To clarify, on top of charity costs, we do look at the cost to the government of implementing the tax (the WHO has a whole table of useful data on the cost of various health interventions); that said, we discount this cost to some extent because it’s counterfactually less valuable than EA expenditure.
Beyond the health impact of the taxes, we only looked at the disvalue people place on reduced freedom of choice. We chose not to model the use of the taxes and took them to be just a lump sump transfer—not just quite accurate given regressivity concerns (since the poor are disproportionately likely to suffer the tax) but this is balanced out (since the poor are disproportionately likely to benefit in terms of improved health and subsequent greater earnings).
Corruption is an interesting issue, but I’m unsure if the evidence is particularly strong, or even which direction it goes—if you have higher revenues you can pay your officials more, reducing the risk of corruption. It’s probably a choice, at the end of the day, whether or not you want to model this as an extraneous adjustment factor—my sense is that we wouldn’t have much confidence in the adjustment, regardless, and so it wasn’t worth trying given lack of time.
To clarify, on top of charity costs, we do look at the cost to the government of implementing the tax (the WHO has a whole table of useful data on the cost of various health interventions); that said, we discount this cost to some extent because it’s counterfactually less valuable than EA expenditure.
Beyond the health impact of the taxes, we only looked at the disvalue people place on reduced freedom of choice. We chose not to model the use of the taxes and took them to be just a lump sump transfer—not just quite accurate given regressivity concerns (since the poor are disproportionately likely to suffer the tax) but this is balanced out (since the poor are disproportionately likely to benefit in terms of improved health and subsequent greater earnings).
Corruption is an interesting issue, but I’m unsure if the evidence is particularly strong, or even which direction it goes—if you have higher revenues you can pay your officials more, reducing the risk of corruption. It’s probably a choice, at the end of the day, whether or not you want to model this as an extraneous adjustment factor—my sense is that we wouldn’t have much confidence in the adjustment, regardless, and so it wasn’t worth trying given lack of time.