Toward Rising Non-Permanent Population Mobility: A case of commuters in Indonesia 1

Toward Rising Non-Permanent Population Mobility: A case of commuters in Indonesia 1 Evi Nurvidya Arifin (enarifin@gmail.com) Universitas Indonesia and Universitas Respati Indonesia Aris Ananta (arisananta@gmail.com) Universitas Indonesia Abstract This paper analyses the first nationwide statistics on commuters, one type of nonpermanent mobility, in Indonesia. This is also the first nationwide statistics on nonpermanent mobility in Indonesia. Currently, most studies on population mobility is related to the permanent population mobility or migration. Yet, non-permanent mobility has been seen as a global emerging phenomenon in both developed and developing countries. Since 2011, the Statistics-Indonesia has made a breakthrough by collecting information on non-permanent population mobility through its Sakernas. This paper uses the data, particularly on commuters during 2011-2014. The study finds a rising trend in number and rate of commuters in Indonesia between 2011 and 2014. Jakarta, Banten and Yogyakarta are the provinces with the highest commuter rates. The contribution of commuters traveling longer time, between 30 and 120 minutes, is increasing. Yet, the distance they travel does not seem to change significantly. Most commuters travel by private transportation and the trend is increasing. Commuters are more likely to be found among urban, male, higher educated, and youth population. Furthermore, commuters are more likely to be observed among ever married workers. A caveat should be noted that the data is limited to the working population aged 15 and above. Further and deeper analyses should be carried on other types of non-permanent population mobility. Keywords: Commuters, Circulars, Non-permanent Population mobility, Indonesia The First Nationwide Statistics on Non-permanent Mobility This paper produces the first analysis on nationwide statistics on commuters, one type of non-permanent mobility, in Indonesia. This is also the first nationwide statistics on non-permanent mobility in Indonesia. It investigates the size, distribution and the rate of commuters in Indonesia. In particular, the results can provide the linkage between population spatial mobility and development, including provision of public and private goods. Currently, most demographic studies on population mobility is related to the permanent population mobility or migration, population movement involving change of residence. As a result, the non-permanent mobility, the population mobility without involving any change of residence, has been understudied due to the lack of the data. 1

Yet, non-permanent mobility has been seen as a global emerging phenomenon in both developed and developing countries. With its Sakernas, since 2011, the Statistics-Indonesia has made a breakthrough by collecting information on non-permanent population mobility. It focuses on commuters and circulars, two types of non-permanent mobility. This paper uses the data from the National Labour Force Survey, particularly the data on commuters, one type of non-permanent population mobility, during 2011-2014. It calculates the rate and prevalence of commuters; pattern of commuters by age, sex, urban-rural, provinces, marital status, educational attainment, employment status and industry, distance and travel time of commuting. The use of data from 2011 to 2014 allows the paper to calculate and analyse the trend of commuting during this period. This paper starts with a brief discussion on stages of population mobility, to show the relative role of no-permanent population mobility in today global demographic challenges. It then discusses how it measure commuters, before showing who the commuters are. Because of space limitation, this paper has not analysed the circulars. There is also an important emerging type of non-permanent mobility which has not been quantified and the information has not been collected in Susenas. This is the wira wiri type, people who move frequently to many places without regular pattern yet they do not change their residences. Hopefully, future surveys can capture this wira-wiri type too. 2 Stages of Population Mobility A theory of population mobility (Zelinsky 1971; Zelinsky and Newman 1974; Skeldon 1990) posited that at an early stage of development population mobility usually occurs at short distance and periods, the so-called non-permanent population mobility. With further development, population mobility takes a longer distance and period. This mobility involves a change of residence, but it does not have to be related to a change of citizenship. Somebody changes residence if he/she stays for longer than a period of time in the destination area. In Indonesia, the threshold is 6 months. This kind of mobility is usually called permanent mobility or migration. As development further accelerates, the theory argues that people will return to non-permanent mobility, population mobility without change of residence. However, this time the non-permanent mobility becomes more complex. It can take a long distance, but travel in a short time as made possible by advancement in transportation system. Currently, non-permanent population mobility has been perceived as an emerging global phenomenon, seen both in developed and developing countries (Kraft, 2014). One of the most often discussed type of population mobility is commuters. (Kraft 2014; Kimbrough 2016; Galvin and Madlener, 2014). There are two possible reasons for people to commute. First is the job-house-skill hypothesis. People decide to commute because miss-match of skill, lack of houses close to the place of work, family matters, and availability of transportation system. Second is the city network hypothesis. Here commuting is not something undesirable. It is enriching the city network. People commute because they love commuting. (Galvin and Madlener, 2014). However, conventional demographic data are mostly on permanent population mobility or migration, population movement with change of residence. Therefore, the data on non-permanent mobility is not available as widely as the ones on migration. Even, in many countries there is no data on non-permanent population mobility. 2

