University:
Boston University
Course:
CAS EC 203 | Empirical Economics 1
Academic year:

2023
Views:

449

Pages:

24
Author:

jaac19736txf

Figure 1: Distribution of BMI by sector of occupation Panel A: Urban working women aged 18 – 60 Panel B: Urban working men aged 18 – 60 Notes: Kernel Density Estimates of BMI presented. Sample in Panel A restricted to urban working women aged 18 – 60 in IHDS1 and IHDS2. In Panel B, sample is restricted to urban working men aged 18 – 60 in IHDS2. See Table A1 for categorization of occupations. Kolmogorov-Smirnov (K-S) test p-value for equality of distribution is 0.00, for both Definitions 1 and 2, for both women and men. Figure 2: Distribution of BMI by activity levels (defined by occupation) Panel A: Urban working women aged 18 – 60 Panel B: Urban working men aged 18 – 60 Notes: Kernel density estimates of BMI presented. Sample in Panel A restricted to urban working women aged 18 – 60 in IHDS1 and IHDS2. In Panel B, sample is restricted to urban working men aged 18 – 60 in IHDS2. See Table A1 for categorization of occupations. In Panel A, for Definition 1, K-S test p-values for equality of distributions are 0.00, 0.00 and 0.01 for the low and high, low and medium, and medium and high activity occupations, respectively. For Definition 2, the corresponding K-S test p-values are 0.00, 0.00 and 0.00. In Panel B, for Definition 1, K-S test p-values for equality of distributions are 0.00, 0.00 and 0.01 for the low and high, low and medium, and medium and high activity occupations, respectively. For Definition 2, the corresponding K-S test p-values are 0.00, 0.00 and 0.00. Figure 3: Lowess plots of BMI on intensity of activity Panel A: Urban working women aged 18 – 60 Panel B: Urban working men aged 18 – 60 Notes: Lowess regression results. Sample in Panel A restricted to urban working women aged 18 – 60 in IHDS1 and IHDS2. In Panel B, sample is restricted to urban working men aged 18 – 60 in IHDS2. See Table A2 for the MET values associated with each occupation. Figure 4: Distribution of BMI by intensity of activity (defined by MET categories) Panel A: Urban working women aged 18 – 60 .08 .06 0 .02 .04 .06 .04 0 .02 Density .08 .1 Definition 2 .1 Definition 1 10 20 30 BMI 40 Light Activity 50 10 20 Moderate Activity 30 BMI 40 50 Vigorous Activity Panel B: Urban working men aged 18 – 60 .1 0 .05 0 .05 Density .1 .15 Definition 2 .15 Definition 1 10 20 30 BMI Light Activity 40 50 10 Moderate Activity 20 30 BMI 40 50 Vigorous Activity Notes: Kernel density estimates of BMI presented. Sample in Panel A restricted to urban working women aged 18 – 60 in IHDS1 and IHDS2. In Panel B, sample is restricted to urban working men aged 18 – 60 in IHDS2. See Table A2 for the MET values associated with each occupation. In Panel A, for Definition 1, K-S test p-values for equality of distributions are 0.00, 0.00 and 0.15 for the light and heavy, light and moderate, and moderate and vigorous intensity occupations, respectively. For Definition 2, the corresponding K-S test p-values are 0.00, 0.00 and 0.03. In Panel B, for Definition 1, K-S test p-values for equality of distributions are 0.00, 0.00 and 0.65 for the light and vigorous, light and moderate, and moderate and vigorous intensity occupations, respectively. For Definition 2, the corresponding K-S test p-values are 0.00, 0.00 and 0.57. Figure 5: Distribution of BMI by work status Panel A: Urban working women aged 18 – 60 Definition 1 Definition 2 1 .8 .8 Cumulative Probability 1 .6 .4 .6 .4 .2 .2 0 0 10 20 30 BMI 40 50 10 Not W orking 20 30 BMI 40 50 40 50 W orking Panel B: Urban working men aged 18 – 60 Definition 1 Definition 2 1 .8 .8 Cumulative Probability 1 .6 .4 .6 .4 .2 .2 0 0 10 20 30 BMI 40 50 Not W orking 10 20 30 BMI W orking Notes: CDF of BMI presented. Sample in Panel A restricted to all urban women aged 18 – 60 in IHDS1 and IHDS2. In Panel B, sample is restricted to all urban men aged 18 – 60 in IHDS2. Table 1: Distribution of urban women and men by working status, occupational groups and physical strenuousness of work. Women Aged 18 – 60 Men Aged 18 – 60 Married Women Married Men Aged Aged 18–60 18–60 Definition Definition Definition Definition Definition Definition Definition Definition 1 2 1 2 1 2 1 2 (1) (2) (3) (4) (5) (6) (7) (8) Working Status Working 17.30 21.01 72.04 76.85 84.66 19.08 83.19 87.54 Non-working 82.70 78.99 27.96 23.15 15.34 80.92 16.81 12.46 Occupation Category (Conditional on Working) Blue Collar 59.43 63.78 61.69 62.86 58.17 63.39 61.86 62.92 occupation White Collar 40.57 36.22 38.31 37.14 41.83 36.61 38.14 37.08 occupation Activity Level of Work (Conditional on Working) Low activity job 40.57 36.22 38.31 37.14 41.83 36.61 38.14 37.08 Medium activity job 23.85 20.97 20.30 20.03 21.25 18.32 20.37 20.01 High activity job 35.58 42.81 41.40 42.83 36.92 45.07 41.49 42.91 BMI Not Working 23.60 23.65 22.72 22.76 23.77 23.83 24.19 24.47 Blue Collar 22.75 22.64 23.03 22.98 22.72 22.59 23.25 23.26 occupation White Collar 24.26 24.21 24.20 24.16 24.50 24.43 24.45 24.44 occupation Low activity job 24.26 24.21 24.20 24.16 24.50 24.43 24.44 24.45 Medium activity job 22.92 22.98 23.46 23.39 22.86 22.89 23.74 23.69 High activity job 22.64 22.47 22.81 22.79 22.64 22.46 23.03 23.05 Intensity of Activity – MET Activity: Light 69.77 63.82 71.68 70.32 71.56 64.14 71.46 70.14 Activity: Moderate 24.97 27.90 25.79 26.75 23.21 27.16 25.78 26.66 Activity: Vigorous 5.26 8.28 2.53 2.93 5.23 8.70 2.76 3.20 BMI Activity: Light 23.73 23.68 23.75 23.70 23.87 23.83 24.00 24.00 Activity: Moderate 22.60 22.50 22.79 22.76 22.55 22.40 22.99 23.01 Activity: Vigorous 22.15 21.88 22.81 22.66 