PROSIDING PERKEM ke-9 (2014) 303 - 310 ISSN: 2231-962X

Does Declining Fertility Impactedon Female Labor Supply? Lesson from Indonesian Demographic Changes Achmad Sjafii Fakulti Ekonomi dan Pengurusan Universiti Kebangsaan Malaysia [email protected] Faridah Shahadan Fakulti Ekonomi dan Pengurusan Universiti Kebangsaan Malaysia [email protected] Doris Padmini Selvaratnam Fakulti Ekonomi dan Pengurusan Universiti Kebangsaan Malaysia [email protected]

ABSTRAK Perubahan demografi Indonesia tempoh masa 1970-2010 telah merubah struktur umur penduduk dengan ketara. Salah satu daripada fenomena yang hadir ialah penurunan kadar kesuburan dan kematian. Indonesia, begitu juga dengan kebanyakan negara-negara telah mengalami penurunan kadar kesuburan dan kematian. Ini akan diikuti oleh beberapa kesan positif kepada percepatan ekonomi. Penurunan kadar kesuburan telah menjadi salah satu punca prestasi ekonomi di beberapa negara terutamanya kes di negara-negara membangun. Penurunan kadar kesuburan akan membawa kepada peningkatan peratusan penduduk umur bekerja. Pengalaman negara-negara maju dan negaranegara perindustrian baru seperti 'the tigers of Asia', nisbah yang lebih kecil anak-anak muda kepada penduduk umur bekerja boleh membuktikan implikasi positif kepada pembangunan ekonomi. Perubahan demografi mempengaruhi perkembangan ekonomi melalui saluran berikut: (1) akibat daripada peralihan struktur umur; (2) akibat daripada kadar kematian yang lebih rendah dan meningkatkan panjang umur. Peralihan struktur umur menyebabkan berkurangan di sebahagian daripada penduduk muda, manakala sebahagian daripada penduduk umur bekerja semakin meningkat. Ini bermakna pergantungan kumpulan umur muda berkurangan berbanding dengan beban kebergantungan total. Ini mempunyai kesan ke arah meningkatkan bekalan tenaga pekerja perempuan dalam pasaran buruh di negara-negara membangun. Beberapa kajian menunjukkan bahawa kejatuhan tanggungan umur belia nisbah menyumbang kepada pertumbuhan ekonomi.Tujuan daripada paper ini adalah untuk mengkaji hubungan sebab dan akibat antara TFR yang merosotdan penyertaan tenaga buruh wanita (FLP) di Indonesia 1990-2012 oleh ujian Granger causality. Kata kunci: Kadar kesuburan, penyertaan buruh perempuan, perubahan demografi

ABSTRACT The Indonesian demographic changes during 1970-2012have change the age structure of the population significantly. One of the phenomenon which appear is a decline both in the fertility and mortality rate. Indonesia, likewise in most of the countries experienced declining of fertility and mortality rate. This will be followed by some positive impact on acceleration of economic. The decline in fertility rates has became one of the causes of the increased economy performance in several countries especially the case in the developing countries. The decline in fertility rate will lead to the increase of the percentage of the working-age population. The experience of developed countries and newly industrialized countries such as ’the tigers of Asia’, small ratio of young people to the working-age population are could prove positive implication to economic development. Demographic changes influenced economic development through the following channels: (1) as a result of an age-structure transition; (2) as a result of lower mortality rate and increasing longevity. An agestructural transition causes decreasing in the portion of young population, while the portion of working-age population is increasing. This means the young age group dependency decreases relative to the total dependency Persidangan Kebangsaan Ekonomi Malaysia ke-9 (PERKEM ke-9) “Urus Tadbir Ekonomi yang Adil : Ke Arah Ekonomi Berpendapatan Tinggi” Kuala Terengganu,Terengganu, 17 – 19 Oktober 2014

304

Achmad Sjafii, Faridah Shahadan, Doris Padmini Selvaratnam

burden. It has impact toward increasing supply of labor in economy, for example increasing female in labor market in developing countries. Studies show stated that falling youth-age dependency ratio contributed to the economic growth. This objective of paper is to examine relationship causality between the declining TFR (due to of demographic changes) and female labor participation (FLP) in Indonesia from 1990 to 2012 by Granger causality test. Keywords:Fertility Rate, Female Labor Participation, Demographic Changes

