The Effects of Cigarette Prices on Obesity

1
The Effects of Cigarette Prices on Obesity

• by
• Charles L. Baum II
• Middle Tennessee State University
• Economics and Finance Department, Box 27
• Murfreesboro, TN 37132
• cbaum_at_mtsu.edu
• W [phone]
• F [phone]

2
Introduction

• 30 of Americans are currently obese (Flegal et
  al., 2002).
• Only 15 of Americans were obese 25 years ago.
• Americans are now more likely to be obese than to
  smoke cigarettes only 22.5 of Americans
  currently smoke (Campaign for Tobacco-Free Kids,
  2004) .

3
Definitions

• Obese is defined as having a Body Mass Index
  (BMI) of 30 or more (CDC).
• BMI is calculated as weight in kilograms divided
  by height in meters squared.
• Overweight is a BMI of 25 to 30 (CDC).

4
Costs of Obesity

• Obesity causes 365,000 premature deaths annually
  smoking causes 435,000 premature deaths (Mokdad
  et al., 2004 Mokdad 2005).
• Obesity is expected soon to overtake tobacco as
  the leading preventable cause of death.
• Medical expenses (Finkelstein et al., 2004).
• Lower wages.

5
Why?

• Americans currently possess more knowledge of the
  consequences of obesity than ever before.
• More Americans are exercising and dieting.
• Americans are estimated to spend over 30 billion
  dollars on weight loss programs annually
  (Philipson and Posner, 1999).
• Biological Factors havent changed substantially.
• Weight is no longer a status symbol.

6
Related Literature

• Philipson (2001), Philipson and Posner (1999),
  and Lakdawalla and Philipson (2002).
• Anderson, Butcher, and Levine (2003a, 2003b).
• Cutler, Glaeser, and Shapiro (2003).
• Chou, Grossman, and Saffer (2004).

7
This Study

• I examine the effects of cigarette prices on
  obesity and BMI.
• I also examine whether cigarette prices affect
  obesity through cigarette smoking.
• I examine dynamic effects of cigarette prices.

8
Effects of Cigarette Prices

• In 1964, the U. S. Surgeon General issued its
  first report relating smoking and health.
• Federal and state governments have since
  increased cigarette taxes (Grossman, 2001).
• The incidence of cigarette smoking has fallen
  from about 45 to 22.5 since the early 1960s.
• Cigarette smoking and obesity seem to be
  inversely related (for example, see Williamson et
  al., 1991).
• The incidence of obesity in the early 1960s was
  13.3 compared to over 30 today (CDC, 2004).

9
Related Literature

• Cigarette prices reduce smoking, with
  elasticities ranging from 0.10 to 0.75.
• For example, Becker, Grossman, and Murphy (1991,
  1994) find that a permanent 10 increase in
  cigarette prices reduces cigarette consumption by
  4 initially and by 7.5 in the long run.

10
Theoretical Model
11
Theoretical Model
12
Theoretical Model
13
Theoretical Model
14
Theoretical Model
15
Theoretical Model
16
Data

• I use National Longitudinal Survey of Youth
  (NLSY) data.
• 12,686 respondents aged 14 to 21 in 1979
  interviewed annually from 1979 to 1994 and
  biennially thereafter.
• Oversamples African-Americans, Hispanics, and
  those with low-incomes.
• Weight information collected in 1981, 1982, 1985,
  1986, 1987, 1989, 1990, 1992, 1993, 1994, 1996,
  1998, and 2000 surveys.
• Generates 101,641 person-year observations across
  the 13 surveys.
• I adjust weight (and height) using Health and
  Nutrition Examination Survey (NHANES) data
  (Cawley, 2000).

17
Figure 1 Body Mass Index
18
Figure 2 Incidence of Obesity
19
Data

• Smoking information collected in 1984, 1992,
  1994, and 1998 NLSY surveys.
• The NLSY identifies state and county of residence
  for each respondent.
• Local cigarette prices acquired from the Tobacco
  Institute.
• Possible border-crossing.
• NLSY limitations.
• Limitations of NHANESIII, NHANES 1999-2004, NHIS,
  and BRFSS data.

20
Table 1 Descriptive Statistics

• Sample means with standard deviations in
  parentheses. The sample size contains 101,641
  person-year observations, 19,506 of which are
  obese and 82,135 of which are not. a The smoking
  sub-sample contains 21,313 observations, 6095 of
  whom smoke and 15,218 of whom do not.

21
Empirical Methodology

• Part 1 Estimate the effects of cigarette prices
  on obesity
• Also examine BMI.
• Examine males and females separately.
• Control for demographic characteristics,
  state-specific effects, national time trends, and
  state-specific time trends.

22
Table 2 The Effects of Cigarette Prices on BMI
and Obesity

• Standard errors are in parentheses and the
  elasticities of cigarette prices are in brackets.
  p are 101,641 observations. The models include
  individual characteristics variables, state dummy
  variables, a national time trend variable, and
  state-specific time trend variables.

23
Table 3 The Effects of Cigarette Prices on BMI
and Obesity

• Standard errors are in parentheses and the
  elasticities of cigarette prices are in brackets.
  p are 101,641 observations. The models include
  individual characteristics variables, state dummy
  variables, a national time trend variable, and
  state-specific time trend variables. a indicates
  that border county respondents are excluded from
  the analysis (with 67,189 in the sample).