Indonesia, a developing country in Southeast Asia and the country with the fourth largest population in the world, is not an exception. It is possible that non-permanent population mobility is neither a new trend nor phenomenon in Indonesia. It may affect the number and composition of population and therefore it brings about social, economic, and political implications in development planning. The issue of the absence of such data has been recognized by Hugo (1982). He explained that the absence of the statistics was due to the stereotype that Indonesians were immobile as the working population were mostly farmers who tend to be born, live and die in the same place. However, the pattern may have changed, where non-permanent mobility may have been intensifying in the last decade partly due to the advancement and changes in transportation and communication technology. Small scale studies were available for a certain area like Medan (Leinbach and Suwarno 1985), the latest in Jakarta (Maimunah and Kaneko 2016). Measurement of Commuters The SAKERNAS (National Labour Force Survey) conducted by the Statistics-Indonesia permits collects data on non-permanent mobility such as commuting and circulating. This survey is conducted annually for every quarter of the year and thus possible to have a quarterto-quarter analysis. This paper focuses the analysis of commuters, taking into account key variables such as age group, gender, urban and rural, provinces, marital status, educational attainment, employment status and industry, distance and travel time of commuting. A caveat should be noted that the data on commuters is only asked to the workers aged 15 and above. This definition excludes the population who commute for not working activities such as students travelling daily from home to school located in other districts, household wives shopping to the market for daily needs to the neighbouring districts. The survey question is as follows Apabila di luar kabupaten/kota tempat tinggal, apakah [NAMA] pergi dan pulang ke/dari tempat kerja setiap hari, setiap minggu, atau setiap bulan? [If the respondent live in different district of residence, does he/she go from and return home every day, every week, or every month?. Here, commuters are the ones answering every day and others are circulars. Those who travel from home and place of work located in the same district are defined as stayers. 3 The Statistics-Indonesia publications provide weighted numbers and percentage distribution of the commuters based on the selected variables. Each survey gathered information from around 200,000 households. Yet, the third quarter data gathered in August have larger samples to enable the calculation of statistics at the district level. To assess the trend across time, this paper utilises the published data on commuters from four consecutive surveys conducted in the August (third quarter) for the years of 2011, 2012, 2013 and 2014 (Badan Pusat Statistik, 2015). In addition, several other publications are used to get more information on the workers by the selected variables. Therefore, this paper produces rates of commuters by several variables. Overall, the commuter rate is calculated as follows CR 15+ = N 15+ W 15+ x 100 Where N15+ is the number of commuters aged 15 years old and over W15+ is the number of workers aged 15 years old and over 3

The commuter rate by age, sex, province and place of residence is calculated as follows CR ij = N ij W ij x 100 Where i = 1, 2, 3 and 4 represents age, sex, province and place of residence and j refers to categories in each of these selected variables. Mobility is highly selective by age and sex. Ideally, age should be grouped into more detailed groups to capture life cycle pattern of commuting. However, the published data presents age broadly into two groups: youth (15-24) and adult (25 and above). Sex is differentiated between male and female. Province is grouped into 33 provinces and place of residence is differentiated between urban and rural residence. Scatter plot is carried out to draw regional variation between commuter rate and percentage of workers who changed job. A scatter plot is also created to examine sex differences of commuting rate among provinces. Who are the Commuters? Table 1 shows that the majority of the workers are stayers, meaning that they are neither commuters nor circulars. However, as mentioned earlier, these stayers do not necessarily mean that they do not move at all. These stayers may include the wira wiri, those who move without any clear pattern and not involving any change of residence, which is not captured by the data. The rate of commuters rose slightly during the four years, from 5.6 % in 2011 to 6.1% in 2014, though the number tends to increase significantly, from 5.98 million in 2011 to 6.94 million in 2014. The number fluctuated with an increasing trend as seen in Table 1. This tremendous increase may have affected public service and the use of transportation infrastructure in particular as well as the daily flow of traffic. The number of commuters also affects the dynamics of population between day and night times. It should be noted that the increase in numbers of commuters may be a result of the rising number of workers in general. Table 1. Commuters and Circulars: Indonesia, 2011-2014 Year Commuters % Circulars % Stayers % Total Workers 2011 5,982,592 5.6 2,138,875 2.0 99,294,842 92.4 107,416,309 2012 6,441,287 5.7 2,354,754 2.1 103,708,827 92.2 112,504,868 2013 6,379,279 5.7 1,955,443 1.7 104,426,350 92.6 112,761,072 2014 6,938,680 6.1 2,241,727 2.0 105,447,619 92.0 114,628,026 Source: compiled and calculated from Badan Pusat Statistik (2015). In addition, there are also workers who go and return to their place of residence in regular times, but longer than a day, called circulars. Yet, as seen in Table 1, the circulars have much smaller rates than commuters. Furthermore, Table 2 indicates that most of those who commute to work use private transportation. Over time, more and more commuters use private transportation, rising from 73.4 percent in 2011 to 77.4 percent in 2014. A study in Jakarta (Maimunah and Kaneko 2016) also found that the use of motorcycle is very high, although the mode of transportation may be linked to the educational level in which the use of cars is more likely among university graduates. Much smaller percentages of them use public transportation. About one in five commuters use public transportation, and declined to 16.9 percent in 2014. This may reflect 4