BACKGROUND In the last two centuries, declining in fertility rate have occurred almost in the entire world, at the first across the high income countries, and then across the middle and low income countries. Generally, the experience of demographic changes both in developed and developing countries can be characterized by the decline in mortality and followed by fertility rate. The decline in fertility rate was followed by increase in percentage change ofthe working-age population. These conditions will cause burden the working-age population to young people became smaller. Bloom, et al. (1999); Birdsall, et al. (2001); McNicoll (2006); Fent, et al. (2008) explain that this condition will provides beneficial to economic growth. On the other hand, the low fertility led to investment in education for the children has increased and therefore, higher per capita output (Prettner and Prskawetz, 2009). Schultz (2001) states that economic explanation for the fertility transition focus on the role of returns to schooling, especially for female, which have encouraged female to obtain more education and facilitated the rise in female’s wages relative to male’s. Moreover, well-educated female causes a decrease in the infant mortality rate which an indicator of good health for the nation (Narayan & Smyth, 2006). In line with the miracle of economic growth in East Asia (1965-1990) as a caused of demographic changes (Bloom and Williamson, 1998), Indonesia has experienced demographic changes also, especially in the decade 1970-2010. The demographic changes during the last four decades have changed the Indonesian population age structure significantly. One of the phenomena which appear on demographic change is a decreasing the total fertility rate (TFR). Generally, countries which experience declining fertility rate will be followed by economic benefit such as abundant supply of labor particularly for female labor supply supporting for economic development. The decreasing fertility rate causes the women raise their participation in the labor force market, in turn, enhances their social status and personal independence. They tend to have more energy and time to contribute both to their families and to the society. Family income can be focused more upon better food or nutrition for infants. Further, effect of nutrition is important to determine the level of attendance of children at school, and improve the quality of education. The quality of education influences the human capital andlabor productivity in the future. So, human capital will affect the labor productivity. Finally, human capital and labor productivity influence economic growth. Therefore, the objective of paper is to examine relationship causality between the declining TFR and female labor participation (FLP) in Indonesia from 1990 to 2012. To achieve the objectives of this paper, three hypotheses have been made,the first hypothesis is (H0.1:FLS not affects TFR and H1.1: FLS affects TFR); the second hypothesis is (H0.2: TFR not affects FLP and H1.2: TFR affects FLP); and the third hypothesis is (H0.3: there is feedback or bilateral causality between FLP and TFR simultaneously and H1.3: there is no bilateral causality or independence between FLP and TFR simultaneously). This study contributes about the empirical evidence in Indonesia concern the demographic changes (i.e. declining in fertility) and female labor participation particularly for developing countries. Meanwhile, based on the author knowledge, the topics of causality relationship between TFR and FLP by age cohort are not widely studied. The findings in this paper can be used as a guideline for the policy makers as well as the various stake holders in Indonesia especially with National Development Planning Agency.

PREVIOUS STUDY Based on Malthus theory (1798, 1976) predicted that a positive relation between population growth and income growth. The hypothesis based on that people marry earlier and have more children when their incomes are greater. However, cross-country evidence over the last hundred years clearly contradicts this prediction. When the industrialized countries’ incomes increased, the fertility rate decreased. This

