1
Youth and Inexperience: Dynamic Inconsistency Among Emerging Adults
______
A Thesis
Presented to
The Honors Tutorial College
Ohio University
______
In Partial Fulfillment of the Requirements for Graduation from the Honors Tutorial College with the degree of
Bachelor of Business Administration
______by
Brian J. Gibbons
April 2014
2
Discovery of Topic
The process of developing the idea for this thesis took a considerable amount of time. The beginnings of my exploration for a topic took place in a class I took with
Dr. Paxton, my thesis advisor, which was entitled The Economics of Altruism. In this class and the accompanying tutorial, we explored the economic reasoning behind why someone might choose to make an altruistic decision rather than a selfish one. This was my introduction to behavioral economics, which I found to be a very interesting subject. During that first tutorial, I explored some of the economics literature concerning an individual’s behavior and the economic rationale behind the making choices and, more specifically, choices that would appear to be economically or behaviorally irrational.
From there, I came across the topic of time inconsistent behavior and hyperbolic discounting which, as I will explain in detail in the following paragraphs, is when a person reverses his or her preferences because of an inconsistent value they place on time. This preference reversal is specific in that the person is impatient with their preferences now but anticipates being more patient in the future. Hyperbolic discounting is one of the most prevalent models in the relatively new field of behavioral economics. Understanding the fundamental concept of hyperbolic discounting, as well as the functional form of the model and the accompanying literature, took almost an entire year of tutorial instruction. During the time I was learning about the hyperbolic discounting model, I was also considering what way I could build upon the literature and develop a topic for my thesis. 3
The initial idea I came up with was to run a study that identified hyperbolic discounters through a survey sent to Ohio University alumni. Using the results from that survey, I would try to point out which people displayed inconsistent preferences and tendencies towards donating money to Ohio University. Once I identified these people, I planned on marketing a commitment device to the sample that would allow an individual to start donating a small amount of money in the present that then escalated to a target amount of donation per period sometime in the future. An initial survey was sent out to over 50,000 alumni through e-mail. From this survey, I gathered over 1,300 responses.
After spending a few months analyzing and cleaning the data, I found that although preferences towards giving was an interesting topic, there was a much stronger story that related hyperbolic discounting tendencies to age. I began reading into the literature surrounding the intertemporal (choices between payoffs at different periods in time) and behavioral characteristics of young adults and members of different generations. What I found was that although there was much written about the various characteristics of emerging adults and members of Gen Y, there was nothing written relating their distinct characteristics to their intertemporal preferences.
Given that I had found a distinct difference between the intertemporal preferences of adults age 21-30 and adults over 30, I felt I had a solid framework to begin my academic article which I later titled, “Youth and Inexperience: Dynamic Inconsistency
Among Emerging Adults.”
4
Explanation of Time Inconsistency
The basis of the topic I explored for my thesis was time inconsistency. Time inconsistency, in very simple terms, can be explained as when an individual prefers a specific future outcome in the present, but then reverses his or her preference towards that outcome and favors a different outcome when the future period becomes the present. These potential outcomes could involve anything ranging from an individual’s time, money, health, or even happiness or satisfaction. People who are time inconsistent are often unable to judge exactly how little value they are placing on the future, so when the future becomes the present they change their mind on plans they had previously made. Therefore they experience a preference reversal on an intertemporal choice.
For instance, imagine someone who is out of shape that one day decides that they would like to improve their physical fitness. This person likely already has a set of commitments that is taking up their time and preventing them from going to the gym to exercise. However, this person is placing a lot of value on getting in to shape, so they decide to make a plan to start working out in a month. There are essentially three categories people that would behave differently in this scenario. The first type of person is not time inconsistent, and is in fact consistent with the value he or she places on future time relative to the present. This person places a very high value on outcomes in future time periods and can actually make decisions for him or herself in the present that accurately match decisions he or she would make in the future. 5
Therefore if this person would commit to working out in a month, they would actually start doing so when that future time period became the present.
The second type of person is a time inconsistent, but we can say that they are sophisticated in their assessment of future plans. Even though they really want to get in to shape, the fact of the matter is they do not place enough value on their time one month from now to stick to their own commitment to getting in shape. What makes this type of person unique, however, is that he or she recognized from the beginning that they probably were not likely to stick to their own plans. A sophisticated time inconsistent person is able to also make decisions concerning future periods as if they were in the future period just like a time consistent person. The only difference is that they know they are going to face a preference reversal and will not end up sticking to the plans they make.
