When Do Pension Reforms Happen? Evidence on Thresholds and Reform Sequencing

Abstract

This paper examines the determinants of pension reform occurrence using a newly

constructed dataset of 790 policy events across 183 countries over the period 2000 to

2024. By classifying each event by type, 112 major structural and parametric reforms

are distinguished from 568 minor administrative amendments and implementation

milestones. The study reveals three key findings. First, the determinants of reform are

highly sensitive to how reform is measured. Under the baseline specification restricted

to major reforms, macroeconomic stability and labour market depth emerge as

significant predictors, while demographic pressure and government debt do not.

Broadening the definition to include minor amendments shifts the set of significant

predictors substantially, indicating that studies pooling all reform events conflate

distinct political economy processes. Second, pension reform is a threshold

phenomenon rather than a linear response to fundamentals. Third, reform is strongly

state dependent: the probability of reform peaks at approximately 10% within two

years of a prior major intervention before declining to roughly 1% after a decade,

making institutional momentum the strongest predictor in the data. Policy simulations

applied to Ghana show that the country historically lay below the reform activation

threshold, with inflation emerging as the binding constraint. While Ghana's recent

disinflation has cleared this barrier, further reform requires parallel measures to

expand formal coverage and strengthen enforcement. To the best of our knowledge,

this paper provides the first systematic cross-country evidence on the determinants of

pension reform and demonstrates that these determinants differ across reform types.