Academy Health Bittersweet: How Prices of Sugar-Rich Foods - - PowerPoint PPT Presentation

Academy Health Bittersweet: How Prices of Sugar-Rich Foods Contribute to the Diet-Related Disease Epidemic in Mexico

Tadeja Graˇ cner RAND Corporation June 27, 2016

1 / 25

Introduction Motivation

Global Public Health Crisis

Obesity:

>1 billion overweight adults > 300 million of them clinically obese

Diabetes:

> 10 % of people diabetic A leading cause of death in Mexico

Hypertension:

> 30 % of people hypertensive A leading cause of premature death

2 / 25

Introduction Motivation

Global Public Health Crisis

Obesity:

>1 billion overweight adults > 300 million of them clinically obese      > 60% in LMIC

Diabetes:

> 10 % of people diabetic A leading cause of death in Mexico      > 80% in LMIC

Hypertension:

> 30 % of people hypertensive A leading cause of premature death      > 80% of CVD deaths in LMIC

LMIC: low- and middle-income countries CVD: cardio-vascular disease

2 / 25

Introduction Motivation

Global Public Health Crisis

Relative decline of prices of unhealthy processed foods Shift towards unhealthy diet

(Cutler et al., 2003; Drewnowski and Popkin, 1997; Lakdawalla and Philipson, 2009)

Increased total calorie intake Change in diet’s nutritional composition

Increased carbohydrates, sugar, fat intake Decreased fiber intake

3 / 25

Introduction Motivation

Global Public Health Crisis

Relative decline of prices of unhealthy processed foods Shift towards unhealthy diet

(Cutler et al., 2003; Drewnowski and Popkin, 1997; Lakdawalla and Philipson, 2009)

Increased total calorie intake Change in diet’s nutritional composition

Increased carbohydrates, sugar, fat intake Decreased fiber intake

⇒ Proposed price-based policies (e.g. sugar/fat tax)

3 / 25

Introduction Motivation

Global Public Health Crisis

Relative decline of prices of unhealthy processed foods Shift towards unhealthy diet

(Cutler et al., 2003; Drewnowski and Popkin, 1997; Lakdawalla and Philipson, 2009)

Increased total calorie intake Change in diet’s nutritional composition

Increased carbohydrates, sugar, fat intake Decreased fiber intake

⇒ Proposed price-based policies (e.g. sugar/fat tax) But...how sensitive is health to price changes of foods rich in these nutrients?

3 / 25

Introduction Research Question

Research Question

What is the effect of changes in prices of foods rich in sugar, fat or combination of other nutrients on the prevalence of Obesity (BMI, abdominal obesity) Type 2 Diabetes Hypertension

4 / 25

Introduction Research Question

Research Question

What is the effect of changes in prices of foods rich in sugar, fat or combination of other nutrients on the prevalence of Obesity (BMI, abdominal obesity) Type 2 Diabetes Hypertension

1 Price changes of foods rich in which nutrients matter? 2 How long does the effect persist? 3 Is there heterogeneity in health response to prices - why?

4 / 25

Introduction Research Question

Research Question

What is the effect of changes in prices of foods rich in sugar, fat or combination of other nutrients on the prevalence of Obesity (BMI, abdominal obesity) Type 2 Diabetes Hypertension

1 Price changes of foods rich in which nutrients matter? 2 How long does the effect persist? 3 Is there heterogeneity in health response to prices - why?

→ Context: Mexico 1996-2010

4 / 25

Introduction Related Literature

Related Literature

Prices Consumption Health

Harding et al. 2014 Dubois et al., 2013 Becker et al., 1991 O’Donoghue and Rabin, 2003 Fletcher et al., 2010 Finkelstein et al. 2010

5 / 25

Introduction Related Literature

Related Literature

Prices Consumption Health

Harding et al. 2014 Dubois et al., 2013 Becker et al., 1991 O’Donoghue and Rabin, 2003 Fletcher et al., 2010 Finkelstein et al. 2010

