Sales Excise Taxes

What the evidence says

The clearest, most consistent thread across these sources is that pass-through of an excise tax to the retail price is not a fixed fraction — it swings from well under 100% to well over 100% depending on product tier, tax design, and market structure. Cigarette taxes are under-shifted in economy-tier brands (pass-through of 0.956) and over-shifted in premium brands (1.140), with mid-price brands landing near complete pass-through (0.990); across the full price distribution, pass-through rises from 0.922 at the 5th percentile to 1.167 at the 95th. Alcohol taxes run 100–200% pass-through (a 10-cent tax increase yields a 10–20 cent price rise). Sports-betting taxes pass through at 76% on average, but that average conceals a design effect: firms that shroud the tax (excluding it from the posted price) pass through 90%, while firms that post tax-inclusive prices pass through only 16%. E-cigarette tax pass-through ranges from 0.07 in China to 1.67 in the US, and in Indonesia and China taxes are sometimes absorbed by firms rather than passed to consumers at all. A 1926–1929 study of German food-staple taxes found the same spread historically: rye flour taxes over-shifted to 190% of the tax, wheat flour to 110%, while slaughter taxes on pork and beef under-shifted to 54% and 82% respectively.

On consumption, the evidence agrees on direction and disagrees only on size. Sugar-sweetened beverage (SSB) taxes decrease purchases of taxed beverages: a global panel estimates a price elasticity of −0.423, Philadelphia saw a 51% decline in taxed-beverage volume sales in the year after tax initiation, Mexico saw purchases fall 6% on average in 2014 (accelerating to 12% by December), and a set of studies compiled in one review found reductions ranging from 6.3% to 40%. Alcohol taxes decrease consumption with elasticities of −0.46 (beer), −0.69 (wine), and −0.80 (spirits), and an aggregate-level partial correlation of −0.44 across 24 studies (19 of 24 finding a relationship distinguishable from zero). Cigarette tax increases in a Serbian simulation decreased total packs sold by 4.7% in year one and 5.3% in year two, cut adult smoking prevalence from 37.9% to 35.8% over two years, and were projected to deter or delay smoking initiation in more than 17,000 youths. E-cigarette demand is comparatively inelastic (elasticity range −0.62 to −0.02 across studies), with taxes producing only a modest reduction in adult use (odds ratio 0.89) and use among 18–24-year-olds (odds ratio 0.87) in the three US studies available.

A second pattern in the SSB literature is substitution: taxes decrease purchases of the taxed product but increase purchases of untaxed substitutes and shift purchases to untaxed jurisdictions. In Mexico, untaxed-beverage purchases rose 4% (largely bottled water) alongside the decline in taxed beverages. In Philadelphia, sales fell inside the city but rose in neighboring counties by 1.14 million ounces per four-week period in supermarkets/grocery/mass-merchandiser channels and 33.0 thousand ounces in pharmacies; cross-border shopping is estimated to offset 25–30% of the total sales-reduction effect. This cross-border pattern is described as general across SSB tax evaluations, "with a few exceptions."

On transportation excise taxes, the evidence covers two distinct levers — fuel/VMT taxes and congestion pricing — and the distributional findings for each are contested or conditional rather than settled (see next section). What is not contested: current fuel tax revenue is described as inadequate to meet Highway Trust Fund infrastructure needs because of inflation and rising fuel efficiency; a revenue-neutral VMT fee would decrease the average tax paid per vehicle, and that decrease would grow further if VMT itself falls as gas prices rise; and the shift in burden for low-income drivers specifically under a VMT fee would be small — a decrease of less than one dollar for most groups.

The evidence is thin or entirely absent on several outcomes the topic invites questions about. No source here quantifies administrative or enforcement costs of any excise tax. The magnitude of "regressive or inequitable effects" of congestion pricing and low-emission-zone rules on commuters and carriers is stated but not sized. Long-run (multi-year) demand elasticity for SSBs is explicitly flagged as unknown — the Mexico estimates cover only the first year after the tax. And revenue from e-cigarette excise taxes is described only as "small compared with traditional tobacco excise revenue," without a figure.

Where the evidence splits

Fuel tax regressivity is reported in opposite directions. One source states fuel and VMT taxes are regressive because the same per-mile or per-gallon cost applies to all taxpayers regardless of income, and a second finds that low-income drivers pay a higher percentage of their income in fuel taxes than non-low-income drivers even though they pay less in absolute dollars. A third source instead reports that "most studies suggest that fuel taxes are nearly proportional...in developed countries." The sources do not explain the disagreement — they draw on different literatures and datasets (a general policy review versus register-based data), but none directly addresses why the other finds a different baseline distribution.

Congestion pricing's distributional effect goes in different directions depending on the city and what happens to the revenue. A meta-analysis of Stockholm's system found no regressive effect, and progressive effects if revenue is invested in transit operation. A separate modeling exercise found Gothenburg's system is more regressive than Stockholm's, with all but the highest-income drivers losing more than they gain. This split is explained by the sources: it turns on the transit-reinvestment assumption and on features specific to each city's charge structure, not on a shared measurement disagreement.

SSB tax pass-through to retail prices was full (100%, exactly the tax amount) in Mexico but 68% in South Africa. The sources report these as findings from different countries without reconciling the gap, so the difference in market or country context is the only distinction the claims themselves supply.

E-cigarette tax pass-through goes from firms absorbing the tax entirely in Indonesia and China to producers over-shifting it (pass-through above 1.0) in several US markets. This one is explained in the sources: it tracks jurisdiction and tax structure — ad valorem tax designs produce lower pass-through and lower effective tax incidence (7 percentage points lower) than specific tax designs.

Thresholds and dose-response

The World Health Organization's recommended threshold for SSB excise taxes is a retail price increase of at least 20%; applying the global elasticity estimate (−0.423), a 20% price increase is projected to reduce global consumption by approximately 8.5%.

Cigarette tax pass-through moves monotonically with the brand's position in the price distribution: pass-through is 0.922 (under-shifting) at the 5th price percentile and rises to 1.167 (over-shifting) at the 95th percentile, with near-complete pass-through (0.990) in the mid-price segment.

Regional pass-through of fuel taxes is roughly 15 percentage points lower in the richest, wealthiest (highest house prices), most densely populated, and most urban areas compared with the opposite extremes on each measure; the resulting regional price disparities were mitigated over the following two to three years.

Tax structure sets a threshold on incidence: states applying ad valorem e-cigarette tax structures show tax incidence 7 percentage points lower than states applying specific tax structures.

Open empirical questions

• No source here quantifies administrative or enforcement costs of any excise tax examined.
• Long-run (multi-year) demand elasticity for sugar-sweetened beverages is not established — the Mexico estimates cover only the first year post-tax, and the source explicitly flags long-term elasticity as unknown.
• The magnitude of "regressive or inequitable effects" of congestion pricing and low-emission-zone rules on commuters and carriers is asserted but not sized in the sources reviewed here.
• One source notes an increase in SSB consumption among teenagers in Mauritius, against the general decline pattern elsewhere, without an explanation for the exception.