Intro GxEcon Empirics Structural Conclusion

Genetic and Economic Interaction in the Formation of Health: The Case of Obesity Pietro Biroli University of Z¨ urich

Grenoble Applied Economics Laboratory Grenoble 29 June 2017

Pietro Biroli

1

Intro GxEcon Empirics Structural Conclusion

Research question Motivation

Outline 1

Introduction Research question Motivation

2

Gene x Econ

3

Empirical Findings ALSPAC Data Results Productivity Effect Preferences Robustness Checks

Replication using FHS 4

Structural Model The Model Simulations

5

Conclusion and Future Work

Pietro Biroli

2

Intro GxEcon Empirics Structural Conclusion

Research question Motivation

Genes, Biology, and Choices What: How genetic differences influence health investments and life-cycle evolution of health via changes in production function ↔ productive efficiency via changes in preferences ↔ allocative efficiency

Pietro Biroli

3

Intro GxEcon Empirics Structural Conclusion

Research question Motivation

Genes, Biology, and Choices What: How genetic differences influence health investments and life-cycle evolution of health via changes in production function ↔ productive efficiency via changes in preferences ↔ allocative efficiency

How: Integrate genes into an life-cycle model of health Micro-foundation of evidence from genetics Genes = measures of heterogeneity

Pietro Biroli

4

Intro GxEcon Empirics Structural Conclusion

Research question Motivation

Genes, Biology, and Choices What: How genetic differences influence health investments and life-cycle evolution of health via changes in production function ↔ productive efficiency via changes in preferences ↔ allocative efficiency

How: Integrate genes into an life-cycle model of health Micro-foundation of evidence from genetics Genes = measures of heterogeneity

Why: Inequality at birth and impact over the life Measure of ‘unobserved ability’ Differential response to prices, taxes, policies Pietro Biroli

5

Intro GxEcon Empirics Structural Conclusion

Research question Motivation

Obesity: major health problem 2nd preventable cause of death and disease (U.S.) [Mokdad et al., 2004] Median American and European are overweight The average French is almost overweight (BMI of 24.3, USA: 28.7) Obese Men

70.0

Overweight Men

60.0 50.0 40.0 30.0 20.0 10.0 France

Estonia

Belgium

Italy

Austria

Latvia

Bulgaria

Romania

Slovakia

Hungary

Germany

Spain

Cyprus

Poland

Greece

Czech Republic

United Kingdom

Malta

Slovenia

0.0

Source: European Health Interview Survey (2010), Eurostat

Cost Consequences Health: (US) $14.3 billion for children, $147 billion for adults, 400k deaths; (US, per year) Economic: lower skills acquisition, wages, labor force, and productivity. ,→ see [Cawley, 2010, Kline and Tobias, 2008, Ng et al., 2014] Pietro Biroli

6

Intro GxEcon Empirics Structural Conclusion

Research question Motivation

Health at a Glance 2011: OECD Indicators - © OECD 2011 2. NON-MEDICAL 2.4.1. DETERMINANTS Children aged 5-17OF years HEALTH who are - Overweight overweightand (including obesityobese), among children latest available estimates Version 1 - Last updated: 28-Oct-2011

Adolescent overweight

Children aged 5-17 years who are overweight (including obese), latest available estimates Girls Boys 37.0 35.9

4.5

50

40

30

20

10

45.0

Greece United States Italy Mexico New Zealand Chile United Kingdom Canada Hungary Iceland Slovenia Australia Spain Portugal OECD Brazil Russian Fed. Sweden Finland India Netherlands South Africa Germany Czech Republic Slovak Republic Denmark France Norway Japan Switzerland Poland Turkey Korea China

30.9 29.0 28.8 27.1 26.6 26.1 25.9 25.5 24.4 24.0 22.9 21.6 21.4 21.1 19.8 19.5 19.1 18.3 17.9 17.7 17.6 16.9 16.2 15.2 14.9 14.7 14.4 13.1 12.4 10.3 9.9

0

% of children aged 5-17 years

Pietro Biroli Source: International Association for the Study of Obesity (2011).

35.0 32.4 28.1 28.2 28.6 22.7 28.9 25.5 22.0 28.7 22.0 32.9 23.5 22.9 23.1 24.2 17.0 23.6 20.6 14.7 13.6 22.6 24.6 17.5 14.1 13.1 12.9 16.2 16.7 16.3 11.3 16.2 5.9

0

10

20

30

40

50

% of children aged 5-17 years

7

Intro GxEcon Empirics Structural Conclusion

Research question Motivation

The obesity ‘epidemic’

BMI increases: 1 with age 2 over time 3 across countries Pietro Biroli

8

Intro GxEcon Empirics Structural Conclusion

Research question Motivation

Inputs into BMI

Behaviors: calories in (diet) & calories out (exercise) ‘Obesogenic Environment’: cheaper food and less exercise [Cutler et al., 2003, Lakdawalla et al., 2005]

Genes: More than 30 genetic variants related to obesity (GWAS) [Speliotes et al., 2010, Berndt et al., 2013]

GxE: Gene-Environment Interaction [Rosenquist et al., 2015, Liu and Guo, 2015, Guo et al., 2015]

Pietro Biroli

9

Intro GxEcon Empirics Structural Conclusion

Research question Motivation

Genes: What and Why? Genetic Variants Human Genome Project (2003) Single Nucleotide Polymorphism (SNPs) ≈ 10 million 2 alleles, e.g.: AA, AT, or TT Why we care? Randomized at birth, fixed for life Causal pathways Cheaper to measure Geneticists at discovering stage

Pietro Biroli

10

Intro GxEcon Empirics Structural Conclusion

Research question Motivation

Related Literature 1

Nature and Nurture (Twin Studies): Use MZ/DZ twins to estimate genetic and environmental heritability. See [Galton, 1874, Taubman, 1976, Kohler et al., 2011]

2

Genoeconomics: Find genetic determinants of economic behaviors: risk aversion, time and social pref., addiction. See [Benjamin et al., 2007, Cesarini et al., 2009, Benjamin et al., 2016]

3

Mendelian Randomization: Genes as IV

See [Davey Smith, 2003, Fletcher and Lehrer, 2009, von Hinke Kessler Scholder et al., 2013, Cawley et al., 2011]

4

Obesity and Health: Evolution of health stock and investment + causes of obesity. See [Grossman, 1972, Lakdawalla et al., 2005, Galama and Kapteyn, 2011, Scholz and Seshadri, 2013, Cawley, 2010, Cutler et al., 2003]

Pietro Biroli

11

Intro GxEcon Empirics Structural Conclusion

Research question Motivation

Related Literature 1

Nature and Nurture (Twin Studies): Use MZ/DZ twins to estimate genetic and environmental heritability. See [Galton, 1874, Taubman, 1976, Kohler et al., 2011]

,→ Contribution: Specific biological mechanisms 2

Genoeconomics: Find genetic determinants of economic behaviors: risk aversion, time and social pref., addiction. See [Benjamin et al., 2007, Cesarini et al., 2009, Benjamin et al., 2016]

,→ Contribution: understand how known genes shape econ choices 3

Mendelian Randomization: Genes as IV

See [Davey Smith, 2003, Fletcher and Lehrer, 2009, von Hinke Kessler Scholder et al., 2013, Cawley et al., 2011]

,→ Contribution: Genes as measure of heterogeneity 4

Obesity and Health: Evolution of health stock and investment + causes of obesity. See [Grossman, 1972, Lakdawalla et al., 2005, Galama and Kapteyn, 2011, Scholz and Seshadri, 2013, Cawley, 2010, Cutler et al., 2003]

,→ Contribution: look at fat mass, exercise, diet over the life-cycle Pietro Biroli

12

Intro GxEcon Empirics Structural Conclusion

Research question Motivation

Preview of Results Raw difference in BMI by FTO-genotype: 0.4 to 0.7 +26% prob overweight Small, as in most GWAS studies, but economically significant and similar to BMI gradient in SES (0.3), education (0.4) Evidence of both productivity and preferences shifts risky-FTO genotype eat 2% more calories convert calories into BMI at 1/3 higher rate

Stronger effect of genes for adults born in later cohort Simulations show that policy responses vary by genotype

Pietro Biroli

13

Intro GxEcon Empirics Structural Conclusion

Research question Motivation

Road map

Genetic and economic framework Empirical Findings Avon Longitudinal Study of Parents and Children (ALSPAC) Framingham Heart Study (FHS)

Structural Model and Simulations

Pietro Biroli

14

Intro GxEcon Empirics Structural Conclusion

Outline 1

Introduction Research question Motivation

2

Gene x Econ

3

Empirical Findings ALSPAC Data Results Productivity Effect Preferences Robustness Checks

Replication using FHS 4

Structural Model The Model Simulations

5

Conclusion and Future Work

Pietro Biroli

15

Intro GxEcon Empirics Structural Conclusion

Simple Model

max U (B, F , `, c; g) E,F

s.t Ω=`+E

(1)

Y = pF F + c

(2)

B = I(F , E; g) + (1 − δ)B 0 + ε

(3)

Utility from BMI B, consumption c, food consumption F , and leisure ` Income Y is devoted to buying food F and non-food consumption c time Ω devoted to exercise E vs leisure ` Genotype g influences:

Cost of investment [disutility: U(.;g)] Productivity of investment [I(.; g)] Pietro Biroli

16

Intro GxEcon Empirics Structural Conclusion

The Model: Genetic-Economic Interaction

Body Mass Index

(a) Shift the production function

BMI Isoquant T-Allele

BMI

Investment 1: (a) Productivity Effect Calories Pietro Biroli

17

Intro GxEcon Empirics Structural Conclusion

The Model: Genetic-Economic Interaction (a) Shift the production function

Body Mass Index

BMI Isoquant A-Risky BMI Isoquant T-Allele

BMI

Investment 1: (a) Productivity Effect Calories Pietro Biroli

18

Intro GxEcon Empirics Structural Conclusion

The Model: Genetic-Economic Interaction

Body Mass Index

BMI Isoquant A-Risky BMI Isoquant T-Allele

BMI

Investment 1: (a) Productivity Effect Calories Pietro Biroli

Sedentary Minutes

(a) Shift the production function (b) Change the utility cost of investment

BMI Isoquant T-Allele

BMI

Indifference Set T-Allele

(b) Cost Effect

Calories

19

Intro GxEcon Empirics Structural Conclusion

The Model: Genetic-Economic Interaction

Body Mass Index

BMI Isoquant A-Risky BMI Isoquant T-Allele

BMI

Investment 1: (a) Productivity Effect Calories Pietro Biroli

Sedentary Minutes

(a) Shift the production function (b) Change the utility cost of investment

BMI Isoquant A-Risky BMI Isoquant T-Allele

BMI

Indifference Set T-Allele

(b) Cost Effect

Indifference Set A-Risky

Calories

20

Intro GxEcon Empirics Structural Conclusion

Examples of genetic effects: MAOA and ALDH2

Productivity: (MAOA gene) x (Maltreatment) = Antisocial Behavior [Caspi et al., 2002]

