Compounding Repression and Electoral Integrity: A Dynamic Panel Study of 142 Countries, 2012–2024, with Policy Implications for Indonesia

Abstract

Background: Elections are increasingly retained while the conditions for verifying them are degraded, yet restriction of digital space and repression of civil-society organisations are almost always estimated separately, so whether they substitute or reinforce is unknown.

Objective: To estimate both instruments jointly and test the derived prediction that a negative interaction identifies two partially substitutable verification routes, whereas a positive interaction identifies a weakest-link chain.

Methods: A balanced panel of 142 electoral democracies and transitional regimes from 2012–2024 (N = 1,704 country-years), constructed from Varieties of Democracy version 14, the Perceptions of Electoral Integrity index, and World Development Indicators, was analysed using two-step system generalized method of moments with the Windmeijer correction.

Results: Digital space restriction reduced the Clean Elections Index by 0.221 points per unit at the mean of the moderator (widest interval attainable over the unreported covariance: 0.154 to 0.288; worst-case p < 0.001), equivalent to 0.058 points per SD; civil-society repression reduced it by 0.254 points per unit (p < 0.001). The interaction was negative (−0.094; 95% CI −0.155 to −0.033; p = 0.002): the damage from closing digital space rose from −0.186 where civil society was least constrained to −0.277 where it was most constrained under every admissible coefficient covariance.

Conclusion: The routes are substitutes: repression destroys redundancy, and the return to opening digital space is largest where civil society is weakest, inverting the usual sequencing intuition for Indonesian reform. Whether the outcome was measured annually or carried forward between elections could not be established, which conditions the dynamic results.