Prosiding Persidangan Kebangsaan Ekonomi Malaysia Ke-9 2014

305

condition is supported by the neoclassical economic theory stated that relationship negatively between fertility rate and income. Even, the recent experience of ‘the tigers of Asia’ and some of developing countries confirms this pattern. So, there are gap between existing theories and facts. For the last four decades the new models were developed to explain the observed negative association between fertility and income. More special, some of them introduced the distinction between the quality and the quantity of children (Becker, 1960; Willis, 1973). The others introduced women’s time allocation decisions and emphasized the opportunity costs of women’s time (Mincer 1963; Becker 1965; Willis 1973). Based on their argument that when income increases may reduce fertility if the income elasticity for the quality of children is sufficiently greater than that for the number of children. Various authors (Ahn and Mira 2002; Brewster and Rindfuss 2000) find that in OECD countries the cross-country correlation between the TFR and FLP turned from a negative value before the 1980s to a positive value thereafter. Bloom, et al. (2009) use a panel of 97 countries over the period 1960–2000 to examine the effect of fertility rate (TFR) on female labor force participation (FLP) during their fertile years. They argue that the effects of TFR reduction on FLP are large, and this effect can contribute significantly to a take off in economic growth during the demographic transition when TFR are falling. The effect of TFR on FLP is negative and statistically significant for all age groups between 20 and 44 years. TFR reduction can increase FLP, as well as investments in each child’s health and education. Mishra, Nielsen & Smyth (2010) using a panel unit root, panel cointegration and Granger causality approach, has found that in the long run Granger causality runs from TFR to FLP in the G7 countries and that there is an inverse relationship between FLP and TFR. The findings of the research that a 1%increases in the TFR rate results in a 0.4% decrease in the FLP for the G7 countries.The studies that have explicitly examined the issue of causation between FLP and TFR. Zimmermann (1985), using German time series data from 1960 to 1979; Cheng (1996b), using USA time series data from 1948 to 1993, and Cheng et al. (1997), employing Japanese time series data from 1950 to 1993, Chevalier, A. &Viitanen, T.K. (2002) within a bivariate Granger causality framework found evidence consistent with short run unidirectional Granger causality running from TFR to FLP. Cheng (1996a) examined causation between fertility and FLP for African-American females. He found unidirectional Granger causality running from FLP to TFR. Engelhardt, et al. (2004) examined the long-run relationship between FLP and TFR within a cointegration and Granger causality framework for France, Italy, Sweden, West Germany and the United Kingdom for 1960–1994. These authors found long run bi-directional Granger causality for all countries except Sweden. Narayan and Smyth (2006) examined the relationship between TFR, FLP and infant mortality rates in Australia over the period 1960–2000 and found that in the long run both the fertility rate and infant mortality rate Granger cause FLP. Hossain & Tisdell (2003) use the multivariate models in Bangladesh (19974-2000) confirm only a unidirectional causality – from FLP to TFR. Meanwhile, study by Dayıoğlu and Kırdar (2010) associated with the determinants of FLP in Turkey find that female with children have lower FLP, particularly in urban areas. The lower fertility rate of younger cohorts of female and the negative correlation between female with children and labor participation imply a higher participation rate for younger female.But otherwise, there are the findings of a study of Cheng (1999) that in fact he did not find causal relation between fertility and FLP in Taiwan. The study finds that education applies a great influence on FLP but not on fertility. This is indicates that working women do not certainly have fewer children, and having small children at home does not always discourage maternal employment in Taiwan. Meanwhile, in developing countries this negative relationship was found only on the job in the modern or formal sector in urban areas. While at work in the informal sector in urban and rural areas, the fertility of women who do not work are not different from those who working. Even in some developing countries women working in the agricultural sector in rural areas have more children than women-working (Todaro, 2006). The investigation by Naqvi and Shahnaz (2002) find that the female economic participation in Pakistan influenced by factors such as age, education, and lower fertility significantly. The employment status of the head of household, presence of male member, and children of ages 0-5 are also important variables that significantly affect female’s labor participation. The study of Gündüzand Smits (2006) in Turkey and observation by Psacharopoulos & Tzannatos (1993) in Latin America also finds coherence. These finding of the studies show that female participated in the formal economy are more educated, have husbands with higher occupations, and have fewer children.