Utility Cost: ALDH2 gene

[[Davey Smith, 2010]]

Alcohol intake induces facial flushing → Higher cost of drinking → Lower alcohol intake → Lower risk of liver cirrhosis Pietro Biroli

21

Intro GxEcon Empirics Structural Conclusion

The gene variant rs9939609 Gene: FTO intron, long-range connection with IRX [Smemo et al., 2014] Risky A-allele connected to obesity by GWAS How?: Regulates appetite Appetite-stimulant hormone (ghrelin) Neural responsiveness to food images Expressed in the hunger-related sites of the brain

⇒ could increase the utility cost of dieting [Karra et al., 2013, Speakman et al., 2008, Fawcett and Barroso, 2010, Wardle et al., 2008, Cecil et al., 2008, Olszewski et al., 2009, Fredriksson et al., 2008, Timpson et al., 2008, Smemo et al., 2014, Claussnitzer et al., 2015] pics

More exercise associated with lower genetic differences in BMI Weight-loss in dieting programs associated with FTO ⇒ could change the productivity of investments [Andreasen et al., 2008, Franks et al., 2008, Kilpel¨ ainen et al., 2011, Huang et al., 2014, Zhang et al., 2012] Pietro Biroli

22

Intro GxEcon Empirics Structural Conclusion

ALSPAC Data Results Replication using FHS

Outline 1

Introduction Research question Motivation

2

Gene x Econ

3

Empirical Findings ALSPAC Data Results Productivity Effect Preferences Robustness Checks

Replication using FHS 4

Structural Model The Model Simulations

5

Conclusion and Future Work

Pietro Biroli

23

Intro GxEcon Empirics Structural Conclusion

ALSPAC Data Results Replication using FHS

ALSPAC Data Avon Longitudinal Study of Parents and Children (ALSPAC) Cohort of children born in 1991-1992 near Bristol (UK) Data from clinic visits Enrolled ≈ 14,000 pregnant mothers, ≈ 8,000 children with genetic data Obesity: Body Mass Index (BMI), ages 1 to 18 Investments: ages 11 and 13 - Child Physical Activity: uni-axial accelerometer MTI Actigraph; see [Mattocks et al., 2008]

- Child Diet: 3-day dietary diary Nutrients with reporting adjustment, see [Noel et al., 2010]

- Genetic data collected at age 7

Pietro Biroli

24

Intro GxEcon Empirics Structural Conclusion

ALSPAC Data Results Replication using FHS

The Children of the 90s

Pietro Biroli

25

Intro GxEcon Empirics Structural Conclusion

ALSPAC Data Results Replication using FHS

Summary Statistics Table: Summary Statistics, Age 11 and 13 by genotype and father SES Genotype T-Allele A-Risky Body Mass Index

Overweight (%)

Kilocalories/day

Sedentary Hours/day

n obs.

19.10 (11.24) [0.07]*** 22.17 (17.26) [0.82]*** 1.89 (0.21) [0.01]** 7.51 (1.54) [0.02] 2562

19.47 (11.07) [0.05]*** 28.48 (20.37) [0.67]*** 1.92 (0.19) [0.01]** 7.55 (1.59) [0.02] 4490

Father SES High Low 19.17 (10.09) [0.06]** 23.39 (17.92) [0.75]*** 1.91 (0.19) [0.01] 7.67 (1.46) [0.02]** 3722

19.39 (11.68) [0.06]** 27.47 (19.93) [0.82]*** 1.90 (0.20) [0.01] 7.43 (1.65) [0.02]** 3330

Total 19.33 (11.17) 26.19 (19.33) 1.91 (0.20) 7.54 (1.58) 7052

2

Mean of Body Mass Index (BMI kg/m ), percentage overweight (BMI greater than 85% pct), sedentary hours, and Kilocalories (x1000), by FTO variant and father SES. Sample variance in parenthesis; mean standard-error in brackets. 49% of the sample is male. 63% of the sample carries one or two A-Alleles in the rs9939609 SNP of the FTO gene (15% are heterozygous AA, Minor Allele Frequency of 0.39 representative of UK population). High SES: manager or professional (47%); low: worker (skilled or unskilled), based on OPCS occupation codes. Pietro Biroli

By Age

Investments

Anthropometrics

Gender

26

Intro GxEcon Empirics Structural Conclusion

ALSPAC Data Results Replication using FHS

Evolution of Body Mass Index

Mom SES Pietro Biroli

Dad Edu

Mom Edu

Income 27

Intro GxEcon Empirics Structural Conclusion

ALSPAC Data Results Replication using FHS

Evolution of Body Mass Index

Mom SES Pietro Biroli

Dad Edu

Mom Edu

Income 28

Intro GxEcon Empirics Structural Conclusion

ALSPAC Data Results Replication using FHS

Evolution of Body Mass Index

Mom SES Pietro Biroli

Dad Edu

Mom Edu

Income 29

Intro GxEcon Empirics Structural Conclusion

ALSPAC Data Results Replication using FHS

Summary Statistics, Environment and Covariates Table: Family Characteristics, by Child FTO genotype

Mother Edu Father Edu Mother SES Father SES Mother BMI Mother age at birth Teen mother (%) Single Mother (%) Parity Birth Weight (kg)

FTO genotype T-Allele A-Risky

Total

3.36 [0.03] 3.32 [0.04] 2.75 [0.02] 2.88 [0.03] 22.74** [0.10] 29.33 [0.12] 1.51 [0.33] 15.85 [0.98] 0.69 [0.02] 3.42 [0.01]

3.34 [0.02] 3.33 [0.02] 2.77 [0.02] 2.86 [0.02] 22.90 [0.06] 29.34 [0.07] 1.88 [0.22] 15.49 [0.58] 0.72 [0.01] 3.42 [0.01]

3.33 [0.02] 3.34 [0.03] 2.78 [0.02] 2.84 [0.02] 23.00** [0.08] 29.35 [0.09] 2.10 [0.29] 15.28 [0.73] 0.73 [0.02] 3.43 [0.01]

Average value of the covariates for the sample used in the main analysis. Pooled across genders and separated by FTO-genotype. Standard errors of means in brackets. Mean difference * significant at 10%; ** significant at 5%; *** significant at 1%. Education ranges from lowest (1 = CSE or less) to highest (5 = degree). Socio-Economic-Status ranges from from highest (1 = professional) to lowest (6 = unskilled). Teen mother is a dummy for mothers who were pregnant before age 19. Single mother is a dummy for a household without a male figure. Pietro Biroli

30

Intro GxEcon Empirics Structural Conclusion

ALSPAC Data Results Replication using FHS

Gene×Calories Interaction

Figure:

Nonparametric local-mean smoothing using Epanechnikov kernel and Silverman’s Rule-of-Thumb bandwidth. Combining information from successive clinical visits, age 11 and 13; excluding outliers in the top and bottom 5% of the distributions of BMI and log(energy intake). Pietro Biroli

Male

Female

31

Intro GxEcon Empirics Structural Conclusion

ALSPAC Data Results Replication using FHS

Gene×Exercise Interaction

Figure:

Nonparametric local-mean smoothing using Epanechnikov kernel and Silverman’s Rule-of-Thumb bandwidth. Combining information from successive clinical visits, age 11 and 13; excluding outliers in the top and bottom 5% of the distributions of BMI and log(sedentary minutes). Pietro Biroli

Male

Female

32

Intro GxEcon Empirics Structural Conclusion

ALSPAC Data Results Replication using FHS

Genetic Productivity Effect Log-linearize a Cobb Douglas production function for obesity

log(Bi,t ) =µ + µg g + αe log(Ei,t ) + αf log(Fi,t )+ + αg×e log(Ei,t ) · g + αg×f log(Fi,t ) · g+ + δlog(Bi,t−1 ) + γb log(Bimom ) + h(Xi,t ) + κt + εi,t

Level effect: µg =

∂f ∂g

Productivity effect: αGxK =

∂f ∂investment



− g=A

∂f investment



g=T

Xi covariates: mom and dad education and SES; mother age at pregnancy; parity; birth weight; age of child at clinic date; dummy for single mother; time dummy; seasonal dummies; month effects; low kilo-calories reporting; late respondent; Pietro Biroli

33

Intro GxEcon Empirics Structural Conclusion

ALSPAC Data Results Replication using FHS

Identification and limitations

Identification: Mendelian Randomization: Mendel’s first law of segregation Genotype random conditional on parental g Dad genotype unobserved → bound using [Altonji et al., 2008] Limitations: Measurement error and misreporting → attenuation Potential endogeneity of investments

Pietro Biroli

34

Intro GxEcon Empirics Structural Conclusion

ALSPAC Data Results Replication using FHS

Reduced Form Table: Gene and Investment Interaction - FTO log(Body Mass Indext ) (2) (3)

(1) Risky FTO Gene

βg

log(Food Intake)

αf

G X Food Intake

αg×f

log(Sedentary min.)

αe

G X Sedentary min.

αg×e

log(BMI)t−1

(1 − δ)

log(BMI)mom

γb

Covariates Mom Gene R2 Observations n

0.019 [0.005]***

0.006 [0.002]***

0.969 [0.007]*** 0.090 [0.007]*** X

0.010 [0.002]*** 0.067 [0.009]*** 0.025 [0.011]** 0.027 [0.009]*** 0.011 [0.011] 0.939 [0.008]*** 0.090 [0.007]*** X

78% 7052 3526

78% 7052 3526

0.32% 7052 3526

(4)

(5)

0.010 [0.003]*** 0.059 [0.010]*** 0.027 [0.011]** 0.028 [0.011]*** 0.010 [0.011] 0.947 [0.013]*** 0.097 [0.012]*** X X 78% 7052 3526

0.010 [0.003]*** 0.069 [0.009]*** 0.026 [0.011]** 0.024 [0.009]*** 0.012 [0.011] 0.967 [0.008]***

78% 7052 3526

Dependent variable: log BMI (kg/m2 ); Risky FTO gene g = 1 if rs9939609 gene variant contains one or more A-Alleles; g = 0 otherwise; Covariates: gender; parity; age of child at clinic date; mom and dad education and SES; mother age at pregnancy; dummy for single mother; reliable dietary report; time dummy; seasonal dummies; late respondent; birth weight. * significant at 10%; ** significant at 5%; *** significant at 1%. Standard error clustered at the individual level in brackets. Pietro Biroli

Sizable effect: ≈ 1/4 kg

35

Intro GxEcon Empirics Structural Conclusion

ALSPAC Data Results Replication using FHS

Utility Cost Second genetic effect: change in the demand for investments A-Allele Higher food intake No differences in activity Table: Utility Cost Effect Male Risky FTO Gene Covariates Observations

Calories Female

(1) 0.020 [0.009]** X 3,347

(2) 0.018 [0.008]** X 3,711

Sedentary Activity Male Female (3) 0.006 [0.007] X 3,347

(4) 0.005 [0.006] X 3,711

Dependent variables: log of daily kilocalories intake (columns (1) and (2)), and log of daily sedentary minutes (columns (3) and (4)). Covariates: log(BMI)t−1 ; log mom BMI during pregnancy; parity; age of child at clinic date; mom and dad education and SES; mother age at pregnancy; dummy for single mother; reliable dietary report; time dummy; month dummies; late respondent; birth weight. * significant at 10%; ** significant at 5%; *** significant at 1%. Standard error clustered at the individual level in brackets.

