Elle Parslow EMPIRICAL ESSAYS ON CHOICES UNDER STRESS, PREFERENCES AND CONSUMPTION
ISBN 978-91-7731-143-0
DOCTORAL DISSERTATION IN ECONOMICS
STOCKHOLM SCHOOL OF ECONOMICS, SWEDEN 2019
Elle Parslow
EMPIRICAL ESSAYS ON CHOICES UNDER STRESS, PREFERENCES AND CONSUMPTION
EMPIRICAL ESSAYS ON CHOICES UNDER STRESS, PREFERENCES AND CONSUMPTION
This doctoral thesis consists of four self-contained chapters on different topics in Behavioural Economics.
“The Impact of Stress on Risky Choice: Preference Shifts or Noise” exam- ines whether risk preferences shift under stress, taking into account the possibility that more mistakes are made in decision-making under stress.
“The Effect of Stress on Consumers’ Estimation of Food Consumption Lev- els: Evidence from a Lab Experiment” examines if consumers overestimate their food consumption levels under stress, which could lead to food waste.
“The Digit Ratio (2D:4D) and Economic Preferences: No Robust Associa- tions in a Sample of 330 Women” examines the effect of the digit ratio on economic preferences with an aim to replicate previous studies.
“Pregnancy and Alcohol Purchases: Evidence from Scanner Data” examines how households adjust their alcohol purchases and fruit and vegetable pur- chases from before, to during and after pregnancy.
ELLE PARSLOW holds a Bachelor of Economics (Honours) and a Bachelor of Laws (Honours) from the University of Queensland. Her main re- search fields are Behavioural Economics, Experi- mental Economics and Applied Econometrics.
Elle Parslow EMPIRICAL ESSAYS ON CHOICES UNDER STRESS, PREFERENCES AND CONSUMPTION
ISBN 978-91-7731-143-0
DOCTORAL DISSERTATION IN ECONOMICS
STOCKHOLM SCHOOL OF ECONOMICS, SWEDEN 2019
Elle Parslow
EMPIRICAL ESSAYS ON CHOICES UNDER STRESS, PREFERENCES AND CONSUMPTION
EMPIRICAL ESSAYS ON CHOICES UNDER STRESS, PREFERENCES AND CONSUMPTION
This doctoral thesis consists of four self-contained chapters on different topics in Behavioural Economics.
“The Impact of Stress on Risky Choice: Preference Shifts or Noise” exam- ines whether risk preferences shift under stress, taking into account the possibility that more mistakes are made in decision-making under stress.
“The Effect of Stress on Consumers’ Estimation of Food Consumption Lev- els: Evidence from a Lab Experiment” examines if consumers overestimate their food consumption levels under stress, which could lead to food waste.
“The Digit Ratio (2D:4D) and Economic Preferences: No Robust Associa- tions in a Sample of 330 Women” examines the effect of the digit ratio on economic preferences with an aim to replicate previous studies.
“Pregnancy and Alcohol Purchases: Evidence from Scanner Data” examines how households adjust their alcohol purchases and fruit and vegetable pur- chases from before, to during and after pregnancy.
ELLE PARSLOW holds a Bachelor of Economics (Honours) and a Bachelor of Laws (Honours) from the University of Queensland. Her main re- search fields are Behavioural Economics, Experi- mental Economics and Applied Econometrics.
Empirical Essays on Choices under Stress, Preferences and Consumption
Elle Parslow
Akademisk avhandling
som för avläggande av ekonomie doktorsexamen vid Handelshögskolan i Stockholm
framläggs för offentlig granskning fredagen den 13 september 2019, kl 13.15,
sal Torsten, Handelshögskolan, Sveavägen 65, Stockholm
Empirical Essays on Choices under Stress,
Preferences and Consumption
Empirical Essays on Choices under Stress, Preferences and Consumption
Elle Parslow
Dissertation for the Degree of Doctor of Philosophy, Ph.D., in Economics
Stockholm School of Economics, 2019
Empirical Essays on Choices under Stress, Preferences and Consumption SSE and Elle Parslow, 2019c
ISBN 978-91-7731-143-0(printed) ISBN 978-91-7731-144-7(pdf)
This book was typeset by the author using LATEX.
Front cover illustration:
Tatiana Shepeleva/Shutterstock.comc Back cover photo:
Juliana Wiklund Printed by:
BrandFactory, G¨oteborg, 2019 Keywords:
Economic preferences, testosterone, 2D:4D, stress response, risk preferences, food choices, household purchases, pregnancy, alcohol consumption.
iii
To my family
Foreword
This volume is the result of a research project carried out at the Department of Economics at the Stockholm School of Economics(SSE).
This volume is submitted as a doctoral thesis at SSE. In keeping with the policies of SSE, the author has been entirely free to conduct and present her research in the manner of her choosing as an expression of her own ideas.