Sugar

Basu et al., 2013 Lustig, 2013 Taubes, 2007

Fat

Barry and Popkin, 1998

5 / 25

Introduction Related Literature

Related Literature

Powell et al., 2009; Grossman et al., 2014; Auld et al., 2009, Lu&Goldman, 2010; Datar&Sturm, 2005; Dasgupta, 1997; Popkin, 2000; Fernald et al., 2008

1 Grouping of foods into healthy vs unhealthy 2 Evidence on body mass index 3 Cross-sections, short individual panels, selected samples 4 Mostly US

This paper Prices Consumption Health

Harding et al. 2014 Dubois et al., 2013 Becker et al., 1991 O’Donoghue and Rabin, 2003 Fletcher et al., 2010 Finkelstein et al. 2010

Sugar

Basu et al., 2013 Lustig, 2013 Taubes, 2007

Fat

Barry and Popkin, 1998

5 / 25

Context Mexico: Variation in Prices

Variation in food prices - food imports

Large reduction and variation in relative prices of processed foods since 1994 Increased import quotas and imports of processed foods

(Clark et al., 2013)

6 / 25

Context Mexico: Variation in Prices

Variation in food prices - retail entry

Large reduction and variation in relative prices of processed foods since 1994 Increased import quotas and imports of processed foods

(Clark et al., 2013)

FDI removal → foreign retail expansion – lower prices (Hawkes, 2006; Atkin et al., 2014)

7 / 25

Context Mexico: Variation in Prices

Variation in food prices - retail entry

1993 2001 2007

Source: Iacovone, 2009. 8 / 25

Context Mexico: Variation in Prices

Variation in food prices - reduced tariffs

Large reduction and variation in relative prices of processed foods since 1994 Increased import quotas and imports of processed foods

(Clark et al., 2013)

FDI removal → foreign retail expansion – lower prices (Hawkes, 2006; Atkin et al., 2014) Reduced tariffs – pass-through a function of distance to US border (Nicita, 2004)

9 / 25

Context Mexico: Data

Data

1 Time series of prices with nutritional information on >25,000 products

→ Prices of nutritionally similar foods (1996-2010)

Example: prices of sugar-rich, protein-rich, sugar-fat rich foods

10 / 25

Context Mexico: Data

Data

1 Time series of prices with nutritional information on >25,000 products

→ Prices of nutritionally similar foods (1996-2010)

Example: prices of sugar-rich, protein-rich, sugar-fat rich foods

Nutrient Clusters

10 / 25

Context Mexico: Data

Data

1 Time series of prices with nutritional information on >25,000 products

→ Prices of nutritionally similar foods (1996-2010)

Example: prices of sugar-rich, protein-rich, sugar-fat rich foods

2 State-level administrative panel dataset on chronic disease incidence

→ Type 2 diabetes, hypertension (1996-2010, 32 states)

3 Nationally representative individual-level panel dataset

→ Obesity (BMI, waist circumference), diabetes, hypertension (2002-2009, >30,000 individuals)

11 / 25

Context Mexico: Data

Data

1 Time series of prices with nutritional information on >25,000 products

→ Prices of nutritionally similar foods (1996-2010)

Example: prices of sugar-rich, protein-rich, sugar-fat rich foods

2 State-level administrative panel dataset on chronic disease incidence

→ Type 2 diabetes, hypertension (1996-2010, 32 states)

3 Nationally representative individual-level panel dataset

→ Obesity (BMI, waist circumference), diabetes, hypertension (2002-2009, >30,000 individuals)

11 / 25

Approach Price Variation

Approach

⇒ Regional price variation over time, conditional on location and year fixed effects

4.4 4.6 4.8 5 5.2 Ln(Price of Sugar) 1995 2000 2005 2010 Year

Prices of sugary foods by states over time

12 / 25

Findings Sugary Foods Prices and Health

Findings

A decrease in the price of sugar-rich foods significantly increases: Obesity, type 2 diabetes and hypertension prevalence:

300 350 400 450 Diabetes II 4.2 4.4 4.6 4.8 5 Log(Pricesugar,t-1)

Diabetes II and Price of Sugary Foods

500 550 600 650 700 Hypertension 4.2 4.4 4.6 4.8 5 Log(Pricesugar,t-1)