The third type of person is naïve in their time inconsistent preferences. This type of person also places a low value on time in the future, but unlike the other two is unable to make decisions as if they were in future periods. Naïve time inconsistent people are only able to make decisions through the lens of the present, and they never stick to the plans they make. In our example, this naïve intertemporal decision maker also wants to get in shape in a month. They judge that even though they are busy now, and likely will be just as busy in a month, they will still be able to make time to go to the gym in one month. They continue believing that they will stick to the plans they have made to get in to shape until one month in the future becomes the present period. 6
When this happens, they are completely surprised that they reverse their preferences and do not go to the gym because they are too busy.
In summary, a time consistent person is someone who always sticks to their future plans, a sophisticated time inconsistent person is someone who knows they won’t stick to their future plans because they do not value the future nearly as much as the present, and a naïve time inconsistent person is someone who is unaware they won’t stick to their future plans because of an inability to judge their own preferences in the future. Keep in mind, however, that this explanation is a simplified version of the concept and its accompanying economic rationale. This paper will also reference hyperbolic discounting, which is a specific type of time inconsistency. The concept of hyperbolic discounting mostly has to do with the steep discount curve that their intertemporal preferences follow. However, this will be explained in more detail in the following sections.
Personal Essay
My research in to the question of whether or not the emerging adults of
Generation Y had different time preferences relative to mature adults was an informative and challenging first glance into the field of behavioral economics research. There were a number of challenges that I faced in designing and writing the paper over the year and a half that I spent working on it. As I mentioned in the section about the discovery of my topic, my introduction to the topic of intertemporal inconsistency and hyperbolic discounting resulted from a class I took called the 7
Economics of Altruism. The first challenge I faced was grasping and comprehending the economic models that explain why individuals make inconsistent intertemporal decisions and why some individuals discount the future at a hyperbolic rate. This took a considerable amount of time, as it was a concept that I was completely unfamiliar with from the start. After I had built an idea of what I wanted my study to look like, I was fortunate enough to have been able to send a survey to a large group of Ohio
University alumni. This was a blessing as it provided a very strong data set with over a thousand observations. With so many points I then was faced with the task of finding and cleaning entries in the data set that were clearly answered insincerely or were obvious outliers.
Following a thorough examination of the data, my next challenge was to familiarize myself with the econometrics necessary to complete the analysis of the data. This was important because a publishable article needs to stand up to criticism of the mathematical models it presents. In order to do so, I read through a book on econometrics that covered the analysis of binomial variables and the concept of probit regressions, of which I was previously unfamiliar. One of the biggest benefits of choosing a topic regarding economics was there was a lot of statistical knowledge to be gained by reading books like the one assigned to me and reading through articles with dense economic theory. It challenged me to take charge of learning something that up to that point I would have felt uncomfortable trying to explain.
After completing the bulk of the work and editing my thesis down to a publishable note with the help of Dr. Paxton, I began the task of submitting the article 8
to a reputable journal of economics. The journal we chose to submit the note to first was Economics Letters, which typically accepts short notes around 5-12 pages. We tailored the article to the characteristics of Economics Letters and submitted the note.
After waiting for about 9 months, however, our article was rejected. Although initially disappointed, I learned that this was simply part of the process of publishing an academic article and it would be a bit far-fetched to have my article accepted by an extremely well regarded journal on my first attempt. In light of that, Dr. Paxton and I decided to look for another journal to publish in that might have previously featured work from authors discussing the economic behavior of emerging adults or Generation
Y. Our research led us to choose the Economic Inquiry, as it appeared that some articles featured in that journal aligned nicely with the findings of my thesis. After once again tailoring the article to the specific format of Economic Inquiry we submitted the article and are awaiting feedback from the journal.
Completing this thesis was a task that took a remarkable amount of thought and effort, as it was the first time I tried to complete anything this academically rigorous. However, from the experience I gained a tremendous deal of insight in to what it takes to explore an idea, complete through research of what has been written about similar topics, design and implement an experiment, craft an article, and ultimately meet the requirements of getting research published in an academic journal.
I believe I was able to make huge strides, both academically and professionally, during the time I spent working on my thesis. As challenges to the limits of my knowledge arose and setbacks occurred during the writing and publishing of the thesis, I was 9
forced to overcome these obstacles in order to complete the project. By doing so, I can definitively say that I have grown as a student and academic and my passion for completing research has grown, as well.
