RepoRt

The rising cost of a healthy dietChanging relative prices of foods in high-income and emerging economies

Steve Wiggins

Sharada Keats

with Euna Han

Satoru Shimokawa

Joel Alberto Vargas Hernández

Rafael Moreira Claro

Report May 2015

Overseas Development Institute203 Blackfriars RoadLondon SE1 8NJ

Tel. +44 (0) 20 7922 0300 Fax. +44 (0) 20 7922 0399 E-mail: info@odi.org.uk

www.odi.org www.odi.org/facebook www.odi.org/twitter

Readers are encouraged to reproduce material from ODI Reports for their own publications, as long as they are not being sold commercially. As copyright holder, ODI requests due acknowledgement and a copy of the publication. For online use, we ask readers to link to the original resource on the ODI website. The views presented in this paper are those of the author(s) and do not necessarily represent the views of ODI.

© Overseas Development Institute 2015. This work is licensed under a Creative Commons Attribution-NonCommercial Licence (CC BY-NC 3.0).

ISSN: 2052-7209

Cover photo: Cover photo credit and caption

developmentprogress.org

• Inhighincomecountriesoverthelast30yearsitseemsthatthecostofhealthyitemsinthediethasrisenmorethanthatoflesshealthyoptions,therebyencouragingdietsthatleadtoexcessweight.

• Itseemsthesamemayapplyinemergingeconomies,suchasBrazil,China,KoreaandMexico,wherepricesoffruitandvegetableshavebeenrisingmorethanmostotherfoods,includingenergy-denseprocessedfoods.

• Astrongcaseemergesforusingtaxesandsubsidiestooffsetthesechangestoencouragemoreconsumptionofhealthyfoodsandlessofunhealthyitems.

Key messages

Contents

Abbreviations & Glossary 6

Acknowledgements 6

Summary 7

1. Introduction 11

2. Research questions and objective 16

3. What is already known? 20

3.1 United States 21

3.2 United Kingdom 29

3.4 Brazil 34

3.5 China 36

3.6 Republic of Korea 36

3.7 Mexico 37

3.8 Summary of literature from emerging economies 38

4. Data and methods 39

5. Results of analysis: price movements 42

5.1 Brazil 43

5.2 China 44

5.3 Republic of Korea 45

5.4 Mexico 46

5.5 United Kingdom 47

5.6 Comparative analysis 48

6. Conclusion and discussion 50

6.1 Changing prices 51

6.2 Do prices matter … and might taxes work? 51

References 54

4 ODI Report

The rising cost of a healthy diet: Changing relative prices of foods in high-income and emerging economies 5

Figures

Figure A: Estimated average annual price changes from 1990 10

Figure 1A1: Foresight’s causal map of overweight and obesity clustered by type of driver 13

Figure 1A2: Extract from Foresight’s causal map of overweight and obesity 13

Figure 1.1: Prevalence of adult overweight and obesity in four countries plus the UK and USA, 1980–2008 14

Figure 1B1: Average diets, six countries, 2010, standardised scores 16

Figure 1B2: Changes in consumption of healthy and unhealthy foods, six countries, 1990–2010, changes to

standardised scores 16

Figure 2A1: Energy density of selected healthy and less healthy snacks, USA, kcal per 100 grams 19

Figure 3.1: Prevalence of overweight and obesity, USA, by age and sex, 2011–12 22

Figure 3.2: Relationship between cost of per unit of energy and energy density for 372 foods in Seattle area

supermarkets, 2006 23

Figure 3.3: Energy density and price per calorie for selected snack foods, fruit and vegetables, USA, 2008 23

Figure 3.4: Price indices for selected foods and beverages in the USA, 1980–2010 24

Figure 3.5: Long-term average annual changes in retail prices for selected foods in the USA 25

Figure 3.6: Fruit and vegetables and ‘junk’ food, unit values and corrected prices relative to total food 30

Figure 3.7: Contribution of different food groups to daily diet costs of healthiest and least healthy group of eaters, adult

women, UK, 1995–98 31

Figure 3.8 Diet costs (per calorie) and energy density, Scotland, 2011 32

Figure 3A1: Kcal changes associated with adjusting meals to healthier options 33

Figure 3A2: Cost of different diets from three types of shop 33

Figure 3.9: Proportion of energy from different food groups, urban household purchases, Brazil, 1996–2009 35

Figure 3.10: Cost per kilocalorie for different food groups in Brazil 36

Figure 3.11: Cooking oil prices, China, 1991–2006 37

Figure 3.12: Vegetable consumption in Republic of Korea, 1969–2009, grams per person per day 38

Figure 3.14: Consumption of soft drinks per person per day, USA, UK, Mexico, Brazil, Korea and China, 2008 38

Figure 3.13: Annual sales of selected ultra-processed foods and drinks, 12 Latin American countries, 1999–2013 38

Figure 5.1: Price of selected foods, São Paulo, Brazil, 1980–2009 44

Figure 5.2: Price of selected foods, China, urban areas, 1989–2006 45

Figure 5.3: Prices of selected foods, Republic of Korea, 1975–2013 46

Figure 5.4: Prices of selected foods, Mexico, 1980–2014 47

Figure 5.5: Prices of selected foods, United Kingdom, 1974–2012 48

Figure 5.6: Estimated average annual price change from 1990 50

Tables

Table 2.1: Causal chain from costs of food to consumption, obesity and ill health 18

Table 3.1: US own-price elasticity of demand, by food and beverage category, 1938–2007 26

Table 3.2: Own-price elasticity of demand for four food groups in the USA, 2007–12 26

Table 3.3: Estimates of food price elasticity in the UK, long-run 32

Table 4.1: Food prices analysed for each food group and country 41

Table 4.2: Data sources and deflators used for country cases 41

Table 5.1: Results of log-linear regression of time on prices 49

Boxes

Box 1A: Mapping drivers of overweight and obesity: spaghetti junctions ahead 13

Box 1B: Changing diets in the four countries, UK and USA 16

Box 2A: Energy density of foods 19

Box 3A: Cheap eats and healthy treats: affordable for families of obese children in the UK? 33

Box 3B: Marketing of processed food in Brazil 35

Box 6A: How feasible are taxes on foods? 54

6 ODI Report

The rising cost of a healthy diet: Changing relative prices of foods in high-income and emerging economies 7

Abbreviations & GlossaryBLS US Bureau of Labour Statistics

BMI Body Mass Index, body mass compared to square of height, kg/m2. BMI of 25 or more indicates overweight, 30 or more indicates obese

CPI Consumer Price Index

Eatwell Plate UK government guide to the composition of a healthy diet

Energy density A measure of calories per unit of food

FAFH Food away from home – food consumed outside people’s homes as in restaurants etc.

FAO Food and Agriculture Organizsation of the United Nations

GDP Gross Domestic Product

HIC High-Income Country

LIC Low-Income Country

MIC Middle- Income Country

Nutrient density A measure of how many ‘healthy’ nutrients are in a food, usually by weight

Potato chips US term, the same as UK potato crisps

Prices, constant or real Nominal prices from which inflation in general price levels have been removed, either by use of a consumer price index or GDP deflator

Prices, indices Prices expressed as ratio of those at a base year, when values are usually set to 100. Indices may or may not be adjusted for inflation. The price indices constructed in this paper have been adjusted for inflation.

Prices, nominal Prices as observed at a given time, with no adjustment for inflation

SNAP Supplemental Nutrition Assistance Program (USA)

SSB Sugar-sweetened beverage (or ‘soda’ in US English)

UMIC Upper Middle Income Country

USDA United States Department of Agriculture

WCRF World Cancer Research Fund

AcknowledgementsManythankstothefollowingfortheircontributiontothisreport:

• ForChina,thereportusesdatafromtheChinaHealthandNutritionSurvey(CHNS),sowethanktheNationalInstituteofNutritionandFoodSafety,ChinaCenterforDiseaseControlandPrevention,CarolinaPopulationCenter(5R24HD050924),theUniversityofNorthCarolinaatChapelHill,theNIH(R01-HD30880,DK056350,R24HD050924,andR01-HD38700)andtheFogartyInternationalCenter,NIHforfinancialsupportfortheCHNSdatacollectionandanalysisfilesfrom1989to2011andfuturesurveys,andtheChina–JapanFriendshipHospital,MinistryofHealthforsupportforCHNS2009;

• ForKorea,HeejoKuandJae-eunMyungprovidedexcellentresearchassistance; • ForMexico,DraAlejandraDonajíBenítezArciniega(UAEMexMedicineFaculty)providedvaluablesupport; • FumiakiImamura(Cambridge)andcolleagueskindlysupplieddataonchangingdietsinthefourcountries; • CorinnaHawkes(WCRF)peerreviewedthereport,andherdetailedandincisivecommentsonanearlierversionhelpedtosharpenthethinking;and,

• DeborahEadecopy-editedthereport.

ThisreportformspartoftheODIproject‘Shockwatch:managingriskandbuildingresilienceinanuncertainworld’,whichissupportedbytheUKDepartmentforInternationalDevelopment(DFID).

Theinterpretationsandopinionsexpressedinthereportarenotnecessarilythoseoftheabovementionedpersons,noroftheOverseasDevelopmentInstitute.Themainauthorsaresolelyresponsibleforanyerrorsandomissions.

Summary

Motivations and questionsIn2014ourpreviousstudy‘FutureDiets’(KeatsandWiggins2014)describedhowacrosstheworldanincreasingshareofthepopulationisoverweightandobese,withtherateofincreaseparticularlypronouncedindevelopingcountries.Nonation,however,hasstemmedtherisingratesofpeoplewhoareoverweightandobese.Effectivepoliciestocombatobesityhaveyettobeproved,ifonlybecausenocountryhasyettestedasufficientlycomprehensivesetofpolicies.Thecausesofexcessweightaremultiple,includingrisingincomes,urbanisationandmoresedentaryoccupations,theinfluenceofmediaandadvertising,andchangingrelativepricesofdifferentfoods.Thislastelementisthefocusofthisreport.

Thereportstartsfromtwoworkinghypotheses:

a.Whentherelativepricesoffoodschange,peoplewillconsumemoreoffoodsthathavebecomerelativelylessexpensive,andlessofthosethathavebecomerelativelymoreexpensive.Peopleonlowincomesareexpectedtobemoresensitivetopricesthanthoseonhigherincomes;and,

b.Whenconsumptionoffoodswithhighcaloriecontentperunitweight(energy-densefoods)increasesattheexpenseoffoodthatislessdenseinenergy,wemayexpecttoseeasignificantincreaseintheprevalenceofoverweightandobesepeople.

Ifthesehypothesesarecorrect,theywouldsuggestthatusingtaxesandsubsidiestoinfluencedietsislikelytobeeffective.

Tothebestofourknowledge,thereisnoexistingstudythatcomparesthechangingcostsoffoodsfromseparatefoodgroupsacrossasampleofdevelopingandemergingeconomies.Thereportfocusesonfourcountries:threeupper-middle-incomecountries(UMICs)–Brazil,ChinaandMexico;andonehigh-incomecountry(HIC),theRepublicofKorea,whichwasstilladevelopingcountryin1990.Theyhavebeenchosensincetheyrepresentemergingeconomiesthataregrowingfasterthanmostindustrialisedcountries,andwheresince1990significantchangesmaybeexpectedinbothdietsandtherelativepricesoffoods.AnalysisofpricesintheUnitedKingdom(UK)hasbeenaddedtoprovidesomecomparison.TheliteraturefromtheUnitedStates(USA)hasbeenreviewedgiventhelargenumberofstudiesthatreportonthepriceoffood,theeffectonfoodconsumptionandinsomecasestheconsequencesforbodyweight.

Whilechangesinpricesofsomefoodssuchasbananas,beverages,cereals,dairyproduce,edibleoilsandsugaroninternationalandnationalmarketsareregularlyreported,

lessisknownabouttheevolutionofnationalretailpricesoffoodintheformpresentedtoconsumers.Hencethecentralquestionsposedinthisreport:

• Whatchangeshavebeenseenintheretailcostoffoodinthefourcountriessince1990?Aretheresystematicdifferencesintheevolutionofpricesfordifferentfoods,andhencechangesinrelativeprices?

• Inparticular,hasprocessedfoodbecomecheaperrelativetounprocessedstaples,fruitandvegetables,meatanddairyproduce?Thismaybeexpectedsincemuchoftheretailcostofprocessedfoodarisesinmanufacturingandlogistics,wheretechnicaladvanceshavereducedunitcosts,perhapsbymorethanadvancesinfarminghavereducedthecostofagriculturalproduce.

Itwaspossibletoexamineonlyasampleofthemanyfoodsonofferinretailoutlets,theaimbeingtohaveatleastoneexamplefromthefollowingfoodgroups:

• Staples Cereals,rootcrops,legumes • Fruitandvegetables Fruitandvegetables • Meat,fishanddairy Minimallyprocessedanimal productsandmilkproducts

• Oils,fats,andsugar Vegetableoilsandfats, animalfats,sugar

• Highlyprocessedfoods Foodsusuallyproduced byindustrialprocesses

What is already known? Insights from the literaturePublishedstudiesforthe UKandtheUSAfrequentlyreportthefollowing,evenifcontraryfindingsandqualificationscanbefoundinmanyotherstudies:

• Moststudiesfindthathealthierfoodscostmorethanlesshealthyones.Moreoverthiseffecthasincreasedoverthelast30–40years,asenergy-dense,processedfoodshavebecomecheaperrelativetolessenergy-densefruitandvegetables.

• Consequentlyhealthydietstendtocostmorethanlesshealthydiets.Thisisnotinevitablyso:choosingcheaperhealthyitemsandsubstitutingthemforcostlierlesshealthyonesmightbothimprovedietandsavemoney.Butformostconsumers,thiswouldrequireboththeabilitytoseethedistinctions,andthedisciplinetofollowaparticulardiet.

• Althoughitseemsthatsomeenergy-denseprocessedfoodshavebecomenotablycheapercomparedtofruitandvegetables,thenatureofthelatterhavechanged

8 ODI Report

The rising cost of a healthy diet: Changing relative prices of foods in high-income and emerging economies 9

–withhigher-valueprepareditemscommoninfoodoutlets,andalsoavailableallyearround.Takingsuchaddedvalueintoaccountthechangeinrelativepricesmaybelessthanisatfirstapparent.

• Consumptionofmostfoodsrespondstopricechanges,althoughformanyfoodstheresponseisrelativelyinelastic–butfarfromperfectlyso.Thoseonlowincomesaremostlikelytorespondtochangingprices.

Studiesoftheimpactsoffoodtaxesthatoftenseemtotriggertinychangesinconsumptionobscurethesefindings,butthisisbecausemoststudiesobserveormodeltheeffectsofverylowtaxes,5%orlessbeingtypical.

• Cross-priceeffectsmatterinassessingthenutritionaleffectsofpricechanges.Taxesonfatorsaltcontentmayaffectconsumptionofother,complementaryfoodsleadingtolessconsumptionofbeneficialnutrients.Usingtaxrevenuestosubsidisesuchcomplementaryfoodswouldcounterthiseffect.

• Studiesoftheimpactsofchangesinpricesonbodyweightproduceasurprisinglystrongconsensusthathigherpricesofunhealthyoptionsreducebodymassindex(BMI),asdocheaperhealthieroptions.‘Surprising’sincebodyweightistheoutcomeofmanyfactors,yetpriceschangescanbeseentomakeadifference.Thestrongesteffectsareseenamongthoseonlowincomeswhoaremostsensitivetothecostoffood.

Therearefewerstudiesonthefouremergingeconomies,buttheyindicatethefollowing:

• Somestudieslinkchangesindiets,aboveallthoseinvolvingmoreconsumptionofprocessedfoods,toprocessedfoodandcookingoilbecomingcheaperthanotherfoods.

• InLatinAmerica,therisingconsumptionofultra-processedfoodsandsugar-sweetenedbeverages(SSBs)isnotable.Someseethisastheconsequenceofheavymarketingbythelargecorporationsthatmanufacturemuchofthisfoodanddrink.

• ThepossibilityofusingtaxestoreduceconsumptionofprocessedfoodandSSBsisactivelybeingstudied,withmostauthorsseeingthepotentialtosignificantlyreduceconsumption.MexicohasalreadyintroducedtaxesonbothSSBsandenergy-densefood.These,whichcameintoeffectinJanuary2014,willbethefocusofintensescrutinytoseewhateffectstheyhave.

Data and methodsKeydataforthisreportareseriesforfoodpricesfrom1990torecentyearsinthefourcountriesplustheUK.Retailpricesweresoughtforrepresentativefoods–thosefrequentlyconsumed–fromthefoodgroupslistedabove.

Inmostcases,directlyobservedretailpriceswereused.ForMexico,however,afoodpriceindexwasusedandcalibratedtopricelevelsfromobservedpricesinMexico

City.IntheUK,householdsurveysreportedbothspendingandquantityfor330foods.Henceitwaspossibletocreateunitpricespaidbydividingexpenditurebythequantity.

Priceserieshavebeendeflatedbyeithertheconsumerpriceindex(CPI)ortheGrossDomesticProduct(GDP)deflatortoremovetheeffectofinflationandallowcomparisonovertime.Oncepricesweredeflated,indicesofthesesdeflatedpriceswereconstructedtoseehowmuchpricesofdifferentfoodsineachcountryhadchangedsincethesamebaseyear.

Moreformaltestingofpricechangeswascarriedoutbyregressingtimeondeflatedandloggedpricesfrom1990toarecentyear,inmostcases2012.Thisallowedatestofwhetherasignificant(log-linear)trendcanbeinferred,andifso,whattheaverageannualpricechangehasbeen.

ResultsThekeyfindingsfromtheanalysiscomefromestimatedannualpricechanges(seeFigureA).

Twothingsarereadilyapparent.Oneisthatpricesoffruitandvegetableshaverisensubstantiallysince1990,mainlybybetween2%and3%ayearonaverage–orby55–91%between1990and2012.Theotheristhatfourofthesixprocessedproductsforwhichestimatesaresignificantshowpricefallssince1990.Mostoftheotherfoodshaveseentheirpricesriseby1–2%ayear,withtheexceptionofthepricefallsforriceinKoreaandchickeninMexico.

DiscussionIfthedetectedtrendsarerealtheypromptquestionsaboutthereasonsforthem.If,forexample,technicalprogressinfarmingwereuniform,sothatunitcostsofproductionwerefallingforallagriculturaloutput,andifadvancesinthelogisticsoffoodwholesalingandretailingweresimilarlyuniform,thenwemightexpectthecostsofmostfoodstomoveroughlyinlinewithoneanother.Butthatisnotthecase.

Sowhyhavefruitandvegetablesbecomemorecostlycomparedtootheritems?Itisnotasthoughtherehavenotbeentechnicaladvancesinhorticulture:onthecontrarysomeofthemostsophisticatedseeds,soilnutrition,watercontrol,andpreventionofpestsanddiseasesareseenpreciselyinthegardensandglasshousesinwhichsomanyfruitandvegetablesaregrown.WhilethereisaworldofdifferencebetweenDutchheatedglasshousesandthetinyplotsofgreenbeansofcentralKenya,inbothcases,comparedtootheragricultureintheirneighbourhoods,thesesystemsarebothmoreintensiveandusemoresophisticatedtechnologythanmostotherlocalfarmenterprises.Moreover,advancesintransportmeanthatfruitandvegetablesaretradedmorethaninthepast,sothatretailmanagersshouldbeabletosourcefromlow-costsuppliersnomatterwheretheymaybe.

Hypothesescanbeimagined:horticulturemaywellhaveasteppedsupplyfunction,sothatwhilesmallquantitiesoffruitandvegetablescanbesuppliedatlowunitcost,onceaparticularvolumeisreached,costsrapidlyescalatetoasignificantlyhigherlevel.Itmayalsobethatthechangesinqualitynotedexplaintheincreasedrelativeprices.Or,itmaynotbeamatterofcostbutofincreaseddemandfromthoseconsumerswhoappreciatethehealthbenefitsoffruitandvegetables.Thesehypothesesmeritaseparatestudy.

Whydoesnotthesameapplytosomeprocessedfoods?Onepossibilityisthatmuchprocessedfooddoesnotrelyoncostlyfarmingredients,butratherismanufacturedfromrelativelycheapingredients,theaddedvaluebeinglargelyinfactoryprocessesofcombiningtheingredientsandenhancingtheirflavour.Advancesinmanufacturingandflavouringprobablyhelpreduceunitcostsinfactory.Thatsaid,processedfoodsarenotuniforminqualityandpricing,sinceforanysub-category,thereareusuallyproductsthatarebranded,soldontheirspecialcharacteristics,usuallywithapricepremium–asapplies,forexample,toSSBs,whichcompetewithcheaper,unbrandedoptions.Thismayexplainwhynotalltheprocessedfoodsconsideredshowdecliningconstantprices.Again,additionalstudiesmightshedlightonthis.

Evidencepresentedintheliteraturereviewsuggeststhatpricesdoaffectconsumption,especiallyforpeople

onlowincomes.Henceitisnosurprisetoseemuchstudyofthepotentialoftaxesonlesshealthyoptionstoreducetheirconsumption,perhapsevenwithsubsidiesonmorehealthyoptionstoraisetheirs.Mostsuchstudiesindicatethatimposingtaxeswouldreduceconsumption.Buttwoqualificationsapply.

Oneisthattheremaybecross-priceeffects,wherebywhentaxesraisethecostofaparticularfood,notonlydoesitsconsumptionfall,butsotoodoesthatofcomplements(foodswhicharetypicallyconsumedtogether,suchasbreadandbutter).Whenthosecomplementscontainvaluednutrientsitisthuspossiblefortaxestoreducethequalityofdiet.Intheorythisproblemcanreadilybetackledbyplacingasubsidyonthevaluedcomplementtooffsetthecross-priceeffect.Inpractice,learningwhichfoodsreallyarecomplements,towhatextent,andthendetermininganoptimallevelofsubsidy,couldleadtoathicketofregulationsthathavetobeadjustedinthelightofemergingevidence,creatinghighadministrativecostsandgivingtheimpressionthatsuchfiscalmeasuresarejusttoodifficulttocontemplate.Thequestionishowstrongcross-priceeffectsareandwhethertheymayberemediedbyothermeasurestoencouragehealthierdiets.

Theotheristheapparentlyseductiveargumentthatsmalltaxeswouldcreateonlysmalleffects:thatconsiderablechangeinconsumptionwouldrequirehigh

10 ODI Report

Figure A: Estimated average annual price changes from 1990

-4 -3 -2 -1 0 1 2 3 4

GrainDried rice

RiceTortilla and maize flour

Vegetable oilSoy oilSugar

Meat, Poultry and its productsFishBeef

ChickenChicken

VegetablesTomatoOrange

CabbageTomato

Fresh vegetablesFresh green vegetables

Cake, biscuit and breadRamen

Sweet biscuitChocolate and snacks

Ready mealsIce cream tub/block

CHKR

BRM

XKR

BRM

XCH

KRBR

MX

UKCH

BRBR

KRM

XM

XUK

CHKR

BRM

XM

XUK

Staples

Oil and sugar

Meat, poultry, fish

Fruit & vegetables

Processed foods

Source: Table 5.1

The rising cost of a healthy diet: Changing relative prices of foods in high-income and emerging economies 11

taxesthatwouldlookdisproportionateandunfair–say,morethantherateofvalue-addedtax(VAT)of20%intheUK.Butalogicalflawapplies.Thepolicyquestionisnotsomuch,‘howlargeataxwouldbenecessarytobringdownconsumptionoflesshealthyfoodXtorecommendedorinsignificantlevels’,but‘howmuchbenefitwouldbederivedfromimposingapoliticallyacceptabletaxonlesshealthyfoodX?’Theanswertotheformermaybeanumbersohighastobedismissedfromthedebate;buttheanswertothelattermaybeasstrikingasthatprovidedbyNnoahametal.(2009)fortheUK:thattaxesandsubsidiesoflessthan20%couldsavenofewerthan6,400prematuredeathsayearfromcoronaryheartdisease(CHD)andcancers.Theargumentabout‘smalltaxes,

smallgains’istantamounttoargumentsthatcondemndoinggoodbecauseperfectionisunattainable.

Intermsofwhatmightbetaxedandsubsidised,thisreportsuggeststhatenergy-densefoodsmightbetaxed,whilefruitandvegetables,whosepricesoftenrisecomparedtootherfoods,mightbesubsidised.

Muchcomesdowntothepoliticalappetitetocontemplatetaxingfoods.EventsinMexicosuggestthatsomeemergingeconomiesmaystealamarchonHICsinthisrespect.TheevidencepresentedinthisreportsuggeststhattheMexicantaxesshouldachieveconsiderablegood,therebyprovidingvaluablelessonsforotherdevelopingandemergingeconomies.

1. Introduction

The rising cost of a healthy diet: Changing relative prices of foods in high-income and emerging economies 13

Box 1A: Mapping drivers of overweight and obesity: spaghetti junctions ahead

Amultitudeoffactorsinfluenceanindividual’schanceofbeingoverweightorobese.TheUKForesightstudyonobesityidentifiedbroadthematicclustersofSocialandIndividualPsychology,FoodProductionandConsumption,Physiology,IndividualPhysicalActivityandPhysicalActivityEnvironment,outlinedinthickdottedlinesontheirdiagramofinteractions(Figure1A1).

