Impact of soda tax on beverage price, sale, purchase, and consumption in the US: a systematic review and meta-analysis of natural experiments

Background As a primary source of added sugars in the US diet, sugar-sweetened beverage (SSB) consumption is presumed to contribute to obesity prevalence and poor oral health. We systematically synthesized and quantified evidence from US-based natural experiments concerning the impact of SSB taxes on beverage prices, sales, purchases, and consumption. Methods A keyword and reference search was performed in PubMed, Web of Science, Cochrane Library, Scopus, and EconLit from the inception of an electronic bibliographic database to Oct 31, 2022. Meta-analysis was conducted to estimate the pooled effect of soda taxes on SSB consumption, prices, passthrough rate, and purchases. Results Twenty-six natural experiments, all adopting a difference-in-differences approach, were included. Studies assessed soda taxes in Berkeley, Oakland, and San Francisco in California, Philadelphia in Pennsylvania, Boulder in Colorado, Seattle in Washington, and Cook County in Illinois. Tax rates ranged from 1 to 2 ¢/oz. The imposition of the soda tax was associated with a 1.06 ¢/oz. (95% confidence interval [CI] = 0.90, 1.22) increase in SSB prices and a 27.3% (95% CI = 19.3, 35.4%) decrease in SSB purchases. The soda tax passthrough rate was 79.7% (95% CI = 65.8, 93.6%). A 1 ¢/oz. increase in soda tax rate was associated with increased prices of SSBs by 0.84 ¢/oz (95% CI = 0.33, 1.35). Conclusion Soda taxes could be effective policy leverage to nudge people toward purchasing and consuming fewer SSBs. Future research should examine evidence-based classifications of SSBs, targeted use of revenues generated by taxes to reduce health and income disparities, and the feasibility of redesigning the soda tax to improve efficiency.


Introduction
Sugar-sweetened beverages (SSBs) are beverages sweetened with various forms of added sugars, including brown sugar, corn syrup, glucose, lactose, and sucrose (1).In the United States, SSB consumption is prevalent, with six in ten youths and five in ten adults consuming SSBs on any given day during the period of 2011-2014 (2,3).This consumption amounts to over 140 kcal from SSBs per day for both youth and adults (2,3).Extensive epidemiologic studies have consistently documented that SSB consumption, as a primary source of added sugars in the US diet, is a significant contributor to the prevalence of obesity, cardiometabolic diseases, and oral health risk (4)(5)(6).These associations have also been observed globally (7,8).
Recent policy interventions aimed at combating obesity have utilized economic incentives to "nudge" (i.e., promote or encourage) people toward a healthier diet choice (9)(10)(11).For example, healthy food subsidies promote fruit and vegetable intake, while excise taxes aim to discourage the consumption of less desirable foods and beverages.Excise taxes, such as soda taxes, are implemented in countries like the US, France, New Zealand, Netherlands, and South Africa (12).A soda tax is a specific excise tax charged on the sale of SSBs to reduce consumption (13).Merchants pay the tax, which is then passed on to consumers through higher prices.The amount of the tax varies across different regions, as it is applied by both state and federal governments.
Existing research on the effect of soda taxes in the US can be categorized into three groups: "proxy," "modeling," and "local" studies.The "proxy" studies have used state sales tax on soda, candy, and other qualified groceries as a substitute for a soda tax due to the absence of soda taxes or related data (19)(20)(21)(22)(23)(24).However, using such a "proxy" can be problematic because a soda tax, as an excise tax, is fundamentally different from a state general or selective sales tax applied as a percentage of the purchase price (24).The "modeling" studies have employed systems science models to simulate the effect of a soda tax using pre-specified parameter values and statistical distributions (25)(26)(27)(28)(29)(30)(31)(32)(33)(34).However, these studies differ from other categories of research in that they are prospective in nature.They aim to identify potential effects of policies that have yet to be enacted and, by necessity, make assumptions about possible retailer and consumer responses.Although the "modeling" studies compare simulated counterfactuals to status quo baselines, they do not provide direct causal inference due to their non-experimental study design.The "local" studies, on the other hand, have utilized quasi-experimental methods to compared soda prices, sales, purchases, or consumption between cities that have implemented a soda tax and neighboring cities without such a tax, or before and after the implementation of a soda tax in a city.Unlike "proxy" or "modeling" studies, the "local" studies directly estimated the impact of soda tax using temporal and geographical variations in tax implementation (35)(36)(37)(38)(39)(40)(41)(42)(43)(44)(45)(46)(47)(48)(49)(50)(51)(52).Thus, these "local" studies serve as natural experiments that provide valuable causal inference.Powell et al. conducted a review of seven local SSB tax implementations and reported an average tax pass-through rate of 70% (53).Another review by Powell et al. found that the demand for SSBs declined by 20%, with an estimated price elasticity of demand of −1.5 following the implementation of soda taxes (54).
This study aims to systematically synthesize and quantify the impacts of soda taxes on beverage prices, sales, purchases, and consumption in the US.By exclusively relying on natural experiments (i.e., "local" studies), our review provides robust causal inferences regarding the effectiveness of soda taxes.Moreover, our approach goes beyond a narrative review by providing quantitative estimates of the magnitude of the tax effect.The findings of this study can inform local, state, and federal policymakers in designing or revising soda tax-related legislation and implementation strategies to effectively combat obesity.

