Body mass index, obesity and labour market: outcomes in Spain

57th ERSA Congress “Social Progress for Resilient Regions” Groningen, 2017 BODY MASS INDEX, OBESITY AND LABOUR MARKET OUTCOMES IN SPAIN Authors: Dr. Nuria Benítez Llamazares Dr. Ricardo Pagán Rodríguez University of Malaga CONTENTS I. INTRODUCTION II. METHODOLOGY III. RESEARCH IV. CONCLUSIONS V. LIMITATIONS OF THE STUDY VI. FUTURE RESEARCH LINES 2 Determinate if body mass index (BMI) –in particular high levels of BMI: overweight and obesity - from working-age males and females living in Spain affects to the probability of being in a specific labour status (inactivity, unemployment or employment). 3 I. INTRODUCTION Use data taken from the European Health Interview Survey (EHIS) from 2009 in Spain:first edition of the survey, comparable at supranational level and with high future prospects. 1. Aim of the study •Overweight and obesity are main World-class problems (WHO,2004 and WHO,2015) (FAO, 2014). •The importance of research on the consequences of overweight and obesity in social relationship, lifestyle an socioeconomic status in people (Bozoyan and Wolbring, 2011; Caliendo and Wang-Sheng Lee, 2013; Cawley, 2004; Cawley, Han and Norton, 2009; Colchero and Bishai, 2007; Han, Norton and Powell, 2011; Lindeboom, Lundborg and Bas Van der Klaaus, 2010; Sabia and Rees, 2012). •Limited national studies focused on the influence that obesity has in the structure of the labour market in Spain. •Main methodology: the European Health Interview Survey (EHIS) from 2009 in Spain. •The results are representative of the socioeconomic structure in Spain. 4 I. INTRODUCTION 2. Selection of the topic of study Overweight and Obesity:“abnormal or excessive fat accumulation that may impair health” (WHO). Key point: both of them are risk factors for the heath of people. 3. Overweight and obesity. Definition and quantification I. INTRODUCTION 5 Adolphe Quételex (1796-1874) defined a relative weight indicator (Quetelex’s index): In 1972, Keys, Fidanza et al. published Indices of relative weight and obesity and renamed Quetelex’s index as Body mass index (BMI). They concluded that BMI is a convenient and reliable indicator of overweight and obesity. BMI = Weight (kg) Height 2(m) I. INTRODUCTION In this study, all levels of obesity have been clustered into one (BMI>30). In addition, the BMI variable has been used in its continuous version. 6 The WHO classifies the nutritional status of adults, both men and women, using as key variable BMI: BMI (kg/m2) Nutritional status Less than 18.5 Underweight From 18.5 to 24.9 Normal weight From 25.0 to 29.9 Overweight From 30.0 to 34.9 Obesity class I From 35.0 to 39.9 Obesity class II Above 39.9 Obesity class III Source: World Health Organization (WHO) 3. Overweight and obesity. Definition and quantification V. METODOLOGÍA 1. Obtención de datos: Encuesta Europea de Salud en España 2009. 7 II. METHODOLOGY 1. Data collection: European Health Interview Survey in Spain (EHIS 2009). Selection of the survey: •NATIONALHEALTH SURVEY IN SPAIN 2011-2012 (8th edition). -Advantages: Comparable backwards in time. -Inconvenient:Not comparable at supranational level or with EHIS 2009. •EUROPEAN HEALTH INTERVIEW SURVEY IN SPAIN 2009 (1st edition). -Advantages: Comparable at European level and with high prospects in the future. -Inconvenient:No possibility to analyse time series backwards. Both are similar in type of questionnaires, population area (residents of main family dwellings), sample size (approximately 24,000 theoretical observations), territory of application (national), and methods of collecting information (face-to-face interview mode: interviewer-PAPI and CAPI). V. METODOLOGÍA 8 II. METHODOLOGY EUROPEAN HEALTH INTERVIEW SURVEY (1ST EDITION) TEMPORARY DISTRIBUTION BY COUNTRIES Year Country 2006 Austria and Estonia. 