Skip to main content
SpringerLink
Log in

Shaking up Foreign Finance: FDI in a Post-Disaster World

  • Research
  • Published:
Economics of Disasters and Climate Change Aims and scope Submit manuscript

Abstract

This paper studies the effects earthquakes have on inward foreign direct investment (FDI) within a country. I use a dynamic difference-in-difference model to estimate the impact of geophysical disaster exposure in 416 Indonesian districts. The effects are only temporary: FDI inflows plummet by 90% on average in the first year after an earthquake before recovering to pre-earthquake levels. The effect is largely driven by shocks through affected upstream industries within local supply chains, and centered within the manufacturing sector. This highlights the importance to also consider indirect earthquake effects through spatial and production networks, besides the direct effects on labor and capital.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Fig. 1
Fig. 2
Fig. 3
Fig. 4

Similar content being viewed by others

Availability of data, code and materials

The datasets generated during and/or analysed during the current study are not publicly available due to copyright reasons of some sub-data sets, but are available from the corresponding author on request.

Notes

  1. Yet flood, as the second most common category, is also assessed for robustness.

  2. Per UNDRR (2022a) definition direct refers to victims:"persons whose goods and/or individual or collective services have suffered serious damage, directly associated with the event" and indirect refers to affected: "people, distinct from victims, who suffer the impact of secondary effects of disasters for such reasons as deficiencies in public services, commerce, work"

  3. The population data is only available until 2014. All years after 2014 are imputed using the district-specific linear population trends

  4. In contrast to the scale, which is measured at the epicenter.

  5. Precisely, the ShakeMap values assigned are values for peak ground acceleration, another metric for velocity of the ground, which in a second step is transformed to the MMI based on values provided by Wald et al. (1999). This conversion is also done by Gignoux and Menéndez (2016)

  6. In relative terms this approach identifies 7% of districts as economic centers. This value reaches 20% in the highest illuminated districts. The value is lower on average due to many non-illuminated cells. As a quintile approach is used the spatial size of an economic center depends on the size of the district. To validate if the nighttime light indeed proxies economic centers, it is cross-referenced against geolocations of cities in Supplementary Information Appendix G2.

  7. The precise definition for district i and year t is

    $$ \text {Shock}_{it}=\underbrace{(\text {MMI}_{it}^{\text {urban}})}_{\text {Capital Shock}} \times \underbrace{(\frac{\text {Deaths}_{it}+\text {Victims}_{it}+\text {Affected}_{it}}{\text {Population}_{i,t-1}^{\text {imp}}})}_{\text {Labor Market Shock}}$$

  8. Indonesia is divided into provinces (level 1), regencies and cities (level 2), districts (level 3) and villages (level 4). So the second administrative level refers to 416 regencies (kabupaten) and 98 cities (kota), but for simplicity all second-level entities are called districts to align the name with the common perception of districts being the level below provinces.

  9. Based on victim records within the Desinventar database. This dummy equals one if more than 20 people are affected by any disaster type other than earthquakes or earthquake-induced tsunamis.

  10. The I-O table is sector-wise aggregated to match the corresponding FDI sectors and converted to the Leontief inverse matrix

  11. Conversion follows Bellemare and Wichman (2020) (equation 11) of arcsin-dummy conversion: \(exp(\beta )-1\)

  12. Two estimators suggest a small pre-trend. This trend is, however, not significant on any standard level.

References

Download references

Acknowledgements

I wish to thank Lisa Chauvet, Sandra Poncet, Jérémie Gignoux, Ariell Reshef, Myriam Ramzy, Rémi Bazillier, Clément Bosquet, Andrea Cinque, Fabio Ascione, Thibault Lemaire, as well as the participants of several seminars and conferences. I am particularly grateful to Lisa Chauvet for the support of this work and actively engaging in its development. I also thank the Centre d’économie de la Sorbonne (CES), Université Paris 1 Panthéon-Sorbonne for support during this research. The responsibility for all conclusions of this paper lie entirely with the author.

Funding

The author declares that no funds, grants, or other support were received during the preparation of this manuscript.

Author information

Authors and Affiliations

  1. Department of Economics, University Paris 1 Panthéon-Sorbonne, Paris, 75005, Île-de-France, France

    Robert Reinhardt

Authors
  1. Robert Reinhardt
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to Robert Reinhardt.

Ethics declarations

Conflict of interest

The author has no relevant financial or non-financial interests to disclose.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

Below is the link to the electronic supplementary material.

Supplementary file 1 (pdf 3241 KB)

Appendix

Appendix

Fig. 5
figure 5

Development of FDI inflows over time by sector The data is based on Indonesian Investment Co-ordinating Board (BKPM) (2021). Author’s own assignement

Full size image
Fig. 6
figure 6

Effects on District-level. The two panels show the effects of Eq. 1 if one extends the model towards the sectoral annual FDI space. Classification of FDI sectors follows broadly ISIC Rev. 3 level 2. The bars depict the 95% intervals and stars indicate *** 1%, ** 5% and * 10% significance. The left panel shows the effect over time, whereas the right one shows the effect solely in the year after the earthquake event (yet using the same regression as before). The treated group includes 38 treated districts and excludes as before treatment reversal and simultaneously occurring disaster events

Full size image
Fig. 7
figure 7

Comparison of different recent DiD estimators. All included estimates allow staggered treatment design, yet exclude the possibility of treatment reversal. Here, each individual regression excludes multiple treatments within unit and uses only a sample of districts, which did not change names over the period (omitting 5% of the sample), which does not affect estimates substantially. The code is inspired by Borusyak (2022) and presents estimators by Sun and Abraham (2021); Sant’Anna and Zhao (2020); De Chaisemartin and d’Haultfoeuille (2020); Borusyak et al. (2021); Wooldridge (2021). The chosen CI bandwidth is the 10% threshold

Full size image
Fig. 8
figure 8

Effects using panel matching. The panels show the effect of earthquake exposure onto FDI inflows, where control units show a similar treatment pattern and are matched. This method is based on Imai et al. (2023). The upper panel excludes districts with more than a single earthquake recorded, the lower excludes districts with only a single earthquake recorded. The three columns show three different matching algorithms for building a control group. The different columns show how many pre-treatment periods are used to build a matching control group. Matching covariates are based on the last three lags of nighttime-light, three lags of arcsin FDI inflow, current log of population and time-invariant composite hazard risk index. All most use matched control group of sample size 10 and standard errors are 1000 times bootstrapped. The confidence intervals shown correspond to a 10% threshold

Full size image
Table 3 Overview on Variables
Full size table
Table 4 Event-Study Type Difference-in-Differences
Full size table
Table 5 Dynamic Difference-in-Differences: Space
Full size table
Table 6 Descriptive Statistics on Shock variables
Full size table

Rights and permissions

Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Reinhardt, R. Shaking up Foreign Finance: FDI in a Post-Disaster World. EconDisCliCha (2024). https://doi.org/10.1007/s41885-024-00148-2

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1007/s41885-024-00148-2

Keywords

Search

Navigation