21.99 21.77 22.87 22.82 Notes: The sample in columns 1 and 2 includes urban women aged 18 – 60 in IHDS1 and IHDS2. The sample in columns 3 and 4 includes urban men aged 18 – 60 in IHDS2. In columns 5–8, the sample is restricted to those that are married. Table reports proportions except in the case of BMI where actual levels are reported. Table 2: OLS regression of BMI in IHDS1 on withdrawal from the labour market in IHDS2 Definition 1 Definition 2 (1) (2) BMI in 2004-05 0.003 0.001 (0.005) (0.004) Constant 1.160** 0.476* (0.420) (0.258) Sample size 884 1,163 Notes: Sample restricted to urban working women aged 18 and above at the time of survey in IHDS1, but aged less than 60 at the time of survey in IHDS2. The dependent variable takes value 1 if a woman stopped working in IHDS2, and takes value 0 if she continue to work in IHDS2. The regressions include IHDS1 individual (age, age square, years of education, marital status, whether or not the individual consumes tobacco, number of children, the average number of hours spent watching television) and household level (dummies for wealth quartiles, whether or not the household has domestic help, whether the household owns a car or a motor cycle, household religions, the share of total expenditure on eating outside) controls, and a set of state dummies. Standard errors clustered at the state level are in parenthesis. Significance: * p<0.10, ** p<0.05, *** p<0.01. Table 3: Regression of BMI on labour market sector of work and intensity. Panel A: Urban Definition 1 Working Women (1) (2) (3) (4) (5) Aged 18 – 60 Blue Collar -0.439** -0.413* (0.193) (0.202) Medium Activity -0.287 (0.291) High Activity -0.537*** (0.160) Physical intensity (in -0.261*** MET) (0.060) Moderate Intensity -0.373 (0.225) Vigorous Intensity -0.796*** (0.161) Constant 13.926*** 13.935*** 14.357*** 13.679*** 13.265*** (0.982) (0.998) (1.018) (0.904) (1.104) High – Medium -0.250 Activity (0.216) Vigorous– Moderate -0.423 Intensity (0.302) Sample size 3,743 3,743 3,743 3,743 4,574 Panel B:Urban Working Men Aged 18 – 60 Blue Collar -0.334*** -0.335*** (0.117) (0.111) Medium Activity -0.066 (0.122) High Activity -0.496*** (0.138) Physical intensity -0.116 (in MET) (0.072) Moderate Intensity -0.279* (0.144) Vigorous Intensity 0.221 (0.326) Constant 15.785*** 15.759*** 15.865*** 15.588*** 15.623*** (0.861) (0.863) (0.920) (0.834) (0.861) High – Medium -0.430*** Activity (0.136) Vigorous – Moderate 0.500 Intensity (0.300) Sample size 5,165 5,165 5,165 5,165 5,491 (6) Definition 2 (7) (8) -0.133 (0.271) -0.553*** (0.190) -0.250*** (0.061) 13.321*** (1.115) -0.420** 13.841*** (1.102) -0.439* (0.243) -0.762*** (0.120) 13.193*** (1.022) (0.188) -0.323 4,574 4,574 (0.210) 4,574 -0.077 (0.128) -0.483*** (0.129) -0.119* (0.066) 15.609*** (0.858) -0.405*** 15.714*** (0.906) -0.275** (0.122) 0.057 (0.278) 15.419*** (0.821) 5,491 0.332 (0.247) 5,491 (0.140) 5,491 Notes: Dependent variable in the regressions is BMI. OLS regressions presented. In Panel A, sample restricted to 18-60-year-old working urban women at the time of survey in IHDS1 and IHDS2. In Panel B, sample restricted to 18-60-year-old working urban men in IHDS2. The regressions include individual (age, age square, years of education, marital status, whether or not the individual consumes tobacco, number of children, the average number of hours spent watching television) and household level (dummies for wealth quartiles, whether or not the household has domestic help, whether the household owns a car or a motor cycle, household religions, the share of total expenditure on eating outside) controls, and a set of state dummies. The regressions in Panel A also include an IHDS2 year dummy. Standard errors clustered at the state level are in parenthesis. Significance: * p<0.10, ** p<0.05, *** p<0.01 Table 4: Regression of BMI on labour market sector of work and intensity (full sample). Definition 1 Definition 2 (1) (2) (3) (4) Panel A: All Urban Women Aged 18 – 60 White Collar -0.222 -0.264 (0.182) (0.165) Blue Collar -0.565*** -0.718*** (0.115) (0.124) Low Activity -0.222 -0.263 (0.182) (0.165) Medium Activity -0.396* -0.400** (0.223) (0.192) High Activity -0.677*** -0.873*** (0.137) (0.148) Constant 11.121*** 11.123*** 11.116*** 11.131*** (0.690) (0.692) (0.693) (0.697) Blue – White -0.343** -0.454*** (0.157) (0.136) Medium – Low Activity -0.174 -0.137 (0.273) (0.234) High – Low Activity -0.455*** -0.610*** (0.156) (0.145) High – Medium Activity -0.281 -0.473* (0.271) (0.236) Sample Size 22,584 22,584 22,584 22,584 Panel B: All Urban Men Aged 18 – 60 White Collar 0.098 0.046 (0.190) (0.167) Blue Collar -0.228** -0.299*** (0.105) (0.091) Low Activity 0.101 0.047 (0.190) (0.168) Medium Activity 0.052 -0.031 (0.139) (0.136) High Activity -0.385*** -0.448*** (0.120) (0.101) Constant 14.330*** 14.323*** 14.338*** 14.337*** (0.804) (0.807) (0.777) (0.778) Blue – White -0.326** -0.345*** (0.120) (0.111) Medium – Low Activity -0.049 -0.078 (0.135) (0.141) High – Low Activity -0.486*** -0.494*** (0.137) (0.122) High – Medium Activity -0.438*** -0.416*** (0.140) (0.141) Sample Size 7,193 7,193 7,193 7,193 Notes: Dependent variable in regressions is BMI. OLS regression results presented. In Panel A, sample restricted to 1860-year-old urban women at the time of survey in IHDS1 and IHDS2. In Panel B, sample restricted to 18-60-year-old urban men in IHDS2. The regressions include individual (age, age square, years of education, marital status, whether or not the individual consumes tobacco, number of children, the average number of hours spent watching television) and household level (dummies for wealth quartiles, whether