Home » A Regressive subsidy, an unequal reform: The Rich benefited, but the poor paid in missed healthcare: Guest post by Xufeng Liu

A Regressive subsidy, an unequal reform: The Rich benefited, but the poor paid in missed healthcare: Guest post by Xufeng Liu

by NNW Bureau
0 comments

This is the 10th in this year’s series of posts by PhD students on the job market.

Economists have long argued that fuel subsidies are expensive and regressive: richer households buy more fuel, so they capture far more of the benefit than poorer ones. When Nigeria’s new president declared on May 29, 2023 that “the fuel subsidy is gone,” pump prices nearly tripled within days. But a subsidy that is regressive in who receives it need not be progressive in who bears its removal.

In my job market paper, my co-author and I find that the short-run disruptions to healthcare access and utilization following Nigeria’s reform fell hardest on poorer and rural households who benefited least from the subsidy while it was in place.

Transportation is part of the price of healthcare, and fuel subsidies help keep healthcare accessible

Early-life healthcare requires numerous, separate visits to health providers. WHO recommends at least eight antenatal contacts, and Nigeria’s immunization schedule asks parents to bring a child back six times in the first 15 months. Additionally, an average African child under 5 experiences a new acute infectious illness every 3 to 5 weeks; more trips are potentially needed if the child is sick and needs curative medical care. While vaccines are free and other maternal care is subsidized at some government health facilities, the journey to the clinic still costs time and often money, with one-third of rural women citing distances as a main barrier to accessing care. 

Fuel subsidies help keep the cost of these trips low by holding petrol prices below market levels. They are also a major form of government intervention with a total of $1 trillion spent worldwide in 2022, twice the annual subsidies on agriculture and ten times the subsidies on industry and technology. In Nigeria, the consequences of removing this support were immediate: between May and June 2023, average within-city bus fares rose by 98 percent. Yet despite the scale of these interventions, we know surprisingly little about what happens when subsidies are removed. Of 77 studies on fossil fuel subsidy we reviewed, 56 rely on simulation; only seven analyze actual removals, and none study health or broader behavioral changes.

Identification based on the pre-reform fuel supply network

Because the reform was nationwide with no Nigerian community untreated, we leverage the differences in communities’ exposure to the reform based on their proximity to the fuel-supply network. Before the reform, the subsidy functioned as a binding national price ceiling that absorbed inland transport costs and kept petrol prices similar nationwide. Therefore, although subsidy removal raised fuel prices nationwide, there were larger increases in communities further from key points in the supply network.

To exploit this spatial gradient, we digitized 375 daily government petroleum-distribution reports to reconstruct the pre-reform network and identify “mother depots,” which receive fuel without relying on roads. We measured road distance from communities in the 2018 and 2023–24 Demographic and Health Surveys to their nearest mother depot. Household-reported and official petrol-price data show that one standard deviation (197 km) of additional distance was associated with approximately 6 percent larger post-reform price increases within states. We then interact depot distance with the post-reform period to estimate a continuous difference-in-differences model. One key identifying assumption is strong parallel trends: absent the reform, changes in healthcare utilization would not have varied systematically with depot distance, conditional on our controls.

We make several checks to support our identification strategy. Before the reform, household characteristics were broadly similar across depot distances, and event studies show no systematic differential healthcare trends. We also assess the sensitivity of our estimates to potential violations of parallel trends using the HonestDiD approach of Rambachan and Roth. Placebos permuting reform dates and depot distances suggest the results are unlikely to reflect chance. We also find no evidence of anticipation by households or fuel suppliers. We include state fixed effects to absorb persistent differences and geopolitical-zone-by-post fixed effects to absorb concurrent macroeconomic shocks, including the foreign-exchange reform.

Figure 1 presents the DiD estimates. We also use the same interaction between depot distance and the post-reform period as an instrument for local petrol prices, yielding broadly similar conclusions.

Preventative care was skipped; time-sensitive care was downgraded

For preventative care, every standard deviation further from a mother depot, the probability that a child received a vaccine dose that had come due fell by 4.7 percentage points after the reform, about 9 percent of the pre-reform rate of 55 percent. Antenatal care followed the same pattern. Any antenatal visit declined by 3.2 percentage points, antenatal care at public primary facilities by 3.6 points, and receipt of three or more doses of preventive malaria medication during pregnancy by 3.4 points.

For acute healthcare needs, households changed the type rather than simply forgoing care. Attendance of a skilled practitioner at birth declined by 2.4 percentage points. Among sick children, parents were more likely to rely on informal providers and symptomatic-only treatment, such as antimotility drugs for diarrhea which clinical guidelines advise against due to serious side effects.

Healthcare disruptions were concentrated among poorer and rural households

We next examine who bore the greatest burden of the reform by comparing healthcare responses across wealth, rural residence, and local health-facility availability. Vaccine-dose receipt fell by 6.9 percentage points among poorer children, while the corresponding estimates for richer and urban children were smaller and statistically indistinguishable from zero. Declines in antenatal care and shifts toward informal treatment were similarly concentrated among poorer, rural households and communities with fewer health facilities. Richer women, by contrast, shifted from private toward public primary facilities for antenatal care, without substantially reducing their overall use. Skilled birth attendance was an exception. It declined by 8.2 percentage points among richer women but little change among poorer women. This likely reflects that richer women had more formal care to lose: before the reform, 75.7 percent of richer women had skilled attendants, versus only 17.5 percent of poorer women.

Health Impacts: cleaner air but mostly in cities

Beyond healthcare utilization, we find no overall effect on neonatal mortality or self-reported health status, but some evidence of health improvements in cities: greater gestational exposure to the post-reform period was associated with lower neonatal mortality among urban births. One possible explanation is cleaner air due to less fuel consumption: relative to comparison countries, PM2.5 declined by about 10 percent in urban Nigeria, compared with 3 percent in rural areas.

Compensation should not solely be based on forgone subsidy benefits

Our findings do not argue for maintaining an expensive and regressive subsidy. Rather, they highlight an asymmetry in the distribution of the reform’s health consequences. In the short run, healthcare utilization fell most among poorer, rural households, while potential air-quality benefits concentrated in cities. This contrasts with the conventional incidence analysis which finds richer and more urban households typically capture the largest benefits. Compensation based solely on foregone subsidy benefits may therefore miss the households most affected by removal.

Finally, our analysis captures short-run effects; longer-term consequences of reduced vaccination and antenatal care require further follow-up. Evaluating fuel-subsidy reform therefore requires looking beyond fiscal savings and the distribution of subsidy benefits to consider who loses access to healthcare, who benefits from improved environmental conditions, and how these changes shape health over time.

read more: https://blogs.worldbank.org/en/impactevaluations/a-regressive-subsidy–an-unequal-reform–the-rich-benefited–but

You may also like