Sizable effect: ≈ 1.5 kg/year Pietro Biroli

36

Intro GxEcon Empirics Structural Conclusion

ALSPAC Data Results Replication using FHS

Decomposition of the Genetic Effect

Decompose the overall effect in difference in parameters and difference in inputs (Oaxaca 1973): BMI g = W g αg ⇒ BMI A − BMI T = W T (αA − αT ) + (W A − W T )αA {z } | | {z } | {z } ∆ BMI

∆ parameters

∆ inputs

Difference in Parameters: 35.4% [26%,39%] → productivity Difference in Inputs:

Pietro Biroli

64.6% [47%,72%] → preferneces

37

Intro GxEcon Empirics Structural Conclusion

ALSPAC Data Results Replication using FHS

Robustness

Check the robustness of the results: → Polygenic Score

Gender+NoUnder

→ Dropping underweight children (≈ 4%) → Different measures of fat-mass

Fat Mass

→ Different measures of investments → Different quantiles

Pietro Biroli

Gender+NoUnder

Food

Quantiles

38

Intro GxEcon Empirics Structural Conclusion

ALSPAC Data Results Replication using FHS

Replication of the Results

Framingham Heart Study (FHS), Offspring Cohort Information on 5,124 individuals, children of the original cohort population (1948) Born over a 60-year period (1905-1965) 8 clinical exams from 1971 to 2008 Genetic info: 1987-1991, 98% consent 4 waves with BMI, caloric intake, and physical activity

Pietro Biroli

39

Intro GxEcon Empirics Structural Conclusion

ALSPAC Data Results Replication using FHS

FHS: Log-Linear Regression Table: FHS: Gene and Investment Interaction - FTO (1) Risky FTO variant

βg

log(Energy Intake)

αf

G X Energy Intake

αg×f

log(Physical Activity)

αe

G X Physical Activity

αg×e

log(BMI)t−1

(1 − δ)

Covariates R2 Observations n

log(Body Mass Indext ) (2) (3) born after 1940

0.024*** [0.007]

0.043*** [0.010]

0.4% 8258 2753

1.2% 4918 1639

0.002 [0.001] 0.013*** [0.004] 0.010** [0.005] -0.005** [0.002] 0.003 [0.003] 0.937*** [0.006] x 85.3% 8258 2753

(4) born after 1940 0.005** [0.002] 0.022*** [0.005] 0.016** [0.006] -0.009*** [0.003] 0.001 [0.004] 0.927*** [0.009] x 84.7% 4642 1547

Dependent variable: log BMI (kg/m2 ); Risky FTO gene g = 1 if rs9939609 gene variant contains one or more A-Alleles; g = 0 otherwise; Covariates: gender; 3-degree polynomial in age; dummies education and income; dummies for marital status; reliable dietary report; time dummies; birth cohort dummies; 20 first principal components of genome. * significant at 10%; ** significant at 5%; *** significant at 1%. Standard error clustered at the individual level in brackets.

Pietro Biroli

40

Intro GxEcon Empirics Structural Conclusion

ALSPAC Data Results Replication using FHS

Birth Year Effects

Pietro Biroli

41

Intro GxEcon Empirics Structural Conclusion

ALSPAC Data Results Replication using FHS

Prices, Income, Food Availability

Pietro Biroli

42

Intro GxEcon Empirics Structural Conclusion

The Model Simulations

Outline 1

Introduction Research question Motivation

2

Gene x Econ

3

Empirical Findings ALSPAC Data Results Productivity Effect Preferences Robustness Checks

Replication using FHS 4

Structural Model The Model Simulations

5

Conclusion and Future Work

Pietro Biroli

43

Intro GxEcon Empirics Structural Conclusion

The Model Simulations

Life-cycle Model

Vt (Bt , Yt , εt ; g) = max u (Bt , Ft , `t , ct ; g) + βEVt+1 (Bt+1 , Yt+1 , εt+1 ; g) E t ,Ft

s.t Ω(Bt ) = `t + E t Yt = pFt Ft + ct Bt+1 = I(Ft , E t ; g) + (1 − δt )Bt + εt

(4) (5) (6)

Utility from BMI Bt , food consumption Ft , and leisure `t Income Yt is devoted to buying food Ft (calories) and non-food consumption ct time Ωt devoted to exercise E t vs leisure `t Genotype g influences:

Cost of investment [disutility: U(.;g)] Productivity of investment [I(.; g)]

Pietro Biroli

44

Intro GxEcon Empirics Structural Conclusion

The Model Simulations

First Order Conditions for Investment Simple 3-period model. The optimal caloric consumption: ∂I(F2 , E 2 ; g) ∂F2 ∂I(F , E 2 2 ; g) = pF Uc02 βE [ϕB3 ] ∂F2

UF0 2 (.; g) = pF Uc02 + βE [−UB0 ]

→ LHS: Productivity effect → RHS: Utility Cost effect Similarly, in the first period: UF0 1 (.; g) = pF Uc01 + βE [ϕB2 + β(1 − δ2 )ϕB3 ]

∂I(F1 , E 1 ; g) ∂F1

Genotype changes the incentive to invest in every period, and this cumulates over time. Pietro Biroli

45

Intro GxEcon Empirics Structural Conclusion

The Model Simulations

Evolution of Body Mass Index

Source: ALSPAC Pietro Biroli

Males

Density 46

Intro GxEcon Empirics Structural Conclusion

The Model Simulations

Calibration Use parameters estimated in reduced form to calibrate: Body Mass Index

20.0 19.5 19.0 18.5

B M 18.0 I 17.5 17.0 16.5

16.0 1

2

3

4

5

6

7

8

9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27

Age

Pietro Biroli

A-Risky

T-Allele

47

Intro GxEcon Empirics Structural Conclusion

The Model Simulations

Policy A: Food Tax Higher food prices Body Mass Index 50% Tax on Calories

20.0 19.5 19.0 18.5

B M 18.0 I 17.5 17.0 16.5

16.0 1

2

3

4

5

6

7

8

9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26

Age

Pietro Biroli

A-Risky

T-Allele

48

Intro GxEcon Empirics Structural Conclusion

The Model Simulations

Policy B: School Eating Reduce caloric consumption by 25% in the first 10 years of life Body-Mass-Index Healthy Eating in School

20.0 19.5 19.0 18.5

B M 18.0 I 17.5 17.0 16.5

16.0 1

2

3

4

5

6

7

8

9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26

Age

Pietro Biroli

A-Risky

T-Allele

49

Intro GxEcon Empirics Structural Conclusion

Outline 1

Introduction Research question Motivation

2

Gene x Econ

3

Empirical Findings ALSPAC Data Results Productivity Effect Preferences Robustness Checks

Replication using FHS 4

Structural Model The Model Simulations

5

Conclusion and Future Work

Pietro Biroli

50

Intro GxEcon Empirics Structural Conclusion

Conclusion: Model of health behaviors and BMI over the life-cycle integrating genetics and economics Genes set the stage for human capital investments Genetic heterogeneity in both productivity and preferences + 1/3 productivity of calories + 2% caloric intake

Effect of genes varies with environment where people are born “Nature and nurture” interact Behaviors can mitigate genetic “risks” Policies can have differential effect depending on genotype

Pietro Biroli

51

Intro GxEcon Empirics Structural Conclusion

Future Work:

Potential Extensions: Preference formation and addiction Parameter uncertainty and learning

Related questions: Genetic architecture of utility parameters cognitive and soft skills ([Rietveld et al., 2014])

Epigenetics: gene expression changing with environment Better identification of “Environment” and G×E

Pietro Biroli

52

Thank You

Pietro Biroli

53

Appendix

Pietro Biroli

54

Genes and DNA Human DNA is composed of an estimated 3 billion base pairs and 20k-25k genes. Genes encode for proteins that regulate the development and functioning of the human body The most common form of genetic variation are SNPs, single neuclotide polymorphisms, and they represent a difference in a single DNA building block (nucleotide). Each SNP contains two alleles, one inherited from your father and the other from your mother. 10 million SNPs are estimated to be in the human genome A genotype is your genetic type at a particular genetic locus, for example a genotype can be AA, AT, or TT. In this paper I consider together those who have at least one A-allele, so that AT and AA are considered as having the same risk (dominant model) A phenotype is the observable outcome of interest, in this case BMI back

Pietro Biroli

55

Cost of Genotyping

Private Genotyping company - 23&me: $100 Back Pietro Biroli

56

Gene × Environment Interaction BUT: responses to environment changes by genotype

Back

Pietro Biroli

57

Parametrization of the Model

Utility u (B, F , `, c; g) = ζB log B + ζF (g) log F + ζ` log ` + ζc log c Production function log Bt+1 = log φ(g)+a(g) log Ft +b(g) log E t + (1 − δ1 − t/T δ2 ) log Bt +εt

10 Parameters ζB , ζF (g), ζ` , ζc , φ(g), a(g), b(g), δ1 , δ2 , σε2




4 vary by genotype back

Pietro Biroli

58

Calibrated Parameters Parameters taken from the literature: β = 0.97 (As in Hubbard, Skinner, and Zeldes (1995); and Engen, Gale, and Uccello (1999)) ζc = 0.36 (As in Scholz and Seshadri (2013))

Calibrated: ζB = 0.4 ζ` = 0.4 ζF (0) = 0.1 ζF (1) = 0.2 a(0) = 0.06 a(1) = 0.09 b(0) = 0.3 b(1) = 0.3 φ(0) = 1.0 φ(1) = 1.1 δ1 = 0.02 δ2 = 0.04 back Pietro Biroli

59

Alternative Parametrization of the Model

Utility o 1−σ n  ρ ρ λ F η(g) `1−η(g) + (1 − λ)B ρ u (B, F , `; g) =

1−σ

+ αct

11 Parameters: λ, η(g), ρ, σ, α, φ(g), a(g), b(g), δ1 , δ2 , σε2 4 vary by genotype back

Pietro Biroli

60

Moments to Match

Match the following moments from the ALSPAC data: Ft , E t at ages 11 and 13 Average, median, and BMI cutoff at different ages Cov (Bt , Ft−1 ) Cov (Bt , E t−1 ) Cov (Bt , Bt−1 ) Cov (Ft , E t ) back

Pietro Biroli

61

Brain Imaging, Appetite, and FTO

Figure:

Brain regions where the TT and AA genotypes exhibited different BOLD responses in fMRI when viewing food/non-food images while fasting (A-F); or comparing interaction between fed/fasting and high-incentive/low-incentive-value food (D-F)

Back Pietro Biroli

62

Brain Imaging, Appetite, and FTO

Figure: Pietro Biroli

Brain regions where the circulating acyl-ghrelin differentially affected brain fMRI responses in TT and AA genotypes

Back 63

Endogeneity of Inputs

So far, assumed Ie , Id ⊥ ⊥ εH Now, consider system of equations:   Bt Id   Ie

= f (Id , Ie , Bt−1 , X ; g) + εH = Id (Ie , Bt−1 , X , Z ; g) + εd = Ie (Id , Bt−1 , X , Z ; g) + εe

Exclusion restriction: Z ⊥ ⊥ εk Lagged investments Ik,t−1 Income Y , family composition, distance to school Mother and Father behaviors back

Pietro Biroli

64

Instrumented Regression Table: Health Production Function - Instrumented Regression

Risky FTO Gene

βg

log(Food Intake)

αf

G X Food Intake

αg×f

log(Sedentary min.)