the low development of reliable and affordable public transportation. Shared transportation is another type used by much smaller percentages of commuters. This leaves to very small percentage of commuters travelling with no transportation, or walking to work (Table 2). Table 2. Modes of Transportation used by Commuters: Indonesia, 2011-2014 Year Public Shared Private Not use 2011 20.54 5.22 73.35 0.89 2012 19.99 4.79 74.52 0.70 2013 16.87 5.04 77.44 0.65 2014 16.87 5.04 77.44 0.65 Source: compiled from Badan Pusat Statistik (2015). More men than women commute to work. As presented in Table 3, the ratio between male and female commuting workers is above 2.3, meaning that the number of male commuting workers are more than double the women, during these four years. The rate of commuters is higher among male workers than among female workers. In each sex group, the rate was rising. Table 3. Commuters by Sex: Indonesia, 2011-2014 Number Commuter Rate Year Male Female Sex Ratio Male Female 2011 4,190,743 1,791,849 2.339 6.29 4.39 2012 4,494,853 1,946,434 2.309 6.44 4.59 2013 4,565,118 1,814,161 2.516 6.51 4.29 2014 4,855,688 2,082,992 2.331 6.79 4.84 Source: Authors calculation. The rate of commuters is also increasing from 5.6 percent in 2011 to 6.1 percent in 2014 as seen in Figure 1. Figure 1 also shows that for every 100 male workers, there are more than 6 people who commute to place of work. This male rate is larger than the female, about 4 to almost 5 female commuters for every 100 female workers. 8.00 7.00 Figure 1. Trend in Rate of Commuting Workers by Sex: Indonesia, 2011-2014 6.00 5.00 4.00 3.00 2.00 2011 2012 2013 2014 Source: drawn by the authors. 5 Male+Female Male Female

Though in some provinces the rate of male commuters is smaller than that of female commuters, in many provinces the rate of male commuters rate is larger than that of female rate. See Figure 2. Figure 2. Scatter Plot of Commuter Rate by Sex among Provinces: Indonesia, 2015 Male 30.0 25.0 DKI 20.0 15.0 10.0 Bali Jabar DIY Banten 5.0 Source: Authors calculation 0.0 0.0 5.0 10.0 15.0 20.0 25.0 30.0 Female Commuters aged 25 and above accounted for more than 82 percent. Thus, the youth (15-24) commuters accounted for about 17 percent (Table 4). This may partly reflect the much higher percentage of workers at age 25 and over. 4 Yet, the youth is more likely to commute, as the rate of commuters is higher among the youth than the older group. Furthermore, commuting is a more urban phenomenon. As can be seen from Table 5, urban commuters account for more than 80 percent of the total commuters. Over time, the number of urban commuters increased faster than rural commuters. The rate of commuters is much larger in urban than in rural areas. It was 9.7 per 100 workers in urban areas as compared to just 1.8 in rural areas. These rates increased to 10.4 and 1.9, respectively, in 2014. This may reflect living and working in different environmental setting where the jobs are more available in places far from the houses. Table 4. Commuters by Age: Indonesia, 2011-2014 Year Number Distribution Commuter Rate 25 and 25 and 25 and 15-24 above 15-24 above 15-24 above 2011 1,015,684 4,966,908 17.0 83.0 6.20 5.46 2012 1,140,508 5,300,779 17.7 82.3 6.73 5.55 2013 1,070,102 5,309,177 16.8 83.2 6.51 5.51 2014 1,224,231 5,714,449 17.6 82.4 7.82 5.77 Source: Authors calculation 6

