ArticlePDF Available

How do Smoking Bans in Bars/Restaurants Affect Alcohol Consumption?

January 2011

January 2011

Authors:

Michael K. Wohlgenant

North Carolina State University

In this study, we employ a rational addiction framework to analyze the effects of smoking bans on alcohol consumption in bars/restaurants. We use pseudo panel data approach which has many advantages compared to panel data. Although cigarettes and alcohol are complements in consumption, our findings suggest that smoking bans in restaurants do not have a significant effect on the restaurant alcohol consumption. It is possible that smoking bans at restaurants cause a decrease in the restaurant alcohol consumption of smokers, but lead to an increase in the restaurant alcohol consumption of nonsmokers. If this is the case, the net effect of smoking bans on overall restaurant alcohol consumption would be zero. These results are just preliminary, and further analyses are required.

Figures - uploaded by Michael K. Wohlgenant

Content may be subject to copyright.

Content uploaded by Michael K. Wohlgenant

Content may be subject to copyright.

Poster Design & Printing by Genigraphics®-

800.790.4001

How do Smoking Bans in Bars/Restaurants Affect Alcohol

Consumption?

Aycan Koksal1, Michael Wohlgenant 2

Department of Agricultural and Resource Economics, North Carolina State University

1akoksal@ncsu.edu, 2michael_wohlgenant@ncsu.edu

Poster prepared for presentation at the Agricultural & Applied Economics Association’s 2011

AAEA & NAREA Joint Annual Meeting, Pittsburgh, Pennsylvania, July 24-26, 2011

Preliminary Results - Please do not cite

Copyright 2011 by Aycan Koksal and Michael Wohlgenant. All rights reserved. Readers may make

verbatim copies of this document for non-commercial purposes by any means, provided that this copyright

notice appears on all such copies.

Aycan Koksal

Ph.D. candidate

Agricultural and Resource Econ.

North Carolina State University

E-mail: akoksal@ncsu.edu

Phone: (919) 389-5908

CONTACT

Poster Design & Printing by Genigraphics®-

800.790.4001

OBJECTIVE

To analyze the effects

of smoking bans on

alcohol consumption

at the restaurants

Aycan Koksal and Michael Wohlgenant

North Carolina State University, Raleigh NC 27695

INTRODUCTION

DATA

DISCUSSION

RESULTS (cont.)

METHOD

CONTACT

RESULTS

How do Smoking Bans in Bars/Restaurants Affect Alcohol Consumption?

Aycan Koksal

Ph.D. candidate

Agricultural and Resource Econ.

North Carolina State University

E-mail: akoksal@ncsu.edu

Phone: (919) 389-5908

If cigarettes and alcohol are complements,

smoking bans at restaurants might decrease

restaurant alcohol consumption but increase

home alcohol consumption .

Thus, we consider restaurant and home

alcohol consumption as two separate goods with

separate habit stocks.

When utility function is quadratic, rational

addiction theory implies following demand

functions (see Bask and Melkersson 2004):

ARit = α1i+ β10+ β11ARit-1+ β12ARit+1+ β13AHit-1+ β14AHit

+ β15AHit+1+ β16Cit-1+ β17Cit + β18Cit+1+ β19PARt

+ γ10 Dt+ γ11 Xi+ u1it

AHit= α2i+ β20+ β21AHit-1+ β22AHit+1+ β23ARit-1+ β24ARit

+ β25ARit+1+ β26Cit-1+ β27Cit + β28Cit+1+ β29PAHt

+γ20 Dt+ γ22 Xi + u2it

Cit = α2i+ β20+ β21Cit-1+ β22Cit+1+ β23ARit-1+ β24ARit

+ β25ARit+1+ β26AHit-1+ β27AHit + β28AHit+1+ β29PCt

+ γ30 Dt+ γ33 Xi +u3it

where ARit is restaurant alcohol consumption

AHit is home alcohol consumption

Cit is cigarette consumption

Dtis a binary variable showing if the state

household resides banned smoking at restaurants

Rational addiction implies βi1 > 0 and βi2 > 0.

A positive (negative) coefficient on the current

consumption of another good suggests

complementarity (substitutability).

We allocate households into cohorts based

on geographic region and gender.

All cohort variables are weighted by the

square root of the number of households in

each cohort. Then fixed effects estimators are

calculated (see McKenzie, 2004).

2002-2008 Consumer Expenditure Diary Survey

Data by Bureau of Labor Statistics is used.