or not the household has domestic help, whether the household owns a car or a motor cycle, household religions, the share of total expenditure on eating outside) controls, and a set of state dummies. The regressions in Panel A also include an IHDS2 year dummy. Standard errors clustered at the state level are in parenthesis. Significance: * p<0.10, ** p<0.05, *** p<0.01. Table 5: BMI on labour market sector of work. Married sample Definition 1 OLS IV (1) (2) Panel A: Urban Working Women Aged 18 – 60 Blue collar -0.553** -1.008** (0.212) (0.429) Constant 13.920*** (1.531) First Stage F-statistic 144.10 [0.000] 0.399 Hansen’s J χ2 statistic [0.891] Sample size 2,353 2,353 Panel B: Urban Working Men Aged 18 – 60 Blue collar -0.651** -1.178*** (0.247) (0.391) Constant 18.322*** (2.755) First Stage F-statistic 59.10 [0.000] 1.339 Hansen’s J χ2 statistic [0.512] Sample size 802 802 Definition 2 OLS IV (3) (4) -0.499** (0.206) 13.779*** (1.603) 3,076 -0.623*** (0.188) 17.656*** (3.422) 1,025 -0.825* (0.443) 195.57 [0.000] 5.247 [0.073] 3,076 -1.307*** (0.472) 77.20 [0.000] 1.199 [0.549] 1,025 Notes: Dependent variable is BMI. OLS and IV regression results presented. In Panel A, sample restricted to 18-60year-old working urban married women at the time of survey in IHDS1 and IHDS2. In Panel B, sample restricted to 18-60-year-old working urban married men in IHDS2. The regressions include individual (age, age square, years of education, marital status, whether or not the individual consumes tobacco, number of children, the average number of hours spent watching television) and household level (dummies for wealth quartiles, whether or not the household has domestic help, whether the household owns a car or a motor cycle, household religions, the share of total expenditure on eating outside) controls, and a set of state dummies. The regressions in Panel A also include an IHDS2 year dummy. English speaking ability and activity status of spouse (working spouses) are used as instruments. Standard errors clustered at the state level are in parenthesis. p-values are reported in square brackets. Significance: * p<0.10, ** p<0.05, *** p<0.01. Table 6: Regression of BMI on labour market intensity. Married sample Definition 1 Definition 2 OLS IV OLS IV (1) (2) (3) (4) Panel A: Urban Working Women Aged 18 – 60 MET -0.174*** -0.514** -0.192*** -0.636*** (0.059) (0.223) (0.064) (0.185) Constant 14.038*** 14.147*** (1.504) (1.678) First Stage F-statistic 50.13 46.05 [0.000] [0.000] 1.429 2.124 Hansen’s J χ2 statistic [0.489] [0.346] Sample size 2,353 2,353 3,076 3,076 Panel B: Urban Working Men Aged 18 – 60 MET -0.010 -0.651** 0.029 -0.632** (0.088) (0.276) (0.072) (0.320) Constant 17.911*** 16.856*** (2.805) (3.571) First Stage F-statistic 30.31 46.09 [0.000] [0.000] 1.541 1.886 Hansen’s J χ2 statistic [0.462] [0.389] Sample size 802 802 1,025 1,025 Notes: Dependent variable is BMI. OLS and IV regression results presented. In Panel A, sample restricted to 18-60year-old working urban married women at the time of survey in IHDS1 and IHDS2. In Panel B, sample restricted to 18-60-year-old working urban married men in IHDS2. The regressions include individual (age, age square, years of education, marital status, whether or not the individual consumes tobacco, number of children, the average number of hours spent watching television) and household level (dummies for wealth quartiles, whether or not the household has domestic help, whether the household owns a car or a motor cycle, household religions, the share of total expenditure on eating outside) controls, and a set of state dummies. The regressions in Panel A also include an IHDS2 year dummy. English speaking ability and activity status of spouse are used as instruments. Standard errors clustered at the state level are in parenthesis. p-values are reported in square brackets. Significance: * p<0.10, ** p<0.05, *** p<0.0 APPENDIX A Figure A1: Average MET values by sector of occupation and activity levels Panel A: Average MET values by sector of occupation Panel B: Average MET values by activity levels Notes: See Table A1 for the categorization of occupations and Table A2 for the corresponding MET values. Table A1: Categorization into type of occupation and physical activity level Categorization Occupational Groups Type of Occupation White collar jobs (non-manual jobs) Professional, technical, and related workers, administrative, executive, and managerial workers, clerical and related workers* Blue collar jobs (manual jobs) Sales workers, Service workers, workers in transport and communications, Farmers, fishermen, hunters, loggers and related workers, Production and related workers* Physical activity level Low (same as white collar jobs) Professional, technical, and related workers, administrative, executive, and managerial workers, clerical and related workers** Medium Sales workers, service workers and workers in transport and communications** High Farmers, fishermen, hunters, loggers and related workers, production, and related workers** Notes: *Occupations coded 00 – 36, 39, 40, 41, 42, 44 and 45 as per NCO 1968 were categorized as white-collar jobs. Occupations coded as 37, 38, 49, 43 and 50 – 99, as per NCO 1968 were categorized as blue-collar jobs. **Occupations coded 00 – 36, 39, 40, 41, 42, 44, 45 as per NCO 1968 were categorized as low activity jobs. This is same as white collar jobs described above. Occupations coded as 37, 38, 43, 49, 86, 98 and 50 – 59 as per NCO 1968 were categorized medium activity jobs. Occupations coded 60 – 85, 87 – 97 and 99 as per NCO 1968 were categorized as high activity jobs Table A2: MET