Table 5. Commuters by Place of Residence: Indonesia, 2011-2014 Year Number Urban Commuter Rate Urban Rural Commuters (%) Urban Rural 2011 4,989,529 993,063 83.4 9.67 1.78 2012 5,317,872 1,123,415 82.6 9.86 1.92 2013 5,279,131 1,100,148 82.8 9.72 1.88 2014 5,803,487 1,135,193 83.6 10.43 1.92 Source: compiled and calculated from Badan Pusat Statistik (2015). The majority of commuters are seen in Java Island, the smallest, mostly densely populated, main island in Indonesia. They accounted for 81 percent of the total commuters, much larger than the percentage of population in Java (56.8 percent in 2015). The nature of the island in tandem with a better transportation system makes possible for workers to commute from one district to another. This commuting phenomenon resulted in heavy traffic on the morning and afternoon peak hours. West Java is the largest contributor of commuters with almost a quarter of total commuter in 2014 or 1.7 million workers. This is followed by Jakarta, East Java, Central Java, Banten and Yogyakarta. Outside Java, Bali and North Sumatra are the only provinces having large numbers of commuters. The commuter rate also varies among provinces. Jakarta stands as having the highest commuter rate with almost 24 out of 100 workers in 2014. The second one is Banten (17 percent) and Yogyakarta (14 percent). The rate in West Java is 9 percent. Outside Java, Bali has the largest commuter rate (9.5 percent), followed by Gorontalo (5.3 percent). Commuters are mostly ever married workers. Although over time, the share of single has slightly increased from 27.0 percent in 2011 to 28.2 percent in 2014 (Table 6). The commuter rate is much higher among ever married than single workers. Table 6. Commuters by Marital Status: Indonesia, 2011-2014 Year Single Ever married Single Ever married 2011 1,617,874 4,364,718 27.0 73.0 2012 1,747,986 4,693,301 27.1 72.9 2013 1,772,474 4,606,805 27.8 72.2 2014 1,955,380 4,983,300 28.2 71.8 Source: Compiled from Badan Pusat Statistik (2015). With regards to educational attainment composition, the largest share is commuters holding senior high school certificate with around 44 percent, followed by those holding tertiary education certificate. Over time, the share of the most educated ones tends to increase between 2011 and 2014 (Table 7). 5 Table 7. Commuters by Educational Attainment: Indonesia, 2011-2014 Year No schooling Incomplete Primary Primary Junior HS Senior HS Tertiary education 2011 0.6 3.6 9.9 18.0 43.8 24.1 2012 0.6 3.6 11.1 14.6 43.1 27.1 2013 0.5 3.0 9.5 13.8 44.9 28.2 2014 0.5 3.3 9.8 12.7 44.8 28.9 Source: Compiled from Badan Pusat Statistik (2015). 7

Trade, restaurants and accommodation are mostly in urban or business district areas. Those commuting to work for this sector account for very significant portion followed by the ones working in social services. Meanwhile, the commuters work in manufacturing are also large. As seen in Table 8, almost 70 percent of commuters work in these three main industries. 6 Table 8. Commuters by Industry: Indonesia, 2011-2014 Year 2011 2012 2013 2014 Agriculture etc. 1.7 2.07 2.10 2.01 Mining and quarrying 0.92 0.88 0.77 0.81 Manufacturing 22.57 25.4 24.17 24.28 Electricity, Gas and Water 0.49 0.59 0.65 0.54 Construction 8.05 8.34 7.49 8.05 Trade, Restaurant and Accommodation 24.34 22.96 22.99 24.65 Transportation, Storage and Communication 7.19 7.21 7.55 6.95 Finance, real estate and business services 9.87 9.14 9.68 8.93 Social services 24.87 23.41 24.62 23.78 Total 100.00 100.00 100.02 100.00 Source: Compiled from Badan Pusat Statistik (2015). How Far and How Long Do They Travel? Almost 70 percent of the commuters travel to work less than 30 km away from their homes with the majority travel in shorter distances, between the range of 10 and 29 km (Table 9). Around 27 percent commute for more than 30 km and the remaining (about 3 percent) do not know their distance to work. The percentages fluctuate and may not show any significant change during the four years. Table 9. Commuting Distance from home (km): Indonesia, 2011-2014 Don't Year <10 10-29 30+ know 2011 22.50 47.34 27.50 2.66 2012 22.61 47.52 26.95 2.92 2013 22.14 47.36 27.27 3.23 2014 22.33 47.54 26.89 3.25 Source: Compiled from Badan Pusat Statistik (2015). Furthermore, Table 10 shows an indication of traveling longer, but not very long. Almost half of them travel between 30 minutes to an hour and the percentage rose. About one fifth travel between one and two hours and the percentage also rose. This is the opposite to those travelling shorter time, less than 30 minutes, with declining percentage. The percentage of those traveling very long, more than 2 hours, also declined. 8