αe

G X Sedentary min.

αg×e

log(BMI)t−1

(1 − δ)

log(BMI)mom

γb

Covariates Observations

(1) OLS

(2) Lagged Invest

(3) Income

(4) Parental Behavior

0.010 [0.002]*** 0.067 [0.009]*** 0.025 [0.011]** 0.027 [0.009]*** 0.012 [0.011] 0.939 [0.008]*** 0.090 [0.007]*** X 7052

0.006 [0.002]** 0.073 [0.039]* 0.029 [0.026] -0.001 [0.030] 0.102 [0.026]*** 0.952 [0.009]*** 0.024 [0.022] X 6264

0.025 [0.020] 0.365 [0.346] 0.354 [0.366] -0.203 [0.248] -0.195 [0.364] 0.927 [0.020]*** 0.005 [0.015] V 7052

0.037 [0.026] 0.660 [0.639] 0.724 [0.580] -0.250 [0.481] -0.086 [0.634] 0.929 [0.034]*** 0.091 [0.008]*** X 7052

Dependent variable: log BMI (kg/m2 ). 3-stage-least-square estimation. Column (1) reports the baseline results from OLS regression in table (3). Column (2) uses lagged values of food intake, protein intake, and sugar intake as instruments for caloric intake; lagged sedentary minutes, moderate to vigorous activity, and counts per minutes as instruments for investment in exercise. Column (3) uses income and financial difficulties, mother and father SES, mother and father education, distance to school, and number of siblings as instruments for both investments. Column (4) uses mother and father Food intake when child was 4-years old, and mother self-reported level of physical activity as instruments for investments. * significant at 10%; ** significant at 5%; *** significant at 1%. Standard error clustered at the individual level and correlated across equations in brackets. Risky FTO gene g = 1 if rs9939609 gene variant contains one or more A-alleles; g = 0 otherwise; Covariates X: gender; parity; age of child at clinic date; mom and dad education and SES; mother age at pregnancy; dummy for single mother; reliable dietary report; time dummy; seasonal dummies; late respondent; birth weight. Covariates V: gender; age of child at clinic date; mother age at pregnancy; reliable dietary report; time dummy; seasonal dummies; late respondent; birth weight. Pietro Biroli

back 65

Utility Cost Effect Table: Investment Equation - Food Intake (2)

(3)

(4)

Risky FTO Gene log(Sed Min)

0.017 [0.006]*** -0.222 [0.052]***

Risky FTO Gene log(Sed Min)

0.109 [0.077] -19.590 [0.128]***

Risky FTO Gene log(Sed Min)

0.112 [0.076] -19.459 [0.184]***

Lagged Food Int. Lagged Protein Int. Lagged Sugar Int.

0.198 [0.021]*** 0.074 [0.014]*** 0.042 [0.009]***

Income

0.009 [0.073] -0.012 [0.045] -0.106 [0.036]*** 0.259 [0.041]*** 0.071 [0.036]** 0.176 [0.041]*** -0.153 [0.048]*** V 7052

Mom Food Int. (age 4) Dad Food Int. (age 4) Mom Physical Activity

0.063 [0.143] 0.018 [0.183] -0.227 [0.099]**

Mom SES Dad SES Mom Edu Dad Edu Distance Num Sibling

Covariates Observations

X 6264

Instrument: Lag Investment

Income and distance to school

X 7052 Parental Behavior

Dependent variable: log(Food Intake). 3-stage-least-square estimation. Column (2) uses lagged values of food intake, protein intake, and sugar intake as instruments for caloric intake; lagged sedentary minutes, moderate to vigorous activity, and counts per minutes as instruments for investment in exercise. Column (3) uses income and financial difficulties, mother and father SES, mother and father education, distance to school, and number of siblings as instruments for both investments. Column (4) uses mother and father Food intake when child was 4-years old, and mother self-reported level of physical activity as instruments for investments. * significant at 10%; ** significant at 5%; *** significant at 1%. Standard error clustered at the individual level and correlated across equations in brackets. Risky FTO gene g = 1 if rs9939609 gene variant contains one or more A-alleles; g = 0 otherwise; Covariates X: gender; parity; age of child at clinic date; log mom BMI during pregnancy; mom and dad education and SES; mother age at pregnancy; dummy for single mother; reliable dietary report; time dummy; seasonal dummies; late respondent; birth weight. Covariates V: gender; age of child at clinic Pietro Biroli back date; mother age at pregnancy; reliable dietary report; time dummy; seasonal dummies; late respondent; birth weight. 66

Utility Cost Effect - 2 Table: Investment Equation - Sedentary Minutes (2)

(3)

(4)

Risky FTO Gene log(Food Intake)

0.008 [0.004]* -0.187 [0.010]***

Risky FTO Gene log(Food Intake)

0.006 [0.004] -0.051 [0.000]***

Risky FTO Gene log(Food Intake)

0.006 [0.004] -0.051 [0.000]***

Lagged Sedentary Min Lagged Vig. Activity Lagged Counts per min

0.191 [0.020]*** 0.021 [0.006]*** -0.140 [0.019]***

Income

0.000 [0.004] -0.001 [0.002] -0.005 [0.002]*** 0.013 [0.002]*** 0.004 [0.002]** 0.009 [0.002]*** -0.008 [0.002]*** V 7052

Mom Food Int. (age 4) Dad Food Int. (age 4) Mom Physical Activity

0.003 [0.007] 0.001 [0.009] -0.012 [0.005]**

Mom SES Dad SES Mom Edu Dad Edu Distance Num Sibling

Covariates Observations

X 6264

Instrument: Lag Investment

Income and distance to school

X 7052 Parental Behavior

Dependent variable: log(Sedentary min.). 3-stage-least-square estimation. Column (2) uses lagged values of food intake, protein intake, and sugar intake as instruments for caloric intake; lagged sedentary minutes, moderate to vigorous activity, and counts per minutes as instruments for investment in exercise. Column (3) uses income and financial difficulties, mother and father SES, mother and father education, distance to school, and number of siblings as instruments for both investments. Column (4) uses mother and father Food intake when child was 4-years old, and mother self-reported level of physical activity as instruments for investments. * significant at 10%; ** significant at 5%; *** significant at 1%. Standard error clustered at the individual level and correlated across equations in brackets. Risky FTO gene g = 1 if rs9939609 gene variant contains one or more A-alleles; g = 0 otherwise; Covariates X: gender; parity; age of child at clinic date; log mom BMI during pregnancy; mom and dad education and SES; mother age at pregnancy; dummy for single mother; reliable dietary report; time dummy; seasonal dummies; late respondent; birth weight. Covariates V: gender; age of child at clinic Pietro Biroli age at pregnancy; reliable dietary report; time dummy; seasonal dummies; late respondent; birth weight. date; mother

back 67

Summary Statistics, Investments Table: Summary Statistics, Food Intake and Exercise FTO genotype T-Allele A-Risky Kilocalories (x1000) Fat Intake (grams/day) Dietary Cholesterol Intake (grams/day) Carbohydrate Intake (grams/day) Total Sugar Intake (grams/day) Physical Activity (Sedentary Hours) Physical Activity (Moderate To Vigorous) Physical Activity (counts per minute) Very Active (self-report)

Total

1.89** [0.01] 75.82** [0.45] 188.66** [1.88] 252.83* [1.30] 114.74 [0.91]

1.92** [0.01] 77.10** [0.33] 193.39** [1.44] 255.58* [0.94] 115.87 [0.64]

1.91 [0.01] 76.64 [0.27] 191.67 [1.15] 254.58 [0.76] 115.46 [0.53]

7.51 [0.02] 23.92 [0.32] 581.96 [3.73] 3.69 [0.02]

7.55 [0.02] 23.68 [0.25] 576.78 [2.84] 3.71 [0.01]

7.54 [0.01] 23.77 [0.20] 578.66 [2.26] 3.7 [0.01]

Average measures of investment in diet, and investment in exercise. Pooled across gender and ages, separated by FTO-genotype. Standard errors of means in brackets. Mean difference * significant at 10%; ** significant at 5%; *** significant at 1%. 3-day dietary records coded using the Diet In Data Out software. Actigraph data: counts per min., min. of sedentary activity, and moderate to vigorous activity. Self-reported activity ranged from 1 (never) to 5 (daily). Pietro Biroli

Back

68

Summary Statistics, by age

Table: Summary Statistics by age, gender, and genotype Body Mass Index Age 8

11

13

Sedentary Hours

Female T-Allele A-Allele

Male T-Allele A-Allele

16.25 (4.71) [0.07] 18.50 (10.39) [0.10] 20.41 (11.84) [0.12]

16.06 (3.37) [0.06] 18.17 (8.56) [0.09] 19.74 (10.29) [0.11]

16.42 (4.57) [0.05] 18.99 (10.80) [0.08] 20.87 (12.56) [0.09]

16.13 (3.59) [0.04] 18.62 (10.29) [0.07] 20.08 (11.68) [0.09]

Female T-Allele A-Allele . . . 7.18 (1.19) [0.04] 8.26 (1.32) [0.05]

. . . 7.25 (1.21) [0.03] 8.24 (1.31) [0.03]

Male T-Allele A-Allele . . . 6.89 (1.27) [0.04] 7.73 (1.50) [0.05]

. . . 6.98 (1.45) [0.03] 7.77 (1.54) [0.04]

Mean of Body Mass Index (BMI kg/m2 ), sedentary hours, and Kilocalories (in thousands), by age, gender, and FTO genotype. Sample variance in parenthesis; mean standard-error in brackets.