Hypertension and Price of Sugary Foods

13 / 25

Findings Food Prices and Health

Findings: Main Effects

DV: Diabetes II Hypertension Log(waist) Abd.Obese Log(Psugar)t

• 0.055
• 0.029
• 0.038
• 0.148

(0.015)*** (0.036) (0.021)* (0.053)*** Log(Pfats)t 0.008 0.015

• 0.033
• 0.040

(0.016) (0.038) (0.021) (0.051) Log(Pfiber)t

• 0.027
• 0.056
• 0.031
• 0.087

(0.017) (0.055) (0.035) (0.095) N 11485 11758 21753 21753

• Adj. R2

0.05 0.04 0.29 0.12 Y Mean 0.11 0.18 90.13 0.66 N Clusters 29 29 32 32

Individual and year FE, and controls included. Robust standard errors in parentheses, clustered at the city level. */**/*** denotes significant at the 10% / 5% / 1% levels.

14 / 25

Findings Fatty Foods Prices and Health

Findings: Main Effects

Prices of foods rich in other nutrients, such as fat, do not seem to matter.

250 300 350 400 450 Diabetes II 4.2 4.4 4.6 4.8 Log(Pricefat,t-1)

Diabetes II and Price of fatty foods

15 / 25

Findings Heterogenous effect

Findings: Heterogenous effects

1 Small response in consumption, but already pre-diabetic or

pre-hypertensive

2 Substantial response in consumption 3 Both

16 / 25

Findings Heterogenous effect

High vs Low risk

DV Diabetes II Hypertension Log(Psugar)t 0.028 0.041 (0.039) (0.057) Med Risk Diab x Log(Psugar)t

• 0.033

(0.065) High Risk Diab x Log(Psugar)t

• 0.194

(0.084)** High Risk Hyper x Log(Psugar)t

• 0.209

(0.089)** N 10524 10727

• Adj. R2

0.06 0.03 Y Mean 0.12 0.20 N Clusters 29 29

Individual, year FE, and controls included. Robust standard errors in parentheses, clustered at the city level. */**/*** denotes significant at the 10% / 5% / 1% levels.

17 / 25

Findings Heterogenous effect

Findings

A decrease in the price of sugar-rich foods significantly increases: Obesity, type 2 diabetes and hypertension prevalence: ≈ 15-20 percent of the type 2 diabetes increase since 1994 Strongest effect on people with:

Highest risk for disease Highest risk for disease and impatience

18 / 25

Conclusion Overview and Discussion

Conclusion

Prices of sugary foods affect short and long-term health: People at highest risk for disease and most impatient more responsive to prices Effects strongest within 1-year price change, up to 3-4 years Prices of sugary foods explain ≈ 15(5) percent of the type 2 diabetes(abdominal obesity) prevalence increase since 1994

19 / 25

Conclusion Overview and Discussion

Conclusion

Prices of sugary foods affect short and long-term health: People at highest risk for disease and most impatient more responsive to prices Effects strongest within 1-year price change, up to 3-4 years Prices of sugary foods explain ≈ 15(5) percent of the type 2 diabetes(abdominal obesity) prevalence increase since 1994 Results applicable to: poorer households in more developed economies, incidentally at highest risk for obesity and related diseases emerging markets: globalization and cheaper unhealthy foods firm behavior (e.g., pricing strategies, product manufacturing)

19 / 25

Conclusion Contact

THANK YOU! Comments and questions welcome at tgracner@rand.org

20 / 25

Conclusion Contact

Risk Assesement Form

TYPE 2 DIABETES RISK ASSESSMENT FORM

Circle the right alternative and add up your points.