306

Achmad Sjafii, Faridah Shahadan, Doris Padmini Selvaratnam

METHODOLOGY OF RESEARCH The data which used in this study is time series (1990-2012), which consist of a total fertility rate (TFR) and female labor supply (FLS). In this study, female labor participation (FLP) is a proxy of female labor supply.TFR obtained from BPS (Statistics Indonesia) and the Indonesian Demographic and Health Survey (IDHS), while the LFP sourced from BPS and ILO. TFR is the average number of live births per 1000 women of reproductive age, usually taken as 15-49 years, in a given year (ILO, 2007). FLP was the number of economically active women (15-64 years) divided by the total female population in the same age cohort (ILO, 2007). This means that they are the actual number of women (15-64years) who conduct economic activities in the labor market. The TFR is calculated as: TFR = (for 5-year age groups), where:ASFRi = agespecific fertility rate for women in age group i (expressed as i rate per woman), i1 (15-19); i2 (20-24); ... i7 (45-49). The formula for calculating the FLP = (LFf/Pf)*100%, where LFf is total female labor force, Pf is total eligible population (female) - both the economically inactive and active populations.FLP is divided into the following age cohorts (FLPi): FLP15+; FLP15-24; FLP15-64; FLP25-54; FLP25-34; FLP35-54; FLP55-64; and FLP65+. To examine the relationship of causality between TFR on FLP, this study uses the Unit Root test and Granger causality test, after thatit is combined by the method of time lags (lag) optimal. The first step is tested the data stationary (stationary stochastic process). This test should be done in the estimation of economic models with time series data. Test the data stationarity examine whether the time series data is stationary or not.Spurious regression can arise if the time series data are not stationary, even though the R2 high and t-statistic significantly. Stationarity test in this study is done by using the Augmented Dickey-Fuller (ADF) by comparing the value ADF statistic with MacKinnon critical value 1%, 5%, and 10%. The Granger Causality equation can be written as the following: ………… (1) ……….… (2) This study contributes about the empirical evidence in Indonesia concern the demographic changes. In general, Granger equation can be interpreted as follows (Gujarati, 2004): [a] Unidirectional causality from FLP to TFR, it means one-way causality from the FLP to TFR occurs if the coefficient lag FLP in equation (2) is statistically significantly different from zero, the coefficient lag TFR in equation (1) is equal to zero; [b] Unidirectional causality from TFR to FLP exists if the set of lagged FLP coefficients is not statistically different from zero and the set of the lagged TFR coefficients is statistically different from zero; [c] Feedback, or bilateral causality, is suggested when the sets of FLP and TFR coefficients are statistically significantly different from zero in both regressions; and [d] Independence, is suggested when the sets of FLP and TFR coefficients are not statistically significant in both the regressions.

THE EMPIRICAL RESULTS AND ANALYSIS The period of decline TFR in Indonesia started on the decade 1970s. Based on data of population census and IDHS show that even though TFR has continued to decline, the rate of decline has slowed down since 1994 when TFR was around 2.9 to 2.6 per woman as found in the 1994-2012 IDHS. It shows that TFR was stabilized at the level of 2.6 children per woman. The graphic shows that the TFR of Indonesia did not decline during the period among 2002-2003 and 2012 IDH surveys. On the other hand, FLP has also increased sharply in the 1980s until the early 1990s. During this period, FLP increase fantastically of 32.6 percent to 51.0 percent in 1991. After this period, the LFP has up and down between 51.0 to 53.2 per cent until 2010 (Figure 1). The decline in the fertility rate in Indonesia has not been separated from the role of family planning programs. The program was launched in three stages. It was starting in 1971 in Java and Bali islands only, and then expanding to include Sumatra in 1975, while the rest of the islands started the programmes in 1979. Since the early 1970’s has conducted strong advocacy programs emphasising small family sizes and legitimating contraceptive use. The use of modern contraceptives among Javanese married women increased rapidly from nearly zero before the 1960’s, to 26 percent in 1976 (Adioetomo, 2006).

