Do short-term unconditional cash transfers change behaviour and preferences? evidence from Indonesia

ABSTRACT

Short-term unconditional cash transfers are used as a temporary mitigation strategy during adverse economic shocks. They can however, cause adverse unintended impacts on behaviour and preferences. We estimate the effect of receiving short-term unconditional cash transfers on behaviour, risk aversion, and intertemporal choice in Indonesia. The country first introduced the program in 2005 and continues to use it whenever adverse economic shocks occur. With 15.5 million beneficiary households, the program remains one of the largest in the world. We use an individual-level longitudinal dataset spanning 1997 – 2014. To identify a causal relationship, we combine coarsened exact matching with difference-in-differences. We find no evidence that the short-term unconditional cash transfer affected beneficiaries’ behaviour or preferences. Together with evidence of its positive impact in mitigating the impact of adverse economic shocks, our findings show that short-term unconditional cash transfers should continue to be part of the government’s portfolio of social protection programs.

KEYWORDS:

• Unconditional cash transfer
• behavioural effects
• intertemporal choice
• risk aversion
• poverty
• Indonesia

JEL CLASSIFICATION:

• D10
• I31
• I38

Acknowledgments

A previous version of this paper was published in the SMERU Working Paper Series (Al-Izzati, Suryadarma & Suryahadi, 2020). We would like to thank Samuel Bazzi, seminar participants at SMERU, two reviewers, and the associate editor for feedback. All remaining errors are ours. The views expressed are the views of the authors and do not necessarily reflect the views or policies of Asian Development Bank Institute (ADBI), Asian Development Bank (ADB), its Board of Directors, or the governments they represent.

Disclosure statement

No potential conflict of interest was reported by the authors.

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/13600818.2023.2204423.

Notes

1 Gentilini (Gentilini, 2019, 2020) shows that 1,042 studies on social protection were published in 2019. Most focus on cash transfers. In 2020, 840 studies on cash transfers were published.

2 The use of the term intertemporal choice follows Handa et al. (2020) and Handa et al. (2014). Another term is time discounting. Following the differentiation between time discounting and time preference in Frederick et al. (2002), our focus is on time discounting. Jung et al. (2021), which use the same dataset we use, use the terms time preference, time discounting, and patience interchangeably.

3 Exclusion error is the proportion of households that were eligible but did not receive the program. Inclusion error is the proportion of households that were ineligible but received the program.

4 Surat Miskin or Surat Keterangan Tidak Mampu (SKTM) is a letter confirming poverty status issued by the village office.

5 Note that the non-recipients in this table cover both the poor who were supposed to receive the program but did not, and the non-poor who were not supposed to receive the program. On the other hand, the recipients were also comprised of the poor who were supposed to receive the program and non-poor who were not supposed to receive the program but did.

6 L1 is an indicator in CEM that shows the multivariate imbalance test result with value ranges from 0 (balance) to 1 (imbalance).

7 We modify Equation 1 and estimate: $y_{it}=\alpha +\beta UC{T}_{it}\times t_t+\theta t_t+{\delta }_i+{\mathrm{\partial }}_d\times t_t+{\epsilon }_{it}$. The variable ${\mathrm{\partial }}_d$ is a binary variable with the value of 1 for Dataset 2 and 0 for Dataset 1.

8 We also estimate the model using panel probit estimation for binary outcomes and report the marginal effects. The estimation results are shown in Table A11. The results are similar.

9 IFLS also collects information for formal insurance ownership such as private insurance (voluntary participation), employer-related insurance (mandatory participation), and social insurance (mandatory and subsidised by the government). However, formal insurance ownership is very small in Indonesia, especially among the poor. In IFLS, the ownership of private insurance is only 2% of the sample. Therefore, we rely on the informal insurance indicator since we want to focus on voluntary insurance.

10 Because of data limitation, we find only two indicators in IFLS that fit the definition of temptation goods: smoking (at the individual level) and alcohol consumption (at the household level). Since our focus is on individuals, we use smoking as the proxy for temptation goods.