24
Empirical Methodology

• Part 2 Examine the cigarette prices-cigarette
  smoking-obesity link
• Estimate the effects of smoking on obesity.
• Potential bias (Cawley, Markowitz, Taurus, 2003).
• Estimate the effects of cigarette prices on
  cigarette smoking.
• Predict cigarette smoking.
• Estimate the effect of predicted cigarette
  smoking on obesity.
• IV Issues Nelson and Startz (1990a,b), Bound,
  Jaeger, and Baker (1995), and Staiger and Stock
  (1997).

25
Table 4 The Effects of Cigarette Smoking on BMI
and Obesity

• Standard errors are in parentheses and the
  marginal effects of cigarette prices are in
  brackets. p 0.01. There are 21,313 observations. The models
  include individual characteristics variables,
  state dummy variables, a national time trend
  variable, and state-specific time trend variables.

26
Table 5 The Effects of Cigarette Prices on the
Incidence of Smoking and Number of Cigarettes
Smoked

• Standard errors are in parentheses and the
  marginal effects of cigarette prices are in
  brackets. p 0.01. There are 21,313 observations. The models
  include individual characteristics variables,
  state dummy variables, a national time trend
  variable, and state-specific time trend
  variables.

27
Table 6 The Effects of Cigarette Smoking on BMI
and Obesity (Predicted Values)

• Standard errors are in parentheses and the
  marginal effects of cigarette prices and
  cigarette taxes are in brackets. p observations. The models include individual
  characteristics variables, state dummy variables,
  a national time trend variable, and
  state-specific time trend variables. R-squared
  values are similar to those reported in table 4.

28
Empirical Methodology

• Part III Examine dynamic effects because current
  weight is not independently determined from past
  weight.
• Examine lagged cigarette prices.
• Examine change in BMI.
• Estimate obesity hazards.

29
Table 7 The Effects of Lagged Cigarette Prices
on BMI and Obesity

• Standard errors are in parentheses and the
  elasticities of cigarette prices are in brackets.
  p are 101,640 observations. The models include
  individual characteristics variables, state dummy
  variables, a national time trend variable, and
  state-specific time trend variables. R-squared
  values are similar to those in table 2.

30
Hazard Specification
31
Likelihood Function
32
Table 8 The Effects of Cigarette Prices on
Change in BMI and the Conditional Probability of
Becoming Obese

• Standard errors are in parentheses and the
  elasticities of cigarette prices are in brackets.
  p are 87,376 observations in the change in BMI
  model. The obesity hazard models use 9763
  respondents, each of whom can provide up to 13
  period-observations. The models include
  individual characteristics variables, state dummy
  variables, a national time trend variable, and
  state-specific time trend variables.

33
Figure 3 Probability of Remaining Non-Obese
(full sample)
34
Discussion and Conclusions

• Cigarette prices increase BMI and the incidence
  of obesity. Estimates suggest increasing the
  price of cigarettes by 0.77 (Health People 2010)
  would
• Increase obesity from 0.192 to 0.198, which is
  0.006 points or a 3.125 change.
• Decrease cigarette smoking from 0.286 to 0.245,
  which is 0.041 points or a 14.33 change.
• Increase obesity nationally from 30 to 31
  (about a 3.125 change) or by 2 million
  Americans.
• Decrease smoking nationally from 22.5 to 19.5
  (about a 14.33 change) or by 6 million
  Americans.

35
Discussion and Conclusions

• The 0.77 cigarette price increase would
• Increase premature deaths due to obesity by
  12,160 annually (365,000/60,000,000 0.00608
  times 2 million additional obese Americans).
• Decrease premature deaths due to cigarette
  smoking by 57,960 annually (435,000/45,000,000
  0.00966 times 6 million fewer smokers).

36
Discussion and Conclusions

• Increasing the price of cigarettes would decrease
  the number of premature deaths by 45,800 annually
  (not 57,960).
• U.S. Surgeon General has consistently said that
  the benefits of quitting smoking outweigh any
  costs, acknowledging that this may include weight
  gain.

37
Discussion and Conclusions

• Responses of the typical American might be
  different.
• Lagged effects on obesity are larger the same
  price increase would increase premature deaths
  due to obesity by 85,120 annually.
• Dynamic effects (over 20 years) are larger, too.
• Lagged effects could be larger on smoking.

38
Discussion and Conclusions

• Cigarette prices decrease cigarette smoking and
  decreased cigarette smoking increases BMI and
  obesity. If cigarette prices were increased
  0.77
• Cigarette smoking would decrease from 0.286 to
  0.245.
• This would increase obesity by about 0.003
  points.
• Remaining effect of cigarette prices could
  operate through food consumption.

39
Extensions

• Confirm magnitudes with nationally representative
  data -- Behavioral Risk Factor Surveillance
  System (BRFSS) data.
• Examine the effects of cigarette prices on
  adolescent obesity.
• Further exploit cigarette prices as an instrument
  with a larger data set across more years.

40
(No Transcript)