Natural disaster and fertility preference: evidence from the tsunami in Indonesia

ABSTRACT

Despite many discussions on the impacts of natural disasters on fertility outcomes, their implications for fertility preferences as a mechanism remain unexplored. This study proposes that natural disasters could promote fertility preferences by reinforcing traditional values. The empirical analysis relies on a post-tsunami longitudinal survey of individuals in the provinces of Aceh and North Sumatra in Indonesia frankenberg2020study. Out of the 28,376 survey respondents, our sample considers around 10,700 individuals aged 15 or more, whose responses are available for the variables used in the estimation. The results from individual-level cross-sectional regressions suggest that the ideal number of children increased with the extent of tsunami damage in that region. The increased fertility preference was more pronounced for people with greater religiosity and those who strengthened their family ties after the tsunami, supporting that the shift in fertility preference was intertwined with the rise of traditional values. For policymakers, our findings raise the need to consider the potentially lasting implications of shifts in cultural attitudes toward fertility beyond immediate post-disaster relief efforts. To address the limitations of our cross-sectional evidence, future research will need to investigate the dynamic relationship between fertility preferences and natural disasters using alternative data.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

¹ The empirical analysis is conducted using Stata 18.

² The questionnaire is as follows: “If you could choose exactly the number of children to have in your whole life, how many would that be?”

³ The average, median, and standard deviation of age are 34.3, 31, and 15.7, respectively.

⁴ All variables take the values recorded in the survey, without further data processing..

⁵ Age 45 is the third quartile in our sample. Choosing a different age threshold does not alter the results.

⁶ The original survey classifies marital status into five groups: never married, married, separated, divorced, and widowed. In Columns 3 and 4, “Never” indicates never married, and “Ever” includes the others.