Methods
The present study was conducted following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (55).

Eligibility criteria
Studies meeting all of the following criteria were eligible for the review: (1) Participants: consumers, stores, or beverage items within US taxing jurisdictions that implemented a soda tax; (2) Interventions: soda tax (SSB excise tax); (3) Comparisons: consumers, stores, or beverage items within and outside the US taxing jurisdictions that implemented a soda tax; (4) Outcomes: beverage prices, sales, purchases, and consumption; (5) Study design: natural experiment; (6) Article type: peer-reviewed original study; (7) Time window of search: from the inception of an electronic bibliographic database to Oct 31, 2022; and (8) Language: English.
Frontiers in Public Health 03 frontiersin.org

Meta-analysis
Meta-analyses were performed to estimate the pooled effects, represented as mean differences, of soda taxes in the US.The six outcomes included: (1) change in prices of taxed beverages (i.e., SSBs); (2) change in prices of untaxed beverages; (3) change in purchases of taxed beverages; (4) change in purchases of untaxed beverages; (5) change in the consumption of taxed beverages; and (6) tax passthrough rate.Out of the 40 studies, 13 were excluded from the meta-analyses due to: (1) non-overlapping outcome measures (36,40,44,(56)(57)(58)(59)(60)(61)(62)(63), or (2) neither standard error nor confidence interval (CI) reported (35,43).To assess heterogeneity among the included studies, we employed the I 2 index, which allows us to quantify the degree of variability between study estimates (64).The I 2 index was interpreted as modest (I 2 ≤ 25%), moderate (25% < I 2 ≤ 50%), substantial (50% < I 2 ≤ 75%), or considerable (I 2 > 75%) (65).Based on the level of heterogeneity observed, we estimated the meta-analyses using either a fixed-effect (FE) model or a random-effect (RE) model.The FE model was utilized when modest or moderate heterogeneity was present, while the RE model was employed when substantial or considerable heterogeneity was observed.To assess publication bias, we conducted Begg's and Egger's tests.These tests allow us to evaluate potential bias in the included studies (66).Additionally, we performed random-effect metaregressions to assess the dose-response effect of alternative soda tax rates on the various outcomes.Meta-analyses were performed using Stata 16.1 MP version.We employed these methods to synthesize the available evidence and provide quantitative estimates of the pooled effects of soda taxes on the specified outcomes.

Study quality assessment
Following Littell et al. (65) and An et al. (67), we designed a study quality assessment tool that rated each study based on ten criteria (Table 5).For each criterion, a score of 1 is assigned if the answer is "yes"; otherwise, a score of 0 is assigned if the answer is "no, " "not applicable" or "not reported." We sum the scores of all ten criteria for a total study-specific score, ranging from 0 to 10.The study quality score was not used as a criterion for inclusion of a study, but as a measure of the strength of the scientific evidence.

