There are many factors that may prevent individuals from achieving consumption smoothing, especially for poor individuals or households. Liquidity barriers for the poor can be overcome by pension programs financed by the government and intergenerational transfers. This research aims at investigating determinants of the family size decisions in poor households. This research uses primary data from survey questionnaires conducted among 447 poor households selected in Family Planning Villages throughout Sintang Regency. The family size decision is only affected by parents’ expectation of an intergenerational transfer. This variable has a significant positive effect, and the possibility of an old-age pension scheme provided by the government does not significantly affect the family size decision. This may be due to the fact that there has been no pension scheme for poor households in Indonesia. The cost of children does not influence the poor household’s family size decision. This is probably due to family package assistance provided by the government which can help reduce the cost of children. The cost of daily living, health and education costs have a significant positive effect on the burden of these costs. The findings of this study indicate that the family assistance package provided by the government does not encourage poor households to have a larger family, meaning that the provision of family assistance packages does not contradict the Family Planning Program. It also provides evidence that the government’s pension scheme for poor households can be also considered as a means of controlling the population growth, because this scheme will lower the expectation of poor households to receive intergenerational transfers.

intergenerational transfer, cost of children, consumption smoothing, pension scheme, poor household

One way to improve people’s welfare and reduce poverty at the macro level is to suppress population growth, for example through a family planning program (Ahlburg 1994). Poverty limits a household’s ability to pay for childcare as well as put aside for retirement, in fact neither an individual nor a household wants a drastic reduction in the consumption level due to a decline in the income at retirement. A certain consumption level is considered acceptable in the future, and households will try to maintain it. In other words, individuals or households hope to achieve consumption smoothing throughout their lives. Research in behavioral economics suggests that consumer preference is measured with regard to a reference point (Kőszegi and Rabin 2006; Siegmann 2002).

A method households can follow to prepare for consumption smoothing at retirement is saving while still productive (Deaton 2019, Bairoliya et al. 2021). However, many factors may prevent individuals from achieving consumption smoothing, especially for poor individuals or households. The consumption smoothing ability reflects an important dimension of the well-being because it marrows the capacity of people to meet their basic needs (Skoufias 2003; Notten 2006). Low-income households fail to succeed in consumption smoothing in retirement due to low welfare levels, which also results in low savings (Shaikh et al. 2018). In cases like this, a public pension system is needed to ensure retirement income levels that prevent the individuals or households from the risk of poverty at an old age (Shang 2014). There has not been much research on policies to mitigate consumption fluctuations for poor households, often the government only pays attention when there is a decline in consumption but there are no government policies that can encourage savings or insurance for poor households when they experience an increase in income (Ajefu 2017). In Indonesia, the pension scheme is only provided to workers in the formal sector. According to data from BKN (2023), in Indonesia the upper retirement age is 70 years. Meanwhile, for non-civil servants, the retirement age limit is 56 years (MediaBUMN 2023).

Liquidity barriers for the poor can be overcome by altruism and philanthropic endowments providing social support for poor households (Akerlof 2007) or programs financed by the government (Jappelli and Pistaferri 2010) and intergenerational transfer (Ioannides and Kan 1996). According to Mussa (2014), intergenerational transfers are expected to make poor households have more children as a source of support in retirement. A family size has an impact on children’s development and formation of human capital with their future economic status (Mussa 2014). Research in the developed and developing countries shows that children from larger families tend to be shorter, less intelligent, and have lower survival rates (Bielicki 1986). Furthermore, a big number of children will reduce the attention given to each child, affecting their academic ability. Therefore, children from poor households of a large size are closer to poverty (Birdsall and Griffifin 1988), with an intergenerational poverty transfer (Moav 2005).

The above statement underscores the assumption that there is a positive causal relationship between poverty and fertility at the national and household levels (Aassve et al. 2005). This implies that at the micro or household level, poor households have a large family size, encouraging them to increase fertility. Mussa (2014) shows that Africa has the highest fertility and poverty rates, but empirically there is no consensus on the nature of the relationship. Poor households that do not have pension funds, hope to be able to maintain their consumption level when they are no longer productive, funds only from their intergenerational transfers to maintain the reference point level of the consumption. Mixed empirical findings were received in Botswana (Chernichovsky1984), positive relationships were found in Sierra Leone and Ethiopia (Ketkar 1979), while negative relationships were identified in Burkina Faso (Langani 1997) as well as in South Sudan (Cohen and House 1994). Meanwhile, in Cameroon, Noumbissi and Sanderson (1998) show that fertility was very high, but the relationship was in the form of an inverted U-shape. This implies that low-income and high-income households have low fertility rates, while middle-income ones have higher fertility. The relationship between the two variables may move from fertility to poverty.

Indonesia is one of the low-income countries with the fourth largest population in the world. The World Bank projects that the poverty rate in Indonesia in 2020 will increase by 10.7% in a mild and 11.6% in severe scenarios. It is estimated that there will be 5.5-8 million new poor people since 2019 (Jayani 2020). A report from Worldometers noted that Indonesian population reached 274.86 million as of December 14^th, 2020. This number ranks Indonesia fourth in terms of the population size in the world (Annu 2020). To overcome complex problems related to poverty and high population growth, the government has carried out several strategies, such as launching a Family Planning Village program since 2016. The implementation aims to bring a family planning program closer to the community, especially in poor, densely populated urban, fishing villages, slums, and other disadvantaged areas (Suratman and Massardi 2020). The Indonesian government hopes that a successful Family Planning program can reduce poverty levels (Central Bureau… 2023).

On the other hand, the government also provides family assistance packages for poor households, with the aim to improve quality of their lives (Byaruhanga and Debesay 2021; Finn and Leibbrandt, 2017), including improving life quality for children from poor households, so that they have a higher opportunity to enter the labor market and will have an impact on increasing welfare and reducing poverty. According to Cabinet secretariat of the Republic of Indonesia, in 2021, the Indonesian government provided family assistance packages to 10 million beneficiaries (Cabinet Secretariat…, 2021). Da Silva et al. (2020) and Matos et al. (2021) found a positive relationship between poverty and family resilience due to exposure to risk of shocks.

Nationally, 5,590,724 people in Indonesia remain in extreme poverty. Meanwhile, the number of extreme poor people in West Kalimantan Province adds up to 73,342 people or 1.41 percent of the population (Central Bureau… 2023). The number of poor people in West Kalimantan in September, 2022, equaled to 6.81 percent, while at the national level it was 9.36 percent. Sintang’s poverty rate in 2022 is above the provincial level, yet below the national one, which is 8.57 percent (Central Bureau… 2023). In the same year, Sintang’s multidimensional poverty rate was the worst (22.27%) in the province of West Kalimantan (Aidha et.al 2020). Based on data from the National Population and Family Planning Agency of Indonesia (BKKBN 2022) Sintang Regency turns out to be a district with the highest number of Family Planning Villages – 44 spread all over its 14 sub-districts. This is accomplished because the desired lower fertility can be only achieved through limiting the number of births (Muhoza et al. 2014).

Education and health care access is expected to increase human resource quality, particularly for children from low-income households, and protect them from falling into poverty. (Cooray et al. 2011; Reza and Widodo 2013). Countries with high fertility rates, such as Indonesia, have “package” policies that facilitate childcare (Galloway and Hart 2015). In addition to the declaration of the Family Planning Village Program, the government also provides social assistance (public assistance) and welfare services to help the poor through Healthy Indonesia and Smart Indonesia Card (Maliangga et al. 2019). The assistance includes monetary benefits, goods, or social services aimed at helping or protecting the most vulnerable individuals, families, and communities to meet basic needs and improve their quality of life. Social assistance and government social investment are conducted through special protection policies for children (Muhtar 2017). Such aids or social package policies will certainly reduce the direct cost of childcare, and this can increase the demand for children in poor households. Family policies can directly affect childbearing behavior by increasing the budget size and decreasing the size of the direct cost of children. A lot of literature supports that the demand for children is a function of their direct costs and individuals’ preferences, for a given level of income (Letablier et al. 2009).

Data show that an increase in the provision of pensions for older people by the state correlates with a decline in fertility (Boldrin et al. 2015). The impact on changing fertility patterns and their relationship with government-provided pensions is a new topic that is quite interesting to be developed. The findings of Ehrlich and Kim (2005) show that a higher scale of the social security system negatively affects fertility.

Based on the above explanation, studying determinants of the family size desired by poor households is interesting. This is determined by the expectation of receiving intergenerational transfers from children in retirement or high direct cost of childcare. The more children, the higher the intergenerational transfer that can be expected, but on the other hand it also increases the cost of children before these children enter the labor market, ceteris paribus. And the limitations of poor households in financing their children’s education will cause their children to also be trapped in poverty because the academic abilities of children from poor households are usually low. Limiting the number of children as a condition for obtaining an old-age pension from the government may be an effective way to control population size and poverty.

Regardless of the community type, most couples of childbearing age balance the cost of having children and the expectations of the economic, social, and psychological benefits (Muhoza et al. 2014). Economic monetization raises concern about higher cost of food, clothing, health, and education associated with a higher number of children. Some socioeconomic factors change as socioeconomic development or family planning campaigns affect the demand for children. Compared to rural areas, urban environments offer more employment and education opportunities and are associated with cultural diversity and openness to new ideas (Garenne and Joseph 2002). Furthermore, this contributes to weakening traditional norms and values, such as the resource sharing system, which makes them more tolerant of small families (Gurmu and Mace 2008). Empirical literature shows a strong negative correlation between education level, preferences, and fertility behavior (Bongaarts 2002).

According to a Nigerian study, poverty can affect economic expectations during crisis, hence they have a lower fertility preference (Odusola 2002). This indicates that the negative impact of poverty on the desired family size occurs when people understand that their economic means cannot support the decision to have more children. There are many studies showing that the desired family size is negatively related to the financial status, where couples of childbearing age with a lower financial status have a higher number of children desired than women with a higher financial status (Hyatt and Milne 1993; Gwatkin et al. 2007). But in some Asian countries, women from poor households are reported to have lower demands than those from rich households (Becker 1991; Rahman et al. 2002).

The individual or household’s ability to respond to uncertainty and cope with income shocks reflects an important life dimension. Furthermore, welfare is enjoyed by individuals or households at a certain point in time. Households with a high risk of being poor are the same as those with low consumption smoothing capabilities. Meanwhile, households with savings for retirement have a lower risk of poverty and a higher ability to smooth consumption. High poverty risks are found in households with a high number of children and dependency burden (Mberu and Reed 2014; Canning et al. 2015; Dribe et al. 2017).

The relationship between demographic structure and household saving rates can be explained through two mechanisms, first, through reducing the number of children who are household dependents to have a higher ability to save. Second, through the composition effect, most of the savings are generated from the difference in income and consumption of the productive population (20 – 65 years old), who mostly receive income from work. Changes in the structure will increase savings in aggregate (Curtis et al. 2015). Changes in the population age structure will affect the aggregate saving level (Gauthier and Hatzius 1997). If a life-cycle saving is driven by the motivation to prepare for retirement, then a low population growth will reduce the aggregate saving (Modigliani and Brumberg 2003). However, when the cost of raising children determines the saving ability, a low population growth will increase the saving rates (Coale and Hoover 1958, Higgins and Williamson 1997).

Fertility behavior can be analyzed within the framework of a neoclassical economic model such as developed by Leibenstein in 1957. This model emphasizes the importance of intergenerational transfer as a form of old-age security and fertility motivation in developing countries. In 1960, Becker further developed this model by reconciling predictions from the demand theory in income to increase the demand for children. This paradox was answered by applying the theory of consumer behavior determined by variations in household income and the opportunity costs of childcare.

The Microeconomic Theory of Household Fertility explains that every family has costs and benefits in determining the desired family size. This theory adopts the conventional consumer behavior theory. Children are considered consumer goods (no profit). A demand for children is a rational economic choice. The desire to have children is influenced by income, the cost of living, and the desire to consume other goods, and the demand for children is positively related to income. This is negatively related to the relative cost of children and preferences for other goods (Todaro, 2000).

Becker (1960) introduced two key elements of household microeconomic theories: preferences and constraints. Hypotheses on the observed behavior are constructed from the comparative static analysis showing how demand varies as budget constraints shift as income. When this methodology is applied to fertility behavior, a clear conflict was noted between theoretical predictions and empirical evidence. As income increases, theory predicts that the quantity of most goods will increase, except for “inferior” goods, which tend to be of a lower class. Empirically, fertility declines as income increases greatly, and there appears to be a negative correlation between family income and the number of children.

Meanwhile, the demand for children has qualitative and quantitative dimensions. Hence, the total expenditure is equal to the number of children multiplied by the quality per child and by a price index reflecting the cost of inputs. In the quality – quantity model, spending on children tends to increase with income. However, this is due to the increased child quality, while fertility does not vary based on income variations.

This research uses primary data collected through survey questionnaires. The population is couples from poor households in Family Planning Villages in Sintang Regency, The Province of West Kalimantan. The total sample includes 447 poor households, and the sample was obtained from 15 Family Planning Villages in Sintang regency. Furthermore, a purposive random sampling was used, where the sample size was determined based on the number of residents of the Family Planning Villages selected. The research samples are presented in Table 1.

All data collected will be analyzed using a regression analysis. In this research, the dependent and independent variables are binary. The variables used are binary or dummy variables with values of 0 and 1. Therefore, the estimation equation used to test the relationship between these variables is as follows:

CC _i = α₀ + α₁SAB_i + α₂SAH_i + α₃SAE_i + ε_i,(1)

FSD_i = β₀ + β₁IT_i + β₂CC_i + β₃PG_i + e_i.(2)

Insert equation (1) in (2), therefore

FSD_i = β₀ + β₁IT_i + β₂(α₀ + α₁SAB_i + α₂SAH_i + α₃SAE_i + ε_i) + β₃PG_i + e_i =

= β₀ + β₁IT_i + β₂α₀ + β₂α₁SAB_i + β₂α₂SAH_i + β₂α₃SAE_i + β₂ε_i + β₃PG_i + e_i =

= (β₀ + α₀β₂) + β₁IT_i + β₂α₁SAB_i + β₂α₂SAH_i + β₂α₃SAE_i + β₃PG_i + (β₂ε_i + e_i) =

= γ₀ + γ₁SAB_i + γ₂SAH_i + γ₃SAE_i + β₁IT_i + β₃PG_i + θ_i.(3)

Where

FSD : Desired family size

IT : Intergenerational Transfer

PG : Pension Scheme provided by the Government

CC : Direct Cost of Childcare

SAB : Government assistance in basic needs packages

SAH : Government assistance in the health sector

SAE : Government Assistance in the field of education

e, θ, ε : Error term

The operational definitions of the variables in this research are presented in Table 2.

Every individual or household is expected to perform consumption smoothing upon retirement. For poor households without retirement savings or pension funds, the source of income can be in the form of intergenerational transfers or an old-age assistance scheme provided by the government. Furthermore, these two variables significantly influence decision-making regarding the desired family size. The expectation of an intergenerational transfer is directly proportional to the desired family size. There are high welfare expectations among poor households at the time of retirement when they have a large family size. Meanwhile, the government pension transfer variable has a negative effect on the desired number of children. The poor households in Family Planning Villages in Sintang Regency have not considered the government pension scheme variable in making family size decisions, probably because there is no pension scheme for the poor in Indonesia. This indicates that providing government pension transfers can be a strategy for controlling the population, but the provision is certainly not without conditions. This requirement to obtain assistance is used to control the population. For example, only households with two children will receive pension assistance from the government. However, data are taken from poor households with children, so this can be biased because family size planning is more appropriate for individuals or households that do not have children. Future research can be conducted on couples without children or individuals from poor households who are matured, yet not married.

Annu CM (2020) Indonesia is 4th rank in country with highest population density. URL: https://databoks.katadata.co.id/datapublish/2020/12/15/indonesia-peringkat-ke-4-negara-berpenduduk-terbanyak-dunia (in Indonesian)

Badan Kepegawaian Negara (2023) Batas usia pension PNS. URL: https://betv.disway.id/read/17293/terbaru-bkn-resmi-tetapkan-batas-usia-pensiun-pns-2023-bukan-56-tahun-lagi-berapa-itu (in Indonesian)

BKKBN (2022) Kampung KB di Indonesia. Badan Koordinasi Keluarga Berencana Nasional.

Cabinet Secretariat of Republic Indonesia. (2021) Gov’t Launches Cash Social Assistance Programs for 2021. URL: https://setkab.go.id/en/govt-launches-cash-social-assistance-programs-for-2021/ (in Indonesian)

Central Bureau of Statistics of Indonesia (2023) Indonesia in figures. URL: https://www.bps.go.id/en

Jayani DH (2020) World Bank Projection for The Indonesian poverty level until 2020. URL: https://databoks.katadata.co.id/datapublish/2020/07/21/proyeksi-bank-dunia-kemiskinan-indonesia-2020-bisa-sampai-11 (in Indonesian)

MediaBUMN (2023) Direksi Anak Perusahaan BUMN, Erick Thohir Hapus Syarat Batas Usia. URL: https://mediabumn.com/direksi-anak-perusahaan-bumn-erick-thohir-hapus-syarat-batas-usia (in Indonesian)

Eddy Suratman – Lecturer and Researcher at Universitas Tanjungpura, Faculty of Economics and Business. Pontianak, The Province of West Kalimantan, 78124, Indonesia. Email: eddy.suratman@ekonomi.untan.ac.id

Restiatun Restiatun – Corresponding author, Lecturer and Researcher at Universitas Tanjungpura, Faculty of Economics and Business. Pontianak, The Province of West Kalimantan, 78124, Indonesia. Email: restiatun@untan.ac.id