Cigarette prices are from Orzechowski&Walker.

For alcohol , we construct Lewbel price indices.

After dropping observations with missing or

recoded state variables, approx. 1200-1400

households remained in each quarter.

In the home alcohol demand equation, current

cigarette consumption has a positive and

significant coefficient which suggests

complementarity relationship.

Smoking ban at restaurants dummy has a

negative coefficient in all three equations, it is

not significantly different from zero.

The results can be explained with the

following scenerio:

- If cigarette and alcohol are complements,

smoking bans at restaurants might cause a

decrease in the restaurant alcohol consumption of

smokers, but might increase restaurant alcohol

consumption of nonsmokers.

-If this is the case, the net effect of smoking

bans on overall restaurant alcohol consumption

will be zero.

These results are just preliminery, and further

analyses are required.

As more states consider smoking bans, it is

necessary to analyze their economic impacts.

If cigarette and alcohol are related in

consumption, as suggested by some studies,

smoking bans can affect alcohol consumption too.

Particularly, smoking bans in bars/restaurants

created a natural experiment to examine the

relationship between smoking and drinking.

We employ a rational addiction framework to

analyze the effect of smoking bans on alcohol

consumption in bars/restaurants.

We use a pseudo panel data approach.

Pseudo panel is disaggregated enough, and it has

main advantages compared with panel data:

- It avoids attrition problem.

- It eliminates difficulties of censoring.

- It has less bias due to measurement error as we

are working with a group average.

Table 1. Smoking bans (at restaurants) over 2002- 2008 period

year

#

states

2002

2

UT, DE

2003

4

UT, DE, NY, FL

2004

7

UT, DE, NY, FL, ME, ID, MA

2005

10

UT, DE, NY, FL, ME, ID, MA, RI, MT, WA

2006

15

UT, DE, NY, FL, ME, ID, MA, RI, MT, WA, NJ, CO, HI, OH, NV

2007

21

UT, DE, NY, FL, ME, ID, MA, RI, MT, WA, NJ, CO, HI, OH, NV, DC, LA, OR,

TN, NH, MN

2008

25

UT, DE, NY, FL, ME, ID, MA, RI, MT, WA, NJ, CO, HI, OH, NV, DC, LA, OR,

TN, NH, MN, IL, MD, IA, PA

Table 2.

Alcohol at Rest

Alcohol at Home

Cigarette

Constnt

41.732

Constnt

60.577

Constnt

-89.025

(0.364)

(0.226)

(<.001)

ARt-1

0.123

AHt-1

-0.009

Ct-1

0.112

(0.077)

(0.907)

(0.107)

ARt+1

0.128

AHt+1

-0.105

Ct+1

0.074

(0.060)

(0.136)

(0.288)

AHt-1

-0.074

ARt-1

0.026

ARt-1

-0.008

(0.259)

(0.739)

(0.844)

AHt

0.064

ARt

0.073

ARt

-0.011

(0.327)

(0.362)

(0.780)

AHt+1

0.014

ARt+1

-0.123

ARt+1

0.045

(0.822)

(0.101)

(0.236)

Ct-1

0.063

Ct-1

-0.105

AHt-1

0.047

(0.611)

(0.435)

(0.195)

Ct

-0.056

Ct

0.243

AHt

0.051

(0.677)

(0.100)

(0.158)

Ct+1

-0.008

Ct+1

0.082

AHt+1

-0.014

(0.951)

(0.545)

(0.699)

PARt

-27.346

PAHt

-40.005

PCt

-3.047

(0.011)

(<.001)

(0.377)

ban

-1.957

ban

-2.268

ban

-1.020

(0.241)

(0.218)

(0.269)

rincome

0.136

rincome

0.018

rincome

0.012

(<.001)

(0.618)

(0.541)

fam.size

-5.888

fam.size

5.373

fam.size

3.511

(0.123)

(0.203)

(0.091)

perslt18

10.129

perslt18

0.304

perslt18

-4.913

(0.036)

(0.955)

(0.067)

age.ref

0.578

age.ref

0.395

age.ref

0.060

(0.003)

(0.063)

(0.574)

white

9.323

white

30.765

white

2.582

(0.211)

(<0.001)

(0.506)

married

-8.009

married

-19.737

married

0.035

(0.411)

(0.066)

(0.995)

widowd

-15.567

widowd

-17.513

widowd

8.347

(0.261)

(0.249)

(0.275)

divorced

-7.417

divorced

-10.770

divorced

3.216

(0.475)