Back

Pietro Biroli

69

Summary Statistics, by age

Table: Summary Statistics by age, gender, and genotype Kilocalories Age 8

11

13

Female T-Allele A-Allele 1.64 (0.08) [0.01] 1.75 (0.13) [0.01] 1.77 (0.21) [0.02]

1.64 (0.08) [0.01] 1.78 (0.12) [0.01] 1.76 (0.18) [0.01]

Whole Sample

Male T-Allele A-Allele 1.75 (0.09) [0.01] 1.92 (0.15) [0.01] 2.12 (0.30) [0.02]

1.79 (0.11) [0.01] 1.97 (0.16) [0.01] 2.15 (0.27) [0.01]

BMI

Sed

Kcal

16.23 (4.08) [0.03] 18.64 (10.23) [0.04] 20.34 (11.92) [0.05]

. . . 7.10 (1.31) [0.02] 8.02 (1.47) [0.02]

1.71 (0.10) [0.00] 1.86 (0.15) [0.01] 1.95 (0.27) [0.01]

Mean of Body Mass Index (BMI kg/m2 ), sedentary hours, and Kilocalories (in thousands), by age, gender, and FTO genotype. Sample Variance in parenthesis; mean standard-error in brackets.

Back

Pietro Biroli

70

Summary Statistics, by gender Table: Summary Statistics, by gender Female T-Allele A-Risky Body Mass Index

Kilocalories/day

Sedentary Hours/day

19.34 (11.80) [0.09]*** 1.77 (0.16) [0.01] 7.70 (1.51) [0.03]

19.77 (11.73) [0.07]*** 1.79 (0.14) [0.01] 7.72 (1.52) [0.03]

Male

Total

T-Allele

A-Risky

18.83 (10.49) [0.09]*** 2.02 (0.23) [0.01]*** 7.29 (1.50) [0.04]

19.14 (10.16) [0.07]*** 2.06 (0.21) [0.01]*** 7.36 (1.62) [0.03]

19.33 (11.17) 1.91 (0.20) 7.54 (1.58)

Mean of Body Mass Index (BMI kg/m2 ), sedentary hours, and Kilocalories (x1000), by gender and FTO variant. Sample variance in parenthesis; mean standard-error in brackets. 49% of the sample is male. 63% of the sample carries one or two A-Alleles in the rs9939609 SNP of the FTO gene (15% are heterozygous AA, Minor Allele Frequency of 0.39, representative of UK population) Back

Pietro Biroli

71

Summary Statistics, Anthropometrics Table: Summary Statistics, Anthropometrics

Height (cm) Weight (kg) BMI kg/cm2 BMI z-score Fat Percentage Overweight (%) Underweight (%) Arm Circ. (cm) Waist Circ. (cm) Waist/Hip ratio

FTO genotype T-Allele A-Risky

Total

154.51 [0.21] 46.22*** [0.24] 19.10*** [0.07] 0.20*** [0.02] 24.31*** [0.19] 22.17*** [0.82] 4.18 [0.40] 23.90*** [0.07] 68.45*** [0.19] 0.82 [0.00]

154.7 [0.13] 46.88 [0.14] 19.33 [0.04] 0.3 [0.01] 25.02 [0.12] 26.19 [0.52] 3.79 [0.23] 24.18 [0.04] 69.05 [0.11] 0.82 [0.00]

154.81 [0.16] 47.26*** [0.18] 19.47*** [0.05] 0.35*** [0.02] 25.42*** [0.15] 28.49*** [0.67] 3.56 [0.28] 24.34*** [0.05] 69.39*** [0.14] 0.82 [0.00]

Body mass index normal z-scores calculated using 1990 British Growth Reference. Fat percentage: ratio of fat mass to total mass. Overweight and Underweight calculated using the BMI z-scores with a cutoff of 5% and 85%. Standard errors of means in brackets. Mean difference * significant at 10%; ** significant at 5%; *** significant at 1%. Back Pietro Biroli

72

Evolution of Body Mass Index

Back Pietro Biroli

Density 73

Evolution of Body Mass Index

Back Pietro Biroli

74

Evolution of Body Mass Index

Back Pietro Biroli

75

Evolution of Body Mass Index

Back Pietro Biroli

76

Evolution of Body Mass Index

Back Pietro Biroli

77

Gene×Calories Interaction

Figure:

Nonparametric local-mean smoothing using Epanechnikov kernel and Silverman’s Rule-of-Thumb bandwidth. Combining information from successive clinical visits, age 11 and 13; excluding outliers in the top and bottom 5% of the distributions of BMI and log(energy intake).

Pietro Biroli

Density

Back

78

Gene×Calories Interaction

Figure:

Nonparametric local-mean smoothing using Epanechnikov kernel and Silverman’s Rule-of-Thumb bandwidth. Combining information from successive clinical visits, age 11 and 13; excluding outliers in the top and bottom 5% of the distributions of BMI and log(energy intake). Pietro Biroli

Density

Back

79

Gene×Exercise Interaction

Figure:

Nonparametric local-mean smoothing using Epanechnikov kernel and Silverman’s Rule-of-Thumb bandwidth. Combining information from successive clinical visits, age 11 and 13; excluding outliers in the top and bottom 5% of the distributions of BMI and log(sedentary minutes).

Pietro Biroli

Density

Back 80

Gene×Exercise Interaction

Figure:

Nonparametric local-mean smoothing using Epanechnikov kernel and Silverman’s Rule-of-Thumb bandwidth. Combining information from successive clinical visits, age 11 and 13; excluding outliers in the top and bottom 5% of the distributions of BMI and log(sedentary minutes).

Pietro Biroli

Density

Back 81

Distribution of BMI, Females

Back

Pietro Biroli

82

Distribution of BMI, Males

Back

Pietro Biroli

83

Distribution of Caloric Intake, Females

Back

Pietro Biroli

84

Distribution of Caloric Intake, Males

Back

Pietro Biroli

85

Distribution of Exercise, Females

Back

Pietro Biroli

86

Distribution of Exercise, Males

Back

Pietro Biroli

87

Polygenic Approach

Consider other genes related to obesity from GWAS studies See [Vimaleswaran and Loos, 2010, Speliotes et al., 2010, Sandholt et al., 2012]

Construct a ‘gene-score’ by adding up the number of obesity-related alleles of 24 different genes, following [Belsky et al., 2013] MC4R TMEM18 FTO TFAP2B BCDIN3D ETV5 BDNF GNPDA2 PPARG THADA IGF2BP2 TCF7L2 NPC1 MTCH2 PCSK1 KCTD15 SH2B1 NRXN3 HHEX LYPLAL1 GCK NEGR1 PTER GCKR

dist

Obtain very similar results by considering the genetic-score The level of interaction is less pronounced, but the results are consistent with the previous tables Back

Pietro Biroli

88

Polygenic Approach Table: Gene and Investment Interaction - Genetic Score Risky Genetic Score log(Energy Intake)

βg

(2)

(3)

(4)

0.009 [0.002]***

0.012 [0.002]*** 0.066 [0.008]*** 0.026 [0.011]** 0.014 [0.008]* -0.003 [0.011] 0.965 [0.008]***

1.05% 7052

0.967 [0.007]*** 0.089 [0.007]*** X 78% 7052

0.012 [0.002]*** 0.065 [0.008]*** 0.025 [0.011]** 0.019 [0.008]** 0.000 [0.011] 0.938 [0.008]*** 0.090 [0.007]*** X 78% 7052

αf

G X Energy Intake

αg×f

log(Sedentary min.)

αe

G X Sedentary min.

αg×e

log(BMI)t−1

(1 − δ)

log(BMI)mom

γb

Covariates R2 Observations

(1) 0.034 [0.005]***

78% 7052

* significant at 10%; ** significant at 5%; *** significant at 1%. Standard error clustered at the individual level in brackets. Dependent variable: log BMI (kg/m2 ); Risky genetic score g = 1 if genetic score > 25; g = 0 otherwise; Covariates: gender; parity; age of child at clinic date; mom and dad education and SES; mother age at pregnancy; dummy for single mother; reliable dietary report; time dummy; seasonal dummies; late respondent; birth weight.

Back Pietro Biroli

89

Utility Cost Effect Effect of the genetic score on the investments: Varies by gender Differences also in activity levels Table: Genetic Effect on Investments - Genetic Score Caloric Consumption Male Female Risky Genetic Score Covariates Observations

(1) 0.011 [0.009] X 3,347

(2) 0.014 [0.008]* X 3,371

Sedentary Minutes Male Female (3) 0.001 [0.007] X 3,347

(4) 0.022 [0.006]*** X 3,371

* significant at 10%; ** significant at 5%; *** significant at 1%. Standard error clustered at the individual level in brackets. Dependent variables: logarithm of daily kilocalories intake (columns (1) and (2)), and logarithm of daily sedentary minutes (columns (3) and (4)). Covariates: log(sedentary min.) in columns (1) and (2), and log(kilocalories) in columns (3) and (4); log(BMI)t−1 ; log mom BMI during pregnancy; parity; age of child at clinic date; mom and dad education and SES; mother age at pregnancy; dummy for single mother; reliable dietary report; time dummy; month dummies; late respondent; birth weight.

→ Must understand better the biological function of the various genes Back Pietro Biroli

90

Heterogeneity by group Table: By Gender and Without Underweight

Risky FTO Gene

βg

log(Food Int.)

αf

G X Food Int.

αg×f

log(Sedentary m.)

αe

G X Sedentary m.

αg×e

Bt−1

(1 − δ)

Controls R2 Observations

(1)

(2)

(3)

Baseline

Males

Females

(4) No Underweight

0.010 [0.002]*** 0.067 [0.009]*** 0.025 [0.011]** 0.027 [0.009]*** 0.012 [0.011] 0.939 [0.008]*** X 78% 7,052

0.006 [0.004] 0.067 [0.013]*** 0.004 [0.016] 0.042 [0.013]*** 0.026 [0.016]* 0.947 [0.012]*** X 79% 3,346

0.010 [0.003]*** 0.082 [0.014]*** 0.044 [0.018]** 0.007 [0.013] -0.007 [0.016] 0.928 [0.011]*** X 79% 3,706

0.011 [0.003]*** 0.069 [0.009]*** 0.030 [0.011]*** 0.028 [0.009]*** 0.009 [0.011] 0.911 [0.008]*** X 77% 6,785

Column (1) reports the baseline estimates (same as table 3). Column (2) and (3) run the model separately for males and females. Column (4) runs the model dropping the children who are below the 5th percentile of the z-BMI standard distribution for the UK (they represent 4% of the sample). * significant at 10%; ** significant at 5%; *** significant at 1%. Standard errors in brackets. Risky FTO gene g = 1 if rs9939609 gene variant contains one or more A-Alleles; g = 0 otherwise. Covariates: gender; parity; age of child at clinic date; log mom BMI during pregnancy; mom and dad education and SES; mother age at pregnancy; dummy for single mother; reliable dietary report; time dummy; seasonal dummies; late respondent; birth weight.

Back Pietro Biroli

91

Measurement of Adiposity Table: Different Measures of Adiposity

Risky FTO Gene

βg

log(Food Int.)

αf

G X Food Int.

αg×f

log(Sedentary m.)

αe

G X Sedentary m.

