Alcohol Consumption and Body Weight J. Catherine Maclean, M.A. Addiction Health Services Research Co

1
Alcohol Consumption and Body WeightJ.
Catherine Maclean, M.A.Addiction Health Services
Research ConferenceOctober 24-26, 2005Santa
Monica, CA
Health Economics Research Group
2
Motivation

• According to the recent data, the majority of
  Americans are now overweight (Must et al, 1999
  Flegal et al, 2002)
• Between 1980 and 2000 the prevalence of
  overweight increased 40 percent and obesity 110
  percent (Flegal, Carroll, Ogden, Johnson, 2002)
• Overweight and obesity increase individual risk
  for a range of negative health conditions (i.e.,
  heart disease, stroke, high blood pressure,
  cancer, and diabetes)
• Elevated rates of chronic disease among the obese
  have increased health care costs to society, in
  2000 the estimated total cost of overweight and
  obesity was 117 billion U.S. dollars (U.S.
  Department of Health and Human Services, 2001)
• Health officials view current obesity rates as
  problematic and seek policies and programs to
  reduce weight
• Physiologically obesity is clearly understood as
  the consequence of a calorie imbalance calories
  consumed gt calories expended
• Ambiguity over the cause (s) of the current high
  levels of obesity

3
Current Study

• Objectives
• Examine relationship between alcohol consumption
  and weight status using an economic framework
• Hypotheses
• Alcohol use is endogenous in the weight status
  specifications
• Alcohol use is positively related to body weight,
  increases the probability of overweight and/or
  obese, and decreases the probability of
  underweight

4
Economic Interpretation of Current Weight Levels

• A number of technological, economic, and social
  changes have taken place over the past few
  decades and altered income levels and relative
  prices (Grossman et al 2002, 2005 Philipson et
  al 2002, 2003 Cutler, Glaiser, Shapiro, 2003)
• These changes have simultaneously decreased the
  relative price of calorie consumption, increased
  the relative price of exercise, and increased
  income levels
• Observed weight levels are the result of
  fully-informed rational individuals responding to
  the new economic environment and maximizing
  utility not evidence of a market failure
• Government intervention to lower weight is not an
  efficient use of resources nor will such action
  increase welfare
• Government action may be welfare enhancing if
• Costs of obesity imposed on society are greater
  than the utility derived by overweight and obese
  individuals from their calorie consumption,
  calorie expenditure, and weight status
• Individuals are not rational and/or are not fully
  informed on the causes and/or consequences of
  obesity

5
Role of Alcohol

• Relative price of alcohol has fallen over past
  several decades
• Theory tells us that individuals will consume
  more alcohol in the new economic environment
• Alcohol is both a high calorie beverage and an
  intoxicating and addictive substance that can
  affect individuals cognitive processes and
  ability to evaluate costs and benefits
• If alcohol use is identified as a risk factor for
  elevated weight status through its influence on
  rationality, policies to reduce alcohol
  consumption may increase social welfare

6
Data

• National Epidemiological Survey of Alcohol and
  Related Conditions (NESARC) Wave 1
• Recent (2001-2002)
• Large (43,083 respondents)
• Nationally representative
• Height and weight information
• Comprehensive alcohol use items
• Over-samples previously under-represented
  sub-groups
• Geographic identifiers

7
Measures

• Body Weight
• Body Mass Index (BMI)
• Underweight (BMI lt 18.5)
• Overweight and/or Obese (BMI gt 25)
• Obese (BMI gt 30)
• Alcohol Use
• Current drinking
• Days drinking to intoxication
• Abuse and/or dependence
• Former drinking included in all specifications
• Standard covariate set (age, sex, race,
  ethnicity, country of birth, region, employment,
  educational attainment, income, marital status)

8
Descriptive Statistics
9
Methods

• Exogeneity of alcohol consumption rejected in
  BMI, overweight and/or obese, and obese
  specifications exogeneity could not be rejected
  in underweight specification
• Statistical models
• BMI specification Two Stage Least Squares
• Overweight and/or obese and obese specifications
  Amemiyas Generalized Least Squares (Newey, 1987)
  and two-stage residual inclusion method (Rivers
  and Vuong, 1988)
• Underweight Probit
• Instrumental Variables
• State-level beer tax and beer tax squared
• Dichotomous indicators for a parent or sibling
  who has/had a problem with alcohol

10
Results

• All 3 current alcohol use measures related to
  weight status outcomes in hypothesized
  directions, most at statistically significant
  levels
• Examples
• Current drinking associated with a 5.9 unit
  increase in BMI
• One additional day drinking to intoxication
  associated with a 10 percentage point increase in
  the probability of obesity
• Diagnosis of alcohol abuse and/or dependence
  associated with a 26 percentage point increase in
  the likelihood of overweight and/or obese

11
Discussion

• Unique contribution of research first study to
  employ an economic framework and econometric
  techniques to examine the alcohol-weight status
  relationship
• Extensions
• Use correction factor for self-reported BMI
  outlined by Lee Sepanski (1995), Cawley (2000),
  and Bound et al (2002)
• Gender specific analyses
• Utilize future waves of NESARC to examine
  longitudinal relationship between alcohol use and
  weight status
• Implications for policy
• Current research documents and quantifies link
  between alcohol consumption and weight status