Figure 1A1: Foresight’s causal map of overweight and obesity clustered by type of driver

Source: Map 5 in Butland et al. 2007

Withineachoftheclustersmultiplefactorsinteract.Belowisthefoodconsumptionclusterwheremanyoftheeconomicfactorscomeinto,seeFigure1A2.

Figure 1A2: Extract from Foresight’s causal map of overweight and obesity

Source: Section of Map in Butland et al. 2007

Theprevalenceofpeoplewhoareoverweightandobeseisincreasingacrosstheworld,especiallysoindevelopingcountries(KeatsandWiggins2014;Ngetal.2014;PopkinandSlining2013;Stevensetal.2012).Someemergingeconomiesnowhavealmostthesameprevalenceofoverweightandobesityasseeninhigh-incomecountries(HICs).Thecostsarehigh.Notonlydoesexcessweightmakepeoplemoresusceptibletoheartdisease,somecancers,strokesandtype-2diabetes,butalsoitleadstoeconomiclossesandhighercostsofhealthcare(Gortmakeretal.2011).

Nonationhasstemmedtherisingratesofpeoplewhoareoverweightandobese.Effectivepoliciestocombatobesityhaveyettobeproved,ifonlybecausenocountryhasyettestedasufficientlycomprehensivesetofpolicies.Causesofexcessweightarewidelyagreedtobemultiple,rangingfromthesimpleeconomicsofrisingincomeandfallingcostsofmanyenergy-denseprocessedfoods,whichallowspeopletoeatmore,tomoresedentarylivesassociatedwithurbanisation,tosocio-culturaleffectsofadvertisingandmediaimages,andeventopotentialaddictiontohighlypalatablefoods(Hawkes2008;Kearney2010;Swinburnetal.2011).Oftentheproblemisseenasarisingfromsystemicchange.Asystemsmapofthedriversofoverweightandobesity,developedintheUKgovernment’sForesightprogramme(Butlandetal.2007)(seeBox1A),showsjusthowcomplex(andinterrelated)thedriversarebelievedtobe.

Thisreportaddressesonepartofthesystem:thewayinwhichpricesoffoodinfluenceschoiceofdiet.Ifdietshavetendedtoincludemoreenergy-densefoods,andespeciallyprocessedfood,isthislargelybecausethesefoodshavebecomerelativelycheapcomparedtofoodlessdensein

energyandoftenunprocessed?Theanswerhasimplicationsforpolicy:namelytheeffectivenessofusingtaxesandsubsidiestoinfluencediets,andhencepublichealth.

Tothebestofourknowledge,noexistingstudycomparesthechangingcostsoffoodsforseparatefoodgroupsacrossasampleofdevelopingandemergingeconomies.Althoughasimpleexercise,itcouldthereforebeofconsiderableinteresttoothersstudyingcausesandpotentialpolicyresponsestorisinglevelsofobesity.

Thisreportaimstounderstandbetterhowrelativepricesofdifferentfoodshavechangedsinceatleast1990,orearlierwhendatapermit,andhowthismayhaveledtodietarychanges.PlentyhasbeendocumentedaboutchangingrelativepricesoffoodsforHICs,aswillbeseenintheliteraturereviewofstudiesoffoodpricesintheUSAandtheUK.Less,however,isknownaboutchangesinpricesindevelopingcountries,aboveallinemergingeconomieswheretheprevalenceofoverweightandobesepeopleisgrowingfastest.Thisreportaimstohelpfillthisknowledgegap.Itfocusesonfourcountries:threeupper-middle-incomecountries(UMICs)–Brazil,China,andMexico–andoneHIC,theRepublicofKorea,whichwasstilladevelopingcountryin1990.Theyhavebeenchosensincetheyrepresentemergingeconomies1thataregrowingfasterthanmostoftheindustrialisedcountries,wheresince1990significantchangesmaybeexpectedinbothdietsandtherelativepricesoffoods.AnanalysisofpricesintheUKhasbeenaddedtoprovidesomecomparison.

Eachofthesefourcountrieshasseenarapidincreaseintheburdenofoverweightandobesitysincethe1980s(seeFigure1.1).AlthoughprevalenceisstillrelativelylowintheAsianexamples,ithasgrownmarkedlyfrom

1 By2015theRepublicofKoreaisnotusuallyregardedasanemergingmarket,butwasseenassuchatthestartofthestudyperiodin1990.

14 ODI Report

Figure 1.1: Prevalence of adult overweight and obesity in four countries plus the UK and USA, 1980–2008

45%

39%36%

25%

14% 13%

23%

71%

64%68%

52%

32%25%

34%

15830 45 64 12

241

1,461

0

200

400

600

800

1000

1200

1400

1600

0%

10%

20%

30%

40%

50%

60%

70%

80%

US UK Mexico Brazil South Korea China GLOBAL

Mill

ions

of a

dults

ove

rwei

ght &

obe

se

Perc

ent o

f adu

lt po

pula

tion

1980 2008 Number overweight and obese in 2008 (M) - R axis

Source: Compiled from data in Stevens et al. 2012

Notes: China: Latest estimates for overweight and obesity in adults 20–60 years over 35% for men and over 30% for women in 2011 – see Figures 12 and

13 in Annex III.4. Rapid increases in excess weight for those under age 20 from 2008 to 2013: for girls from 4.5% to 13%, for boys from 7% to 23%. (He

2013). Mexico: between 2002 and 2012 much of the increase in people overweight and obese came from already overweight people becoming heavier.

The rising cost of a healthy diet: Changing relative prices of foods in high-income and emerging economies 15

the1980s.Oftheselectedcountries,Chinahasseenthesmallestproportionalincreaseinoverweightandobesity,thoughthesheersizeofChina’spopulationmeanstheburdenofoverweightandobeseadults,thoughrepresentingonly25%ofthepopulation,outstripsbysome50milliontheburdenintheUKandUSA.

Themedicalcostsofexcessweightarealreadyconsiderableandrising:

• Brazil:annualcostofdiseaserelatedtoexcessweightisestimatedatUS$2.1billion,ofwhichUS$21millionisdirectlyattributabletooverweightandobesity(Bahiaetal.2012).

• China:directmedicalcostsofadultchronicdiseasesattributabletooverweightandobesitywereestimatedataroundYuan21billionin2003[US$2.74billion](Zhaoetal.2008).Economiclossesfromprematuredeathsfromheartdisease,strokes,anddiabeteswereestimatedin2005atUS$18billion(WHOFactsheetat:http://www.who.int/chp/chronic_disease_report/media/china.pdf?ua=1).

• Mexico:DiabeteshasbeenaleadingcauseofdeathinMexicosince2002,followedbycardiovasculardisease(CVD)andbraindisease.Prevalenceoftype-2diabetesinadultsisupfrom6%in2000toover9%in2012(ValdésRamos2012).

Thefourcountrieshavealsoseensignificantchangesintheirdietsforatleastthelast20years,forsimilarreasonstothoseseeninHICs:risingincomes,urbanisation,moresedentaryoccupations,changesincostsoffoodandtheinfluenceofmarketingandmediaontastes.Box1Breportsonhowwhiledietsacrosstheworldhavecometocontainmorehealthyfoods,inthefouremergingeconomiestheyhavealsotendedtoincludemoreunhealthyones.

Therestofthereportissetoutasfollows.Section2statestheresearchquestionsandobjectives,andpresentsthehypothesisedcausalchainfromfoodpricestoobesity.Section3reviewstheliteratureonfoodpricesandtheirrelationtoconsumption.ThisbeginswiththequitelargeUSliterature,reviewedbecausemoreisknownabouttheobesityintheUSAthananywhereelse.FollowingthatsomeoftheliteraturefortheUKandthefouremergingeconomiesisreviewed.Section4recordsthesourcesofdataandthemethodsusedtoanalyseit.Section5presentsthefindings.FinallySection6concludesanddiscussestheresults.

MoredetailedinformationfortheemergingeconomiescanbefoundinAppendixIII.

16 ODI Report

Fromdetailedsurveysofdietsacrosstheworld,dietshavebeenassessedfortheextenttowhichhealthyandunhealthyfoodsareconsumed(Imamuraetal.2015).Thefoodswerecategorisedasfollows:

• Healthyitems:fruits,vegetables,beansandlegumes,nutsandseeds,wholegrains,milk,totalpolyunsaturatedfattyacids,fish,plantomega-3sanddietaryfibre.

• Unhealthyitems:unprocessedredmeats,processedmeats,sugar-sweetenedbeverages(SSBs),saturatedfat,transfat,dietarycholesterolandsodium.

Eachcountrywasgivenastandardised(1–100)scoreaccordingtointakerelativetotheglobalsample.Forunhealthyoptions,higherscoresrepresentlessconsumptionofthese.

Ofthesixcountries,dietsinMexico,theUKandBrazilcontainthemosthealthyoptions,whileChinaandKoreahavetheleast(seeFigure1B1,panelA).Thedifferences,however,arenotthatlarge.Morestrikingarethedifferencesinconsumptionofunhealthyfoods(seepanelB),whereChina,KoreaandMexicohaveleastconsumption,whileBrazilandtheUSAhavethehighest.

Figure 1B1: Average diets, six countries, 2010, standardised scores

(a) Healthy options (b)

Unhealthy options

Source: Fumiaka Imamura, personal communication

Evenmorestriking,however,arethechangesinthesescoresseensince1990(seeFigure1B2).Allcountrieshaveseenanincreaseinconsumptionofhealthyitems,exceptforChina.Forunhealthyfoods,theUKandtheUSAhaveseenlessconsumptionoftheseitems,butalltheemergingeconomieshaveseenrisingconsumptionofunhealthyfoods.

Figure 1B2: Changes in consumption of healthy and unhealthy foods, six countries, 1990–2010, changes to standardised scores

Source: Fumiaka Imamura, personal communication

Insum,itseemsthatinmostofthesixcountries,dietscontainincreasingquantitiesofhealthyitems,offsetinthefouremergingeconomiesbyincreasingquantitiesofunhealthyitems.

0

20

40

60

BRA CHN KOR MEX UK USA

0

20

40

60

80

BRA CHN KOR MEX UK USA

Box 1B: Changing diets in the four countries, UK and USA

-10 -5 0 5 10

BRA

CHN

KOR

MEX

UK

USA

Quality Change,healthy foods

Quality Change,unhealthy foods

2. Research questions and

objective

Thisreportstartsfromtwoworkinghypotheses:

a)Whentherelativepricesoffoodschange,peoplewillconsumemoreoffoodsthathavebecomerelativelylessexpensive,andlessofthosethathavebecomerelativelymoreexpensive.Peopleonlowincomesareexpectedtobemoresensitivetopricesthanthoseonhigherincomes;and,

b)Whenconsumptionoffoodswithhighcaloriecontentperunitweight(energy-densefoods)increasesattheexpenseoffoodlessdenseinenergy,thenwemayexpecttoseeasignificantincreaseintheprevalenceofoverweightandobesepeople.

ThecausalchainthatrunsfromfoodcoststoconsumptiontoobesityandillhealthissetoutinTable2.1.Thelogicrunsfromfactorsthataffectfoodprices–bothchangesincoststhataffectthemdirectlyandotherfactors–tochangesinfoodpriceswhich,togetherwithfactorssuchasincomes,leadtochangesinthepurchaseandconsumptionoffood,whichtogetherwithotherfactorssuchasphysicalactivity,leadtochangesinbodyweightthatfinallyleadtoillhealth.ThischainhasbeenusedtostructurethereviewofliteraturefortheUSA,wheresomelimitthemselvestolevel1,otherstrytoconnect1to2,orto2,3and4orgostraightfrom1to3or4.

Whilechangesinthepricesofsomefoodssuchasbananas,beverages,cereals,dairyproduce,edibleoilsandsugaroninternationalandnationalmarketsareregularlyreported,lessisknownabouttheevolutionofretailpricesinnationalmarketsforfoodintheformpresentedtoconsumers.Hencethecentralquestionsposedhereare:

• Whatchangeshavebeenseenintheretailcostoffoodinthefourcountriessince1990?Aretheresystematicdifferencesintheevolutionofpricesfordifferentfoods,andhencechangesinrelativeprices?

• Inparticular,hasprocessedfoodbecomecheaperrelativetounprocessedstaples,fruitandvegetables,meatanddairyproduce?Thismaybeexpectedsincemuchoftheretailcostofprocessedfoodarisesinmanufacturingandlogistics,wheretechnicaladvanceshavereducedunitcosts,perhapsbymorethanadvancesinfarminghavereducedthecostofagriculturalproduce.

Itwaspossibletoexamineonlyasampleofthemanyfoodsonofferinretailoutlets,theaimbeingtohaveatleastoneexamplefromthefollowingfoodgroups:

• Staples Cereals,rootcrops,legumes • Fruitandvegetables Fruitandvegetables • Meat,fishanddairy Minimallyprocessedanimal productsandmilkproducts

• Oils,fats,andsugar Vegetableoilsandfats, animalfats,sugar

• Highlyprocessedfoods Foodsusuallyproduced byindustrialprocesses

Thegroupshavebeenconstructedlargelyfortheirnutritionalcharacteristics,withstaplesbeingaprimesourceofenergy;fruitandvegetablesfortheirvitaminsandminerals;meat,fishanddairyforprotein;and,oils,fatsandsugarsforenergyinparticularlyconcentratedform.Totheseconventionalgroupshavebeenaddedhighlyprocessedfoods,thatisthosethat

Table 2.1: Causal chain from costs of food to consumption, obesity and ill health

Level Direct links Other factors

0 Increases in productivity … in• Agriculture;• Foodprocessing;and,• Foodtransport,storage,distributionandretailing… that reduce unit costs of food.

Publicpolicy:Taxes,SubsidiesPricingasmarketingstrategy–offers,lossleaders,etc.

1 Changeinfoodprices,constant,netofinflation• Absolute,comparedtoothergoodsandservices• Relative,onefoodorgroupoffoods,comparedtoanother

IncomesTransfers–foodstampsPreferences–influencedbyadvertising,publiceducation,media,peersAvailability–e.g.food‘deserts’

2 Changein(purchaseand)consumption Geneticandphysicalfactors–abilitytouseenergy,storefatPhysicalactivity–influencedbywork,transport,sportsandleisureoptions,etc.

3 Changeinweight(BMI,fatcomposition) Remedialmedicalinterventionssuchasprescribeddrugs

4 Diseaseandillheath:• Prematuredeath,disability,illness• Costsofhealthcare• Economiclosses

18 ODI Report

The rising cost of a healthy diet: Changing relative prices of foods in high-income and emerging economies 19

Box 2A: Energy density of foods

Foodsvaryconsiderablyintheirenergydensity,asFigure2A1forselectedsnacksintheUSAshows.

Figure 2A1: Energy density of selected healthy and less healthy snacks, USA, kcal per 100 grams

0 100 200 300 400 500 600

CeleryTomatoes, grape/cherry

Broccoli floretsWatermelonRed peppersStrawberries

CantaloupeCarrots, baby

PlumOranges, navel

ApplesTangerines

Peaches, cannedFruit cocktail, canned

Pineapple, cannedGrapes

Applesauce, jarredBananas

Sweet potatoes, cookedPudding, ready-to-eat

Ice creamPizza, from frozen

RaisinsCupcakes

DanishFruit rolls

Toaster pastriesPretzels

Granola barsMuffinsDonuts

Graham crackersCookies

CrackersSandwich crackers

Tortilla chipsCorn chips

Potato chipsChocolate candy (off scale)

Calories per 100g

1008

125kcal/100g - WCRF average dietary density recommendation

275kcal/100g - Mexican energy-dense tax threshold

Source: Constructed with data from USDA (thresholds added)

Foodsinlowinenergyforweightareseenasbeingpartofahealthydiet.TheBritishNutritionFoundationclassifiesfoodbyenergydensityasfollows:

• Verylow <60kcalper100grams • Low 60to150kcalper100grams • Medium 150to400kcalper100grams • High >400kcalper100grams

Source: http://www.nutrition.org.uk/healthyliving/fuller/what-is-energy-density.htmls

TheWorldCancerResearchFund(WCRF)recommendsdietsinwhichmostitemscontainlessthan125kcalper100grams.Mexicodiscouragesconsumptionoffoodswithmorethan275kcalper100gramsthroughtheimpositionofatax

(seesection3.6).Energydensityisjustonedimensionofahealthydiet.TheImamuraetal.(2015)studyoutlinedinBox1Breflects

amedicalconsensusontenhealthyfoodsorcategoriesandtenunhealthyoptions(thestudywasabletomeasureonlysevenoftheten).

Locketal.(2010)summariseaconsensusonhealthydietsasfollows:

• Asystematicreviewofdietaryrecommendationsdefinedbyexpertpanelsandpublishedbetween1990and2004forthepreventionofnutritionaldeficienciesandinfectiousandchronicdiseases,identifiedabroadconsensusacross94reports.Consensualexpertopinionsuggeststhathealthydietsshouldcontainlargeamountsofcereals,vegetables,fruits,andpulses,whilelimitingtheamountofredandprocessedmeat,resultinginahighintakeofdietaryfibreandmicronutrientsandalowintakeoffats,saturatedfattyacids,addedsugars,andsalt.(WCRF2007)

• Inadditiontomaintenanceofenergybalance(totalcaloricintakevstotalenergyexpenditure)andhealthyweight,ahealthydiettoprovideadequatepopulationnutritionandreducechronicdiseaseriskconsistsof:15–30%oftotalenergyasfat,ofwhichsaturatedfatshouldbelessthan10%andtransfattyacidslessthan1%;55–75%oftotalenergyastotalcarbohydrate,ofwhichaddedsugarsshouldbelessthan10%;10–15%oftotalenergyasproteinfrommainlyplantsources;lessthan5gperdayofsalt;andmorethan400gperdayoffruitsandvegetables.(WHO/FAO2003)

havebeenproducedindustrially,oftenaddingtothemainingredientsadditionalfatoroil,salt,sugarandflavouringstoenhancetasteandpalatability.

Threeofthesefoodgroupsareofparticularinterestfortheirpotentialroleindietsthatleadtoobesity.Foodsthataretypicallydenseinenergy(seeBox2A)makeitpossibletoeatlargeamountsofcaloriesbeforefeelingfull.Theyincludethetwocategoriesoffats,oilsandsugar;andhighlyprocessedfoods.Fruitandvegetablesconstitutetheothergroup,partlybecausemostarenotdenseinenergy,andpartlybecauseitiswidelyconsideredthatdietsinHICsshouldincludemoreoftheseitems(see,forexample,USDA2015).Ifitwerethecasethatpricesofoils,fats,sugarsandprocessedfoodswerefallingrelativetofruitandvegetables,thiswouldbeacauseforconcernsinceitwouldbeaneconomicencouragementtoselectlesshealthyitemsindiets.

Additionalanalysiswouldbenecessarytotracetheseeffectsalongthecausalchain.Forthetimebeing,theaimiskeptsimpleandstraightforward:torecordchangesinthepricesoffoodsfromdifferentcategories.Insubsequentresearchwehopetofollowupthisquestionbyaddressingtheprobableconsequencesofchangesinrelativepricesoffoods,asking:

• Dochangesintherelativepricesoffoodcorrelatewithchanginglevelsofconsumptionofthemainfoodgroups?

• Dotheycorrelatewithchanginglevelsofoverweightandobesity?

• Whereenergy-densefoodshavebecomerelativelycheaperthanotherfoods,havetherebeenmorerapidincreasesinlevelsofoverweightandobesity?

20 ODI Report

3. What is already known?

3.1 United States

3.1.1 Background: the high rates of overweight and obese people in the USA

OverweightandobesityratesintheUSAareamongthehighestintheworld,andthehighestamongOECDcountries(seeFigure3.1).Some74%ofUSadultmaleswereestimatedtobeoverweightorobesein2008,afigureexceededonlybyratesineightsmallPacificislandnations2andKuwait.USadultfemaleshadslightlylowerestimatedratesofoverweightandobesity,some68%in2008,3althoughstillextremelyhigh.

In2008,thedirectmedicalcostofoverweightandobesityintheUSAwasestimatedatUS$113.9billion,about0.77%ofUSGDPthatyearaccordingtoWorldBankdata(Tsaietal.2011).4

ObesityhasbeenrisingintheUSA:

Inspiteofincreasedrecognitionandmediaattentiontotheproblem,theobesityepidemiccontinuestoworsen.Justbetween2000and2005,theprevalenceofobesityintheUSAincreasedby24%,whilethenumberofseverelyobese[bodymassindex(BMI)>40]cases

increasedby50%,high-lightingtheimportanceofchangeovertime.Whenlookingattrenddata,changesinBMIappeartobeverysimilaracrossallpopulationgroups,althoughtheprevalenceatanypointishighestamonggroupswithlowerincomeandeducation,andsomeethnicminorities.(Sturm2008)

Whileobesityisanindividualcondition,itswidespreadprevalenceandincreasesuggestthatthedriversmustincludefactorsthataffectmanypeople,notjustindividuals.

USdietsarefarfromthenutritionist’sideaofwellbalanced.Theycontaintoomuchoil,protein-richfood,cerealsandsugar,andtoofewfruitsandvegetables.

Tomeet2005Dietary Guidelines,typicalAmericanswouldneedtomorethandoubletheircurrentintakeofvegetablesandwhole-grainfoodswhilereducingtheirintakeofsolidfatsandaddedsugarsbyhalf(U.S.DepartmentofHealthandHumanServicesandU.S.DepartmentofAgriculture,2005).(KuchlerandStewart2008)

Thirtyyearsago,dietswerenotquitesounhealthy:calorieavailabilityintheUSAhasrisenmarkedlysincethe1970s.Sturm(2008)reportsthattheincreaseinavailableenergycomesalmostentirelyfromcarbohydrates,especiallySSBsandsnackfoods:

Theavailabilityofsugar-sweetenedbeveragesandsnackitemshasincreasedparticularlyquickly.Between1970and2005,caloricsweetenersincreasedby20poundspercapitaperyear;sweetsandconfectionarygoodsincreasedby3pounds…Theavailabilityofsugar-sweetenedbeveragesincreasedby8.5gallonspercapitaperyearfrom1985to2005;40%ofthisincreasewasduetofruit-flavoreddrinksandsportsdrinks,andtheremainderwasduetocarbonatedsoftdrinks.

Sturmconsiderstheseincreasestobemorethanenoughtoexplaintheepidemicofobesity,evenwhenconsideringwaste.

Overthe40yearsfrom1970to2009,totalenergyavailabilityincreasedalmost11%percapita,accordingtoCardenandCarr(2013).Energyfromprotein,carbohydratesandfatincreasedbyaround5%,10%and15%respectively.5

2 Nauru,CookIslands,Tonga,FrenchPolynesia,Samoa,Palau,Kuwait,KiribatiandtheMarshallIslands(datafromStevensetal.2012).

3 OnthisindicatortheUSAlagsbehindsomeothercountriesinprevalence(thosealreadymentionedforthecaseofmen,aswellastheFederatedStatesofMicronesia,SaintKittsandNevis,PuertoRico,NetherlandsAntilles,Egypt,Belize,Barbados,SouthAfrica,Fiji,UAE,Bahrain,Dominica,Bahamas,Qatar,SolomonIslands,MexicoandSaudiArabia)(datafromStevensetal.2012).

4 Bywayofcomparison,theUSAspentaroundUS$49.1billiononforeignaidin2008.

5 Theybelievethattoomuchemphasisisplacedontheincreasedintakeoffructoseforwhichtheyseetheyseenoincreasefrom1970to2009:increasesinglucoseandfatwerethemaincontributors.

Figure 3.1: Prevalence of overweight and obesity, USA, by age and sex, 2011–12

Source: Compiled from data from Tables 3 and 4 in Ogden et al. 2014

22 ODI Report

The rising cost of a healthy diet: Changing relative prices of foods in high-income and emerging economies 23

StudiesoffoodpricesintheUSAtypicallyposethefollowingquestions:

• Arehealthierfoodsmorecostlythanlesshealthyfoods? • Howhavethepricesofdifferentfoodschangedovertimerelativetooneanother?

• Howdoesconsumptionrespondtopricechanges–includingthoseinducedbytaxesandsubsidies?

• Whateffectdopricechangeshaveonlevelsofobesityandassociatedillhealth?

3.1.2 Costs of food: healthy and less healthy foods compared

Acommonwaytodefinethehealthinessoffoodsisbyreferencetoenergydensity,sinceitisassumedthatadietofenergy-densefoodswillleadtoover-consumptionbeforetheappetiteissated.Studiescomparingthecostpercalorieacrossfoodsusuallyshowthatthosedenseinenergyhavethelowestcostspercalorie.

Forexample,when372foodsinSeattlein2006weremeasuredforpriceandenergycontent,aclearinverserelationbetweenenergydensityandcostperunitofenergycouldbeseen(seeFigure3.2)(MonsivaisandDrewnowski2007).Whenfoods,otherthanbeverages,weresortedintofivequintilesbytheirenergydensity,theaveragecostper1,000kcalvariedfromUS$1.76forthemostenergy-densefoodtoUS$18.16fortheleast.

PlottingthecostsofdifferentsnackfoodsintheUSAagainsttheirenergydensityshowsasimilarpattern(seeFigure3.3).Fruitandvegetablesarenotablymorecostlypercaloriethanmostprocessedsnacks.

Connelletal.(2012)confirm,drawingondatafromtheLowerMississippidelta,thatfats,oilsandsweetsaremuchcheaperpercaloriecomparedtofruitandvegetables.