values of occupation. Two- digit Occupations occupation code 00 Physical Scientists 01 Physical Science Technicians 02 Architects, Engineers, Technologists and Surveyors 03 Engineering Technicians 04 Aircraft and Ships Officers 05 Life Scientists 06 Life Science Technicians 07 Physicians and Surgeons (Allopathic Dental and Veterinary Surgeons) 08 Nursing and other Medical and Health Technicians 09 Scientific, Medical and Technical Persons, Other 10 Mathematicians, Statisticians and Related Workers 11 Economists and Related Workers 12 Accountants, Auditors and Related Workers 13 Social Scientists and Related Workers 14 Jurists 15 Teachers 16 Poets, Authors, Journalists and Related Workers 17 Sculptors, Painters, Photographers and Related Creative Artists 18 Composers and Performing Artists 19 Professional Workers, NEC 20 Elected and Legislative Officials 21 Administrative and Executive Officials Government and Local Bodies 22 Working Proprietors, Directors and Managers, Wholesale and Retail Trade 23 Directors and Managers, Financial Institutions 24 Working Proprietors, Directors and Managers Mining, Construction, Manufacturing and Related Concerns 25 Working Proprietors, Directors, Managers and Related Executives, Transport, Storage and Communication 26 Working Proprietors, Directors and Managers, Other Service 29 Administrative, Executive and Managerial Workers, NEC 30 Clerical and Other Supervisors 31 Village Officials 32 Stenographers, Typists and Card and Tape Punching Operators Two- digit MET value Intensity of occupation Sector of Occupation Type of Activity 1.80 2.50 1.60 Light Light Light White Collar White Collar White Collar Low Activity Low Activity Low Activity 2.39 2.00 2.10 2.50 2.35 Light Light Light Light Light White Collar White Collar White Collar White Collar White Collar Low Activity Low Activity Low Activity Low Activity Low Activity 2.42 Light White Collar Low Activity 2.50 Light White Collar Low Activity 1.50 Light White Collar Low Activity 1.50 1.50 Light Light White Collar White Collar Low Activity Low Activity 1.94 Light White Collar Low Activity 1.50 2.50 1.50 Light Light Light White Collar White Collar White Collar Low Activity Low Activity Low Activity 3.00 Light White Collar Low Activity 2.33 2.20 1.50 2.00 Light Light Light Light White Collar White Collar White Collar White Collar Low Activity Low Activity Low Activity Low Activity 1.50 Light White Collar Low Activity 1.50 Light White Collar Low Activity 1.90 Light White Collar Low Activity 1.50 Light White Collar Low Activity 1.58 Light White Collar Low Activity 1.50 Light White Collar Low Activity 1.75 1.50 1.65 Light Light Light White Collar White Collar White Collar Low Activity Low Activity Low Activity 33 34 35 36 37 38 39 40 41 42 43 44 45 49 50 51 52 53 54 55 56 57 59 60 61 62 63 64 65 66 67 68 71 72 73 74 Book-keepers, Cashiers and Related Workers Computing Machine Operators Clerical and Related Workers, NEC Transport and Communication Supervisors Transport Conductors and Guards Mail Distributors and Related Workers Telephone and Telegraph Operators Merchants and Shopkeepers, Wholesale and Retail Trade Manufacturers, Agents Technical Salesmen and Commercial Travellers Salesmen, Shop Assistants and Related Workers Insurance, Real Estate, Securities and Business Service Salesmen and Auctioneers Money Lenders and Pawn Brokers Sales Workers, NEC Hotel and Restaurant Keepers House Keepers, Matron and Stewards (Domestic and Institutional) Cooks, Waiters, Bartenders and Related Worker (Domestic and Institutional) Maids and Other House Keeping Service Workers NEC Building Caretakers, Sweepers, Cleaners and Related Workers Launderers, Dry-cleaners and Pressers Hair Dressers, Barbers, Beauticians and Related Workers Protective Service Workers Service Workers, NEC Farm Plantation, Dairy and Other Managers and Supervisors Cultivators Farmers other than cultivators Agricultural Labourers Plantation Labourers and Related Workers Other Farm Workers Forestry Workers Hunters and Related Workers Fishermen and Related Workers Miners, Quarrymen, Well Drillers and Related Workers Metal Processors Wood Preparation Workers and Paper Makers Chemical Processors and Related Workers 1.75 Light White Collar Low Activity 1.50 1.72 Light Light White Collar White Collar Low Activity Low Activity 2.17 Light White Collar Low Activity 2.00 3.33 Light Moderate Blue Collar Blue Collar Medium Activity Medium Activity 1.50 Light White Collar Low Activity 1.50 Light White Collar Low Activity 1.75 2.00 Light Light White Collar White Collar Low Activity Low Activity 2.50 Light Blue Collar Medium Activity 1.83 Light White Collar Low Activity 1.50 2.50 2.00 3.50 Light Light Light Moderate White Collar Blue Collar Blue Collar Blue Collar Low Activity Medium Activity Medium Activity Medium Activity 2.75 Light Blue Collar Medium Activity 4.50 Moderate Blue Collar Medium Activity 3.25 Moderate Blue Collar Medium Activity 2.67 Light Blue Collar Medium Activity 2.50 Light Blue Collar Medium Activity 2.83 2.67 3.00 Light Light Moderate Blue Collar Blue Collar Blue Collar Medium Activity Medium Activity High Activity 4.8 4.28 6.00 6.00 Moderate Moderate Vigorous Vigorous Blue Collar Blue Collar Blue Collar Blue Collar High Activity High Activity High Activity High Activity 3.25 5.00 3.00 5.00 4.25 Moderate Moderate Moderate Moderate Moderate Blue Collar Blue Collar Blue Collar Blue Collar Blue Collar High Activity High Activity High Activity High Activity High Activity 2.98 2.92 Light Light Blue Collar Blue Collar High Activity High Activity 2.71 Light Blue Collar High Activity 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 