Prosiding Persidangan Kebangsaan Ekonomi Malaysia Ke-9 2014

307

FLP of Indonesia by age cohort which are highest age cohort (25-54) and (35-54) years, both of participation rate are (58.7 and 61.6 percent) respectively. These two age cohorts are the prime working age. Meanwhile the lower FLP 15-24 age cohort reached 39.1 to 44.1 percent in the same period (Figure 2). This condition can be understood considering the majority of this age cohort is still in school (not including economically active population). While the lowest participation rate occurred in the age cohort of 65+ which is equal 23.0 percent due to this age cohort too old to work. To determine the causal relationship between TFR and FLP, the first step is test the stationarity of data to look at data stationary or not by using unit root test, or also called the ADF test. First, the ADF test conducted on the level. Based on the unit root test results are shown in Table 1, that variable FLP and TFR in the age cohort 15+; 15-24; 15-64; 25-54; 25-34; 35-54; and 55-64 is not stationary at the level of significance (the value of the ADF test statistic is smaller than the critical value Mackinon at all levels of significance, either 1%, 5%, or 10%), so the variable FLP and TFR in all age cohorts contain a unit root. Whereas the variable FLP age cohort 65+ have been stationary at the level, but the TFR is not stationary in the level. Therefore, so that the data of FLP to be tested further at the level of first difference. Based on the unit root test results are shown in Table 1. It needs to be note that TFR and FLP for all age cohorts have been stationary at the first difference level with the ADF test statistic greater than the critical value Mackinon at significance level of 1%, 5%, and 10%. The next step is determining optimal lag. Determination of the optimal number of lag in this paper using information criteria VAR estimated with the method recommended by Aike Information Criterion (AIC), Final Prediction Error (FPE), Hannan Quinn (HQ) and Schwarz Information Criterion (SIC). Determination of the optimal lag length can be known by the number of asterisks (*) in Table VAR Lag Order Selection Criteria. The result of optimal lag length is lag 1, which means that the variables TFR and FLP in all age cohorts have a better causality at lag length 1. Lag length 1 shows the independent variables one year ago influenced the dependent variable in current year, or the independent variable on current year affects the dependent variable the next year. Based on the first hypothesis (H0.1: FLP not affects TFR and H1.1: FLP affects TFR), after Granger causality test showed that H0.1 is rejected if the probability value on the results of Granger causality test is less than the critical value (1%, 5%, or 10%) and the mean FLS affect (cause) changes in TFR, while H0.1 is not rejected when the probability value on the results of Granger causality test is greater than the critical value which indicates that FLP does not affect the (not cause) of TFR. The second hypothesis (H0.2: TFR not affects FLP and H1.2: TFR affects FLP), H0.2 is rejected if the probability value on granger causality test result is less than the critical value (1%, 5%, or 10%) and H0.2 is not rejected when the probability value on granger causality test results are greater than the critical value. This hypothesis is applies to all age cohorts (i.e. 15+; 15-24; 15-64; 25-54; 25-34; 35-54; 5564; and 65+). Based on Table 3, result of granger causality test for the first hypothesis on the age cohort 15+ and 15-64 showed that FLS affect (cause) TFR, because the value of the probability of the age cohort 15+ and 15-64 for the first hypothesis is 0.03668 and 0.03958 (less than 0.10 and 0.05), then H0.1 is rejected and H1.1 is not rejected, which means increased FLS in the age cohort 15-64 year and 15+ year led to a decrease in TFR. While the probability value of the age cohort15-24 year (0.13966); 25-54 year (0.24726); 25-34 year (0.31984); 35-54 year (0.33850); 55-64 year (0.31309); and 65+ (0.45392) are greater than 0.01, 0.05, or 0.10, then H0.1 is not rejected and H1.1 is rejected. This means that the increasing FLP in the age cohorts did not affect (not causes) decreasing TFR. Based on the second hypothesis (H0.2: TFR not affects FLP and H1.2: TFR affects FLP), after Granger causality test, shows that the TFR for all age cohorts does not affect (not cause) FLP.The probability for the age cohort 15+ year (0.32245), 15-24 year (0.96706), 15-64 year (0.34808), 25-54 year (0.72585), 25-34 year (0.56795), 35-54 year (0.13316), 55-64 year (0.45918) and 65+ year (0.16000) are greater than 0.01, 0.05, or 0.10, then H0.2 is not rejected and H1.2 is rejected which means declining TFR does not(not cause) affect increasing FLP for all age cohorts. The Granger causality test for the third hypothesis (H0.3: there is feedback or bilateral causality between FLP and TFR simultaneously and H1.3: there is no bilateral causality (independence) between FLP and TFR simultaneously) shows that there is no bilateral causality in all age groups between the FLP and TFR. In the other word, the causality between the variables FLP and TFR for all age cohorts are independence. It was shown by the lag coefficients FLS and TFR was not statistically significant different from zero simultaneously on both equations (1) and (2) above. The finding of the paper emphasized by study by Barro (1991), Mincer (1963); Becker (1965); Willis (1973) that increases in the quantity of human capital tend to lead to higher rates of investment in human and physical capital, and hence, to higher per capita growth. A supporting force is that more human capital per person reduces fertility rates, because human capital is more productive in producing goods and additional human capital rather than more children. The brief said that when income

308

Achmad Sjafii, Faridah Shahadan, Doris Padmini Selvaratnam

increases may reduce fertility if the income elasticity for the quality of children is sufficiently greater than that for the number of children.