Study selection
Figure 1 shows the study selection flow diagram.We identified 4,574 articles through the keyword and reference search.After removing duplicates, 3,655 articles underwent title and abstract screening, in which 3,562 articles were excluded.The remaining 93 articles were reviewed of full texts against the eligibility criteria.Of these, 53 articles were excluded, including 12 "modeling" studies, 14 "proxy" studies, 17 studies that did not examine soda taxes, six commentaries, and four non-US-based studies.The remaining 40 studies that examined the impact of soda taxes on beverage prices, sales, purchases, and consumption in the US were included in the review (35-52, 56-63, 68-81).

Study findings
Table 3 reports the estimated effects of soda taxes.Four key findings have emerged.First, soda taxes led to increased SSB prices in the taxing jurisdictions relative to the nontaxing jurisdictions.In Berkeley, a 1.0 ¢/oz.soda tax increased the price for taxed beverages by 0.65¢/oz.relative to non-Berkeley stores (51).Across all brands and sizes of SSBs, the overall price of taxed beverages increased by 0.43 ¢/ oz.relative to neighboring cities (37).The price increase was 0.47 ¢/ oz. for small-size SSBs (≤33.8 oz), 0.46 ¢/oz.for a 2-liter SSB bottle, and 0.49 ¢/oz.for a multipack of soda relative to that in neighboring cities (45).In other taxing jurisdictions implementing a 1.0 ¢/oz.(76).An increase in the soda tax rate by 1 ¢/oz.decreased monthly household purchases of taxed beverages by 53.0 oz. or 12.2% (41).The average quantity of taxed beverages purchased per shopping trip decreased in Philadelphia and increased in neighboring cities, which resulted in a relative decrease of 8.5 oz. of SSBs purchased per shopping trip in Philadelphia (40).
Third, soda taxes reduced the consumption of SSBs in taxing jurisdictions relative to nontaxing jurisdictions.Consumption of SSBs was found to decrease by 25% in Berkeley relative to neighboring cities (46).Soda intake was reduced by 0.81 servings per week in Philadelphia compared with all other comparison cities two years after tax implementation (60).The odds of daily consumption of regular soda and energy drinks were 40 and 64% lower, respectively, in Philadelphia relative to the comparison cities two-month post-tax (52).Adults in Philadelphia were found to consume regular soda 10.4 fewer times per month after soda tax implementation, denoting a reduction of approximately 30% (40).People in Philadelphia were more likely to reduce their frequency of SSB consumption, and their monthly SSB consumption declined by 51.65 oz.compared to those residing in neighboring cities (58).
1. Pass-through was complete in large chain supermarkets (+1.07¢/oz., p = 0.001) and small chain supermarkets and chain gas stations (1.31¢/oz., p = 0.004), partial in pharmacies (+0.45¢/oz., p = 0.03), and negative in independent corner stores and independent gas stations (−0.64¢/oz., p = 0.004).Soda sales fell significantly compared with control beverage groups in the period immediately following the election, decreasing by between 10 and 20% compared with precampaign levels.On-campus soda sales continued to fall when the tax was implemented in the city but not on campusdecreasing by 18-36% compared with the precampaign period-and remained at this depressed level after the tax implementation on campus.17 The volume sold of taxed beverages declined by 38.9% in Philadelphia after soda tax implementation compared to Baltimore.

18
No evidence showed that substantial changes in the overall consumption of SSBs or added sugars consumed through beverages for either adults or children after the tax.
There was a slight decrease in the volume of SSBs purchased per shopping trip in Oakland and a small increase in purchases at stores outside of the city, resulting in a decrease in purchases of 11.33 oz., but it was not statistically significant.
Roughly 60% of the tax passed on to consumers in the form of higher prices.
(Continued) 2. Prices of some non-taxed beverages also increased while the prices of healthy foods generally did not.

22
The soda tax was associated with reductions of taxed beverage purchases at 3 and 6 month but not 12 month.Analyses aggregating all 6 weeks of post-tax time points showed significant reductions (−203.7 oz., 95% CI = -399.6,−7.8).