• 1. Age

0 p. Under 45 years 2 p. 45–54 years 3 p. 55–64 years 4 p. Over 64 years

• 2. Body-mass index

(See reverse of form) 0 p. Lower than 25 kg/m2 1 p. 25–30 kg/m2 3 p. Higher than 30 kg/m2

• 3. Waist circumference measured below the ribs

(usually at the level of the navel) MEN WOMEN 0 p. Less than 94 cm Less than 80 cm 3 p. 94–102 cm 80–88 cm 4 p. More than 102 cm More than 88 cm

• 4. Do you usually have daily at least 30 minutes
• f physical activity at work and/or during leisure

time (including normal daily activity)? 0 p. Yes 2 p. No

• 5. How often do you eat vegetables, fruit or

berries? 0 p. Every day 1 p. Not every day

• 6. Have you ever taken medication for high

blood pressure on regular basis? 0 p. No 2 p. Yes

• 7. Have you ever been found to have high blood

glucose (eg in a health examination, during an illness, during pregnancy)? 0 p. No 5 p. Yes

• 8. Have any of the members of your immediate

family or other relatives been diagnosed with diabetes (type 1 or type 2)? 0 p. No 3 p. Yes: grandparent, aunt, uncle or first cousin (but no own parent, brother, sister

• r child)

5 p. Yes: parent, brother, sister or own child Total Risk Score The risk of developing type 2 diabetes within 10 years is Lower than 7 Low: estimated 1 in 100 will develop disease 7–11 Slightly elevated: estimated 1 in 25 will develop disease 12–14 Moderate: estimated 1 in 6 will develop disease 15–20 High: estimated 1 in 3 will develop disease Higher Very high: than 20 estimated 1 in 2 will develop disease

Test designed by Professor Jaakko Tuomilehto, Department of Public Health, University of Helsinki, and Jaana Lindström, MFS, National Public Health Institute.

Please turn over

21 / 25

Conclusion Contact

Changes in tariffs - HFCS ban (2002-2005)

4.5 4.6 4.7 4.8 Log(Pricesugar,t) 1995 2000 2005 2010 Year South North

22 / 25

Conclusion Contact

Impatience: Abdominal Obesity

DV: Log(waistline) Abdominally Obese (1) (2) (3) (4) Log(Psugar)t

• 0.008
• 0.019
• 0.073
• 0.081

(0.019) (0.018) (0.057) (0.051) Log(Psugar)t x Impatient(5)

• 0.046
• 0.112

(0.016)*** (0.061)* Log(Psugar)t x Impatient(+4)

• 0.024
• 0.087

(0.016) (0.053) N 16627 16627 16627 16627

• Adj. R2

0.30 0.30 0.13 0.13 Y Mean 90.46 0.67 N Clusters 32 32 32 32

Individual, year FE, and controls included. Robust standard errors in parentheses, clustered at the city level. */**/*** denotes significant at the 10% / 5% / 1% levels.

23 / 25

Conclusion Contact

Impatience: Diabetes II

DV: Diabetes II All High Risk (1) (2) (3) (4) Log(Psugar)t

• 0.047
• 0.046
• 0.045
• 0.002

(0.021)** (0.021)** (0.063) (0.057) Log(Psugar)t x Impatient(5) 0.012

• 0.118

(0.025) (0.069) Log(Psugar)t x Impatient(4+) 0.009

• 0.177

(0.029) (0.073)** N 9095 9095 1965 1965

• Adj. R2

0.06 0.06 0.06 0.07 Y Mean 0.11 N Clusters 29 29 29 29

Individual, year FE and controls included. Robust standard errors in parentheses, clustered at the city level. */**/*** denotes significant at the 10% / 5% / 1% levels.

24 / 25

Conclusion Contact

Impatience: Hypertension

DV: Hypertension All High Risk (1) (2) (3) (4) Log(Psugar)t 0.042 0.042 0.038 0.030 (0.039) (0.040) (0.064) (0.058) Log(Psugar)t x Impatient(5)

• 0.069
• 0.188

(0.039)* (0.085)** Log(Psugar)t x Impatient(4+)

• 0.062
• 0.160

(0.037) (0.076)** N 9276 9276 3026 3026

• Adj. R2

0.04 0.04 0.06 0.06 Y Mean 0.19 N Clusters 29 29 29 29

Individual, year FE and controls included. Robust standard errors in parentheses, clustered at the city level. */**/*** denotes significant at the 10% / 5% / 1% levels.

25 / 25