Mariza Montes de Oca Leon, Achim Hagen, Franziska Holz in this IMF paper:
Is the removal of fuel subsidies as politically costly as many assume? Removing fossil fuel subsidies (FFS) is a compelling measure to fight climate change, as it improves public finances, reduces greenhouse gas emissions, and decreases local air pollutants (Sterner, 2007; Davis, 2014). Despite these benefits, subsidies remain politically stable and entrenched (Strand, 2013).
This paradox is manifest in Latin America, where mass protests in Ecuador in 2019 led to a decline in President Lenin Moreno’s popularity, and President Evo Morales’ approval ratings in Bolivia dropped after announcing a fuel price increase in 2010. Similar experiences in Europe, such as Macron’s popularity drop after the yellow vest crisis, suggest that the phenomenon is not unique to Latin America (Douenne and Fabre, 2022). With the energy and cost of living crisis in 2022, fossil fuel subsidies have reached all-time highs and are estimated to be as high as $1 trillion (IEA, 2023b) to $1.3 trillion (Black et al., 2023) in 2022.
How significant are the political costs of subsidy removal for elected officials, and which income groups disapprove most of these measures — the wealthy elites or the middle and low-income voters?
To explain this seeming paradox of FFS removal, this paper combines empirical analysis with theoretical modeling to estimate the effect of gasoline subsidy removal on presidential approval ratings in two Latin American cases: Mexico and Bolivia in the early 2010s.
Findings?
Our results suggest that removal is politically costly. Yet, these costs seem moderate and short-lived. We analyze what explains these political costs by using a probabilistic voting model. Our theoretical results suggest that high-income groups drive the loss in political support in low-income countries where subsidies are regressive. Our model incorporates trust in the president, which explains why other income groups, particularly low-income groups may prefer subsidies as a distribution mechanism.
Using micro-level survey data for Mexico, a difference-in-differences model confirms our theoretical predictions, namely that high income groups’ approval of the president decreases as a result of the subsidy removal and that low trust in the president by low income groups can cause a decrease of approval in reaction to the removal in these groups as well.
Our study offers insights for shaping effective policies. If reelection incentives of politicians lead to the lock-in of FFS that are regressive and environmentally harmful, feasible political strategies for a phase-out have to account for this challenge. Higher trust in the president is key to increase acceptance of a removal of regressive subsidies across all income groups. Trust moderates the negative effect of subsidy removal on popularity, and it can, in some cases, even revert the negative impact.
Mostly Economics 