SSE is grateful for the financial support provided by the Jan Wallander and Tom Hedelius Foundation which has made it possible to carry out the project.
G¨oran Lindqvist Tore Ellingsen
Director of Research Professor and Head of the Stockholm School of Economics Department of Economics
Stockholm School of Economics
Acknowledgements
Firstly I would like to thank my main supervisor, Anna Dreber Almenberg, for her helpful advice, guidance, and pragmatic attitude. I am especially appre- ciative of her enthusiasm for my research ideas, no matter how outside the box they could be. Most important of all, she was able to guide me successfully to finishing this thesis. I always left our meetings more confident than when I had entered. Thank you also to my co-supervisor, Magnus Johannesson, for his excellent feedback, level of knowledge and attention to detail.
I would also like to thank Rickard Sandberg for providing advice on my re- search and studies. He also helped me to organise the research visit to Stanford and I am thankful for that experience.
During the PhD I have benefited greatly from the friendship and support of my cohort. I would also like to thank all of my current and former PhD col- leagues at SSE for making the days at the office much more enjoyable. I es- pecially would like to thank Domenico, Svante, Camilli, Thomas, Aljoscha, Nadiia and Elin. You each have always been extremely generous with your help and advice.
Two of the chapters in this thesis are based on a lab experiment run together with Julia Rose. I am grateful that we found shared research interests and I highly appreciate all her time and effort that went into planning and running the experiment together.
I would also like to thank the administrative staff at the Department of Eco- nomics Rasa, Malin, Ritva, Pia, Idha and Lyudmila for all of their kind help and assistance with many matters. I am also grateful to Anders who has pro- vided comments on parts of this thesis.
And finally I wish to thank the people who have supported me in immeasur- able ways. My parents have been a constant source of support, as well as my brother Jeremy, Nikki and the girls. I am especially grateful to my sambo, Andreas. When we met in that salsa class I am sure you had no idea that you
viii STRESS, PREFERENCES AND CONSUMPTION
would become a large contributor to the completion of this thesis. Lucky for me you have been there to support me in the toughest times of the PhD and for this I am extremely thankful.
Stockholm, June 11th, 2019 Elle Parslow
Contents
Introduction 1
1 The Impact of Stress on Risky Choice 5
1.1 Introduction . . . 6
1.2 Literature on Stress and Risky Choice . . . 9
1.3 Experimental Design and Procedures . . . 17
1.3.1 Stress Induction . . . 17
1.3.2 Measurement of Stress Response . . . 19
1.3.3 Risk Task . . . 21
1.3.4 Controls and Demographics . . . 24
1.3.5 Participants and Sessions . . . 25
1.4 Results . . . 28
1.4.1 Descriptive Statistics . . . 28
1.4.2 Stress Response . . . 29
1.4.3 Risk taking - Descriptive Results . . . 33
1.4.4 Risk taking - Structural Estimation . . . 38
1.5 Discussion . . . 43
1.A Appendix . . . 44
1.A.1 Experimental Procedure . . . 44
1.A.2 Experimental Instructions . . . 47
1.A.3 Supplementary Data and Analyses . . . 56
1.A.4 Robustness Checks . . . 62
Bibliography 74 2 The Effect of Stress on Consumption 85 2.1 Introduction . . . 86
2.2 Experimental Design and Procedures . . . 90 ix
x STRESS, PREFERENCES AND CONSUMPTION
2.2.1 Stress Induction . . . 91
2.2.2 Measurement of Stress Response . . . 93
2.2.3 Consumption Task . . . 94
2.2.4 Participants and Sessions . . . 97
2.3 Results . . . 100
2.3.1 Descriptive Statistics . . . 101
2.3.2 Stress Response . . . 101
2.3.3 Consumption Task Results . . . 105
2.3.4 Robustness Checks . . . 108
2.4 Discussion . . . 110
2.A Appendix . . . 112
2.A.1 Experimental Procedure . . . 112
2.A.2 Experimental Instructions . . . 115
2.A.3 Supplementary Data and Analyses . . . 123
2.A.4 Additional Robustness Checks . . . 125
Bibliography 128 3 The Digit Ratio and Economic Preferences 135 3.1 Introduction . . . 136
3.1.1 Dictator Game Giving . . . 139
3.1.2 Risk Taking . . . 144
3.1.3 Competitiveness . . . 151
3.2 Method . . . 153
3.2.1 Experimental Procedures and Design . . . 153
3.2.2 Collection of 2D:4D . . . 155
3.3 Results . . . 155
3.4 Discussion . . . 161
3.A Appendix . . . 163
Bibliography 165 4 Pregnancy and Alcohol Purchases 173 4.1 Introduction . . . 174
4.2 Data . . . 178
4.2.1 Identification of Households with a Pregnant House- hold Member . . . 179
CONTENTS xi
4.2.2 Sample of Pregnant Households . . . 181
4.2.3 Outcome Variables . . . 183
4.3 Empirical Strategy . . . 185
4.3.1 Event Study Approach . . . 186
4.3.2 Primary Quasi-Experiment . . . 186
4.4 Results . . . 189
4.4.1 Event Study . . . 189
4.4.2 Difference-in-Differences Estimation . . . 192
4.5 Discussion . . . 198
4.A Appendix . . . 200
4.A.1 Identification of Pregnancy . . . 200
4.A.2 Event Study . . . 201 4.A.3 Quantitative Result on Changes in Alcohol Purchases 204
Bibliography 206
Introduction
This thesis consists of four self-contained chapters on different topics in be- havioural economics. The first and second chapters examine choices under stress using data from a lab experiment with 194 participants. The first exam- ines whether risk preferences shift under stress taking into account the possi- bility that more mistakes could be made in decision-making under stress. The second examines if consumers estimate higher quantities of food consumption under stress. The third chapter examines the effect of the digit ratio(2D:4D) on economic preferences, with an aim to replicate previous studies, and uses data from a lab experiment with 330 participants. The final chapter examines how households adjust their alcohol purchases and fruit and vegetable purchases from before, to during and after pregnancy, using a consumer panel scanner dataset.