I particularly enjoyed trying to discover and interpret the story that was hidden behind the numbers in my dataset. This might have been the most difficult obstacle, and it was certainly the part of the thesis that took me the longest, but I really enjoyed being able to use my findings to create a snapshot of the behavior of my sample population. This is because that snapshot ultimately added, in one way or another, to the bigger picture of what is known about the topic of the behavior of current emerging and mature adults. The evolution this story was the result of many discussions between Dr. Paxton and I regarding the literature and how it relates our data. Finally, we were able come to the conclusion that we had found a trend that was significant, substantive, and fit nicely into what others have been saying about the topic. For me, this was the most rewarding part of the entire experience.
I would like to thank the support I received from my Director of Studies, Dr.
Raymond Frost, as he helped me along my way to discovering my passion for this subject, and I would also like to thank Dr. Julia Paxton, my thesis advisor, for her patient guidance and knowledgeable instruction, without which I would not have been able to accomplish this work.
10
YOUTH AND INEXPERIENCE: DYNAMIC INCONSISTENCY AMONG EMERGING ADULTS BRIAN J. GIBBONSa and JULIA PAXTONb aHonors Tutorial College, Ohio University bDepartment of Economics, Ohio University
Building on research on dynamic inconsistency and age related preferences; this paper introduces the concept that inconsistent intertemporal preferences are directly related to age. The findings from observations of a large sample of college graduates indicate that emerging adults are more likely to exhibit hyperbolic discounting behavior than adults over the age of 30. Among emerging adults respondents, lower incomes and more naïve self-assessments are determinants of hyperbolic discounting behavior. Keywords: Hyperbolic discounting, Emerging adults, Generation Y, Dynamic inconsistency, Intertemporal preferences “There is nothing constant in this world but inconsistency.” -- Jonathan Swift
1. Introduction
A large subset of behavioral economic research in the past has fallen under the broad heading of hyperbolic discounting. At its most basic level, hyperbolic discounting is the act of an individual placing inconsistent valuation on his or her future time following a steep hyperbolic discount curve, rather than an exponential or linear discount curve that place a more consistent value on time. This observation of time inconsistency is of particular relevance to economists, as it can explain why both humans and animals prefer to engage in actions that their future self would not choose in which to engage. The hyperbolic discounting model was introduced by Phelps and
Pollak (1968) as they described time preferences and the national savings rate. The concept of hyperbolic discounting, although disputed in its definition (Rasmusen 11
2008) and its functional form (Rubinstein 2003), has been applied to a multitude of topics.
The literature surrounding hyperbolic discounting is particularly broad in scope and application. Although the applications of hyperbolic discounting have been extensively researched, a gap in the literature exists in regards to intertemporal choice and hyperbolic discounting and their relation to age. The literature has established a connection between intertemporal choice and gender (Breman 2011), however an obvious gap exists that fails to relate hyperbolic discounting tendencies to age cohorts, specifically emerging adults. The present study fills this gap by establishing a relationship between intertemporal choice and the characteristics of different age cohorts.
2. Literature Review
2.1 Hyperbolic Discounting and Intertemporal Preferences
The literature surrounding hyperbolic discounting is particularly broad in scope and application. Early literature regarding intertemporal choice was largely centered on field studies to estimate individual discount rates. However, it was found that it is difficult to draw inferences between an individual’s savings behavior and discount rate. A large amount the experiments done regarding discount rate preferences are now conducted in the laboratory (Benzion, Rapoport, and Yagil 1989). Much debate has taken place over the interpretation of the theory behind hyperbolic discounting, with a different interpretation of the subject resulting from a number of dissenting opinions. 12
Rubinstein (2003) argues that theoretical economists may have abandoned the constant discount function in favor of the hyperbolic discounting too hastily. He showed in his study that by using experimental observations framed in a specific way, one can disprove the applicability of the function to anomalous behavior just as easily as it has been proven in other experiments.