Unusually,DavisandCarlson(2012)suggestthattherelationshipbetweenpricepercalorieandenergydensityisaspuriouscorrelation,usingasampleofover4,000differentfoods,althoughencompassingfewerfoodgroupsthanMonsivaisandDrewnowski(2007)–forinstance,excludingoils,from2003/04.Aregressionofenergydensityonprice,however,forthedatatheypresentinTable1,givesanR-squareof0.41(withhighlysignificantF-statistic).

Otherstudieshaveassessedfoodsforawiderrangeofhealthycharacteristics.Forexample,Drewnowski(2010)foundthatcarbohydrates,sugarandfathadlowerpricepergram,whileprotein,fibre,vitaminsandmineralshadhigherpricespergram.

CarlsonandFrazão(2012)definedlesshealthyfoodsasthosehighinsaturatedfat,addedsugarand/orsodium,orthatcontributelittletomeetingdietaryrecommendations.Againtheyconfirmedthathealthyfoodscostmorethanlesshealthyfoodswhenmeasuredintermsofthepriceoffoodenergy.

Giventheevidencethathealthierfoodscostmorethanlesshealthyfoods,itisnosurprisethatstudies report that healthier diets cost more than less healthy ones.Forexample,peopleinthePugetSoundregionwhoeatmorevitamins,minerals,andfibrehavedietsthatcostmore(Aggarwaletal.2012).Higherintakesofdietaryfibre,vitaminsA,C,D,E,andB12,betacarotene,folate,iron,calcium,potassium,andmagnesiumwereassociatedwithhigherdietcosts–especiallysowhenthemainsourceofthenutrientcamefromfruitandvegetables.Conversely,higherintakesofsaturatedfats,transfatsandaddedsugarswereassociatedwithlowerdietcosts.Figure 3.2: Relationship between cost of per unit of

energy and energy density for 372 foods in Seattle area supermarkets, 2006

Source: Figure 2 in Monsivais and Drewnowski 2007

Note: Linear regression gives R2= 0.38.

Figure 3.3: Energy density and price per calorie for selected snack foods, fruit and vegetables, USA, 2008

Source: Constructed with data from USDA ERS database available

at http://www.ers.usda.gov/data-products/fruit-and-vegetable-prices.

aspx#.UuphpLS7Tm4

Note: Horizontal axis is a log scale.

Basedoncurrenteatinghabits,compliancewithdietaryguidelinesislikelytoentailhigherdietcostsfortheconsumer.(Drewnowski2010)

Womenwithhigh-nutrientdietswerefoundtohavemorecostlydietsonaveragethanmenwithhigh-nutrientdiets,reflectingwomen’srelativelyhigherconsumptionoffruitandvegetables,andmen’srelativelyhigherconsumptionofmeats.

Ifhealthierdietscostmorethanlesshealthyones,thenthoseonlowincomes,beingsensitivetofoodprices,arelikelytochooselesshealthyoptions.Forthoseonverylowincomes,healthydietssimplybecomeunaffordable.

Thefactthatenergy-densefoods(megajoules/kilogram)costlesspermegajoulethandonutrient-densefoodsmeansthatenergy-densedietsarenotonlycheaperbutmaybepreferentiallyselectedbythelower-incomeconsumer.Inotherwords,thelowcostofdietaryenergy(dollars/megajoule),ratherthanspecificfood,beverage,ormacronutrientchoices,maybethemainpredictorofpopulationweightgain.(Drewnowski2007)

Not all studies find that healthier diets are more expensive.Forexample,DrewnowskiandEichelsdoerfer(2009)lookedatwhetheraMediterraneandiet–consideredtoberelativelyhealthysinceitisrichinvegetables,fruits,beans,wholegrains,oliveoilandfish–costmorethanpeople’scurrentdiets.Whiletheirfindingssuggestedsomenutrient-richlow-energy-densityfoodsassociatedwiththeMediterraneandietwereexpensive,othersthatalsofitwithintheMediterraneandietarypatternwerenot.

WhenBernsteinetal.(2010)studiedthedietsofUSnursesgradedbytheAlternativeHealthyEatingIndex(AHEI),whichhasbeenlinkedtolowerratesofcardiovasculardisease,theyfoundthatsignificantincreasesinhealthinessindicescouldbeobtainedfromspendingmoreonnuts,soyandbeans,andwholegrains,whilereducingspendingonredandprocessedmeats.Insimilarvein,DrewnowskiandRehm(2013)reportthatschoolmealscanbemademorenutritious,aboveallinpotassiumandfibre,byincludingmorepotatoandbeans.

3.1.3 Changes in prices over timeThekeyissuesarechangesinpricesofdifferentfoodsrelativetooneanother.Several studies show that energy-dense food has become cheaper over time, compared to foods less dense in energy.Forexample,Gelbachetal.(2009)examinedtrendsin‘healthy’foods–freshfruitandvegetables,sodacrackers,wholemilk,lighttunaandyoghurt–comparedto‘unhealthy’foods–processedcheese,butter,cola,margarine,potatocrisps,etc.–from1982to1995.Theyfoundpricesoftheunhealthyfoodsfellfartherthanthoseforhealthyfoods.Similarly,WendtandTodd(2011)foundrisingrealpricesforfruitandvegetables,andfallingrealpricesforcarbonateddrinksfrom1980to2010(seeFigure3.4).

ThestudybyPowelletal.(2013)confirmsthesetrends:‘between1980and2011itbecame2.2timesmoreexpensivetopurchasefreshfruitsandvegetablescomparedtopurchasingcarbonatedbeverages’.Overthetwoyearsfrom2004to2006pricesofpricesofthehighestquintileofenergy-densefoodsinSeattledroppedby1.8%,whilepricesoftheleastenergy-densefoodsgrewby19.5%(MonsivaisandDrewnowski2007).

Changes in quality, seasonality and processing qualify these findings.KuchlerandStewart(2008)studiedmonthlypricesfromthe1980stothemid-2000sinUScities.Theycomparedchangesinpricesforfourprocessedfoodstopricesfor11fruitsandvegetables(seeFigure3.5).

Inseveralcases,littledifferencewasseeninpricechangesoffoodsinthetwogroups,althoughtheirattentionwasdrawntotomatoesandbroccoli,whichhadriseninpricewhilethepricesofotherfruitandvegetablesandtheprocessedfoodshadfallen.

Thismight,however,beexplainedbythreechangesinquality.One,somevarietiesofproducehavechangedovertime.Tomatoes,forexample,areincreasinglymarketedasvinetomatoesandotherspecialisttomatoes,whicharesoldatpremiumprices:theaveragetomatopriceovertimewillthusbebiasedupwards.Two,whileinthepastsomefruitandvegetableswereonlyavailableseasonally,bythemid-2000smostwereontheshelvesalltheyearround.The1980averageannualpriceofstrawberrieswasbasedonproductionduringnomorethanfourmonthsoftheUSseason:by2006theaverageyear-roundpriceincludedoff-seasonfruitproducedathighercostorair-freightedtotheUSA,withadditionaltransportcosts.Three,somevegetableswereincreasinglysoldwashed,cutandbagged,asinthecaseofbroccoli,withhighervalueaddedfortheconsumer.

Figure 3.4: Price indices for selected foods and beverages in the USA, 1980–2010

Source: Figure 1 in Wendt and Todd 2011 (based on data from the

Bureau of Labor Statistics). Note: Prices are annual average for urban

consumers. ‘All fruits and vegetables’ include fresh, canned, and

frozen. Base period 1982-84=100.

24 ODI Report

The rising cost of a healthy diet: Changing relative prices of foods in high-income and emerging economies 25

…valueaddedthroughtransportation,processing,wholesaling,andretailinghasgrowntoaccountforaboutthree-fourthsoftheretailpriceoffruitsandvegetables,onaverage,comparedwithabouttwo-thirdsintheearly1980s(Stewart,2006).Theseservicesservetwopurposes:increasedconvenienceandvariety.(KuchlerandStewart2008)

3.1.4 Price changes and consumption

Themostfamiliarstudiesdealingwiththeeffectofpricesonconsumptionlookatown-priceandcross-priceelasticityofdemand.Andreyevaetal.(2010)reviewedUSliteratureonhowpricechangesaffectdemandformajorfoodcategories,lookingatatotalofsome160studiesofpriceelasticityformajorfoodcategories,publishedbetween1938and2007,mostlyinthelatterpartofthisperiod.Theyfoundown-priceelasticitiesforfoodsandnon-alcoholicbeveragesrangedfrom0.27to0.81(absolutevalues),whichisrelativelyinelastic(see

Table3.1).Foodawayfromhome,softdrinks,juice,andmeatsweremostresponsivetopricechanges(0.7–0.8).

Powelletal.(2013)extendtheperiodbyreviewingstudiesofpriceelasticityfrom2007to2012(seeTable3.2).Althoughmostfoodshaveinelasticown-priceelasticityofdemand,demandforSSBsisrelativelyelastic.

Morespecificstudiesofpricelookatpriceresponse,butinthecontextofotherfactorsaffectingconsumption,andsomedifferentiatebyeconomicandsocialstatusofdifferentgroupsofconsumers:

• Powelletal.(2009a)estimatedthelinkbetweenyoungadults’consumptionoffruitandvegetablesandthepricesofthosefruitandvegetables,pricesofotherfoodconsumedathome,andfastfood,aswellastheavailabilityofrestaurantsandfoodshops.Higherconsumptionoffruitandvegetableswasassociatedwithlowerfruitandvegetableprices,withapriceelasticityof−0.32.Thisown-priceeffectwasrobusttotheinclusionofotherfoodpricesandtheavailabilityoffoodoutlets.

Figure 3.5: Long-term average annual changes in retail prices for selected foods in the USA

Source: Tables 3 and 4 in Kuchler and Stewart 2008 (based on ERS calculations using Bureau of Labor Statistics), US City average price data

and urban CPI.

Note: suspended series are those with prices not recorded between 2000 and 2006 (by 2000 bagged versions and prepared salads accounted for

large share of produce sold).

Youngadultswithlowerincomesandlowerlevelsofeducation,thosewithlowereducatedmothers,andmiddle-incomeparentswerethemostpricesensitive.Theyfoundnostatisticallysignificantcross-priceeffectsonfruitandvegetableconsumptionfromothergroceryprices(meat,dairyandbread)orfastfoodprices.

• PowellandHan(2011)sawsomepriceresponseamongadolescentstothecostoffastfoods,butonlyamongthoseonlowincomes.

• French(2005)lookedathoweffectiveprice-basedinterventionsmightbetopromoteconsumptionofhealthierfoodsinworkplacesandschools.Reviewingseveralstudies,shefoundpricereductionsoflower-fatsnackoptionsof10%,25%and50%ledsalesofthesesnackstoincreaseby9%,39%and93%respectively.Salesoffreshfruitandvegetablesalsoincreasedwhentheirpriceswerehalved.

• Khanetal.(2012)reportedthata10%increaseinthepriceoffastfoodwasassociatedwith5.7%lowerfrequencyofweeklyconsumptionoffastfoodamongchildreningrades5and8.

• SturmandDatar(2011)examinedthevaryingpricesoffoodsacrossUSmetropolitanareas.Amongchildreningrade5(averageageof11),theyfoundlowerrealpricesforvegetablesandfruitspredictsignificantlyhigherfrequencyofintake.Higherdairypricespredictlower

frequencyofmilkconsumption,whilehighermeatpricespredictincreasedmilkconsumption.Similarpriceeffectswerenotfoundforfastfoodorsoftdrinkconsumption.

• Powelletal.(2013)showedcleartrendstowardsincreasedintakesofSSBsandfoodeatenawayfromhome(particularlyfastfood),forwhichpriceshadfallen;andlowandlittle-changingconsumptionoffruitandvegetablesforwhichpriceshadrisenoverthelasttwodecades.TheyalsosawstrongereffectsonthoseonlowincomesandontheSupplementalNutritionalAssistanceProgram(SNAP)thatistheUSA’sbiggestnationalfood-stampwelfareprogrammeforthoseonlowincomesand/orunabletowork.

Table 3.1: US own-price elasticity of demand, by food and beverage category, 1938–2007

Food and Beverage Category (a) Absolute Value of Mean Price Elasticity Estimate (95% CI) Range No. of Estimates

Foodawayfromhome 0.81 (0.56,1.07) 0.23-1.76 13

Softdrinks 0.79 (0.33,1.24) 0.13-3.18 14

Juice 0.76 (0.55,0.98) 0.33-1.77 14

Beef 0.75 (0.67,0.83) 0.29-1.42 51

Pork 0.72 (0.66,0.78) 0.17-1.23 49

Fruit 0.70 (0.41,0.98) 0.16-3.02 20

Poultry 0.68 (0.44,0.92) 0.16-2.72 23

Dairy 0.65 (0.46,0.84) 0.19-1.16 13

Cereals 0.60 (0.43,0.77) 0.07-1.67 24

Milk 0.59 (0.40,0.79) 0.02-1.68 26

Vegetables 0.58 (0.44,0.71) 0.21-1.11 20

Fish 0.50 (0.30,0.69) 0.05-1.41 18

Fats/oils 0.48 (0.29,0.66) 0.14-1.00 13

Cheese 0.44 (0.25,0.63) 0.01-1.95 20

Sweets/sugars 0.34 (0.14,0.53) 0.05-1.00 13

Eggs 0.27 (0.08,0.45) 0.06-1.28 14

Source: Table 1 in Andreyeva et al..2010

Note. Values were calculated based on the 160 studies reviewed. Absolute values of elasticity estimates are reported: the estimated elasticities

are all negative in that quantity demanded falls with rising prices. The price elasticity of demand measures the percentage change in purchased

quantity or demand with a 1% change in price. aIncluding restaurant meals and fast food.

Table 3.2: Own-price elasticity of demand for four food groups in the USA, 2007–12

Food item Price elasticity of demand Range

Sugar-sweetenedbeverages -1.21 -0.71to-3.87

Fastfood -0.52 -0.47to-0.57

Fruits -0.49 -0.26to-0.81

Vegetables -0.48 -0.26to-0.72

Source: Data from Table 2 in Powell et al. 2013

26 ODI Report

The rising cost of a healthy diet: Changing relative prices of foods in high-income and emerging economies 27

Severalofthesestudiesweremotivatedbyaninterestinthepotentialeffectsoftaxesandsubsidiesondifferentfoods.Inasynthesisof24studiesfocusingonadolescentsaged12to17andadultsaged18andoverpredominantlyintheUSA6,An(2013)broughttogetherevidencefrom20distinctfieldinterventions–pricediscountsorvouchersforhealthierfoods:fruits,vegetables,orlow-fatsnackssoldinsupermarkets,cafeterias,vendingmachines,farmers’markets,orrestaurants–toassesstheeffectivenessofsubsidiesonpromotinghealthierfoodpurchasesandconsumption.Inallcasesbutone,subsidiesonhealthierfoodssignificantlyincreasedtheirpurchaseandconsumption.Theonenullfindingwasowingtoitssmallfinancialincentive(50centstowardsthepurchaseofanyfruitorvegetable).

3.1.5 Food prices and body weightSeveralstudieslookattheimpactofdifferentpricingonconsumptionandlinkittoweightoutcomes.Most report that higher prices for unhealthy, and lower prices for healthy options, lead to lower body weights.

Forexample,inalongitudinalstudy(WendtandTodd2011)followedanationallyrepresentativecohortofkindergartenchildrenenteringin1998/99tograde8(2007)todeterminetheinfluenceofchangingpricesofcertainfoodsonchildren’sBMIs.Theevolutionofrelativepricestheytrackedshowedpricesofwholemilk,non-alcoholicbeverages,andcarbonateddrinksfallinginrealtermsoverthelast30years,whilepricesoffruitandvegetablesroseconsiderably.Moreover,withunhealthyoptionsbecomingcheaper,consumptionrose.Theyfound:

• A10%decreaseinthepriceoflow-fatmilkinthepreviousquarterwasassociatedwithadecreaseinBMIof0.35%,.

• A10%decreaseinthepriceofdarkgreenvegetables–spinach,broccoli–inthepreviousquarterwasassociatedwithadecreaseinBMIof0.28%.

• A10%increaseinthepriceofcarbonatedbeverages(oneyearprior)wasassociatedwithadeclineinBMIof0.42%,withastrongereffectonchildreninlow-incomehouseholds.

• A10%priceincreaseinfruitjuices(100%juice)orstarchyvegetables–potato,maize–(alsooneyearprior)decreasedBMIby0.3%.

• AdecreaseinthepriceofsweetsnacksinthepreviousquarterincreasedBMIby0.27%,althoughsometimesobservedchangesweremoredelayed.

Powelletal.(2007)reportthepricesoffastfoodtobeimportantdeterminantsofadolescents’bodyweightandeatinghabits:a10%increaseinthepriceofafastfoodmealleadingtoa3%increaseintheprobabilityoffrequentfruitandvegetableconsumption,a0.4%decreaseinBMI,

anda5.9%decreaseinprobabilityofoverweight.Pricesoffruitandvegetables,aswellasdensityofrestaurantoutlets,weredeemedlessimportantdeterminants.Nonetheless,changesinallobservedeconomicandsocio-demographiccharacteristicstogetheronlyexplainedroughly25%ofthechangeinmeanBMIand20%ofthechangeinoverweightbetween1997and2003.

Duffeyetal.(2009)sawthatfromthemid-1980stothemid-2000sthepricesofsoda(SSB)andpizzafellwhilemilkpricesrose.Theyfoundthata10%riseinthepriceofsodaorpizzawaslinkedtoa7.12%or11.5%decreaseinenergyintakefromthesefoodsrespectively.Priceincreasesinbothfoodsreducedbodyweightandsusceptibilitytodiabetes.

Morrisseyetal.(2014)foundthatforunder-fivesinlow-incomehouseholds,morecostlyfruitandvegetableswerelinkedtohigherBMIsamongchildren,arelationshipdrivenbypricesoffreshfruitandvegetablesratherthanoffrozenorcannedoptions.Higherpricesforsoftdrinkswerealsolinkedtoalowerlikelihoodofthechildrenbeingoverweight.Counter-intuitively,however,higherfast-foodpriceswerelinkedtoagreaterlikelihoodofchildrenbeingoverweight.

Lookingatapopulationofadultsofover60yearsofage,Goldmanetal.(2011)founda10%dropinpricepercaloriewasassociatedwithaBMIincreaseofapproximately0.26units,ora0.77%risewithintwoyears.ThiseffectoffoodpricesonBMIwasstatisticallysimilaracrossobeseandnon-obesepopulations,whilenosignificantdifferencewasestablishedacrosspoorandnon-poorpopulations.

Althoughtheshort-termeffectofpricepercalorieonBMIappearsrelativelysmall,thelong-termeffectmaybelarger.Aftertenyears,apermanent10%reductioninpricepercalorieislinkedtoBMIincreasingby1.05units(2.5%).Overthefullspanofthestudy,thisequatestoariseinBMIof2.2units,or5.1%:asignificantcontributiontototalgrowthofmeanBMIovertheperiod(Goldmanetal.2011).

Inastudyunusualforitsinclusionofmeasuresofthepercentageofbodyfat(PBF)aswellasBMI,Grossmanetal.(2013)lookedattheinfluenceoffoodpricesonclinicalobesity–measuredbyBMIandPBF.Controllingforcontextualvariables,suchasethnicity,age,familyincome,householdtypeandsize,andeducation,theyfoundthatincreasesinrealfoodprices(percalorie)–forhomeconsumption,andintherealpriceoffast-food–ledtolowerobesityinyouthsaged12to18years,whileincreasesinrealpricesoffruitandvegetablesledtohigherobesity.

Percentagebodyfat(PBF)measureswerenolesssensitive–andinsomecasesmoresensitivetosuchpricechangesthanBMImeasures.PricesoffruitandvegetablesweremoreimportantindeterminingfemalePBFthanmalePBF.A10%riseinfruitandvegetablepricescausesPBFrisesof9%forfemalesand7%formales(significantonlyforfemales).On

6 Whilethisstudyassessesinterventionsinsevencountries,mostofthemareintheUSA(14),withoneeachinFrance,Germany,theNetherlands,SouthAfricaandtheUK.

theotherhand,thepriceofacalorieinfoodconsumedathomeorinfast-foodrestaurantsplaysamoreimportantroleinmalethaninfemalePBF(Goldmanetal.2011).

Powell(2009)andPowelletal.(2010)lookedatpricesaswelltheavailabilityoffast-foodoutlets.7Theyfoundthatthepriceoffastfood,butnotavailabilityoffast-foodrestaurants,hadasignificantinfluenceonBMIamongteenagers,withpriceelasticityof−0.08(comparedwiththepriceelasticityof−0.10estimatedusingacross-sectionalmodel).Theweightofteenagersinlowertomiddle-socio-economicstatusfamilieswasmostsensitivetofast-foodprices.

Asisterstudy(PowellandBao2009)foundfruitandvegetablepricesalmostequallystronglylinkedtochildren’sBMI:a10%increaseinthepriceoffruitandvegetableswaslinkedtoa0.7%increaseinchildren’sBMI.Theinfluenceoffast-foodpriceswasnotstatisticallysignificantinthefullsample,butweaklynegativelyassociatedwithBMIamongadolescents,withanestimatedelasticityof-0.12.Moreover,associationsoffruitandvegetableandfast-foodpriceswithBMIweresignificantlystronger(economicallyandstatistically)amongchildrenfromlow-incomehouseholds.Estimatedfruitandvegetableandfast-foodpriceelasticitieswere0.14and-0.26,respectively,amonglow-incomechildrenand0.09and-0.13,respectively,amongchildrenwithlesseducatedmothers.

SturmandDatar(2005)alsoconsideredfoodpricesandfood-outletdensityinexaminingchangesintheBMIofUSprimaryschoolchildren.LowerrealpricesforfruitandvegetablespredictedasignificantlysmallergaininBMIbetweenkindergartenandgrade3,halfofitoccurringbetweengradeskindergartenandgrade1.LowermeatpricesraisedBMI,thoughgenerallybyasmallermagnitude,whiletheeffectwasnotsignificantforBMIgainoverthreeyears.Effectsweremeaningfullylargerforchildrenlivinginpoverty,childrenalreadyatriskofbeingoverweightoralreadyoverweightinkindergarten,andAsianandHispanicchildren.Nosignificanteffectsfordairyorfast-foodpriceswerefound.

Beydounetal.(2011)lookedatwhatinfluencepriceindicesoffastfoodsandthoseforfruitandvegetableshadondietaryintakeandtheBMIofUSchildrenandadolescentsagedtwoto18years.Amongtwotonine-year-olds,ahigherfast-foodpriceindex(byUS$1)wasassociatedwithlowerfast-foodconsumption,healthiereatingpatterns,andhigherintakeoffibre,calcium,dairy,andfruitandvegetables.ThefruitandvegetablepriceindexwasrelatedtolowerfibreintakeandhigherBMIs.Theirfindingsfor10to18-year-oldswerelessconsistent.Significantassociationswerealmostequallybalancedbetweenlow-andhigh-incomegroups,withsomesignificantinteractionsbetween

foodpricesandfamilyincomeobserved,particularlyamongtheyoungergroupofchildren.

Chouetal.(2004)foundthata10%increaseinpricesatfast-foodrestaurantswouldreducetheprobabilityofobesityby0.65%,whilea10%increaseinpricesatfull-servicerestaurantswouldreducetheprobabilityofobesityby0.67%,anda10%increaseinthepriceoffoodathomewouldreducetheprobabilityofobesityby0.62%.

Powelletal.(2013)observedhigherfast-foodpriceswereassociatedwithlowerweight,particularlyamongadolescents.Lowerfruitandvegetablepricesweregenerallyfoundtolinktolighterbodyweightsamonglow-incomechildrenandadults.Theyconclude:

Thegrowingevidencebaseassessedhereinindicatesthatchangesintherelativepricesoflesshealthyandhealthierfoodsandbeveragescansignificantlychangeconsumptionpatternsand,mayhavesignificantimpactsonweightoutcomesatthepopulationlevel,particularlyamongpopulationsmostatriskforobesityanditsconsequences.Raisingthepricesoflesshealthyoptionsbytaxingthemhastheaddedbenefitofgeneratingconsiderablerevenuesthatcanbeusedtosupportcostlyprogramsandotherinterventionsaimedatimprovingdiets,increasingactivity,andreducingobesity,includingsubsidiesforhealthierfoodsandbeverages.(Powelletal.2013)

Other studies have found insignificant or negligible impacts of relative prices on body weights. ThestudybyGelbachetal.(2009)foundthatpricesofunhealthyfoodsfellfrom1982to1995morethanthepricesofhealthyfoods,butthatalthoughchangesinBMIwerecausallyrelatedtorelativefoodprices,thedegreeofinfluencewassmall.A100%taxon‘unhealthy’foodswouldreduceaverageBMIbylessthan1%andreduceincidenceofoverweightby2%andobesityby1%.

OtherstudieshavenotbeenabletofindasignificantinfluenceofrelativefoodpricesonBMIs.HanandPowell(2011),inalongitudinalstudyofover10,000youngadults,wereunabletofindasignificanteffectoffoodpricesontheprevalenceofobesityamongyoungwomen.Foryoungmen,a10%increaseinthepriceoffastfoodwaslinkedtoa13%dropinthechanceofbeingobese–althoughthisfindinglostitseconomicandstatisticalsignificancewhenindividualfixedeffects(e.g.work,maritalstatus,andschoolenrolment)wereintroduced.