Spinners, Weavers, Knitters, Dyers and Related Workers Tanners, Fellmongers and Pelt Dressers Food and Beverage Processors Tobacco Preparers and Tobacco Product Makers Tailors, Dress Makers, Sewers, Upholsterers and Related Workers Shoe makers and Leather Goods Makers Carpenters, Cabinet and Related Wood Workers Stone Cutters and Carvers Blacksmiths, Tool Makers and Machine Tool Operators Machinery Fitters, Machine Assemblers and Precision Instrument Electrical Fitters and Related Electrical and Electronic Workers Broadcasting Station and Sound Equipment Operators and Cinema Projectionists Plumbers, Welders, Sheet Metal and Structural Metal Preparers and Erectors Jewellery and Precious Metal Workers and Metal Engravers (Except Printing) Glass Formers, Potters and Related Workers Rubber and Plastic Product Makers Paper and Paper Board Products Makers Printing and Related Workers Painters Production and Related Workers, NEC Bricklayers and Other Constructions Workers Stationery Engines and Related Equipment Operators, Oilers and Greasors Material Handling and Related Equipment Operators, Loaders and Unloaders Transport Equipment Operators Labourers NEC 2.73 Light 3.25 Blue Collar High Activity Blue Collar High Activity 2.70 2.92 Light Light Blue Collar Blue Collar High Activity High Activity 2.29 Light Blue Collar High Activity 2.50 Light Blue Collar High Activity 3.43 Moderate Blue Collar High Activity 2.83 3.31 Light Moderate Blue Collar Blue Collar High Activity High Activity 3.08 Moderate Blue Collar High Activity 2.94 Light Blue Collar High Activity 2.00 Light Blue Collar Medium Activity 4.42 Moderate Blue Collar High Activity 1.50 Light Blue Collar High Activity 2.50 Light Blue Collar High Activity 3.10 Moderate Blue Collar High Activity 2.83 Light Blue Collar High Activity 1.85 2.88 2.90 Light Light Light Blue Collar Blue Collar Blue Collar High Activity High Activity High Activity 4.35 Moderate Blue Collar High Activity 3.10 Moderate Blue Collar High Activity 3.69 Moderate Blue Collar High Activity 2.73 5.00 Light Moderate Blue Collar Blue Collar Medium Activity High Activity Notes: MET values are matched using Tudor-Locke et al. (2011). For occupation codes 61, 62 and 66 we use https://sites.google.com/site/compendiumofphysicalactivities/Activity-Categories/occupation (accessed on August 30, 2017) Table A3: First stage IV regression results. Blue collar job Definition 1 Definition 2 (1) (2) Panel A: Urban Married Working Women Aged 18 – 60 Spoken English (Moderate or Fluent) Spouse employed in Medium Activity Spouse employed in High Activity First Stage F-statistic Hansen’s J χ2 statistic Sample size Panel B: Urban Married Working Men Aged 18 – 60 Spoken English (Moderate or Fluent) Spouse employed in Medium Activity Spouse employed in High Activity First Stage F-statistic Hansen’s J χ2 statistic Sample size MET Definition 1 (3) Definition 2 (4) -0.114*** (0.036) 0.319*** (0.022) 0.373*** (0.021) 144.10 [0.000] 0.399 [0.891] 2,353 -0.142*** (0.030) 0.318*** (0.022) 0.371*** (0.023) 195.57 [0.000] 5.247 [0.073] 3,076 0.114* (0.065) 0.325*** (0.044) 0.750*** (0.065) 50.13 [0.000] 1.429 [0.489] 2,353 0.047 (0.085) 0.342*** (0.051) 0.836*** (0.072) 46.05 [0.000] 2.124 [0.346] 3,076 -0.059** (0.029) 0.421*** (0.038) 0.465*** (0.035) 59.10 [0.000] 1.339 [0.512] 802 -0.068** (0.029) 0.417*** (0.031) 0.436*** (0.028) 77.20 [0.000] 1.199 [0.549] 1,025 -0.048 (0.074) 0.563*** (0.109) 0.913*** (0.096) 30.31 [0.000] 1.541 [0.462] 802 -0.084 (0.071) 0.460*** (0.106) 0.849*** (0.083) 46.09 [0.000] 1.886 [0.389] 1,025 Notes: Dependent variables are blue collar job (column 1 and 2) and MET (column 3 and 4) respectively. In Panel A, Sample restricted to 18-60-year-old married working urban women at the time of survey. In Panel B, sample restricted to 18-60-year-old married working urban men. The regressions include individual (age, age square, years of education, marital status, whether or not the individual consumes tobacco, number of children, the average number of hours spent watching television) and household level (dummies for wealth quartiles, whether or not the household has domestic help, whether the household owns a car or a motor cycle, household religions, the share of total expenditure on eating outside) controls, and a set state dummies. The regressions in Panel A also include an IHDS2 year dummy. English speaking ability and activity status of spouse are used as instruments. Standard errors clustered at the state level are in parenthesis. p-values are reported in square brackets. Significance: * p<0.10, ** p<0.05, *** p<0.00. Table A4: Instrument Validity. Regressions of instrument on BMI. Married Women Sample. Definition 1 Definition 2 (1) (2) (3) (4) (5) (6) (7) Spoken English 0.59 0.148 -0.113 -0.027 (Moderate or Fluent) (0.218) (0.218) (0.270) (0.269) Spouse employed in -0.082 -0.190 0.234 Medium Activity (0.327) (0.279) (0.251) Spouse employed in -0.233 -0.304 -0.302 High Activity (0.231) (0.192) (0.205) Blue collar -0.548** (0.209) -0.471* (0.271) MET Age Age square Years of education Smokes tobacco Hours watching Television Number of children Second wealth quartile Third wealth quartile Fourth wealth quartile Household hire servant Household possessing car or bike Muslim household Other religion household Share of expenditure on eating out Year (2011-12) Constant Sample size -0.512** (0.196) -0.143** (0.060) 0.384*** (0.066) -0.004*** (0.001) 0.039*** (0.014) -0.232 (0.648) 0.091 0.372*** (0.072) -0.004*** (0.001) 0.040** (0.016) -0.509 (0.393) 0.087** (8) 0.149 (0.219) -0.332* (0.169) -0.423* (0.234) 0.383*** (0.065) -0.004*** (0.001) 