Published on July 1, 2022
Job titles and other details are as of the time of publication.
(The interview was conducted in Japanese and was thereafter translated into English.)

Decision theory: Using observable data to “visualize” the unobservable

My field of research, decision theory, examines how individuals make decisions under various conditions, such as when facing risk or making intertemporal choices. Consumer theory in microeconomics is also part of this field.

Elementary consumer theory typically assumes the existence of utility functions that represent individual preferences, but these utility functions cannot be directly observed. Instead, individual consumer behavior is analyzed to infer the utility functions assumed to be held by individuals and to reveal the principle of utility maximization, which aims to achieve the maximum level of satisfaction within budget constraints. The approach that uses observable data to test unobservable decision-making rules is known as revealed preference theory, or axiomatic decision theory.

Utility maximization is a basic theory studied in the first year of economics programs. In these courses, utility maximization is often assumed without much explanation or justification. In contrast, revealed preference theory aims to uncover the underlying preferences based on observed choice behavior or demand functions. It was included as a topic in textbooks when I was an undergraduate. However, it was rarely explained in details during lectures. Moreover, textbooks I read provided little discussion on the purpose or background of why revealed preference theory is studied, so I did not fully grasp its significance at that time. Yet, through a series of events, it has now become one of my primary research areas.

My research also explores intertemporal choices, which consider consumer choice behavior over time. Individual decision-making involves many factors that require a long-run perspective. Plans for the future are formulated to maximize lifetime utility through consumption, savings, and investment. The theory abstracting these dynamic choices is referred to as intertemporal choice theory. However, humans are not always capable of faithfully implementing their planned actions. Temptations and various other factors often lead individuals to deviate from their original plans, necessitating mechanisms to enforce self-discipline. I am particularly interested in incorporating these human factors (behavioral biases) into intertemporal choice theory.

Encountering examples of behavioral biases while struggling with the motivation behind abstract models

I decided to pursue research in decision theory during my graduate studies at the University of Rochester in the United States. My fascination with the subject began in microeconomics class taught by Professor Larry Epstein, one of the world’s leading researchers in decision theory. Under his guidance, I began my research with him as my advisor.

Before this, I had studied general equilibrium theory, a major branch of mathematical economics, during my undergraduate and graduate years in Japan. My passions for mathematics and logical reasoning had driven my interest in economics, and general equilibrium theory captivated me with its highly abstract framework for analyzing how market prices are determined without the specifics of the real economy. However, while writing my master’s thesis on general equilibrium theory and advancing to a doctoral program in Japan, I found it increasingly difficult to justify my focus on abstract models, not only to others but also to myself.

It was around that time when I learned about the concept of dynamic inconsistency from a fellow graduate student who had temporarily returned from the University of Rochester, where he studied that topic.

For instance, consider the choice between receiving one apple today or two apples tomorrow. Many people tend to choose the former. However, when faced with the choice of receiving one apple a year from now or two apples a year and a day from now, many people opt for the latter. In both cases, waiting one more day offers an additional apple, yet preferences shift depending on whether the choice is immediate or distant in the future. This tendency is known as present bias, or more generally, non-stationary preference. Moreover, assuming that the preferences of today and a year later are invariant, people often think that waiting for an extra day a year from now is the better deal. However, when that moment arrives, they may instead prefer to take the single apple immediately. This reversal in decision making, known as dynamic inconsistency, prevents individuals from following through their initial plans.

My first encounter with the concept of dynamic inconsistency deeply intrigued me. The theory described a real aspect of human nature with minimal abstraction, using a familiar example of intertemporal choices to present a compelling argument. That was how I decided to shift the focus of my research and follow in the footsteps of the graduate student mentioned earlier, ultimately enrolling at the University of Rochester.

Building a foundation as a researcher through overseas study

Studying abroad was one of the best decisions I have ever made. The university’s curriculum was well designed: first-year students were taught the fundamentals in depth, second-year students advanced to the frontlines of research through specialized courses taught by faculty members, and third-year students were paired with advisors for intensive research training. Advisors provided candid feedback on their students’ research ideas, which helped me gradually grasp the difference between compelling and less compelling research topics.

Decision theory, in particular, adopts a distinctive research method known as the axiomatic approach. I struggled at first to produce interesting results, and I remember Professor Epstein criticizing my research ideas every time I presented them. Although he was strict about research, his critiques were always accompanied by suggestions for relevant research papers, which helped to broaden my understanding. I still recall the immense joy I felt when he finally approved my research proposal after I had worked tirelessly to convince him of my ideas. I am deeply grateful to Professor Epstein for his patience in listening to my poor English and for his invariable guidance. I hold him in the highest regard. My graduate school years in the US laid the foundation for my research career and continue to drive my academic pursuit to this day.