Whilefoodeatenawayfromhomeneednotbeunhealthy,datasuggestthateatingonemealaweekawayfromhomefortheaverageUSconsumerleavesthemroughlytwopounds(approximately1kilogram)heaviereachyear(Toddetal.2010).TheshareofUSfoodspendingoutsidethehomehasrisentoalmost50%,comparedwith25%in1960(Kumcu2011).

7 Low-tomiddle-incomeareashave1.25–1.3timesasmanyfast-foodrestaurantsashigh-incomeareas.Theproportionoffast-foodrestaurantscomparedtothetotalnumberofrestaurantsintheUSAwentfrom17%in1997to30%in2006.Fast-foodrestaurantsandconveniencestoresarereadilyavailablearoundUSsecondaryschools,especiallythoseinlargercitiesand/orlow-incomeneighbourhoods.

28 ODI Report

The rising cost of a healthy diet: Changing relative prices of foods in high-income and emerging economies 29

3.1.6 Effects of taxes and subsidies on consumption and body weight

Taxesandsubsidiesonfoodanddrinkseekexplicitlytochangepricesandtherebytoinfluenceconsumption.

Studies of impacts of taxes on items such as soda typically find small effects, if any.Lookingatdataonadolescents,Powelletal.(2009)foundnostatisticallysignificantassociationsbetweenstate-leveltaxesonsoda(SSBs)andadolescentBMI.AweakeconomicandstatisticallysignificanteffectwasfoundbetweenratesoftaxonsodasoldinvendingmachinesandBMIamongteenagersatriskofbecomingoverweight.

PowellandChaloupka(2009)reviewedliteraturepublishedbetween1990and2008toexaminewhethertaxesorsubsidieswouldleadtosufficientchangesinpatternsoffoodconsumptionandoveralldiettoreducepeople’sweights.Whenstatisticallysignificantlinkswerefoundbetweenfoodandrestaurantprices(taxes)andweightoutcomes,effectsweregenerallysmall.Largereffectswereseenamongpopulationsoflowersocio-economicstatus,aswellasforthoseatriskofoverweightorobesity.Theyconcludedthatthe(limited)evidenceavailable(dealingentirelywithsmalltaxesorsubsidies)indicatedthatsmallincentiveswouldnotbelikelytoyieldsignificantchangesinBMIortheprevalenceofobesity,thoughstrongerinterventionsmayhavesomemeasurableimpactonweightoutcomesintheUSA,especiallyforchildren,adolescents,populationsoflowsocio-economicstatus,andthosemostatriskforoverweight(PowellandChaloupka2009).

Likewise,Sturmetal.(2010),lookingatchildrenfromkindergartentograde5,foundexistingtaxesonsoda(typicallynotmuchhigherthan4%)didnotsubstantiallyaffectoveralllevelsofsodaconsumptionorobesityrates.Somesub-groupsofat-riskchildren–thosealreadyoverweight,fromlow-incomefamilies,orAfricanAmericans–werefoundtobemoresensitivethanotherstosodataxes,particularlywhensuchSSBswereavailableinschools.

Fletcheretal.(2010)foundnoevidencethattaxesonsoftdrinksorrestrictionsonvendingmachinesaffectedtheBMIofchildreningrades5and8.Analysesoftherelationshipbetweensodataxestoweightoutcomesshowedminimalimpacts(Powelletal.2013).

Schroeteretal.(2006)havegonesofarastosuggestataxonfoodawayfromhome–acategoryofconsumptionimplicatedintheriseofobesity–mightactuallyleadtoincreasedobesity.Theirmodelshowedthatwhiletaxingmealsawayfromhomemayreducethefrequencywithwhichpeopleconsumethem,thesubstituteofeatingathomemayactuallyleadtomorecaloriesbeingconsumed,sincesomuchofthefoodeatenathomeisenergy-rich.

Thesestudies,however,lookedattheimpactsofrelativelylowtaxes,8typically5%orless,onasinglefoodthatmightaccountforlessthan10%ofcalorieintake.

3.1.7 Summary of the US literatureAconsensusemergesfromthisliterature,evenifcontraryfindingsandqualificationscanbefoundinthemanystudiesreviewed,asfollows:

• Moststudiesfindthathealthierfoodscostmorethanlesshealthyones.Moreoverthiseffecthasincreasedoverthelast30–40years,asenergy-dense,processedfoodshavebecomecheaperrelativetolessenergy-densefruitandvegetables.

• Consequentlyhealthydietstendtocostmorethanlesshealthydiets.Thatisnotinevitable:choosinglowercosthealthyitemsandsubstitutingthemforthemorecostlylesshealthyonesmightbothimprovedietandsavemoney.Butformostconsumers,thiswouldrequireboththeabilitytoseethedistinctions,andthedisciplinetofollowaparticulardiet.

• Althoughitseemssomeenergy-denseprocessedfoodshavebecomenotablycheapercomparedtofruitandvegetables,thenatureofthelatterhavechanged–withhigher-valueprepareditemscommoninfoodoutlets,andavailablealltheyearround.Takingthisaddedvalueintoaccountthechangeinrelativepricesmaybelessthanisatfirstapparent.

• Consumptionofmostfoodsrespondstopricechanges,althoughformanyfoodstheresponseisrelativelyinelastic.Thoseonlowincomesaremostlikelytorespondtochangingprices.

Thisfindinghasbeendilutedbystudiesoftheimpactsoffoodtaxesthatoftenseemtotriggertinychangesinconsumption.Althoughfoodindustrylobbyistsusesuchstudiestoargueagainsttaxes,thetaxesstudiedormodelledarealmostalwaysverysmall,5%orlessbeingtypical.Ithasneverbeenexpectedthata5%taxonSSBs,forexamplewouldcausealargereductionintheirconsumption.

• StudiesoftheimpactsofchangesinpricesonbodyweightproduceasurprisinglystrongconsensusthathigherpricesofunhealthyoptionsreduceBMI,asdocheaperhealthieroptions.‘Surprising’sincebodyweightistheoutcomeofmanyfactors,yetpricechangescanbeseentomakeadifference.Thestrongesteffectsareseenamongthoseonlowincomes,whoaremostsensitivetothecostoffood.

8 Around3%to5%(someacademicssuggestSSBtaxshouldbeat20%–seehttps://www.fmhs.auckland.ac.nz/en/soph/global-health/projects/informas/government-healthy-food-environment-policy-index.html).

3.2 United Kingdom

3.2.1 Background: obesity in the UKOverweightandobesityratesintheUKarealmostashighasUSratesandsignificantlyhigherthaninmostEuropeancountries.In2008,around68%ofadultmalesintheUKwereestimatedtobeoverweightorobese,while61%ofadultfemalesfellintothiscategory(datafromStevensetal.2012).9Some9%ofUKchildrenarealreadyclassifiedasobesewhentheystartschool,risingto19%inyear6(LGA2014).Children’sobesityvariesbyinequality,witharound25%ofthemostdeprivedchildren(byquintileofdeprivation)agedfrom10to11inLondonin2007/08obese,comparedtoaround13%oftheleastdeprivedquintile(seeFigure11inTheMarmotReview2010).

In2006/07,overweightandobesitycosttheNationalHealthService(NHS)£5.1billion10–morethanthecostofsmoking(£3.3billion),alcohol(another£3.3billion)orphysicalinactivityalone(£0.9billion)(Scarboroughetal.2011).ItisestimatedthatoneinsevenhospitalbedsintheUKisoccupiedbyapatientwithdiabetes,whilethenumberofadmissionstoNHShospitalswithaprimarydiagnosisofobesityrosefrom1,019in2001/02,to11,736in2011/12(LGA2014).

TypicalUKdietsarenotbalancedinaccordancewithdietaryrecommendations,withexcessiveconsumptionofgrainsandotherstarchyfoods,protein-richfoods,oils,fats,andsugar–coupledwithparticularlylowintakeoffruitandvegetables.

StudieshaveshowndecliningenergyintakesintheUK.PrenticeandJebb(1995)reporteddeclinesinsomemeasuresofenergyintakeandmorerecentlyGriffithetal.(2013)confirmedthetrendforlowerconsumptionofdietaryenergyintheUKfrom1980to2009,withsignificantreductionsincaloriesboughtforhomeconsumptiononlypartlyoffsetbytherisingshareofeatingoutinaggregateconsumption.PrenticeandJebb(1995)therebyinferredthatrisingobesitywastheresultoftoolittleactivity:‘moderninactivelifestylesareatleastasimportantasdietintheaetiologyofobesityandpossiblyrepresentthedominantfactor’.Debateonthiscontinues,somearguingthatrepresentingthemaindriversas‘sloth’or‘gluttony’is‘overlysimple’(Robertoetal.2015).

3.2.2 UK trends in relative prices The relatively few studies in the UK show healthy diets to be more costly than less healthy diets

Capaccietal.(2012)usehouseholddatafortheUKfrom1997to2009toestimatepricesof‘healthy’and‘unhealthy’bundlesoffoods.Theformerwererestrictedforsimplicitytofruitandvegetables,excluding

potatoes;whilethelatterwerethe‘bigsix’foodgroups–confectionery,softdrinks,crisps/savourysnacks,fastfood,pre-sugaredbreakfastcerealsandpre-preparedconveniencefoods;thatis,thosecategorisedbytheFoodStandardsAgency(FSA)ashighinfats,sugarandsalt(HFSS)forthepurposesofadvertisingregulationssetbytheOfficeofCommunications(Ofcom),apublicregulator.

Theyestimatethatpricesoffruitandvegetablesincreasedbyabout7%relativetoallfoodsoverthe13-yearperiod,whilethepricefor‘junk’foodrelativetoallfoodsfellbyabout15%(seeFigure3.6).

Sincetheyanalysedunitvalues,thatisspendingdividedbyquantity,ratherthanobservedprices,theylaterproducedacorrectiontoinferprices.Thedifferencearisessincewhenaspricesofagroupoffoodsrises,consumerstendtoswitchtowardscheaperitemswithinthegroupincludinglowerqualityitems,sotheiraverageunitvaluedoesnotriseinlinewithprices.

Jonesetal.(2014)classified94foodsandbeveragesintheUKas‘morehealthy’or‘lesshealthy’accordingtoanutrient-profilingmodeldevelopedbytheFSA.Meanpricesin2012were£2.50forlesshealthyand£7.49formorehealthyitemsbyunit,whileby1,000kcallesshealthyitemscost£0.29,whilemorehealthyitemscost£1.27.Moreover,whileallpricesrosefrom2002to2012,thepricesofmorehealthyitemsrosesignificantlyfasterthanlesshealthyonesinabsoluteterms:£0.17comparedto£0.07/1,000kcalperyearonaverageformoreandlesshealthyitems,respectively.

9 Interestingly,whileinmentheprevalenceofobesitygrewacrossdifferentjobcategoriesfrom1997to2007inasimilarfashion,inwomen,prevalenceofobesityincreasedacrossalljobcategoriesexcept‘professional’,whereitbarelyshiftedatalloverthetimeperiod–seeFigure11.10inAnnexII.

10 Incomparison,England’sspendingonpreventivehealthinterventionsforobesity,dietandlifestylein2006/07wasonly£116million(seeTable4.3inTheMarmotReview2010).

Figure 3.6: Fruit and vegetables and ‘junk’ food, unit values and corrected prices relative to total food

Source: Figure 1 in Capacci et al. 2012

30 ODI Report

The rising cost of a healthy diet: Changing relative prices of foods in high-income and emerging economies 31

Comparingthedifferencesin‘healthy’and‘unhealthy’dietcostsforagroupofover15,000UKwomenaged35to69,11Cadeetal.(1999)classifieddietsintoeightgradesofhealthiness,basedonWHOrecommendations.12Womeninthehealthiestgroupspentonaverage64%moreontheirfoodthanwomenintheleasthealthygroup(£2.33perdaycomparedto£3.81perday).Theyalsospentalmosthalfoftheirfoodbudgetsonfruitandvegetablescomparedtolessthanathirdintheleasthealthygroup,andfarlessonmeat—indeed,womeninthehealthiestdietgroupwerealmostfourtimesaslikelytobevegetarian.Figure3.7showsabreakdownoftheirdifferentdietcostsbyfoodgroup.

Alaterstudyusedthesamewomen’scohortstudydatafor1995–98toinvestigateasimilarquestion,classifyingdietsinaslightlydifferentway13accordingtohowwelltheyadheredtotheUKDepartmentofHealth’s‘EatwellPlate’,aguidetothecompositionofahealthydiet(Morrisetal.2014).Theysimilarlyconcludedahealthydietwasmoreexpensivethanalesshealthyone,withthehealthiestcostingtwicethepriceoftheleasthealthy,£6.63perdayand£3.29perday,respectively.

WriedenandBarton(2011)lookedspecificallyatenergy density ofdietsinScotland,andcomparedcostsofenergy-denseandlessenergy-densediets.Separatinghouseholdsintoquintilesofenergydensityshowsthe

quintilewiththeleastenergy-densedietconsumesapproximately123kcalper100goffoodandmilk,whilethoseinthequintilewiththemostenergy-densedietsconsume231kcalper100goffoodandmilk.TheWCRFrecommendsnotexceedingadietaryenergydensityonaverageof125kcalper100goffood.

Moreover,thecostpaidper2,000kcalforhouseholdsinthelowestquintileofenergydensityisalmost£5,whiletheequivalentcostforhouseholdsinthehighestquintileis£3.76(seeFigure3.8).

Dietsweremoreenergy-denseonaverageforsingle-parenthouseholds(183kcal/100g)andotherhouseholdswithchildren(177kcal/100g)thanforhouseholdswithoutchildren(single-personhouseholds,forinstance,ateanaverageof169kcal/100g).Meanenergydensityforfoodandmilkconsumedinthe309householdsmeetinghealthtargetsforfatconsumption(<=35%offoodenergy)andfruitandvegetableconsumption(>400g/day)was136kcalper100g.Forthe3,859householdsnotmeetingthesetargets,theequivalentfigurewas175kcal/100g.

IfhealthierdietscostmoreintheUK,thefearisthatpeopleonlowincomesmaynotbeabletoaffordthem.Banksetal.(2012)reportthisisnotnecessarilythecase,andindeedthetypeoffoodretailerplaysmoreofaroleinthecostoffoodthandodifferencesinfoodchoicebetweenhealthyandlesshealthyoptions(seeBox3A).

11 DatafromtheUKWomen’sCohortStudycollected1995–1998.

12 WHOhealthydietindicatorcomponentsincluded:percentageoftotalenergyfromsaturatedfattyacids,frompolyunsaturatedfattyacids,fromprotein,fromcomplexcarbohydrates,andfromfreesugars(excludingfructoseandlactose);gramsofdietaryfibre,offruitandvegetables,ofpulses,nuts,andseeds.

13 Theylookedatsevendietarypatterns,withadiethealthinessscorefrom1to5.Worstdietsweredescribedas‘Monotonouslowqualityomnivore’(scoreof1),followedby‘Traditionalmeatchipsandpuddingeater’(2);‘Conservativeomnivore’and‘Lowdiversityvegetarian’(3);‘Higherdiversitytraditionalomnivore’and‘Highdiversityvegetarian’(4);and‘Healthconscious’(5).

Figure 3.7: Contribution of different food groups to daily diet costs of healthiest and least healthy group of eaters, adult women, UK, 1995–98

0.19 0.33

0.64

1.870.42

0.05

£0.00

£0.50

£1.00

£1.50

£2.00

£2.50

£3.00

£3.50

£4.00

Least healthy dietgroup

Most healthy dietgroup

Diet

cos

t per

day

Alcoholic beverages

Non-alcoholic beverages

Sauces/soups, savourysnacks, sweetsDairy, eggs, margarine,spreadsFish

Meat

Grains/nuts/seeds

Vegetables, fruit

Bread, cereals, potatoes,pasta, rice

9 8

29

49

17

1

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Least healthy dietgroup

Most healthy dietgroup

Prop

ortio

n of

die

t cos

t

Source: Data from Table 3 in Cade et al. 1999

3.2.3 Price changes and consumptionArecentrathercomprehensiveanalysis,correctingshortcomingsinearlierwork,hasassessedown-andcross-priceelasticitiesforalargesetoffoodgroups,acrossincomeandgeographicalsub-setsoftheUKpopulationfrom2001/02to2009(Tiffinetal.2011).Own-priceelasticitiesforfoodgroupsrangefromtheelastic-1.38forfruitandnutstotherelativelyinelastic-0.58formeat.(seeTable3.3).

Theauthorsthenusedtheseelasticitiestoseehowchangesinfoodpricesandfoodexpenditurewouldaffecttheintakeofnutrients–wherecross-priceeffectsbecomesignificant.Forexample,higherpricesoffatandstarch,dairyandeggproductswouldleadhouseholdstoconsumelessfatandenergy,asmightbeexpected;buttheywouldalsoleadtolowerconsumptionofvitaminsandmicronutrients.A10%increaseinthepriceof

dairyandeggproductsreducestheintakeofbeneficialnutrientssuchascalcium(-3.8%),iron(-3.5%),vitaminD(-4.6%)andzinc(-8.4%).Theeffectsofsubsidiescanbemodelled.Subsidisingvegetablesby5%increasesvegetableconsumptionoflow-incomehouseholdsby3.23%,butalsoraisesconsumptionofmeatby0.45%,decreasesalcoholby0.62%,decreasesfishby0.23%,andhasanegligibleeffectondairy,eggs,fatsandstarches.Itwouldincreasetheintakeofcarotenes(+2%),VitaminC(+0.8%)andfibre(+0.5%).

Somestudieshavelookedattheinfluenceonpurchasingofsalespromotionsthataffectprices.Hawkes(2009a)synthesisedtheliterature(notonlyfromtheUK)ontheinfluenceofsalespromotionsonfoodconsumption.Althoughpromotionsleadtosignificantsalesincreasesovertheshortterm,itwasnotclearifchangesinfoodconsumptionwouldpersistafterthepromotion,owingtolackofinformationonlonger-termeffects.

Figure 3.8 Diet costs (per calorie) and energy density, Scotland, 2011

£4.97£4.84

£4.40

£4.28

£3.76123

151

167

187

231

100

120

140

160

180

200

220

240

£3.50

£3.70

£3.90

£4.10

£4.30

£4.50

£4.70

£4.90

1 (least dense) 2 3 4 5 (most dense)

Quintile of energy density (food and milk)

Kcal

per

100

g

£ pe

r 200

0kca

l

Household and eating out combined cost Food and milk energy density - R axis

WCRF energy density recommendation - R axis

Source: Compiled fom data in Table 10 in Wrieden and Barton 2011

Notes: WCRF = World Cancer Research Fund. Kcal = kilocalories

Table 3.3: Estimates of food price elasticity in the UK, long-runChange in consumption in response to price change

Dairy & Egg Meat Fish Fruit & Nuts Veg. Fats & Starches Alcohol Expend.

Dairy&Eggs -1.00 -0.01 0.04 0.20 0.05 -0.13 0.01 0.85

Meat -0.09 -0.58 0.04 -0.03 0.08 -0.48 -0.09 1.14

Fish 0.56 0.29 -0.70 -0.43 -0.01 -0.26 -0.09 0.64

Fruits&Nuts 0.38 0.01 -0.06 -1.38 -0.01 0.38 -0.07 0.76

Veg. 0.12 0.12 0.00 -0.01 -0.65 -0.32 0.03 0.72

Fats&Starches -0.09 -0.09 -0.01 0.07 -0.10 -0.83 0.04 1.02

Alcohol -0.27 -0.28 -0.05 -0.35 -0.10 -0.28 -1.12 2.46

Source: Table 4 Tiffin et al. 2011

32 ODI Report

The rising cost of a healthy diet: Changing relative prices of foods in high-income and emerging economies 33

Box 3A: Cheap eats and healthy treats: affordable for families of obese children in the UK?

Affordabilityisoftencitedasafactorinpeople’sunhealthydiets.Arecentrandomisedtrialofobesechildrenagedfiveto16lookedatthisquestion,comparingtheiractualdietstoatheoretical,healthydietbasedontheEatwellPlate.Thehealthydietswereonlyslightlyadjustedversionsofthechildren’sexistingdiets–amoreacceptableoptionthanpricinganalternativedietfarremovedfromexistingones.Figure3A1showssomeexamplesofadjustmentsbetweenexistingandhealthydiets,andtheimpacttheywouldhaveonenergyintake.

Figure 3A1: Kcal changes associated with adjusting meals to healthier options

-300 -200 -100 0 Original meal Healthier option

2sliceswhitetoast,paté,crisps,cola,largeportionHoneynutCheerios®,semi-skimmedmilk

2sliceswhitetoast,low-fatbutter,apple,semi-skimmedmilk,1portionCheerios®,banana

Portionfriedchicken Portionroastchicken

Crisps,3sliceswhitebread,butter 4sliceswhitebread,roastchicken,LFbutter,salad

3slicesofpizza 2slicesvegetablepizza,mixedsalad,orangejuice

1packready-madelasagne ½packoflasagne,carrots,apple

Cola Zerocola

½pintblackcurrantsquash ½pintno-sugarblackcurrantsquash

1sliceofcheesecake,cream 1 fruit ice pop

1packofHaribo®Jellies 1portionricepuddingwithSSmilk

Source: Data from Table 1 in Banks et al. 2012

Bothdietswerepricedatthreeshops:aneighbourhoodmid-rangesupermarket,abudgetsupermarket,andthelocalhighstreet.Thechildren’sactualdietboughtatabudgetsupermarketwasthecheapest(£2.48/day).Thehealthieroptionatthesameshopcostanadditional33pence/day(£2.81).Thesameexerciseinamid-rangesupermarketincurredanadditionalcostof4penceperday(£3.40versus£3.44).

Shiftingfromtheunhealthyoptionboughtatamid-rangesupermarkettothehealthier,budget-outletoptioncouldsave59penceperday.Thehealthieroptionwascheaperthantheexistingdietifpurchasedonthehighstreet(£3.58versus£3.75),althoughforbothmenusthehighstreetwasthemostexpensiveoption–seeFigure3A2.Evenifswitchingfromtheexistingtoabudgetsupermarkethealthydiet,theextracostwouldbeonlyabout£10permonth.

Figure 3A2: Cost of different diets from three types of shop

2.48

3.4

3.75

2.81

3.443.58

£2.00

£3.00

£4.00

Budgetsupermarket

Supermarket Independentshops

Existing diet

Healthy diet

Source: Data from Table 3 in Banks et al. 2012

Theyconcludedthatformanyofthefamiliesofobesechildren,theextracostofeatinghealthilywouldnotnecessarilybeprohibitive,thoughcostmaybeabarrierforthemostdisadvantaged.Source: Banks et al. 2012

Studyingthespecificcaseoftheimpactofpromotionsonsugaryfoodpurchases–take-homeconfectionary,frozenconfectionaryandicecream,andnon-dietsoftdrinks–inScotlandbetween2006and2011,Revoredo-GihaandAkaichi(2014)foundthatconsumersdidindeedrespondtopromotionsonsugaryfoods.Inparticular,familieswithchildrensignificantlyincreasedpurchasesofsugaryproductsoverthesixyearsreviewed,owinginlargeparttopricepromotions.Moreover,theuseofsuchpromotionsbythefourlargestScottishsupermarketsincreasedovertime,suggestingthatretailersusedpromotionstokeeppeoplespendingovertherecession.

3.2.4 Effect of price changes – including taxes – on consumption and health in the UK

Severalstudieshavemodelledthelikelyimpactoftaxesandsubsidiesoneatinghabits,weightorhealth.

Briggsetal.(2013)modela20%taxonSSBsontheprevalenceofoverweightandobesityofpeopleaged16andoverintheUK.Own-priceelasticityforSSBis-0.92forconcentratedand-0.81fornon-concentrateddrinks,sothata20%taxonSSBdecreasesconsumptionoftheformerby15%andthelatterby16%.Tocompensate,consumptionofotherdrinksrises.Consumersonhigherincomestendtoswitchtowater,whilethoseonlowerincomesswitchtodietsoftdrinks,milk,andfruitjuice.

Annualrevenuefromsuchataxwasestimatedat£276million.Totalexpenditureondrinkswouldrisebyfor2.1%forthehighestincometercile,by1.7%forthemiddletercile,andby0.8%forthelowesttercile.ThetaxwouldreducethenumberofobeseadultsintheUKby1.3%(or180,000people),andthenumberofoverweightby0.9%(285,000people).Effectsonobesitywerestrongeratyoungerages,leadingtheauthorstoconclude:‘Taxationofsugarsweeteneddrinksisapromisingpopulationmeasuretotargetpopulationobesity,particularlyamongyoungeradults’(Briggsetal.2013).

Myttonetal.(2007)examinetheeffectsonnutrition,healthandexpenditureofextendingVATtoawiderrangeoffoodsintheUK.ConsumptionpatternsandelasticitydataweretakenfromtheNationalFoodSurveyofGreatBritain,14whilethehealtheffectsofchangingsaltandfatintakewerefrompreviousmeta‐analyses.Threescenarioswereconsidered:ataxonthemainsourcesofsaturatedfatindiets;ataxonfoodsdefinedasunhealthybyanutrient-profilingmodeldevelopedfortheFSA;andataxonfoodstoobtainthebesthealthoutcome.