αg×e

Bt−1

(1 − δ)

(1) Prob Overweight

(2) BMI and Height

Weight

zBMI

Fat %

0.228 [0.065]*** 0.500 [0.224]** 0.091 [0.274] 0.554 [0.218]** 0.082 [0.252] 2.101 [0.052]***

0.010 [0.002]*** 0.060 [0.009]*** 0.025 [0.011]** 0.026 [0.009]*** 0.012 [0.011] 0.934 [0.008]*** 0.106 [0.021]*** X 78% 7,050

0.012 [0.003]*** 0.072 [0.011]*** 0.030 [0.013]** 0.031 [0.011]*** 0.009 [0.013] 0.761 [0.008]*** 0.92 [0.031]*** X 88% 7,048

0.081 [0.019]*** 0.490 [0.070]*** 0.199 [0.083]** 0.189 [0.067]*** 0.076 [0.080] 0.869 [0.008]***

-0.011 [0.019] 0.036 [0.078] 0.029 [0.093] 0.141 [0.068]** 0.021 [0.081] 0.306 [0.022]***

X 77% 7,052

X 55% 5,305

log(Height) Controls R2 Observations

X 7,052

(3)

(4)

(5)

Column (1) runs a probit model on the probability of being obese. Column (2) uses Bt =log(weight) as dependent variable, controlling for log(height). Column (3) uses z-BMI as dependent variable. Column (4) uses the estimated percentage of body fat as dependent variable. * significant at 10%; ** significant at 5%; *** significant at 1%. Standard errors in brackets. Risky FTO gene g = 1 if rs9939609 gene variant contains one or more A-Alleles; g = 0 otherwise. Covariates: gender; parity; age of child at clinic date; log mom BMI during pregnancy; mom and dad education and SES; mother age at pregnancy; dummy for single mother; reliable dietary report; time dummy; seasonal dummies; late respondent; birth weight. Pietro Biroli Back

92

Measurement of Food Intake Table: Different Measures of Food Intake - FTO gene

Risky FTO Gene

βg

log(Food)

αf

G X Food

αg×f

log(Sedentary min.)

αe

G X Sedentary min.

αg×e

log(BMI)t−1

(1 − δ)

Covariates R2 Observations

(1)

(2)

(3)

(4)

(7)

Proteins

Fat

Carbs

(5) Dietary Cholesterol

(6)

Calories

Sugar

Starch

0.010 [0.002]*** 0.067 [0.009]*** 0.025 [0.011]** 0.027 [0.009]*** 0.012 [0.011] 0.939 [0.008]*** X 78% 7052

0.010 [0.002]*** 0.046 [0.007]*** 0.027 [0.009]*** 0.025 [0.009]*** 0.010 [0.011] 0.939 [0.008]*** X 78% 7052

0.009 [0.002]*** 0.037 [0.007]*** 0.015 [0.008]* 0.027 [0.009]*** 0.013 [0.011] 0.944 [0.008]*** X 78% 7052

0.008 [0.002]*** 0.047 [0.008]*** 0.013 [0.010] 0.026 [0.009]*** 0.011 [0.011] 0.942 [0.008]*** X 78% 7052

0.008 [0.002]*** 0.010 [0.004]*** 0.009 [0.005]* 0.024 [0.009]** 0.010 [0.011] 0.945 [0.008]*** X 78% 7051

0.007 [0.002]*** 0.011 [0.005]** 0.002 [0.006] 0.024 [0.009]** 0.010 [0.011] 0.946 [0.008]*** X 78% 7052

0.008 [0.002]*** 0.046 [0.007]*** 0.011 [0.009] 0.027 [0.009]*** 0.011 [0.011] 0.943 [0.008]*** X 78% 7052

Column (1) reports the baseline estimates (same as table 3). The different measures of dietary intake used are: Food intake (kilocalories/day - column 1); protein intake (grams/day - column 2); fat intake (grams/day - column 3); carbohydrate intake (grams/day - column 4); dietary cholesterol intake (mg/day - column 5); total sugar intake (grams/day - column 6); starch intake (grams/day column 7); non-starch polysaccharide (fibre) intake (grams/day - column 8); factor score of all the dietary measures (column 9); * significant at 10%; ** significant at 5%; *** significant at 1%. Standard error clustered at the individual level in brackets. Dependent variable: log BMI (kg/m2 ); Covariates: gender; parity; age of child at clinic date; log mom BMI during pregnancy; mom and dad education and SES; mother age at pregnancy; dummy for single mother; reliable dietary report; time dummy; seasonal dummies; late respondent; birth weight.

Back Pietro Biroli

93

Measurement of Exercise Table: Different Measures of Physical Activity - FTO gene

Risky FTO Gene

βg

log(Food Intake)

αf

G X Food Intake

αg×f

log(Exercise)

αe

G X Exercise

αg×e

log(BMI)t−1

(1 − δ)

Covariates R2 Observations

(1) Sedentary min

MVPA

(2)

(3) Counts per min

Factor Score

(4)

0.010 [0.002]*** 0.067 [0.009]*** 0.025 [0.011]** 0.027 [0.009]*** 0.012 [0.011] 0.939 [0.008]*** X 78% 7052

0.009 [0.002]*** 0.068 [0.009]*** 0.021 [0.011]* -0.011 [0.002]*** -0.001 [0.002] 0.934 [0.008]*** X 79% 7043

0.009 [0.003]*** 0.069 [0.009]*** 0.024 [0.011]** -0.028 [0.005]*** -0.009 [0.006] 0.936 [0.008]*** X 79% 7052

0.009 [0.002]*** 0.069 [0.009]*** 0.023 [0.011]** -0.008 [0.002]*** -0.002 [0.002] 0.936 [0.008]*** X 79% 7043

Column (1) reports the baseline estimates (same as table 3). The different measures of exercise used are: sedentary minutes (column 1); moderate to vigorous physical activity (MVPA - column 2); counts per minute (column 3) factor score of all the exercise measures (column 4); * significant at 10%; ** significant at 5%; *** significant at 1%. Standard error clustered at the individual level in brackets. Dependent variable: log BMI (kg/m2 ); Covariates: gender; parity; age of child at clinic date; log mom BMI during pregnancy; mom and dad education and SES; mother age at pregnancy; dummy for single mother; reliable dietary report; time dummy; seasonal dummies; late respondent; birth weight.

Back Pietro Biroli

94

Quantile Regression

Back Pietro Biroli

95

Quantile Regression

Back Pietro Biroli

96

According to Mendel’s laws of independent assortment, we expect a bell-shaped distribution

back

Pietro Biroli

97

References I

Altonji, J. G., Elder, T. E., and Taber, C. R. (2008). Using Selection on Observed Variables to Assess Bias from Unobservables when Evaluating Swan-Ganz Catheterization. American Economic Review: Papers & Proceedings, 98(2):345–350. Andreasen, C. H., Stender-petersen, K. L., Mogensen, M. S., Torekov, S. S., Wegner, L., Andersen, G., Nielsen, A. L., Albrechtsen, A., Borch-johnsen, K., Rasmussen, S. S., Clausen, J. O., Sandbæk, A., Lauritzen, T., Hansen, L., Jorgensen, T., Pedersen, O., and Hansen, T. (2008). Low Physical Activity Accentuates the Effect of the FTO rs9939609 Polymorphism on Body Fat Accumulation. Diabetes, 57(January):95–101. Belsky, D. W., Moffitt, T. E., Sugden, K., Williams, B. S., Houts, R. M., McCarthy, J., and Caspi, A. (2013). Development and evaluation of a genetic risk score for obesity. Biodemography and social biology, 59(1):85–100. Benjamin, D. J., Caplin, A., Cesarini, D., Thom, K., and Turley, P. (2016). Smoking, Genes, and Health. Working Paper, pages 1–27. Benjamin, D. J., Chabris, C. F., Glaeser, E. L., Gudnason, V., Harris, T. B., Laibson, D. I., Launer, L. J., and Purcell, S. M. (2007). Genoeconomics. In Weinstein, M., Vaupel, J. W., and Wachter, K. W., editors, Biosocial Surveys, chapter 15, pages 304–335. The National Academies Press, Washington, D.C.