CONCLUSION Based on the overall results above, it can be concluded that there is a unidirectional causality between FLP and TFR in the age cohorts 15+ and 15-64 years. It means there is only one direction the relationship (unidirectional) of one variable to another variable that is the increasing FLP affects decreasing TFR. Unidirectional causality between FLP and TFR in this study in the age cohorts 15+ and 15-64 years showed increasing the FLP affect decreasing the TFR and not works in reverse. The age cohort 15+ and 15-64 years, showed indications that their role in supplying labor to the labor market is still very important. This condition would appear from the number of female labor supply, especially in rural areas in Indonesia. Neoclassical theory has been stated previously that seems more in line with the empirical conditions of employment in Indonesia, especially for female labor supply in the labor market.An increase in education due to the increase in human capital has an enormous influence on female's participation in the labor market. The female labor force participation has changed the thinking of women to better appreciate their time and pay more attention to quality than quantity of children. In the end, female laborparticipation which has been influenced the decision of the number of children (family size) in the household.

REFERENCE Adioetomo, S.M. (2006). Age-structural transitions and their implications: the case of Indonesia over a century, 1950-2050. CICRED, 129-157. Ahn, N. & Mira, P. (2002). A note on the changing relationship between fertilityand female employment rates in developed countries. Journal of Population Economics, 15, 667–682. Becker, G.S. (1960). An economic analysis of fertility, In: Demographic and Economic Change in Developed Countries. Universities-NBER, Conference Series No.11, Princeton UniversityPress, Princeton Becker, G.S. (1965). A Theory of The Allocation of Time. Economic Journal, 75, 493–517. Bloom, D.E., Canning, D., Fink, Günther, and Finlay, J.E.(2009). Fertility, Female Labor Force Participation, and the Demographic Dividend. Journal of Economic Growth 14 (2), 79–101. Cheng, B.S. (1996a).The causal relationship between African American fertility and female labor supply: policy implications. Rev Black Political Economic, 25, 77–88. Cheng, B.S. (1996b). An investigation of cointegration and causality between fertility and female labor forceparticipation. Appl Econ Letter, 3, 29–32. Cheng, Benjamin S. (1999). Cointegration and Causality between Fertility and Female Labor Participation in Taiwan: A Multivariate Approach. Atlantic Economic Journal, 27, 4. Chevalier, A. & Viitanen, T.K. 2002.The causality between female labour forceparticipation and the availability of childcare. Applied Economics Letters, 9, 915-918. Dayıoğlu, M. and Kırdar, M.G. (2010). Determinants of and Trends in Labor Force Participation of Women in Turkey. State Planning Organization of the Republic of Turkey and World Bank, Working Paper No.5 Engelhardt H. & Prskawetz, A. (2004).On the Changing Correlation betweenFertility and Female Employment over Space and Time.MPIDRWorking Paper, WP 2002-052. Gündüz, A.H. & Smits, J. (2006).Variation in labor market participation of married women in Turkey. NICE Working Paper. Hossain, M. & Tisdell , C. (2003). Fertility and Female Work Force Participation in Bangladesh: Causality and Cointegration, Social Economics, Policy and Development (Working Paper). Malthus, T.R. (1798). An Essay on the Principle of Population, 1st. Ed. (reprint 1976) Norton, NY. Mishra, V.& Smyth, R. (2010).Female labor force participation and total fertility rates inthe OECD: New evidence from panel cointegration and Granger causality testing. Journal of Economics and Business, 62, 48-64. Naqvi, Z.F. and Shahnaz, L. (2002). How Do Women Decide to Work in Pakistan? The Pakistan Development Review,41 (4), 495–513.

Prosiding Persidangan Kebangsaan Ekonomi Malaysia Ke-9 2014

309

Narayana, P.K. & Smyth, R. (2006). Female labor force participation, fertility and infant mortality in Australia: some empirical evidence from Granger causality tests. Applied Economics, 38, 563572. Todaro, M.P. & Smith, S.C. (2006). Economic Development. 9th ed. Pearson Education Ltd, UK. Willis R.J. (1973).a New Approach to the Economic Theory of Fertility Behavior. Journal of Political Economy, 81, 2, S14–S64. Zimmermann, K.F. (1985). Family Economy, Springer, Berlin.

