaviour remained beyond the reach of colonial rulers. Meanwhile, spatial demographic diversity tended to increase rather than decrease.

    Original languageDutch
    QualificationDoctor of Philosophy
    Awarding Institution
    • van der Woude, A.M., Promotor
    Award date5 Oct 1994
    Place of PublicationS.l.
    Print ISBNs9789054853008
    Publication statusPublished - 1994


    • demography
    • migration
    • settlement
    • social structure
    • colonialism
    • history
    • indigenous people
    • censuses
    • netherlands east indies

    Cite this

    Gooszen, A. J. (1994). Een demografisch mozaiek, Indonesie 1880 - 1942. S.l.: Gooszen.