23
1.There was a 82% tax pass-through for bottled regular soda one year after tax implementation, with a price increase of 8%.    2. Volume sold of taxed beverages fell to a greater extent for family-versus individual-size beverages (31% vs. 10%) and fell to a greater extent for soda (29%) compared to all other beverage types.
3. Moderate substitution to untaxed beverages was found-volume sold of untaxed beverages increased by 4% (p < 0.05).
4. There was no significant increase in the overall volume sold of taxed beverages in the 2-mile border area of Seattle.
On average, in the first year of post-tax implementation, prices of taxed beverages rose by 1.03 ¢/oz.(p < 0.001), corresponding to a 59% tax passthrough rate.27 1.There was a 119% tax pass-through rate across all taxed beverages in Cook County compared to its comparison site.
2. This price increase was 34% for taxed beverages.
4. Pass-through was higher for individual-size (126%) compared to family-size (117%) beverages and higher for energy drinks (145%) compared to other sweetened beverages.
5. Based on the baseline prices of different categories and sizes of beverages, the effective percentage increase in beverage prices ranged from a 52% increase for family-size soda to a 10% increase for family-size energy drinks.

28
People in Philadelphia were more likely to decrease their frequency of SSB consumption (39.2% vs. 33.5%),and less likely to increase their frequency of SSB consumption (38.9% vs.The tax was largely, but not completely, passed through to consumers.In both the hand-collected store data and restaurant data, pass-through is slightly less than 75%, whereas pass-through is just over 50% using the scanner data; consumers bear most, but not all, of the largest tax on sugar-sweetened beverages.2. There was no significant difference in 100% juice or milk intake. 3. In subgroup analyses, the tax was associated with a reduction of 1.13 servings per week in Hispanic/Latinx adolescents (95%CI = −2.04,−0.23 servings; p = 0.01) and 1.2 servings per week in adolescents with obesity (95% CI = −2.33,−0.13 servings; p = 0.03).

32
Taxed beverages sales fell by 14%, but 46% of this decrease is offset with an increase in the border area.
The taxed beverage prices increased by 0.49 ¢/oz. in Oakland relative to Sacramento.Tax pass-through is 49%.
(95% CI = 0.05,0.75) in pharmacies but did not change in other store types.Prices of taxed individual-size soda specifically increased in all store types, by 0.91-2.39¢/oz.
3. There were not statistically significant price effects for unsweetened beverages and fountain drinks.2. Declines were larger for familysize (29%) compared to individual-size (10%) beverages; particularly for soda (36% decrease for family-size compared to no change for individual-size).
3. There was no change in volume sold of taxed beverages in Seattle's 2-mile border area, suggesting no cross-border shopping.
Prices of taxed beverages increased by 1.04 ¢/oz.2-year post-tax, with 59% tax pass-through rate.

36
A 5% volume reduction in Washington but fail to detect an effect in Berkeley.
Prices in Washington reacted sharply and promptly (often by a larger magnitude than the tax), whereas retail prices in Berkeley reacted marginally (by less than 30% the magnitude of the tax).The tax is passed through at an average rate of 97%, leading to a 34% price increase.

38
Each cent per ounce of taxes causes the sales quantity of taxed beverages to decrease in a range of 5.1-14.4%.
Each cent per ounce of taxes causes the price of the taxed beverages to increase in a range from 0.47 to 0.98 ¢/oz.2. There was a 9% increase in volume sold of taxed beverages in the twomile border area surrounding Oakland relative to the Sacramento border area, driven by a 12% increase for family-size taxed beverages.
3. After accounting for this cross-border shopping, there was a net decrease of 6% in taxed beverage volume sold in Oakland.
4. There was no significant change in untaxed beverage volume sold in either Oakland or its border area relative to their respective comparison sites.
Taxed beverage prices increased by 0.