A short summary of each chapter follows.
The Impact of Stress on Risky Choice: Preference Shifts or Noise (with J. Rose)
We analyse the impact of stress on risky choice in a large-scale between-subjects design with 194 participants. In particular, the main contribution of our work is that we are able to disentangle errors in decision making from an actual shift in preferences between our experimental treatment and control conditions.
Acute stress is induced using the Trier Social Stress Test for Groups, and risk preferences are elicited using the multiple price list method. By additionally controlling for cognitive reflection, we analyse whether our results are driven by different levels of cognitive ability as the driver for observed noise within and between conditions. We find that, controlling for noise, risk attitudes are stable between conditions. The overall number of decision errors is low on average and does not increase under stress. Additionally, we find statistically
1
2 STRESS, PREFERENCES AND CONSUMPTION
significant evidence for lower cognitive abilities being correlated with more noise in decision-making in general, however, independent of condition.
The Effect of Stress on Consumers’ Estimation of Food Consumption Levels: Evidence from a Lab Experiment
In a pre-registered experiment, I investigate empirically whether stress affects consumers’ estimation of their future consumption levels of common grocery items. I hypothesise that stress may lead to higher estimated quantities of food consumption, and this may be a mechanism behind the problem of food waste.
I use a between-subjects design with 194 participants, where subjects are ran- domised to either stress treatment or a control group. Acute stress is induced using the Trier Social Stress Test for Groups, and consumers’ estimation of their food consumption levels is measured by a binary quantity choice task. I am unable to reject the null hypothesis that acute stress does not have a signifi- cant effect on average on consumers’ estimation of their consumption levels. I also am unable to reject the null hypotheses of no gender difference in the treat- ment effect, and no effect of hunger on estimated consumption quantities. By providing initial empirical evidence that consumers do not overestimate their food consumption under stress, this study contributes to increased understand- ing of the role of stress in consumer choice behaviour.
The Digit Ratio(2D:4D) and Economic Preferences: No Robust Associations in a Sample of 330 Women
(with E. Ranehill, N. Zethraeus, L. Blomberg, B. von Schoultz, A. Lind´en Hirschberg, M. Johannesson and A. Dreber)
Many studies report on the association between 2D:4D, a putative marker for prenatal testosterone exposure, and economic preferences. However, most of these studies have limited sample sizes and test multiple hypotheses (without preregistration). In this study we mainly replicate the common specifications found in the literature for the association between the 2D:4D ratio and risk taking, the willingness to compete, and dictator game giving separately. In a sample of 330 women we find no robust associations between any of these eco- nomic preferences and 2D:4D. We find no evidence of an effect for sixteen of the eighteen total regressions we run. The two regression specifications which are significant have not previously been reported and the associations are not
CONTENTS 3 in the expected direction, and therefore they are unlikely to represent a real effect.
Pregnancy and Alcohol Purchases: Evidence from Scanner Data (with A. Janssen)
We analyse household-level changes in consumption in response to pregnancy.
Using scanner data, we identify households with a pregnant household mem- ber. We use an event study as well as a difference-in-differences approach to estimate the effect during and after pregnancy on (1) alcohol purchases and (2) relative expenditure on fruit and vegetables. We find that during preg- nancy, households reduce their alcohol purchases by 22% on average. After pregnancy, purchases of alcohol are 27% lower than in the months of pre- pregnancy. In contrast, the relative expenditure on fruit and vegetables does not increase during pregnancy but decreases post-pregnancy by 19%. One pos- sible interpretation of the results is that a reduction in alcohol consumption during pregnancy reduces consumption in the long term by a change of gen- eral consumption habit. Although, we do not rule out other explanations such as reduced alcohol intake as a result of breastfeeding, or that having children increases incentives for healthy behaviour.