Pietro Biroli

98

References II Berndt, S. I., Gustafsson, S., M¨ agi, R., Ganna, A., Wheeler, E., Feitosa, M. F., Justice, A. E., Monda, K. L., Croteau-Chonka, D. C., Day, F. R., Esko, T., Fall, T., Ferreira, T., Gentilini, D., Jackson, A. U., Luan, J., Randall, J. C., Vedantam, S., Willer, C. J., Winkler, T. W., Wood, A. R., Workalemahu, T., Hu, Y.-J., Lee, S. H., Liang, L., Lin, D.-Y., Min, J. L., Neale, B. M., Thorleifsson, G., Yang, J., Albrecht, E., Amin, N., Bragg-Gresham, J. L., Cadby, G., den Heijer, M., Eklund, N., Fischer, K., Goel, A., Hottenga, J. J., Huffman, J. E., Jarick, I., Johansson, ˚ A., Johnson, T., Kanoni, S., Kleber, M. E., K¨ onig, I. R., Kristiansson, K., Kutalik, Z., Lamina, C., Lecoeur, C., Li, G., Mangino, M., McArdle, W. L., Medina-Gomez, C., M¨ uller-Nurasyid, M., Ngwa, J. S., Nolte, I. M., Paternoster, L., Pechlivanis, S., Perola, M., Peters, M. J., Preuss, M., Rose, L. M., Shi, J., Shungin, D., Smith, A. V., Strawbridge, R. J., Surakka, I., Teumer, A., Trip, M. D., Tyrer, J., Van Vliet-Ostaptchouk, J. V., Vandenput, L., Waite, L. L., Zhao, J. H., Absher, D., Asselbergs, F. W., Atalay, M., Attwood, A. P., Balmforth, A. J., Basart, H., Beilby, J., Bonnycastle, L. L., Brambilla, P., Bruinenberg, M., Campbell, H., Chasman, D. I., Chines, P. S., Collins, F. S., Connell, J. M., Cookson, W. O., de Faire, U., de Vegt, F., Dei, M., Dimitriou, M., Edkins, S., Estrada, K., Evans, D. M., Farrall, M., Ferrario, M. M., Ferri` eres, J., Franke, L., Frau, F., Gejman, P. V., Grallert, H., Gr¨ onberg, H., Gudnason, V., Hall, A. S., Hall, P., Hartikainen, A.-L., Hayward, C., Heard-Costa, N. L., Heath, A. C., Hebebrand, J., Homuth, G., Hu, F. B., Hunt, S. E., Hypp¨ onen, E., Iribarren, C., Jacobs, K. B., Jansson, J.-O., Jula, A. M., K¨ ah¨ onen, M., Kathiresan, S., Kee, F., Khaw, K.-T., Kivim¨ aki, M., Koenig, W., Kraja, A. T., Kumari, M., Kuulasmaa, K., Kuusisto, J., Laitinen, J. H., Lakka, T. A., Langenberg, C., Launer, L. J., Lind, L., Lindstr¨ om, J., Liu, J., Liuzzi, A., Lokki, M.-L., Lorentzon, M., Madden, P. A., Magnusson, P. K. E., Manunta, P., Marek, D., M¨ arz, W., Mateo Leach, I., McKnight, B., Medland, S. E., Mihailov, E., Milani, L., Montgomery, G. W., Mooser, V., M¨ uhleisen, T. W., Munroe, P. B., Musk, A. W., Narisu, N., Navis, G., Nicholson, G., Nohr, E. A., Ong, K. K., Oostra, B. A., Palmer, C. N. A., Palotie, A., Peden, J. F., Pedersen, N., Peters, A. H., Polasek, O., Pouta, A., Pramstaller, P. P., Prokopenko, I., P¨ utter, C., Radhakrishnan, A., Raitakari, O. T., Rendon, A., Rivadeneira, F., Rudan, I., Saaristo, T. E., Sambrook, J. G., Sanders, A. R., Sanna, S., Saramies, J., Schipf, S., Schreiber, S., Schunkert, H., Shin, S.-Y., Signorini, S., Sinisalo, J., Skrobek, B., Soranzo, N., Stanˇ c´ akov´ a, A., Stark, K., Stephens, J. C., Stirrups, K., Stolk, R. P., Stumvoll, M., Swift, A. J., Theodoraki, E. V., Thorand, B., Tregouet, D.-A., Tremoli, E., Van der Klauw, M. M., van Meurs, J. B. J., Vermeulen, S. H., Viikari, J., Virtamo, J., Vitart, V., Waeber, G., Wang, Z., Wid´ en, E., Wild, S. H., Willemsen, G., Winkelmann, B. R., Witteman, J. C. M., Wolffenbuttel, B. H. R., Wong, A., Wright, A. F., Zillikens, M. C., Amouyel, P., Boehm, B. O., Boerwinkle, E., Boomsma, D. I., Caulfield, M. J., Chanock, S. J., Cupples, L. A., Cusi, D., Dedoussis, G. V., Erdmann, J., Eriksson, J. G., Franks, P. W., Froguel, P., Gieger, C., Gyllensten, U., Hamsten, A., Harris, T. B., Hengstenberg, C., Hicks, A. A., Hingorani, A., Hinney, A., Hofman, A., Hovingh, G. K., Hveem, K., Illig, T., Jarvelin, M.-R., J¨ ockel, K.-H., Keinanen-Kiukaanniemi, S. M., Kiemeney, L. A., Kuh, D., Laakso, M., Lehtim¨ aki, T., Levinson, D. F., Martin, N. G., Metspalu, A., Morris, A. D., Nieminen, M. S., Njølstad, I., Ohlsson, C., Oldehinkel, A. J., Ouwehand, W. H., Palmer, L. J., Penninx, B. P., Power, C., Province, M. A., Psaty, B. M., Qi, L., Rauramaa, R., Ridker, P. M., Ripatti, S., Salomaa, V., Samani, N. J., Snieder, H., Sørensen, T. I. A., Spector, T. D., Stefansson, K., T¨ onjes, A., Tuomilehto, J., Uitterlinden, A. G., Uusitupa, M., van der Harst, P., Vollenweider, P., Wallaschofski, H., Wareham, N. J., Watkins, H., Wichmann, H.-E., Wilson, J. F., Abecasis, G. R., Assimes, T. L., Barroso, I., Boehnke, M., Borecki, I. B., Deloukas, P., Fox, C. S., Frayling, T. M., Groop, L. C., Haritunian, T., Heid, I. M., Hunter, D., Kaplan, R. C., Karpe, F., Moffatt, M. F., Mohlke, K. L., O’Connell, J. R., Pawitan, Y., Schadt, E. E., Schlessinger, D., Steinthorsdottir, V., Pietro Biroli

99

References III Strachan, D. P., Thorsteinsdottir, U., van Duijn, C. M., Visscher, P. M., Di Blasio, A. M., Hirschhorn, J. N., Lindgren, C. M., Morris, A. P., Meyre, D., Scherag, A., McCarthy, M. I., Speliotes, E. K., North, K. E., Loos, R. J. F., and Ingelsson, E. (2013). Genome-wide meta-analysis identifies 11 new loci for anthropometric traits and provides insights into genetic architecture. Nature genetics, 45(5):501–12. Caspi, A., McClay, J., Moffitt, T. E., Mill, J., Martin, J., Craig, I. W., Taylor, A., and Poulton, R. (2002). Role of genotype in the cycle of violence in maltreated children. Science, 297(5582):851–4. Cawley, J. (2010). The economics of childhood obesity. Health affairs, 29(3):364–71. Cawley, J., Han, E., and Norton, E. C. (2011). The validity of genes related to neurotransmitters as instrumental variables. Health Economics, 20(May):884–888. Cecil, J. E., Tavendale, R., Watt, P., Hetherington, M. M., and Palmer, C. N. A. (2008). An obesity-associated FTO gene variant and increased energy intake in children. The New England Journal of Medicine, 359:2558–2566. Cesarini, D., Dawes, C. T., Johannesson, M., Lichtenstein, P., and Wallace, B. (2009). Genetic Variation in Preferences for Giving and Risk Taking. Quarterly Journal of Economics, 124(2):809–842. Claussnitzer, M., Dankel, S. N., Kim, K.-H., Quon, G., Meuleman, W., Haugen, C., Glunk, V., Sousa, I. S., Beaudry, J. L., Puviindran, V., Abdennur, N. A., Liu, J., Svensson, P.-A., Hsu, Y.-H., Drucker, D. J., Mellgren, G., Hui, C.-C., Hauner, H., and Kellis, M. (2015). FTO Obesity Variant Circuitry and Adipocyte Browning in Humans. New England Journal of Medicine, 373(10):895–907.

Pietro Biroli

100

References IV Cutler, D. M., Glaeser, E. L., and Shapiro, J. M. (2003). Why Have Americans Become More Obese? Journal of Economic Perspectives, 17(3):93–118. Davey Smith, G. (2003). ’Mendelian randomization’: can genetic epidemiology contribute to understanding environmental determinants of disease? International Journal of Epidemiology, 32(1):1–22. Davey Smith, G. (2010). Mendelian Randomization for Strengthening Causal Inference in Observational Studies: Application to Gene x Environment Interactions. Perspectives on Psychological Science, 5(5):527–545. Fawcett, K. A. and Barroso, I. (2010). The genetics of obesity: FTO leads the way. Trends in genetics, 26(6):266–74. Fletcher, J. M. and Lehrer, S. F. (2009). The effects of adolescent health on educational outcomes: causal evidence using genetic lotteries between siblings. Forum for Health Economics & Policy, 12(2). Franks, P. W., Jablonski, K. A., Delahanty, L. M., McAteer, J. B., Kahn, S. E., Knowler, W. C., and Florez, J. C. (2008). Assessing gene-treatment interactions at the FTO and INSIG2 loci on obesity-related traits in the Diabetes Prevention Program. Diabetologia, 51(12):2214–23. Fredriksson, R., H¨ agglund, M., Olszewski, P. K., Stephansson, O., Jacobsson, J. A., Olszewska, A. M., Levine, A. S., Lindblom, J., and Schi¨ oth, H. B. (2008). The obesity gene, FTO, is of ancient origin, up-regulated during food deprivation and expressed in neurons of feeding-related nuclei of the brain. Endocrinology, 149(5):2062–71.

Pietro Biroli

101

References V Galama, T. and Kapteyn, A. (2011). Grossman’s missing health threshold. Journal of Health Economics, 30(5):1044–56. Galton, F. (1874). English men of science: Their nature and nurture. McMillan & Co, London. Grossman, M. (1972). On the concept of health capital and the demand for health. Journal of Political Economy, 80(2):223–255. Guo, G., Liu, H., Wang, L., Shen, H., and Hu, W. (2015). The Genome-Wide Influence on Human BMI Depends on Physical Activity, Life Course, and Historical Period. Demography, 52(5):1651–1670. Huang, T., Qi, Q., Li, Y., Hu, F. B., Bray, G. A., Sacks, F. M., Williamson, D. A., and Qi, L. (2014). FTO genotype, dietary protein, and change in appetite: the Preventing Overweight Using Novel Dietary Strategies trial. The American journal of clinical nutrition, 99(5):1126–30. Karra, E., O’Daly, O. G., Choudhury, A. I., Yousseif, A., Millership, S., Neary, M. T., Scott, W. R., Chandarana, K., Manning, S., Hess, M. E., Iwakura, H., Akamizu, T., Millet, Q., Gelegen, C., Drew, M. E., Rahman, S., Emmanuel, J. J., Williams, S. C. R., R¨ uther, U. U., Br¨ uning, J. C., Withers, D. J., Zelaya, F. O., and Batterham, R. L. (2013). A link between FTO, ghrelin, and impaired brain food-cue responsivity. The Journal of clinical investigation, 123(8):3539–51.

Pietro Biroli

102

References VI Kilpel¨ ainen, T. O., Qi, L., Brage, S., Sharp, S. J., Sonestedt, E., Demerath, E. W., Ahmad, T., Mora, S., Kaakinen, M. A., Sandholt, C. H., Holzapfel, C., Autenrieth, C. S., Hypp¨ onen, E., Cauchi, S., He, M., Kutalik, Z., Kumari, M., Stanˇ c´ akov´ a, A., Meidtner, K., Balkau, B., Tan, J. T., Mangino, M., Timpson, N. J., Song, Y., Zillikens, M. C., Jablonski, K. A., Garcia, M. E., Johansson, S., Bragg-Gresham, J. L., Wu, Y., van Vliet-Ostaptchouk, J. V., Onland-Moret, N. C., Zimmermann, E., Rivera, N. V., Tanaka, T., Stringham, H. M., Silbernagel, G., Kanoni, S., Feitosa, M. F., Snitker, S., Ruiz, J. R., Metter, J., Larrad, M. T. M., Atalay, M., Hakanen, M., Amin, N., Cavalcanti-Proenc¸a, C., Grøntved, A., Hallmans, G., Jansson, J.-O., Kuusisto, J., K¨ ah¨ onen, M., Lutsey, P. L., Nolan, J. J., Palla, L., Pedersen, O., P´ erusse, L., Renstr¨ om, F., Scott, R. A., Shungin, D., Sovio, U., Tammelin, T. H., R¨ onnemaa, T., Lakka, T. A., Uusitupa, M., Rios, M. S., Ferrucci, L., Bouchard, C., Meirhaeghe, A., Fu, M., Walker, M., Borecki, I. B., Dedoussis, G. V., Fritsche, A., Ohlsson, C., Boehnke, M., Bandinelli, S., van Duijn, C. M., Ebrahim, S., Lawlor, D. A., Gudnason, V., Harris, T. B., Sørensen, T. I. A., Mohlke, K. L., Hofman, A., Uitterlinden, A. G., Tuomilehto, J., Lehtim¨ aki, T., Raitakari, O. T., Isomaa, B., Njølstad, P. R., Florez, J. C., Liu, S., Ness, A., Spector, T. D., Tai, E. S., Froguel, P., Boeing, H., Laakso, M., Marmot, M. G., Bergmann, S., Power, C., Khaw, K.-T., Chasman, D., Ridker, P. M., Hansen, T., Monda, K. L., Illig, T., J¨ arvelin, M.-R., Wareham, N. J., Hu, F. B., Groop, L. C., Orho-Melander, M., Ekelund, U., Franks, P. W., and Loos, R. J. F. (2011). Physical activity attenuates the influence of FTO variants on obesity risk: a meta-analysis of 218,166 adults and 19,268 children. PLoS medicine, 8(11):e1001116. Kline, B. and Tobias, J. L. (2008). The wages of BMI: Bayesian analysis of a skewed treatment-response model with nonparametric endogeneity. Journal of Applied Econometrics, 23(6):767–793. Kohler, H.-P., Behrman, J. R., and Schnittker, J. (2011). Social Science Methods for Twins Data: Integrating Causality, Endowments, and Heritability. Biodemography and Social Biology, 57(1):88–141. Lakdawalla, D., Philipson, T. J., and Bhattacharya, J. (2005). Welfare-Enhancing Technological Change and the Growth of Obesity. American Economic Review, 95(2):253–257. Liu, H. and Guo, G. (2015). Lifetime Socioeconomic Status, Historical Context, and Genetic Inheritance in Shaping Body Mass in Middle and Late Adulthood. American Sociological Review, 80(4):705–737. Pietro Biroli