Outcome
Studies included in the metaanalysis

Study quality assessment
Table 5 reports criterion-specific and overall ratings from the study quality assessment.The included studies, on average, scored six out of ten, with scores ranging from 3 to 9. All studies collected and analyzed both pre-and post-tax outcomes for the intervention group (i.e., retailers or consumers residing in the taxing jurisdiction) and included a control group (i.e., retailers or consumers living outside the taxing jurisdiction or beverages not subject to the soda tax).The majority of studies (n = 36) had comparable intervention and control groups.Additionally, most studies had a sample representative of the stores or the population residing in the taxing jurisdiction (n = 17) (35, 36, 38-42, 44, 45, 50, 51, 60, 70-72, 74, 75).Furthermore, most studies included the change in beverage prices (n = 28) (36-39, 42, 43, 45, (36, 40, 42, 44, 46, 47, 50-52, 58, 60).Additionally, several studies assessed the impact of cross-border sales or purchases (n = 17) (36, 37, 39, 40, 42, 46, 48-50, 57, 62, 69, 74, 77, 79-81).However, we noted some limitations in the quality of the included studies.Only two studies provided a sample size justification (58,62), and nine studies performed adequate statistical procedures to adjust for temporal trends of the outcomes (37,44,47,49,50,52,60,73,79).To enhance the rigor of future research in this area, we recommend that studies provide clear justifications for sample sizes and utilize appropriate statistical procedures to account for temporal trends.Additionally, researchers should aim to improve reporting practices, including the transparent reporting of study limitations and potential sources of bias.

Discussion
A few previous meta-analyses estimated the effect of soda taxes or SSB prices on beverage sales, purchases, and consumption.Nakhimovsky et al. reviewed studies in middle-income countries and estimated that a 10% increase in SSB prices was associated with a reduction in SSB consumption from 1.2-9.3kcal per person per day (83).Teng et al. estimated that a 10% soda tax was associated with a decline in SSB purchases and consumption by 10% (84).Escobar et al. (14) and Powell et al. (85) estimated the price-elasticity of SSBs to be −1.30and − 1.21, indicating that a 10% increase in SSB prices was associated with a reduction in demand for SSBs by 13.0 and 12.1%, respectively.Afshin et al. (12) reported that a 10% increase in SSB prices was associated with a reduction in SSB consumption by 7%.Those estimates from the previous meta-analyses were comparable to the estimated impact of soda taxes on SSB purchases and consumption in this review.However, the key differences are: previous reviews and meta-analyses were nearly exclusively based on the "proxy" or "modeling" studies due to lacking a soda tax in reality and collected data from multiple other countries besides the US.This review contributed to the literature by comprehensively assessing the efficacy of soda taxes based on evidence reported exclusively from natural experiments (the "local" studies).In addition, the review provided quantitative estimates of the magnitude of the tax effect.Review findings can potentially inform policymakers in designing and implementing soda taxes to curb the obesity epidemic.
Several main criticisms of soda taxes are worth noting.First, substitution effects in addition to cross-border shopping may partially offset the intended effect of soda taxes on SSB purchases and consumption (49,50).A nationwide soda tax adoption could eliminate cross-border shopping (49).Second, the soda tax could be regressive and the impact of soda consumption is greatest in low income populations.However, a systematic review revealed only modest differences-0.1-1.0%and 0.03-0.60% of annual household income paid for soda taxes for low-and high-income households, respectively (86).Third, revenues collected by soda taxes were rarely earmarked for healthy diet promotion, such as in the form of healthy food subsidies paid to low-income households (87).Fourth, relatively arbitrary classifications of SSBs result in taxing some low-sugar beverages but exempting some high-sugar ones (88).The beverage industry has used these policy loopholes to lobby against soda tax bills (88).Finally, some researchers questioned the efficiency of the volumetric soda tax and suggested taxing the amount of sugar in a drink rather than the volume of liquid that accompanied the sugar to boost the soda tax's health benefits and overall economic gains (89).
Despite the various criticisms of soda taxes, it remains a potentially effective policy leverage to nudge people toward reducing purchases and consumption of SSBs.There is also preliminary evidence linking soda tax to reduced body weight, BMI, and overweight or obesity (14,15).The revenues collected from soda taxes could support other health interventions, such as nutrition education campaigns or healthy food subsidies, which may reinforce the long-term sustainability of soda taxes on dietary pattern changes.Such policy arrangements might also alleviate health disparities at the population level.
This study comprehensively reviews evidence reported from natural experiments on soda taxes implemented across US local taxing jurisdictions.The strengths of the included studies encompassed realworld policy interventions providing causal inferences, the inclusion of comparison cities, an elucidation of the local soda tax implementation processes, and longitudinal changes in price, consumption, and purchases pre-and post-tax.However, a few limitations of this review and the included studies should be noted.First, all studies were natural experiments focused on one or a few US cities implementing a soda tax.The limited geographic coverage and nationally non-representative sample have confined the generalizability of study findings.Therefore, caution should be exercised when extrapolating the results to other cities or jurisdictions.Second, DID is a quantitative method to reveal causal references, but the resulting evidence should not be overstated due to lacking a randomization design (82).Third, it is important to consider the magnitude of the tax rates examined in the included studies.The range of tax rates observed in the reviewed studies was relatively narrow, typically between 1 and 2 ¢/oz.Thus, the effects reported in the included studies may not necessarily apply to different tax magnitudes.Non-linearities in the effects on tax pass-through or purchases may exist, and these non-linearities could be influenced by factors such as company profit considerations or consumer price sensitivity.Additionally, we acknowledge that the slight increase in prices of untaxed beverages following soda tax implementations is an interesting finding.However, the reasons behind this increase are not entirely clear from available evidence.It could be a result of temporal changes in the prices of untaxed beverages or it may reflect substitution effects, where higher demand for untaxed beverages leads to a price increase.Further research is needed to understand the underlying factors contributing to this observation.

Conclusion
This study systematically reviewed evidence from natural experiments regarding the impact of soda taxes on beverage prices, sales, purchases, and consumption in the US.Soda tax implementation was associated with increased prices of taxed beverages and reduced purchases and consumption of taxed beverages.Soda taxes could be effective policy leverage to nudge people toward purchasing and  For each criterion, a score of one was assigned if "yes" was the response, whereas a score of zero was assigned otherwise.A study-specific global score ranging from 0 to 10 was calculated by summing up scores across all criteria.The study quality assessment helped measure the strength of scientific evidence but was not used to determine the inclusion of studies.
consuming fewer SSBs.Future research should examine evidencebased classifications of SSBs, more targeted use of revenues generated by taxes to reduce health and income disparities, and the feasibility of redesigning the soda tax to improve its efficiency.
soda tax, Falbe et al. documented that the average price of SSBs increased by 0.92 ¢/oz.(95%CI = 0.28, 1.56; p < 0.01) in Oakland 10 months post-tax and 1 ¢/oz.(95%CI = 0.35, 1.65; p < 0.01) in San Francisco 4 months post-tax in comparison to Richmond and San Jose (70).Studies reported that the taxed beverage prices increased by 0.49 ¢/oz.(77), and the regular soda price increased by 0.82 ¢/oz.(72) in Oakland relative to Sacramento one year post-tax.The taxed beverage prices increased by 0.67 ¢/oz.(80), and bottled regular soda prices increased by 1.44 ¢/oz.(78) in Oakland relative to Sacramento two year post-tax.Powell and Leider found that the taxed beverage prices increased by 1.13 ¢/oz. in Cook County relative to St. Louis (73).Powell et al. found the prices of all taxed beverage types increased by 1.19 ¢/oz. in Cook County relative to St. Louis (75).In Philadelphia, a 1.5 ¢/oz.soda tax raised prices by 1.58 ¢/oz.for all taxed beverages combined relative to comparison untaxed stores outside Philadelphia (39).Coary et al. found that SSB prices increased by 1.53 ¢/oz.relative to those outside the city (43

3 ,
152 prices for 2016 2. 2,763 prices for 2017 Prices from both taxed beverage and untaxed beverage DID Changes in beverage prices Data collected in person at retail stores (Continued)

2 .
No statistically significant change in prices was found in either time period for taxed and untaxed fountain drinks and untaxed bottled diet soda.24 Volume sold of taxed beverages in Cook County compared with St Louis exhibited a posttax implementation level decrease of 25.7% (β = −0.297;95%CI = −0.415,−0.179) and a posttax repeal level increase of 30.5%(β = 0.266, 95% CI = 0.124, 0.408), with no net change in volume sold from pretax to 8 months after repeal.Compared with St Louis, posttax implementation in Cook County resulted in a level increase in taxed beverage prices of 1.13 ¢/oz.(95% CI = 1.01, 1.25), followed by a posttax repeal level decrease of −1.19 ¢/oz.(95%CI = −1.33,−1.04), with no resulting change pretax to posttax repeal.(Continued)

43 . 1 . 2 .
0%) relative to those residing in untaxed cities. of a soda tax on consumption Effects of a soda tax on sales or purchases Effects of a soda tax on prices 29 Purchases of taxed beverages declined by 6.1 fl oz.(95% CI = −9.9,−2.4; p < 0.001), corresponding to a 42% decline in Philadelphia compared with Baltimore; there were no significant changes in purchases of nontaxed beverages.Although there was no significant moderation by neighborhood income or customer education level, declines in taxed beverage purchases were larger among customers shopping in low-income neighborhoods (−7.1 fl oz.; 95%CI = −13.0,−1.1; p = 0.001) and individuals with lower education levels (−6.9 fl oz.; 95% CI = −12.5,−1.3; p = 0.001).Taxed beverage prices increased 2.06 ¢/oz.(95% CI = 1.75, 2.38; p < 0.001), with 137% of the tax passed through to prices 2 years after tax implementation in Philadelphia compared to Baltimore 30

31 1 .
Intake of soda reduced by 0.81 servings per week in Philadelphia compared with all other comparison cities 2 years after tax implementation.

37 1 .
The total volume of taxed beverages per store sold in Philadelphia decreased by 46%.2.A large amount of cross-shopping to stores outside of Philadelphiaoff-sets more than half of the reduction in sales in the city and reduces the net decrease in sales of taxed beverages to only 22%.

39 1 .
Taxed beverage volume sold decreased by 18% in Oakland relative to Sacramento, with a larger decrease for family-size beverages (23%) relative to individual-size beverages (8%).

1 . 3 .
Did the study collect and analyze both pre-and post-tax outcomes for the intervention group (i.e., retailers or consumers residing in the taxing jurisdiction)?Were the intervention and control groups similar in all aspects except for their soda tax implementation status?If not,

TABLE 1
(76)eiler et al. reported that the soda tax passed through at a rate of 97%, leading to a price increase by 1.45 ¢/ oz.relative to the surrounding area of Philadelphia(50).Bleich et al. reported that the relative price of taxed beverages increased by 1.81 ¢/ oz.1-year post-tax (69) and 2.06 ¢/oz.two-yearpost-tax(76)inPhiladelphia compared to Baltimore, resulting in a tax passthrough rate of 120.4% (69) and 137% (76), respectively.Petimar et al. reported that taxed beverage prices increased by 1.02 ¢/oz., with 68% of the tax passed through to prices 2 years after tax implementation in Philadelphia compared to Baltimore (81).Cawley et al. reported that Characteristics of soda taxes assessed in the studies included in the review.

TABLE 2
Sample characteristics, sample size, and measures of the studies included in the review.
(49)nd, soda taxes reduced SSB sales and purchases in the taxing jurisdictions relative to the nontaxing jurisdictions.SSB sales in Berkeley declined by 2.7% relative to neighboring cities(51).The total sales volume of taxed beverages decreased by 48.7%(49), or 38.9% (69) in Philadelphia relative to Baltimore following soda tax implementation.Similarly, the volume sold of taxed beverages decreased by 27% (74) or 25.7% (73) in Cook County relative to St. Louis.Sales volume of taxed SSBs reduced by 14% (77) or 18% (80) in Oakland compared to Sacramento.Taxed beverage volume sales decreased by 50% in Philadelphia relative to Baltimore two years post-tax (81).Taxed beverage sales fell by 22% in Seattle relative to Portland (79).Compared with the surrounding metropolitan area, sales quantity of taxed SSB fell by 10.6% in Berkeley, 14.4% in Oakland, 5.1% in Boulder, 8.3% in Philadelphia, 10.7% in Cook County, and 5.6% in Seattle, respectively (63).Concerning purchases, taxed beverage purchases declined by 6.1 fl oz., corresponding to a 42% decline in Philadelphia relative to Baltimore (68)price of taxed beverages increased by 0.83 ¢/oz.relative to untaxed beverages, and 55.3% of the tax was passed through to consumers at the Philadelphia International Airport(68).In Seattle, a 1.75 ¢/oz.soda tax raised prices of taxed beverages by 1.03 ¢/oz.relative to Portland, corresponding to a 59% tax passthrough rate (48).Jones-Smith et al. found that the soda tax was associated with a price increase of 1.58 ¢/oz.relative to its comparison area south of Seattle, and 90% of the tax was passed through to consumers (71).
Roberto et al. estimated that approximately a quarter of the decrease in the taxed beverage sales in Philadelphia could be offset by increases in sales in bordering areas (49).Seiler et al. estimated that cross-shopping to stores outside Philadelphia offset over half of the reduction in SSB sales in the city (50).Petimar et al. estimated that 30% of the reduction in taxed beverage sales in Philadelphia was offset by cross-border shopping

TABLE 3
Estimated effects of a soda tax on price, consumption, purchases, or sales of beverages in the studies included in the review.For small-size beverages (≤ 33.8 oz), price increases in Berkeley relative to those in comparison cities were 0.69 ¢/oz.(95% CI = 0.36, 1.03) for soda, 0.47 ¢/ 3 1.Across all brands and sizes of products examined, 43.1% (95% CI = 27.7,58.4%) of the Berkeley tax was passed on to consumers.2. For each mile of distance between the store and the closest store selling untaxed SSBs, pass-through rose 33.3% for 2-liter bottles and 25.8% for 12-packs of 12-oz cans.4 Households living in Alameda County increased their consumption of sugar-sweetened beverages by 8.89 oz.relative to other households in the U.S., increased their consumption of soda by 26.56 oz.relative to other households in the U.S., and by 37.15 oz.relative to residents of San Mateo, a neighboring county.

TABLE 3 (
Continued) 3. Changes in consumption inBerkeley were significantly different from those in the comparison group, which had no significant changes.151.In Philadelphia and Baltimore, taxed beverage volume sales in 4 weeks decreased in all store types.2.Compared with Baltimore, there were significantly more substantial declines in the volume of taxed beverages sold in the after-tax period in Philadelphia.Total volume sales of taxed beverages in Philadelphia decreased by 1.3 billion oz.(from 2.475 billion to 1.214 billion) or by 51.0% following tax implementation.1.In Philadelphia and Baltimore, the mean price per oz. of taxed beverages increased at all stores in the after-tax periods.2.Compared with Baltimore, Philadelphia experienced significantly greater increases in taxed beverage prices.For supermarkets, the prices increased 0.65 ¢/oz., 95% CI = 0.60, 0.69; For mass merchandize stores, the price increased 0.87 ¢/oz., 95% CI = 0.72, 1.02;For pharmacies, the price increased 1.56 ¢/oz., 95% CI = 1.50, 1.62.16 The tax was fully passed through to consumers via higher retail prices.2.For all taxed beverages combined, the 1.5 ¢-per-oz tax raised prices by1.582 3. Diet soda exhibited a higher price increase in taxed cities. 21 1.The soda tax was associated with an average price increase of 1.58 ¢/oz.among Seattle retailers.Nearly the full cost of the tax was passed through to consumers in Seattle.

TABLE 3 (
Continued) Taxed beverages sales fell by 22% in Seattle relative to Portland.

TABLE 3 (
(79)inued) Leider et al. estimated that two-thirds of the overall decrease in the volume sold of taxed beverages in Oakland was offset by crossborder shopping(80).Léger and Powell estimated that 46% of Oakland's taxed beverage sales decrease was offset by cross-border shopping(77).Powell and Leider reported that there was no significant change in taxed beverage sales in Seattle's two-mile border area, suggesting there was no cross-border shopping for taxed beverages(79).

TABLE 4
Results from meta-analyses and publication bias tests.

TABLE 5
Study quality assessment.