During my time studying abroad, I had the opportunity to engage in discussions with students from diverse nationalities and backgrounds. These experiences gave me the confidence to adapt to and thrive in any environment. I encourage students at Hitotsubashi University to actively take advantage of exchange programs and other study abroad opportunities during their undergraduate years, rather than waiting until graduate school.

Visiting Boston to develop an intertemporal choice theory explaining various anomalies

Since April 2022, I have been engaged in international joint research at Boston University, focusing on developing an intertemporal choice theory. This theory incorporates a time aspect into consumer choice behavior, emphasizing the role of time discount rates, which are measured by discounting (or perceiving as lower) the value of outcomes based on subjectivity judgement.

Put simply, if an individual considers 10,000 yen a year from now to be equivalent to 9,000 yen today, the discount factor for one year is 9,000/10,000 = 0.9, the value referred to as the time discount factor. In economics, the standard approach to discounted utility involves this factor to calculate the present value of a future utility stream. However, experimental findings, known as the magnitude effect, indicate that discount rates vary depending on the magnitude of the money involved. For example, when considering a payment one year from now, the discount rate differs between receiving 100 yen and receiving 10,000 yen. In the former case, the discount rate tends to be higher, as the small amount may be perceived as insignificant. In the latter case, the discount rate tends to be lower, as the large amount is seen as desirable.

The standard discounted utility model, however, performs analyses under the strong assumption of constant discount rates, which fails to account for the magnitude effect, where discount rates depend on the amount of money involved. My research aims to address this limitation by introducing minimal revisions to the standard model, incorporating these elements to develop a framework capable of explaining various anomalies (behavioral patterns that cannot be explained by traditional economic models).

Balancing tractability and generality in model analysis

I have frequently mentioned the term “model” when discussing my research, as model analysis represents an important aspect of economics.

Real-world consumer behavior and social phenomena are so complex that it is impossible to include all elements when creating forecasts. Therefore, the first step is to simplify non-essential elements and model the essential ones using mathematics. This allows us to analyze what kinds of conclusions can be drawn. Once predictions and implications are derived, statistical and experimental data are gathered to assess the validity of these implications. This aspect of economics as an empirical science is becoming increasingly vital in economics. This is why, even though economics is often categorized as a “Humanity and Social Sciences” subject in Japan, the ability to think logically using mathematics is absolutely important.

However, economic models rely on the strong assumption of sorting essentials and non-essentials. The stronger the assumption and the higher the level of simplification, the more tractable the model becomes, allowing for sharper conclusions and insights into complex realities. At the same time, predictions based on such strong assumptions may conflict with statistical and experimental data, making it difficult to justify the model’s explanations.

To align the model with the data, the assumptions need to be weakened to generalize the model. However, overemphasizing generalization to explain the data can result in a model with an extremely high degree of freedom that can explain anything. Such a model will not be useful for forecasting as it does not lead to specific implications. The challenge lies in weakening assumptions just enough to avoid contradictions with robust empirical data while maintaining an appropriate level of generality without losing tractability. Creating a model that strikes this delicate balance is the most exciting part of my research.

Model analysis as an essential methodology for understanding society

Model analysis, a defining characteristic of economics, allows any type of social phenomena to become the subject of economic research. While economics is often associated with money, it is not limited to such matters. Recently, behavioral economics has gained significant attention as a field that applies realistic approaches to induce “good” behavior. It leverages various cognitive tendencies (behavioral biases and anomalies) related to risks, intertemporal choices, altruism, and other relevant factors. Such practical application of behavioral economics is known as nudging, which uses these insights to subtly guide decision making. Moreover, economic model analysis has been adopted in fields unrelated to money or markets. For example, the method for efficiently exchanging goods without monetary transfers is being applied to match organ donors with transplant recipients.

I highly encourage Hitotsubashi University students studying economics to master the methodology of model analysis. This approach provides a powerful framework for understanding policies and social systems, whether students pursue carrers as policymakers, analysts, marketers, or in other professions. Studying economics is also useful for fulfilling one’s role as a member of society, for instance, making informed decision when voting, with the vision of an ideal society in mind. My hope is that students will acquire economic literacy to actively and thoughtfully participate in society as informed citizens.