APPENDIX

Sources: Census, IDHS, Various Years. FIGURE 1: Trend in TFR and FLP, Indonesia, 1971-2010

Sources: ILO, Key Indicators of the Labor Market. FIGURE 2: Trend in FLP by Age Cohort, 1990-2012

310

Achmad Sjafii, Faridah Shahadan, Doris Padmini Selvaratnam

TABLE1: ADF Unit Root Test ADF unit root test Age Cohort 15+ 15-24 15-64 25-54 25-34 35-54 55-64 65+

FLP Level -2.827671 -1.825797 -2.767615 -1.751695 -1.598077 -2.220367 -2.316421 -3.875427**

Differences -4.035261** -5.022979* -4.013953** -4.446187** -4.342896** -3.832961** -5.056028* -3.277492***

Level -3.090612 -3.090612 -3.090612 -3.090612 -3.090612 -3.090612 -3.090612 -3.090612

TFR Differences -5.651155* -5.651155* -5.651155* -5.651155* -5.651155* -5.651155* -5.651155* -5.651155*

∗, ∗∗, ∗∗∗ Statistical significance at the 10%, 5% and 1% levels respectively

TABLE 2:Lag Length Criteria Age Cohort 15+ 15-24 15-64 25-54 25-34 35-54 55-64 65+

Lag 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1

LogL -21.95069 -2.538270 -31.93058 -16.03728 -21.95069 -2.538270 -29.63242 -8.452489 -38.24672 -16.20650 -25.67086 -4.201321 -32.46500 -13.70669 -21.62183 -0.890064

LR NA 3.53055* NA 7.45206* NA 3.53055* NA 36.58352* NA 38.06947* NA 37.08375* NA 32.40072* NA 35.80941*

FPE 0.030249 0.007475* 0.074943 0.025503* 0.030249 0.007475* 0.060813 0.012798* 0.133076 0.025899* 0.042422 0.008696* 0.078674 0.020634* 0.029358 0.006435*

AIC 2.177336 0.776206* 3.084598 2.003389* 2.177336 0.776206* 2.875675 1.313863* 3.658793 2.018773* 2.515533 0.927393* 3.133182 1.791517* 2.147439 0.626369*

SC 2.276521 1.073763* 3.183784 2.300947* 2.276521 1.073763* 2.974861 1.611420* 3.757978 2.316330* 2.614718 1.224950* 3.232367 2.089074* 2.246624 0.923927*

HQ 2.200701 0.846302* 3.107964 2.073485* 2.200701 0.846302* 2.899040 1.383958* 3.682158 2.088868* 2.538898 0.997488* 3.156547 1.861613* 2.170804 0.696465*

* indicates lag order selected by the criterion LR : sequential modified LR test statistic (each test at 5% level) FPE : Final prediction error AIC : Akaike information criterion SC : Schwarz information criterion HQ : Hannan-Quinn information criterion

TABLE 3: The Result ofGranger Causality test Age Cohort 15+ 15-24 15-64 25-54 25-34 35-54 55-64 65+

H0 : No Causality FLP does not Granger Cause TFR TFR does not Granger Cause FLP FLP does not Granger Cause TFR TFR does not Granger Cause FLP FLP does not Granger Cause TFR TFR does not Granger Cause FLP FLP does not Granger Cause TFR TFR does not Granger Cause FLP FLP does not Granger Cause TFR TFR does not Granger Cause FLP FLP does not Granger Cause TFR TFR does not Granger Cause FLP FLP does not Granger Cause TFR TFR does not Granger Cause FLP FLP does not Granger Cause TFR TFR does not Granger Cause FLP

X2 5.05118 1.03203 2.37661 0.00175 4.88370 0.92565 1.42509 0.12666 1.04350 0.33777 0.96407 2.46166 1.07383 0.57086 0.58456 2.13838

Probability 0.03668 0.32245 0.13966 0.96706 0.03958 0.34808 0.24726 0.72585 0.31984 0.56795 0.33850 0.13316 0.31309 0.45918 0.45392 0.16000