What the evidence says
Across every source in this set, sugar-sweetened beverage (SSB) demand moves in the same direction with price: consumption decreases as price rises. Pooled and meta-analytic estimates cluster in a band. A pooled analysis across interventional and prospective observational studies puts the elasticity at −0.67 (95% CI: −0.31 to −1.04), meaning a 10% price increase is associated with a 7% decrease in consumption. A separate pooled estimate combining tax-simulation modeling with real-world tax studies puts it higher, at −1.30 (95% CI: −1.09 to −1.51), and a systematic review restricted to US simulation studies lands at −0.79 (95% CI: −0.33 to −1.24). An umbrella review spanning four SSB meta-analyses reports a median demand reduction of 9.1% for a 10% price increase, with elasticity estimates across those four meta-analyses ranging from −0.67 to −1.59.
At the level of individual studies the spread widens further. A review of 14 studies found a mean elasticity of −0.79 with a range from −0.13 to −3.18; a review of 12 studies found a mean of −1.21, ranging from −0.71 to −3.87; a panel-data study using instrumental variables estimated an elasticity of about −1.4; and a cross-sectional compensated own-price elasticity was estimated at −0.805. One source states plainly that this wide range is not explained by the paper reporting it.
SSB demand also stands out relative to other food categories: one cross-sectional demand-system estimate reports that most food categories are price-inelastic, but SSBs (along with dairy and non-SSB beverages) are exceptions to that pattern — though the specific SSB elasticity coefficient is not given in that claim. The same source finds that SSBs and non-SSB beverages behave as substitutes, so that a positive cross-price elasticity applies: as SSB prices rise, some consumers shift toward non-SSB beverages, though no magnitude for that cross-price elasticity is reported.
For the long-run horizon specifically, only one source explicitly labels its estimates as based on "long-run" input data. That source reports several distinct long-run own-price elasticities depending on product category and household income: carbonated SSBs at −0.73 overall (ranging from −0.49 to −1.02 by income quartile), total SSBs at −1.29 for high-income versus −0.95 for low-income households, non-concentrated SSBs at −0.79 to −0.85 across income thirds, and regular carbonated soft drinks at −1.03 to −1.04 regardless of income. A habit-formation model applied to the same long-run data produces elasticities of −1.06 to −1.22 for low-income households and −1.44 to −1.54 for high-income households, depending on whether the model assumes rational or myopic behavior. Outside this one source, none of the other elasticity estimates in this set specify whether they measure short-run or long-run response — that distinction is absent from the pooled, meta-analytic, and simulation-based estimates above.
Where the evidence splits
The three pooled/meta-analytic estimates from the same review disagree on magnitude: −0.67 for interventional/observational studies, −1.30 for the mixed tax-simulation/real-tax sample, and −0.79 for simulation-only studies. The source attributes at least part of this to differences in study composition — simulation versus empirical, and tax/price-change studies versus interventional-observational studies — but does not fully reconcile the three figures.
The 14-study review (mean −0.79) and the 12-study review (mean −1.21), both cited by the same paper alongside a third panel-IV estimate of about −1.4, disagree on central tendency by a wide margin. The source states this heterogeneity is not explained in the underlying papers.
Within the single source that reports long-run, income-differentiated elasticities, the pattern of income difference itself is contested. For total SSBs and for the habit-formation model, high-income households show more elastic demand than low-income households, and the source states this difference was tested and found. For non-concentrated SSBs and for regular carbonated soft drinks, the same source reports elasticities that are similar across income groups, with no difference found between them. The source does not explain why the income gradient appears for some SSB categories and not others.
The umbrella review reports elasticity estimates ranging from −0.67 to −1.59 across four meta-analyses and suggests this "may relate to methodological differences across studies," without specifying which differences produce which estimates.