Along the same lines as Rubinstein, Rasmusen (2008) attempts to dispel points that he believes to be common misconceptions about the hyperbolic discounting function. He enumerates six misconceptions that he finds to be the most common: “(1)
Hyperbolic discounting is not about the discount rate changing over time. A constant discount rate is not essential for time consistency, nor does a varying discount rate create time inconsistency. (2) Hyperbolic discounting does not, as commonly used, mean discounting using a hyperbolic function. (3) Hyperbolic discounting really isn't about the shape of the discount function anyway. (4) Hyperbolic discounting is not about someone being very impatient. (5) Hyperbolic discounting is not necessarily about lack of self-control, or irrationality. (6) Hyperbolic discounting does not depend delicately on the length of the time period”. His most important observation is that hyperbolic discounting needs to be modeled in terms of relativistic time rather than absolute time. This implies that when examining a discount rate for three years from now one would classify that point as 3 years from the present rather than simply using the absolute classification of 2017, for example. The shape of the discount curve is irrelevant according to Rasmusen; rather the thing that is most important is that as a 13
future period approaches the present an individual changes his or her consumption preferences in an inconsistent manner.
This conclusion is supported by the findings of Zauberman, Kim, Malkoc, and
Bettman (2009) who explain, from a psychological standpoint, that perception of duration of time can also have a causal effect on intertemporal choices and discount rates. Kim (2006) poses a counter argument to point (5), stating that low self-control is a function of hyperbolic time preferences and that past and present failures of self- control in sticking to a resolution can damage self-confidence. According to Kim, this previously observed psychological occurrence can be so severe that an individual will actually avoid engaging in resolutions they find important in order to avoid the negative outcome if they were to once again fail.
A significant relationship towards divergence from the linear probability- weighting model at the individual level has been established (Epper, Fehr-Duda, and
Bruhin 2011). In other words, individuals are particularly bad at sticking to a linear discount utility function because they are not able to properly predict the probability of effects on them from internal and environmental factors. These findings reinforce
Rasmusen’s conclusion that the shape of the hyperbolic discount curve is irrelevant and Zauberman et al.’s theory by demonstrating that individual probability distortions, rather than underlying individual preferences, have a significant effect on the discount rate. Simply put, an individual may engage in present oriented hyperbolic discounting because of perceived environmental uncertainty (i.e. economic, legal, political climate) rather than to satisfy a pure present-biased time preference (i.e. drug addicts). 14
Understanding that an individual’s discount utility function can be influenced by probability distortions leads to the interesting conclusion that policy could have a significant effect on such function.
Once again, the applicability of intertemporal choice and hyperbolic discounting to a multitude disciplines must be emphasized. If policy can have a substantive effect on discount rates, much could be done with legislation to change individual utility by increasing the present value of future outcomes. Hepburn,
Duncan, and Papachristodoulou (2010) show through their research with fisheries that policy interventions through commitment devices in an environmental context can increase total welfare over time. Time-based preference reversals often are present in environmental contexts. This is because in these markets present gains outweigh future societal repercussions because negative economic consequences from environmental damage are not immediately considered by or enforced upon those creating the negative welfare or externalities.
In regards to health, Scharff (2009) argues that policy established in the United
States has been unsuccessful at fighting the trend of obesity because it does not inherently require any sort of commitment mechanism. He identifies the behavior of people classified as obese to be consistent with hyperbolic discounting tendencies and suggests that a commitment device must be established to “transfer the present value of the long term costs of obesity closer to the point of consumption”. By doing so one would more likely be able to overcome the propensity to consume in the short run. It may be that the most important application of the findings of hyperbolic discounting 15
literature to disciplines outside of theoretical economics is to identify the contexts in which hyperbolic discounting is having a negative effect on welfare and institute commitment products to normalize the present value of long term discount rates.
A recent field experiment done by Ashraf, Karlan, and Yin (2006) is of particular relevance to the current study and will provide a large amount of the theoretical framework on which this study is built. In their study, Ashraf, Karlan, and
Yin employ theoretical survey questions similar to those posed in previous studies
(Benzion et. al, 1989) such as, “Would you prefer to receive P200 (Philippine pesos) guaranteed today, or P300 guaranteed in 1 month?” and “Would you prefer to receive
P200 guaranteed in 6 months, or P300 guaranteed in 7 months?” in order to measure time-preferences, as well as time preference reversals. Using this survey the researchers are able to measure three characteristics among their sample set: impatience, present-biased time inconsistency, and future-biased time inconsistency.
After identifying those among the sample who display time inconsistent preferences, a savings commitment device was marketed to a trial group. A significant effect for the adoption of the product was found among women who were classified as hyperbolic discounters. These women experienced a positive increase in saving, yet there was an indeterminable effect on welfare for the household.