2007 Slovenia and Switzerland. 2008 Belgium, Bulgaria, France, Cyprus, Latvia, Malta, Czech Republic, Romania and Turkey. 2009 Germany, Slovakia, Spain, Greece, Hungary and Poland. Source: Methodology of the European Health Interview Survey, Eurostat. EUROPEAN HEALTH INTERVIEW SURVEY IN SPAIN 2014, 2ND EDITION (EHIS 2014): results available in November 2015. 1. Data collection: European Health Interview Survey in Spain (EHIS 2009). V. METODOLOGÍA 2. Data and definition of variable 9 II. METHODOLOGY DATA VARIABLES Cross -sectional Dependent variables (LDV) Independent variables Household questionnaire + Adult questionnaire = 22.188 theoretical obs. (theoretical sample size). 20.891 obs. valid. SITU : Occupational status of individuals. Values (probit model): =0, INACTIVE. =1, UNEMPLOYED =2, PART-TIME EMPLOYED =3, FULL-TIME EMPLOYED Qualitative: Most of them. Level of studies, Region (CC.AA), Size of municipality, Disability, Household composition, Etc. Finally selected: 15.055 obs.(working -age individuals, 16 -64 years old). Note : SITU is obtained from the number of working hours of individuals, so that: I NACTIVE and UNEMPLOYED = 0 hours of work per week. PAR -TIME EMPLOYED = less than 35/40 hours of work per week. FULL -TIME EMPLOYED = 35/40 hours of work per week. Quantitative • Discrete Age Number of children • Continuous BMI Household income level 2. Econometric Analysis 16 III. RESEARCH MARGINAL EFECTS OF VARIABLES BMI AND BMI2 HYPOTHESIS 1: DETERMINATE IF BMI INFLUENCES THE OCCUPATIONALSTATUS Dependent variable: SITU Dependent variable: SITU BMI BMI2 BMI BMI2 Total effect Inactive Unemployed Part-time employed Full-time employed Women Total effect Inactive Unemployed Part-time employed Full-time employed Men Maximum value of SITU Maximum value of SITU BMI2 BMI2 17 III. RESEARCH 2. Econometric analysis HYPOTHESIS 1: DETERMINATE IF BMI INFLUENCES THE OCCUPATIONALSTATUS Hola Hola Hola Hola BMI Probability of being full-time employed Men BMI = 26.5 Women BMI = 25.5 VI. INVESTIGACIÓN 2. Análisis econométrico 18 III. RESEARCH HYPOTHESIS 2: DETERMINATE IF DEPENDENT VARIABLE OCCUPATIONALSTATUS (SITU) IS TRUNCATED OR CENSORED Estimation of the labour insertion equation with a tobit model: Goodness of fit: Good. Wald test and Pseudo R2 test: Both significant. Maximum value of SITU BMI2 BMI BMI2 Table 2. Labour insertion equation estimated with a Tobit model (0 = NOT WORKING, 1 = WORKING) Independent variables Men Women Coefficients Coefficients 19 III. RESEARCH 2. Econometric analysis Hola Hola Hola Hola BMI Probability of being employed Men BMI = 26.5 Women BMI = 24.4 HYPOTHESIS 2: DETERMINATE IF DEPENDENT VARIABLE OCUPATIONALSTATUS (SITU) IS TRUNCATED OR CENSORED •It is proved that BMI influences the occupational status of individuals, especially in obese women who are up to 9.5% less likely to be employed (Cawley, 2004). 20 IV. CONCLUSIONS CONTRAST OF HYPOTHESIS Hypothesis 1 ✔ Hypothesis 2 ✔ 1. LIMITATIONS OF THE VARIABLE BODY MASS INDEX (BMI). 2. RELATIVELY OUTDATED SURVEYAND LIMITED DATA. 3. LIMITATIONS OF VARIABLESAND ESTIMATION METHODS: •Limited and qualitative dependent variables (LDV). •Probit and Tobit models are asymptotic (consistent, asymptotically normal and asymptotically efficient). •Importance of the initial value distribution of the variables. 4. REVERSE CAUSALITY. 21 V. LIMITACIONS OF THE STUDY •Analyse the 2nd edition of the European Health Survey in Spain (2014). •Compare the results of EHIS09 and EHIS14 and work out time series. •Expand the study to other countries of the European Union. •Analyse the effect of overweight and obesity on other variables of the labour market (e.g. Salaries). 22 VI. FUTURE RESEARCH LINES THANK YOU VERY MUCH FOR YOUR ATTENTION! 23