(0.345)

(0.578)

seperatd

-16.361

seperatd

-34.641

seperatd

7.524

(0.441)

(0.137)

(0.522)

college

2.372

college

-2.302

college

0.864

(0.709)

(0.741)

(0.808)

R2

0.579

R2

0.569

R2

0.676

Table 3.

seperate

system

εAR,AR

-3.357

(0.012)

-6.684

(<.001)

εAH,AH

-2.523

(0.001)

-4.981

(<.001)

εC,C

-0.538

(0.362)

-0.747

(0.226)

εAR,AH

-0.018

(0.973)

3.595

(0.069)

εAR,C

-0.001

(0.999)

1.293

(0.163)

εAH,AR

0.039

(0.879)

3.866

(0.001)

εAH,C

-0.063

(0.507)

-0.290

(0.588)

εC,AR

-0.127

(0.674)

-0.927

(0.484)

εC,AH

-0.441

(0.231)

1.061

(0.419)

εAR,Y

0.970

(<.001)

0.849

(<.001)

εAH,Y

0.067

(0.559)

0.277

(0.029)

εC,Y

0.146

(0.290)

0.094

(0.518)

ResearchGate has not been able to resolve any citations for this publication.

ResearchGate has not been able to resolve any references for this publication.

Article

A Note on the Interpretation of the Rational Addiction Model

January 1996

Brian S Ferguson

The Rational Addiction Model of Becker and Murphy (1988) has become a standard tool in the analysis of the demand for drugs, cigarettes, alcohol and other potentially addictive goods. In this paper, we shall argue that the success of the model has been more apparent than real, and that its empirical implementation must be carefully reconsidered.

Article

Full-text available

Alcoholic Beverages and Cigarettes: Complements or Substitutes? A Pseudo Panel Approach

January 2011

Copyright 2011 by Aycan Koksal and Michael Wohlgenant. All rights reserved. Readers may make verbatim copies of this document for non-commercial purposes by any means, provided that this copyright notice appears on all such copies.

Article

Rational addiction and cultural goods: the case of the Italian theatregoer

April 2016 · Journal of Cultural Economics

This paper tests whether demand for theatre in Italy is consistent with the model of rational addiction presented in Becker and Murphy (J Polit Econ 96(4):675–700, 1988). Data from a novel 34-year panel on regional annual theatre attendance are used to estimate market demand. Four models are applied to investigate the demand function, and all of these also include per capita income and other ... [Show full abstract] control variables as regressors. The first two models are estimated to check whether theatregoers are myopically addicted to theatre. The results suggest that the theatre is an addictive good because past consumption (and prices) significantly raises the marginal utility of current consumption. The third model tests the rational addiction hypothesis, which assumes that future attendance also influences current attendance, whilst past and future prices influence current attendance only indirectly through their impact on past and future attendances. However, our most highly specified model, introducing past and future prices, demonstrates that Italian theatregoers are not myopic but fully rational as outlined in Becker and Murphy (1988). The results demonstrate that the rational addiction hypothesis is applicable not only to ‘‘harmful’’ addictions such as alcohol, cigarettes and drug consumption, but also to ‘‘beneficial’’ addictions, such as theatre attendance. This result has important policy implications because theatre is one of the most subsidised performing arts in Italy; if theatregoers are fully rational, policy makers can influence theatre attendance using alternative policy instruments (price and income), thereby reducing government expenditure on theatre subsidies.

Article

Full-text available

Testing for the Rational Addiction Hypothesis in Turkish Cigarette and Alcohol Consumption

December 2017

Raif Cergibozan

In this study, the relationship between simultaneous and intertemporal cigarette and alcohol consumption for Turkey is analyzed using GMM in the context of the Rational Addiction Model for the 1994-2016 period. Firstly, the demand of alcohol and that of cigarettes are estimated separately in the study. According to the results of the analysis, alcohol and cigarette demand are both in accordance ... [Show full abstract] with the Rational Addiction Model. The price elasticity of demand for both substances is negative. It is also found that the demand for alcohol is more elasticated than that for cigarettes. In addition, the cross-price elasticities are not statistically significant. Based on the results of this study, it is suggested that public policy combating these addictive substances be determined separately for these goods. Based on model results, the optimal strategy for increasing the income of the government should be to increase the price of cigarettes prices rather than of alcohol. Finally, no evidence is found indicating that the previous consumption of one of these substances influences the current consumption of the other good.