0.031** (0.013) -0.213 (0.630) 0.099 0.382*** (0.066) -0.004*** (0.001) 0.032** (0.014) -0.198 (0.652) 0.098 -0.175*** (0.060) 0.384*** (0.066) -0.004*** (0.001) 0.036*** (0.010) -0.258 (0.623) 0.089 0.368*** (0.071) -0.004*** (0.001) 0.032** (0.012) -0.474 (0.404) 0.085** -0.192*** (0.064) 0.371*** (0.072) -0.004*** (0.001) 0.040*** (0.014) -0.525 (0.395) 0.082* -0.154** (0.058) 0.368*** (0.071) -0.004*** (0.001) 0.035*** (0.011) -0.487 (0.405) 0.081* (0.063) -0.085 (0.085) 0.753*** (0.249) 1.305*** (0.257) 1.853*** (0.301) 0.847** (0.410) 0.044 (0.064) -0.085 (0.085) 0.731*** (0.255) 1.271*** (0.269) 1.809*** (0.308) 0.846** (0.399) 0.041 (0.065) -0.082 (0.085) 0.706** (0.258) 1.280*** (0.271) 1.861*** (0.331) 0.891** (0.394) 0.068 (0.065) -0.082 (0.085) 0.688** (0.259) 1.241*** (0.280) 1.802*** (0.330) 0.879** (0.384) 0.059 (0.039) -0.155* (0.079) 0.784*** (0.225) 1.345*** (0.178) 1.994*** (0.266) 0.777* (0.404) 0.209 (0.039) -0.158* (0.079) 0.751*** (0.227) 1.289*** (0.186) 1.916*** (0.285) 0.793* (0.389) 0.216 (0.041) -0.154* (0.078) 0.710*** (0.232) 1.267*** (0.190) 1.955*** (0.302) 0.828** (0.389) 0.230 (0.041) -0.157* (0.078) 0.690*** (0.229) 1.222*** (0.195) 1.881*** (0.311) 0.832** (0.377) 0.232 (0.252) 0.673 (0.427) 0.255 (0.242) 0.666 (0.430) 0.276 (0.247) 0.626 (0.420) 0.252 (0.236) 0.623 (0.420) 0.289 (0.202) 0.851** (0.311) 0.166 (0.198) 0.850** (0.316) 0.160 (0.202) 0.796** (0.306) 0.162 (0.198) 0.804** (0.309) 0.167 (0.338) -0.035*** (0.336) -0.035*** (0.353) -0.036*** (0.351) -0.035*** (0.385) -0.025*** (0.384) -0.026*** (0.394) -0.026*** (0.393) -0.026*** (0.012) 0.863*** (0.234) 13.917*** (1.530) 2,353 (0.011) 0.878*** (0.223) 14.034*** (1.539) 2,353 (0.012) 0.824*** (0.235) 14.047*** (1.509) 2,353 (0.012) 0.851*** (0.222) 14.175*** (1.505) 2,353 (0.009) 0.895*** (0.178) 13.781*** (1.600) 3,076 (0.009) 0.905*** (0.169) 13.959*** (1.614) 3,076 (0.009) 0.858*** (0.186) 14.144*** (1.685) 3,076 (0.009) 0.877*** (0.174) 14.257*** (1.674) 3,076 Notes: Dependent variable is BMI. OLS results are presented. Sample restricted to 18-60-year-old working urban married women at the time of survey in IHDS1 and IHDS2.The regressions also include a set of state dummies not reported here. English speaking ability and activity status of spouse are used as instruments. Standard errors clustered at the state level are in parenthesis. p-values are reported in square brackets. Significance: * p<0.10, ** p<0.05, *** p<0.01. Table A5: Regression of BMI on labour market sector of work and intensity. Married sample with working and non-working spouses. Definition 1 Definition 2 OLS IV OLS IV (1) (2) (3) (4) Panel A: Urban Working Men Aged 18 – 60 Blue collar -0.309*** -2.145*** -0.332*** -2.330*** (0.091) (0.661) (0.073) (0.770) Constant 17.253*** 17.158*** (0.961) (0.909) First Stage F-statistic 58.18 77.38 [0.000] [0.000] Hansen’s J Chi-squared statistic 7.776 7.013 [0.051] [0. 071] Sample size 4,264 4,264 4,463 4,463 Panel B: Urban Working Men Aged 18-60 MET -0.103 -0.900*** -0.096 -0.943*** (0.067) (0.303) (0.060) (0.314) Constant 17.323*** 17.188*** (1.026) (0.982) First Stage F-statistic 28.75 55.16 [0.000] [0.000] Hansen’s J Chi-squared statistic 8.068 7.161 [0.044] [0.067] Sample size 4,264 4,264 4,463 4,463 Notes: Dependent variable is BMI. OLS and IV regression results presented. In Panel A and Panel B, sample restricted to 18-60-year-old working urban married men in IHDS2. The regressions include individual (age, age square, years of education, marital status, whether or not the individual consumes tobacco, number of children, the average number of hours spent watching television) and household level (dummies for wealth quartiles, whether or not the household has domestic help, whether the household owns a car or a motor cycle, household religions, the share of total expenditure on eating outside) controls, and a set of state dummies. English speaking ability and activity status of spouse (including non-working spouses as well) are used as instruments. Standard errors clustered at the state level are in parenthesis. p-values are reported in square brackets. Significance: * p<0.10, ** p<0.05, *** p<0.01. APPENDIX B. FULL SET OF RESULTS Table B1: Regression of BMI on labour market sector of work and intensity. Urban Working Women aged 18–60. Blue Collar (1) -0.439** (0.193) Medium Activity Definition 1 (2) (3) Physical intensity (in MET) -0.261*** -0.250*** (0.060) (0.061) Married Smoke tobacco Hours watching Television Number of children Second wealth quartile Third wealth quartile Fourth wealth quartile Household hire servant Household possessing car or bike Muslim household Other religion household Share of expenditure on eating out Year (2011-12) Constant Sample size (8) 0.338*** (0.052) -0.003*** (0.001) 0.023 (0.016) 0.115 (0.161) 0.101 (0.401) 0.060* (0.034) -0.110 (0.073) 0.630*** (0.161) 1.091*** (0.222) 1.810*** (0.314) 0.914** (0.422) 0.261 0.337*** (0.053) -0.003*** (0.001) 0.024 (0.016) 0.136 (0.153) 0.108 (0.406) 0.064* (0.034) -0.111 (0.073) 0.617*** (0.169) 1.077*** (0.223) 1.799*** (0.317) 0.914** (0.421) 0.267 0.346*** (0.053) -0.003*** (0.001) 0.022 (0.014) 0.091 (0.158) 0.074 (0.407) 0.054 (0.035) -0.105 (0.074) 0.546*** (0.165) 0.993*** (0.220) 1.748*** (0.331) 0.964** (0.409) 0.273 -0.373 (0.225) -0.796*** (0.161) 0.344*** (0.052) -0.003*** (0.001) 0.026* (0.015) 0.093 (0.157) 0.086 (0.404) 0.053 (0.035) -0.104 (0.073) 0.594*** (0.172) 1.081*** (0.216) 1.839*** (0.328) 0.954** (0.412) 0.282* (0.165) 0.319 (0.250) -0.060 (0.248) -0.041*** (0.165) 0.338 (0.243) -0.061 (0.247) -0.041*** (0.165) 0.250 (0.232) -0.036 (0.239) -0.041*** (0.161) 