Simulating the Health Impact of Updating Alcohol Tax Rates in the Philippines

Dante Salvador, Jr., Michael Mo, and Reneepearl Kim Sales

The Alliance for Improving Health Outcomes was tapped by the Department of Health – Health Policy Development and Planning Bureau to conduct simulation studies to assess the health impact of updating alcohol tax rates in the Philippines. This is the first study in the Philippines that utilizes a simple, replicable modelling framework to estimate health effects of alcohol tax intervention. Unlike previous estimates, if any, this study is linked to tax structure, actual consumption, and relative risk functions. 

A two-step economic modelling approach was used to simulate the impact of tax and price changes of alcohol to revenue and health outcomes starting from 2019. This study aimed to simulate the public health and economic impact of updating alcohol tax rates in the Philippines in two tax scenarios:

  1. Department of Health-Department of Finance (DOH-DOF): higher specific tax and 10% annual indexation rate
  2. House of Representatives (HOR): higher specific tax and 7% annual indexation rate

Main Results

Cumulative consumption

Both the DOH-DOF and HOR proposed tax structures will result in a reduction of alcohol consumption (Table 1). The DOH-DOF tax structure will result in a greater reduction of consumption across all alcohol types compared with the HOR proposed rate, owing to the deterrent effect of higher prices brought about by higher taxes. With the DOH-DOF proposal, the consumption of spirits, beer, and wine will decrease by 18.3%, 20.5%, and 4.68% respectively after a four-year implementation. 

Cumulative lives saved

The simulation demonstrates that increasing tax rates for alcoholic beverages from the current schedule to the proposed structure of the DOH-DOF or HOR will result in population gains in terms of lives saved (Table 2). The proposed tax schedule of DOH-DOF will result in almost 35,000 more lives saved compared with the HOR proposed rates after a four-year implementation (57,020 vs 22,152 lives saved). In both simulations, the tax intervention will result in more lives saved for males, demonstrating that the intervention greatly benefits the male population due to the greater prevalence of alcohol consumption among males.

Economic implications

Assuming that the price elasticity of alcohol consumption remains constant or increasing across the entire range of prices, government revenues are projected to increase with the implementation of either the DOH-DOF and HOR proposals (Table 3). Since the price elasticity of alcohol is less than one (inelastic demand), the regime with the higher tax rates (DOH-DOF) will yield higher revenues.

Model Assumptions

  1. Only the annual increase in specific tax rates were changed. No changes were made on ad valorem tax.
  2. Complete or 100% tax pass through was assumed.
  3. This model assumes that no cross-drinking occurs in the population.
  4. Closed population is projected without incorporating effects of fertility on population.
  5. Fixed mortality rate figures were used to project future mortality.
  6. Binge drinking and its adverse consequences cannot be modelled using this methodology. 
  7. Other effects such as increased income or interventions which increase or decrease alcohol consumption (marketing or health promotion efforts) and their interactions were not modelled.

Conclusions and recommendations

Increasing alcohol taxation will result in more lives saved compared with the current tax structure, in addition to generating additional revenues. Our findings are  in line with the conclusions of the study by Stacey et al. (2018), where the shift in disease burden as a result of tax increase produces life gains as well as raise revenues. We recommend that the government push for higher alcohol tax rates in light of this new evidence. To further strengthen the current evidence base, we recommend that impact on alcohol-related morbidity and its implications on health expenditure be modelled alongside mortality effects.

Interested in this study?