Inthefirstscenario,theyfindthattaxingonlythekeysourcesofdietarysaturatedfatisnotlikelytoreduceprevalenceofcardiovasculardisease(CVD),mainlybecauseafallinsaturatedfatisoffsetbyariseinsalt

consumption.Inthesecondscenario,theyfindthattaxingunhealthyfoodsmightpreventsome2,300deathsperyear,chieflybyreducingsaltintake.Finally,inthethirdscenario,theyfindthatataxonawiderrangeoffoodscouldpreventupto3,200CVD-relateddeathsintheUKperyear,areductionof1.7%.Theauthorsconcludethatwhileataxoncertainfoodscanleadtounpredictablehealtheffectsifcross-priceelasticitiesofdemandarenottakenintoconsideration,acarefullytargeted‘fattax’couldhaveamodestbutmeaningfulinfluenceonpeople’sdietsandontheincidenceofCVD.

Inarelatedstudy,Nnoahametal.(2009)exploresimilartaxscenarios,aswellastax-subsidyscenarios,toassesstheimpactonmortalityfromnotonlyCVD,butalsocancer.TheyfindthatataxontheprincipalsourcesofdietarysaturatedfatisunlikelytoreduceCVD15orcancermortality.Indeed,ratherthanpreventingdeaths,itleadstoextradeathssincealthoughfatconsumptionfalls,sotoodoestheconsumptionoffruitandvegetablesowingtocross-priceelasticities.Ataxon‘lesshealthy’foods(definedaccordingtotheFSAnutrient-profilingmodel)couldevenincreaseCVDandcancerdeathsby35to1,300ayear,forsimilarreasons.

Incontrast,ataxon‘lesshealthy’foodscombinedwithasubsidyonfruitandvegetablesof17.5%couldavertupto2,900CVDandcancerdeathseveryyear,whiletaxing‘lesshealthy’foodsandusingalltaxrevenuetosubsidisefruitandvegetablescouldavertupto6,400CVDandcancerdeathsayear.Eachscenariowouldplaceahighereconomicburdenonlower-incomefamilies.Inthelasttwoscenarioshowever,manyofthelivessavedthroughtax-subsidyschemeswouldbethoseofpoorerpeople.

2.5 Summary of the UK literature

• Healthydietscostmorethanlesshealthydiets. • Overthelast10–20years,thecostoffruitandvegetableshasrisencomparedtootherfoods,andespeciallyprocessedfood.

• Cross-priceeffectsmatterinassessingthenutritionaleffectsofpricechanges.Taxesonfatorsaltcontentmayaffectconsumptionofother,complementaryfoods,leadingtolowerintakeofbeneficialnutrients.Usingtaxrevenuestosubsidisesuchcomplementaryfoodswouldcounterthiseffect.

14 TheelasticitiesusedhereandinNnoahametal.(2009)describedbelowarethesameasthosecriticisedbyTiffinetal.(2011)forbeingcalculatedonthebasisofoutdatedeconomicandstatisticalapproaches.

15 Thoughitreducescoronaryheartdisease(CHD),itincreasesstrokesbyagreaterextent.

34 ODI Report

The rising cost of a healthy diet: Changing relative prices of foods in high-income and emerging economies 35

3.4 Brazil

3.4.1 Consumption of (ultra) processed foodConsumptionofultra-processedfoodanddrink–thatis,ready-to-eatordrinkandfoodsthathavebeenindustriallypreparedfromingredients,typically‘energy-dense,fatty,sugaryorsalty,andformulatedtobehyper-palatable’(Monteiroetal.2012)–hasbeenrising,fromjustover80kgpercapitaperyearin1999toaround110kgpercapitaperyearby2013(Moubarac2014).Consumptionofsugarydrinks,asmodelledfromEuromonitordata,wasamongthehighestformiddle-andlow-incomehouseholds(Basuetal.2013).Someultra-processedfoods,suchasbreadandsausages,havelongformedpartofBraziliandiets,whileotherssuchascrisps,biscuits,energybars,andsugarydrinksaremorerecentadditions,encouragedbywidespreadmarketing(seeBox3B).

From1990to1996householdsinSãoPauloincreasedtheirspendingon‘industrialised’(highlyprocessed)food,whilespendinglessonsemi-preparedandnon-processedfoods(BarrettoandCyrillo2001).Morerecently,between1996and2009,surveysshowconsumersgettingsmallersharesofenergyfrommoretraditional,minimallyprocessedfoods,asthecontributionofultra-processedproductshasgrown(seeFigure3.9).

Energy-densefoodstendtobecheaperpercaloriethanotherfoods(seeFigure3.10).Mostoftheraworminimallyprocessedfoods–suchasfruitsandvegetables–havehighercostspercaloriethandothemoderatelyprocessedfoods–suchassugarandoils.Intermsofcostperkilocalorie,manyofthehighlyprocessedfoodswerealsorelativelycheap,exceptforreadymeals(RicardoandClaro2012).

3.4.2 Effect of food prices and incomes on consumption

Lookingattheinfluenceofthepriceoffruitandvegetablesontheircontributiontopeople’sdietsinBrazil,ClaroandMonteiro(2010)showedthata1%

fallintheirpricewouldincreasetheircontributiontocalorieintakeby0.79%.ForSãoPauloalowerelasticityofdemandwasfound,a1%pricefallleadingtoonlya0.2%increaseinconsumptionbycalorie(Claroetal.2007).Theyalsorecordedasmallcross-priceelasticity:a1%increaseinthepriceofotherfoodswouldincreasefruitandvegetablecontributiontocalorieintakeby0.07%.

Incomeelasticitieswerealsoreported.ForBrazilasawhole,theincomeelasticityofdemandforfruitandvegetableswasestimatedat0.27(ClaroandMonteiro2010),andforSãoPauloat0.04(Claroetal.2007):thatis,highlyinelasticresponsestoincome,althoughinbothcasesresponsesweregreateramongthoseonlowincomes,aswouldbeexpected.

Pricesareonlyoneinfluenceonfoodchoice:astudyofthedietsofwomenaged20to60yearsinurbanSãoLeopoldoinsouthernBrazil,(Lenzetal.2009)foundthathealthydietsweremorefrequentamongwomenwithhigherincomesandeducationallevel,whilewomenwithlowerincomesandlesseducationweremorelikelytoconsumedietswithhigherhealthrisks.

Figure 3.9: Proportion of energy from different food groups, urban household purchases, Brazil, 1996–2009

Source: Monteiro 2013

Box 3B: Marketing of processed food in Brazil

MonteiroandCannon(2012)seepowerfultransnationalfoodandsnackcompanies–dubbed‘BigFood’and‘BigSnack’–playingamajorroleinpeople’sincreasingconsumptionofultra-processed,fastorconveniencefoods.SuchfoodsareincreasinglydifficulttoavoidincontemporaryBrazil:

…wewentforlunchtoaworkers’restaurantneartheUniversityofSãoPaulo,whereatraditionalfreshlycookedmealofrice,beans,andachoiceofmeat,togetherwithmixedsalad,costtheequivalentof$US6.Wenoticedthatthebottledwaterofferedwas‘made’byaonceBraziliancompanynowownedbyCoca-Cola,andthattheartisanalwater-basedicelolliescontainingfruitjuice,whicharestillsoldbypedlarsonBrazilianbeachesandsuppliedbytraderstosimplerestaurants,hadbeenreplacedbyfatty,sugarybrandsofNestléicecream.Thesesameicecreams,togetherwithotherNestlé‘popularlypositionedproducts’,whichare‘targetedatandboughtbylowincomeconsumers’,arenowbeingsolddoor-to-doorintheoutskirtsofseveralBraziliancities,ontrainsandsubwaystations,inretailchainsthatsellelectronicandhouseappliances,andalsoon‘floatingsupermarkets’thattakeNestléproductstoremoteAmazonianvillages.(MonteiroandCannon2012)

4.3 Food taxesTaxesonlesshealthyfoodsmaybeparticularlyeffectiveinBrazil.InvestigatingtheeffectofataxonSSBs,Claroetal.(2012)usedhouseholdfood-consumptiondatacollectedin2002–2003forasampleofover48,000householdsinBrazil.Controllingfordemographicvariables,income,andpricesofotherfoodsanddrinks,theyfoundthatanincreaseinthepriceofsugarydrinksledtoreducedconsumption:a1%priceincreaseledtoa0.85%reductionincaloriesfromsugarydrinks.Thiseffectwasstrongerforpoorpeople(1.03%reduction),thanfornon-poor(0.63%reduction).Suchhighelasticitiessuggestthattaxesmighthaveastrongeffect,particularlyamongpoorerpeople:ataxof30%ontheaveragepriceofsugarydrinkswouldreduceaverageconsumptionbyabout25%(Claroetal.2012).

Figure 3.10: Cost per kilocalorie for different food groups in Brazil

$5.14

$3.45

$0.65

$1.90

$0.27 $0.34 $0.56 $0.39

$1.59

$2.38

$5.18

$2.06 $1.90

$0.00

$1.00

$2.00

$3.00

$4.00

$5.00

$6.00

Frui

ts, v

eget

able

s an

d gr

eens

Mea

ts, e

ggs,

and

milk

Stap

les

(cer

eals

, gra

ins,

tube

rs a

ndbe

ans)

Aver

age

for g

roup

1

Suga

r

Oils

and

fats

Flou

rs a

nd p

asta

Aver

age

for g

roup

2

Brea

ds, b

iscu

its, a

nd c

akes

Proc

esse

d m

eats

Read

y m

eals

Othe

r pro

cess

ed fo

ods

Aver

age

for g

roup

3

Group 1 (raw or minimallyprocessed)

Group 2 (moderatelyprocessed)

Group 3 (highly processed)

Braz

ilian

Rea

l per

1,0

00 k

cal

Source: Table 3 in Ricardo and Claro 2012

36 ODI Report

The rising cost of a healthy diet: Changing relative prices of foods in high-income and emerging economies 37

3.5 China

3.5.1 Food prices in ChinaCookingoilhasbecomecheapercomparedtootherfoodsinChina(seeFigure3.11)(LuandGoldman2010).

3.5.2 Effect of food prices on weight in ChinaToassesstheeffectsofrelativefoodpricesonbodyweightandbodyfatovertime,LuandGoldman(2010)usedacohortstudyof15,000adultsfromover200communitiesinChinafromtheChinaHealthandNutritionSurvey(CHNS)from1991to2006.Theyfoundthatdecreasesinthepriceofenergy-densefoodsledtogreaterbodyfat;aneffecttheycouldnotalwaysfindforbodyweight,leadingthemtoconclude:

…changesinfoodconsumptionpatternsinducedbyvaryingfoodpricescanincreasepercentagebodyfattoriskylevelsevenwithoutsubstantialweightgain.Inaddition,foodpricesandsubsidiescouldbeusedtoencouragehealthierfoodconsumptionpatternsandtocurbobesity.(LuandGoldman2010)

AlsousingdatafromtheCHNS,He’s2013studyofchildobesityfromthemid-1980stothemid-2000sfoundanobesogenicenvironmenttobeamuchmoreimportantfactorinshapingobesity-relatedriskbehaviourthanwasthemoreindividualdeterminant‘willpowerbasedonknowledge’.Interestingly,givenhowoftenruralparentsmigrateforworkandleavechildreninthecareofelderlyfamilymembers,shealsofound:

Childreninthecareofgrandparentsarehealthier,probablyduetothegenerallylowdegreeofaccesstoobesogenicfoodsandacloserintergenerationalrelationshipthatfacilitateseffectivecommunicationandpromoteshealthylifestyleformation.(He2013)

James(2008)seesthelong-termfallinpricesoffatsandoilscomparedtomoreexpensiveproductslikefish,meat,dairy,andfruitandvegetables,ascontributortogrowingobesityinChina.

3.6 Republic of Korea

3.6.1 Consumption trendsSofar,theRepublicofKoreahastakenwhatsomedescribeasauniquetrajectorythroughthe‘nutritiontransition’(seePopkinetal.2012),asaresultofstrongeffortstomaintainatraditionalKoreandietlowinfatandhighinvegetablesinthemidstofrapideconomicgrowthandWesternisation(Leeetal.2002).Thoughlargeshiftscanbeseenoverthepastdecade,drivenbyanincreasingopennesstoimportingfood,veryhighvegetableandmodestfatconsumptionseemto

havepersisted(Leeetal.2012).Estimatesofvegetableconsumptionperperson,particularlykimchi,haveremainedinmostyearsbetween260and290gramsadaysince1969(seeFigure3.12).

Somelesspositivetrendshavebeenspotted,however,withaveragedailyalcoholintakerisingfrom39kcalto82kcalperpersonbetween1998and2009,aswellasenergyintakefromSSBrisingamongteenagers(Leeetal.2012).Hanetal.(2013)documentedrisingprevalenceofSSBconsumptionamongadolescentsandallolderagegroupsbetween2001and2009:

SSBconsumptionprevalenceincreasedto38%,69%,70%,and50%by2009upfrom31%,66%,63%,and32%in2001amongadolescents,youngadults,adults,andtheelderly,respectively.

TheprevalenceofSSBconsumptionwashigheramongindividualsofrelativelyhighsocio-economicstatus,particularlyinthecaseoffruitdrinksandmiscellaneousSSBs–sports/energydrinks,coffee/teaproductsandflavouredmilk–ratherthansoda.

Astudyofchildandadolescentconsumptionfor2008–2011foundthatonaverage,childrenagedsevento12drankaround65mlofSSBsaday,whilechildrenaged13–18drank120mladay.Some12%ofchildrenandadolescents,however,weredrinkingmorethan300mlofSSBaday.Leeetal.(2013)foundhigherconsumptionwasassociatedwithhighoverallenergyintake,butlowconsumptionofmilk,fruitandvegetables(notmeetingthe400gadayrecommendationforthelatter).Furthermore,theyfoundthatbeingoverweightandobesewassignificantlyassociatedwithgreateroddsofhighSSBconsumptionamongboysaged7–12years.

Figure 3.11: Cooking oil prices, China, 1991–2006

Source: Figure 1 in Lu and Goldman 2010

3.7 Mexico

3.7.1 Consumption of ultra-processed food and drink

Mexico’sconsumptionofprocessed,oftenunhealthy,foodhasskyrocketed(seeFigure3.13).MexicohasledLatinAmericainpercapitaconsumptionofultra-processedfoodanddrinks.

MexicanconsumptionofsoftdrinksperpersonismuchhigherthanmightbepredictedbyitspercapitaGDP(Basuetal.2013)(seeFigure3.14):in2008,dailyconsumptionwasalmostonethirdmorethanintheUSA.

Sternetal.(2014),using24-hourdietrecallstudiesfrom1999and2012withnationallyrepresentativesamples,foundthattotaldailyenergyfrombeveragesincreasedbyabout45kcalforchildrenagedfiveto11,by57kcalforgirlsandyoungwomenaged12to19,andby96kcalforadultwomenaged20to49.

3.7.2 Food prices in Mexico

TherealpriceofmanybeveragesdecreasedovertimeinMexico,correspondingtolargeincreasesinMexicanconsumptionofenergy-containingbeverages–wholemilk,carbonatedandnon-carbonatedSSBs,fruitjuicewithvarioussugarandwatercombinationsadded,andalcohol–between1999and2006,sothatby2006,adultsandadolescentsobtained22%and20%respectivelyoftheirenergyintakefromthesebeverages(Barqueraetal.2008).

Figure 3.12: Vegetable consumption in Republic of Korea, 1969–2009, grams per person per day

58 79 91 109 96 80124 102 114 113 117

213 158

191 153 177

144

162181 177

214169

271

237

281262 273

224

286 283 291

327

286

0

50

100

150

200

250

300

350

1969 1973 1977 1981 1985 1989 1995 1998 2001 2005 2009

g/ca

pita

/day

Other vegetables Kimchi Total vegetables

Source: Figure 2 in Lee et al. 2012 Note: Values are presented as three-year or four-year moving averages. Kimchi intake is presented as a mean

from age>=1 between 1969 and 1995, and from age>=2 from 1998 to 2002.

Figure 3.13: Annual sales of selected ultra-processed foods and drinks, 12 Latin American countries, 1999–2013

Source: Figure 1 in Moubarac 2014 (based on Euromonitor Passport

Global Market Information Database (2014) and WHO Global

Burden of Disease)

Note: Ultra-processed food and drink products include: carbonated drinks,

fruit and vegetable juices, ‘sports’ and ‘energy’ drinks, breakfast cereals,

sweet and savoury snacks, confectionery, ice creams, biscuits, spreads,

sauces and ready meals. Quantity in litres was converted into kilograms.

Figure 3.14: Consumption of soft drinks per person per day, USA, UK, Mexico, Brazil, Korea and China, 2008

2.8

1.6

3.7

1.0 0.8

0.4

1.3  

0.7  

1.2  

0.7  

0.2  0.1  0  

0.5  

1  

1.5  

2  

2.5  

3  

3.5  

4  

US UK Mexico Brazil Korea, Rep China

Cups

per

per

son

per d

ay

Soft drinks

Of which, carbonates

Source: Constructed from USDA data and population estimates from

FAOSTAT. Original source: Euromonitor 2009

38 ODI Report

The rising cost of a healthy diet: Changing relative prices of foods in high-income and emerging economies 39

3.7.3 Food and drink taxes in MexicoInDecember2013,aftermuchdebate,thelegislatureinMexicoapprovedtwotaxes:apeso(aboutUS$0.07)perlitreonsugarydrinksanda5%taxonenergy-densefoodswithmorethan275kcalper100grams(seeBox2Aforanillustrationoffoodsabovethisthreshold).

ThetaxonSSBsalonecouldhelpprevent515,000newcasesofdiabetesby2030andleadtoUS$14billioninsavingsforthehealthsystem(ArantxaColchero,NationalInstituteofPublicHealth,reportedinMartinandCattan(2013)).

EarlyreportssuggestthetaxmaybehelpingtostempurchasesofsomeSSBsandsnackfoods.AccordingtotheNationalObesitySurveyof1,500adultsinMexicoinAugust2014,52%ofMexicansreducedintakeofSSBsin2014(EFE2014).Moreover,‘some98percentofrespondentssaidconsumingsoftdrinkscontributedtoobesityandcausedpeopletogetdiseaseslikediabetes,thesurvey’.PepsiCosnacksalesvolumedroppedby3%,whileCoca-Colaalsoreportedadeclineinsalesoverthefirsthalfof2014–thoughMexicostillhastheworld’shighestconsumptionofCoca-Colapercapita(RT.com2014).

3.8 Summary of literature from emerging economies

• Somereportslinkchangesindiets,aboveallthoseinvolvingmoreconsumptionofprocessedfoods,toprocessedfoodandcookingoilbecomingcheaperthanotherfoods.

• InLatinAmerica,therisingconsumptionultra-processedfoodsandSSBsiscomingunderthespotlight.Someseethisconsumptionastheconsequenceofheavymarketingbythelargecorporationsthatmanufacturemuchofthisfoodanddrink.

• ThepossibilityofusingtaxestoreduceconsumptionofprocessedfoodandSSBsisactivelybeingstudied,withmostauthorsseeingthepotentialtosignificantlyreduceconsumption.MexicohasintroducedtaxesonbothSSBsandenergy-densefood.These,whichcameintoeffectinJanuary2014,willbethefocusofintensescrutinytoseewhateffectstheyhave.

4. Data and methods

The rising cost of a healthy diet: Changing relative prices of foods in high-income and emerging economies 41

Keydataforthisreportareseriesforfoodpricesfrom1990torecentyears.Theaimwastocompilecomparableseriesofretailpricesforrepresentativefoodsacrossasampleofcountriesovertimefromatleast1990onwards,andwherepossiblefrom1980.Retail prices were sought for representative foods fromthefoodgroupssetoutinSection2,namely:staples,fruitsandvegetables;meat,seafood,anddairy;oils,fats,andsugar;andhighlyprocessedfoods.

Table4.1liststhefoodschoseninthefourcountriesplustheUK.Theaimwastoselectatleastonerepresentativefromeachfoodgroup,preferablyonewithlargeconsumption,and–otherthanfortheitemsintheprocessedcategory–foodsthatwereminimallyprocessedifatall.Inpractice,choicewaslimitedtotherangeofpriceseriesalreadycollectedineachcountry.Insomecountries,mostnotablyChina,priceswereavailableonlyforgroups

offoods,suchas‘cake,biscuitandbread’ratherthantheindividualproducts.

Ideallywewouldhavelikedtohavetheretailpricesforthesamefoodsacrossthefourcountries,butpublisheddatadidnotallowthat.Inanycase,theaimwastohaveatleastonefoodfromeachfoodgroup,choosingthosethataremosttypicallyconsumed.Sincedietsvarybycountry,relevantfoodsalsodiffer:forexample,thestaplefoodinKoreaisrice,whileinMexicoitismaize.

Justonepriceforeachfoodhasbeenused,eithernationalaveragesorthepricespaidinamajorcity.Itwasnotpossibletoobtainmoredetailonhowpricesvarywithincountriesbetween,forexample,urbanandruralareas,orbyfood.

Forthefourcountrystudies,throughourcollaborativegroup,dataweretakenfrompubliclyavailabledatabases,assetoutinTable4.2.

Table 4.1: Food prices analysed for each food group and country

Brazil China Korea, Rep Mexico UK

Staple Rice Grain Rice Tortillaandmaizeflour Flour(wheat)

Fruit or vegetables OrangesTomatoes Vegetables Dried and fresh fruits

Cabbage Tomato Freshvegetables

Freshgreenvegetables

Fats or sugar Sugar Soyoil

Oilandfat Vegetableoil Sugar

Oilsandediblefats Sugar

Sugar

Meat or dairy Beef Meat,poultryandproducts

Fish Chicken

Chicken Chicken

Highly processed Regularsausage Sweetbiscuit

Cake,biscuitandbread Ramen(noodles) Chocolateandsnacks Readymeals

Icecreamtub/block

Table 4.2: Data sources and deflators used for country cases

Country Deflator Data source

Brazil CPI • FoodpricedatacollectedbytheInstitutodeEconomiaAgrícola(InstituteofAgriculturalEconomics)ofSãoPauloStateforthemetropolitanareaofSãoPaulo,Brazil’slargestcity[2010populationof12.5millioninthemunicipality–themetropolitanareacontainsmore],from1980to2009.

• CPIdatafromFundaçãoInstitutodePesquisasEconômicas/UniversidadedeSãoPaulo,from1980to2009.Owingtohyper-inflationinthe1980sandearly1990s,significantcurrencychangestookplaceinMarch1986,January1989,August1993andJuly1994andhenceappropriatecorrectionshadtobemadetocomparepricesacrosscurrencyregimes,usinginformationfromFundaçãoInstitutodePesquisasEconômicas/UniversidadedeSãoPaulo(InstituteofEconomicResearch/UniversityofSãoPaulo).

China GDP • FoodpricedatafromChinaHealthandNutritionSurvey,andfoodpriceindexdatafromChinaStatisticalYearbooks.• GDPdeflatorfromIMF

Korea,Republic

CPI • FoodpriceindexdatafromtheConsumerPriceIndexforKorea(StatisticsKorea).Price-leveldatafromthemonthlyReportofCostofLivinginKoreaforJanuary2006.

• CPIdeflatorfromStatisticsKorea.CPIbyItem(Commodities&Services) http://kosis.kr/statHtml/statHtml.do?orgId=101&tblId=DT_1J0A112&conn_path=I3

Mexico CPI • FoodpriceindexdatafromINEGI(Mexico).Price-leveldataforMexicoCityfromProcuraduríaFederaldelConsumidor(Mexico)• CPIdeflatorfromWorldBankWDI

UK GDP • Unitfoodpricedata(nationalaverage)imputedfromsurveydataonspendingandconsumptionfromtheAdjustedNationalFoodSurveydata1974to2000,ExpenditureandFoodSurvey2001–02to2007andLivingCostsandFoodSurvey2008onwards.

• GDPdeflatorfromWorldBankWDI

Inmostcases,directlyobservedretailpriceswereused.ForMexico,however,afoodpriceindexwasusedandcalibratedtopricelevelsfromobservedpricesinMexicoCity.IntheUK,householdsurveysreportedbothspendingandquantityfor330foods.Henceitwaspossibletocreateunitpricesbydividingspendingbythequantity.Arguablythisproducesabettermeasureofprice,sinceunitpricescapturetheeffectsofthefrequentpricepromotionsseeninUKfoodretailing,andnotjusttherecommendedprice.

PriceserieshavebeendeflatedbyeithertheCPIortheGDPdeflatortoremovetheeffectofinflationandallowcomparisonovertime.Insomecases,suchastheUK,theCPIandtheGDPgivealmostidenticalresults.Inthreecases,however,theGDPdeflatorissignificantlystrongerthantheCPI:between1990and2012theGDPdeflatorexceededtheCPIaccountofinflationby21%for

China,by25%forMexico,andby59%forBrazil.Thosedifferencesmakenodifferencetocomparisonofchangesinrelativepriceswithincountry,butclearlyaffectcross-countrycomparisons,oraninterpretationofcomparablepricelevelsovertime.

Oncepricesweredeflated,indicesofthesesdeflatedpriceswereconstructedtoseehowmuchpricesofdifferentfoodsineachcountryhadchangedovertimefromthesamebaseyear.