103

References VII Mattocks, C., Ness, A. R., Leary, S. D., Tilling, K., Blair, S. N., Shield, J., Deere, K., Saunders, J., Kirkby, J., Davey Smith, G., Wells, J. C., Wareham, N. J., Reilly, J. J., and Riddoch, C. J. (2008). Use of accelerometers in a large field-based study of children: protocols, design issues, and effects on precision. Journal of physical activity & health, 5 Suppl 1:S98–111. Mokdad, A. H., Marks, J. S., Stroup, D. F., and Gerberding, J. L. (2004). Actual causes of death in the United States, 2000. JAMA, 291(10):1238–45. Ng, M., Fleming, T., Robinson, M., Thomson, B., Graetz, N., Margono, C., Mullany, E. C., Biryukov, S., Abbafati, C., Abera, S. F., Abraham, J. P., Abu-Rmeileh, N. M. E., Achoki, T., AlBuhairan, F. S., Alemu, Z. A., Alfonso, R., Ali, M. K., Ali, R., Guzman, N. A., Ammar, W., Anwari, P., Banerjee, A., Barquera, S., Basu, S., Bennett, D. A., Bhutta, Z., Blore, J., Cabral, N., Nonato, I. C., Chang, J.-C., Chowdhury, R., Courville, K. J., Criqui, M. H., Cundiff, D. K., Dabhadkar, K. C., Dandona, L., Davis, A., Dayama, A., Dharmaratne, S. D., Ding, E. L., Durrani, A. M., Esteghamati, A., Farzadfar, F., Fay, D. F. J., Feigin, V. L., Flaxman, A., Forouzanfar, M. H., Goto, A., Green, M. A., Gupta, R., Hafezi-Nejad, N., Hankey, G. J., Harewood, H. C., Havmoeller, R., Hay, S., Hernandez, L., Husseini, A., Idrisov, B. T., Ikeda, N., Islami, F., Jahangir, E., Jassal, S. K., Jee, S. H., Jeffreys, M., Jonas, J. B., Kabagambe, E. K., Khalifa, S. E. A. H., Kengne, A. P., Khader, Y. S., Khang, Y.-H., Kim, D., Kimokoti, R. W., Kinge, J. M., Kokubo, Y., Kosen, S., Kwan, G., Lai, T., Leinsalu, M., Li, Y., Liang, X., Liu, S., Logroscino, G., Lotufo, P. A., Lu, Y., Ma, J., Mainoo, N. K., Mensah, G. A., Merriman, T. R., Mokdad, A. H., Moschandreas, J., Naghavi, M., Naheed, A., Nand, D., Narayan, K. M. V., Nelson, E. L., Neuhouser, M. L., Nisar, M. I., Ohkubo, T., Oti, S. O., Pedroza, A., Prabhakaran, D., Roy, N., Sampson, U., Seo, H., Sepanlou, S. G., Shibuya, K., Shiri, R., Shiue, I., Singh, G. M., Singh, J. A., Skirbekk, V., Stapelberg, N. J. C., Sturua, L., Sykes, B. L., Tobias, M., Tran, B. X., Trasande, L., Toyoshima, H., van de Vijver, S., Vasankari, T. J., Veerman, J. L., Velasquez-Melendez, G., Vlassov, V. V., Vollset, S. E., Vos, T., Wang, C., Wang, S. X., Weiderpass, E., Werdecker, A., Wright, J. L., Yang, Y. C., Yatsuya, H., Yoon, J., Yoon, S.-J., Zhao, Y., Zhou, M., Zhu, S., Lopez, A. D., Murray, C. J. L., and Gakidou, E. (2014). Global, regional, and national prevalence of overweight and obesity in children and adults during 1980-2013: a systematic analysis for the Global Burden of Disease Study 2013. Lancet, 384(9945):766–781. Noel, S. E., Mattocks, C., Emmett, P. M., Riddoch, C. J., Ness, A. R., and Newby, P. K. (2010). Use of accelerometer data in prediction equations for capturing implausible dietary intakes in adolescents. American Journal of Clinical Nutrition, 92(6):1436–1445.

Pietro Biroli

104

References VIII Olszewski, P. K., Fredriksson, R., Olszewska, A. M., Stephansson, O., Alsi¨ o, J., Radomska, K. J., Levine, A. S., and Schi¨ oth, H. B. (2009). Hypothalamic FTO is associated with the regulation of energy intake not feeding reward. BMC neuroscience, 10:129. Rietveld, C. A., Esko, T., Davies, G., Pers, T. H., Turley, P., Benyamin, B., Chabris, C. F., Emilsson, V., Johnson, A. D., Lee, J. J., de Leeuw, C., Marioni, R. E., Medland, S. E., Miller, M. B., Rostapshova, O., van der Lee, S. J., Vinkhuyzen, A. A. E., Amin, N., Conley, D. C., Derringer, J., van Duijn, C. M., Fehrmann, R. S. N., Franke, L., Glaeser, E. L., Hansell, N. K., Hayward, C., Iacono, W. G., Ibrahim-Verbaas, C., Jaddoe, V. W. V., Karjalainen, J. M., Laibson, D. I., Lichtenstein, P., Liewald, D. C., Magnusson, P. K. E., Martin, N. G., McGue, M., McMahon, G., Pedersen, N. L., Pinker, S., Porteous, D. J., Posthuma, D., Rivadeneira, F., Smith, B. H., Starr, J. M., Tiemeier, H., Timpson, N. J., Trzaskowski, M., Uitterlinden, A. G., Verhulst, F. C., Ward, M. E., Wright, M. J., Davey Smith, G., Deary, I. J., Johannesson, M., Plomin, R., Visscher, P. M., Benjamin, D. J., Cesarini, D., and Koellinger, P. D. (2014). Common genetic variants associated with cognitive performance identified using the proxy-phenotype method. Proceedings of the National Academy of Sciences, 111(38):13790–13794. Rosenquist, J. N., Lehrer, S. F., O’Malley, A. J., Zaslavsky, A. M., Smoller, J. W., and Christakis, N. A. (2015). Cohort of birth modifies the association between FTO genotype and BMI. Proceedings of the National Academy of Sciences, 112(2):354–9. Sandholt, C. H., Hansen, T., and Pedersen, O. (2012). Beyond the fourth wave of genome-wide obesity association studies. Nutrition and Diabetes, 2(7):e37. Scholz, J. K. and Seshadri, A. (2013). Health and Wealth In a Lifecycle Model. Working Paper, (July). Smemo, S., Tena, J. J., Kim, K.-H., Gamazon, E. R., Sakabe, N. J., G´ omez-Mar´ın, C., Aneas, I., Credidio, F. L., Sobreira, D. R., Wasserman, N. F., Lee, J. H., Puviindran, V., Tam, D., Shen, M., Son, J. E., Vakili, N. A., Sung, H.-K., Naranjo, S., Acemel, R. D., Manzanares, M., Nagy, A., Cox, N. J., Hui, C.-C., Gomez-Skarmeta, J. L., and Nobrega, M. A. (2014). Obesity-associated variants within FTO form long-range functional connections with IRX3. Nature. Pietro Biroli

105

References IX Speakman, J. R., Rance, K. A., and Johnstone, A. M. (2008). Polymorphisms of the FTO gene are associated with variation in energy intake, but not energy expenditure. Obesity, 16(8):1961–5. Speliotes, E. K., Willer, C. J., Berndt, S. I., Monda, K. L., Thorleifsson, G., Jackson, A. U., Lango Allen, H., Lindgren, C. M., Luan, J., M¨ agi, R., Randall, J. C., Vedantam, S., Winkler, T. W., Qi, L., Workalemahu, T., Heid, I. M., Steinthorsdottir, V., Stringham, H. M., Weedon, M. N., Wheeler, E., Wood, A. R., Stefansson, K., North, K. E., McCarthy, M. I., Hirschhorn, J. N., Ingelsson, E., and Loos, R. J. F. (2010). Association analyses of 249,796 individuals reveal 18 new loci associated with body mass index. Nature genetics, 42(11):937–48. Taubman, P. (1976). Earnings, education, genetics, and environment. Journal of Human Resources, 11(4):447–461. Timpson, N. J., Emmett, P. M., Frayling, T. M., Rogers, I. S., Hattersley, A. T., McCarthy, M. I., and Davey Smith, G. (2008). The fat mass- and obesity-associated locus and dietary intake in children. The American Journal of Clinical Nutrition, 88(4):971–8. Vimaleswaran, K. S. and Loos, R. J. F. (2010). Progress in the genetics of common obesity and type 2 diabetes. Expert reviews in molecular medicine, 12(February):e7. von Hinke Kessler Scholder, S., Davey Smith, G., Lawlor, D. A., Propper, C., and Windmeijer, F. (2013). Child height, health and human capital: Evidence using genetic markers. European Economic Review, 57:1–22. Wardle, J., Carnell, S., Haworth, C. M. A., Farooqi, I. S., O’Rahilly, S., and Plomin, R. (2008). Obesity associated genetic variation in FTO is associated with diminished satiety. The Journal of clinical endocrinology and metabolism, 93(9):3640–3.

Pietro Biroli

106

References X

Zhang, X., Qi, Q., Zhang, C., Smith, S. R., Hu, F. B., Sacks, F. M., Bray, G. A., and Qi, L. (2012). FTO genotype and 2-year change in body composition and fat distribution in response to weight-loss diets: the POUNDS LOST Trial. Diabetes, 61(11):3005–11.

Pietro Biroli

107