Table 10. Commuters travelling time from home: Indonesia, 2011-2014 Year <30 31-59 60-119 120+ 2011 29.74 45.05 20.89 4.32 2012 29.70 46.55 20.13 3.61 2013 28.21 46.67 21.04 4.08 2014 28.33 46.36 21.44 3.86 Source: Compiled from Badan Pusat Statistik (2015). Concluding Remarks This is a preliminary study using the recently available data on non-permanent population mobility in Indonesia, particularly on commuters. This study is expected to be followed by further and deeper studies using the data, to fill in the absence of quantitative analysis on non-permanent population mobility in Indonesia. This study finds the existence of variation of commuters by age-sex-educational status-marital status and location (urban-rural residence, Java vs outside Java Island, and provinces). This variation may affect number and composition of population, which are needed for development planning. It also finds a rising trend of commuting during the fouryear period. Therefore, to make better development planning, time has come to widen demographic analysis to include impact of non-permanent mobility. Non-permanent population mobility is expected to play a more important role in future development planning. Further data collection and deeper analyses should be carried out on commuters and other types of non-permanent population mobility. References Ananta, Aris and Evi Nurvidya Arifin. Emerging Patterns of Indonesia s International Population Mobility, Malaysian Journal of Economic Studies, 51 (1), pp.29-41, 2014. Badan Pusat Statistik. 2015. Statistik Mobilitas Penduduk dan Tenaga Kerja 2015. Jakarta: Badan Pusat Statistik. Galvin, Raymond and Reinhard Madlener. Determinants of Commuter Trends and Implications of Indirect Rebound Effects: a case study of Germany s Largest Federal State of NRW, 1994-2013, FCN Working Paper no 9/ 2014. Published by Institute of Future Energy Consumer Needs and Behavior, RWTH Aachen University, Aachen, Germany. Hugo, Graeme J. 1982. Circular Migration in Indonesia, Population and Development Review, vol. 8, no. 1 (March), pp. 59-83. Leinbach, Thomas and Suwarno. 1985. Commuting and Circulation Characteristics in the Intermediate Sized City: The Example of Medan, Indonesia. Singapore Journal of Tropical Geography, Vol. 6, No. 1: 35-47. Maimunah, Siti and Shinji Kaneko. 2016. Commuters Behaviors and Attitudes toward Current Transport Mode Chosen. Journal of International Development and Cooperation, Vol.22, No.1 & No.2, 2016: 91-105. 9

Kimbrough, Gray. Measuring Commuting in the American Time Use Survey, Working Paper 15-20, May 2016. University of North Carolina at Greensboro. Kraft, Stanislav. Daily spatial mobility and transport behaviour in the Chech Republic: pilot study in the Pisek and Bystrice and Pernsteynem Region, Human Geography, Journal of Studies and Research in Human Geography, vol. 8, no. 2, Nov. 2014. Skeldon, Ronald. Population Mobility in Developing Countries. Belhaven, North Carolina, US: Belhaven Press, 1993 Zelinsky, Wilbur. The Hypothesis of the Mobility Transition, Geographical Review, vol. 61, no. 2, Apri, 1971, pp 219-249 Zelinsky, Wilbur and James L. Newman. Hypothesis Revisited, Annals of the Association of the American Geographers, vol. 64, no. 1, March 1974, pp. 185-187. Notes 1 This paper is presented at the XXVIII IUSSP Conference, 30 October 4 November 2017, Cape Town, South Africa. 2 The word wira-wiri was first used in Ananta and Arifin (2014). 3 It should be noted, as mentioned earlier, that the Sakernas has not included wira-wiri, those people who move frequently without any regular pattern and yet do not involve change of residence. In Sakernas, this wira wiri people are treated as stayers. 4 Further studies should make a better age categorization. This paper is not able to do a better age grouping as this paper used the published data. Future publication may make a better age classification. For example, the classification can use 15-24, 25-34, 35-44, 45-54, 55-64, and 65 + age groups. 5 In this paper we do not calculate the commuter rate by education because the data are not easily available. 6 In this paper we do not calculate the commuter rate by industry because the data are not easily available. 10