Moreformaltestingofpricechangeswascarriedoutbyregressingtimeondeflatedandloggedpricesfrom1990toarecentyear,inmostcases2012.Thatallowedatestofwhetherasignificant(log-linear)trendcanbeinferred,andifso,whattheaverageannualpricechangehasbeen.Thekeyfindingsfromtheanalysisherecomefromtheseestimatedannualpricechanges.

42 ODI Report

5. Results of analysis: price movements

ThissectionreportswhatthedatashowforthefouremergingeconomiesandtheUK,thencomparestheseinsightstoderivebroaderinsights.AnnexIIIhasmoredetailsforeachofthecountries.Foreachcountry,pricesarepresentedintwoways.One,asconstantpricesperunitweight.Two,asindicesofthesepriceswithabasesettotheearly1980s;exceptforChinawherethemid-1990shadtobeusedsincetheseriesfortheonlyprocessedfoodinthesamplebeganthen,andforKoreawhereitwaspossibletotaketheindexbacktothelate1970s.Graphsofindicesmakeiteasiertocomparethesizeofchangesbetweenfoodswhentheyhaveverydifferentabsolutecostsperunitweight.

5.1 BrazilPricesofkeyfoodsinSãoPaulofrom1980to2009appearinFigure5.1,firstinconstant2009values,secondinconstantvaluesbutindexedtoacommon1980/82base.

Brazilianpricesarequitevolatileovertime,especiallyduringthe1980s.Inpartthesharpmovementsinpricesreflecttimesofhighinflationthatendedonlyinthemid-1990s.Despitethevariance,sometrendscanbeseen.Pricesoffruitandvegetables16havegainedthemost.Thelowestincreasesseenwereforrice,sausageandsoyoil.Sweetbiscuits,whichalongwithsausageareaprocessedfood,alsogainedquitealot.

Figure 5.1: Price of selected foods, São Paulo, Brazil, 1980–2009

(a) Constant 2009 prices

(b) Price indices, 1980/82 = 100

80

100

120

140

160

180

200

220

240

1980

1981

1982

1983

1984

1985

1986

1987

1988

1989

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

Inde

x: 1

980-

1982

=10

0

Tomato Orange Sweet biscuit

Beef Sugar Rice

Soy oil Regular sausage

Source: Data from Rafael Claro (Original source: data collected by the Instituto de Economia Agrícola (Institute of Agricultural Economics) of

São Paulo State for the metropolitan area of São Paulo, deflated to 2009 levels with CPI data and corrected for currency changes.

44 ODI Report

The rising cost of a healthy diet: Changing relative prices of foods in high-income and emerging economies 45

Allpricesriseintheseseries.Thatmaypartlyreflectthechoiceofdeflator,takenastheCPI.HadtheGDPdeflatorbeenused,somepriceswouldhavefallen,becauseitregistersafull60%moreinflationbetween1990and2012thantheCPI.Whilethisaffectsthestrengthoftrends,itdoesnotaffecttherelativechangesinpricesbetweenfoods.

5.2 ChinaPricesofselectedfoodsinChinafrom1989to2006appearinFigure5.2,firstinconstant2006values,secondinconstantvaluesbutindexedtoacommon1993base.

FoodpricesinChinahavebeennotablylessvariablefromyeartoyearthaninBrazil.Since1993,thepricesofriceandvegetableshaverisen;thoseforcake,biscuitandbreadhaveremainedalmostthesame;whilethoseforchicken,oilsandsugarhavefallen.

16 Orangesandtomatoeswerechosenasthesewerethemostcommonlyconsumedofthefruitandvegetablesintermsofgramspercapitainarecentconsumptionsurvey(datafromRafaelClaro).

Figure 5.2: Price of selected foods, China, urban areas, 1989–2006

(a) Constant 2006 prices

0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

4

6

8

10

12

14

16

18

20

1989

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

Yuan

per

kg

Yuan

per

kg

Sugar Oils

Chicken Cake, biscuit, bread

Rice - R axis Vegetables (commonly eaten) - R axis

(b) Price indices, urban China 1984–2012, 1994/96 = 100

405060708090

100110120130140150160170

Inde

x (1

994

to 1

996

= 1

00)

Grain Oil and fat

Meat, Poultry and its products Vegetables

Dried and Fresh Fruits Cake, biscuit and bread

Source: Data from Satoru Shimokawa, based on China Statistical Yearbooks. Indices were adjusted to constant 2012 levels by China’s GDP

deflator from the IMF.

5.3 Republic of KoreaPricesofselectedfoodsinKoreafrom1975to2013appearinFigure5.3,firstinconstant2006values,secondinconstantvaluesbutindexedtoacommon1975–77base.

Onlytwooftheselectedfoodsroseinconstantpricefrom1975–77:fish,themostconsumedanimalproduct,17byalmostfourtimes,andcabbage–thekeyingredientofkimchi,Korea’snationaldish,by60%.Allotherfoodsbecamecheaper,aboveallsugarandvegetableoil.

17 In2011,averagepercapitasupplyofdemersalandpelagicfishintheRepublicofKoreawasaround37kg/capita.Incomparison,around16kg/capitaofpoultrymeatwassupplied(FAOSTAT).

Figure 5.3: Prices of selected foods, Republic of Korea, 1975–2013

(a) Constant 2006 prices

2007

2008

2009

2010

2011

2012

2013

0

200

400

600

800

1000

1975

1976

1977

1978

1979

1980

1981

1982

1983

1984

1985

1986

1987

1988

1989

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

Kore

an W

on p

er 1

00g

(or 1

00m

l for

oil)

Rice Ramen Fish Chicken

Cabbage Vegetable oil Sugar

(b) Prices indexed to 1975/77=100

50

100

150

200

250

300

350

400

450

20

40

60

80

100

120

140

160

180

1975

1977

1979

1981

1983

1985

1987

1989

1991

1993

1995

1997

1999

2001

2003

2005

2007

2009

2011

2013

Inde

x of

con

stan

t 200

6 fis

h pr

ice

(197

5-19

77 =

100

)

Inde

x of

con

stan

t 200

6 pr

ices

(197

5-19

77

= 1

00)

Dried rice Ramen Chicken Cabbage

Vegetable oil Sugar Fish - R axis

Source: Data from Euna Han based on the Monthly Report of Cost of Living in Korea for January 2006 and the CPI for individual food items

between 1975 and 2013, deflated to 2006 values using the CPI.

46 ODI Report

The rising cost of a healthy diet: Changing relative prices of foods in high-income and emerging economies 47

5.4 MexicoPricesofkeyfoodsinMexicofrom1980to2014appearinFigure5.4,firstinconstant2010values,secondinconstantvaluesbutindexedtoacommon1980–82base.

Despiteconsiderableshort-termvariabilityinfoodprices,significanttrendscanbeseen.Somefoodcostshaverisensignificantly.By2014,maizeflourandtortillacostalmosttwicewhattheydidintheearly1980s.Policy

changeexplainsmuchofthissinceuntil1998tortillapriceswerecontrolled.AsFigure5.4shows,beforethatyear,priceshadnotrisensincetheearly1980sbutoncepriceswereliberalised,theydoubledwithinadecade.Tomatoesandotherfreshvegetablesalsoroseincost.Significantfallsinpricebyaround40%canbeseenforchicken,oilsandfats.Readymeals,chocolateandsnackshaveseensmallerpricesfalls.Sugarpriceshavebarelychanged.

Figure 5.4: Prices of selected foods, Mexico, 1980–2014

(a) Constant 2010 prices

60

70

80

90

100

110

120

0

10

20

30

40

50

60

1980

1981

1982

1983

1984

1985

1986

1987

1988

1989

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

2012

2013

2014

Peso

s pe

r kg

Peso

s pe

r kg

Tortilla and maize flour Oils and edible fats

Tomato Chicken

Fresh vegetables Chocolate and snacks - R axis

(b) Priced indexed to 1980–82=100

25

50

75

100

125

150

175

200

225

1980

1981

1982

1983

1984

1985

1986

1987

1988

1989

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

2012

2013

2014

Inde

x: 1

980-

1982

=10

0

Tortilla and maize flour Chocolate and snacks Ready meals

Oils and edible fats Tomato Fresh vegetables

Chicken Sugar

Source: Data from Joel Alberto Vargas Hernández, deflated by Mexican CPI from World Bank WDI. Original data on price indices for selected

foods for Mexico from INEGI Mexico, combined with data on actual food prices for 2015 from Mexico City (Procuraduría Federal del

Consumidor (México), accessed 25 February 2015.

Note: Food price series were constructed using national average price indices and price levels in early 2015 for Mexico City – these were

assumed as 2014 price levels for the purposes of simplicity. Resulting price trends indicate prices rather than actual observed prices.

5.5 United KingdomPricesofselectedfoodsintheUKfrom1974to2012appearinFigure5.5,firstinconstant2005–06values,thenaspriceindices;followedbyindicesforfoodgroups,weightedaccordingtorelativeweightofconsumption,indexedtoacommon1984–86base.

Panels(a)and(b)showthatthepriceoffreshgreenvegetableshasbeenrisinginconstanttermssincethe

1980s,whilethatoficecreamhasfallensignificantly.Pricesfortheotherthreeproductssawonlysmallchanges.

Forthefoodgroupsinpanel(c),pricesofstapleshaveriseninconstanttermsby50%sincethe1980s–largelyduetotherisingcostofpotatoes–whilethoseoffruitandvegetableshaverisenby30%.Pricesforfatsandsugarshavecomedownslightly;whilethoseformeatanddairyandprocessedfoodshavefallenby25%sincethe1980s.

Figure 5.5: Prices of selected foods, United Kingdom, 1974–2012

(a) Constant 2010 prices (b) Prices indexed to 1984–86=100

0

50

100

150

200

250

300

1974

1977

1980

1983

1986

1989

1992

1995

1998

2001

2004

2007

2010

Penc

e pe

r 100

g Flour

Ice cream tub/block

Sugar

Fresh green vegetables

Chicken

(c) Prices by food group, weighted and indexed to 1984–86=100

50

70

90

110

130

150

170

190

210

230

250

1974

1975

1976

1977

1978

1979

1980

1981

1982

1983

1984

1985

1986

1987

1988

1989

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

2012

Staples Processed Fats, sugars

Fruits and vegetables Meat and dairy

Source: Constructed from data from DEFRA

Note: Staples index includes flour, fresh potatoes, dried rice, and oatmeal/oat products. Processed index includes ice cream tub/block, chocolate

biscuits, chips (frozen or not), and ready meals and convenience meat. Fats and sugars index includes butter, vegetable oils excluding olive oil,

and sugar. Fruits and vegetables index includes fresh green vegetables, fresh onions (including leeks and shallots), fresh tomatoes, oranges and

bananas. Meat and dairy index includes eggs, chicken (uncooked, whole or in pieces), and whole milk.

48 ODI Report

The rising cost of a healthy diet: Changing relative prices of foods in high-income and emerging economies 49

5.6 Comparative analysisTocompareacrossthefivecountries,priceshavebeenanalysedfor1990to2012asfaraspossible:forChina,twofoodseriesbeginonlyin1993,whilealltheBrazilianpriceseriesendin2009.Forthisperiodasimpleregressionoftimeonloggedprices–itmakeslittledifferenceifthe

pricesarenotlogged–wascarriedout.Thisallowedtheseriestobecheckedtoseeiftherewasindeedasignificanttrendthroughtime,orwhetherthevariationsseenwerestochastic.Italsoallowedanestimatetobemadeofthedirectionandmagnitudeofthetrendinprices,averagedasalog-linearrelation.Table5.1setsouttheresults.

Table 5.1: Results of log-linear regression of time on prices

Country Product Period R-square F-stat Sig Slope t-stat

BRA Rice 1990-2009 0.39 11.39 0.0034 1.64% 3.37

BRA Sugar 1990-2009 0.00 0.07 0.8001 -0.15% -0.26

BRA Soyoil 1990-2009 0.56 22.55 0.0002 1.70% 4.75

BRA Beef 1990-2009 0.75 54.75 < 0.0001 1.86% 7.40

BRA Regularsausage 1990-2009 0.01 0.11 0.7474 0.07% 0.33

BRA Sweetbiscuit 1990-2009 0.54 20.94 0.0002 1.28% 4.58

BRA Tomato 1990-2009 0.87 122.82 < 0.0001 2.31% 11.08

BRA Orange 1990-2009 0.83 88.70 < 0.0001 2.51% 9.42

CHN Grain 1990-2012 0.37 12.49 0.0020 1.85% 3.53

CHN Oilandfat 1990-2012 0.11 2.64 0.1192 -0.51% -1.62

CHN Meat,Poultry&products 1990-2012 0.55 25.80 < 0.0001 1.30% 5.08

CHN Vegetables 1990-2012 0.94 319.45 < 0.0001 2.98% 17.87

CHN DriedandFreshFruits 1993-2012 0.08 1.51 0.2357 0.39% 1.23

CHN Cake,biscuitandbread 1993-2012 0.26 6.29 0.0219 -0.67% -2.51

KOR Dried rice 1990-2012 0.49 19.99 0.0002 -1.02% -4.47

KOR Ramen 1990-2012 0.92 254.11 < 0.0001 1.46% 15.94

KOR Chicken 1990-2012 0.04 0.92 0.3481 0.26% 0.96

KOR Fish 1990-2012 0.70 49.64 < 0.0001 1.83% 7.05

KOR Cabbage 1990-2012 0.85 114.55 < 0.0001 3.04% 10.70

KOR Vegetableoil 1990-2012 0.33 10.36 0.0041 1.06% 3.22

KOR Sugar 1990-2012 0.12 2.98 0.0988 0.57% 1.73

MEX Tortillaandmaizeflour 1990-2012 0.90 196.30 < 0.0001 3.73% 14.01

MEX Chocolateandsnacks 1990-2012 0.21 5.68 0.0266 -0.32% -2.38

MEX Readymeals 1990-2012 0.21 5.62 0.0274 -0.84% -2.37

MEX Oilsandediblefats 1990-2012 0.00 0.04 0.8408 -0.09% - 0.20

MEX Tomato 1990-2012 0.64 37.59 < 0.0001 2.04% 6.13

MEX Freshvegetables 1990-2012 0.77 69.61 < 0.0001 1.13% 8.34

MEX Chicken 1990-2012 0.54 24.66 < 0.0001 -1.79% -4.97

MEX Sugar 1990-2012 0.52 22.84 0.0001 1.39% 4.78

UK Flour 1990-2012 0.16 3.97 0.0594 0.65% 1.99

UK Icecreamtub/block 1990-2012 0.79 78.22 < 0.0001 -3.31% - 8.84

UK Sugar 1990-2012 0.01 0.25 0.6209 -0.13% -0.50

UK Freshgreenvegetables 1990-2012 0.97 694.18 < 0.0001 3.26% 26.35

UK Chicken 1990-2012 0.71 51.73 < 0.0001 0.71% 7.19

Source: Regressions of time on logged prices. Data from sources in Section 4.

Note: Cells in grey are insignificant estimates. Cells in italics are marginally significant estimates. Slope estimates expressed as percentages.

Nineofthe34estimates(showningreyinTable5.1)werenotsignificantatthe5%level,whileanotherthree(showninitalics)werenotsignificantatthe1%level.Removingthenineinsignificantestimatesandrearrangingthestatisticsbyfoodgroup,presentingtheslopeestimatesinachartgivesFigure5.6.

Twothingsarereadilyapparent.Oneisthatpricesoffruitandvegetableshaverisensubstantiallysince1990,mainlybybetween2%and3%ayearonaverage–orby55–91%between1990and2012.Theotheristhatfourofthesixprocessedproductsforwhichestimatesaresignificantshowpricefallssince1990.18Mostoftheotherfoodshaveseentheirpricesriseby1–2%ayear,withtheexceptionsofthepricefallsseenforriceinKoreaandchickeninMexico.

Figure 5.6: Estimated average annual price change from 1990

-4 -3 -2 -1 0 1 2 3 4

GrainDried rice

RiceTortilla and maize flour

Vegetable oilSoy oilSugar

Meat, Poultry and its productsFishBeef

ChickenChicken

VegetablesTomatoOrange

CabbageTomato

Fresh vegetablesFresh green vegetables

Cake, biscuit and breadRamen

Sweet biscuitChocolate and snacks

Ready mealsIce cream tub/block

CHKR

BRM

XKR

BRM

XCH

KRBR

MX

UKCH

BRBR

KRM

XM

XUK

CHKR

BRM

XM

XUK

Staples

Oil and sugar

Meat, poultry, fish

Fruit & vegetables

Processed foods

Source: Table 5.1

18 WhileramenpriceshaveriseninKoreasincetheearly1990s,theyhadfallenconsiderablyfromthemid-1970sto1991,sothatby2013theywerestillcheaperthan40yearsago.

50 ODI Report

6. Conclusion and discussion

6.1 Changing pricesThesinglemoststrikingfindingisthatinemergingeconomiesthepricesoffruitandvegetableshaverisensince1990(andinsomecasesforlongerthanthat)andmorerapidlythanthoseofmostotherfoods.ThisreplicateswhatcanbeseenforbothUSAandUKpricesoffruitandvegetables.

Limitedevidencealsosuggeststhatpricesofprocessedfoodshaveeitherfallenorhaveincreasedslowlyintheemergingeconomies,aresultthatagainreflectswhatcanbeseenfortheUSAandtheUK.

Twoqualificationsapply,however.First,pricetrendsseenareuneven,varyingovertime,betweencountriesandbetweensimilarfoodswithincountries.Whiletrendscanbeseen,otherfactorsmustbeatwork–includingthedisconcertingpossibilitythatpricesarenotcarefullyobservedandreported.

Second,thequalityandnatureoftheobservedfoodsmaychangeovertime,apointwelldocumentedfortheUSAwheremostfruitandvegetableshavemoreaddedvalueonthesupermarketshelfthantheyhad30ormoreyearsago–cutandtrimmed,bagged,washed,andavailableallyearround,nomatterwhattheseason.Mightthisalsobethecaseintheemergingeconomies?Thiswouldrequiremorestudy.

Ifthetrendsdetectedarereal,somethingotherthanillusionsseeninnoisydatadistortedbyhedonicchanges,theypromptquestionsaboutthereasonsforthem.If,forexample,technicalprogressinfarmingwereuniformsothatunitcostsofproductionwerefallingforallagriculturaloutput,andifadvancesinthelogisticsoffoodwholesalingandretailingweresimilarlyuniform,thenwemightexpectthecostsofmostfoodstomoveroughlyinlinewithoneanother.Butthatisnotthecase.

Sowhyhavefruitandvegetablesbecomemorecostlycomparedtootheritems?Itisnotasthoughtherehavenotbeentechnicaladvancesinhorticulture:onthecontrarysomeofthemostsophisticatedseeds,soilnutrition,watercontrol,andpreventionofpestsanddiseasesareseenpreciselyinthegardensandglasshousesinwhichsomanyfruitandvegetablesaregro.WhilethereisaworldofdifferencebetweenDutchheatedglasshousesandthetinyplotsofgreenbeansofcentralKenya,inbothcases,comparedtootheragricultureintheirneighbourhoods,thesesystemsarebothmoreintensiveandusemoresophisticatedtechnologythanmostotherlocalfarmenterprises.Moreover,advancesintransportmeanthatfruitandvegetablesaretradedmorethaninthepast,sothatretailmanagersshouldbeabletosourcefromlow-costsuppliersnomatterwheretheymaybe.

Hypothesescanbeimagined:horticulturemaywellhaveasteppedsupplyfunction,sothatwhilesmallquantitiesoffruitandvegetablescanbesuppliedatlowunitcost,onceaparticularvolumeisreached,costsrapidlyescalatetoasignificantlyhigherlevel.Itmayalsobethatthechangesinqualitynotedexplaintheincreasedrelativeprices.Or,it

maynotbeamatterofcostbutofincreaseddemandfromthoseconsumerswhoappreciatethehealthbenefitsoffruitandvegetables.Thesehypothesesmeritaseparatestudy.

Whydoesnotthesameapplytosomeprocessedfoods?Onepossibilityisthatmuchprocessedfooddoesnotrelyoncostlyfarmingredients,butratherismanufacturedfromrelativelycheapingredients,theaddedvaluebeinglargelyinfactoryprocessesofcombiningtheingredientsandenhancingtheirflavour.Advancesinmanufacturingandflavouringprobablyhelpreduceunitcostsinfactory.Thatsaid,processedfoodsarenotuniforminqualityandpricing,sinceforanysub-category,thereareusuallyproductsthatarebranded,soldontheirspecialcharacteristics,usuallywithapricepremium–asapplies,forexample,tosoftdrinksthatcompetewithcheaper,unbrandedoptions.ThismayexplainwhynotalltheprocessedfoodsconsideredinTable5.8showdecliningconstantprices.Onceagain,additionalstudiesmightshedlightonthis.

6.2 Do prices matter … and might taxes work?

Evidencepresentedintheliteraturereviewsuggeststhatpricesdoaffectconsumption,especiallyforpeopleonlowincomes.Own-priceelasticityofdemandforfoodmayberelativelyinelastic,butnotthatinelastic–USestimatesofown-priceelasticityformostfoodsareabove(minus)0.45.Hencepricechangesdoinfluencediets.Moreover,itisstrikingtoseethatintheUSA,wheremanyconsumerscanaffordmostfoodtheywish,consumptionissensitivetoprice,especiallyforSSBs–presumablybecausethereareplentyofalternativesoftdrinksifonebecomesmoreexpensivethantherest.

Henceitisnosurprisetoseemuchstudyofthepotentialoftaxesonlesshealthyoptionstoreducetheirconsumption,perhapsevenwithsubsidiesonmorehealthyoptionstoraisetheirs.Mostsuchstudiesindicatethatimposingtaxeswouldreduceconsumption.Buttwoqualificationsapply.

Oneisthattheremaybecross-priceeffects,wherebywhentaxesraisethecostofaparticularfood,notonlydoesitsconsumptionfall,butsotoodoesthatofcomplements(foodswhicharetypicallyconsumedtogether,suchasbreadandbutter).Whenthosecomplementscontainvaluednutrientsitisthuspossiblefortaxestoreducethequalityofdiet.Intheorythisproblemcanreadilybetackledbyplacingasubsidyonthevaluedcomplementtooffsetthecross-priceeffect.Inpractice,learningwhichfoodsreallyarecomplements,towhatextent,andthendetermininganoptimallevelofsubsidy,couldleadtoathicketofregulationsthathavetobeadjustedinthelightofemergingevidence,creatinghighadministrativecostsandgivingtheimpressionthatsuchfiscalmeasuresarejusttoodifficulttocontemplate.Thequestionishowstrongcross-priceeffectsareandwhethertheymayberemediedbyothermeasurestoencouragehealthierdiets.

52 ODI Report

The rising cost of a healthy diet: Changing relative prices of foods in high-income and emerging economies 53

Thesecondistheapparentlyseductiveargumentthatsmalltaxeswouldcreateonlysmalleffects:thatconsiderablechangeinconsumptionwouldrequirehightaxesthatwouldlookdisproportionateandunfair–say,morethantherateofVATof20%intheUKcase.Butalogicalflawapplieshere.Thepolicyquestionisnotsomuch,‘howlargeataxwouldbenecessarytobringdownconsumptionoflesshealthyfoodXtorecommendedorinsignificantlevels’,but‘howmuchbenefitwouldbederivedfromimposingapoliticallyacceptabletaxonlesshealthyfoodX?’Theanswertotheformerquestionmaybeanumbersohighastobedismissedfromthedebate;buttheanswertothelatterquestionmaybeasstrikingasthatprovidedbyNnoahametal.(2009)fortheUK:thattaxesandsubsidiesoflessthan20%couldsavenofewerthan6,400prematuredeathsayearfromcoronaryheartdiseaseandcancers.If6,400peopleweretodieinacatastrophicaccident,massiveresourceswouldbedeployed

topreventarepetition.RoaddeathsintheUKarefarfewerthanthisnumber(lessthan2,000ayear),toavoidwhichallmannerofregulationsapplytoroadusers.Theargumentaboutsmalltaxes,smallgainsistantamounttoargumentsthatcondemndoinggoodbecauseperfectionisunattainable–‘le mieux est l’ennemi du bien’,asVoltaireputit.

Intermsofwhatmightbetaxedandsubsidised,thisreportsuggeststhatenergy-densefoodsmightbetaxed,whilefruitandvegetableswhosepricesoftenrisecomparedtootherfoods,mightbesubsidised.

Muchcomesdowntothepoliticalappetitetocontemplatetaxingfoods(seeBox6A).EventsinMexicosuggestthatsomeemergingeconomiesmaystealamarchonHICsinthisrespect.TheevidencepresentedinthisreportsuggeststhattheMexicantaxesshoulddoconsiderablegood,therebyprovidingvaluablelessonsforotherdevelopingandemergingeconomies.

Box 6A: How feasible are taxes on foods?