2.2 Characteristics of Generation Y
Comparable experiences such as economic, political and social events among age groups form similar characteristics, values, and opinions among generational 16
cohorts (Petroulas, Brown, and Sundin 2010). Research in the fields of marketing and management has established details on the general characteristics of Generation Y, however relatively little research exists applying these concepts to the field of economics. Petroulas et. al (2010) stated that Generation Y has a lack of organizational loyalty and a short term goal focus. This categorization is reinforced by
Southard and Lewis (2004) who state that the cohort prefers instant gratification rather than the long-term investment of time and effort. Tulgan (2009) submits that
Generation Y harbors a sense of immediacy as a result of growing up with a never before seen amount of technology.
Another characteristic established through research is that Generation Y has been influenced by well-educated baby boomer parents (Martin and Tulgan 2002).
Twenge and Campbell (2008) find that members of Generation Y are higher in narcissism, higher in self-esteem, have less need for social approval, and have a higher external locus of control. Generation Y is known as the most educated and culturally diverse generation in history. This generation is exceptionally tolerant toward a variety of different diverse lifestyles.
A recent study shows that emerging adults in their 20’s and 30’s have 40 percent less real accrued wealth than their parents did at the same age, despite the fact that income has doubled in the past 25 years (Steuerle, et. al., 2013). With easy access to credit and college loans, recent college graduates are struggling to pay off debt and are delaying the purchase of homes and cars (Van Horn 2012). Psuedo-panel data has also shown that the members of Generation Y rely on the use of credit more heavily 17
and repay less of their balance than members of older generations (Jiang and Dunn
2013).
Taking in to consideration all of these characteristics, it is clear that Generation
Y has entered early adulthood in a world much different from that of previous generations. Experiences during their development have caused them to form distinct characteristics that could potentially have some effect on time preference. The academic observations of Gen Y from the literature concerning their mounting debt repayments, their short term goal focus, their higher external locus of control, and their experiences growing up with technology influenced the current study by providing a rationale behind why Generation Y might potentially have different intertemporal preferences than mature adults.
2.3 Age Related Time Preference
A growing wealth of literature has emerged regarding the characteristics of
“emerging adults”, or those individuals ages eighteen to twenty-five and beyond who do not classify themselves as an adult, including their risk preferences, financial behavior, and time preferences. The study of these characteristics is of particular relevance to those interested in the topic of hyperbolic discounting. Arnett (2001) identifies emerging adults as those ages 18-25 who have less stable financial situations, interpersonal relationships, living arrangements, cognitive and emotional development, and religious beliefs. He also states that the transition from adolescence to adulthood is a slow and incremental process for Americans, extending in many 18
cases into the late twenties. Arnett finds in his study that there is a significant difference in opinion between adolescents and those already in midlife adulthood over whether social norm violations were a part of the transition to adulthood. Adolescents and emerging adults viewed violations of norm compliance as an insignificant factor in the classification as an adult, whereas young to midlife adults viewed compliance with societal norms to be paramount to the classification as an adult.
In a recent study, a correlation between sensation-seeking scores and problematic financial behavior for emerging adults was established (Worthy,
Jonkman, and Blinn-Pike 2010). Robb (2011) presents similar findings showing that increasing financial knowledge has a direct impact on prudent debt management through credit card use in college students. The findings of these two studies are solidified through a national survey that controlled for demographic and financial backgrounds (Gutter and Copur 2011). This survey and the resulting study present the idea that time orientation towards the future and risk aversion can increase social and economic welfare, and once again found that higher levels of financial knowledge and risk aversion were positively related to financial welfare among emerging adults.
Consequently, overall welfare was higher as it is a function of financial welfare.
Developing the ideas established by the literature further, one could make the observation that current emerging adults and recently self-classified adults of
Generation Y have developed particularly poor financial habits relative to other generations because of increasing access to credit and underdeveloped financial literacy. 19
In so much as can be observed through the literature, recent research has failed to explain if these present-biased trends have had an effect on the discount utility of the cohort as a whole. Nelson and Barry (2005) establish that those who self-classify as “adults” engage in less risk-taking behaviors compared to their “emerging adult” peers who are self-classified using the same criteria. When taking this into account with the initial proposal, one finds that a very interesting and possibly significant tendency towards present-oriented intertemporal choice among the emerging adults and young adults of Generation Y.
3. Dynamic Inconsistency Theory
Traditional models of time consistent financial behavior assume an exponential discount function. The exponential discount factor is given as: