Household Participations and Sustainable Development Programs: Social Impact of Government Assistance in Indonesia

A comprehensive policy impact evaluation study was conducted by Hidrobo et al. [4] in Latin America. Researchers also specifically examined the impact of providing various public transfers in the form of cash, food vouchers and food aid on three important indicators of social capital, namely mutual trust, discrimination and community participation. The results showed that public transfers were negatively correlated with discrimination and positively related to community participation. Other studies that discuss social capital in Indonesia and the calculation of the social capital index are summarized in Table 1.

Empirical studies on the social impact of transfer grants are also starting to be carried out in several developing Asian countries, including Indonesia. Taking a slightly different point of view, Cameron and Shah [5] examined the social effects of cash assistance if leakage was found in the distribution of aid. More specifically, the researchers examined the linkage of Cash Transfer Assistance (CTA) leaks in Indonesia with the crime rate. Using Indonesia Household Social Economic Survey 2006 data, the researchers showed that targeting errors in CTA, especially leakage, were positively associated with increased crime experienced by households. Researchers also confirmed that the existence of CTA leakage was negatively correlated with household members' participation in community activities.

The importance of social capital as a determinant of the economic development of a society has long been discussed by several experts. Conceptually, social capital is often associated with social values prevailing in society, mutual trust and the participation of individuals or households in activities or social organizations. Ponthieux [6] critically reviews the concept of social capital starting from Coleman's social structure, participation in organizations according to Putnam to the concept of trust by Fukuyama.

In empirical studies, experts have also introduced the concept of calculation related to social capital from the household to the community level [7]. Several other researchers such as Hjollund and Svendsen [8] specifically introduced a calculation method to create a kind of social capital index using a factor analysis approach. In this case the researcher limits the concept of social capital to the dimensions of trust, cooperation and social networks. In general, the dimensions of social capital are summarized in two major indicators, namely trust indicators and voluntary organization indicators. In the aspect of group collaboration, the researcher uses Putnam's concept of social capital which is formulated as the number of memberships in social organizations.

Table 1. Summary of some previous research

The relationship between social capital and aggregate economic growth is analyzed by García et al. [9]. Using time series data from 1970 to 2001 for 23 OECD countries, researchers found a positive and significant relationship between social capital and economic growth. Social capital accounts for 7 to 10 percent of the economic growth of the sample countries. This finding further emphasizes the importance of the contribution of social capital to economic growth in addition to other explanatory factors.

Basically, research on social capital in Indonesia has also been carried out since the early decades of 2000. Miguel et al. [10] used SUPAS, PODES and SUSENAS data to examine the impact of industrialization on social capital in Indonesia with the study period between 1985 and 1997. The results of the study found that industrialized districts had relatively high indicators of social capital. Meanwhile, districts located near industrialized areas experienced a decline in observed social capital indicators, such as the number of out-migration and relatively little community participation in community gathering activities.

Other research related to social capital in Indonesia was conducted by Grootaert [11] using data from a Local Level Institution survey in three provinces: Jambi, Central Java and East Nusa Tenggara. The data used consists of multilevel levels starting from the household, community to district levels. Several indicators of social capital used are membership density, heterogeneity index, meeting attendance and decision-making index. The results found that households with higher social capital enjoyed higher expenditure, had more assets, had more savings and had higher access to credit. Researchers also found a mechanism for the positive effects of social capital on welfare indicators through three channels, namely (1) sharing information among activity group members, (2) reducing opportunistic behavior and (3) improving group decision-making processes.

So far, studies on social capital are still concerned with the issue of measuring and calculating the impact of social capital on welfare indicators and poverty indicators. The results of the study largely reinforce the hypothesis that social capital with the various indicators used is proven to have a positive impact on household welfare and the economy more generally. In line with the increasing number of government programs in various developing countries, studies on the indirect effects of public transfers have been carried out.

The evaluation of the impact of public transfers in the form of cash transfers has been widely studied, including the effects of mistargeting of beneficiaries. Stoffler [12] conducted a simulation to test the effect of cash transfers on farmer household consumption and production using Taiwanese household data. The results of several simulations show that transfers have a positive impact on increasing consumption and production. The positive effects of transfers were also felt by non-recipient (non-target) households and recipient households even though they were not poor (mistarget).

Regarding target errors, Weiss [13] identified that this phenomenon is commonplace in developing countries, including countries that have implemented poverty alleviation programs for a long time, such as India and the People's Republic of China (PRC). The existence of leakage and undercover is also found in countries that are relatively new to implementing public transfer programs such as Indonesia and Thailand. In general, it is concluded that despite the leakage, poverty alleviation programs have had a positive impact.

In a broader scope, Coady et al. [14] conducted a review of the targeting program in several developing countries. Several targeting methods are discussed in detail and indexed to measure targeting performance in several countries (including Indonesia) [15]. Evaluation is carried out not only for the transfer program but also includes subsidy programs and job creation programs.

The evaluation of the impact of the Direct Cash Assistance (CTA) program which was implemented in 2005 to 2006 was comprehensively carried out by Bazzi et al. [16]. Although CTA was explicitly provided to anticipate the impact of rising world oil prices, several other welfare indicators such as education, health and employment were also evaluated. The provision of public transfers in the form of cash to some extent has a positive effect on the welfare indicators tested.

The impact of public transfers is not only evaluated on household welfare indicators such as consumption expenditure, health and education. Several researchers have also examined the indirect impact of providing public transfers on social capital. Attanasio et al. [17] with an experimental approach found evidence that the level of community cooperation who received the program was relatively higher than those who did not get the program. Meanwhile, Ressler [18] with a qualitative approach found evidence that public transfers actually strengthened existing social networks. Researchers found this evidence after conducting interviews with several urban and rural households in Kenya.

The indirect effect of public transfer policies on non-recipient households but in the program, areas was reported to be positive. Angelucci and De Giorgi [19] found that public transfers increased the consumption of non-recipient households by 10 percent. This increase occurred through increasing loans, private transfers between families and families and through a decrease in savings.

Previous studies so far have found evidence that poverty alleviation programs in the form of cash transfers have a positive effect on recipient households. This positive influence can also be felt by non-recipient households. The positive effect of the public transfer program is not only in the form of increasing the level of welfare of recipient households, but also in the form of strengthening social capital, which is shown by the increasing participation of the community in formal and informal activities.

One of the problems encountered in the public transfer program is the existence of target errors in the form of leaks and undercover. The target error will certainly cause the effectiveness of achieving the program to be not optimal. How does the target error affect social capital in the community? Research on this topic is relatively rare. One of the studies using household micro data in Indonesia was conducted by Cameron and Shah [20]. By using SUSENAS and IFLS data, the researcher stated that the target error, especially the leakage, would result in an increase in crime. Furthermore, the researchers also found a negative relationship between leakage and undercover with community participation.

The main problem with the estimation method used by Cameron and Shah [20] is that there is no control over the unobserved nature of the household but it has an important effect on the relationship between social capital and public transfers. This problem can be resolved if a household data panel is available that contains data on household behavior related to community participation and the status of receiving public transfers [21].

4.1 Result

The data used in this study were the publication of the survey of IFLS wave 3 (2000) and IFLS wave 4 (year 2007). IFLS-4 has specifically provided a questionnaire regarding government programs including Direct Cash Assistance (CTA). As many as 12,979 households were surveyed, almost a quarter (22.35 percent to be exact) stated that they received CTA. By using the per capita monthly expenditure limit of Rp. 175.000, - for the category of poor households, then about 7 percent of the households surveyed were categorized as poor households. Of the 2,901 households that received CTA, 2,436 were not included as poor households (around 18.76 percent of total households). Meanwhile, there were around 449 poor households that had not received CTA (around 3.4 percent of total households). This shows that cases of leakages are more dominant than cases that have not received assistance (under coverages).

Some household data related to this study can be seen in Table 2. The data used basically consists of two units of analysis, namely at the household level and the community (village) level. At the household level, it consists of data regarding the status of obtaining CTA (yes or no), income level, age of the head of the household, gender of the head of the household and marital status (married or not). The data regarding household participation in social activities is calculated based on the number of household members involved in community participation activities such as ROSCAS, community meetings, cooperative activities, community service and activities in the context of village improvement.

At the village level, the character of the village is represented by variables that indicate village infrastructure such as the presence of terminals, articles, telecommunications stalls and post offices. One of the important explanatory variables in this study are leakages and under coverages. Leakage is calculated based on the share of non-poor households that received CTA. Meanwhile under coverages were the portion of poor households that did not receive CTA. These two main indicators are calculated at the village level.

According to the methodology used, household panel data is required to isolate the effects of unobservable factors. For this reason, the third wave of household data (IFLS-3) and fourth (IFLS-4) is used to compile the household panel data. Data on household participation is available for both IFLS waves so it is very suitable for the analysis needs. To control the determinants of community participation at the community level, Fixed Effect techniques will be used at the appropriate community level. In this case the Fixed Effect will be carried out at the village level (Enumeration Area). Given that leakage and under coverages variables are measured at the village level, the use of the village level Fixed Effect (FE) will cause these two important variables to be removed from the model (omitted). For this reason, besides at the village level, the FE approach will also be applied at higher community levels such as the district / city (district) level.

Table 2. Statistical description of some variables

Table 3 presents the results of estimating the impact of CTA on cooperative activities using different control variables. Column (1) and column (2) display the OLS estimation results. The first-row column (1) shows that there is a negative correlation between BLT households and cooperative activities. This negative correlation disappears after controlling for influencing factors at the village level (note column 2). Column (3) presents the FD estimation results for the same case. As a result, CTA households were positively correlated with cooperative activities. After controlling for the influence of other factors at the village level, this positive effect becomes stronger (more significant with a higher coefficient).

The estimation results indicate that apart from being influenced by the character of the household itself, cooperative activities are also very much determined by the facilities available in each area (village). CTA households were more active in participating in village cooperative activities. This shows the positive impact of the public transfer program on community participation activities.

Furthermore, testing the social impact of CTA was also carried out for other types of community participation activities, namely community service. The first row shows that the cash transfer program has no significant impact on community service activities. Factors that consistently influence household activities following cooperative activities are the age of the head of the household, the gender of the head of the household, the number of household members and the level of income. More household members will participate in voluntary work if the household head is older. Households with fewer female heads of household participate in community service activities. Community service activities increase if there are more household members. Finally, household income is also an important factor in determining household activity in community service activities.

Similar results can be found in the estimation of the social impact of CTA on community participation in the form of village improvement program activities. There was no indication that CTA had a significant impact on these activities. Several factors that have a significant influence on village improvement activities are the age and sex of the head of the household and the number of household members concerned. The estimation results are relatively consistent using both the OLS and FD approaches and the use of control over other factors at the village level which are time-invariant in nature.

To further examine the link between public transfers and community participation, the same model estimation was carried out but in a slightly different setting. The public transfer program is not only viewed in terms of the status of CTA households and not CTA households, but from the leakages and under coverages of this program. As already mentioned, leakage is calculated based on the ratio of non-poor CTA recipient households to all households in a village. While under coverages are calculated based on the ratio of poor non-CTA households to all households in the village.

Table 3. Impact of CTA on community participation (cooperatives)

Note: Figures in parentheses are a robust standard error. (***), (**), (*) indicate the level of significance of 1 percent, 5 percent and 10 percent, respectively.

The existence of leakage and under coverages is an indication of mistargeting in the distribution of CTA. The OLS test results show a negative relationship between CTA leakage and cooperative activities. However, this result tends to be biased because it has not controlled for unobserved factors at the household level. After controlling for this factor, evidence can be found that the leakage of CTA has actually caused a large number of household members to join the village activities. Not only that, the phenomenon of under coverages also resulted in the same conclusion, namely that there was a positive relationship between village activities and the number of poor households that had not been reached by CTA.

This finding is actually interesting to note, considering that the impact of providing public transfers in the form of cash will increase the involvement of household members in participating in village activities even though there is a leak in the distribution of CTA. Table 4 presents the estimated impact of the CTA leakage on other types of social activities, namely cooperative activities. In the absence of control for unobserved factors, a negative association was found between BLT leakage and cooperative participation. After controlling for all relevant factors, evidence is found that there is no significant impact of CTA leakage on cooperative activities. With leaks and under coverages, CTA recipient households are still relatively more actively involved in cooperative activities. In essence, cooperative activities are not disturbed by leaks in the distribution of CTA funds.

4.2 Discussion

As discussed in the previous section, in general this study supports the finding that there is a positive relationship between public transfers and beneficiary social activities. With an experimental approach, Attanasio et al. [17] found a significant relationship between public transfers and social activities. The results of Ressler's [18] interviews with the Kenyan community also found evidence that public transfers strengthen existing social networks.

The results of recent studies using household micro data also yield roughly the same general conclusions. In the case of several African countries, Babajanian [3] found that public transfers are positively related to individual behavior in social activities. The same finding was also reported by Hidrobo et al. [4] in the case of Latin American countries. Empirical evidence shows that there is a positive relationship between recipients of public transfers and community participation.

This research specifically examines the effect of public transfers in the form of Cash Transfer Assistance (CTA) on several community participation activities such as community meeting activities, cooperative activities and so on. The results showed that CTA recipient households were more active in participating in village meeting. The relationship between CTA and village meeting is a new finding, although previously there has been research on village meeting behavior in Indonesia using IFLS data. Unlike the approach used in previous research, this study uses a consistent approach to isolate all observed and unobservable determinants of village meeting. Vanadharajan's [23] research used IFLS-2 (1997) cross-section data, while Lasagni and Lollo [24] used IFLS-3 (2000) and IFLS-4 (2007). The two studies on village meeting in Indonesia did not anticipate the possibility of an unobserved factor effect that could potentially produce a biased estimator.

In theory, the determinants of village meeting can also come from unobserved factors such as social sanctions [25] and the nature of households that tend to be individualistic (selfish families) or tend to socialize with the surrounding community. The characteristics of the household can influence the decision to participate in village meeting and can also correlate with the status of the household in obtaining public transfers. To isolate the influence of these unobserved factors, this study uses a First Difference (FD) approach with a combination of Fixed Effects (FE) at the community level. The test results consistently show that CTA has a positive relationship with village activity.

Table 4. Impact of leakage CTA on participation in cooperative activities

Note: Figures in parentheses are a robust standard error. (***), (**), (*) indicate the level of significance of 1 percent, 5 percent and 10 percent, respectively.

The finding of a positive correlation between CTA and village meeting is interesting to be investigated further. Several explanations can be proposed to describe the relationship between CTA and village meeting. First, it is in accordance with Geertz's [26] finding that village meeting is carried out to strengthen harmony between community members. The results of a qualitative study conducted by Hosain et al. [27] found that giving CTA to some community members inevitably caused social jealousy which could disturb the harmony of the local community. However, the social friction was only temporary in nature and was more of a vertical conflict between residents who did not receive CTA and local officials. Social friction does not lead to a significant increase in crime as found by Cameron and Shah [5]. Improve harmony between members of this community, village meeting can be used as a means of strengthening social ties that have been disturbed. Empirical evidence shows that the intensity of the social gathering is still high even though there were cases of mistargeting in the provision of CTA.

Second, as hypothesized in Ambec and Treich's [25] social pressure model, village meeting can be used by individuals to anticipate community pressure from “social obligations” to share income. Households that received CTA used village meeting as a medium for sharing with fellow citizens without having to lose a certain amount of money but had to be willing to postpone a portion of their current consumption expenditure to get bigger results in the future.

This research uses IFLS data from 2007. Is the IFLS questionnaire still relevant to answer research questions about the relationship between government assistance and social activities? Unfortunately, IFLS data is available until 2014, so it cannot be used to answer this question. To test the relevance of the research results, alternative data were used from the results of the Rural Household Survey (RHS) which was released in 2017. This survey was conducted by the Center for Population and Policy Studies at Gadjah Mada University. Like the IFLS, this survey also collects detailed data on Indonesian households. However, the scope of data taken is relatively more limited.

Using the same analysis model, the results of research using different alternative data can be presented in Table 5 as follows. The dependent variable used is the involvement of household members in social activities in the village. While the main independent variable is the status of households receiving government assistance. In addition, characteristic variables are also used as control variables.

Table 5. Impact of government assistance on social activities (RHS data 2017)

Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1

The results of the study show that by using more recent data, empirical evidence shows that there is a significant positive relationship between government assistance and social activities. Model estimation shows consistent results with several variations of estimation strategies: OLS, Fixed Effect (FIXED) and Random Effect (RANDOM). Thus, it can be shown that the results of the IFLS data review are still relevant for use in more recent cases. The estimation results in Table 5 also show interesting results. First, the role of women in social activities was more than that of men. Married household members have relatively reduced social activities. The variables of age and employment status do not have a significant impact on social activities.