Sugar,rum,andtobacco,arecommoditieswhicharenowherenecessariesoflife,whicharebecomeobjectsofalmostuniversalconsumption,andwhicharethereforeextremelypropersubjectsoftaxation.AdamSmith, The Wealth of Nations

InboththeUSAandtheUK,althoughratesofoverweightandobesityhavelongbeenrecognisedinhealthcarecirclesashavingreachedepidemicproportions,policy-makershavebeenslowtorecognisetheseverityoftheissue.LangandRayner(2007)suggestedthismaystemfromsloworineffectiveadvocacyworkbypublichealthproponents;fromevidencethatisnoteasilytranslatedintopolicy;orfromalackofpoliticalchampions.Someofthesefactorsarechanging,thoughprogressisslow.IntheUSAinparticularpoliticalchampionshaveemergedinrecentyears:MichelleObama’scampaignagainstchildobesity‘Let’sMove’waslaunchedin2010,alongsidetherelatedPartnershipforaHealthierAmerica(PHA);19whilethethenmayorofNewYork,MichaelBloomberg,attempted,unsuccessfully,20togettougherregulationsonthesizeoffizzydrinktotakeeffectin2013.ChampionsarelessvisibleintheUK,mostofthosebeingcelebritycookssuchasJamieOliverratherthanpoliticians.Byandlarge,economicmeasurestocontrolobesityarenotcommoninfooddebatesintheUK:

…beyondthewidelydebatedfoodtaxes,theuseoffinancialmechanismstoencouragehealthierdietshasnotbeenavisiblepartofthepolicydebateabouthealthyeatingandobesityintheUnitedKingdom.(Hawkes2009b)

TowhichLangandRayner(2009)add:

ThereisapowerfultemptationinGovernmenttolimitactionstoachoice-based,personalizationapproach,inpartbecausethisstyleofinterventionisalignedtothecommercialsector’sowncustomermanagementandmarketingmethods,butalsobecauseacross-societyapproachappearssobiginconceptionthatfailureisassumed.

Yetthismaybechanging:

Oncedispatchedtothebottomdrawofpolicyoptionstoaddressunhealthyeating,foodtaxesnowseembackintheouttrayofEuropeanpolicymakers.EvenDavidCameronmadeanoffhandquiprecentlysuggestingthatthisissomethingtheBritishGovernmentmightexplore.(Hawkes2012)

IntheUSAtheDietaryGuidelinesAdvisoryCommittee’sreporttotheSecretaryofHealthandHumanServicesandtheSecretaryofAgriculture(USDA,2015)hasjustbeenreleased,whichrecommendstaxandtax-subsidypolicies:

AlignnutritionalandagriculturalpolicieswithDietaryGuidelinesrecommendationsandmakebroadpolicychangestotransformthefoodsystemsoastopromotepopulationhealth,including the use of economic and taxing policiestoencouragetheproductionandconsumptionofhealthyfoodsandtoreduceunhealthyfoods.Forexample,earmark tax revenues from sugar-sweetened beverages, snack foods and desserts high in calories, added sugars, or sodium, and other less healthy foods for nutrition education initiatives and obesity preventionprograms.(USDA2015;emphasisadded)

Althoughindustrylobbiesmaycampaignhardagainsttaxes–theAmericanBeverageAssociationspentUS$7.7millionopposingaproposedtaxonSSBsinSanFrancisco(Rt.com2014)–thepublicmaybemorebehindtax/subsidyinitiativesthanpoliticiansrealise.ApollofNewYorkresidentsfoundthat:

…52%supporteda‘sodatax,’butthenumberroseto72%whenrespondentsweretoldthattherevenuewouldbeusedforobesityprevention.(BrownellandFrieden2009)

MonteiroandCannon(2012)arguethattaxesmaybemoreacceptablesociallyandpoliticallyinBrazilandotherdevelopingcountries,comparedtosomeHICs:

‘…theviewsofmanycommentatorsandpolicy-makersintheSouthareinsharpcontrastwiththeircounterpartsintheNorth.IncountriesliketheUS,thegeneraltendencyistodealwithfood,nutrition,andpublichealthinisolationasmatterslargelyofinformation,education,and“individuallifestyleadjustments”designedtoreducetheriskofvariousdisabilitiesanddiseases.ButinBrazilandothercountriesintheSouth,foodisseenbymostindependentscholarsandpolicy-makersaspartofamuchbroaderdiscoursethatinvolvesgeneralwell-being,thefamily,friendship,commensality,culture,sustainablelivelihoods,environmentalpreservation,nationalidentityandsovereignty,aswellaspersonalandpublichealth.(MonteiroandCannon2012)

Theyarenotaloneinseeingthegovernmentsofemergingeconomiesasmorepreparedtoactforbetterhealth,asJames(2008)arguesforChina:

Chinahastraditionallybeenfarmoreresponsivetothevalueofpolicieswhichlimitinequalitiesandestablishstandardsofcarethanmanywesterngovernments,whohaveyettorecognizethattheindividualisticfree-marketapproachtoobesitypreventionisguaranteedtofail.Chinacouldthereforeleadtheway:ifitfollowswesternapproaches,thehealthandeconomicburdenwillbecomeunsustainable.

19 RecentinitiativeslinkedtothisincludethePHA’s(ForinfoonPHAseehttp://ahealthieramerica.org/about/about-the-partnership/)‘DrinkUp’campaign(seehttp://www.youarewhatyoudrink.org/about/)toencouragepeopletodrinkmorewater,andtheplannedFNV(FruitandVegetables–seehttp://www.fnv.com/)celebrity-backedmarketingcampaigntopromoteconsumptionofFNVasabrand,usingsophisticatedadvertisingtechniques.

20 Ultimatelythisplanfailed:seehttp://en.wikipedia.org/wiki/Sugary_Drinks_Portion_Cap_Rule/.

54 ODI Report

The rising cost of a healthy diet: Changing relative prices of foods in high-income and emerging economies 55

ReferencesAggarwal,A.,P.MonsivaisandA.Drewnowski(2012)‘Nutrientintakeslinkedtobetterhealthoutcomesareassociated

withhigherdietcostsintheUS’,PLoS ONE 7(5):e37533.doi:10.1371/journal.pone.0037533.An,R.(2013)‘EatingBetterforLess:EffectivenessofFinancialIncentivesinModifyingDietaryandGroceryShopping

Behavior’.RandCorporationDissertation.RGSD-311(http://www.rand.org/pubs/rgs_dissertations/RGSD311.html).Andreyeva,T.,M.W.LongandK.D.Brownell(2010)‘Theimpactoffoodpricesonconsumption:asystmaticreview

ofresearchonthepriceelasticityofdemandforfood’,American Journal of Public Health100(2):216-222.doi:10.2105/AJPH.2008.151415.

Arnoult,M.H.,R.TiffinandW.B.Traill(2008)‘ModelsofNutrientDemand,TaxPolicy&PublicHealthImpact’.Report03in‘ImplicationsofaNutritionDrivenFoodPolicyforLandUseandtheRuralEnvironment’.UniversityofReading,UK.

Auld,M.C.,andL.M.Powell(2009)‘Economicsoffoodenergydensityandadolescentbodyweight’,Economica76:719-740.

Bahia,L.,E.S.Coutinho,L.A.Barufaldi,A.AbreuGde,T.A.Malhão,C.P.deSouzaandD.V.Araujo(2012)‘Thecostsofoverweightandobesity-relateddiseasesintheBrazilianpublichealthsystem:cross-sectionalstudy’,BMC Public Health18(12):440.doi:10.1186/1471-2458-12-440.

Banks,J.,J.Williams,T.Cumberlidge,T.Cimonetti,D.J.Sharp,andJ.P.Shield(2012)‘Ishealthyeatingforobesechildrennecessarilymorecostlyforfamilies?’British Journal of General Practice62(594):e1-5.doi:10.3399/bjgp12X616300.

BarqueraS.,L.Hernández-Barrera,M.L.Tolentino,J.EspinosaJ.,S.W.Ng,J.A.RiveraandB.M.Popkin(2008)‘EnergyintakefrombeveragesisincreasingamongMexicanadolescentsandadults’,Journal of Nutrition 138(12):2454-2461.

BarrettoS.AandD.C.Cyrillo(2001)‘AnalysisofhouseholdexpenditureswithfoodinthecityofS.Paulointhe1990s’,Revista de Saúde Pública 35:52-9.

Basu,S.,P.Yoffe,N.Hills,andR.H.Lustig(2013)‘Therelationshipofsugartopopulation-leveldiabetesprevalence:aneconometricanalysisofrepeatedcross-sectionaldata’,PLoS ONE 8(2):e57873.doi:10.1371/journal.pone.0057873.

Bernstein,A.M.,D.E.Bloon,B.A.Rosner,M.FranzandW.C.Willett(2010)‘RelationoffoodcosttohealthfulnessofdietamongUSwomen’,American Journal of Clinical Nutrition (http://ajcn.nutrition.org/content/92/5/1197.long).

Beydoun,M.A.,L.M.Powell,andY.Wang(2008)‘Theassociationoffastfood,fruitandvegetablepriceswithdietaryintakesamongUSadults:istheremodificationbyfamilyincome?’Social Science and Medicine 66(11):2218-29.doi:10.1016/j.socscimed.2008.01.018.

Beydoun,M.A.,L.M.Powell,X.ChenandY.Wang(2011)‘Foodpricesareassociatedwithdietaryquality,fastfoodconsumption,andbodymassindexamongU.S.childrenandadolescents’,Journal of Nutrition 141(2):304-11.doi:10.3945/jn.110.132613.

Blanchflower,D.G.,A.J.OswaldandS.L.Stewart-Brown(2012)‘Ispsychologicalwell-beinglinkedtotheconsumptionoffruitandvegetables?’WorkingPaper.WarwickEconomicsResearchPaperSeries(TWERPS)996.Coventry:EconomicsDepartment,UniversityofWarwick.(PermanentWRAPurl:http://wrap.warwick.ac.uk/53081).

Bleich,S.,D.Cutler,C.MurrayandA.Adams(2007)‘Whyisthedevelopedworldobese?’NBERWorkingPaper12954.Cambridge,MA:NationalBureauofEconomicResearch(http://www.nber.org/papers/w12954).

Brambila-Macias,J.,B.Shankar,S.Capacci,M.Mazzocchi,F.J.A.Perez-Cueto,W.VerbekeandW.B.Traill(2011)‘Policyinterventionstopromotehealthyeating:areviewofwhatworks,whatdoesnot,andwhatispromising’,Food and Nutrition Bulletin 32(4).365–375

Briggs,A.D.M.,O.T.Mytton,A.Kehlbacher,R.Tiffin,M.RaynerandP.Scarborough(2013)‘Overallandincomespecificeffectonprevalenceofoverweightandobesityof20%sugarsweeteneddrinktaxinUK:econometricandcomparativeriskassessmentmodellingstudy’,British Medical Journal 347:f6189.doi:10.1136/bmj.f6189.

Butland,B.,S.Jebb,P.Kopelman,K.McPherson,S.Thomas,J.MardellandV.Parry(2007)‘Foresight:TacklingObesities:FutureChoices-ObesitySystem’.London:UKGovernmentOfficeforScience,UKDepartmentofInnovation,Universities,andSkills(https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/295153/07-1177-obesity-system-atlas.pdf).

Cade,J.,H.Upmeier,C.Calvert,andD.Greenwood(1999)‘Costsofahealthydiet:analysisfromtheUKWomen’sCohortStudy’Public Health Nutrition 2(4):505-512.

Campos,P.,A.Saguy,P.Ernsberger,E.OliverandG.Gaesser(2006)‘Theepidemiologyofoverweightandobesity:publichealthcrisisormoralpanic?’International Journal of Epidemiology 35:55-60.

Capacci,S.,M.MazzocchiandB.Shankar(2012)‘TheregionalpriceofjunkfoodsrelativetohealthyfoodsintheUK:indirectestimationofatimeseries,1997–2009’.Paperpresentedat86thAnnualConferenceoftheAgriculturalEconomicsSociety,UniversityofWarwick,Coventry,16–18April2012

Caraher,M.andG.Cowburn(2005)‘Taxingfood:implicationsforpublichealthnutrition’,Public Health Nutrition 8(8):1242-1249.

Carden,T.J.andT.P.Carr(2013)‘Foodavailabilityofglucoseandfat,butnotfructose,increasedintheU.S.between1970and2009:analysisoftheUSDAfoodavailabilitydatasystem’,Nutrition Journal 12:130.doi:10.1186/1475-2891-12-130.

Carlson,A.,andE.Frazão(2012)‘Arehealthyfoodsreallymoreexpensive?Itdependsonhowyoumeasuretheprice’,USDA ERS Bulletin96.Washington,DC:UnitedStatesDepartmentofAgriculture.

Chan,R.S.M.andJ.Woo(2010)Preventionofoverweightandobesity:howeffectiveisthecurrentpublichealthapproach’,International Journal of Environmental Research and Public Health 7:765-783.doi:10.3390/ijerph7030765.

Chou,S.,M.GrossmanandH.Saffer(2004)‘Aneconomicanalysisofadultobesity:resultsfromtheBehavioralRiskFactorSurveillanceSystem’,Journal of Health Economics 23:565-587.

Christian,T.andI.Rashad(2009)‘TrendsinU.S.foodprices,1950-2007’,Economics and Human Biology 7(1):113-20.Claro,R.M.andC.A.Monteiro(2010)‘Familyincome,foodprices,andhouseholdpurchasesoffruitsandvegetablesin

Brazil’,Revista Saúde Pública 44(6).Claro,R.M,H.C.E.doCarmo,F.M.S.MachadoandC.A.Monteiro(2007)‘Income,foodprices,andparticipationof

fruitandvegetablesinthediet’,Revista Saúde Pública 41(4).Claro,R.M.,R.B.Levy,B.M.PopkinandC.A.Monteiro(2012)‘Sugar-sweetenedbeveragetaxesinBrazil’,American

Journal of Public Health 102(1):178-183.doi:10.2105/AJPH.2011.300313.Connell,C.L.,J.M.Zoellner,M.K.Yadrick,S.C.Chekuri,L.B.CrookandM.L.Bogle(2012)‘EnergyDensity,Nutrient

Adequacy,andCostperServingCanProvideInsightintoFoodChoicesintheLowerMississippiDelta’.SocietyforNutritionEducationandBehavior.PublishedbyElsevierInc.

Courtemanche,C.(2009)‘Longerhoursandlargerwaistlines:therelationshipbetweenworkhoursandobesity’,Forum for Health Economics and Policy 12(2).

Crino,M.,G.Sacks,S.Vandevijvere,B.SwinburnandB.Neal(2015)‘TheInfluenceonPopulationWeightGainandObesityoftheMacronutrientCompositionandEnergyDensityoftheFoodSupply’.CurrentObesityReports.TopicalCollectiononEtiologyofObesity.SpringerScienceandBusinessMedia,NewYork.

Darmon,N.,A.BriendandA.Drewnowski(2004)‘Energy-densedietsareassociatedwithlowerdietcosts:acommunitystudyofFrenchadults’,Public Health and Nutrition 7(1):21-7.

Davis,G.C.andA.Carlson(2012)‘HowSpuriousistheRelationshipBetweenFoodPriceandEnergyDensity?ASimpleProcedureandStatisticalTest’.PaperforpresentationattheAgriculturalandAppliedEconomicsAssociationAnnualMeeting,Seattle,WA(http://ageconsearch.umn.edu/bitstream/124716/2/SpuriousAAEAFinal%202.pdf).

Doak,C.M.,L.S.Adair,C.MonteiroandB.M.Popkin(2000)‘OverweightandunderweightcoexistwithinhouseholdsinBrazil,ChinaandRussia’,Journal of Nutrition 130:2965-2971.

Dowler,E.A.andY.O.Seo(1985)‘Assessmentofenergyintake:estimatesoffoodsupplyvmeasurementoffoodconsumption’,Food Policy 1(3):278-288.

Drewnowski,A.(2007)‘Therealcontributionofaddedsugarsandfatstoobesity’,Epidemiologic Reviews 29(1):160-171.

Drewnowski,A.(2010)‘ThecostofUSfoodsasrelatedtotheirnutritivevalue’,American Journal of Clinical Nutrition 92(5):1181-1188.doi:10.3945/ajcn.2010.29300.

Drewnowski,A.andC.D.Rehm(2013)‘Vegetablecostmetricsshowthatpotatoesandbeansprovidemostnutrientsperpenny’,PLoS ONE 8(5):e63277.doi:10.1371/journal.pone.0063277.

Drewnowski,A.andP.Eichelsdoerfer(2009)‘TheMediterraneandiet:doesithavetocostmore?’Public Health Nutrition 12(9A):1621-1628.doi:10.1017/S1368980009990462.

Du,S.,T.A.Mroz,F.ZhaiandB.M.Popkin(2004)‘RapidincomegrowthadverselyaffectsdietqualityinChina–particularlyforthepoor!’Social Science & Medicine 59(7):1505-1515.

Duffey,K.J.,P.Gordon-Larsen,J.M.Shikany,D.Guilkey,D.R.JacobsJr.andB.M.Popkin(2010)‘Foodpriceanddietandhealthoutcomes:20yearsofTheCARDIAStudy’, Archives of Internal Medicine 170(5):420-426.doi:10.1001/archinternmed.2009.545.

EFE(2014)‘MexicansCutDownonConsumptionofSugaryDrinks,SurveyShows’,14October(http://laht.com/article.asp?ArticleId=2356488&CategoryId=14091).

Elliott,E.,O.ParryandJ.Ashdown-Lamber(2006)‘Evaluationofcommunityfoodco-opspilotinWales’.WorkingPaper85.SchoolofSocialSciences,CardiffUniversity(http://www.cf.ac.uk/socsi/resources/wrkgpaper-85.pdf).

Eyles,H,C.N.Mhurchu,N.NghiemandT.Blakely(2012)‘Foodpricingstrategies,populationdiets,andnon-communicabledisease:asystematicreviewofsimulationstudies’,PLoS Med 9(12):e1001353.doi:10.1371/journal.pmed.1001353.

56 ODI Report

The rising cost of a healthy diet: Changing relative prices of foods in high-income and emerging economies 57

Ezzati,M.,S.VanderHoorn,C.M.M.Lawes,R.Leach,W.P.T.James,A.D.Lopez,A.RodgersandC.J.L.Murray(2005)‘Rethinkingthe“diseasesofaffluence”paradigm:globalpatternsofnutritionalrisksinrelationtoeconomicdevelopment’,PLoS Med 2(5):e133.doi:10.1371/journal.pmed.0020133.

FAO(2001)FOOD BALANCE SHEETS: A handbook.Rome:FoodandAgricultureOrganization(http://www.fao.org/docrep/003/x9892e/X9892e01.htm#P85_16086).

Fletcher,J.M,D.FrisvoldandN.Tefft(2010)‘Taxingsoftdrinksandrestrictingaccesstovendingmachinestocurbchildhoodobesity’,Health Affairs 29(5):1059-66.doi:10.1377/hlthaff.2009.0725.

Formisano,A.,M.Hunsberger,K.Bammann,B.Vanaelst,D.Molnar,L.A.Moreno,M.Tornaritis,T.Veidebaum,L.Lissner,G.BarbaandA.Siani(2014)‘Familystructureandchildhoodobesity:resultsoftheIDEFICSProject’,Public Health Nutrition 17(10):2307-2315.

Fox,G.andV.W.Ruttan(1983)‘AguidetoLDCfoodbalanceprojections’,European Review of Agricultural Economics 10:325-356.

French,S.A.(2005)‘Publichealthstrategiesfordietarychange:schoolsandworkplaces’,Journal of Nutrition 135(4):910-912.

Friel,S.,O.Walsh,andD.McCarthy(2006)‘Theironyofarichcountry:issuesoffinancialaccesstoandavailabilityofhealthyfoodintheRepublicofIreland’,Journal of Epidemiology and Community Health 60:1013-1019.doi:10.1136/jech.2005.041335.

Gelbach,J.B.,J.KlickandT.Stratmann(2009)‘CheapDonutsandExpensiveBroccoli:TheEffectofRelativePricesonObesity’.AvailableatSSRN:http://ssrn.com/abstract=976484

Gigante,D.P.,C.G.Victora,A.Matijasevich,B.L.HortaandF.C.Barros(2013)‘AssociationoffamilyincomewithBMIfromchildhoodtoadultlife:abirthcohortstudy’,Public Health and Nutrition 16(2):233-239.

Goldman,D.,D.LakdawallaandY.Zheng(2011)‘FoodPricesandtheDynamicsofBodyWeight’.NBERWorkingPaper15096.Cambridge,MA:NationalBureauofEconomicResearch.

Gortmaker,S.L.,B.A.Swinburn,D.Levy,R.Carter,P.L.Mabry,D.T.Finegood,M.L.andMoodie(2011)‘Changingthefutureofobesity:science,policy,andaction’,The Lancet 378(9793):838-847.

Griffith,R.,andM.O’Connell(2010)‘Publicpolicytowardsfoodconsumption’,Fiscal Studies 31(4):481-507.Griffith,R.,L.NesheimandM.O’Connell(2009)‘Empiricalestimatesoftheimpactofafattax’.London:Institutefor

FiscalStudies,UniversityCollegeLondon.Griffith,R.,R.LluberasandM.Lührmann(2013)‘GluttonyinEngland?Long-termchangeindiet’.IFSBriefingNote

BN142.London:InstituteofFiscalStudies,UniversityCollegeLondon.Grossman,M.,E.TakinandR.Wada(2013)‘FoodPricesandBodyFatnessamongYouths,2013’.IZADiscussionPaper

No.7465(http://ftp.iza.org/dp7465.pdf).Gruber,J.,andM.Frakes(2006)‘Doesfallingsmokingleadtorisingobesity?’Journal of Health Economics 25:183-197.Han,E.andL.M.Powell(2011)‘Effectoffoodpricesontheprevalenceofobesityamongyoungadults’,Public Health

125(3):129-35.doi:10.1016/j.puhe.2010.11.014Han,E.,T.H.KimandL.M.Powell(2013)‘Beverageconsumptionandindividual-levelassociationsinSouthKorea’,

BMC Public Health 13:195(http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3626885/pdf/1471-2458-13-195.pdf).Hawkes,C.(2009a)‘Salespromotionsandfoodconsumption’,Nutrition Reviews 67(6):333-42.doi:

10.1111/j.1753-4887.2009.00206.x.Hawkes,C.(2009b)‘Financialincentivesanddiscincentivestoencouragehealthyeating’.PaperpreparedforWHICH?.Hawkes,C.(2012)‘Foodtaxes:whattypeofevidenceisavailabletoinformpolicydevelopment?’Nutrition Bulletin

37(1):51-56.doi:10.1111/j.1467-3010.2011.01949.x.He,W.(2013)‘FamilyStructureandChildMalnutritioninChina:ThreeEssays’.Dissertation,DukeUniversity,Durham,

NC(http://oatd.org/oatd/record?record=handle\:10161\%2F7139).HouseofCommonsHealthCommittee(2004)‘Obesity:ThirdReportofSession2003-04’.Volume1.HC23-I.London:

HouseofCommons.Imamura,F.,R.Micha,S.Khatibzadeh,S.Fahimi,P.Shi,J.PowlesandD.Mozaffarian,withtheGlobalBurdenof

DiseasesNutritionandChronicDiseasesExpertGroup(NutriCoDE)(2015)‘Dietaryqualityamongmenandwomenin187countriesin1990and2010:asystematicassessment’,Lancet Global Health (3):e132-42.

Jacobs,K.andD.A.Sumner(2002)‘TheFoodBalanceSheetsoftheFoodandAgricultureOrganization:AReviewofPotentialWaystoBroadentheAppropriateUsesoftheData’.AReviewSponsoredbyFAO,UCDavis.

James,W.P.T.(2008)Thefundamentaldriversoftheobesityepidemic’,Obesity Reviews 9(suppl.1):6-13.Jensen,J.D.andS.Smed(2007)‘Cost-effectivedesignofeconomicinstrumentsinnutritionpolicy’,International Journal

of Behavioral Nutrition and Physical Activity 4:10.JocelinePomerleau,K.Lock,M.McKeeandD.R.Altmann(2004)‘Thechallengeofmeasuringglobalfruitandvegetable

intake’,The Journal of Nutrition 134(5):1175-80.

Johansson,G.,A.Wikman,A.M.Ahrén,G.HallmansandI.Johansson(2001)‘Underreportingofenergyintakeinrepeated24-hourrecallsrelatedtogender,age,weightstatus,dayofinterview,educationallevel,reportedfoodintake,smokinghabitsandareaofliving’,Public Health Nutrition 4(4):919-27.

JonesN.R.V.,A.I.Conklin,M.SuhrckeandP.Monsivais(2014)‘Thegrowingpricegapbetweenmoreandlesshealthyfoods:analysisofanovellongitudinalUKdataset’.PLoS ONE 9(10):e109343.doi:10.1371/journal.pone.0109343.

Kearney,J.(2010)‘FoodConsumptionTrendsandDrivers’.Philosophical Transactions of The Royal Society 365No.15542793-2807(http://rstb.royalsocietypublishing.org/content/365/1554/2793.full).

Keats,S.,andS.Wiggins,2014.Futurediets:Implicationsforagricultureandfoodprices.ODIReport.http://www.odi.org/sites/odi.org.uk/files/odi-assets/publications-opinion-files/8776.pdf

Khan,T.,L.M.PowellandR.Wada(2012)‘Fastfoodconsumptionandfoodprices:evidencefrompaneldataon5thand8thgradechildren’,Journal of Obesity.doi:10.1155/2012/857697.

Kuchler,F.andH.Stewart(2008)‘PriceTrendsAreSimilarforFruits,Vegetables,andSnackFoods’.USDAERSEconomicResearchReportNumber55.Washington,DC:UnitedStatesDepartmentofAgriculture.

Kumcu,A.(2011)‘IntheLongRun:FoodSpendingDippedDuringRecession’.USDAAmberWaves.Washington,DC:UnitedStatesDepartmentofAgriculture(http://www.ers.usda.gov/amber-waves/2011-march/in-the-long-run.aspx#.VHiN4cmYZws).

Lang,T.andG.Rayner(2007)‘Overcomingpolicycacophonyonobesity:anecologicalpublichealthframeworkforpolicymakers’,Obesity Reviews 8(s1):165-181.

Lee,H.S.,K.J.DuffeyandB.M.Popkin(2012)‘SouthKorea’sentrytotheglobalfoodeconomy:shiftsinconsumptionoffoodbetween1998and2009’,Asia Pacific Iournal of Clinical Nutrition 21(4):618-629.

LeeS.K.andJ.Sobal(2003)‘Socio-economic,dietary,activity,nutritionandbodyweighttransitionsinSouthKorea’,Public Health Nutrition 6(7):665-674.

Lee,H-S.,S-O.KwonandY.Lee(2013)‘Weightstatusanddietaryfactorsassociatedwithsugar-sweetenedbeverageintakeamongKoreanchildrenandadolescents-KoreaNationalHealthandNutritionExaminationSurvey,2008-2011’,Clinical Nutrition Research 2(2):135-142.

Lee,M.J.,B.M.PopkinandS.Kim(2002)‘TheuniqueaspectsofthenutritiontransitioninSouthKorea:theretentionofhealthfulelementsintheirtraditionaldiet’Public Health and Nutrition 5(1A):197-203.

Lenz,A.,M.T.A.Olinto,J.S.Dias-da-Costa,A.L.Alves,M.Balbinotti,M.P.PattussiandD.G.Bassani(2009)‘Socioeconomic,demographic,andlifestylefactorsassociatedwithdietarypatternsofwomenlivinginSouthernBrazil’Cadernos de Saúde Pública 25(6)RiodeJaneiro.

Li,M.,M.J.Dibley,D.W.SibbrittandH.Yan(2010)‘Dietaryhabitsandoverweight/obesityinadolescentsinXi’anCity,China’,Asia Pacific Journal of Clinical Nutrition 19(1):76-82.

LocalGovernmentAssociation(LGA)(2014)‘Tacklingthecausesandeffectsofobesity:Investinginournation’sfuture:Thefirst100daysofthenextgovernment.Report’(http://www.local.gov.uk/documents/10180/6341755/100+Days+Obesity+publication/b650d6cb-289b-4f8c-a823-3c10380d75ff).

Lock,K.,R.D.Smith,A.D.Dangour,M.Keogh-Brown,G.Pigatto,C.Hawkes,andZ.Chalabi(2010)‘Health,agricultural,andeconomiceffectsofadoptionofhealthydietrecommendations’,The Lancet 376(9753):1699-1709.

Lu,Y.andD.Goldman(2010)‘TheEffectsofRelativeFoodPricesonObesity–EvidencefromChina:1991-2006’.AmericanAssociationofWineEconomicsAAWEWorkingPaperNo.15720.

Macdiarmid,J.andJ.Blundell(1998)‘Assessingdietaryintake:who,whatandwhyofunder-reporting’,Nutrition Research Reviews 11:231-253.doi:10.1079/NRR19980017.

Maillot,M.,N.DarmonandA.Drewnowski(2009)‘Arethelowest-costhealthfulfoodplansculturallyandsociallyacceptable?’Public Health Nutrition (http://depts.washington.edu/uwcphn/pubs/author/docs/2010phn.pdf).

Martin,E.andN.Cattan(2013)‘Mexicotacklesobesityepidemicwithtaxonjunkfood’,BloombergNews(http://www.bloomberg.com/news/articles/2013-10-29/mexico-tackles-obesity-epidemic-with-tax-on-junk-food).

Mendez,M.A.,B.M.Popkin,G.Buckland,H.Schroder,P.Amiano,A.Barricarte,J.M.Huerta,J.R.Quirós,M.J.SánchezandC.González(2011)‘Alternativemethodsofaccountingforunderreportingandoverreportingwhenmeasuringdietaryintake-obesityrelations’,American Journal of Epidemiology 15;173(4):448-58.

Monsivais,P.andA.Drewnowski(2007)‘Therisingcostoflow-energy-densityfoods’,Journal of the American Dietetic Association 107(12):2071-2076.doi:10.1016/j.jada.2007.09.009.

Monteiro,C.A.,G.Cannon,R.B.Levy,R.M.ClaroandJ.-C.Moubarac(2012)‘Thefoodsystem.Processing.Thebigissuefordisease,goodhealth,well-being’,World Nutrition 3(12):527-569.

Monteiro,C.A.(2013)‘Thenewroleofindustrialfoodprocessinginfoodsystemsanditsimpactonnutritionandhealth–aperspectivefromtheSouth’.PresentationatUN-SCNMeetingoftheMindsonNutritionImpactofFoodSystems.Geneva,25-28March.

58 ODI Report

The rising cost of a healthy diet: Changing relative prices of foods in high-income and emerging economies 59

Monteiro,C.A.andG.Cannon(2012)‘TheimpactofTransnational“BigFood”CompaniesontheSouth:aviewfromBrazil’,PLoS Med 9(7):e1001252.doi:10.1371/journal.pmed.1001252.

Monteiro,C.A.,R.BertazziLevy,R.MoreiraClaro,I.R.RibeirodeCastroandG.Cannon(2010)‘Increasingconsumptionofultra-processedfoodsandlikelyimpactonhumanhealth:evidencefromBrazil’,Public Health Nutrition 14(1):5-13.doi:10.1017/S1368980010003241.

MoreiraClaro,R.,H.C.E.doCarmo,F.M.SartiMachadoandC.A.Monteiro(2007)‘Income,foodprices,andparticipationoffruitandvegetablesinthediet’,Rev. Saúde Pública 41(4)(http://www.scielo.br/scielo.php?script=sci_arttext&pid=S0034-89102007000400009&lng=en&nrm=iso&tlng=en).

Morris,M.A.,C.Hulme,G.P.Clarke,K.L.EdwardsandJ.E.Cade(2014)‘Whatisthecostofahealthydiet?UsingdietdatafromtheUKWomen’sCohortStudy’,Journal of Epidemiology and Community Health 68:1043-1049.

Morrissey,T.W.,A.JacknowitzandK.Vinopal(2014)‘Localfoodpricesandtheirassociationswithchildren’sweightandfoodsecurity’,Pediatrics 133(3):422-30.doi:10.1542/peds.2013-1963.

Moubarac,J.-C.(2014)‘Consumptionofultra-processedfoodanddrinkproductsinLatinAmerica:Trends,impactonobesity,andpolicyimplications’.FactSheet:PanAmericanHealthOrganization.Washington,DC:PAHO.

Muhammad,A.,J.L.Seale,Jr.,B.MeadeandA.Regmi(2011)‘InternationalEvidenceonFoodConsumptionPatterns:AnUpdateUsing2005InternationalComparisonProgramData’.USDAERSTechnicalBulletinNumber1929.Washington,DC:UnitedStatesDepartmentofAgriculture.

Musaiger,A.O.(2011)‘OverweightandobesityineasternMediterraneanregion:prevalenceandpossiblecauses’,Journal of Obesity.doi:10.1155/2011/407237.

Mytton,O,A.GrayM.RaynerandH.Rutter(2007)‘Couldtargetedfoodtaxesimprovehealth?’Journal of Epidemiology & Community Health 61(8):689-694.

Ng,M.,T.Fleming,M.Robinson,B.Thompsonetal.(2014)‘Global,regional,andnationalprevalenceofoverweightandobesityinchildrenandadultsduring1980–2013:asystematicanalysisfortheGlobalBurdenofDiseaseStudy2013’,The Lancet 384(9945):766-781.doi:10.1016/S0140-6736(14)60460-8.

Ng,S.W.,C.N.Mhurchu,S.A.Jebb,andB.M.Popkin(2011)‘PatternsandtrendsofbeverageconsumptionamongchildrenandadultsinGreatBritain,1986–2009’,British Journal of Nutrition 108(3):536-551.

Nnoaham,K.E.,G.Sacks,M.Rayner,O.MyttonandA.Gray(2009)‘Modellingincomegroupdifferencesinthehealthandeconomicimpactsoftargetedfoodtaxesandsubsidies’,International Journal of Epidemiology 38:1324-1333.

Offer,A.,R.PecheyandS.Ulijaszek(2010)‘Obesityunderaffluencevariesbywelfareregimes:theeffectoffastfood,insecurity,andinequality’.UniversityofOxfordDiscussionPapersinEconomicandSocialHistoryNumber82.Oxford:

Ogden,C.L.,M.D.Carroll,B.K.KitandK.M.Flegal(2014)‘PrevalenceofchildhoodandadultobesityintheUnitedStates,2011-2012’,Journal of the American Medical Association 311(8):806-814.doi:10.1001/jama.2014.732.

Osman,M.andGalal,O.M.(2002)‘ThenutritiontransitioninEgypt:obesity,undernutritionandthefoodconsumptioncontext’,Public Health Nutrition:5(1A):141-148.

Parsons,A.C.,M.Shraim,J.Inglis,P.AveyardandP.Hajek(2009)‘Interventionsforpreventingweightgainaftersmokingcessation’,Cochrane Database of Systematic Reviews Issue1.Art.No.:CD006219.doi:10.1002/14651858.CD006219.pub2.

Pieroni,L.D.LanariandL.Salmasi(2010)‘FoodPricesandOverweightPatternsinItaly’.MPRAPaper(http://mpra.ub.uni-muenchen.de/23744/1/Food_prices_and_Overweight_Patterns_in_Italy_Pieroni_et_al_pdf).

Popkin,B.M.(2006)‘Technology,transport,globalizationandthenutritiontransitionfoodpolicy’,Food Policy 31(6):554-569.

Popkin,B.M.,L.S.AdairandS.W.Ng(2012)‘NOWANDTHEN:theglobalnutritiontransition:thepandemicofobesityindevelopingcountries’,Nutrition Reviews 70(1):3-21.

Popkin,B.M.andM.Slining(2013)‘Newdynamicsinglobalobesityfacinglow-andmiddle-incomecountries’,Obesity Reviews 14.S2:11-20.

PowellL.M.,J.F.Chriqui,T.Khan,R.WadaandF.J.Chaloupka(2013)‘Assessingthepotentialeffectivenessoffoodandbeveragetaxesandsubsidiesforimprovingpublichealth:asystematicreviewofprices,demandandbodyweightoutcomes’,Obesity Reviews 14(2):110-28.doi:10.1111/obr.12002.

Powell,L.M,andE.Han(2011)‘ThecostsoffoodathomeandawayfromhomeandconsumptionpatternsamongU.S.adolescents’,Journal of Adolescent Health 48(1):20-26.

Powell,L.M,andF.J.Chaloupka(2007)‘Foodpricesandobesity:evidenceandpolicyimplicationsfortaxesandsubsidies’,The Milbank Quarterly 87(1):229-257.doi:10.1111/j.1468-0009.2009.00554.x.

Powell,L.M,andY.Bao(2009)‘Foodprices,accesstofoodoutletsandchildweight’,Economics and Human Biology 7(1):64-72.doi:10.1016/j.ehb.2009.01.004.

Powell,L.M,M.C.Auld,F.J.Chaloupka,P.M.O’MalleyandL.D.Johnston(2007)‘Associationsbetweenaccesstofoodstoresandadolescentbodymassindex’,American Journal of Preventive Medicine 33(4):SupplementS301-S307.doi:10.1016/j.amepre.2007.07.007.

Powell,L.M.(2009)‘Fastfoodcostsandadolescentbodymassindex:evidencefrompaneldata’,Journal of Health Economics 28(5):963-70.doi:10.1016/j.jhealeco.2009.06.009.

Powell,L.M.andY.Bao,2009.Foodprices,accesstofoodoutletsandchildweight.Economics and Human Biology 7(1):64-72,doi:10.1016/j.ehb.2009.01.004

Powell,L.M.,E.HanandF.J.Chaloupka(2010)‘Economiccontextualfactors,foodconsumption,andobesityamongU.S.adolescents’,The Journal of Nutrition 140(6):1175-1180.

Powell,L.M.,J.ChriquiandF.J.Chaloupka(2009)‘Associationsbetweenstate-levelsodataxesandadolescentbodymassindex’,The Journal of Adolescent Health 45(3Suppl):S57-63.doi:0.1016/j.jadohealth.2009.03.003.

Powell,L.M.,M.C.Auld,F.J.Chaloupka,P.M.O’MalleyandL.D.Johnston(2007)‘Accesstofastfoodandfoodprices:relationshipwithfruitandvegetableconsumptionandoverweightamongadolescents’,Advances in Health Economics and Health Services Research 17:23-48.

Powell,L.M.,Z.ZhaoandY.Wang(2009)‘FoodpricesandfruitandvegetableconsumptionamongyoungAmericanadults’,Health & Place 15(4):1064-1070.

Rao,M.,A.Afshin,G.SinghandD.Mozaffarian(2013)‘Dohealthierfoodsanddietpatternscostmorethanlesshealthyoptions?Asystematicreviewandmeta-analysis’,BMJ Open 3:e004277.doi:10.1136/bmjopen-2013-004277.

Rayner,M.,P.ScarboroughandT.Lobstein(2009)‘TheUKOfcomNutrientProfilingModel:Defining“healthy”and“unhealthy”foodsanddrinksforTVadvertisingtochildren’(http://www.dph.ox.ac.uk/bhfhprg/publicationsandreports/acad-publications/bhfcpnppublished/nutrientprofilemodel).

Revoredo-Giha,C.andF.Akaichi(2014)‘Retailers’promotionsandthedemandforsugaryproductsinScotland’.PosterpaperpreparedforpresentationattheEAAE2014Congress‘Agri-FoodandRuralInnovationsforHealthierSocieties’,Ljubjana,26–29August.

Ricardo,C.Z.andR.M.Claro(2012)‘CustodaalimentaçãoedensidadeenergéticadadietanoBrasil,2008-2009’[CostandenergydensityofdietinBrazil,2008-2009].Cad. Saúde Pública,RiodeJaneiro28(12):2349-2361.

Roberto,C.A.,B.Swinburn,C.Hawkes,T.T-K.Huang,S.A.Costa,M.Ashe,L.Zwicker,J.H.CawleyandK.D.Brownell(2015)‘Patchyprogressonobesityprevention:emergingexamples,entrenchedbarriers,andnewthinking’,The Lancet.PublishedOnlineFebruary18,2015,http://dx.doi.org/10.1016/S0140-6736(14)61744-X

Rodríguez-Ramírez,S.,M.A.Méndez,S.Hernández-Cordero,T.GonzálezdeCossíoandB.M.Popkin(2013)‘Implausibleenergyintakereportinginoverweightwomenthroughoutaweightlosstrial’,The Journal of the Federation of American Societies for Experimental Biology 27(622):19.

Rosin,O.(2008)‘Theeconomiccausesofobesity:asurvey’,Journal of Economic Surveys 22(4):617-647.RT.com(2014)‘PepsiandCoketakehammeringinMexicoasjunkfoodtaxstartstobite’,9October(http://rt.com/

business/194644-sugary-drinks-tax-mexico/).Rydén,P.J.andL.Hagfors(2011)‘Dietcost,dietqualityandsocio-economicposition:howaretheyrelatedandwhat

contributestodifferencesindietcosts?’Public Health Nutrition 14(9):1680-92.doi:10.1017/S1368980010003642.Santaliestra-Pasías,A.M.,T.Mouratidou,V.Verbestel,K.Bammann,D.Molnar,S.Sieri,A.Siani,T.Veidebaum,S.Mårild,

L.Lissner,C.Hadjigeorgiou,L.Reisch,I.deBourdeaudhuijandL.A.Moreno(2013)‘PhysicalactivityandsedentarybehaviourinEuropeanchildren:theIDEFICSstudy’,Public Health Nutrition 8:1-12.

Scarborough,P.,P.Bhatnagar,K.K.Wickramasinghe,S.Allender,C.FosterandM.Rayner(2011)‘Theeconomicburdenofillhealthduetodiet,physicalinactivity,smoking,alcoholandobesityintheUK:anupdateto2006-07NHScosts’,Journal of Public Health 33(4):527-535.

Scarborough,P.,P.Bhatnagar,K.K.Wickramasinghe,S.Allender,C.Foster,andM.Rayner,2011.TheeconomicSeiden,A.,N.Hawley,D.Schulz,S.RaifmanandS.T.McGarvey(2012)‘Long-termtrendsinfoodavailability,food

prices,andobesityinSamoa’,American Journal of Human Biology 24(3):286-295.doi:10.1002/ajhb.22237.Serra-Majem,L.,D.MacLean,L.Ribas,D.Brulé,W.Sekula,R.Prattala,R.Garcia-Closas,A.Yngve,M.Lalondeand

A.Petrasovits(2003)‘Comparativeanalysisofnutritiondatafromnational,household,andindividuallevels:resultsfromaWHO-CINDIcollaborativeprojectinCanada,Finland,Poland,andSpain’,Journal of Epidemiology and Community Health 57:74-80

Shemilt,I.,G.J.Hollands,T.M.Marteau,R.Nakamura,S.A.Jebbetal.(2013)‘Economicinstrumentsforpopulationdietandphysicalactivitybehaviourchange:asystematicscopingreview’,PLoS ONE 8(9):e75070.doi:10.1371/journal.pone.0075070.

Smil,V.(1987)Energy, Food, Environment: Realities, Myths, Options.Oxford:ClarendonPress.Steenhuis,I.H.M,W.E.WaterlanderandA.deMul(2011)‘Consumerfoodchoices:theroleofpriceandpricing

strategies’,Public Health Nutrition 14(12):2220-2226.

60 ODI Report

The rising cost of a healthy diet: Changing relative prices of foods in high-income and emerging economies 61

Stern,D.,C.Piernas,S.Barquera,J.A.RiveraandB.M.Popkin(2014)‘CaloricbeveragesweremajorsourcesofenergyamongchildrenandadultsinMexico,1999-2012’,Journal of Nutrition 144(6):949-56.doi:10.3945/jn.114.190652.

Stevens,G.A.,G.M.Singh,Y.Lu,G.Danaeietal.(2012)‘National,regional,andglobaltrendsinadultoverweightandobesityprevalences’,Population Health Metrics 10:22.doi:10.1186/1478-7954-10-22.

Sturm,R.(2008)‘Stemmingtheglobalobesityepidemic:whatcanwelearnaboutsocialandeconomictrends?’Public Health 122(8):739-746.

Sturm,R.andA.Datar(2005)‘Bodymassindexinelementaryschoolchildren,metropolitanareafoodpricesandfoodoutletdensity’,Public Health 119:1059-1068.

Sturm,R.andA.Datar(2011)‘Regionalpricedifferencesandfoodconsumptionfrequencyamongelementaryschoolchildren’,Public Health 125(3):136-41.doi:10.1016/j.puhe.2010.11.016.

Sturm,R.andR.An(2014)‘Obesityandeconomicenvironments’,Obesity and CA: A Cancer Journal for Clinicians 64(5):337-350.

Sturm,R.,L.M.Powell,J.F.ChriquiandF.J.Chaloupka(2010)‘Sodataxes,softdrinkconsumption,andchildren’sbodymassindex’,Health Affairs 29(5):1052-8.doi:10.1377/hlthaff.2009.0061.

Swinburn,B.A.,G.Sacks,K.D.Hall,K.McPherson,D.T.Finegood,M.L.MoodieandS.L.Gortmaker(2011)‘Theglobalobesitypandemic:shapedbyglobaldriversandlocalenvironments’,The Lancet 378(9793):804-814.

Temple,N.J.andN.P.Steyn(2011)‘Thecostofahealthydiet:aSouthAfricanperspective’, Nutrition 27:505-508.TheMarmotReview(2010)‘FairSociety,HealthyLives’:TheMarmotReviewExecutiveSummary.StrategicReviewof

HealthInequalitiesinEnglandpost-2010’.London:UniversityCollegeLondonTiffin,R.,K.Balcombe,M.SaloisandA.Kehlbacher(2011)‘EstimatingFoodandDrinkElasticities’.Reading:University

ofReading,DEFRA.Todd,J.E.,E.LeibtagandC.Penberthy(2011)‘GeographicDifferencesintheRelativePriceofHealthyFoods’.USDA

ERSEconomicInformationBulletinNumber78.Washington,DC:UnitedStatesDepartmentofAgriculture.Todd,J.E.,L.MancinoandB.Lin(2010)‘TheImpactofFoodAwayfromHomeonAdultDietQuality’.USDA

EconomicResearchReport90.Washington,DC:UnitedStatesDepartmentofAgriculture(http://ers.usda.gov/media/136609/err90_1_.pdf).

Tsai,A.G.,D.F.WilliamsonandH.A.Glick(2011)‘DirectmedicalcostofoverweightandobesityintheUnitedStates:aquantitativesystematicreview’,Obesity Review 12(1):50-61.

UKGOfS(GovernmentOfficeforScience)(2007)‘Foresight:TacklingObesities:FutureChoices–ObesitySystemAtlas’.London:UKDepartmentofInnovation,Universities,andSkills(https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/295153/07-1177-obesity-system-atlas.pdf).

Urban,L.E.,G.E.Dallal,L.M.Robinson,L.M.Ausman,E.Saltzman,andS.B.Roberts(2009)TheAccuracyofStatedEnergyContentsofReduced-energy,CommerciallyPreparedFoods.Journal of the American Dietetic Association 110(1):116-123.

USDA(2015)‘ScientificReportofthe2015DietaryGuidelinesAdvisoryCommittee.AdvisoryReporttotheSecretaryofHealthandHumanServicesandtheSecretaryofAgriculture’.Washington,DC:UnitedStatesDepartmentofAgriculture.

ValdésRamos,R.(2012)‘NutritionandHealthConditionsofMexicans’.Presentation(http://www.odi.org/sites/odi.org.uk/files/odi-assets/events-presentations/1562.pdf).

Volpe,R.,E.RoegerandE.Leibtag(2013)‘HowTransportationCostsAffectFreshFruitandVegetablePrices’.USDAEconomicResearchReportNumber160.Washington,DC:UnitedStatesDepartmentofAgriculture.

Wang,J.,M.Williams,E.Rush,N.Crook,N.G.ForouhiandD.Simmons(2010)‘MappingtheavailabilityandaccessibilityofhealthyfoodinruralandurbanNewZealand–TeWaioRona:DiabetesPreventionStrategy’,Public Health Nutrition 13(7):1049-55.

Waterlander,W.E.,W.E.deHaas,I.vanAmstel,A.J.Schuit,J.W.R.Twisk,M.Visser,J.C.SeidellandI.H.M.Steenhuis(2010)‘Energydensity,energycostsandincome–howaretheyrelated?’Public Health Nutrition (10):1599-608.doi:10.1017/S1368980009992989.

WCRF(2007)Food,nutrition,physicalactivityandthepreventionofcancer:aglobalperspective(chapter10).Washington,DC:WorldCancerResearchFund.

Wendt,M.andJ.E.Todd(2011)‘TheEffectofFoodandBeveragePricesonChildren’sWeights’.ERR-118.Washington,DC:UnitedStatesDepartmentofAgriculture,EconomicResearchService.

WHO/FAO(2003)‘Diet,nutritionandthepreventionofchronicdiseases’.TechnicalReportSeries916.GenevaandRome:WorldHealthOrganizationandFoodandAgricultureOrganization.

Wrieden,W.L.andK.L.Barton(2011)‘TheScottishDiet:EstimationsofEnergyDensityandExpenditure’.Aberdeen:FoodStandardsAgency(Scotland).

Xu,X.,J.N.Variyam,Z.ZhaoandF.J.Chaloupka(2014)‘RelativefoodpricesandobesityinU.S.metropolitanareas:1976-2001’,PLoS One,9(12):e114707.doi:10.1371/journal.pone.0114707.

Zhao,W.,Y.Zhai,J.Hu,J.Wang,Z.Yang,L.KongandC.Chen(2008)‘Economicburdenofobesity-relatedchronicdiseasesinMainlandChina’,Obesity Reviews 9Suppl1:62-7.doi:10.1111/j.1467-789X.2007.00440.x.

Zhou,S.andT.O.Awokuse(2012)‘Urbanization,NutritionTransition,andObesity:EvidencefromChina’.SelectedPaperpreparedforpresentationattheAgricultural&AppliedEconomicsAssociation’s2014AAEAAnnualMeeting,Minneapolis,MN,27-29July2014.

62 ODI Report

The rising cost of a healthy diet: Changing relative prices of foods in high-income and emerging economies 63

ODI is the UK’s leading independent think tank on international development and humanitarian issues.

Readers are encouraged to reproduce material from ODI Reports for their own publications, as long as they are not being sold commercially. As copyright holder, ODI requests due acknowledgement and a copy of the publication. For online use, we ask readers to link to the original resource on the ODI website. The views presented in this paper are those of the author(s) and do not necessarily represent the views of ODI.© Overseas Development Institute 2015. This work is licensed under a Creative Commons Attribution-NonCommercial Licence (CC BY-NC 3.0).ISSN: 2052-7209

All ODI Reports are available from www.odi.org

Overseas Development Institute203 Blackfriars Road London SE1 8NJTel +44 (0)20 7922 0300 Fax +44 (0)20 7922 0399

www.odi.org