Targeting populations at health-risk: relevance of a food price policy

(1)

HAL Id: hal-02753848

https://hal.inrae.fr/hal-02753848

Submitted on 3 Jun 2020

HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

Targeting populations at health-risk: relevance of a food price policy

Patrice Bertail, France Caillavet

To cite this version:

Patrice Bertail, France Caillavet. Targeting populations at health-risk: relevance of a food price policy.

19. International Congress of Nutrition: Nutrition Security for all, Oct 2009, Bangkok, Thailand. �hal- 02753848�

(2)

Targeting populations at health-risk:

relevance of a food price policy

2- D

^ata ^anD

V

^ariables

TNS Worldpanel 2004: Household purchases for at-home consumption

Dependent variables of the demand systems: Budget shares for 26 food groups including:

Healthy patterns

• : Fruits & Vegetables (Fresh and Processed) Unhealthy patterns

• : Sugar-rich foods (Sugar, Confectionary & chocolate, Non alcoholic and Alcoholic beverages); Fat-rich foods (Meat products, Cheese and animal Fats, Convenience dishes)

Explanatory Variables: Prices (quality-adjusted unit values Cox & Wohlgenant), Socio-demo: age, BMI, education, occupation, household size, region, residence area…

1-i

^Dentifying

7

homogeneous segments Among them, evidence of 3 target segments:

S Segment “Overweight” dominated with overweighed or obese panelists: 23.6% of the sample

2/3 of panelists are overweighed or obese, mostly women

•

Income/unit of consumption

• m mean, lower education level, older

Western region, rural

•

S Segment “Poor 1”: 22.3% of the sample

Lower income/unit of consumption, fewer overweight or obese panelists

•

Younger households, larger household size, members < 13 years,

•

Rural area

•

S Segment “Poor 2”: 15.7% of the sample

Low income/unit of consumption, younger households

•

Urban area

•

2- t

^axing

u

^nhealthy

f

^ooD

P

^atterns

: b

^est ^canDiDate

?

Higher price elasticities for fat-rich than sugar-rich products : Considering price elasticities, i.e. the response of quantities purchased to a variation of price, we find the highest sensitivity to price for meat products, cheese and animal fats, confectionary and chocolate, alcoholic drinks.

Best candidate for food taxation: meat products

Consensus on meat products price elasticities: high levels for the 3 target

•

segments of poor and overweight

But these foods are over-consumed by the overweight segment only

• m Relevance on

the poor segments?

Cheese and animal fats: Moderate sensitivity to price and for the overweight

•

segment only.

3- s

^ubsiDizing

h

^ealthy

f

^ooD

P

^atterns

: b

^est ^canDiDate

?

Budget intervention on Fruit is more relevant than a price intervention: expenditure elasticity, i.e. the response of quantities purchased to a variation in food budget, is higher than price elasticity; moreover, a budget intervention can be targeted only to poor segments.

Best candidate for food subsidies: vegetables

We find a high price elasticity and a healthy substitution pattern (substitution with meat products) on the 3 target segments : among poor segments and, in a lesser extent, on the overweight segment.

Bertail, P., Caillavet F. 2008. Fruit and vegetable consumption patterns: a segmentation approach. American Journal of Agricultural Economics 90:827-842.

Mytton O., Gray A., Rayner M., Rutter H. 2007. Could targeted food taxes improve health? J. Epidemiology Community Health 61:689-694.

Smed S., Jensen J.D., Denver S. 2007. Socioeconomic characteristics and the effect of taxation as a health policy instrument. Food Policy 32: 624-639.

French National Research Agency (ANR), Programme National de Recherche en Alimentation et nutrition humaine, project ANR-05-PNRA, 012.

4- i

^mPact ^of

2

^Price ^Policies

: s

^ubsiDies ^to

V

^egetables ^with ^or ^without

t

^axation ^of

m

^eat ^ProDucts

Price policy and improvement of the diet: Combining Subsidies to Vegetables and Taxation of Meat products (policy 2) reaches a higher increase of fruit and vegetable consumption and reduction of fat-rich products than one only of these policies.

Taking into account inequality:

Subsidies to Vegetables alone (policy 1) are to be advocated.

Food taxation is regressive since poorer households dedicate a larger share of their budget to food.

S To identify the target of a price policy in France, i.e. population groups at health-risk and sensitive to prices of healthy or unhealthy foods.

S To assess the impact of modifying the price of foods according to their nutritional properties on these population groups.

S Populations sensitive to food prices include some poor and overweight groups.

S The more efficient price policy to improve the diet of these groups combines subsidies on vegetables with taxation on meat products.

S Efficiency vs Equity: Taking into account inequalities, subsidies on vegetables alone are recommended.

1- m

ôDelling ^fooD ^DemanD ânD îts

heterogeneity

Hypothesis: existence of different food demands in the population

A segmentation approach: using a clustering method without prior criteria (income, BMI…)

Estimation of a demand system for different homogeneous categories of population with a notion of homogeneity :

based on food consumption behaviour (quantities purchased,

•

budget shares)

defined by the parameters of the demand system (sensitivity

•

to price and budget variations).

OBJECTIVES

CONCLUSIONS

METHODS

RESULTS

MAJOR REFERENCES

FUNDING

A FINITE MIXTURE REGRESSION MODEL

Assumptions : Existence of a finite number of unknown classes in which food demand is described by a stable demand system (homogeneity of consumption behaviour) :

Choice of a flexible specification: non linear Almost Ideal Demand System (cf Deaton &

•

Muellbauer).

The parameters of the demand system are different on each class (mixture model cf McLachlan,

•

De Sarbo and Cron).

Goal : Estimate simultaneously the classes and the parameters on each class, taking into account the endogeneity of expenditure, then explain a posteriori the probabilities membership.

Procedure:

E-M algoritm : affecting each household to the class for which he has the largest posterior

•

probability

Choice of the number of classes: with information criteria (BIC, Akaike…)

•

Probabilities membership modelled as a function of sociodemographic characteristics of the

•

household : age, BMI, education, household income, residence area… (Logit-type specification)

Patrice BERTAIL

¹

, France CAILLAVET

²

(1) ModalX - Paris X University and CREST-LS, (2) INRA/ALISS UR 1303

Graphique 1: Uncompensated price elasticities of demand for fat-rich products on target segments

0 0,5 1 1,5 2 2,5 3 3,5 4

Meat products Cheese & an fat

Poor 1 Poor 2 Overweight

0 10 20 30 40 50 60 70 80 90

Poor 1 Poor 2 Overweight total

before

after policy 1 after policy 2

-0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4 0.5

F&V Sugar -rich products Fat-rich products

% Variation of consumption after

10% price decrease of vegetables and 10% increase of meat products

total

Overweight Poor 2 Poor 1

Graphique 2: Uncompensated price elasticities of demand for vegetables (fresh/processed) on target segments

0 0,5 1 1,5 2 2,5 3 3,5

Fresh veg Proc veg