Ars.els-cdn.com



Online AppendixAppendix Table 1. Characteristics of Exit Interview RespondentsPhiladelphiaComparison Communities20162017Difference20162017DifferenceAge39.9739.67-0.3038.3238.890.58(0.95)(0.88)Female0.540.600.060.560.570.01(0.04)(0.04)Hispanic0.270.18-0.090.150.190.04(0.03)(0.03)African-0.560.610.050.590.53-0.06 American(0.04)(0.04)White0.150.170.020.230.260.03(0.03)(0.03)Other Race or 0.280.22-0.060.180.210.03 Multi-Racial(0.04)(0.03)Household 4.204.270.074.504.40-0.10 Size(0.14)(0.12)Income below0.710.720.010.710.720.01 185% FPL(0.04)(0.03)Observations6007631,3637057381,443Notes: The figures represent the mean characteristics of the exit interview respondents at Philadelphia stores and stores in comparison communities prior to the tax (2016) and after implementation of the tax (2017). Standard errors, in parentheses, are estimated from linear regressions with no covariates.Appendix Table 2. Characteristics of Household Survey Respondents PhiladelphiaComparisonCommunitiesDifferenceAge (adult)39.5837.861.73(1.61)Age (child)8.147.900.24(0.62)Female (adult)0.660.610.05(0.07)Female (child)0.500.57-0.07(0.07)Hispanic0.240.180.06(0.06)African-American0.660.630.03(0.07)White0.180.24-0.06(0.05)Other Race0.160.130.04(0.05)Household Size4.284.61-0.33(0.27)Household income below 0.610.540.07 poverty level(0.07)Greater than a High School 0.480.470.01 Education(0.07)SNAP Recipient0.560.450.10(0.07)WIC Recipient0.160.20-0.04(0.06)Free or Reduced-Price Lunch0.320.42-0.10 Recipient(0.07)Observations241199440Notes: This table shows the mean characteristics of adults and children living in Philadelphia and in the comparison communities who responded to both waves of the household survey and the differences between the two groups. Standard errors, in parentheses, are estimated from linear regressions with no covariates.Appendix Table 3. Characteristics of Exit Interview and Household Survey Respondents in Comparison to Philadelphia Residents in the American Community SurveyExit InterviewSurveyACS (Households with Children)ACS (All Households)Age (adult)39.9739.5838.6745.04(0.73)(1.02)(0.22)(0.05)Age (child)8.148.11(0.44)(0.68)Female (adult)0.540.660.600.54(0.03)(0.05)(0.01)(0.00)Female (child)0.500.49(0.05)(0.01)Hispanic0.270.240.190.13(0.03)(0.04)(0.01)(0.000)African-American0.560.660.440.40 (0.03)(0.04)(0.01)(0.00)White0.150.180.350.43(0.02)(0.03)(0.01)(0.00)Other Race or 0.280.160.210.17 Multi-Racial(0.03)(0.04)(0.01)(0.00)Household Size4.204.283.892.27 (0.11)(0.18)(0.04)(0.02)Low-Income0.710.610.540.43(0.03)(0.05)(0.01)(0.01)Greater than a High0.480.460.51 School Education(0.05)(0.01)(0.01)SNAP Recipient0.560.430.26(0.05)(0.01)(0.01)Observations6002411,9477,082Notes: Standard errors are shown in parentheses. Means and standard errors incorporate the sampling weights for all samples and the sampling design of the ACS data. The income measures are not identical across samples. For the exit interviews, low-income is defined as having a household income below 185% of the federal poverty guidelines. For the household survey, low-income is defined as having a household income below 100% of the federal poverty guidelines. For the ACS data, low-income is defined as a household income below 200% of the federal poverty guidelines. In the ACS data, 28.1 percent of households with children have income below 100% of the federal poverty guidelines. For income, due to missing data, the sample sizes in the ACS data are 922 households with children and 3,938 total households.Appendix Table 4. Mean Consumption Before and After Implementation of the Beverage Tax in Philadelphia Residents of PhiladelphiaResidents of Comparison CommunitiesDiD20162017Diff.20162017Diff.AdultsAdded Sugars41.3937.35-4.0435.7241.425.69-9.73(4.18)(3.76)(5.62)[214][214][214][178][178][178][392]Monthly FrequencyTaxed 54.0750.48-3.5945.7150.234.52-8.11??? Beverages(5.60)(5.14)(7.60)[222][222][222][182][182][182][404]Regular Soda34.4127.18-7.2223.8834.1610.28-17.50(5.97)(4.56)(7.51)[232][232][232][193][193][193][425]Diet Soda3.972.58-1.395.884.84-1.04-0.35(1.07)(1.71)(2.02)[234][234][234][192][192][192][426]Untaxed161.30165.974.67141.90161.9820.09-15.42??? Beverages(13.16)(13.41)(18.78)[183][183][183][163][163][163][346]Bottled Water61.7976.9515.1656.9171.6314.720.44(6.54)(6.68)(9.35)[224][224][224][187][187][187][411]Coffee28.2532.103.8528.6833.294.61-0.76(2.97)(3.63)(4.69)[224][224][224][188][188][188][412]ChildrenAdded Sugars19.0816.82-2.2621.4119.80-1.61-0.65(3.20)(2.70)(4.19)[189][189][189][170][170][170][359]Monthly FrequencyTaxed 37.7833.11-4.6742.7737.82-4.950.28??? Beverages(6.05)(4.11)(7.32)[200][200][200][171][171][171][371]Regular Soda16.0611.94-4.1216.4311.82-4.610.47(4.12)(2.12)(4.64)[215][215][215][183][183][183][398]Sweetened 21.8619.65-2.2121.7820.82-0.97-1.24??? Fruit drinks(3.99)(3.65)(5.40)[211][211][211][176][176][176][387]Untaxed172.84171.98-0.86154.32173.5619.23-20.09??? Beverages(11.24)(15.20)(18.90)[180][180][180][153][153][153][333]Bottled water55.6857.481.8041.8355.7113.87-12.08(6.91)(8.39)(10.87)[210][210][210][176][176][176[386Milk62.0054.56-7.4451.4454.242.80-10.24(5.13)(7.03)(8.70)[207][207][207][175][175][175][382]Notes: The figures are the unconditional mean values for the outcome variables (frequency of consumption per month by beverage type and grams of added sugars) for adults and children by time and location. Standard errors, in parentheses, are estimated from linear regressions with no covariates. Sample sizes are in brackets. DiD = difference-in-differences.Appendix Table 5. Unweighted Estimates of the Impact of the Beverage Tax on ConsumptionAdultsChildrenGrams of Added Sugars-4.04-0.46 (2.77)(2.44)[392][359]Monthly FrequencyAll Taxed Beverages-3.860.47(4.55)(4.22)[404][371]Regular Soda-7.800.29(3.07)(2.32)[425][398]Diet Soda0.120.31(1.27)(0.91)[426][406]Sweetened fruit drinks2.900.96(2.65)(2.31)[418][387]Energy Drinksn.a.-0.53(0.59)[213]Untaxed Beverages11.70-3.04(11.31)(11.01)[346][333]Bottled Water7.920.81(5.45)(4.85)[411][386]Coffee2.80n.a.(3.52)[412]Juice6.25-0.91(2.88)(3.11)[414][387]Milk6.43-0.55(4.28)(5.42)[398][382]Notes: All specifications control for gender, age, race, education, and household income. The estimates represent the change in the outcome for households in Philadelphia relative to those in the comparison communities. Standard errors are in parentheses. Sample sizes are in brackets. The outcomes are grams of added sugars and the frequency of monthly consumption by beverage type. All regressions were estimated without survey weights. Standard errors are clustered at the store level.n.a. = not applicable. We did not ask adults separate questions about consumption of energy drinks; energy drinks are included in the questions about sweetened fruit drinks. We collected information about coffee consumption among children, but very few responded that they drink any coffee. Appendix Table 6. Additional Estimates of the Impact of the Beverage Tax by Subgroups?Grams of Added Sugars per Day?Frequency of Regular Soda Consumption per MonthAdultsChildrenAdultsChildrenIndividual CharacteristicsHousehold Member Receives SNAP -4.281.84-5.143.68 Benefits(7.53)(4.81)(10.90)(4.77)[386][355][419][393]43.0320.6234.1017.30Household Member Receives WIC -6.571.25-12.996.80 Benefits(9.00)(6.49)(10.79)(5.74)[388][357][420][396]36.1314.2322.7210.38Sample Adults Have Greater than a 0.25-1.27-6.98-2.42 High School Education(6.17)(3.35)(5.93)(3.19)[392][359][425][398]24.4516.5618.8311.96Child is Over Age 10 n.a.-8.19n.a.-1.72(9.15)(5.89)[359][398]36.4824.01Attitudes Toward the Tax and SSBsSample Adult Supports Tax at 2.87n.a.4.18n.a. Baseline(8.42)(18.55)[341][371]32.6425.85Sample Adult Agrees that SSBs are -8.42n.a.-12.47n.a. Bad for Health(3.56)(4.88)[390][423]37.7227.32Sample Adult is Trying to Drink -4.31n.a.-10.81n.a. Fewer SSBs(4.65)(6.14)[392][425]37.3126.56Cross-border shoppingNever Shops Outside Philadelphia-12.94n.a.-9.21n.a.(10.37)(13.51)[245][266]34.8227.58Lives Far from Philadelphia Border -7.8211.81-16.8011.48 (Top Quartile Based on Distance)(7.76)(4.47)(8.44)(5.40)[384][352][414][390]55.899.4148.515.14Notes: All regressions control for baseline consumption, gender, age, race, education, and household income. The mean of the dependent variable is shown in italics below the sample size in brackets. We report the impact for the subgroup as the sum of the coefficient for Phil and the coefficient for the interaction of Phil and the subgroup indicator in equation (1). For additional notes, see Table 5. Appendix Figure 1. Impact of the Tax on Monthly Consumption of Regular Soda (Adults)Notes: The solid line shows the impact of the tax for various levels of baseline monthly consumption of regular soda. The shaded portion represents the 95 percent confidence interval for the impacts.Appendix Figure 2. Impact of the Tax on Monthly Consumption of Regular Soda (Children)Notes: The solid line shows the impact of the tax for various levels of baseline monthly consumption of regular soda. The shaded portion represents the 95 percent confidence interval for the impacts. ................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download