Stanford Research Communication Program
  Home   Researchers Professionals  About
 
Archive by Major Area

Engineering
Humanities

Social Science

Natural Science

Archive by Year

Fall 1999 - Spring 2000

Fall 2000 - Summer 2001

Fall 2001 - Spring 2002

Fall 2002 - Summer 2003


 

 

 


A Step Towards Better Retirement

Aparna Gupta
Decision Sciences Department
Rensselaer Polytechnic Institute
agupta@sccm.stanford.edu
June 2001

We all have the potential to obtain a high quality of life in our retirement years; we just need to spend some time and effort planning for it. We need to make conscious, deliberate decisions and then act on them. My research involves developing a tool that will help people better plan for and make more informed decisions regarding their retirement.

Irrespective of our professions, we all look forward to the time in our lives when we can afford to be free of the daily hustle of life and spend time doing things that we find pleasurable. For most of us, that time is when we retire from our primary occupations. In order for our dreams for retirement to be realized, however, we need to be strong and secure, both financially and health-wise, when the time comes. This means we not only need to begin saving well in advance, but that we also need to come up with a strategy to guarantee a good support mechanism for health-related needs. The natural question that arises is how to best allocate financial resources and invest wisely in order to assure the best possible outcome at retirement. Clearly, though, we cannot predict the future in terms of the returns we earn on investments or how our needs change with time depending on our health-conditions and other changing needs. In my research, I develop a framework in which retirement objectives are formally formulated with the intent to obtain the “best” decisions to meet these objectives.

Many problems that arise in making good decisions can be framed as mathematical programming problems, in which each decision is treated as a variable (assuming it can be numerically expressed). The decisions, as would be expected, affect the final outcome that we as planners care about. The mathematical programming approach also models the planner’s preference for different final outcomes numerically, assigning a higher number to the preferred outcome than to the less preferred outcome. As a simple example of how preferences are measured, we could play the following game Two coins will be flipped; when both are heads the planner gets $2, if one is heads the planner gets $1, and if neither are heads the planner gets nothing. The planner will enter the game by paying $1. I could then ask the planner how happy the outcome of “two heads,” “one head,” and “no heads” will make them, thereby forming a depiction of their preferences. A similar example from the retirement planning perspective would be the following Assume that the planner saves $1,000 today and invests the money in IBM stock. One year from now one share of IBM stock could be worth a range of different values, and as a result the planner’s investment will have a range of different values. The question is how will the planner view the different levels their investment may take a year from now. In reality, when making investment decisions there are many more investment choices the planner faces.

In addition to an investment strategy, the planner also has a number of other decisions to make, such as deciding how much to contribute to retirement savings on a regular basis, what type of health insurance (long-term care, supplemental Medicare, etc.) to purchase as they get older so that their health-related needs are best met at retirement, and how to spend the savings after retirement so that, while they are able to do all the things they planned for, they do not deplete their savings too quickly. To develop a comprehensive decision-support tool, all these different decisions need to be modeled in the problem.

An important additional issue to be considered in relation to this decision-making problem is a cognitive one. The planner evaluates the outcomes of the decisions they make based on their cognitive capacity for understanding what the future will be like. Research done through psychological experiments indicates that people are not always rational and consistent in evaluating and comparing different outcomes. Our capacity to think critically and analytically, while one of nature’s marvels, is occasionally hampered by the circumstances under which we make these judgments. For example, it so happens that when the same set of options is presented to a person in two different ways, the response in preferring one outcome to the other systematically differs. This is called the “framing effect” in behavioral psychology. For instance, presenting a patient with a treatment option in terms of its effectiveness results in responses that are different from when the same option is presented in terms of its failure. The question that arises is whether there is a way that the planner may overcome any cognitive hindrances in making retirement planning related decisions. I also attempt to address such issues in this research.

The framework for using this tool is as follows The planner, after a reasonable amount of interaction with the tool, comes up with a numerical representation of their value-system, for which the different potential outcomes of their decisions can be compared. For example, let’s assume that the result of the preliminary interaction with the tool reveals that the planner cares for the total amount of wealth saved at the time they retire. This wealth, denoted by “W,” is used to provide their living expenses after retirement. The quantity “W” is determined in terms of savings, investment decisions, insurance purchase decisions, etc. Through the preliminary interaction it is also inferred that the numerical representation of their value-system is best modeled as E[ln(W)], where E[.] is the expectation operator. This forms the objective of the mathematical program. Using the mathematical programming tool, it is possible to obtain the corresponding “best” decisions. If upon analysis of these decisions the planner (the person preparing for the future) observes some idiosyncrasies in the decisions recommended, the tool will provide the capability to make a more detailed analysis of the decisions and the underlying rules for comparing outcomes. This will give the planner an improved insight into their preferences. On the other hand, if the decisions recommended look reasonable, the planner will have solved a very important problem, not only by “soul-searching” and identifying their retirement objectives, but also by making good decisions that will help them meet their retirement goals; hopefully much better than they would otherwise have been met without undergoing a rigorous decision-making process.