0.262 (0.231) -0.046 (0.245) -0.042*** (0.147) 0.455** (0.202) -0.105 (0.245) -0.035*** (0.146) 0.484** (0.200) -0.102 (0.243) -0.036*** (0.146) 0.380* (0.186) -0.083 (0.244) -0.036*** (0.144) 0.383** (0.183) -0.093 (0.246) -0.037*** (0.013) 0.697*** (0.229) 13.926*** (0.982) 3,743 (0.013) 0.693*** (0.230) 13.935*** (0.998) 3,743 (0.013) 0.678*** (0.223) 14.357*** (1.018) 3,743 (0.013) 0.672*** (0.220) 13.679*** (0.904) 3,743 (0.012) 0.730*** (0.199) 13.265*** (1.104) 4,574 (0.012) 0.718*** (0.201) 13.321*** (1.115) 4,574 (0.012) 0.706*** (0.200) 13.841*** (1.102) 4,574 (0.012) 0.705*** (0.198) 13.193*** (1.022) 4,574 Vigorous Intensity Years of education Definition 2 (6) (7) -0.133 (0.271) -0.553*** (0.190) Moderate Intensity Age square (5) -0.413* (0.202) -0.287 (0.291) -0.537*** (0.160) High Activity Age (4) 0.367*** (0.056) -0.003*** (0.001) 0.025* (0.015) 0.048 (0.172) -0.198 (0.267) 0.073*** (0.023) -0.129* (0.069) 0.694*** (0.166) 1.180*** (0.199) 1.978*** (0.262) 0.900** (0.390) 0.299* 0.362*** (0.057) -0.003*** (0.001) 0.026* (0.015) 0.094 (0.163) -0.182 (0.272) 0.079*** (0.024) -0.130* (0.069) 0.675*** (0.172) 1.155*** (0.195) 1.955*** (0.262) 0.901** (0.389) 0.311** 0.368*** (0.056) -0.003*** (0.001) 0.021 (0.013) 0.060 (0.172) -0.203 (0.275) 0.069** (0.025) -0.125* (0.070) 0.600*** (0.173) 1.062*** (0.192) 1.890*** (0.278) 0.953** (0.384) 0.313** -0.439* (0.243) -0.762*** (0.120) 0.368*** (0.056) -0.003*** (0.001) 0.024 (0.014) 0.047 (0.172) -0.193 (0.271) 0.067** (0.025) -0.124* (0.069) 0.645*** (0.178) 1.136*** (0.198) 1.974*** (0.280) 0.943** (0.382) 0.314** Notes: Dependent variable is BMI. OLS regression results presented. Sample restricted to 18-60-year-old working urban women at the time of survey in IHDS1 and IHDS2. The regressions also control for a set of state dummies. Standard errors clustered at the state level are in parenthesis. Significance: * p<0.10, ** p<0.05, *** p<0.01. Table B2: Regression of BMI on labour market sector of work and intensity. Married Urban Working Women aged 18–60. Definition 1 Definition 2 OLS IV OLS IV OLS IV OLS IV (1) (2) (3) (4) (5) (6) (7) (8) Blue collar -0.553** -1.008** -0.499** -0.825* (0.212) (0.429) (0.206) (0.443) MET -0.174*** -0.514** -0.192*** -0.636*** (0.059) (0.223) (0.064) (0.185) Age 0.383*** 0.386*** 0.385*** 0.395*** 0.371*** 0.373*** 0.371*** 0.377*** (0.065) (0.063) (0.066) (0.062) (0.071) (0.070) (0.071) (0.069) Age square -0.004*** -0.004*** -0.004*** -0.004*** -0.004*** -0.004*** -0.004*** -0.004*** (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) Years of education 0.034** 0.017 0.043*** 0.022 0.035*** 0.023 0.038*** 0.003 (0.015) (0.019) (0.014) (0.019) (0.012) (0.018) (0.011) (0.018) Married -0.210 -0.202 -0.249 -0.307 -0.515 -0.516 -0.526 -0.556 (0.640) (0.630) (0.632) (0.607) (0.403) (0.393) (0.405) (0.399) Smokes tobacco 0.099 0.104 0.090 0.083 0.086** 0.087** 0.081* 0.074 (0.064) (0.064) (0.065) (0.068) (0.038) (0.039) (0.040) (0.046) Hours watching -0.085 -0.084 -0.083 -0.076 -0.155* -0.154** -0.154* -0.148* Television (0.084) (0.084) (0.084) (0.085) (0.079) (0.078) (0.078) (0.078) Number of children 0.751*** 0.715*** 0.702** 0.519 0.788*** 0.766*** 0.711*** 0.457 (0.248) (0.264) (0.258) (0.319) (0.222) (0.232) (0.230) (0.295) Second wealth quartile 1.302*** 1.211*** 1.275*** 1.008*** 1.352*** 1.295*** 1.268*** 0.872*** (0.258) (0.313) (0.272) (0.370) (0.175) (0.214) (0.189) (0.283) Third wealth quartile 1.857*** 1.726*** 1.875*** 1.596*** 1.988*** 1.890*** 1.952*** 1.524*** (0.295) (0.360) (0.323) (0.403) (0.262) (0.331) (0.297) (0.388) Fourth wealth quartile 0.850** 0.838** 0.899** 0.968*** 0.768* 0.758* 0.826** 0.922*** (0.409) (0.400) (0.392) (0.350) (0.399) (0.393) (0.383) (0.344) Household has 0.047 0.013 0.078 0.055 0.204 0.186 0.228 0.221 domestic help (0.248) (0.236) (0.242) (0.240) (0.200) (0.195) (0.200) (0.199) Household possessing 0.671 0.648 0.624 0.474 0.851** 0.836*** 0.796** 0.617** car or bike (0.430) (0.427) (0.421) (0.418) (0.311) (0.312) (0.305) (0.309) Muslim household 0.260 0.257 0.266 0.268 0.156 0.161 0.159 0.187 (0.337) (0.317) (0.353) (0.349) (0.385) (0.369) (0.395) (0.393) Other religion -0.035*** -0.035*** -0.036*** -0.035*** -0.025*** -0.025*** -0.026*** -0.025*** household (0.012) (0.011) (0.012) (0.013) (0.009) (0.009) (0.009) (0.009) Share of expenditure 0.865*** 0.904*** 0.826*** 0.844*** 0.892*** 0.921*** 0.857*** 0.877*** on eating out (0.232) (0.212) (0.234) (0.228) (0.176) (0.152) (0.185) (0.186) Constant 14.038*** 14.147*** (1.504) (1.678) Sample size 2,353 2,353 2,353 2,353 3,076 3,076 3,076 3,076 Notes: Dependent variable is BMI. OLS and IV regression results presented. Sample restricted to 18-60-year-old working urban married women at the time of survey in IHDS1 and IHDS2. The regressions also include a set of state dummies not reported here. English speaking ability and activity status of spouse are used as instruments. Standard errors clustered at the state level are in parenthesis. p-values are reported in square brackets. Significance: * p<0.10, ** p<0.05, *** p<0.01. Table B3: Regression of BMI on labour market sector of work and intensity. Urban Men aged 18–60. Definition 1 Definition 2 (1) (2) (3) (4) (5) (6) (7) Urban Working Men Blue Collar -0.334*** -0.335*** (0.117) (0.111) Medium Activity -0.066 -0.077 (0.122) (0.128) High Activity -0.496*** -0.483*** (0.138) (0.129) Physical intensity -0.116 -0.119* (in MET) (0.072) (0.066) Moderate Intensity -0.279* (0.144) Vigorous Intensity 0.221 (0.326) Age 0.273*** 0.276*** 0.273*** 0.272*** 0.287*** 0.290*** 0.288*** (0.034) (0.034) (0.034) (0.034) (0.035) (0.035) (0.035) Age square -0.003*** -0.003*** -0.003*** -0.003*** -0.003*** -0.003*** -0.003*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Years of education 0.062*** 0.061*** 0.066*** 0.069*** 0.066*** 0.064*** 0.069*** (0.016) (0.016) (0.015) (0.014) (0.015) (0.016) (0.015) Married 0.680*** 0.670*** 0.675*** 0.671*** 0.736*** 0.727*** 0.730*** (0.229) (0.226) (0.228) (0.229) (0.235) (0.232) (0.234) Smokes tobacco -0.486*** -0.485*** -0.481*** -0.488*** -0.484*** -0.484*** -0.479*** (0.110) (0.111) (0.111) (0.112) (0.107) (0.108) (0.107) Hours watching 0.015 0.015 0.013 0.013 0.009 0.010 0.008 Television (0.072) (0.071) (0.073) (0.072) (0.063) (0.062) (0.063) Number of children -0.043 -0.043 -0.045 -0.039 -0.041 -0.040 -0.043 (0.044) (0.045) (0.045) (0.046) (0.040) (0.040) (0.040) Second wealth quartile 0.657*** 0.627*** 0.632*** 0.654*** 0.557*** 0.530*** 0.531*** (0.107) (0.102) (0.100) (0.096) (0.109) (0.109) (0.110) Third wealth quartile 0.943*** 0.895*** 0.918*** 0.953*** 0.905*** 0.859*** 0.880*** (0.194) (0.194) (0.192) (0.189) (0.193) (0.194) (0.192) Fourth wealth quartile 1.135*** 1.091*** 1.135*** 1.176*** 1.060*** 1.020*** 1.058*** (0.316) (0.314) (0.307) (0.307) (0.319) (0.317) (0.311) Household hire servant 1.094*** 1.102*** 1.100*** 1.092*** 0.992*** 0.998*** 1.000*** (0.355) (0.359) (0.352) (0.351) (0.303) (0.308) (0.301) Household possessing 0.794*** 0.796*** 0.804*** 0.812*** 0.842*** 0.848*** 0.852*** car or bike (0.209) (0.207) (0.209) (0.212) (0.191) (0.190) (0.194) Muslim household 0.236 0.243 0.232 0.243 0.206 0.214 0.200 (0.288) (0.284) (0.289) (0.287) (0.284) (0.280) (0.283) Other religion household 0.147 0.162 0.166 0.170 0.183 0.192 0.200 (0.208) (0.208) (0.210) (0.207) (0.190) (0.190) (0.192) Share of expenditure -0.002 -0.003 -0.002 -0.002 -0.004 -0.005 -0.004 on eating out (0.012) (0.012) (0.012) (0.012) (0.011) (0.011) (0.011) Constant 15.647*** 15.641*** 15.706*** 15.431*** 15.322*** 15.321*** 15.392*** (0.761) (0.766) (0.829) (0.734) (0.763) (0.761) (0.819) High – Medium Activity -0.430*** -0.405*** (0.136) (0.140) (8) -0.275** (0.122) 0.057 (0.278) 0.288*** (0.035) -0.003*** (0.000) 0.072*** (0.013) 0.727*** (0.235) -0.486*** (0.107) 0.007 (0.063) -0.038 (0.041) 0.551*** (0.106) 0.911*** (0.189) 1.095*** (0.312) 0.996*** (0.302) 0.858*** (0.195) 0.211 (0.280) 0.204 (0.190) -0.003 (0.011) 15.097*** (0.725) Vigorous – Moderate Intensity Sample size 5,165 5,165 5,165 0.500 (0.300) 5,165 5,491 5,491 5,491 0.332 (0.247) 5,491 Notes: Dependent variable is BMI. OLS regression results presented. Sample restricted to 18-60-year-old working urban men in IHDS2. The regressions also control for a set of state dummies. Standard errors clustered at the state level are in parenthesis. Significance: * p<0.10, ** p<0.05, *** p<0.01. Table B4: Regression of BMI on labour market sector of work and intensity. Married Urban Working Men 18–60. Definition 1 Definition 2 OLS IV OLS IV OLS IV OLS (1) (2) (3) (4) (5) (6) (7) Blue collar -0.651** -1.178*** -0.623*** -1.307*** (0.247) (0.391) (0.188) (0.472) MET -0.010 -0.651** 0.029 (0.088) (0.276) (0.072) Age 0.159 0.157 0.155 0.173 0.172 0.173 0.174 (0.121) (0.125) (0.125) (0.135) (0.141) (0.135) (0.143) Age square -0.002 -0.002 -0.002 -0.002 -0.002 -0.002 -0.002 (0.001) (0.001) (0.001) (0.001) (0.002) (0.001) (0.002) Years of education 0.023 0.040 0.061 0.027 0.048 0.027 0.069* (0.035) (0.037) (0.040) (0.033) (0.036) (0.033) (0.036) Smokes tobacco -0.099 -0.095 -0.089 -0.250 -0.257 -0.250 -0.266 (0.420) (0.439) (0.437) (0.277) (0.291) (0.277) (0.288) Hours watching 0.077 0.081 0.087 -0.077 -0.078 -0.077 -0.080 Television (0.125) (0.135) (0.137) (0.100) (0.105) (0.100) (0.106) Number of children 0.036 0.035 0.033 0.059 0.052 0.059 0.046 (0.145) (0.156) (0.162) (0.082) (0.087) (0.082) (0.092) Second wealth quartile 0.355 0.380 0.408 0.145 0.175 0.145 0.213 (0.351) (0.366) (0.358) (0.293) (0.308) (0.293) (0.301) Third wealth quartile 1.027** 1.070** 1.117** 0.744** 0.807* 0.744** 0.883** (0.408) (0.442) (0.415) (0.356) (0.401) (0.356) (0.374) Fourth wealth quartile 1.814*** 1.958*** 2.126*** 1.451*** 1.666*** 1.451*** 1.890*** (0.529) (0.594) (0.528) (0.489) (0.569) (0.489) (0.506) Household has 0.802* 0.827* 0.858* 0.497 0.520 0.497 0.536 domestic help (0.427) (0.469) (0.502) (0.477) (0.497) (0.477) (0.505) Household possessing 0.612* 0.640* 0.674** 0.645* 0.670* 0.645* 0.691* car or bike (0.326) (0.324) (0.313) (0.362) (0.366) (0.362) (0.356) Muslim household 1.170* 1.224 1.286* 0.797 0.855 0.797 0.924 (0.694) (0.726) (0.701) (0.755) (0.772) (0.755) (0.737) Other religion 0.409 0.392 0.371 0.283 0.253 0.283 0.223 household (0.599) (0.609) (0.598) (0.529) (0.529) (0.529) (0.525) Share of expenditure -0.022 -0.024 -0.026 -0.012 -0.013 -0.012 -0.014 on eating out (0.020) (0.022) (0.022) (0.023) (0.025) (0.023) (0.025) Constant 17.285*** 16.708*** 17.656*** 16.856*** (2.862) (2.981) (3.422) (3.571) Sample size 802 802 802 802 1,025 1,025 1,025 IV (8) -0.632** (0.320) 0.136 (0.142) -0.002 (0.002) 0.020 (0.039) -0.171 (0.263) -0.064 (0.112) 0.055 (0.081) -0.029 (0.259) 0.439 (0.356) 1.243*** (0.428) 0.679 (0.479) 0.724** (0.361) 0.574 (0.762) 0.300 (0.573) -0.011 (0.024) 1,025 Notes: Dependent variable is BMI. OLS and IV regression results presented. Sample restricted to 18-60-year-old working urban married men in IHDS2. The regressions also include a set of state dummies not reported here. English speaking ability and activity status of spouse are used as instruments. Standard errors clustered at the state level are in parenthesis. p-values are reported in square brackets. Significance: * p<0.10, ** p<0.05, *** p<0.01.